Thesis Nadine Offermans by Universitaire Pers

Occupational asbestos exposure and cancer risk A population-based approach using job-exposure matrices

Nadine Offermans

Copyright Nadine Offermans, Maastricht 2015 No part of this book may be reproduced or transmitted in any form or by any means, without prior permission in writing by the author, or when appropriate, by the publishers of the publications. Layout: Tiny Wouters Cover: Ellen Offermans Production: Datawyse – Universitaire Pers Maastricht ISBN: 978‐94‐6159‐447‐1

Occupational asbestos exposure and cancer risk A population‐based approach using job‐exposure matrices

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Universiteit Maastricht, op gezag van de Rector Magnificus, Prof. Dr. L.L.G. Soete volgens het besluit van het College van Decanen, in het openbaar te verdedigen op woensdag 10 juni 2015 om 14.00 uur door Nadine Sibilla Maria Offermans

UNIVERSITAIRE

PERS MAASTRICHT

Promotor Prof. dr. P.A. van den Brandt

Copromotores

Dr. R.A. Bausch‐Goldbohm (TNO, Leiden) Dr. R.C.H. Vermeulen (Institute for Risk Assessment Sciences, UU; Julius Center for Health Sciences and Primary Care, UMC Utrecht)

Beoordelingscommissie Prof. dr. F.J. van Schooten (chairman) Dr. A. Dingemans Prof. dr. H.I. Grabsch Prof. dr. D. Heederik (Institute for Risk Assessment Sciences, UU) Dr. T. Pal (Coronel Instituut voor Arbeid en Gezondheid, AMC) The studies in this thesis were supported by the Netherlands Organization for Health Research (ZonMw) under grant 50‐50‐500‐98‐6153.

Contents Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6

General introduction

Comparison of expert and job‐exposure matrix‐based retrospective exposure assessment of occupational carcinogens in the Netherlands Cohort Study

Occupational asbestos exposure and risk of pleural mesothelioma, lung cancer, and laryngeal cancer in the prospective Netherlands Cohort Study

Occupational asbestos exposure and risk of esophageal, gastric and colorectal cancer in the prospective Netherlands Cohort Study

Occupational asbestos exposure and risk of oral cavity and pharyngeal cancer in the prospective Netherlands Cohort Study

General discussion

107

Summary

135

Valorisation

141

Dankwoord

151

Curriculum Vitae

157

List of publications

161

Chapter 1

General introduction

ASBESTOS

The term asbestos is applied to several silicate minerals when they occur in a fibrous form known as ‘asbestiform’, characterized by bundles of thin, long, and separable fibers.1,2 Six minerals (or fiber types) are denoted as asbestos: the fibrous serpentine mineral chrysotile and the five fibrous amphibole minerals actinolite, amosite, anthophyllite, crocidolite and tremolite.2 All fiber types are composed of silicon (Si) and oxygen (O), and depending on the fiber type they also contain hydrogen (H) and one or more cations (primarily magnesium (Mg), iron (Fe), calcium (Ca), and sodium (Na)). In theory, chrysotile has a chemical composition of Mg3Si2O5(OH)4, but variations in chemical composition exist. It occurs as a curled sheet silicate, which wraps around itself in a spiral, forming a hollow tubular fiber. The amphibole forms of asbestos consist of two chains based on Si4O11 units, linked by a band of cations whose ratios determine the fiber type.3 Evidence has been found that asbestos use dates back to 2500 B.C., when it was used for strengthening clay pottery.4 Asbestos minerals possess a number of properties useful in commercial applications: heat stability, thermal and electrical insulation, wear and friction characteristics, tensile strength, the ability to be woven, and resistance to chemical and biological degradation.1,2 As the modern industrial era gained momentum in the 1880s, it were these properties that resulted in a burgeoning of the applications of asbestos.4

1.1 Recognition of toxicity and carcinogenicity and consequent legislation It was gradually recognized that using asbestos is not only advantageous but is also associated with the occurrence of serious health consequences, which resulted in increasingly strict measures to reduce the exposure to asbestos. A number of adverse health outcomes is now established as causally associated with exposure to asbestos, regardless of fiber type: asbestosis, mesothelioma (i.e., cancers of the pleura and peritoneum), lung, laryngeal, and ovarian cancer.5,6 Adverse health outcomes related to the use of asbestos were not recognized simultaneously in all countries. Asbestosis (progressive pulmonary fibrosis) was already recognized as an occupational disease in The United Kingdom in 1931,7 but it was not until the 1964 paper by Selikoff and colleagues8 that the international scientific community was convinced of the carcinogenicity of asbestos for lung and mesothelial tissue.9 In this paper, large increases in the risk of lung cancer and mesothelioma were observed in a population of 632 insulation workers occupationally exposed to asbestos in New York and New Jersey. Only in 2009, asbestos has been causally related to the occurrence of laryngeal and ovarian cancer.5 Because of the known toxicity, asbestos

Chapter 1

has been banned in more than 50 countries nowadays, including all European Union member states.10 In general, these bans are fairly recent, with most bans in Europe being enforced in the 1990s. As there is no worldwide ban on asbestos, numerous countries still use, import and export asbestos and asbestos‐containing products. These are almost all countries in Asia, Eastern Europe, Latin America, and Africa.11 In the Netherlands, asbestosis was acknowledged as an occupational disease in 1949, while consensus on the causal relation between asbestos exposure and lung cancer and mesothelioma had to wait until the work of Zielhuis in 196812 and Stumphius in 1969,13 respectively.9 Following these publications, specific guidelines were proposed for working with asbestos in guidance note P‐116 of the Labor Inspectorate (1971). Although these guidelines were not enforced by any specific legislation, they certainly increased the pressure on industry to control asbestos exposure. Consequently, in several branches of industry asbestos control measures to reduce exposure were introduced around 1973‐1976.14 Measures included local exhaust ventilation, wet working techniques, closed machines, automated bag cutters and sound housekeeping procedures.15 The first legislation was enforced in 1978 with a ban on crocidolite and a threshold limit value of 2 fibers per ml (f/ml) for all other asbestos fibers. During the 1980s, exposure levels decreased again, probably due to increasingly stringent legal standards and reductions in the threshold limit values in 1988 (1 f/ml) and 1993 (0.3 f/ml). In 1994, a total ban on the use of asbestos products was enforced.14

1.2 Asbestos production and use Natural deposits of asbestos are found all over the world. Chrysotile asbestos has been commercially exploited in many countries, among others in Russia, Canada (Quebec and Ontario), South Africa, the United States (Vermont and California), Australia, Zimbabwe, and Italy.16 Amphibole asbestos deposits have been commercially exploited particularly in South Africa, Australia, and Finland.1 Overall, the largest and commercially most important deposits are the ones in Russia, Canada and South‐Africa. In the 20th century, these countries were responsible for more than 80% of the world production.17 Cumulative world production from 1900 through 2003 amounts roughly to 181 million tons,4 with a peak in industrial use of asbestos in the late 1960s and 1970s when more than 3000 industrial applications or products were listed.3 In 2012, the annual world production of asbestos was estimated to be more than 2 million tons with Russia being the largest producer of asbestos followed by China, Brazil, Kazakhstan, and Canada. Since 1995, the leading asbestos‐using nations have been Russia and China, followed by Brazil, Thailand, Kazakhstan, India, Ukraine, and Iran.18 In the Netherlands, the asbestos industry was virtually non‐existent in the years before 1945. While there are no asbestos mines in the Netherlands,16 the manufacture of asbestos products increased sharply after 1945; the most important products being

General introduction

asbestos cement, insulation material, friction materials and brake linings, boards and floor coverings.9,19 The asbestos industry in the Netherlands was quite substantial as about eight million tons of asbestos‐containing products were manufactured and used in the 20th century,20 with a peak in the annual import of asbestos in the mid‐1970s.21 Of the imported asbestos, 2.4% was processed before World War II, 15.9% in the 1950s, 25.4% in the 1960s, 43.6% in the 1970s, 10.7% in the 1980s, and 2.0% in the 1990s. Almost 54% of all asbestos was used in asbestos cement products.22 Other leading asbestos‐using industries were the insulation industry and the ship building and maintenance industry.20 Workers also frequently exposed to asbestos were engineering workers, maintenance workers, firemen, furnace workers, carpenters, construction workers, glassblowers, power station operators, and workers in the primary metal branch and metal product branch, petrol refining plants, chemical industry, and electronic equipment industry.9 In 2010, the total population ever occupationally exposed to asbestos was estimated to amount to at least 330,000 employees.20 Occupational asbestos exposure can still take place when homes and other buildings are demolished, when soil purification activities are undertaken, and when ships, drilling platforms and other machines with asbestos insulation are repaired or pulled down. In addition, non‐occupational exposure may also have occurred ‐ and may still occur ‐ in the context of building renovations and if asbestos is present in the environment,20 where is was used to harden dirt tracks, yards, and driveways.23 Furthermore, domestic asbestos exposure occurred when women washed the clothes of their occupationally exposed husbands.24

1.3 Asbestos burden Worldwide, the annual number of asbestos‐related cancer deaths in workers was recently estimated to be 100,000–140,000,11 and until a worldwide ban on asbestos is achieved, the asbestos cancer pandemic may take as many as 10 million lives.25 As asbestos is still being produced and used on a large scale, these figures may underestimate the true impact of this pandemic.11 In the Netherlands, a country with one of the highest mesothelioma‐related mortality rates, at least 13,000 mesothelioma deaths and another 13,000 deaths due to asbestos‐ related lung cancer are expected in the period 2000‐2028, with a peak of approximately 500 mesothelioma cases annually in 2017.26 The reason why this peak is expected in the years to come is attributable to the facts that the risk of occupational asbestos exposure was highest in the mid‐seventies and that asbestos has an average latency period between first asbestos exposure and diagnosis of mesothelioma of approximately 45 years.27

Chapter 1

ASBESTOS CARCINOGENICITY

2.1 Evaluation of the evidence on asbestos‐related cancers In the sections below, we present a short summary of the evidence, both epidemiological and experimental, for those cancers definitely or potentially linked to asbestos exposure: i. Respiratory cancers (i.e., pleural mesothelioma, lung and laryngeal cancer) ii. Gastrointestinal cancers (i.e., esophageal, gastric and colorectal cancer) iii. Oral cavity and pharyngeal cancer iv. Ovarian cancer. In the evaluation of carcinogenicity, we follow the classification by the International Agency for Research on Cancer (IARC). The overall evaluation of human carcinogenicity by IARC is based on epidemiological, animal experimental and other relevant evidence on genotoxicity, mutagenicity, metabolism or mechanisms. Until now, there have been a number of evaluations on asbestos carcinogenicity (IARC Monographs) with the most recent evaluation dating from 2012. 2.1.1 Respiratory cancers Epidemiological evidence Mesothelioma and lung cancer As mesothelioma and lung cancer share ‐to some extent‐ the same body of evidence, they are considered together here. Both cancers have been studied extensively and the range in relative risks reported in the literature is wide. In 1955, Sir Richard Doll was the first to demonstrate an excess risk of lung cancer among asbestos textile workers.28 In 1960, a study among crocidolite miners conducted by Wagner and colleagues was the first in which a possible association between occupational asbestos exposure and mesothelioma was revealed.29 Since then, excesses of mesothelioma and lung cancer have been observed in a large number of cohort and case‐control studies and in a variety of different industries.6 Two authoritative meta‐analyses are worth mentioning, the paper by Hodgson and Darnton (2000) and the one by Berman and Crump (2008).30‐32 Both studies estimated the average excess risk of mesothelioma and lung cancer per unit increase (in fiber‐years) of asbestos exposure (expressed as the potency or the so‐called KM and KL value, for mesothelioma and lung cancer, respectively). Despite the substantial heterogeneity in risk estimates between the studies included, both meta‐analyses show an order‐of‐magnitude agreement in excess risk per fiber‐ year. Most of the studies included in these meta‐analyses are not very recent and involved merely workers highly exposed to asbestos. Currently, at least in developed countries with an asbestos ban, most individuals are no longer exposed to these high levels.

General introduction

Therefore, the focus has shifted towards estimating excess risk accurately for levels encountered at the lower end of the exposure distribution. For mesothelioma, there is still controversy whether low levels of exposure are carcinogenic. Some studies found evidence of a threshold level,33,34 though others could not subscribe to the existence of a threshold level for asbestos‐related mesothelioma.35,36 For lung cancer, there is currently the notion that relative risks at the lower end of the exposure distribution might be higher than predicted by downward linear extrapolation from highly exposed occupational cohorts.37‐39 Risk of mesothelioma may differ for the pleura and peritoneum in that pleural mesotheliomas may be induced by lower levels of exposure than peritoneal mesotheliomas.6 For lung cancer, previous studies showed asbestos to be associated preferentially with lung adenocarcinoma,40,41 though this association has not been reported consistently throughout the literature.42,43 Laryngeal cancer Laryngeal cancer has not been studied as extensively in relation to asbestos as mesothelioma and lung cancer.5 A meta‐analysis of 35 cohort and 18 case‐control studies conducted by the Institute of Medicine (IOM) reported a significantly increased summary relative risk (RR) of 1.4 for ever versus never exposed to asbestos. Furthermore, there is some evidence from both cohort and case‐control studies that risk follows an exposure‐response pattern; the aggregate RR in the most highly exposed subjects compared to those never exposed to asbestos ranged from 1.38 to 2.57.1 Few studies investigated the risk of subtypes of laryngeal cancer and those that did were mostly the ones that did not find an association between asbestos and laryngeal cancer. These studies showed that risk was more or less comparable for glottis and supraglottis cancer, with values of around 144‐46 as well as non‐significantly increased risk estimates,47,48 though one study revealed a threefold higher risk for supraglottis than glottis cancer.49 Other evidence Asbestos fibers have been detected in the pleura and lung samples have confirmed that asbestos fibers accumulate in lung tissue.6 Only a few studies investigated if asbestos fibers deposit in the larynx and although two of them were confirmative, contamination from other tissues cannot be excluded.1 In numerous animal experiments, mainly in rodents, malignant mesothelioma and lung tumors have been observed after exposure to asbestos.6 Studies in rats and Syrian hamsters found that asbestos inhalation, at levels sufficient to cause mesothelioma in both species and lung cancer in rats, did not induce laryngeal cancer.1

Chapter 1

Evaluation of evidence by IARC According to IARC, a large body of evidence has been accumulated that asbestos causes mesothelioma and lung cancer. In 2009, the relation between asbestos and laryngeal cancer has been classified as causal.5 2.1.2 Gastrointestinal cancers Epidemiological evidence Esophageal cancer Several occupational cohort studies have observed elevated cancer rates after asbestos exposure, and have suggested that risk might be exposure‐dependent.6 The IOM meta‐ analysis, including 25 occupational cohorts, reported a summary relative risk (RR) of 0.99(0.79‐1.27) for any versus no asbestos exposure.1 In contrast, a meta‐analysis examining studies with heavier asbestos exposure reported an elevated summary standardized mortality ratio (SMR) of 2.38(1.45‐3.68) in exposed workers.50 As esophageal cancer is a relatively rare cancer, only few studies looked at subtypes. For both histological types, esophageal squamous cell carcinoma (ESCC, occurring largely in the upper two‐thirds of the esophagus) and esophageal adenocarcinoma (EAC, occurring in the lower one‐third of the esophagus), results are mixed, with a suggestion of an exposure‐response relation only for EAC.51‐54 Gastric cancer Increased risks and exposure‐response relations have been observed in occupational cohorts with higher asbestos exposure.1,6,55,56 The IOM meta‐analysis including 42 occupational cohorts reported a summary RR of 1.17(1.07‐1.28) for any versus no asbestos exposure, while results for case‐control studies were inconsistent.1 Regarding gastric cancer subtypes, no increased risk of gastric cardia adenocarcinoma or gastric non‐cardia adenocarcinoma has been observed.53,57,58 Colorectal cancer Occupational cohorts fairly consistently show increased colorectal cancer risks and exposure‐response relations after asbestos exposure,6 while results from case‐control studies are less consistent.1 The IOM meta‐analysis including 41 occupational cohorts reported a summary RR of 1.15(1.02‐1.31) for any versus no asbestos exposure, with the largest excesses of colorectal cancer being observed among the earliest North American and British insulation workers.1 For colorectal cancer subtypes, there is some suggestion in the literature that the association with asbestos might be stronger for colon than for rectal cancer.6 Other evidence Despite technical difficulties, there is some evidence that asbestos fibers disseminate to and persist in the gastrointestinal tract. The evidence from experimental studies is, however, rather limited and does not indicate that asbestos is carcinogenic to

General introduction

esophageal or gastric tissue. For the colorectum, there is some evidence for the induction of benign colon polyps in rats.1 Evaluation of evidence by IARC For gastric and colorectal cancer, the evidence is classified as limited, though the working group involved at IARC was evenly divided as to whether the evidence was strong enough to justify classification as sufficient for colorectal cancer.6 For esophageal cancer, no IARC classification existed and we followed the recent meta‐ analysis conducted by the Institute of Medicine (IOM) on which the IARC evaluation of 2012 draws. IOM considered the evidence to be inadequate.1 2.1.3 Oral cavity and pharyngeal cancer Epidemiological evidence Results of previous cohort studies on the association between asbestos and oral cavity and pharyngeal cancer (OCPC) are rather consistent and show modestly increased risks with a meta‐RR of 1.44(1.04‐2.00), while case‐control studies are rather limited in number and show inconsistent results.1 Data on exposure‐response patterns from both types of study are limited and tend towards lower risks for the more extreme exposures.1 A Finnish study found no exposure‐response association either,59 but a meta‐analysis stratifying results on exposure circumstance showed a RR of 1.63(1.27‐ 2.09) for asbestos miners and millers with the highest exposures.60 There are only a few studies that have investigated the asbestos‐related risk of oral cavity cancer (OCC) separately. These studies showed non‐significantly increased risks61,62 as well as decreased risks,51,63 with a recent meta‐analysis revealing relative risk (RR) estimates of 1.13(0.81‐1.57) and 1.15(0.84‐1.57) for low and high exposure, respectively, based on five studies.60 For pharyngeal cancer (PhC), risk estimates of around one51,62 as well as increased risks have been reported after asbestos exposure in both overall pharyngeal cancer63 and hypopharyngeal cancer.44,61 A recent meta‐analysis showed increased risks of PhC without evidence of an exposure‐response relation (RRs of 1.26(0.96‐1.66) and 1.27(0.98‐1.66) for low and high exposure, respectively).60 Other evidence There are no reports on recovery of asbestos fibers from the oral cavity or pharynx; the absence of such data neither supports nor disproves the possibility that fibers accumulate at these sites. No increase in tumors has been observed in animals chronically exposed to asbestos by inhalation.1 Evaluation of evidence by IARC IARC classified the evidence for pharyngeal cancer as limited,6 while no separate conclusion was drawn on oral cavity cancer.

Chapter 1

2.1.4 Ovarian cancer Epidemiological evidence The published literature examining the association between occupational asbestos exposure and ovarian cancer is relatively sparse as the workforce exposed to asbestos has been predominantly male. Yet, cohort studies of women who were heavily exposed to asbestos in their jobs consistently reported strong positive associations between asbestos exposure and ovarian cancer.3,6,64 In addition, studies on environmental asbestos exposure also showed increased risks, though non‐significant.6 However, misclassification of disease may have biased these results as most studies used the cause of death from the death certificate while it was particularly difficult to distinguish between peritoneal mesothelioma and ovarian serous carcinoma.65 Other evidence Asbestos fibers have been detected in the ovaries of exposed women, though possible sample contamination cannot be excluded.66,67 Furthermore, there are only a few animal studies investigating the association between asbestos exposure and ovarian cancer with no conclusive evidence.68 Evaluation of evidence by IARC In 2009, IARC concluded that sufficient evidence is available to classify the association between asbestos and ovarian cancer as causal.5

2.2 Exposure routes Asbestos exposure mainly occurs through inhalation and ingestion. Dermal absorption of asbestos is negligible, but dermal exposure may lead to secondary inhalation or ingestion of asbestos.3 After inhalation, asbestos fibers may deposit in the respiratory tract, or be cleared by the mucociliary escalator or alveolar macrophages. If retained, inhaled fibers may reach the pleura or other extrapulmonary sites. Several mechanisms have been proposed for extrapulmonary translocation of asbestos ranging from direct transpleural penetration, followed by movement to and transport via lymphatics and the bloodstream, to swallowing of asbestos fibers cleared from the lower respiratory tract by the mucociliary escalator.1 For the ovaries, retrograde movement of asbestos fibers through the reproductive tract has also been suggested as a possible route of exposure.66,69 This may be effected by perineal talc exposure that may have contained asbestos in the past.69,70

2.3 Mechanisms of carcinogenicity The carcinogenic potential of asbestos minerals is probably directly and exclusively related to their ability to form fibrous dust particles.2 Asbestos fibers differ with respect to size (length and diameter) and chemical composition. These disparities influence their deposition, movement, and clearance from the body and their carcinogenic

General introduction

potency. Fiber diameter is the most important factor affecting penetration and deposition in the lungs. Thin fibers have the highest inhalation potential and deposit deep within the lungs.3,6 The physicochemical properties most relevant to carcinogenesis are surface chemistry and reactivity, surface area, fiber dimensions, and biopersistence.5 Fibers longer than 8 μm with a diameter of less than 1.5 μm are considered the most potent carcinogens.3,6 The carcinogenic mechanisms of asbestos involve a complex interplay between the fibers and target cells. Direct and indirect mechanisms have been proposed, and established mechanistic events include impaired fiber clearance leading to macrophage activation, inflammation, generation of reactive oxygen and nitrogen species, tissue injury, genotoxicity, aneuploidy and polyploidy, epigenetic alteration, activation of signaling pathways, and resistance to apoptosis.5

RATIONALE AND AIM OF THESIS

Although the health effects of asbestos are one of the best documented in occupational health, they are also one of the most controversial as there is still considerable debate around asbestos carcinogenicity. Firstly, questions pertain to the risk at the lower end of the exposure distribution for mesothelioma and lung cancer, in order to set acceptable exposure limits and substantiate compensation claims. Secondly, it still needs to be determined if asbestos is causally related to gastrointestinal cancers and oral cavity and pharyngeal cancer. Thirdly, most studies have been carried out in an industry setting, resulting in the possibility of uncontrolled confounding due to lifestyle factors as smoking and alcohol for especially laryngeal, pharyngeal, and oral cavity cancer. Fourthly, except for lung cancer the number of cases was generally too small in most industry‐based studies to investigate the association with subtypes of cancer. Finally, the possible interaction between asbestos and smoking in relation to cancer has only been thoroughly examined for lung cancer with considerable variability in the magnitude of joint effects varying between additive and multiplicative. Population‐based studies are well suited to address these questions given their overall wide range in exposure levels including those at the lower end of the exposure distribution (i.e. exposure levels in jobs outside asbestos mining, insulation, cement and textile manufacturing, and other more highly exposed jobs), the possibility to control for potential confounders, and the often larger number of cases. The research described in this thesis was carried out within the framework of a population‐based study: the Netherlands Cohort Study (NLCS).71 Within this prospective cohort study, we aimed to assess:

Chapter 1

The overall association between occupational asbestos exposure and the risk of respiratory, gastrointestinal, and oral cavity and pharyngeal cancer, with special attention to: i. risk differences between relatively low and high exposure ii. risk associated with the lower end of the exposure distribution iii. the existence of an exposure‐response relation iv. potential confounding 2. The association between occupational asbestos exposure and subtypes of the cancers of interest 3. The possible additive or multiplicative interaction between smoking and asbestos in relation to the cancers of interest We were not able to look into the relation between occupational asbestos exposure and ovarian cancer and only studied men as the proportion of long‐term employed women was rather low in the NLCS. In most population‐based studies, actual exposure levels remain unknown, due to the absence of exposure measurements related to the actual work environments of the study participants.72 Moreover, study subjects are employed in a wide variety of jobs and workplaces, which renders (retrospective) assessment of occupational exposure extremely difficult for most, if not all population‐based studies.73,74 Several methodologies exist for estimating occupational exposures retrospectively from the available information on occupational histories,72 of which case‐by‐case expert assessment and job‐exposure matrices (JEMs) are commonly used.75 In case‐by‐case expert assessment, occupational experts assign exposure to every single job code based on job history information and/or subject reported information on jobs and exposure determinants. A JEM is a cross‐classification of job codes and/or industries on one axis and exposure agents on the other, with the cells of the matrix indicating the presence, intensity, and/or prevalence of exposure to a specific agent in a specific job code/industry. In addition, some JEMs contain a third axis indicating time period.72 Case‐by‐case expert assessment is generally considered the best possible method for assessing occupational exposure in population‐based studies,72,76 but this methodology is not always feasible as it is time‐consuming and dependent on the availability of expertise on (historical) occupational exposure and information ascertained within the study. As misclassification of exposure might bias and attenuate the risk estimates, reliable and valid assessment of occupational exposure to carcinogens is critical in the conduct of occupational epidemiological studies.72,77 Therefore, an additional aim of this thesis was to evaluate several candidate JEMs in terms of reliability, using case‐by‐case expert assessment as the ‘gold standard’ for the selection of the most appropriate JEM. The JEMs compared were the Asbestos JEM15 and DOMJEM76, both developed in the

General introduction

Netherlands, and the Finnish FINJEM.78 The reason for considering a Finnish JEM was that the FINJEM has already been used in the NLCS and contains quantitative exposure estimates.79

STUDY DESIGN

The research described in this thesis took place within the framework of the Netherlands Cohort Study.71 In brief, the NLCS started in September 1986 when 58 279 men and 62 573 women aged 55‐69 years, originating from 204 municipalities in the Netherlands with computerised population registries, were enrolled in the cohort. At baseline, participants completed a self‐administered questionnaire on dietary habits and lifestyle, occupational history and other potential risk factors for cancer.71 For reasons of efficiency in questionnaire processing and follow‐up, the case‐cohort approach was used.80 Incident cases were enumerated from the entire cohort, whereas the accumulated person‐years at risk in the entire cohort were estimated from a random subcohort of 5000 subjects (2411 men and 2589 women), selected immediately after baseline. This subcohort is being followed‐up for vital status information while the entire cohort is being monitored for incident cancer by annual record linkage to the Netherlands Cancer Registry and the Dutch Pathology Registry (PALGA).81,82 The research described in this thesis was based on 17.3 years of follow‐up of the cohort (baseline to December 2003). Completeness of incident cancer coverage was estimated to be almost 100%.83

OUTLINE OF THESIS

In order to determine which JEMs are suitable for use in the NLCS, chapter 2 describes a study in which several candidate JEMs were evaluated in terms of reliability, using the case‐by‐case expert assessment of the subcohort as the ‘gold standard’. In chapter 3, we examined the association between occupational asbestos exposure and the risk of respiratory cancers: pleural mesothelioma, lung and laryngeal cancer. In chapter 4, we studied the association between occupational asbestos exposure and the risk of gastrointestinal cancers: esophageal, gastric and colorectal cancer. In chapter 5, we looked into the association between occupational asbestos exposure and risk of tumors of the oral cavity and pharynx. Finally, chapter 6 is a discussion of the main findings of occupational asbestos research in the population‐based NLCS in light of some important strengths and limitations ‐ especially when compared to industry‐based studies ‐ and recommendations for future research.

Chapter 1

REFERENCES 1. 2.

3. 4. 5.

7. 8. 9. 10. 11. 12. 13. 14.

15.

16. 17. 18. 19. 20.

21. 22.

23.

NAS. Asbestos: Selected Cancers. Washington, DC: The National Academies Press, 2006. Case BW, Abraham JL, Meeker G, Pooley FD, Pinkerton KE. Applying definitions of "asbestos" to environmental and "low‐dose" exposure levels and health effects, particularly malignant mesothelioma. J Toxicol Environ Health B Crit Rev. 2011;14:3‐39. National Toxicology Program. Asbestos. Rep Carcinog. 2011;12:53‐6. Virta RL. Worldwide Asbestos Supply and Consumption Trends From 1900 Through 2003. U.S. Geological Survey Circular 1298, 2006. Straif K, Benbrahim‐Tallaa L, Baan R, Grosse Y, Secretan B, El Ghissassi F, Bouvard V, Guha N, Freeman C, Galichet L, Cogliano V; WHO International Agency for Research on Cancer Monograph Working Group. A review of human carcinogens‐‐part C: metals, arsenic, dusts, and fibres. Lancet Oncol. 2009;10:453‐4. IARC. Asbestos (Chrysotile, Amosite, Crocidolite, Tremolite, Actinolite and Anthophyllite). In: A Review of Human Carcinogens: Arsenic, Metals, Fibres, and Dusts. Lyon, France: International Agency for Research on Cancer (IARC), 2012. 219‐309. Swuste PHJJ, Burdorf A, Klaver JAM. Asbest, het inzicht in de schadelijke gevolgen in de periode 1930‐ 1969 in Nederland. Delft: Delftse Universitaire Pers, 1988. Selikoff IJ, Churg J, Hammond EC. Asbestos Exposure and Neoplasia. JAMA. 1964;188:22‐6. Swuste P, Burdorf A, Ruers B. Asbestos, asbestos‐related diseases, and compensation claims in The Netherlands. Int J Occup Environ Health. 2004;10:159‐65. International Ban Asbestos Secretariat. http://ibasecretariat.org/alpha_ban_list.php Accessed February 4, 2015. Ramazzini C. Asbestos is still with us: repeat call for a universal ban. J Occup Environ Med. 2010;52: 469‐72. Zielhuis RL. [Biological effects of asbestos. 2d International Conference, Dresden, 22‐25 April, 1968]. Ned Tijdschr Geneeskd. 1968;112:1494‐6. Stumphius J. Asbest in een bedrijfsbevolking. Assen: Van Gorcum & Comp, 1969. Burdorf A, Swuste PHJJ, Looman CWN. Estimating the future developments in mesothelioma mortality due to historical asbestos use and its consequences for public health. In: Peters GB Peters (eds). Sourcebook on asbestos diseases. Charlottesville, VA: Reed Elsevier, 1998. 145‐67. Swuste P, Dahhan M, Burdorf A. Linking expert judgement and trends in occupational exposure into a job‐exposure matrix for historical exposure to asbestos in the Netherlands. Ann Occup Hyg. 2008;52:397‐403. Ross M. The geologic occurrences and health hazards of amphibole and serpentine asbestos. Rev Mineral Geochem. 1981;9A:279‐323. Ruers RF. Macht en tegenmacht in de Nederlandse asbestregulering. Den Haag: Boom Juridische uitgevers, 2012. U.S. Geological Survey. Mineral commodity summaries 2012. http://minerals.usgs.gov/minerals/ pubs/mcs/2012/mcs2012.pdf Accessed February 4, 2015. Burdorf A, Swuste P. An expert system for the evaluation of historical asbestos exposure as diagnostic criterion in asbestos‐related diseases. Ann Occup Hyg. 1999;43:57‐66. Gezondheidsraad. Asbestos: risks of environmental and occupational exposure. The Hague, the Netherlands: Health Council of the Netherlands, report 2010/10E. http://www.gezondheidsraad.nl/ sites/default/files/201010E.pdf. Accessed February 4, 2015. Burdorf A, Jarvholm B, Englund A. Explaining differences in incidence rates of pleural mesothelioma between Sweden and the Netherlands. Int J Cancer. 2005;113:298‐301. Harmsma S, Mulder HFHM. Asbest in Kaart, historisch onderzoek asbestgebruik, methode asbestkansenkaart. Groningen: Register Historisch onderzoekbureau, Senter‐Novem/Bodem+ Landelijk Informatiebeheer Bodem, 2006. Driece HA, Siesling S, Swuste PH, Budorf A. Assessment of cancer risks due to environmental exposure to asbestos. J Expo Sci Environ Epidemiol. 2010;20:478‐85.

General introduction

24. Ferrante D, Bertolotti M, Todesco A, Mirabelli D, Terracini B, Magnani C. Cancer mortality and incidence of mesothelioma in a cohort of wives of asbestos workers in Casale Monferrato, Italy. Environ Health Perspect. 2007;115:1401‐5. 25. LaDou J. The asbestos cancer epidemic. Environ Health Perspect. 2004;112:285‐90. 26. Segura O, Burdorf A, Looman C. Update of predictions of mortality from pleural mesothelioma in the Netherlands. Occup Environ Med. 2003;60:50‐5. 27. Marinaccio A, Binazzi A, Cauzillo G, Cavone D, Zotti RD, Ferrante P, Gennaro V, Gorini G, Menegozzo M, Mensi C, Merler E, Mirabelli D, Montanaro F, Musti M, Pannelli F, Romanelli A, Scarselli A, Tumino R; Italian Mesothelioma Register (ReNaM) Working Group. Analysis of latency time and its determinants in asbestos related malignant mesothelioma cases of the Italian register. Eur J Cancer. 2007;43:2722‐8. 28. Doll R. Mortality from lung cancer in asbestos workers. Br J Ind Med. 1955;12:81‐6. 29. Wagner JC, Sleggs CA, Marchand P. Diffuse pleural mesothelioma and asbestos exposure in the North Western Cape Province. Br J Ind Med. 1960;17:260‐71. 30. Hodgson JT, Darnton A. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure. Ann Occup Hyg. 2000;44:565‐601. 31. Berman DW, Crump KS. Update of potency factors for asbestos‐related lung cancer and mesothelioma. Crit Rev Toxicol. 2008;38 Suppl 1:1‐47. 32. Berman DW, Crump KS. A meta‐analysis of asbestos‐related cancer risk that addresses fiber size and mineral type. Crit Rev Toxicol. 2008;38 Suppl 1:49‐73. 33. Ilgren EB, Browne K. Asbestos‐related mesothelioma: evidence for a threshold in animals and humans. Regul Toxicol Pharmacol. 1991;13:116‐32. 34. Gibbs GW, Berry G. Mesothelioma and asbestos. Regul Toxicol Pharmacol. 2008;52:S223‐31. 35. Iwatsubo Y, Pairon JC, Boutin C, Ménard O, Massin N, Caillaud D, Orlowski E, Galateau‐Salle F, Bignon J, Brochard P. Pleural mesothelioma: dose‐response relation at low levels of asbestos exposure in a French population‐based case‐control study. Am J Epidemiol. 1998;148:133‐42. 36. Hillerdal G. Mesothelioma: cases associated with non‐occupational and low dose exposures. Occup Environ Med. 1999;56:505‐13. 37. van der Bij S, Koffijberg H, Lenters V, Portengen L, Moons KG, Heederik D, Vermeulen RC. Lung cancer risk at low cumulative asbestos exposure: meta‐regression of the exposure‐response relationship. Cancer Causes Control. 2013;24:1‐12 38. Gustavsson P1, Nyberg F, Pershagen G, Schéele P, Jakobsson R, Plato N. Low‐dose exposure to asbestos and lung cancer: dose‐response relations and interaction with smoking in a population‐based case‐ referent study in Stockholm, Sweden. Am J Epidemiol. 2002;155:1016‐22. 39. De Matteis S, Consonni D, Lubin JH, Tucker M, Peters S, Vermeulen RCh, Kromhout H, Bertazzi PA, Caporaso NE, Pesatori AC, Wacholder S, Landi MT. Impact of occupational carcinogens on lung cancer risk in a general population. Int J Epidemiol. 2012;41:711‐21. 40. Johansson L, Albin M, Jakobsson K, Mikoczy Z. Histological type of lung carcinoma in asbestos cement workers and matched controls. Br J Ind Med. 1992;49:626‐30. 41. Raffn E, Lynge E, Korsgaard B. Incidence of lung cancer by histological type among asbestos cement workers in Denmark. Br J Ind Med. 1993;50:85‐9. 42. Lee BW, Wain JC, Kelsey KT, Wiencke JK, Christiani DC. Association of cigarette smoking and asbestos exposure with location and histology of lung cancer. Am J Respir Crit Care Med. 1998;157:748‐55. 43. Paris C, Benichou J, Saunier F, Metayer J, Brochard P, Thiberville L, Nouvet G. Smoking status, occupational asbestos exposure and bronchial location of lung cancer. Lung Cancer. 2003;40:17‐24. 44. Marchand JL, Luce D, Leclerc A, Goldberg P, Orlowski E, Bugel I, Brugère J. Laryngeal and hypopharyngeal cancer and occupational exposure to asbestos and man‐made vitreous fibers: results of a case‐control study. Am J Ind Med. 2000;37:581‐9. 45. Elci OC, Akpinar‐Elci M, Blair A, Dosemeci M. Occupational dust exposure and the risk of laryngeal cancer in Turkey. Scand J Work Environ Health. 2002;28:278‐84. 46. Ramroth H, Ahrens W, Dietz A, Becher H. Occupational asbestos exposure as a risk factor for laryngeal carcinoma in a population‐based case‐control study from Germany. Am J Ind Med. 2011;54:510‐4. 47. Muscat JE, Wynder EL. Tobacco, alcohol, asbestos, and occupational risk factors for laryngeal cancer. Cancer. 1992;69:2244‐51.

Chapter 1

48. De Stefani E, Boffetta P, Oreggia F, Ronco A, Kogevinas M, Mendilaharsu M. Occupation and the risk of laryngeal cancer in Uruguay. Am J Ind Med. 1998;33:537‐42. 49. Hinds MW, Thomas DB, O'Reilly HP. Asbestos, dental X‐rays, tobacco, and alcohol in the epidemiology of laryngeal cancer. Cancer. 1979;44:1114‐20. 50. Morgan RW, Foliart DE, Wong O. Asbestos and gastrointestinal cancer. A review of the literature. West J Med. 1985;143:60‐5. 51. Gustavsson P, Jakobsson R, Johansson H, Lewin F, Norell S, Rutkvist LE. Occupational exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus: a case‐control study in Sweden. Occup Environ Med. 1998;55:393‐400. 52. Parent ME, Siemiatycki J, Fritschi L. Workplace exposures and oesophageal cancer. Occup Environ Med. 2000;57:325‐34. 53. Jansson C, Johansson AL, Bergdahl IA, Dickman PW, Plato N, Adami J, Boffetta P, Lagergren J. Occupational exposures and risk of esophageal and gastric cardia cancers among male Swedish construction workers. Cancer Causes Control. 2005;16:755‐64. 54. Santibañez M, Vioque J, Alguacil J, Barber X, García de la Hera M, Kauppinen T; PANESOES Study Group. Occupational exposures and risk of oesophageal cancer by histological type: a case‐control study in eastern Spain. Occup Environ Med. 2008;65:774‐81. 55. Gamble J. Risk of gastrointestinal cancers from inhalation and ingestion of asbestos. Regul Toxicol Pharmacol. 2008;52:S124‐53. 56. Frumkin H, Berlin J. Asbestos exposure and gastrointestinal malignancy review and meta‐analysis. Am J Ind Med. 1988;14:79‐95. 57. Jansson C, Plato N, Johansson AL, Nyrén O, Lagergren J. Airborne occupational exposures and risk of oesophageal and cardia adenocarcinoma. Occup Environ Med. 2006;63:107‐12. 58. Sjödahl K, Jansson C, Bergdahl IA, Adami J, Boffetta P, Lagergren J. Airborne exposures and risk of gastric cancer: a prospective cohort study. Int J Cancer. 2007;120:2013‐8. 59. Tarvainen L, Kyyrönen P, Kauppinen T, Pukkala E. Cancer of the mouth and pharynx, occupation and exposure to chemical agents in Finland [in 1971‐95]. Int J Cancer. 2008;123:653‐9. 60. Paget‐Bailly S, Cyr D, Luce D. Occupational exposures to asbestos, polycyclic romatic hydrocarbons and solvents, and cancers of the oral cavity and pharynx: a quantitative literature review. Int Arch Occup Environ Health. 2012;85:341‐51. 61. Reid A, Ambrosini G, de Klerk N, Fritschi L, Musk B. Aerodigestive and gastrointestinal tract cancers and exposure to crocidolite (blue asbestos): incidence and mortality among former crocidolite workers. Int J Cancer. 2004;111:757‐61. 62. Purdue MP, Järvholm B, Bergdahl IA, Hayes RB, Baris D. Occupational exposures and head and neck cancers among Swedish construction workers. Scand J Work Environ Health. 2006;32:270‐5. 63. Langevin SM, O'Sullivan MH, Valerio JL, Pawlita M, Applebaum KM, Eliot M, McClean MD, Kelsey KT. Occupational asbestos exposure is associated with pharyngeal squamous cell carcinoma in men from the greater Boston area. Occup Environ Med. 2013;70:858‐63. 64. Camargo MC, Stayner LT, Straif K, Reina M, Al‐Alem U, Demers PA, Landrigan PJ. Occupational exposure to asbestos and ovarian cancer: a meta‐analysis. Environ Health Perspect. 2011;119:1211‐7. 65. Reid A, de Klerk N, Musk AW. Does exposure to asbestos cause ovarian cancer? A systematic literature review and meta‐analysis. Cancer Epidemiol Biomarkers Prev. 2011;20:1287‐95. 66. Heller DS, Gordon RE, Westhoff C, Gerber S. Asbestos exposure and ovarian fiber burden. Am J Ind Med. 1996;29:435‐9. 67. Langseth H, Johansen BV, Nesland JM, Kjaerheim K. Asbestos fibers in ovarian tissue from Norwegian pulp and paper workers. Int J Gynecol Cancer. 2007;17:44‐9. 68. Bunderson‐Schelvan M, Pfau JC, Crouch R, Holian A. Nonpulmonary outcomes of asbestos exposure. J Toxicol Environ Health B Crit Rev. 2011;14:122‐52. 69. Baan R, Straif K, Grosse Y, Secretan B, El Ghissassi F, Cogliano V; WHO International Agency for Research on Cancer Monograph Working Group. Carcinogenicity of carbon black, titanium dioxide, and talc. Lancet Oncol. 2006;7:295‐6. 70. Langseth H, Hankinson SE, Siemiatycki J, Weiderpass E. Perineal use of talc and risk of ovarian cancer. J Epidemiol Community Health. 2008;62:358‐60.

General introduction

71. van den Brandt PA, Goldbohm RA, van 't Veer P, Volovics A, Hermus RJ, Sturmans F. A large‐scale prospective cohort study on diet and cancer in The Netherlands. J Clin Epidemiol. 1990;43:285‐95. 72. Teschke K, Olshan AF, Daniels JL, De Roos AJ, Parks CG, Schulz M, Vaughan TL. Occupational exposure assessment in case‐control studies: opportunities for improvement. Occup Environ Med. 2002;59:575‐ 93; discussion 594. 73. Kromhout H, Heederik D, Dalderup LM, Kromhout D. Performance of two general job‐exposure matrices in a study of lung cancer morbidity in the Zutphen cohort. Am J Epidemiol. 1992;136:698‐711. 74. Kromhout H, Vermeulen R. Application of job‐exposure matrices in studies of the general population: some clues to their performance. Eur Respir Rev. 2001;11:80‐90. 75. Goldberg M, Imbernon E. The use of job exposure matrices for cancer epidemiology research and surveillance. Arch Public Health. 2002;60:173‐85. 76. Peters S, Vermeulen R, Cassidy A, Mannetje A', van Tongeren M, Boffetta P, Straif K, Kromhout H; INCO Group. Comparison of exposure assessment methods for occupational carcinogens in a multi‐centre lung cancer case‐control study. Occup Environ Med. 2011;68:148‐53. 77. Benke G, Sim M, Forbes A, Salzberg M. Retrospective assessment of occupational exposure to chemicals in community‐based studies: validity and repeatability of industrial hygiene panel ratings. Int J Epidemiol. 1997;26:635‐42. 78. Kauppinen T, Toikkanen J, Pukkala E. From cross‐tabulations to multipurpose exposure information systems: a new job‐exposure matrix. Am J Ind Med. 1998;33:409‐17. 79. Preller L, van den Bosch LM, van den Brandt PA, Kauppinen T, Goldbohm A. Occupational exposure to silica and lung cancer risk in the Netherlands. Occup Environ Med. 2010;67:657‐63. 80. Prentice RL. A case‐cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika. 1986;73:1‐11. 81. Casparie M, Tiebosch AT, Burger G, Blauwgeers H, van de Pol A, van Krieken JH, Meijer GA. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol. 2007;29:19‐24. 82. Van den Brandt PA, Schouten LJ, Goldbohm RA, Dorant E, Hunen PM. Development of a record linkage protocol for use in the Dutch Cancer Registry for Epidemiological Research. Int J Epidemiol. 1990;19:553‐8. 83. Goldbohm RA, van den Brandt PA, Dorant E. Estimation of the coverage of Dutch municipalities by cancer registries and PALGA based on hospital discharge data. Tijdschr Soc Gezonheidsz. 1994;72:80‐4.

Chapter 1

Chapter 2

Comparison of expert and job‐exposure matrix‐based retrospective exposure assessment of occupational carcinogens in the Netherlands Cohort Study Nadine S.M. Offermans Roel Vermeulen Alex Burdorf Susan Peters R. Alexandra Goldbohm Tom Koeman Martie van Tongeren T. Kauppinen Ijmert Kant Hans Kromhout Piet A. van den Brandt Occupational and Environmental Medicine. 2012;69:745‐51

Chapter 2

ABSTRACT Objectives Reliable retrospective exposure assessment continues to be a challenge in most population‐based studies. Several methodologies exist for estimating exposures retrospectively, of which case‐by‐case expert assessment and job‐exposure matrices (JEMs) are commonly used. This study evaluated the reliability of exposure estimates for selected carcinogens obtained through three JEMs by comparing the estimates with case‐by‐case expert assessment within the Netherlands Cohort Study (NLCS). Methods The NLCS includes 58 279 men aged 55‐69 years at enrolment in 1986. For a subcohort of these men (n=1630), expert assessment is available for exposure to asbestos, polycyclic aromatic hydrocarbons (PAHs) and welding fumes. Reliability of the different JEMs (DOMJEM (asbestos, PAHs), FINJEM (asbestos, PAHs and welding fumes) and Asbestos JEM (asbestos) was determined by assessing the agreement between these JEMs and the expert assessment. Results Expert assessment revealed the lowest prevalence of exposure for all three exposures (asbestos 9.3%; PAHs 5.3%; welding fumes 11.7%). The DOMJEM showed the highest level of agreement with the expert assessment for asbestos and PAHs (ks=0.29 and 0.42, respectively), closely followed by the FINJEM. For welding fumes, concordance between the expert assessment and FINJEM was high (k=0.70). The Asbestos JEM showed poor agreement with the expert asbestos assessment (k=0.10). Conclusions This study shows case‐by‐case expert assessment to result in the lowest prevalence of occupational exposure in the NLCS. Furthermore, the DOMJEM and FINJEM proved to be rather similar in agreement when compared with the expert assessment. The Asbestos JEM appeared to be less appropriate for use in the NLCS.

Retrospective occupational exposure assessment

INTRODUCTION Reliable and valid assessment of occupational exposure to carcinogens is critical in the conduct of occupational epidemiological studies, as misclassification of exposure might bias and attenuate the risk estimate.1,2 Several methodologies exist for estimating occupational exposures retrospectively from the available information on occupational histories,2 of which case‐by‐case expert assessment and job‐exposure matrices (JEMs) are commonly used.3 Case‐by‐case expert assessment is generally considered the best possible method for assessing occupational exposure in population‐based studies.2,4 Availability of detailed descriptions of industry, company, working environment, tasks and products used enables experts to allow for local differences in material usage, production processes, control measures and/or calendar period of exposure.4 This enables them to take account of exposure differences between individuals with similar jobs, which can result in less misclassification of exposure. Possible disadvantages of this approach are the learning phenomenon (as experts’ skills evolve during the process)1,5 and inter‐expert variability within studies and between studies (as their training, experience and work field might be different).4,6 In addition, case‐by‐case expert assessment requires considerable resources.5 A JEM is a cross‐classification of job codes and/or industries on one axis and exposure agents on the other, with the cells of the matrix indicating the presence, intensity and/or prevalence of exposure to a specific agent in a specific job code/industry. In addition, some JEMs contain a third axis indicating time period.2 Unlike expert assessment, JEMs allocate the same exposure estimate to all workers within a job code, thereby disregarding possible inter‐individual variability within job codes. This is a major drawback since there may be large differences in exposure levels between individuals with the same job in the same company and/or different companies. Conversely, the main advantage of using a JEM lies in the more standardised way in which jobs are being translated into specific exposures (i.e., computerised linkage of exposure estimates to job codes), representing a more efficient and reproducible methodology.4 In addition, some JEMs, like the FINJEM, might also provide a more transparent methodology compared with case‐by‐case expert assessment as exposure measurements (though small in number) underlie the exposure estimate of (several of) the matrix cells.7 In most population‐based studies, actual exposure levels remain unknown due to the absence of exposure measurements,2 though recent efforts have made use of existing large amounts of exposure data.8,9 Moreover, study subjects are employed in a wide variety of jobs and workplaces, which renders (retrospective) assessment of

Chapter 2

occupational exposure extremely difficult for most, if not all population‐based studies.10,11 One of these population‐based studies is the prospective Netherlands Cohort Study (NLCS), which started in 1986 among 120 852 men and women.12,13 For a random sample of this cohort (the subcohort), case‐by‐case expert assessment, executed in the framework of a previous study in the NLCS,13 is available for exposure to asbestos, polycyclic aromatic hydrocarbons (PAHs) and welding fumes. As expert assessment is not possible for the cohort as a whole, the objective of the present study was to evaluate several candidate JEMs in terms of reliability, using the case‐by‐case expert assessment of the subcohort as the ‘gold standard’ for the selection of the most appropriate JEM. Therefore, in this study, reliability refers to method agreement: ‘Do two techniques used to measure a particular variable, in otherwise identical circumstances, produce the same result?’.14 The JEMs compared were the Asbestos JEM15 and DOMJEM.4 developed in the Netherlands, and the Finnish FINJEM.7 The reason for considering a Finnish JEM was that the FINJEM has been used in a previous study in the NLCS.16

METHODS Study population The study design and data collection strategies have been described in detail previously. In brief, the NLCS started in September 1986 when 58 279 men and 62 573 women aged 55‐69 years, originating from 204 municipalities with computerized population registries, were enrolled in the cohort. At baseline, participants completed a self‐administered questionnaire on dietary habits and lifestyle, occupational history and other potential risk factors for cancer.12 Following the case‐cohort approach, a random subcohort was selected immediately after baseline to estimate person‐years at risk accumulated in the entire cohort.12 As the number of long‐term employed women was rather low, the current analyses were restricted to men. For these male subcohort members (N=1630), case‐by‐case expert assessment was available in the framework of a previous study in the NLCS for exposure to asbestos, PAHs and welding fumes.13 Therefore, the current study was conducted on this subcohort only. The NLCS was approved by the institutional review boards of the Netherlands Organisation for Applied Scientific Research TNO (Zeist) and Maastricht University (Maastricht).

Methods of assessment of occupational exposure Information on lifetime occupational history until 1986 was obtained from the questionnaire filled in at study enrolment. Questions concerned the job title, name and

Retrospective occupational exposure assessment

type of the company, products made in the department and period of employment. Based on these questions, occupations were coded according to the Standard Occupational Classification of 1984 of the Dutch Central Bureau of Statistics (CBS‐84), supplemented by a three‐digit code assigned within the NLCS (CBS‐NLCS) based on the job title. Subjects could enter a maximum of five occupations which was generally sufficient to cover the lifetime occupational history for the large majority of the cohort, as cohort subjects held on average 1.9 job codes during their working life. For all subjects, the job code was assessed for each of the maximally five occupations held between starting work and 1986. Case‐by‐case expert assessment For the male subcohort members in the NLCS, the combined case‐by‐case expert assessment of two independent expert raters (an occupational hygienist, IJ Kant, and an occupational epidemiologist, Gerard MH Swaen) is available, among others, for exposure to asbestos, PAHs and welding fumes.13 This expert assessment consisted of three steps. First, both experts excluded, independent of each other, all CBS‐NLCS job codes with no potential exposure to these carcinogens. Second, for all job codes with potential exposure, the complete self‐reported information on job title, name and type of the company, products made in the department and period of employment was used for scoring the prevalence (P) of exposure to these carcinogens in the Netherlands, as the prevalence may differ between companies and periods. This scoring was also carried out independent of each other. Four exposure prevalence categories were defined: no exposure to the specific carcinogen, possible exposure (prevalence of exposure estimated to be <30%), probable exposure (prevalence of exposure of 30%‐ 90%) and nearly certain exposure (prevalence of exposure >90%). In a final expert meeting, all job codes for which at least one of the experts suspected relevant exposure were re‐evaluated and a consensus prevalence of exposure was assigned to each of the job codes within the NLCS. In this assessment, the intensity (I) of exposure was not taken into account. It may be expected that those job codes with a high prevalence of exposure also entail a high intensity of exposure, but a secondary analysis of FINJEM for asbestos and PAHs showed that this is not an unequivocal pattern. For job codes with a low prevalence of exposure, intensity of exposure is often less clear. Table 1 shows an overview of relevant characteristics. JEMs The Dutch Asbestos JEM and DOMJEM and the Finnish FINJEM differ from each other regarding their axes. For an overview of relevant characteristics, see Table 2.1.

Chapter 2

Table 2.1

Characteristics of the case‐by‐case expert assessment, Asbestos JEM, DOMJEM, and FINJEM.

Case‐by‐case expert Asbestos JEM assessment Occupational axis CBS‐NLCS CBS‐84 (1036 job codes (123 job codes) present in the subcohort)

DOMJEM ISCO‐68 (1842 job codes)

Exposure axis; included agents Time axis

Asbestos PAHs Welding fumes

Asbestos

Asbestos PAHs

Exposure is assigned for the length of an occupation, taking into consideration the time periods encompassed.

No time scale present, exposure is an estimation of the average exposure over time.

Matrix cells; Measures of exposure

P (%): 0=0 1<30 2=30‐90 3>90

1945‐1949 1950‐1954 1955‐1959 1960‐1964 1965‐1969 1970‐1974 1975‐1979 1980‐1984 1985‐1989 * P (%): 0=0 1=5‐20 2=20‐80 3>80 † 3 I (fibres/cm ): 0=0 1=0‐0.5 2=0.5‐1 3=1‐2 4=2‐5 5=5‐10 6=10‐20

FINJEM The Combined Occupational Classification of Finnish Censuses in 1970‐85 (311 job codes) Asbestos Benzo‐a‐pyrene (i.e. proxy for PAHs) Welding fumes 1945‐1959 1960‐1974 1975‐1984 1985‐1994

PI P (%): General cut‐off points 0‐100 are not defined for † the DOMJEM. I : Continuous scale, PI: exposure specific units 0 = no exposure 1 = low exposure PI: (low Phigh I / Continuous scale, high Plow I) exposure specific units 2 = high exposure (high Phigh I)

†

P refers to prevalence of exposure, i.e., agent‐specific prevalence of exposure in an occupation. I refers to the mean intensity of exposure among the exposed. JEM, job‐exposure matrix.

The Asbestos JEM is a partly disease‐oriented JEM as it is based on known asbestos exposure through available asbestos measurements (partly from the FINJEM Database) and verified cases of mesothelioma within occupations.15 On the occupational axis, the Asbestos JEM follows CBS‐84 and includes only the 123 job codes with definite exposure to asbestos, out of the total of 914 job codes included in CBS‐84. The exposure axis contains the agent asbestos, for which estimates of the prevalence (P) and intensity (I) of exposure are present in a categorical manner (four categories for P; seven categories for I). The time axis in the Asbestos JEM consists of ten 5‐year time periods covering the period 1945‐1994.

Retrospective occupational exposure assessment

The DOMJEM is a generic JEM developed by experts in the Netherlands for application in general population studies. Given that specific occupational exposures are relatively rare, the slightest deviation of perfect specificity will lead to an (marked) underestimation of the degree of association. Therefore, the experts scored occupations, based on their knowledge and experience, with an emphasis on specificity rather than sensitivity.4 The DOMJEM uses ISCO‐68 (International Standard Classification of Occupations of 1968) and contains 1842 different job codes. The relevant agents for this study included in the exposure axis are asbestos and PAHs, and the accompanying measure of exposure is a combined measure of PI, for which three categories are available in the DOMJEM. As exposure is an estimation of the average exposure over time, this JEM does not possess a time axis. However, asbestos exposure levels assigned pertain to the period before the ban. The FINJEM was constructed for exposure assessment in large register‐based studies and is based on both expert assessment and exposure measurements.7 The FINJEM contains all 311 classes of the Combined Occupational Classification of Finnish Censuses in 1970‐1985. The relevant agents covered by the exposure axis are asbestos, PAHs and welding fumes for which three continuous measures of exposure are available: P, I as well as a combined measure of PI. This JEM includes a time axis consisting of four time periods, dissimilar in length, covering the period 1945‐1994. FINJEM also includes five additional 3‐year periods covering the years 1995‐2009, but these were not used in the present study.

Comparison of exposure assessment methods The main focus of the comparison was to study if the JEMs assign exposure to subjects similar to the reference method, case‐by‐case expert assessment and to each other. As the risk analyses within NLCS are primarily based on estimates of cumulative exposure (CE), CE per subject was chosen as the unit of comparison. Although several exposure measures are available (i.e., prevalence and intensity), none of them are available for all methods: the DOMJEM only contains PI and the case‐by‐case expert assessment only includes P. Given that case‐by‐case expert assessment only includes P, this measure was compared with PI in the JEMs. For the case‐by‐case expert assessment, the CE per subject is a combined measure of the prevalence (P) and duration (years) of exposure. To calculate the CE for the expert assessment, a weight was assigned to each category of exposure prevalence. Each weight corresponds to the midpoint of prevalence in each exposure category: no exposure, weight 0; possible exposure, weight 0.15; probable exposure, weight 0.60 and nearly certain exposure, weight 0.95. For each carcinogen, the CE per subject was calculated by multiplying the weight given to each exposure category by the number of years exposed in a certain job code and subsequently summing up all weighted exposures. 13

Chapter 2

For the JEMs, the CE is a combined measure of the prevalence, intensity (I) and duration (years) of exposure. To arrive at the CE for the JEMs, first the CBS‐NLCS codes had to be recoded to the job classification used in the job axis of the JEMs (CBS‐84 for the Asbestos JEM, ISCO‐68 for the DOMJEM and Finnish codes for the FINJEM) in order to be able to link the JEMs. For that reason, the CBS‐NLCS codes in the subcohort were merged into the broader classes of CBS‐84. Furthermore, an algorithm recoding from CBS‐NLCS to the five‐digit ISCO‐68 has been developed for the NLCS.17 Finally, an algorithm recoding ISCO‐68 to the Finnish job codes had already been developed within the framework of the INTEROCC Study (Van Tongeren M, Kincl L, Richardson L, et al., paper presented at EPICOH 2011, Oxford). After recoding of the job codes, the actual linkage with the JEMs was carried out. In order to obtain the CE for the Asbestos JEM and FINJEM, first PI per job code was estimated using the time‐specific exposure information in the time axis of both JEMs, before summing PI over the maximum of five occupation subjects could enter. For those workers who started working before 1945, exposure estimates from the earliest time period in both JEMs were used, since the period before 1945 was absent in both JEMs. For the DOMJEM, the CE was estimated by summing the product of PI and duration for the maximum of five job codes. The DOMJEM scores of no, low and high exposure for PI were arbitrarily assigned values of 0, 1 and 4 to mirror the log‐normal (multiplicative) nature of occupational exposure levels. The weighting was based on reported levels for semi‐quantitatively scored exposure, thereby assuring a balanced weighting between intensity and duration in the calculation of CE. 18

Statistical analyses In the subcohort, 1455 of the 1630 men completed the self‐reported occupational history sufficiently to assign job codes. For these male subcohort members, the agreement in exposure assessment between the JEMs mutually and the case‐by‐case expert assessment was estimated. First, the prevalence of exposure among the male subcohort members according to the different methods was calculated as described in the previous section. The CE was categorised in four categories: non‐exposed and three categories for the exposed based on the tertiles of CE. In order to determine the prevalence of exposure, the tertiles of CE were combined into one category of exposed (i.e., CE>0). Second, weighted Cohen’s κs and 95% CIs were determined for the four categories of CE using Stata SE V.10.0 (Stata). Linear weights were applied to take into account the extent of disagreement. These weights are proportional to the number of categories, with perfect agreement along the main diagonal (weight of 1) and smaller weights for the other cells (0.5 for ‘no vs. low exposure’ and ‘low vs. high exposure’), the smallest weight being for the biggest disagreement (0 for ‘no exposure vs. high exposure’). ks were interpreted using the following arbitrary cut‐off points: <0.4 poor, 0.4‐0.75 moderate to good and >0.75 excellent agreement.

Retrospective occupational exposure assessment

The following comparisons were made: 1. Case‐by‐case expert assessment versus FINJEM (for exposure to asbestos, PAHs and welding fumes). 2. Case‐by‐case expert assessment versus DOMJEM (for exposure to asbestos and PAHs). 3. Case‐by‐case expert assessment versus Asbestos JEM (for exposure to asbestos). 4. FINJEM versus DOMJEM (for exposure to asbestos and PAHs). 5. FINJEM versus Asbestos JEM (for exposure to asbestos). 6. DOMJEM versus Asbestos JEM (for exposure to asbestos).

Sensitivity analyses Besides the main analysis where the unit of comparison concerned the individual CE for all male subcohort members, agreement between the different methods was also assessed restricted to exposed subjects in order to determine how the methods agree on the actual level of exposure. In order to determine if a subject was exposed, case‐by‐ case expert assessment was used. Second, a comparison was made solely on P for FINJEM and the Asbestos JEM with the expert assessment. Third, individual job codes were used as the unit of comparison instead of the male subcohort members, with the CE per job code as the measure of interest for all methods. Fourth, an advantage of this approach included the possibility to study how agreement changes by time period, by assessing agreement on CE per job code for the periods 1929‐1944, 1945‐1959, 1960‐ 1974 and 1975‐1986 separately instead of focusing on the cumulative lifetime occupational exposure. As exposure prevalence in the expert assessment was estimated for the years in which the male subcohort members practiced their jobs, it was not possible to calculate a mean expert‐based exposure for the abovementioned periods. Therefore, only the job codes of which the start and end dates concurred with the abovementioned time periods could be taken into the analyses.

RESULTS Subject characteristics The 1455 male NLCS subcohort members (mean age of 61.3 years) worked on average 16.6 years in the same job code (CBS‐NLCS) and held on average 1.9 job codes during their working life, resulting in 1036 different CBS‐NLCS codes from a total of 14 154 job codes. These CBS‐NLCS codes were recoded into the respective ISCO‐68 code (n=508), Finnish code (n=236) and CBS‐84 code (n=78 of the 123 codes that entail asbestos exposure) to allow linkage to the DOMJEM, FINJEM and Asbestos JEM, respectively. The timeframe during which subjects were occupationally active, as reported in the

Chapter 2

questionnaire, includes the period 1929‐1986. Furthermore, 40.8% of the male subcohort members started working before 1945.

Exposure prevalence For exposure to asbestos, case‐by‐case expert assessment had the lowest prevalence of exposure (9.3%), followed by the FINJEM (28.3%), DOMJEM (30.0%) and Asbestos JEM (32.9%). For exposure to PAHs, case‐by‐case expert assessment also showed the lowest prevalence of exposure (5.3%), more closely followed by the DOMJEM (8.5%) and FINJEM (10.1%). The lowest prevalence of exposure to welding fumes was also attained by the case‐by‐case expert assessment (11.7%), with FINJEM resulting in a prevalence of 15.6%.

Agreement in exposure assessment: weighted Cohen’s κ Weighted Cohen’s κs and 95% CIs for exposure assigned to the male subcohort members by the different methods are shown in Table 2.2. Table 2.2

Weighted Cohen’s κs and 95% CIs for cumulative exposure assigned to the male subcohort members (N=1455) by the different methods.

Case‐by‐case Case‐by‐case Case‐by‐case Asbestos JEM Asbestos JEM DOMJEM expert expert expert vs. DOMJEM vs. FINJEM vs. FINJEM assessment assessment assessment vs. Asbestos JEM vs. DOMJEM vs. FINJEM Asbestos 0.10 (0.05 ‐ 0.13) 0.29 (0.23 ‐ 0.32) 0.23 (0.19 ‐ 0.29) 0.25 (0.20 ‐ 0.29) 0.42 (0.38 ‐ 0.47) 0.50 (0.47 ‐ 0.55) PAHs 0.42 (0.29 ‐ 0.52) 0.40 (0.32 ‐ 0.52) 0.51 (0.44 ‐ 0.57) Welding 0.70 (0.65 ‐ 0.74) fumes *

Cumulative exposure is based on Pjob duration for the expert assessment and on PIjob duration for the JEMs. JEM, job‐exposure matrix.

For exposure to asbestos, the method showing the highest level of agreement with case‐by‐case expert assessment was the DOMJEM (Cohen’s κ=0.29, 95% CI 0.23 to 0.32), followed by the FINJEM (Cohen’s κ=0.23, 95% CI 0.19 to 0.29). The agreement between the expert assessment and the Asbestos JEM was low (Cohen’s κ=0.10, 95% CI 0.05 to 0.13). Furthermore, the comparison between DOMJEM and FINJEM revealed the highest concordance between methods (Cohen’s κ=0.50, 95% CI 0.47 to 0.55). Finally, the level of agreement with the Asbestos JEM was higher for FINJEM (Cohen’s κ=0.42, 95% CI 0.38 to 0.47) than for DOMJEM (Cohen’s κ=0.25, 95% CI 0.20 to 0.29). Patterns of agreement were comparable for exposure to PAHs. The DOMJEM showed the highest level of agreement with the case‐by‐case expert assessment (Cohen’s κ=0.42, 95% CI 0.29 to 0.52), as was the concordance between DOMJEM and FINJEM highest (Cohen’s κ=0.51, 95% CI 0.41 to 0.57).

Retrospective occupational exposure assessment

For exposure to welding fumes, the level of agreement between case‐by‐case expert assessment and FINJEM was high (Cohen’s κ=0.70, 95% CI 0.66 to 0.74). Spearman’s correlation analyses on the continuous CE measures revealed a similar relative ranking of the different methods with respect to each other, though absolute numbers were somewhat higher (results not shown).

Sensitivity analyses The agreement between the different methods when restricting analyses to exposed subjects only was overall lower than in the main analyses, with the FINJEM and Asbestos JEM revealing hardly any agreement with the case‐by‐case expert assessment for asbestos (Cohen’s κ=0.09, 95% CI 0.03 to 0.16 and Cohen’s κ=0.04, 95% CI 0.06 to 0.12, respectively). For PAHs, FINJEM showed the highest agreement with the expert assessment (Cohen’s κ=0.32, 95% CI 0.22 to 0.44) and equal agreement with the DOMJEM compared with the main analyses (Cohen’s κ=0.51, 95% CI 0.36 to 0.64). For welding fumes, agreement between the FINJEM and expert assessment was lower compared with the main analyses (Cohen’s κ=0.42, 95% CI 0.37 to 0.54) (see Table 2.3). Table 2.3

Sensitivity analysis: weighted Cohen’s κs and 95% CIs for cumulative exposure assigned by the † different methods to exposed subjects only .

Case‐by‐case Case‐by‐case expert assessment expert assessment vs. Asbestos JEM vs. DOMJEM Asbestos 0.04 (‐0.06 ‐ 0.12) 0.26 (0.13 ‐ 0.34) (n=135) PAHs 0.23 (0.11 ‐ 0.43) (n=77) Welding fumes (n=170)

Case‐by‐case Asbestos JEM Asbestos JEM DOMJEM expert assessment vs. DOMJEM vs. FINJEM vs. FINJEM vs. FINJEM 0.09 (‐0.03 ‐ 0.16) ‐0.03 (‐0.12 ‐ 0.07) 0.19 (0.12 ‐ 0.31) 0.36 (0.25 ‐ 0.47) 0.32 (0.22 ‐ 0.44)

0.51 (0.36 ‐ 0.64)

0.42 (0.37 ‐ 0.54)

Cumulative exposure is based on Pjob duration for the expert assessment and on PIjob duration for the † JEMs. Exposed (whether low, medium, or high) is based on the case‐by‐case expert assessment. JEM, job‐ exposure matrix.

When comparing the FINJEM and Asbestos JEM on P with the expert assessment, patterns of agreement were similar to the comparison on PI in the main analyses (results not shown). Also when studying concordance between the methods when using individual job codes as the unit of comparison, instead of subjects as in the main analyses, differences were negligible (results not shown). Finally, results were again largely comparable to the main analyses if the agreement was determined for specific time periods instead of the cumulative lifetime occupational exposure (results not shown). Only for the last period 1975‐1986, agreement between all methods was lower, though numbers were low because not all job codes could be included as the start and end dates of a job code did not necessarily concur with the time periods used.

Chapter 2

DISCUSSION Results of epidemiological studies should be interpreted in the light of the quality of the exposure assessment methods used and, if available, information on the validity of the assessment method should be included in the scientific report.19,20 As validation of retrospective exposure assessment is not easily accomplished because no gold standard is available, this study focused on reliability. Given the size of the study population, future exposure assessment in the NLCS will have to be based on JEMs. As reliability of a JEM may depend on the study design, exposure of interest and quality of work history/exposure information available, it is of paramount importance to apply different JEMs in the same cohort in order to evaluate consistency of findings in relation to generalisibility. Therefore, this study scored several JEMs on reliability in the NLCS by assessing agreement between JEMs and case‐by‐case expert assessment by means of Cohen’s κ and the prevalence of exposure. To this end, case‐by‐case expert assessment was available for the subcohort in the NLCS. First, results of this study show that the case‐by‐case expert assessment assigned the lowest prevalence of exposure. Second, the main analyses show variable agreement in exposure assignment, when comparing the JEMs with the expert assessment and with each other. For the comparison with the case‐by‐case expert assessment, the JEMs show poor agreement for asbestos (all JEMs), moderate agreement for PAHs (DOMJEM and FINJEM) and high agreement for welding fumes (FINJEM). For the JEM‐JEM comparisons, the agreement is moderate to good when comparing the DOMJEM and FINJEM for asbestos and PAHs, moderate when comparing the Asbestos JEM with the FINJEM and low when comparing the Asbestos JEM with the DOMJEM. Third, sensitivity analyses showed comparable results to the main analyses, except for the analyses restricted to exposed subjects only for which agreement between methods tended to be (much) lower as expected. Case‐by‐case expert assessment, when based on detailed questionnaires or interviews, is often considered the best possible method for retrospective exposure assessment in population‐based cohort studies. 2,21 Yet, studies comparing expert assessment to objective measurements in urine or the work environment show strongly varying results. The same is true for studies examining agreement between experts, which found ks from 0 to 1.0 with a median of about 0.6.2 As such, expert assessment should not generally be regarded as a gold standard. In the case‐by‐case expert assessment in this study, only limited information was available, which may have further hampered appropriate assessment of exposure patterns. However, Post et al.22 found in their study that exposure estimates hardly improved when actual measurements became available to experts and Teschke et al.2 found variable results in their review as not all

Retrospective occupational exposure assessment

studies showed an increase in the quality of the expert assessments with an increase in the amount of data available. It is of interest to note that the expert assessment in this study showed the lowest exposure prevalence. Given that occupational exposure prevalence to carcinogens is low in most population‐based studies, this would seem preferable, though lower prevalence estimates do not guarantee that the risk of false‐ positive estimates is also lower. As a recent study of Bhatti et al.23 found the opposite, that is, exposure prevalence determined by expert assessment was higher than assessed by means of a JEM, this again points out the importance of study‐specific comparison of different retrospective exposure assessment methods. The rather low agreement between the expert assessment and the Asbestos JEM could be explained by the concept behind this latter JEM, as it was originally designed to aid subjects with asbestos‐related diseases in pursuing compensation in legal trials. This JEM is based on 86 cases of asbestosis and 710 mesothelioma cases within occupations in the Netherlands.24 Since mesothelioma is almost exclusively due to asbestos exposure,25 this JEM only includes those 123 job codes with definite historical exposure to asbestos. Furthermore, administrative employees (i.e., white‐collar workers) in companies manufacturing asbestos products were also among the mesothelioma cases, an occupational group normally not expected to be exposed to asbestos. Consequently, this partly disease‐oriented JEM might be more sensitive than the two general population JEMs, which are more aimed at specificity than sensitivity because of the low exposure levels in the general population.4,7 Results were in line with this expectation as exposure prevalence was highest for the Asbestos JEM, though differences with the other JEMs were small. Lack of a time axis could also be responsible for the variable agreement,2,4,10,26 as exposure levels have generally decreased over time because of better understanding of occupational hazards and subsequent regulations.27,28 Nonetheless, the DOMJEM, which lacks a time axis, scored equally compared with the FINJEM on agreement with the expert assessment and better than the Asbestos JEM. Possibly, time trends in exposure levels were smaller compared with the difference between low and high intensity, which made inclusion of a time axis of less importance. Finally, the level of agreement between the FINJEM and case‐by‐case expert assessment might be influenced by true differences in technology and use of materials between countries.21,29 Finland had a more stringent occupational policy than the Netherlands, had an extensive asbestos textile industry as opposed to the Netherlands and had an asbestos mine that might explain the low level of agreement for asbestos between the FINJEM and expert assessment.

Chapter 2

This study had several limitations. First, no data were available on inter‐expert agreement. Therefore, although the case‐by‐case expert assessment proved to be the most stringent method in exposure assignment, no conclusions could be drawn as to the actual reliability of the expert assessment itself. Second, the NLCS is a population‐ based study with accordingly low exposure prevalence. This has consequences for calculating Cohen’s κ, yielding imprecise estimates when exposure prevalence is low.6 Though agreement in the sense of Cohen’s κ is informative about the reliability of a method, agreement depends on how exposure is defined. As expected, when restricting agreement to exposed subjects only, Cohen’s κs were lower with some comparisons revealing no agreement at all. As already known, agreement between methods is better on the dichotomy exposed versus non‐exposed than on actual levels of exposure. Moreover, as the relation of assessment by either method (case‐by‐case expert assessment or JEM) to true exposure is unknown, it is not possible to quantify the amount of exposure misclassification for each method, though misclassification will by definition be non‐differential in cohort studies. Therefore, when using either method for risk prediction, it should be kept in mind that risk estimates will be attenuated. In order to further assess the reliability of the JEMs, their predictive ability will be studied using the case‐cohort approach in a future study on exposure to asbestos and cancer risk in the NLCS.

CONCLUSION In conclusion, this study shows case‐by‐case expert assessment to assign the lowest prevalence of exposure in the NLCS. Furthermore, the DOMJEM and FINJEM proved to be rather similar in agreement when compared with this expert assessment. Finally, the Asbestos JEM appeared to be less appropriate for use in the NLCS.

Retrospective occupational exposure assessment

REFERENCES 1.

3. 4.

5. 6.

7. 8.

10. 11. 12. 13.

14. 15.

16. 17.

18.

19.

Benke G, Sim M, Forbes A, Salzberg M. Retrospective assessment of occupational exposure to chemicals in community‐based studies: validity and repeatability of industrial hygiene panel ratings. Int J Epidemiol 1997;26:635‐42. Teschke K, Olshan AF, Daniels JL, De Roos AJ, Parks CG, Schulz M, Vaughan TL. Occupational exposure assessment in case‐control studies: opportunities for improvement. Occup Environ Med 2002;59:575‐ 93; discussion 594. Goldberg M, Imbernon E. The use of job exposure matrices for cancer epidemiology research and surveillance. Arch Public Health 2002;60:173‐85. Peters S, Vermeulen R, Cassidy A, Mannetje A', van Tongeren M, Boffetta P, Straif K, Kromhout H; INCO Group. Comparison of exposure assessment methods for occupational carcinogens in a multi‐centre lung cancer case‐control study. Occup Environ Med 2011;68:148‐53. McGuire V, Nelson LM, Koepsell TD, Checkoway H, Longstreth WT Jr. Assessment of occupational exposures in community‐based case‐control studies. Annu Rev Public Health 1998;19:35‐53. Correa A, Min YI, Stewart PA, Lees PS, Breysse P, Dosemeci M, Jackson LW. Inter‐rater agreement of assessed prenatal maternal occupational exposures to lead. Birth Defects Res A Clin Mol Teratol 2006;76:811‐24. Kauppinen T, Toikkanen J, Pukkala E. From cross‐tabulations to multipurpose exposure information systems: a new job‐exposure matrix. Am J Ind Med 1998;33:409‐17. Friesen MC, Coble JB, Lu W, Shu XO, Ji BT, Xue S, Portengen L, Chow WH, Gao YT, Yang G, Rothman N, Vermeulen R. Combining a job‐exposure matrix with exposure measurements to assess occupational exposure to Benzene in a population cohort in Shanghai, China. Ann Occup Hyg 2012;56:80‐91. Peters S, Vermeulen R, Olsson A, Van Gelder R, Kendzia B, Vincent R, Savary B, Williams N, Woldbæk T, Lavoué J, Cavallo D, Cattaneo A, Mirabelli D, Plato N, Dahmann D, Fevotte J, Pesch B, Brüning T, Straif K, Kromhout H. Development of an exposure measurement database on five lung carcinogens (ExpoSYN) for quantitative retrospective occupational exposure assessment. Ann Occup Hyg 2012;56:70‐9. Kromhout H, Heederik D, Dalderup LM, Kromhout D. Performance of two general job‐exposure matrices in a study of lung cancer morbidity in the Zutphen cohort. Am J Epidemiol 1992;136:698‐711. Kromhout H, Vermeulen R. Application of job‐exposure matrices in studies of the general population: some clues to their performance. Eur Respir Rev 2001;11:80‐90. van den Brandt PA, Goldbohm RA, van 't Veer P, Volovics A, Hermus RJ, Sturmans F. A large‐scale prospective cohort study on diet and cancer in The Netherlands. J Clin Epidemiol 1990;43:285‐95. van Loon AJ, Kant IJ, Swaen GM, Goldbohm RA, Kremer AM, van den Brandt PA. Occupational exposure to carcinogens and risk of lung cancer: results from The Netherlands cohort study. Occup Environ Med 1997;54:817‐24. Petrie A, Sabin C. Medical Statistics at a Glance. Oxford: Blackwell Publishing Ltd, 2005. 105‐7. Swuste P, Dahhan M, Burdorf A. Linking expert judgement and trends in occupational exposure into a job‐exposure matrix for historical exposure to asbestos in the Netherlands. Ann Occup Hyg 2008;52:397‐403. Preller L, van den Bosch LM, van den Brandt PA, Kauppinen T, Goldbohm A. Occupational exposure to silica and lung cancer risk in the Netherlands. Occup Environ Med 2010;67:657‐63. Koeman T, Offermans NS, Christopher‐de Vries Y, Slottje P, Van Den Brandt PA, Goldbohm RA, Kromhout H, Vermeulen R. JEMs and incompatible occupational coding systems: effect of manual and automatic recoding of job codes on exposure assignment. Ann Occup Hyg 2013;57:107‐14. Stewart PA, Herrick RF, Blair A, Checkoway H, Droz P, Fine L, Fischer L, Harris R, Kauppinen T, Saracci R. Highlights of the 1990 Leesburg, Virginia, International Workshop on retrospective exposure assessment for occupational epidemiology studies. Scand J Work Environ Health 1991;17:281‐5. Tielemans E, Heederik D, Burdorf A, Vermeulen R, Veulemans H, Kromhout H, Hartog K. Assessment of occupational exposures in a general population: comparison of different methods. Occup Environ Med 1999;56:145‐51.

Chapter 2

20. Vlaanderen J, Vermeulen R, Heederik D, Kromhout H; ECNIS Integrated Risk Assessment Group, European Union Network Of Excellence. Guidelines to evaluate human observational studies for quantitative risk assessment. Environ Health Perspect 2008;116:1700‐5. 21. Mannetje A, Fevotte J, Fletcher T, Brennan P, Legoza J, Szeremi M, Paldy A, Brzeznicki S, Gromiec J, Ruxanda‐Artenie C, Stanescu‐Dumitru R, Ivanov N, Shterengorz R, Hettychova L, Krizanova D, Cassidy A, van Tongeren M, Boffetta P. Assessing exposure misclassification by expert assessment in multicenter occupational studies. Epidemiology 2003;14:585‐92. 22. Post W, Kromhout H, Heederik D, Noy D, Smit Duijzentkunst R. Semiquantitative estimates of exposure to methylene chloride and styrene: the influence of quantitative exposure data. Appl Occup Environ Hyg 1991;6:197‐204. 23. Bhatti P, Stewart PA, Linet MS, Blair A, Inskip PD, Rajaraman P. Comparison of occupational exposure assessment methods in a case‐control study of lead, genetic susceptibility and risk of adult brain tumours. Occup Environ Med 2011;68:4‐9. 24. Burdorf A, Dahhan M, Swuste P. Occupational characteristics of cases with asbestos‐related diseases in The Netherlands. Ann Occup Hyg 2003;47:485‐92. 25. Burdorf A, Swuste P. An expert system for the evaluation of historical asbestos exposure as diagnostic criterion in asbestos‐related diseases. Ann Occup Hyg 1999;43:57‐66. 26. Hawkes AP, Wilkins JR 3rd. Assessing agreement between two job‐exposure matrices. Scand J Work Environ Health 1997;23:140‐8. 27. Siemiatycki J, Richardson L, Straif K, Latreille B, Lakhani R, Campbell S, Rousseau MC, Boffetta P. Listing occupational carcinogens. Environ Health Perspect 2004;112:1447‐59. 28. Creely KS, Cowie H, Van Tongeren M, Kromhout H, Tickner J, Cherrie JW. Trends in inhalation exposureea review of the data in the published scientific literature. Ann Occup Hyg 2007;51:665‐78. 29. Kauppinen T, Heikkilä P, Plato N, Woldbaek T, Lenvik K, Hansen J, Kristjansson V, Pukkala E. Construction of job‐exposure matrices for the Nordic occupational cancer study (NOCCA). Acta Oncol 2009;48:791‐800.

Chapter 3

Occupational asbestos exposure and risk of pleural mesothelioma, lung cancer, and laryngeal cancer in the prospective Netherlands Cohort Study Nadine S.M. Offermans Roel Vermeulen Alex Burdorf R. Alexandra Goldbohm Timo Kauppinen Hans Kromhout Piet A. van den Brandt Journal of Occupational and Environmental Medicine. 2014;56:6‐19

Chapter 3

ABSTRACT Objectives To study the association between occupational asbestos exposure and pleural mesothelioma, lung cancer, and laryngeal cancer, specifically addressing risk associated with the lower end of the exposure distribution, risk of cancer subtypes, and the interaction between asbestos and smoking. Methods Using the Netherlands Cohort Study (n=58,279 men, aged 55 to 69 years), asbestos exposure was estimated by linkage to job‐exposure matrices. After 17.3 years of follow‐up, 132 pleural mesothelioma, 2324 lung cancer, and 166 laryngeal cancer cases were available. Results The multivariable‐adjusted model showed overall positive associations between all levels of asbestos exposure and mesothelioma, lung cancer, and laryngeal cancer. Lung adenocarcinoma and glottis cancer showed only a positive association after prolonged higher asbestos exposure (hazard ratio per 10 years increment, 1.43 (95% confidence interval, 1.06 to 1.93) and 1.95 (95% confidence interval, 1.36 to 2.80), respectively). There was no statistically significant interaction between asbestos and smoking. Conclusions Asbestos levels encountered at the lower end of the exposure distribution may be associated with an increased risk of pleural mesothelioma, lung cancer, and laryngeal cancer.

Occupational asbestos exposure and respiratory tract tumors

INTRODUCTION In 2006, globally, an estimated 125 million people were still occupationally exposed to asbestos1 with its use even increasing in parts of Asia, South America, and the former Soviet Union.2 In the Netherlands, despite being banned in 1993, asbestos is still a public health concern with respect to asbestos removal and site cleaning in the general environment.3 Asbestos research has been ongoing for decades and evidence has been accumulated that, regardless of fiber type, asbestos causes mesothelioma and lung, laryngeal, and ovarian cancer.4,5 Nevertheless, there remain questions around asbestos carcinogenicity that pertain to risk at the lower end of the exposure distribution,5–7 the possibility of uncontrolled confounding due to smoking and drinking for especially laryngeal cancer,8 the association with subtypes of lung9 and laryngeal cancer,8 and the interaction between asbestos and smoking in relation to lung10 and laryngeal cancer.8 Population‐based studies are well suited to address these questions given their overall wide range in exposure levels, including those at the lower end of the exposure distribution (i.e., exposure levels in jobs outside asbestos mining, insulation, cement and textile manufacturing, and other more highly exposed jobs), the possibility to control for potential confounders, and large size. The prospective Netherlands Cohort Study (NLCS) started in 1986 among 120,852 men and women of the general population.11,12 Besides questions on occupational history, the NLCS contains extensive information on dietary habits and lifestyle factors. Given the large study size and long follow‐up, many cases of asbestos‐related cancer have emerged in the NLCS. Therefore, the primary objectives of this study were to assess the following: 1. The overall association between occupational asbestos exposure and the risk of pleural mesothelioma, lung cancer, and laryngeal cancer, with special attention to risk associated with the lower end of the exposure distribution and to potential confounding 2. The association between occupational asbestos exposure and subtypes of lung and laryngeal cancer 3. The possible additive or multiplicative interaction between smoking and asbestos in relation to pleural mesothelioma, lung cancer, and laryngeal cancer Because the proportion of long‐term employed women was rather low, this study was conducted only among men. We previously evaluated several methodologies for retrospective occupational exposure assessment within the NLCS. The job‐exposure matrices (JEMs) DOMJEM and FINJEM showed rather similar agreement with case‐by‐case expert assessment and showed moderate agreement among each other.13 To provide insight into the methodological uncertainty associated with the choice of JEM, we present the main risk analyses using both DOMJEM and FINJEM.

Chapter 3

MATERIALS AND METHODS Study Population and Cancer Follow‐Up The study design and data collection strategies for the NLCS have been described in detail previously.11 In brief, the NLCS started in September 1986 when 58,279 men and 62,573 women aged 55 to 69 years, originating from 204 municipalities in the Netherlands with computerized population registries, were enrolled in the cohort. At baseline, participants completed a self‐administered questionnaire on dietary habits and lifestyle, occupational history, and other potential risk factors for cancer.11 For reasons of efficiency in questionnaire processing and follow‐up, the case–cohort approach was used.14 Incident cases were enumerated from the entire cohort, whereas the accumulated person‐years at risk in the entire cohort were estimated from a random subcohort of 5000 subjects (2411 men and 2589 women), selected immediately after baseline. This subcohort is being followed up for vital status information, while the entire cohort is being monitored for incident cancer by annual record linkage to the Netherlands Cancer Registry and the Dutch Pathology Registry (PALGA).15,16 For these analyses, a total of 17.3 years of follow‐up (baseline to December 2003) was available. Completeness of incident cancer coverage was estimated to be almost 100%.17 The NLCS was approved by the institutional review boards of the Netherlands Organisation for Applied Scientific Research TNO (Zeist) and Maastricht University (Maastricht). All prevalent cases at baseline other than skin cancer were excluded, leaving 2336 male subcohort members, 160 pleural mesothelioma cases (International Classification of Diseases for Oncology, Third Edition (ICD‐O‐3) code C384), 2932 lung cancer cases (ICDO‐3 code C34), and 216 laryngeal cancer cases (ICD‐O‐3 codes C32.0 and C32.1, which refer to cancer of the glottis and supraglottis, respectively). The reason for considering only pleural mesothelioma and cancer of the glottis and supraglottis is the very low number of cases for peritoneal mesothelioma (n=10) and cancer of the subglottis (n=4). Subjects without any, or only uncodable, information on occupational history or who never worked professionally were omitted from the analyses. As a result, 2107 male subcohort members, 145 pleural mesothelioma cases, 2592 lung cancer cases, and 184 laryngeal cancer cases were available for analyses after 17.3 years of follow‐up.

Occupational asbestos exposure and respiratory tract tumors

Standard Occupational Classification of 1984 of the Dutch Central Bureau of Statistics, supplemented by a three‐digit code assigned within the NLCS based on the job title. Subjects could enter a maximum of five occupations, which was generally sufficient to cover the lifetime occupational history for the large majority of the cohort, because cohort subjects held on average 1.9 job codes during their working life up to 1986. For all subjects, the job code was assessed for each of the maximally five occupations held between starting work and 1986. Job‐exposure matrices We applied two JEMs, the DOMJEM from the Netherlands and the Finnish FINJEM, as described previously.13 Briefly, DOMJEM is a generic JEM developed by occupational exposure experts in the Netherlands for application in general population studies. It contains a combined measure of the probability × intensity of exposure, which is semiquantitative (no, low, or high exposure) with a weighting of 0, 1, and 4, respectively.18 FINJEM was constructed for exposure assessment in large register‐based studies, is based on both expert assessment and exposure measurements, and contains a time axis.19 Although FINJEM was constructed for Finland, exposure estimates were not adapted to Dutch occupational circumstances before applying it in the NLCS. Asbestos exposure variables Several exposure variables were defined: ever versus never exposed to asbestos (yes/no), duration of exposure (years), cumulative exposure (CE; fiber‐years/ml (f‐y/ml) (FINJEM) or unit‐years (DOMJEM), see explanation hereafter), and duration of high exposure (years). The CE is a combined measure of the probability (P), intensity (I), and duration (years) of exposure. Ever versus never exposed is based on the CE, in that subjects were classified as being ever exposed to asbestos when CE>0. For occupations with P × I > 0, duration of employment was summed to obtain the duration of exposure. For DOMJEM, the CE measure was estimated by summing the product of P × I and duration over the reported occupations. The DOMJEM scores of no, low, and high exposure for P × I were arbitrarily assigned values of 0, 1, and 4, respectively, to mirror the log‐normal (multiplicative) nature of occupational exposure levels, hence the expression in unit‐years. The weighting was based on reported levels for semiquantitatively scored exposure, thereby ensuring a balanced weighting between intensity and duration in the calculation of CE.20 To arrive at the CE for FINJEM, first the P × I per job code was estimated using the time‐ specific exposure information in the time axis of this JEM, before summing P × I over the reported occupations. For those workers who started working before 1945, exposure was set to zero, because there was hardly any asbestos industry in the Netherlands in the period before 1945.

Chapter 3

For the duration of high exposure, first, the P × I per occupation was categorized into no, low, or high exposure on the basis of the distribution in the subcohort. Second, duration of employment was summed for those occupations with a high P × I to obtain the duration of high exposure. Participants were classified into never‐exposed subjects and tertiles of those exposed to asbestos on the basis of the distribution among the subcohort for the duration of exposure and CE (reference group is the never exposed) and for the duration of high exposure (reference group is the never highly exposed). Continuous variables were also used. For the duration of (high) exposure, an increment of 10 years was used; while for the CE, an increment of 1 f‐y/ml (FINJEM) or unit‐year (DOMJEM) was used.

Statistical analyses Age‐adjusted (for lung cancer, also adjusted for family history of lung cancer (yes/no)) and multivariable‐adjusted hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) were estimated by using Cox proportional hazards (PH) models. Because age is the natural time scale in studies of disease occurrence, attained age was used as date of entry in the analyses.21 In models with (attained) age as date of entry, baseline age should also be considered a covariate when there is a cohort effect or when the proportional hazards assumption is violated,22 which was the case in our analyses. Because most PH assumptions was no longer violated after adjustment for baseline age, we included baseline age as a covariate in our analyses. The total person‐years at risk were estimated from the subcohort,23 and standard errors were estimated by using a robust covariance matrix estimator to account for increased variance due to sampling the subcohort from the entire cohort.24 The covariates included in the multivariate models were either a priori selected risk factors based on the literature or variables that changed the age‐adjusted regression coefficients by at least 10% (using a backward stepwise procedure). For mesothelioma, only current smoking (yes/no), the number of cigarettes smoked per day, and years of smoking cigarettes were considered potential confounders. Although we can assume that mesothelioma is solely caused by asbestos, we added these covariates for confirmation purposes only. For lung cancer, the full covariate model consisted of current smoking (yes/no), the number of cigarettes smoked per day, years of smoking cigarettes, and occupational exposure to crystalline silica and polycyclic aromatic hydrocarbons. For laryngeal cancer, alcohol consumption was entered in the model in addition to current smoking (yes/no), the number of cigarettes smoked per day, and years of smoking cigarettes. Risk factors that were added to the model but did not satisfy the 10% rule and were not mentioned earlier were fruits and vegetables for all cancers and nickel, chromium, and welding fumes for lung cancer. All covariates were entered into the models as continuous variables, except for current smoking (yes/no). The number of cigarettes smoked per day and years of smoking

Occupational asbestos exposure and respiratory tract tumors

cigarettes were added to the model as both centered and noncentered variables because of possible problems with collinearity. As results of both analyses were comparable (not shown), we only presented the noncentered variables. To enable comparison, the age‐adjusted (for mesothelioma and laryngeal cancer) and family history of lung cancer–adjusted (for lung cancer) analyses were restricted to subjects included in the multivariable‐adjusted analyses (ie, with no missing values on confounding variables), which left 1962 subcohort members and 132 cases for pleural mesothelioma, 1962 subcohort members and 2324 cases for lung cancer, and 1931 subcohort members and 166 cases for laryngeal cancer for analyses. For each analysis, the proportional hazards assumption was tested by using the scaled Schoenfeld residuals.25 Trends for all subjects and only including the exposed were evaluated with the Wald test by assigning subjects the median value for each level of the categorical variable among the subcohort members, and this variable was entered as a continuous term in the Cox regression model. In addition, we evaluated the association with subtypes of lung cancer (n=379 for small cell lung cancer, 350 for large cell, 931 for squamous cell carcinoma, and 493 for adenocarcinoma) and laryngeal cancer (n=122 for glottis and 44 for supraglottis). Furthermore, we studied the interaction between asbestos exposure (yes/no) and smoking (no/former/current) for all three cancers both on a multiplicative and an additive scale. First, we examined whether the joint effect of asbestos and smoking was closer to additivity or multiplicativity for each cancer endpoint. Second, we tested for statistically significant departure from multiplicativity by including an interaction term in the Cox regression model. Third, we tested for statistically significant departure from additivity using the CI of the relative excess risk of cancer due to interaction, according to methods described by Knol and VanderWeele,26 which were adapted for use in Stata (Stata Corporation, College Station, TX). All analyses were performed using the Stata statistical software package (intercooled Stata, version 10). All tests were two‐tailed, and differences were regarded as statistically significant at P<0.05. Finally, population‐attributable fractions (PAFs) were calculated for all three cancers on the basis of the HRs for all asbestos exposure variables in the multivariable‐adjusted models, using the formula: pi(RRi − 1)/(pi(RRi − 1) + 1), where p is the fraction of exposed subjects in the subcohort, i indexes the exposure level, and RR values are based on HRs.27

RESULTS The distribution of asbestos exposure and potential confounders among male subcohort members and cancer cases in the NLCS is given in Table 3.1.

Chapter 3

Table 3.1

Age at baseline (yrs) (mean, SD) Family history of lung cancer (%) Cigarette smoking (%) Never Former Current Number of cigarettes per day†‡ (mean, SD) Years of smoking†‡ (yrs) (mean, SD) Alcohol consumption‡ (g/d) (mean, SD) DOMJEM Never exposed§ (%) Ever exposed§ (%) Duration of exposure¶ (yrs) T1 (median: 4) (%) T2 (median: 18) (%) T3 (median: 37) (%) Cumulative probability×intensity of exposure (unit‐years) T1 (median: 4) (%) T2 (median: 20) (%) T3 (median: 38) (%) Ever highly exposed¶ (%) Duration of high exposure¶ (yrs) T1 (median: 4) (%) T2 (median: 11) (%) T3 (median: 31) (%) FINJEM Never exposed§ (%) Ever exposed§ (%) Duration of exposure¶ (yrs) T1 (median: 7) (%) T2 (median: 25) (%) T3 (median: 37) (%) Cumulative probability×intensity of exposure (f‐y/ml) T1 (median: 0.20) (%) T2 (median: 1.58) (%) T3 (median: 6.57) (%) Ever highly exposed¶ (%) Duration of high exposure¶ (yrs) T1 (median: 6) (%) T2 (median: 20) (%) T3 (median: 35) (%) *

Distribution of potential confounders and asbestos exposure among male subcohort members and cancer cases in the NLCS, 1986‐2003. Subcohort (n=2107) n Mean/% 2107 61.3 204 9.7 264 12.5 1080 51.3 763 36.2 1722 17.1 1807 2066 1498 609 214 200 195 217 192 200 49 17 16 16 1561 546 186 178 182 182 182 182 292 102 95 95

Pleural mesothelioma cases (n=145) SD n Mean/% SD 4.2 145 61.1 4.1 n/a n/a 18 12.4 75 51.7 52 35.9 10.6 115 17.2 11.0

33.6 11.9 125 15.0 16.9 140 71.1 69 28.9 76 35.2 10 32.8 30 32.0 36 35.7 7 31.5 21 32.8 48 2.3 28 34.6 6 32.7 7 32.7 15 74.1 73 25.9 72 34.1 20 32.6 21 33.3 31 33.4 21 33.3 25 33.3 26 13.9 42 35.0 12 32.5 12 32.5 18

33.3 16.2 47.6 52.4 13.1 39.5 47.4 9.2 27.6 63.2 19.3 21.4 25.0 53.6 50.3 49.7 27.8 29.2 43.0 29.2 34.7 36.1 29.0 28.6 28.6 42.8

Lung cancer cases (n=2592) n Mean/% SD 2592 62.0 4.1 321 12.4 100 3.9 827 31.9 1663 64.2 2243 19.8 10.6

12.1 2454 16.4 2534 1760 832 232 293 307 234 269 329 110 33 27 50 1732 860 261 323 276 245 280 335 521 146 186 189

40.7 18.4 67.9 32.1 27.9 35.2 36.9 28.1 32.3 39.6 4.2 30.0 24.6 45.4 66.8 33.2 30.3 37.6 32.1 28.5 32.6 38.9 20.1 28.0 35.7 36.3

Laryngeal cancer cases (n=184) n Mean/% SD 184 61.8 3.9 n/a n/a 4 2.2 63 34.2 117 63.6 166 18.3 9.7

9.2 180 19.4 180 127 57 17 16 24 15 18 24 9 1 2 6 130 54 20 18 16 17 19 18 29 8 9 12

39.6 10.2 22.3 25.8 69.0 31.0 29.8 28.1 42.1 26.3 31.6 42.1 4.9 11.1 22.2 66.7 70.7 29.3 37.1 33.3 29.6 31.5 35.2 33.3 15.8 27.6 31.0 41.4

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. ‡ Among former and current smokers only. Sum of categories deviates from total number because of § ¶ missing values. Exposure based on the cumulative probability×intensity of exposure (unit‐years or f‐y/ml). Exposure based on the probability×intensity of exposure (unit‐years or f‐y/ml) per job. f‐y/ml, fiber‐years/ml; n/a, not applicable; NLCS, Netherlands Cohort Study. †

Occupational asbestos exposure and respiratory tract tumors

Lung cancer cases more often reported a family history of lung cancer. In addition, they more often smoked cigarettes, and smoked more cigarettes per day and for more years than subcohort members, as did the laryngeal cancer cases. Cases for all three cancers consumed more alcohol per day and were, on average, more often and longer (highly) exposed to asbestos. Although FINJEM generally revealed a somewhat lower percentage of ever‐exposed subjects, DOMJEM showed the lowest percentage of highly exposed subjects. When studying the association between asbestos exposure and covariates in the subcohort (Table 3.2), subjects exposed to asbestos more often smoked cigarettes and for more years, with the JEMs showing different patterns over the tertiles. Alcohol consumption was higher for those never exposed to asbestos than for those exposed, as classified according to both JEMs. There was an association between exposure to silica and polycyclic aromatic hydrocarbons and asbestos exposure, with DOMJEM showing an increasing pattern over the tertiles of exposure, while the pattern for FINJEM was less clear. Overall, all asbestos exposure variables were positively associated with risk of mesothelioma, lung cancer, and laryngeal cancer, using both DOMJEM and FINJEM, and for the age‐adjusted (mesothelioma and laryngeal cancer) and for family history of lung cancer–adjusted (yes/no) (lung cancer) models as well as the multivariable‐adjusted models (Tables 3.3 to 3.5). Adjusting for potential confounders had no influence on the association with mesothelioma, and little to no effect on the association with lung and laryngeal cancer. Therefore, only multivariable‐adjusted results are mentioned in the text hereafter. Overall, DOMJEM revealed a higher HR (95% CI) for the duration of high exposure (tertile 3 vs. never: HR=13.66 (95% CI, 5.86 to 31.84) for mesothelioma; HR=2.99 (95% CI, 1.39 to 6.41) for lung cancer; and HR=6.36 (95% CI, 2.18 to 18.53) for laryngeal cancer) than FINJEM (HR=3.28 (95% CI, 1.82 to 5.92) for mesothelioma; HR=1.74 (95% CI, 1.20 to 2.54) for lung cancer; and HR=1.49 (95% CI, 0.75 to 2.97) for laryngeal cancer). FINJEM showed a higher HR (95% CI) for ever versus never exposed (HR=3.02 (95% CI, 2.11 to 4.34) for mesothelioma; HR=1.50 (95% CI, 1.27 to 1.78) for lung cancer; and HR=1.42 (95% CI, 0.99 to 2.03) for laryngeal cancer) than DOMJEM (HR=2.62 (95% CI, 1.82 to 3.76] for mesothelioma; HR=1.19 (95% CI, 1.02 to 1.40) for lung cancer; and HR=1.20 (95% CI, 0.84 to 1.72) for laryngeal cancer). For mesothelioma (Table 3.3), associations generally reached statistical significance and showed a clear dose–response relation when including the never‐exposed subjects (Ptrend <0.001). When only including the exposed subjects, only DOMJEM showed a significant trend (Ptrend <0.001).

Chapter 3

Table 3.2

Association between category of asbestos exposure and potential confounders in the subcohort in the NLCS, 1986‐2003.

Characteristic DOMJEM Age at baseline (yrs) (mean) Family history of lung cancer (%) Cigarette smoking (%) Never Former Current Number of cigarettes per day* (mean) Years of smoking* (yrs) (mean) Alcohol consumption (g/d) (mean) Silica† (%) Never exposed T1 (Median: 7) T2 (Median: 27) T3 (Median: 47.5) PAHs† (%) Never exposed T1 (Median: 4) T2 (Median: 18.5) T3 (Median: 36) FINJEM Age at baseline (yrs) (mean) Family history of lung cancer (%) Cigarette smoking (%) Never Former Current Number of cigarettes per day* (mean) Years of smoking* (yrs) (mean) Alcohol consumption (g/d) (mean) Silica† (%) Never exposed T1 (Median: 0.33) T2 (Median: 1.45) T3 (Median: 4.80) PAHs† (%) Never exposed T1 (Median: 0.08) T2 (Median: 0.27) T3 (Median: 0.77) *

Asbestos exposure based on the cumulative probability×intensity Never exposed T1 T2 T3 Median: Median: Median: 4 unit‐years 20 unit‐years 38 unit‐years 61.4 61.1 61.3 60.9 9.9 9.7 10.4 7.5 14.0 8.8 9.9 8.0 50.9 55.8 48.4 51.5 35.1 35.4 41.7 40.5 17.1 17.4 16.6 16.7 33.5 32.6 34.8 34.5 15.5 14.1 12.9 15.0 88.4 78.4 82.3 86.5 3.2 16.1 3.1 3.5 4.2 4.6 9.4 4.5 4.2 0.9 5.2 5.5 96.2 81.6 82.8 79.5 0.8 14.7 3.7 4.5 1.5 2.8 10.9 3.5 1.5 0.9 2.6 12.5 Median: Median: Median: 0.20 f‐y/ml 1.58 f‐y/ml 6.57 f‐y/ml 61.4 61.1 60.4 61.0 10.1 8.2 13.2 4.4 13.4 12.1 6.6 11.0 51.4 53.3 55.5 44.0 35.2 34.6 37.9 45.0 17.0 16.0 18.5 16.6 33.4 32.8 34.2 36.0 15.7 12.0 14.9 12.9 98.8 81.3 68.7 41.2 0.1 14.8 17.6 6.6 0.4 1.7 9.3 25.3 0.7 2.2 4.4 26.9 97.8 80.2 58.8 68.2 0.5 6.6 13.2 14.8 0.8 3.3 18.1 7.7 0.9 9.9 9.9 9.3 †

Among former and current smokers only. Exposure based on the cumulative probability×intensity of exposure (unit‐years or f‐y/ml), categorized in never‐exposed and tertiles of exposed in the subcohort. f‐y/ml, fiber‐years/ml; NLCS, Netherlands Cohort Study; PAHs, polycyclic aromatic hydrocarbons; T, tertile.

Table 3.3

Occupational asbestos exposure and respiratory tract tumors

Hazard ratios (HRs), 95%CIs and PAFs for pleural mesothelioma for categories of asbestos * † exposure , estimated with DOMJEM and FINJEM in the NLCS, 1986‐2003 .

DOMJEM Never exposed¶ Ever exposed¶ Duration of exposure# (yrs) T1 (median: 4) T2 (median: 18) T3 (median: 37) P for trend (over the exposed only) Continuous, per 10 yrs Cumulative probability×intensity of exposure (unit‐years) T1 (median: 4) T2 (median: 20) T3 (median: 38) P for trend (over the exposed only) Continuous, per 1 unit‐year Duration of high exposure# (yrs) Never highly exposed T1 (median: 4) T2 (median: 11) T3 (median: 31) P for trend (over the exposed only) Continuous, per 10 yrs FINJEM Never exposed¶ Ever exposed¶ Duration of exposure# (yrs) T1 (median: 7) T2 (median: 25) T3 (median: 37) P for trend (over the exposed only) Continuous, per 10 yrs Cumulative probability×intensity of exposure (f‐y/ml) T1 (median: 0.20) T2 (median: 1.58) T3 (median: 6.57) P for trend (over the exposed only) Continuous, per 1 fiber‐year Duration of high exposure# (yrs) Never highly exposed T1 (median: 6) T2 (median: 20) T3 (median: 35) P for trend (over the exposed only) Continuous, per 10 yrs

Person y- ears No. of in subcohort cases 107557 41688 15469 12615 13604 149245 15580 11991 14117 149245 145705 1257 1180 1103 149245 112078 37167 13127 11759 12281 149245 12440 12541 12187 149245 129206 7217 6375 6447 149245

66 66 10 26 30 132 7 19 40 132 110 6 6 10 132 67 65 19 19 27 132 20 22 23 132 95 11 11 15 132

Pleural mesothelioma HR‡ (95%CI) HR§ (95%CI)

Population attributable fraction (PAF) (%) 1 (ref) 1 (ref) 2.61 (1.83‐3.73) 2.62 (1.82‐3.76) 31.9 1.02 (0.52‐2.02) 1.03 (0.52‐2.02) 3.55 (2.19‐5.78) 3.56 (2.18‐5.81) 33.1 3.67 (2.32‐5.80) 3.69 (2.31‐5.87) <0.001 <0.001 (<0.001) 1.46 (1.31‐1.63) 1.46 (1.31‐1.63) 0.71 (0.32‐1.56) 0.71 (0.32‐1.56) 2.78 (1.62‐4.78) 2.80 (1.62‐4.83) 32.7 4.69 (3.07‐7.15) 4.71 (3.06‐7.26) <0.001 <0.001 (<0.001) 1.02 (1.02‐1.03) 1.02 (1.02‐1.03) 1 (ref) 1 (ref) 6.82 (2.57‐18.07) 6.85 (2.58‐18.15) 7.26 (2.71‐19.45) 7.19 (2.66‐19.40) 15.9 13.61 (5.91‐31.35) 13.66 (5.86‐31.84) <0.001 <0.001 (<0.001) 2.25 (1.73‐2.93) 2.26 (1.74‐2.94) 1 (ref) 1 (ref) 3.02 (2.11‐4.32) 3.02 (2.11‐4.34) 34.3 2.47 (1.45‐4.21) 2.47 (1.45‐4.20) 2.88 (1.68‐4.94) 2.88 (1.67‐4.95) 34.4 3.73 (2.31‐6.01) 3.74 (2.31‐6.06) <0.001 <0.001 (0.167) 1.42 (1.27‐1.58) 1.42 (1.27‐1.58) 2.69 (1.60‐4.52) 2.69 (1.60‐4.53) 3.04 (1.82‐5.10) 3.04 (1.80‐5.12) 34.5 3.36 (2.03‐5.57) 3.37 (2.03‐5.59) <0.001 <0.001 (0.441) 1.08 (1.05‐1.12) 1.08 (1.04‐1.12) 1 (ref) 1 (ref) 2.14 (1.09‐4.20) 2.13 (1.08‐4.20) 2.46 (1.26‐4.79) 2.45 (1.25‐4.78) 18.3 3.28 (1.81‐5.92) 3.28 (1.82‐5.92) <0.001 <0.001 (0.436) 1.41 (1.22‐1.63) 1.41 (1.22‐1.63)

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. 47 Analyses have also been performed using the Asbestos JEM. Hazard ratios for mesothelioma were somewhat lower than for DOMJEM and FINJEM as might have been expected based on a previous study comparing expert and job‐exposure matrix‐based retrospective exposure assessment of occupational 13 ‡ § carcinogens. . Age‐adjusted model. Adjusted for age (yrs), current smoking (yes/no), number of cigarettes ¶ smoked per day, and years of smoking cigarettes. Exposure based on the cumulative probability×intensity of # exposure (unit‐years or f‐y/m ). Exposure based on the probability×intensity of exposure (unit‐years or f‐ y/ml) per job. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; PAF, population‐attributable fraction; ref, reference. †

Chapter 3

For lung cancer (Table 3.4), not all associations were statistically significant, though tests for trend were significant when including the never‐exposed subjects (Ptrend <0.05). When only the exposed subjects were considered, trends were no longer significant. Results by histology of lung cancer were fairly comparable to overall lung cancer apart from adenocarcinoma, for which associations with most exposure variables were weaker or absent. The only exception was the duration of high exposure for DOMJEM, which showed a statistically significant association for the continuous variable (HR=1.43 (95% CI, 1.06 to 1.93)) and positive trend (Ptrend =0.047) for risk of adenocarcinoma. For laryngeal cancer (Table 3.5), risk by location showed usually stronger associations for supraglottis than glottis cancer, except for the duration of high exposure for DOMJEM, which was statistically significant for both the highest tertile (HR=7.09 (95% CI, 2.31 to 21.74)) and the continuous variable, as was the dose–response relation (Ptrend =0.002). The joint effect of asbestos and smoking for each cancer endpoint was assessed by comparing the HRs of current smokers not exposed to asbestos and never‐smokers exposed to asbestos with the HR of current smokers exposed to asbestos (Table 3.6). Given space limitations and because HRs for ever versus never exposed were overall higher for FINJEM than for DOMJEM, we will only present results for FINJEM. For mesothelioma, the observed HR of 4.18 for the combined exposure category is lower than the product of HRs for asbestos and smoking (5.39×1.35=7.28), indicating that the joint effect is less than multiplicative. The observed HR is also lower than expected when using an additive model (5.39+1.35−1=5.74). For lung cancer, the observed HR of 10.21 is lower than the product of asbestos and smoking (7.48×1.79=13.39), indicating that the joint effect is less than multiplicative. The observed HR is higher than what would have been expected by using an additive model (7.48+1.79−1=8.27). For laryngeal cancer, the observed HR of 20.73 is lower than the product (16.95×2.27=38.48), indicating that the joint effect is considerably less than multiplicative. The observed HR is somewhat higher than what would have been expected by using an additive model (16.95+2.27−1=18.22). When testing for departure from multiplicativity (by including an interaction term) or additivity (using the CI of relative excess risk of cancer due to interaction), there was no statistically significant multiplicative or additive interaction between asbestos and smoking for any of the cancers. Relative excess risk of cancer due to interaction and 95% CIs are presented in Table 3.6.

Person years in sub‐ cohort 107450 41644 15455 12587 13602 149094 15566 11963 14114 149094 145556 1256 1179 1103 149094

DOMJEM § Never exposed § Ever exposed ¶ Duration of exposure (yrs) T1 (median: 4) T2 (median: 18) T3 (median: 37) P for trend (over the exposed only) Continuous, per 10 yrs Cumulative probability × intensity of exposure (unit‐years) T1 (median: 4) T2 (median: 20) T3 (median: 38) P for trend (over the exposed only) Continuous, per 1 unit‐year ¶ Duration of high exposure (yrs) Never highly exposed T1 (median: 4) T2 (median: 11) T3 (median: 31) P for trend (over the exposed only) Continuous, per 10 yrs

1581 743 217 260 266 2324 217 243 283 2324 2229 30 24 41 2324

No. of cases

1 (ref) 1.25 (1.09‐1.43) 0.94 (0.76‐1.16) 1.53 (1.23‐1.90) 1.37 (1.11‐1.70) < 0.001 1.12 (1.06‐1.17) 0.93 (0.75‐1.14) 1.52 (1.21‐1.90) 1.42 (1.15‐1.74) < 0.001 1.01 (1.00‐1.01) 1 (ref) 1.73 (0.91‐3.28) 1.55 (0.81‐2.98) 2.84 (1.49‐5.39) < 0.001 1.44 (1.17‐1.75)

1 (ref) 1.19 (1.02‐1.40) 0.93 (0.73‐1.18) 1.54 (1.20‐1.98) 1.22 (0.95‐1.57) 0.008 (0.101) 1.09 (1.02‐1.15) 0.91 (0.72‐1.16) 1.55 (1.20‐2.00) 1.26 (0.99‐1.61) 0.005 (0.052) 1.01 (1.00‐1.01) 1 (ref) 1.97 (0.93‐4.17) 1.34 (0.66‐2.74) 2.99 (1.39‐6.41) 0.002 (0.313) 1.46 (1.15‐1.87)

HR (95%CI)

‡

Lung cancer HR (95%CI)

†

Population No. attributable of fraction cases (PAF) (%) 257 5.2 122 38 6.1 40 44 379 37 6.1 36 49 379 359 4 2.5 3 13 379 1 (ref) 1.19 (0.91‐1.55) 0.98 (0.65‐1.47) 1.46 (0.97‐2.20) 1.21 (0.80‐1.83) 0.132 (0.453) 1.09 (0.99‐1.20) 0.94 (0.62‐1.41) 1.42 (0.93‐2.16) 1.32 (0.89‐1.95) 0.082 (0.211) 1.01 (1.01‐1.02) 1 (ref) 1.59 (0.46‐5.44) 1.02 (0.28‐3.77) 6.38 (2.42‐16.78) 0.001 (0.074) 1.79 (1.30‐2.45)

HR (95%CI)

‡

Small cell lung cancer

231 119 30 45 44 350 27 46 46 350 334 7 5 4 350

1 (ref) 1.29 (0.99‐1.68) 0.88 (0.57‐1.35) 1.83 (1.25‐2.68) 1.34 (0.89‐2.02) 0.018 (0.106) 1.13 (1.03‐1.24) 0.77 (0.50‐1.20) 2.03 (1.38‐2.97) 1.36 (0.91‐2.04) 0.010 (0.039) 1.01 (1.00‐1.02) 1 (ref) 2.96 (1.07‐8.22) 1.93 (0.64‐5.82) 1.51 (0.33‐6.82) 0.119 (0.661) 1.19 (0.78‐1.82)

622 309 82 113 114 931 87 108 114 931 898 13 6 14 931

1 (ref) 1.29 (1.07‐1.57) 0.91 (0.67‐1.23) 1.73 (1.29‐2.32) 1.40 (1.04‐1.89) 0.001 (0.032) 1.12 (1.04‐1.20) 0.95 (0.70‐ 1.27) 1.76 (1.30‐ 2.37) 1.35 (1.00‐ 1.82) 0.002 (0.070) 1.01 (1.00‐1.02) 1 (ref) 2.03 (0.84‐4.93) 0.86 (0.31‐2.34) 2.87 (1.19‐6.88) 0.027 (0.631) 1.43 (1.07‐1.91)

354 139 45 47 47 493 46 41 52 493 475 3 7 8 493

1 (ref) 0.96 (0.75‐1.22) 0.83 (0.58‐1.20) 1.20 (0.83‐1.73) 0.90 (0.61‐1.33) 0.865 (0.828) 1.00 (0.92‐1.10) 0.84 (0.58‐1.20) 1.14 (0.77‐1.68) 0.97 (0.67‐1.39) 0.961 (0.630) 1.00 (1.00‐1.01) 1 (ref) 0.99 (0.27‐3.61) 1.73 (0.66‐4.56) 2.36 (0.88‐6.29) 0.047 (0.769) 1.43 (1.06‐1.93)

Large cell lung cancer Squamous cell Adenocarcinoma carcinoma ‡ ‡ ‡ No. HR (95%CI) No. HR (95%CI) No. HR (95%CI) of of of cases cases cases

Hazard ratios (HRs), 95%CIs and PAFs for overall lung cancer and subtypes for categories of asbestos exposure*, estimated with DOMJEM and FINJEM in the NLCS, 1986‐2003.

Table 3.4

Occupational asbestos exposure and respiratory tract tumors

1 (ref) 1.57 (1.37‐1.81) 1.39 (1.12‐1.72) 1.88 (1.50‐2.34) 1.49 (1.19‐1.85) <0.001 1.16 (1.10‐1.21) 1.32 (1.06‐1.65) 1.51 (1.21‐1.88) 1.91 (1.54‐2.38) <0.001 1.06 (1.03‐1.08) 1 (ref) 1.35 (1.02‐1.79) 1.96 (1.47‐2.62) 1.85 (1.39‐2.46) <0.001 1.22 (1.14‐1.31)

1558 766 242 281 243 2324 222 251 293 2324 1865 132 162 165 2324

†

HR (95%CI)

1 (ref) 1.50 (1.27‐1.78) 1.47 (1.15‐1.87) 1.58 (1.21‐2.07) 1.46 (1.12‐1.90) <0.001(0.977) 1.13 (1.06‐1.20) 1.44 (1.12‐1.86) 1.40 (1.09‐1.79) 1.76 (1.30‐2.38) <0.001(0.223) 1.04 (1.01‐1.07) 1 (ref) 1.23 (0.90‐1.69) 1.76 (1.23‐2.51) 1.74 (1.20‐2.54) <0.001(0.126) 1.19 (1.08‐1.31)

‡

HR (95%CI)

Lung cancer

No. of cases

Population No. attributable of fraction cases (PAF) (%) 256 11.5 123 30 11.5 52 41 379 20 12.1 46 57 379 296 23 7.3 30 30 379 1 (ref) 1.56 (1.18‐2.05) 1.16 (0.75‐1.79) 1.93 (1.29‐2.87) 1.64 (1.08‐2.50) 0.001 (0.194) 1.19 (1.08‐1.31) 0.81 (0.49‐1.36) 1.68 (1.14‐2.46) 2.60 (1.65‐4.09) <0.001 (<0.001) 1.07 (1.02‐1.11) 1 (ref) 1.41 (0.85‐2.32) 2.32 (1.35‐3.98) 2.52 (1.47‐4.31) <0.001 (0.053) 1.34 (1.17‐1.54)

‡

HR (95%CI)

Small cell lung cancer

229 121 32 44 45 350 32 39 50 350 277 15 31 27 350

No. of cases 1 (ref) 1.62 (1.23‐2.15) 1.33 (0.87‐2.04) 1.78 (1.17‐2.69) 1.82 (1.21‐2.74) <0.001 (0.357) 1.20 (1.09‐1.31) 1.41 (0.92‐2.15) 1.50 (1.00‐2.26) 2.16 (1.35‐3.47) <0.001 (0.130) 1.03 (0.99‐1.07) 1 (ref) 0.97 (0.53‐1.77) 2.40 (1.43‐4.02) 2.00 (1.13‐3.55) 0.001 (0.111) 1.26 (1.09‐1.46)

604 327 109 119 99 931 113 109 105 931 755 55 57 64 931

1 (ref) 1.68 (1.36‐2.06) 1.71 (1.27‐2.28) 1.77 (1.29‐2.43) 1.53 (1.10‐2.11) <0.001 (0.592) 1.15 (1.07‐1.24) 1.94 (1.45‐2.60) 1.53 (1.13‐2.06) 1.52 (1.04‐2.22) <0.001 (0.377) 1.02 (0.99‐1.06) 1 (ref) 1.22 (0.82‐1.80) 1.47 (0.95‐2.27) 1.54 (0.97‐2.42) 0.018 (0.488) 1.13 (1.01‐1.28)

359 134 47 45 42 493 40 40 54 493 409 23 32 29 493

1 (ref) 1.09 (0.84‐1.42) 1.19 (0.82‐1.72) 1.07 (0.71‐1.61) 0.99 (0.66‐1.50) 0.792 (0.568) 1.01 (0.92‐1.11) 1.08 (0.73‐1.60) 0.96 (0.65‐1.41) 1.31 (0.83‐2.08) 0.398 (0.313) 1.05 (1.00‐1.09) 1 (ref) 1.00 (0.61‐1.64) 1.52 (0.90‐2.55) 1.37 (0.77‐2.43) 0.140 (0.207) 1.11 (0.96‐1.29)

Squamous cell Adenocarcinoma carcinoma ‡ ‡ ‡ HR (95%CI) No. HR (95%CI) No. HR (95%CI) of of cases cases

Large cell lung cancer

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. † Adjusted for age (yrs) and family history of lung cancer (yes/no). ‡ Adjusted for age (yrs), family history of lung cancer (yes/no), current smoking (yes/no), number of cigarettes smoked per day, years of smoking cigarettes, and occupational exposure to silica and PAHs (based on cumulative probability×intensity of exposure (unit‐years or f‐y/ml)). § Exposure based on the cumulative probability×intensity of exposure (unit‐years or f‐y/ml). ¶ Exposure based on the probability×intensity of exposure (unit‐years or f‐y/ml) per job. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; PAF, population‐attributable fraction; ref, reference.

Person years in sub‐ cohort FINJEM § Never exposed 111973 § Ever exposed 37121 ¶ Duration of exposure (yrs) T1 (median: 7) 13117 T2 (median: 25) 11894 T3 (median: 37) 12111 P for trend (over the exposed only) Continuous, per 10 yrs 149094 Cumulative probability × intensity of exposure (f‐y/ml) T1 (median: 0.20) 12433 T2 (median: 1.58) 12535 T3 (median: 6.57) 12152 P for trend (over the exposed only) Continuous, per 1 fiber‐year 149094 ¶ Duration of high exposure (yrs) Never highly exposed 129095 T1 (median: 6) 7211 T2 (median: 20) 6344 T3 (median: 35) 6444 P for trend (over the exposed only) Continuous, per 10 yrs 149094

(continued)

Table 3.4

Chapter 3

DOMJEM § Never exposed § Ever exposed Duration of exposure¶ (yrs) T1 (median: 4) T2 (median: 18) T3 (median: 37) P for trend (over the exposed only) Continuous, per 10 yrs Cumulative probability × intensity of exposure (unit‐years) T1 (median: 4) T2 (median: 20) T3 (median: 38) P for trend (over the exposed only) Continuous, per 1 unit‐year ¶ Duration of high exposure (yrs) Never highly exposed T1 (median: 4) T2 (median: 11) T3 (median: 31) P for trend (over the exposed only) Continuous, per 10 yrs

105823 41025 15110 12389 13526 146848 15150 11836 14040 146848 143376 1189 1180 1103 146848

Person years in subcohort 114 52 5 14 23 166 13 16 23 166 157 1 2 6 166

No. of cases

Glottis cancer ‡ Population No. of HR (95%CI) attributable fraction cases (PAF) (%) 1 (ref) 90 1 (ref) 1.20 (0.84‐1.72) 5.5 32 0.94 (0.61‐1.44) 0.89 (0.50‐1.58) 10 0.76 (0.39‐1.50) 1.20 (0.66‐2.18) 5.7 5 0.54 (0.21‐1.37) 1.57 (0.95‐2.58) 17 1.46 (0.83‐2.56) 0.107 (0.101) 0.652 (0.111) 1.12 (0.99‐1.27) 122 1.06 (0.91‐1.24) 0.76 (0.41‐1.40) 8 0.61 (0.29‐1.28) 1.46 (0.83‐2.59) 6.1 8 0.91 (0.42‐1.94) 1.51 (0.92‐2.49) 16 1.33 (0.75‐2.36) 0.086 (0.057) 0.608 (0.065) 1.01 (1.01‐1.02) 122 1.01 (1.00‐1.02) 1 (ref) 115 1 (ref) 1.00 (0.12‐7.94) 1 1.33 (0.17‐10.57) 1.53 (0.33‐7.02) 4.2 1 1.11 (0.14‐8.72) 6.36 (2.18‐18.53) 5 7.09 (2.31‐21.74) 0.002 (0.113) 0.002 (0.408) 1.87 (1.32‐2.65) 122 1.95 (1.36‐2.80)

Laryngeal cancer ‡ HR (95%CI)

1 (ref) 1.20 (0.85‐1.70) 0.91 (0.52‐1.58) 1.12 (0.63‐2.01) 1.61 (1.00‐2.60) 0.091 1.13 (1.00‐1.27) 0.78 (0.43‐1.40) 1.36 (0.78‐2.37) 1.56 (0.97‐2.51) 0.071 1.01 (1.01‐1.02) 1 (ref) 0.84 (0.11‐6.40) 1.71 (0.39‐7.57) 5.49 (2.05‐14.71) 0.002 1.77 (1.30‐2.42)

†

HR (95%CI) 24 20 5 9 6 44 5 8 7 44 42 0 1 1 44

1 (ref) 2.14 (1.13‐4.04) 1.25 (0.44‐3.58) 3.79 (1.69‐8.54) 1.99 (0.77‐5.13) 0.009 (0.522) 1.27 (1.05‐1.53) 1.23 (0.42‐3.58) 3.70 (1.58‐8.65) 2.22 (0.90‐5.46) 0.007 (0.419) 1.02 (1.00‐1.03) 1 (ref) ‐ 2.53 (0.32‐19.79) 4.24 (0.53‐33.70) 0.153 (<0.001) 1.56 (0.82‐2.98)

Supraglottis cancer ‡ No. of HR (95%CI) cases

Hazard ratios (HRs), 95%CIs and PAFs for overall laryngeal cancer and subtypes for categories of asbestos exposure*, estimated with DOMJEM and FINJEM in the NLCS, 1986‐2003.

Table 3.5

Occupational asbestos exposure and respiratory tract tumors

56 110575 36273 12760 11383 12130 146848 11922 12392 11959 146848 127107 7148 6298 6295 146848

Person years in subcohort 115 51 20 17 14 166 17 18 16 166 139 8 8 11 166

No. of cases

Glottis cancer Supraglottis cancer Population No. of HR‡ (95%CI) No. of HR‡ (95%CI) attributable fraction cases cases (PAF) (%) 1 (ref) 91 1 (ref) 24 1 (ref) 1.42 (0.99‐2.03) 9.8 31 1.09 (0.71‐1.68) 20 2.66 (1.41‐5.03) 1.75 (1.03‐2.97) 13 1.43 (0.77‐2.66) 7 3.10 (1.23‐7.84) 1.39 (0.80‐2.42) 10.0 8 0.84 (0.39‐1.80) 9 3.31 (1.45‐7.60) 1.14 (0.62‐2.07) 10 1.02 (0.51‐2.03) 4 1.58 (0.53‐4.75) 0.224 (0.256) 0.987 (0.525) 0.011 (0.255) 1.07 (0.95‐1.21) 122 1.00 (0.85‐1.17) 44 1.24 (1.03‐1.49) 1.59 (0.91‐2.79) 12 1.41 (0.74‐2.69) 5 2.31 (0.82‐6.47) 1.44 (0.83‐2.48) 10.0 12 1.21 (0.64‐2.29) 6 2.35 (0.88‐6.24) 1.25 (0.71‐2.21) 7 0.70 (0.31‐1.57) 9 3.18 (1.43‐7.10) 0.143 (0.463) 0.803 (0.171) 0.001 (0.710) 1.00 (0.96‐1.05) 122 0.97 (0.91‐1.04) 44 1.04 (0.99‐1.10) 1 (ref) 107 1 (ref) 32 1 (ref) 1.05 (0.49‐2.28) 4 0.70 (0.25‐1.96) 4 2.23 (0.72‐6.96) 1.12 (0.52‐2.41) 2.9 4 0.75 (0.27‐2.10) 4 2.20 (0.74‐6.55) 1.49 (0.75‐2.97) 7 1.23 (0.54‐2.82) 4 2.39 (0.81‐7.11) 0.272 (0.363) 0.994 (0.230) 0.025 (0.914) 1.11 (0.94‐1.32) 122 1.02 (0.82‐1.27) 44 1.32 (1.03‐1.68)

Laryngeal cancer HR‡ (95%CI)

1 (ref) 1.41 (0.99‐2.00) 1.56 (0.94‐2.59) 1.55 (0.90‐2.65) 1.13 (0.63‐2.02) 0.171 1.08 (0.96‐1.22) 1.39 (0.81‐2.39) 1.46 (0.86‐2.49) 1.36 (0.79‐2.36) 0.080 1.01 (0.97‐1.06) 1 (ref) 1.09 (0.52‐2.33) 1.23 (0.58‐2.61) 1.63 (0.84‐3.14) 0.138 1.16 (0.98‐1.36)

HR† (95%CI)

* Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. † Age‐adjusted model. ‡ Adjusted for age (yrs), current smoking (yes/no), number of cigarettes smoked per day, years of smoking cigarettes, and alcohol consumption (g/d). § Exposure based on the cumulative probability×intensity of exposure (unit‐years or f‐y/ml). ¶ Exposure based on the probability×intensity of exposure (unit‐years or f‐y/ml) per job. CI, confidence interval; f‐y/ml, fiber‐ years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; PAF, population‐attributable fraction; ref, reference.

FINJEM § Never exposed Ever exposed§ ¶ Duration of exposure (yrs) T1 (median: 7) T2 (median: 25) T3 (median: 37) P for trend (over the exposed only) Continuous, per 10 yrs Cumulative probability × intensity of exposure (f‐y/ml) T1 (median: 0.20) T2 (median: 1.58) T3 (median: 6.57) P for trend (over the exposed only) Continuous, per 1 fiber‐year Duration of high exposure¶ (yrs) Never highly exposed T1 (median: 6) T2 (median: 20) T3 (median: 35) P for trend (over the exposed only) Continuous, per 10 yrs

(continued)

Table 3.5

Chapter 3

8 10 71 29 2 1

16349 4112 16349 4113 16098 3904

Never Person‐ No. of years cases

58484 19520

40 18

508 261 5.85 (1.40‐24.49) 6.53 (1.50‐28.51) 1.12 (0.63‐1.98)

2.15 (1.60‐2.90) 3.42 (2.45‐4.76) 1.52 (1.22‐1.91)

1.49 (0.68‐3.24) 3.82 (1.71‐8.51) 2.61 (1.60‐4.26)

35145 14681

37140 13488

37212 13531

72 33

979 476

20 23

Current Person No. of ‐years cases

7.48 (5.55‐10.08) 5.57 (3.38‐9.16)

HR (95%CI) for current cigarette smoking within strata of asbestos exposure 1.35 (0.58‐3.12) 0.78 (0.35‐1.73)

0.41 0.50 0.85

P for multiplicative interaction

16.95 (4.10‐70.03) 5.85 (1.40‐24.49) 16.95 (4.10‐70.03) 20.73 (4.89‐87.91) 2.67 (0.34‐20.90) 8.56 (1.09‐67.43) 1.23 (0.78‐1.94)

7.48 (5.55‐10.08) 2.15 (1.60‐2.90) 10.21 (7.26‐14.35) 1.86 (1.13‐3.06) 1.39 (1.10‐1.76)

HR (95%CI) for former cigarette smoking within strata of asbestos exposure 1.35 (0.58‐3.12) 1.49 (0.68‐3.24) 4.18 (1.79‐9.75) 0.72 (0.33‐1.56) 3.07 (1.64‐5.76) HR (95%CI)

Measure of interaction on additive scale: Relative excess risk due to interaction, RERI (95%CI) Former cigarette smoking Current cigarette smoking ‐2.06 (‐10.53‐1.42) ‐1.56 (‐10.42‐2.27) 0.47 (‐0.65‐1.55) 1.94 (‐0.13‐4.89) ¶ ¶ ‐ ‐

§ 1 (ref) 54580 2.27 (0.20‐25.52) 22441 2.27 (0.20‐25.52)

1 (ref)† 1.79 (1.04‐3.08) 1.79 (1.04‐3.08)

39 32

Cigarette smoking Former Person No. of HR (95%CI) ‐years cases

* 1 (ref) 58517 5.39 (2.02‐14.39) 19525 5.39 (2.02‐14.39)

HR (95%CI)

Age‐adjusted model. † Adjusted for age (yrs), family history of lung cancer (yes/no), and occupational exposure to silica and PAHs (based on cumulative probability×intensity of exposure (f‐y/ml)). ‡ Based on DOMJEM due to an empty cell for FINJEM. § Adjusted for age and alcohol consumption (g/d). ¶ Estimate was unstable due to low numbers. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Pleural mesothelioma Lung cancer Laryngeal cancer

Pleural mesothelioma Never exposed Ever exposed HR (95%CI) for asbestos exposure within strata of cigarette smoking Lung cancer Never exposed Ever exposed HR (95%CI) for asbestos exposure within strata of cigarette smoking ‡ Laryngeal cancer Never exposed Ever exposed HR (95%CI) for asbestos exposure within strata of cigarette smoking

Asbestos Exposure

Hazard ratios (HRs) and 95%CIs for pleural mesothelioma, lung and laryngeal cancer for asbestos exposure (yes/no), by smoking; estimated with FINJEM in the NLCS, 1986‐2003.

Table 3.6

Occupational asbestos exposure and respiratory tract tumors

Chapter 3

Population‐attributable fractions were highest for mesothelioma and ranged from 15.9% to 34.5%, depending on the asbestos exposure variable and JEM used. Population‐attributable fractions for lung and laryngeal cancer were more or less comparable and were in the range of 2.5% to 12.1% for lung cancer and 2.9% to 10.0% for laryngeal cancer (Tables 3.3 to 3.5).

DISCUSSION This study was able to confirm the well‐established associations between asbestos exposure and mesothelioma, lung cancer, and laryngeal cancer, revealing elevated risks at the lower end of the exposure distribution in the population‐based NLCS. Associations with lung cancer subtypes were generally comparable to overall lung cancer, except for adenocarcinoma, which showed only a weak positive association after prolonged higher asbestos exposure. For laryngeal cancer, associations were usually stronger for supraglottis than glottis cancer, which showed only a positive association after prolonged higher asbestos exposure. There was no statistically significant interaction on an additive or multiplicative scale between asbestos and smoking in relation to pleural mesothelioma, lung cancer, and laryngeal cancer. The estimated joint effect of asbestos and smoking was both lower than additive and multiplicative for mesothelioma. For lung cancer, the joint effect was between additivity and multiplicativity, while for laryngeal cancer the joint effect was closer to additivity than multiplicativity. Finally, results were robust against the use of different JEMs, though DOMJEM showed higher HRs for the duration of high exposure and FINJEM revealed somewhat higher HRs for ever versus never exposed. The observed positive associations between asbestos and mesothelioma, lung cancer, and laryngeal cancer in the NLCS are qualitatively comparable to many other studies.5,8 Quantitatively, results are more difficult to compare, because the range in relative risks reported in the literature is wide, certainly for lung cancer and mesothelioma, which have been studied extensively. Furthermore, the NLCS is a population‐based study with a wide range in exposure levels, including those at the lower end of the exposure distribution. Low asbestos exposure levels are currently present in most industrialized countries, and the magnitude of the cancer risks associated with them is of importance for setting acceptable exposure limits. For mesothelioma, HRs were significantly elevated in this study, even for the lowest tertile of CE (median, 0.20 f‐y/ml) based on FINJEM (HR=2.69 (95% CI, 1.60 to 4.53)). When using DOMJEM, no increased HR was observed for the lowest tertile. As the time since first exposure to asbestos was more than 20 years for all but one worker with mesothelioma, the length of follow‐up was long enough for mesothelioma to develop. Some previous studies found no evidence of a threshold level for asbestos‐related

Occupational asbestos exposure and respiratory tract tumors

mesothelioma,28,29 though others did find evidence of a threshold level.30,31 Because there is no uniform definition of what low level exposure entails, and because exposure assessment in our study is JEM‐based possibly entailing nondifferential exposure misclassification, we cannot subscribe on the basis of our data to the (non)existence of a threshold level for asbestos‐related mesothelioma. For lung cancer, this study showed a significantly increased HR even for the lowest tertile of CE based on FINJEM (HR=1.47 (95% CI, 1.15 to 1.87)). Again, when using DOMJEM, no increased HR was observed for the lowest tertile. The FINJEM estimate was comparable to or higher than the relative risks presented in the meta‐analysis by van der Bij et al.,32 depending on the model they used, for a higher exposure level of 4 f‐y/ml. In this meta‐analysis, they proposed that the asbestos‐related increase in relative risk of lung cancer may be larger than expected from previous meta‐analyses.32 Also, Gustavsson et al.6 found, in their population‐based study, a higher relative risk of lung cancer at the lower end of the exposure distribution than predicted by downward linear extrapolation from highly exposed occupational cohorts. Possibly, relative risks at the lower end of the exposure distribution are higher than expected. Laryngeal cancer has not been studied as extensively in relation to asbestos as mesothelioma and lung cancer and has only recently been linked causally to asbestos.4 For FINJEM, our results were comparable to the summary relative risk of 1.43 for ever versus never exposed to asbestos in a meta‐analysis of case–control studies, while for DOMJEM, the HR for ever versus never exposed was slightly lower.8 Adjusting for alcohol consumption and smoking did not alter these results, which is in line with several previous studies on asbestos and laryngeal cancer.8 Population‐attributable fractions discussed hereafter are based on ever versus never exposed for ease of comparison with other studies. In the NLCS, PAFs for mesothelioma (31.9% for DOMJEM and 34.3% for FINJEM) are lower than reported in the literature.33‐35 Because virtually all mesothelioma is due to (occupational) asbestos, PAFs of around 100% might have been expected. The NLCS is, however, a population‐ based study with a wide range in exposure levels and a modest exposure prevalence, which both determine the PAF. Moreover, exposure assessment using JEMs could entail nondifferential exposure misclassification leading to attenuated HRs and consequently PAFs. Indeed, the percentage never exposed according to both JEMs is around 50%. If we assume mesothelioma to be solely caused by asbestos, this percentage of 50% may point to nonoccupational asbestos exposure but probably also to incidental occupational asbestos exposures in jobs not captured by the JEMs. Population‐attributable fractions for lung and laryngeal cancer were only slightly lower than reported in the literature. For lung cancer, Albin et al.33 reported PAFs in the range of 10% to 20%, though other studies found lower estimates.36,37 An earlier publication in the NLCS found a PAF of 11.6%,12 which is comparable to the PAF for FINJEM (11.5%) in the present study and twice as high as the PAF for DOMJEM (5.2%). PAFs in this

Chapter 3

earlier study were, however, based on case‐by‐case expert assessment, a shorter follow‐up, and the probability of exposure (without including duration). For laryngeal cancer, the PAF of 8.3% calculated by Nurminen and Karjalainen35 was higher than the PAF of 5.5% for DOMJEM, but lower than the PAF of 9.8% for FINJEM in the NLCS. This study indicated adenocarcinoma to be the lung cancer subtype showing a weak association only after prolonged higher asbestos exposure. The weaker association between asbestos and adenocarcinoma is of interest, because previous studies showed asbestos to be associated preferentially with adenocarcinoma,38,39 though this association has not been reported consistently throughout the literature.40,41 Because asbestos and smoking may have similar exposure routes and related pathophysiological pathways,5 and because smoking is only weakly associated with adenocarcinoma,40,41 this might explain why only after prolonged higher asbestos exposure a positive association with adenocarcinoma was observed. Another reason might be that adenocarcinoma is more frequently observed among nonsmokers,40 and because most subjects exposed to asbestos were smokers, the true association between asbestos and adenocarcinoma may have been stronger. When stratifying by smoking, the association with adenocarcinoma for never‐smokers ever exposed to asbestos compared with never‐smokers never exposed to asbestos was indeed higher (HR=2.43 (95% CI, 0.83 to 7.11)), though numbers were low (FINJEM results, not shown). Relative risk by location of laryngeal cancer in the NLCS is comparable to some, though not all studies.8 The NLCS showed overall stronger associations for supraglottis than glottis cancer. Because the location of the supraglottis seems more readily exposed to tobacco,42 and because exposure routes and pathophysiological pathways of tobacco and asbestos are believed to be related,5 the higher HR for cancer of the supraglottis might have been expected. Unfortunately, we could not check whether and to what degree this association between asbestos and cancer of the supraglottis was driven by smoking, because all asbestos‐exposed men were smoking. In addition, the numbers were low, certainly for supraglottis cancer. As expected, this study showed that the HR of mesothelioma was not influenced by smoking as opposed to HRs of lung and laryngeal cancer. Although smoking is paramount in the asbestos–lung cancer association, smoking is no risk factor for mesothelioma.43 A synergistic effect between asbestos and smoking in relation to lung cancer is supported by a number of systematic reviews,44–46 though the degree of synergism remains uncertain because results range from additivity to supramultiplicativity.6,10 Our study found no statistically significant interaction on an additive or multiplicative scale for lung cancer, and joint effects were between additivity and multiplicativity. For laryngeal cancer, there was also no statistically significant interaction on an additive or multiplicative scale, though joint effects were closer to additivity than

Occupational asbestos exposure and respiratory tract tumors

multiplicativity. Results have to be interpreted carefully, as the numbers were low most notably among the reference category of the never‐smokers not exposed to asbestos. Previous studies also observed a joint effect closer to additivity than multiplicativity, though one study found a supramultiplicative joint effect of asbestos and smoking.8 This study had the advantage of using two JEMs for studying asbestos–cancer associations to get insight into the methodological uncertainty associated with the choice of JEM. Overall, there was a high degree of similarity in results between both JEMs, which reinforced our confidence in the results presented. Nevertheless, both JEMs may have their particular strengths. Possibly, DOMJEM is better in selecting the prolonged highly exposed subjects as reflected by higher HRs for the duration of high exposure and a lower exposure prevalence of the ever highly exposed. FINJEM may be better in discriminating between ever and never exposed, as reflected by somewhat higher HRs for ever versus never exposed and a lower exposure prevalence of the ever exposed. Furthermore, PAFs were affected by this particular difference between both JEMs and were overall higher (up to threefold) for FINJEM than for DOMJEM. Therefore, some caution seems appropriate when judging PAFs that are JEM‐based. Because exposure assessment in population‐based studies like the NLCS has to be based on rather brief occupational histories and JEMs, nondifferential exposure misclassification could have biased the exposure–response relationships. Both JEMs were, however, able to confirm the well‐known association between asbestos exposure and mesothelioma. Strengths of this study included the prospective design; the long, nearly complete follow‐up and large study size; and the possibility to correct for several lifestyle confounders such as smoking and alcohol. The prospective design reduced the potential for recall bias, most notably on potential confounders as smoking and alcohol, and the nearly complete follow‐up of cases and subcohort members made selection bias unlikely. The length of follow‐up ensured a long enough latency to develop mesothelioma and enough power to study subtypes and interaction between asbestos and smoking. Finally, the NLCS is a population‐based study with a wide range in exposure levels, including those at the lower end of the exposure distribution, which are nowadays present in most industrialized countries.

CONCLUSION This study was able to confirm the well‐established associations between asbestos exposure and mesothelioma, lung cancer, and laryngeal cancer, revealing elevated risks at the lower end of the exposure distribution in the population‐based NLCS, also after adjusting for several lifestyle confounders. Relative risks by lung cancer histology are

Chapter 3

comparable to overall lung cancer results, except for adenocarcinoma, which showed only a weak association after prolonged higher asbestos exposure. Supraglottis cancer seems to have a stronger association with asbestos exposure than glottis cancer. There was no evidence of statistically significant additive or multiplicative interaction between asbestos and smoking in relation to pleural mesothelioma, lung cancer, and laryngeal cancer. Nevertheless, joint effects of asbestos and smoking were closer to additivity than multiplicativity for laryngeal cancer. Finally, results were robust against the use of different JEMs, with DOMJEM and FINJEM resulting in essentially similar HRs.

Occupational asbestos exposure and respiratory tract tumors

REFERENCES 1. 2. 3.

7. 8. 9.

10. 11. 12.

13.

14. 15.

16.

17. 18.

19. 20.

WHO. Elimination of asbestos‐related diseases. http://whqlibdoc.who.int/hq/2006/WHO_SDE_OEH_ 06.03_eng.pdf. Accessed October 15, 2012. LaDou J. The asbestos cancer epidemic. Environ Health Perspect. 2004;112:285‐90. Gezondheidsraad. Asbestos: risks of environmental and occupational exposure. The Hague, the Netherlands: Health Council of the Netherlands, report 2010/10E. http://www.gezondheidsraad.nl/ sites/default/files/201010E.pdf. Accessed November 12, 2013. Straif K, Benbrahim‐Tallaa L, Baan R, Grosse Y, Secretan B, El Ghissassi F, Bouvard V, Guha N, Freeman C, Galichet L, Cogliano V; WHO International Agency for Research on Cancer Monograph Working Group. A review of human carcinogens‐‐part C: metals, arsenic, dusts, and fibres. Lancet Oncol. 2009;10:453‐4. IARC. Asbestos (Chrysotile, Amosite, Crocidolite, Tremolite, Actinolite and Anthophyllite). In: A Review of Human Carcinogens: Arsenic, Metals, Fibres, and Dusts. Lyon, France: International Agency for Research on Cancer (IARC), 2012. 219‐309. Gustavsson P, Nyberg F, Pershagen G, Schéele P, Jakobsson R, Plato N. Low‐dose exposure to asbestos and lung cancer: dose‐response relations and interaction with smoking in a population‐based case‐ referent study in Stockholm, Sweden. Am J Epidemiol. 2002;155:1016‐22. Ogden TL. Canadian chrysotile report released‐‐at last. Ann Occup Hyg. 2009;53:307‐9. NAS. Asbestos: Selected Cancers. Washington, DC: The National Academies Press, 2006. 173‐92. Gonzalez M, Vignaud JM, Clement‐Duchene C, Luc A, Wild P, Bertrand O, Thiberville L, Martinet Y, Benichou J, Paris C. Smoking, occupational risk factors, and bronchial tumor location: a possible impact for lung cancer computed tomography scan screening. J Thorac Oncol. 2012;7:128‐36. IARC. Tobacco smoke. In: A Review of Human Carcinogens: Tobacco Smoke and Involuntary Smoking. Lyon, France: International Agency for Research on Cancer (IARC), 2004. 917–9. van den Brandt PA, Goldbohm RA, van 't Veer P, Volovics A, Hermus RJ, Sturmans F. A large‐scale prospective cohort study on diet and cancer in The Netherlands. J Clin Epidemiol. 1990;43:285‐95. van Loon AJ, Kant IJ, Swaen GM, Goldbohm RA, Kremer AM, van den Brandt PA. Occupational exposure to carcinogens and risk of lung cancer: results from The Netherlands cohort study. Occup Environ Med. 1997;54:817‐24. Offermans NS, Vermeulen R, Burdorf A, Peters S, Goldbohm RA, Koeman T, van Tongeren M, Kauppinen T, Kant I, Kromhout H, van den Brandt PA. Comparison of expert and job‐exposure matrix‐based retrospective exposure assessment of occupational carcinogens in the Netherlands Cohort Study. Occup Environ Med. 2012;69:745‐51. Prentice RL. A case‐cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika. 1986;73:1‐11. Casparie M, Tiebosch AT, Burger G, Blauwgeers H, van de Pol A, van Krieken JH, Meijer GA. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol. 2007;29:19‐24. Van den Brandt PA, Schouten LJ, Goldbohm RA, Dorant E, Hunen PM. Development of a record linkage protocol for use in the Dutch Cancer Registry for Epidemiological Research. Int J Epidemiol. 1990;19:553‐8. Goldbohm RA, van den Brandt PA, Dorant E. Estimation of the coverage of Dutch municipalities by cancer registries and PALGA based on hospital discharge data. Tijdschr Soc Gezonheidsz. 1994;72:80‐4. Peters S, Vermeulen R, Cassidy A, Mannetje A', van Tongeren M, Boffetta P, Straif K, Kromhout H; INCO Group. Comparison of exposure assessment methods for occupational carcinogens in a multi‐centre lung cancer case‐control study. Occup Environ Med. 2011;68:148‐53. Kauppinen T, Toikkanen J, Pukkala E. From cross‐tabulations to multipurpose exposure information systems: a new job‐exposure matrix. Am J Ind Med. 1998;33:409‐17. Stewart PA, Herrick RF, Blair A, Checkoway H, Droz P, Fine L, Fischer L, Harris R, Kauppinen T, Saracci R. Highlights of the 1990 Leesburg, Virginia, International Workshop on Retrospective Exposure Assessment for Occupational Epidemiology Studies. Scand J Work Environ Health. 1991;17:281‐5.

Chapter 3

21. Korn EL, Graubard BI, Midthune D. Time‐to‐event analysis of longitudinal follow‐up of a survey: choice of the time‐scale. Am J Epidemiol. 1997;145:72‐80. 22. Gail MH, Graubard B, Williamson DF, Flegal KM. Comments on 'Choice of time scale and its effect on significance of predictors in longitudinal studies' by Michael J. Pencina, Martin G. Larson and Ralph B. D'Agostino, Statistics in Medicine 2007; 26:1343‐59. Stat Med. 2009;28:1315‐7. 23. Volovics A, van den Brandt PA. Methods for the analysis of case‐cohort studies. Biomed J. 1997;39: 195‐214. 24. Barlow WE, Ichikawa L, Rosner D, Izumi S. Analysis of case‐cohort designs. J Clin Epidemiol. 1999;52:1165‐72. 25. Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69:239‐ 41. 26. Knol MJ, VanderWeele TJ. Recommendations for presenting analyses of effect modification and interaction. Int J Epidemiol. 2012;41:514‐20. 27. Steenland K, Armstrong B. An overview of methods for calculating the burden of disease due to specific risk factors. Epidemiology. 2006;17:512‐9. 28. Iwatsubo Y, Pairon JC, Boutin C, Ménard O, Massin N, Caillaud D, Orlowski E, Galateau‐Salle F, Bignon J, Brochard P. Pleural mesothelioma: dose‐response relation at low levels of asbestos exposure in a French population‐based case‐control study. Am J Epidemiol. 1998;148:133‐42. 29. Hillerdal G. Mesothelioma: cases associated with non‐occupational and low dose exposures. Occup Environ Med. 1999;56:505‐13. 30. Ilgren EB, Browne K. Asbestos‐related mesothelioma: evidence for a threshold in animals and humans. Regul Toxicol Pharmacol. 1991;13:116‐32. 31. Gibbs GW, Berry G. Mesothelioma and asbestos. Regul Toxicol Pharmacol. 2008;52:S223‐31. 32. van der Bij S, Koffijberg H, Lenters V, Portengen L, Moons KG, Heederik D, Vermeulen RC. Lung cancer risk at low cumulative asbestos exposure: meta‐regression of the exposure‐response relationship. Cancer Causes Control. 2013;24:1–12. 33. Albin M, Magnani C, Krstev S, Rapiti E, Shefer I. Asbestos and cancer: An overview of current trends in Europe. Environ Health Perspect. 1999;107 Suppl 2:289‐98. 34. Driscoll T, Nelson DI, Steenland K, Leigh J, Concha‐Barrientos M, Fingerhut M, Prüss‐Ustün A. The global burden of disease due to occupational carcinogens. Am J Ind Med. 2005;48:419‐31. 35. Nurminen M, Karjalainen A. Epidemiologic estimate of the proportion of fatalities related to occupational factors in Finland. Scand J Work Environ Health. 2001;27:161‐213. 36. Nicholson WJ, Perkel G, Selikoff IJ. Occupational exposure to asbestos: population at risk and projected mortality‐‐1980‐2030. Am J Ind Med. 1982;3:259‐311. 37. Morabia A, Markowitz S, Garibaldi K, Wynder EL Lung cancer and occupation: results of a multicentre case‐control study. Br J Ind Med. 1992;49:721‐7. 38. Johansson L, Albin M, Jakobsson K, Mikoczy Z. Histological type of lung carcinoma in asbestos cement workers and matched controls. Br J Ind Med. 1992;49:626‐30. 39. Raffn E, Lynge E, Korsgaard B. Incidence of lung cancer by histological type among asbestos cement workers in Denmark. Br J Ind Med. 1993;50:85‐9. 40. Lee BW, Wain JC, Kelsey KT, Wiencke JK, Christiani DC. Association of cigarette smoking and asbestos exposure with location and histology of lung cancer. Am J Respir Crit Care Med. 1998;157:748‐55. 41. Paris C, Benichou J, Saunier F, Metayer J, Brochard P, Thiberville L, Nouvet G. Smoking status, occupational asbestos exposure and bronchial location of lung cancer. Lung Cancer. 2003;40:17‐24. 42. Schottenfeld D, Fraumeni J. Cancer Epidemiology and Prevention. New York: Oxford University Press, 2006. 627–37. 43. Kanarek MS. Mesothelioma from chrysotile asbestos: update. Ann Epidemiol. 2011;21:688‐97. 44. Erren TC, Jacobsen M, Piekarski C. Synergy between asbestos and smoking on lung cancer risks. Epidemiology. 1999;10:405‐11. 45. Lee PN. Relation between exposure to asbestos and smoking jointly and the risk of lung cancer. Occup Environ Med. 2001;58:145‐53. 46. Liddell FD. The interaction of asbestos and smoking in lung cancer. Ann Occup Hyg. 2001;45:341‐56.

Occupational asbestos exposure and respiratory tract tumors

47. Swuste P, Dahhan M, Burdorf A. Linking expert judgement and trends in occupational exposure into a job‐exposure matrix for historical exposure to asbestos in the Netherlands. Ann Occup Hyg. 2008;52:397‐403.

Chapter 3

Occupational asbestos exposure and gastrointestinal tract tumors

Chapter 4

Occupational asbestos exposure and risk of esophageal, gastric and colorectal cancer in the prospective Netherlands Cohort Study Nadine S.M. Offermans Roel Vermeulen Alex Burdorf R. Alexandra Goldbohm András P. Keszei Susan Peters Timo Kauppinen Hans Kromhout Piet A. van den Brandt International Journal of Cancer. 2014;135:1970‐7

Chapter 4

ABSTRACT Objectives The evidence for an association between occupational asbestos exposure and esophageal, gastric and colorectal cancer is limited. We studied this association specifically addressing risk differences between relatively low and high exposure, risk associated with cancer subtypes, the influence of potential confounders and the interaction between asbestos and smoking in relation to cancer risk. Methods Using the Netherlands Cohort Study (n=58,279 men, aged 55–69 years at baseline), asbestos exposure was estimated by linkage to a job‐exposure matrix. After 17.3 years of follow‐up, 187 esophageal, 486 gastric and 1,724 colorectal cancer cases were available for analysis. Results The models adjusted for age and family history of cancer showed that mainly (prolonged) exposure to high levels of asbestos was statistically significantly associated with risk of esophageal adenocarcinoma (EAC), total and distal colon cancer and rectal cancer. For overall gastric cancer and gastric non‐cardia adenocarcinoma (GNCA), also exposure to lower levels of asbestos was associated. Additional adjustment for lifestyle confounders, especially smoking status, yielded non‐significant associations with overall gastric cancer and GNCA in the multivariable‐adjusted model, except for the prolonged highly exposed subjects (tertile 3 vs. never: HR 2.67, 95% CI: 1.11–6.44 and HR 3.35, 95% CI: 1.33–8.44, respectively). No statistically significant additive or multiplicative interaction between asbestos and smoking was observed for any of the studied cancers. Conclusions This prospective population‐based study showed that (prolonged) high asbestos exposure was associated with overall gastric cancer, EAC, GNCA, total and distal colon cancer and rectal cancer.

Occupational asbestos exposure and gastrointestinal tract tumors

INTRODUCTION The International Agency for Research on Cancer evaluated the evidence for an association between asbestos exposure and gastric and colorectal cancer as limited, though for colorectal cancer they were evenly divided as to whether the evidence was strong enough to justify classification as sufficient.1 For esophageal cancer, a relatively rare cancer, results of epidemiological studies on associations with asbestos are mixed and together with the fact that animal experiments do not support biological activity of asbestos at this site, the evidence was considered to be inadequate.2 Therefore, the question remains whether asbestos entails an increased risk of developing gastrointestinal tumors, and if so, whether risk differs for relatively low and high exposure levels. As there are numerous other risk factors for gastrointestinal cancers, an additional question relates to the influence of potential confounders. Furthermore, cancer subtypes may have different etiologies and should be studied separately if possible. Finally, as for lung cancer, the question arises if interaction between asbestos and smoking is present in relation to gastrointestinal cancers.1,2 Population‐based studies are well suited to address these questions given their overall wide range in exposure levels, possibility to control for potential confounders and often larger number of cases. One of these population‐based studies is the prospective Netherlands Cohort Study (NLCS), conducted among 120,852 men and women of the general population.3 Within the framework of this study, we had the following objectives: 1. To investigate the overall association between occupational asbestos exposure and risk of esophageal, gastric and colorectal cancer with special attention to risk differences between relatively low and high exposure and to potential confounding; 2. To study the association between occupational asbestos exposure and subtypes of esophageal, gastric and colorectal cancer; 3. To examine the presence of additive or multiplicative interaction between asbestos and smoking in relation to esophageal, gastric and colorectal cancer. As the proportion of long‐term employed women was rather low, this study was only conducted among men.

MATERIALS AND METHODS Study population and cancer follow‐up In brief, the NLCS started in September 1986 when 58,279 men from the Netherlands aged 55–69 years were enrolled in the cohort. At baseline, participants completed a

Chapter 4

self‐administered questionnaire on dietary habits and lifestyle, occupational history and other potential risk factors for cancer.3 For reasons of efficiency in questionnaire processing and follow‐up (a total of 17.3 years of follow‐up was available at the time of analysis), the case–cohort approach was used.4 End‐points for this study were incident, microscopically confirmed esophageal, gastric and colorectal cancer cases, obtained by record linkage to cancer registries and classified by anatomic site or histological type. Accompanying codes are available online (Supporting Information Table S4.1). Esophagus carcinomas included squamous cell carcinomas (ESCC) and adenocarcinomas (EAC). Gastric cancer was classified as cardia adenocarcinomas (GCA) and non‐cardia adenocarcinomas (GNCA). Colorectal cancer cases were classified as proximal colon, distal colon, rectosigmoid or rectal tumors. All prevalent cases at baseline other than skin cancer and subjects without any, or only uncodable, information on occupational history or who never worked professionally were omitted from the analyses.

Occupational exposure assessment Information on lifetime occupational history until 1986 was obtained from the questionnaire completed at study enrolment. Questions concerned the job title, name and type of the company, products made in the department and period of employment. Subjects could enter a maximum of five occupations. This was generally sufficient to cover the lifetime occupational history for the large majority of the cohort, as cohort subjects held on average 1.9 jobs during their working life. Occupational asbestos exposure was estimated by linkage to a job‐exposure matrix (JEM), DOMJEM,5 as described previously.6,7 Several exposure variables were defined: ever versus never exposed to asbestos (yes/no), duration of exposure (years), cumulative exposure (CE; unit‐years), ever versus never highly exposed to asbestos (yes/no) and duration of high exposure (years). Ever versus never exposed is based on the CE, in that subjects were classified as being ever exposed to asbestos when CE > 0. For occupations with P X I > 0, duration of employment was summed in order to obtain the duration of exposure. The CE measure was estimated by summing the product of P X I and duration over the reported occupations. The DOMJEM scores of no, low and high exposure for P X I were arbitrarily assigned values of 0, 1 and 4 to mirror the log‐normal (multiplicative) nature of occupational exposure levels, hence the expression in unit‐years. The weighting was based on reported levels for semi‐quantitatively scored exposure, thereby assuring a balanced weighting between intensity and duration in the calculation of cumulative exposure.8 For the duration of high exposure, first the P X I per occupation was categorized into no, low or high exposure based on the distribution in the subcohort. Second, duration

Occupational asbestos exposure and gastrointestinal tract tumors

of employment was summed for those occupations with a high P X I in order to obtain the duration of high exposure. For 10% of the population, some information on occupational history could not be coded. In these cases, asbestos exposure was set to zero for the job or period with lacking occupational history information.

Statistical analyses Hazard ratios (HR) and corresponding 95% confidence intervals (95% CI) were estimated by using Cox proportional hazards models. The models were adjusted for age and family history of cancer (yes/no), as well as multivariable‐adjusted. The covariates included in the multivariable‐adjusted models were either a priori‐selected risk factors based on the literature or variables that changed the age‐adjusted regression coefficients by at least 10% (using a backwards stepwise procedure). To enable comparison, the models adjusted for age and family history of cancer were restricted to subjects included in the multivariable‐adjusted analyses (i.e., with no missing values on confounding variables), which left 1,866 subcohort members and 187 esophageal (61 ESCC and 126 EAC), 486 gastric (143 GCA and 343 GNCA) and 1,724 colorectal (1,113 total colon (503 proximal and 568 distal), and 425 rectum) cancer cases for analyses. For each analysis, the proportional hazards assumption was tested by using the scaled Schoenfeld residuals.9 Trends for all subjects and only including the exposed were evaluated with the Wald test by assigning subjects the median value for each level of the categorical variable among the subcohort members, and this variable was entered as a continuous term in the Cox regression model. We tested for a possible interaction between asbestos exposure (yes/no) and smoking status (never/former/current) in relation to gastrointestinal cancers. Statistically significant departure from multiplicativity was tested by including an interaction term in the Cox regression model. Statistically significant departure from additivity was tested using the confidence interval of the relative excess risk of cancer due to interaction (RERI), according to methods described by Knol and VanderWeele,10 which were adapted for use in Stata. All tests (2‐tailed) were performed using Stata (version 10), and differences were regarded as statistically significant at P<0.05.

Sensitivity analysis In order to provide insight into the methodological uncertainty associated with the choice of JEM, we also used a Finnish JEM (FINJEM)11 as described previously.6,7 Results are presented in the Supporting Information Tables S4–S6, as well as a brief comparison of results of both JEMs.

Chapter 4

RESULTS The distribution of asbestos exposure and potential confounders among male subcohort members and cancer cases in the NLCS is given in Supporting Information Table S4.2. On average, cases for all three cancers were more often, longer and higher exposed to asbestos. Except for (prolonged) higher asbestos exposure, most asbestos exposure variables showed no statistically significant increased risk of esophageal, gastric and colorectal cancer in the multivariable‐adjusted model (Tables 4.1–4.3). Adjusting for potential confounders was generally of minor influence, except for overall gastric cancer and GNCA. Therefore, only multivariable‐adjusted results are presented below, unless mentioned otherwise. For esophageal cancer (Table 4.1), no statistically significant results were observed, though the HR (95% CI) for ever versus never highly exposed and for the duration of high exposure (continuous variable; per 10 years) was borderline significant (HR 2.22 (1.00–4.94) and 1.45(1.00–2.10), respectively). Results by histology of esophageal cancer showed a statistically significant association with EAC, not only for ever versus never highly exposed (HR 2.52 (1.01–6.26)), but also for other asbestos exposure variables. Additionally, a significant exposure–response relation for the duration of exposure and the cumulative exposure (Ptrend<0.05) was observed, which disappeared when only the exposed were considered. For gastric cancer (Table 4.2), several asbestos exposure variables revealed statistically significant increased HRs in the model adjusted for age and family history of gastric cancer. These associations mostly disappeared in the multivariable‐adjusted model, especially after adjusting for smoking status. The same was true for the exposure– response relations. The only exception was the duration of high exposure which remained significant (tertile 3 vs. never: HR 2.67 (1.11–6.44); Ptrend<0.05, also when only the exposed were considered). Results by location of gastric cancer showed statistically significant associations with most exposure variables for GNCA in the model adjusted for age and family history of gastric cancer (results not shown). Again, in the multivariable‐adjusted model only the association with the duration of high exposure remained significant [tertile 3 vs. never: HR 3.35 (1.33–8.44); Ptrend<0.05, also when only the exposed were considered). GCA showed no statistically significant associations or exposure–response relations.

Person years in subcohort 19216 7565 2863 2209 2493 26781 2891 2069 2606 26781 26145 636 220 221 194 26781 122 65 27 15 23 187 25 17 23 187 179 8 3 2 3 187

No. of cases 1 (ref) 1.40 (1.02‐1.93) 1.52 (0.97‐2.37) 1.13 (0.64‐1.98) 1.51 (0.94‐2.43) 0.066 1.11 (0.99‐1.25) 1.39 (0.88‐2.19) 1.37 (0.80‐2.34) 1.45 (0.90‐2.34) 0.052 1.01 (1.00‐1.02) 1 (ref) 2.20 (1.00‐4.83) 2.19 (0.62‐7.75) 1.48 (0.33‐6.62) 2.40 (0.68‐8.53) 0.091 1.42 (0.99‐2.03)

1 (ref) 1.35 (0.96‐1.89) 1.45 (0.90‐2.32) 1.10 (0.62‐1.95) 1.44 (0.88‐2.36) 0.113 1.10 (0.97‐1.24) 1.31 (0.81‐2.12) 1.35 (0.78‐2.32) 1.40 (0.85‐2.28) 0.088 1.01 (1.00‐1.02) 1 (ref) 2.22 (1.00‐4.94) 2.27 (0.63‐8.21) 1.44 (0.32‐6.51) 2.50 (0.69‐9.02) 0.083 1.45 (1.00‐2.10)

Esophageal cancer HR† (95%CI) HR‡ (95%CI) 44 17 7 5 5 61 7 5 5 61 59 2 1 0 1 61

No. of cases 1 (ref) 1.08 (0.58‐1.99) 1.12 (0.48‐2.64) 1.20 (0.46‐3.14) 0.93 (0.35‐2.51) 0.962 0.96 (0.76‐1.23) 1.11 (0.47‐2.63) 1.26 (0.48‐3.33) 0.91 (0.34‐2.44) 0.977 1.00 (0.98‐1.02) 1 (ref) 1.64 (0.37‐7.28) 2.46 (0.30‐20.27) ‐ 2.64 (0.31‐22.61) 0.526 1.45 (0.72‐2.93)

ESCC HR‡ (95%CI)

78 48 20 10 18 126 18 12 18 126 120 6 2 2 2 126

No. of cases 1 (ref) 1.54 (1.03‐2.29) 1.71 (1.00‐2.93) 1.07 (0.54‐2.13) 1.77 (1.01‐3.11) <0.05 (0.175) 1.16 (1.01‐1.33) 1.50 (0.86‐2.61) 1.40 (0.74‐2.65) 1.70 (0.97‐2.98) <0.05 (0.753) 1.01 (1.00‐1.02) 1 (ref) 2.52 (1.01‐6.26) 2.26 (0.49‐10.39) 2.09 (0.46‐9.52) 2.27 (0.49‐10.45) 0.093 1.41 (0.94‐2.11)

EAC HR‡ (95%CI)

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. † Adjusted for age (yrs) and family history of esophageal cancer (yes/no). ‡ Adjusted for age (yrs), family history of esophageal cancer (yes/no), smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower vocational, secondary and medium vocational, and higher vocational/university), BMI (kg/m2), and alcohol consumption (g/d). § Exposure based on the cumulative probabilityintensity of exposure (unit‐years). ¶ Exposure based on the probabilityintensity of exposure (unit‐years) per job. # Trends over the exposed subjects were added only if trends over all subjects were statistically significant in the full covariate model. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Never exposed§ Ever exposed§ Duration of exposure¶ (yrs) T1 (median:4) T2 (median:18) T3 (median:37) P for trend (over the exposed only)# Continuous, per 10 yrs Cumulative probabilityintensity of exposure (unit‐years) T1 (median:4) T2 (median:20) T3 (median:38) P for trend (over the exposed only)# Continuous, per 1 unit‐year Never highly exposed Ever highly exposed Duration of high exposure¶ (yrs) T1 (median:4) T2 (median:10.5) T3 (median:30.5) P for trend Continuous, per 10 yrs

Hazard ratios (HRs) and 95% CIs for overall esophageal cancer and subtypes for categories of asbestos exposure*, estimated with DOMJEM in the NLCS, 1986‐2003.

Table 4.1

Occupational asbestos exposure and gastrointestinal tract tumors

Never exposed§ Ever exposed§ Duration of exposure¶ (yrs) T1 (median:4) T2 (median:18) T3 (median:37) P for trend Continuous, per 10 yrs Cumulative probabilityintensity of exposure (unit‐years) T1 (median:4) T2 (median:20) T3 (median:38) P for trend Continuous, per 1 unit‐year Never highly exposed Ever highly exposed Duration of high exposure¶ (yrs) T1 (median:4) T2 (median:10.5) T3 (median:30.5) P for trend (over the exposed only)# Continuous, per 10 yrs

Person years in subcohort 19216 7565 2863 2209 2493 26781 2891 2069 2606 26781 26145 636 220 221 194 26781 No. of cases 341 145 33 56 56 486 37 47 61 486 468 18 2 7 9 486 †

Gastric cancer HR† (95%CI) HR‡ (95%CI) 1 (ref) 1 (ref) 1.13 (0.91‐1.41) 1.02 (0.81‐1.29) 0.68 (0.46‐1.01) 0.63 (0.42‐0.94) 1.52 (1.09‐2.12) 1.38 (0.98‐1.94) 1.31 (0.94‐1.82) 1.15 (0.81‐1.61) <0.05 0.211 1.10 (1.01‐1.19) 1.06 (0.97‐1.15) 0.75 (0.51‐1.09) 0.69 (0.47‐1.02) 1.36 (0.95‐1.93) 1.22 (0.85‐1.75) 1.39 (1.01‐1.91) 1.22 (0.87‐1.70) <0.05 0.207 1.01 (1.00‐1.01) 1.01 (1.00‐1.01) 1 (ref) 1 (ref) 1.84 (1.04‐3.25) 1.72 (0.96‐3.07) 0.56 (0.13‐2.49) 0.56 (0.12‐2.50) 2.15 (0.87‐5.31) 2.07 (0.83‐5.13) 2.86 (1.20‐6.80) 2.67 (1.11‐6.44) <0.05 <0.05 (<0.05) 1.43 (1.10‐1.86) 1.41 (1.08‐1.85) No. of cases 102 41 8 18 15 143 8 15 18 143 138 5 1 3 1 143 1 (ref) 1.01 (0.67‐1.51) 0.53 (0.25‐1.13) 1.48 (0.86‐2.56) 1.12 (0.61‐2.06) 0.429 1.07 (0.93‐1.25) 0.53 (0.25‐1.11) 1.32 (0.72‐2.40) 1.31 (0.75‐2.30) 0.324 1.00 (1.00‐1.01) 1 (ref) 1.63 (0.62‐4.30) 0.96 (0.12‐7.63) 2.89 (0.84‐9.99) 1.05 (0.14‐8.19) 0.284 1.11 (0.71‐1.74)

GCA HR‡ (95%CI)

No. of cases 239 104 25 38 41 343 29 32 43 343 330 13 1 4 8 343

1 (ref) 1.02 (0.78‐1.34) 0.67 (0.42‐1.05) 1.34 (0.90‐2.00) 1.16 (0.79‐1.70) 0.288 1.05 (0.95‐1.15) 0.76 (0.50‐1.18) 1.18 (0.77‐1.80) 1.19 (0.81‐1.74) 0.341 1.01 (1.00‐1.01) 1 (ref) 1.79 (0.92‐3.47) 0.41 (0.05‐3.16) 1.70 (0.53‐5.45) 3.35 (1.33‐8.44) <0.05 (<0.05) 1.51 (1.13‐2.02)

GNCA HR‡ (95%CI)

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. Adjusted for age (yrs) and family history of gastric cancer ‡ (yes/no). Adjusted for age (yrs), family history of gastric cancer (yes/no), smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower vocational, secondary and medium vocational, and higher vocational/university), 2 § ¶ BMI (kg/m ), and alcohol consumption (g/d). Exposure based on the cumulative probabilityintensity of exposure (unit‐years). Exposure based on the # probabilityintensity of exposure (unit‐years) per job. Trends over the exposed subjects were added only if trends over all subjects were statistically significant in the full covariate model. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Hazard ratios (HRs) and 95% CIs for overall gastric cancer and subtypes for categories of asbestos exposure , estimated with DOMJEM in the NLCS, 1986‐2003.

Table 4.2

Chapter 4

Total colorectal cancer HR† (95%CI) HR‡ (95%CI) No. of cases 1 (ref) 1 (ref) 836 0.98 (0.84‐1.14) 0.95 (0.81‐1.11) 277 0.79 (0.62‐1.00) 0.77 (0.60‐0.98) 89 1.10 (0.86‐1.40) 1.05 (0.81‐1.35) 87 1.10 (0.87‐1.40) 1.07 (0.84‐1.36) 101 0.490 0.612 1.03 (0.97‐1.10) 1.02 (0.96‐1.09) 1113 0.79 (0.63‐1.00) 0.78 (0.61‐0.99) 93 1.05 (0.82‐1.36) 1.01 (0.77‐1.31) 78 1.14 (0.90‐1.43) 1.10 (0.87‐1.40) 106 0.435 0.547 1.00 (1.00‐1.01) 1.00 (1.00‐1.01) 1113 1 (ref) 1 (ref) 1084 1.60 (1.05‐2.42) 1.53 (1.00‐2.34) 29 1.71 (0.86‐3.41) 1.69 (0.84‐3.40) 7 1.25 (0.59‐2.64) 1.22 (0.57‐2.61) 7 1.96 (0.97‐3.95) 1.87 (0.93‐3.77) 15 <0.05 <0.05 (0.833) 1.25 (1.00‐1.55) 1.23 (0.99‐1.53) 1113 1 (ref) 0.89 (0.74‐1.06) 0.73 (0.55‐0.97) 0.97 (0.72‐1.30) 1.00 (0.76‐1.32) 0.819 1.00 (0.93‐1.07) 0.74 (0.56‐0.98) 0.95 (0.70‐1.29) 1.01 (0.77‐1.33) 0.824 1.00 (1.00‐1.01) 1 (ref) 1.25 (0.76‐2.05) 0.93 (0.37‐2.36) 0.89 (0.35‐2.29) 2.19 (1.04‐4.62) 0.107 1.24 (0.97‐1.59)

Colon HR‡ (95%CI)

†

Proximal colon No. HR‡ (95%CI) of cases 370 1 (ref) 133 0.94 (0.75‐1.19) 36 0.65 (0.44‐0.96) 49 1.24 (0.87‐1.77) 48 1.05 (0.74‐1.49) 0.617 503 1.03 (0.94‐1.12) 38 0.68 (0.46‐0.99) 45 1.23 (0.85‐1.79) 50 1.06 (0.75‐1.49) 0.627 503 1.00 (1.00‐1.01) 490 1 (ref) 13 1.28 (0.67‐2.45) 3 0.89 (0.25‐3.17) 4 1.13 (0.36‐3.55) 6 1.97 (0.73‐5.33) 0.254 503 1.19 (0.85‐1.66)

No. of cases 431 137 53 37 47 568 55 32 50 568 553 15 3 3 9 568

1 (ref) 0.87 (0.69‐1.10) 0.87 (0.62‐1.21) 0.80 (0.54‐1.19) 0.94 (0.65‐1.34) 0.473 0.97 (0.89‐1.07) 0.87 (0.63‐1.22) 0.76 (0.50‐1.15) 0.95 (0.67‐1.35) 0.489 1.00 (1.00‐1.01) 1 (ref) 1.26 (0.68‐2.33) 0.79 (0.22‐2.80) 0.76 (0.21‐2.74) 2.54 (1.09‐5.93) 0.117 1.32 (0.99‐1.75)

Distal colon HR‡ (95%CI)

No. of cases 293 132 39 44 49 425 37 41 54 425 404 21 10 7 4 425

1 (ref) 1.05 (0.82‐1.34) 0.83 (0.56‐1.21) 1.22 (0.85‐1.77) 1.17 (0.82‐1.67) 0.255 1.06 (0.97‐1.16) 0.78 (0.53‐1.14) 1.21 (0.82‐1.78) 1.24 (0.88‐1.75) 0.170 1.01 (1.00‐1.01) 1 (ref) 2.15 (1.23‐3.77) 2.90 (1.25‐6.76) 2.23 (0.88‐5.67) 1.31 (0.42‐4.12) <0.05 (0.183) 1.22 (0.91‐1.62)

Rectum HR‡ (95%CI)

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. Adjusted for age (yrs) and family history of colorectal cancer ‡ (yes/no). Adjusted for age (yrs), family history of colorectal cancer (yes/no), smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower vocational, secondary and medium vocational, and higher vocational/university), 2 § ¶ BMI (kg/m ), and alcohol consumption (g/d). Exposure based on the cumulative probabilityintensity of exposure (unit‐years). Exposure based on the # probabilityintensity of exposure (unit‐years) per job. Trends over the exposed subjects were added only if trends over all subjects were statistically significant in the full covariate model. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Person No. years in of subcohort cases § Never exposed 19022 1258 Ever exposed§ 7508 466 Duration of exposure¶ (yrs) T1 (median:4) 2842 146 T2 (median:18) 2197 148 T3 (median:37) 2469 172 P for trend Continuous, per 10 yrs 26530 1724 Cumulative probabilityintensity of exposure (unit‐years) T1 (median:4) 2869 150 T2 (median:20) 2052 132 T3 (median:38) 2588 184 P for trend Continuous, per 1 unit‐year 26530 1724 Never highly exposed 25898 409 Ever highly exposed 632 57 ¶ Duration of high exposure (yrs) T1 (median:4) 220 21 T2 (median:10.5) 222 15 T3 (median:30.5) 190 21 P for trend (over the exposed only)# Continuous, per 10 yrs 26530 1724

Hazard ratios (HRs) and 95% CIs for overall colorectal cancer and subtypes for categories of asbestos exposure*, estimated with DOMJEM in the NLCS, 1986‐2003.

Table 4.3

Occupational asbestos exposure and gastrointestinal tract tumors

Chapter 4

For colorectal cancer (Table 4.3), no statistically significant results were noted, though the HR for ever versus never highly exposed was borderline significant (HR 1.53(1.00‐2.34)). Results by location of colorectal cancer showed statistically significant associations for the duration of high exposure with total colon and distal colon cancer (tertile 3 vs. never: HR 2.19 (1.04–4.62) and 2.54 (1.09–5.93), respectively), and for ever versus never highly exposed with rectal cancer (HR 2.15 (1.23–3.77)). For proximal colon cancer, none of the associations or exposure–response relations was statistically significant. When testing for departure from multiplicativity or additivity (Supporting Information Table S4.3), esophageal cancer showed a p for multiplicative interaction of 0.09. For the other cancer (sub)types, there was no indication of a multiplicative or additive interaction between asbestos and smoking.

DISCUSSION Mainly after (prolonged) exposure to high levels of asbestos, this population‐based study showed multivariable‐adjusted statistically significant increased HRs for overall gastric cancer, EAC, GNCA, total and distal colon and rectal cancer in the NLCS. Adjusting for potential confounders, especially smoking status, yielded non‐significant associations with overall gastric cancer and GNCA for those exposed to lower levels of asbestos. There was no statistically significant additive or multiplicative interaction between asbestos and smoking in relation to esophageal, gastric and colorectal cancer.

Esophageal cancer Several occupational cohort studies have observed elevated cancer rates after asbestos exposure, and have suggested that risk might be dose‐dependent.1 A meta‐analysis conducted by the Institute of Medicine (IOM) reported a summary relative risk (RR) of 0.99 (0.79–1.27) for any versus no asbestos exposure.2 In contrast, a meta‐analysis examining studies with heavier exposures reported an elevated summary standardized mortality ratio (SMR 2.38 (1.45–3.68)) in asbestos exposed workers.12 Our study also showed an increased HR for the (prolonged) highly exposed subjects, which was borderline significant. This may support the notion that only higher asbestos exposure levels entail an increased risk of esophageal cancer. Our study showed significantly increased HRs for EAC. As esophageal cancer is a relatively rare cancer, only few studies looked at subtypes. For both histological types, results are mixed, with a suggestion of an exposure–response relation only for EAC.13–16 As such, the analyses reported here provide some further support for an association with EAC but not ESCC.

Occupational asbestos exposure and gastrointestinal tract tumors

Gastric cancer For overall gastric cancer, increased risks and exposure–response relations have been observed in occupational cohorts with higher asbestos exposure,1,2,17,18 which is comparable to our study. The IOM meta‐analysis of occupational cohorts reported a summary RR of 1.17 (1.07–1.28) for any versus no asbestos exposure, while results for case–control studies were inconsistent.2 We hypothesize that, also for gastric cancer only prolonged exposure to higher asbestos levels may be associated. To our knowledge, this is the first prospective population‐based cohort study to investigate the association between asbestos and both gastric cancer subtypes. In other study designs, no increased risk of GCA has been observed,15,19 which is comparable to our study in which there was only an increased risk of GNCA. Diverging trends and marked geographic variation suggest GCA and GNCA to be separate disease entities with different etiologies.20 This may be supported by these diverging results for asbestos exposure. An industry‐based study, however, did not observe an association with GNCA after either moderate or high asbestos exposure.21

Colorectal cancer Occupational cohorts fairly consistently show increased risks of colorectal cancer after asbestos exposure and exposure–response relations,1 while results from case–control studies are less consistent.2 Meta‐analyses reported increased risks after asbestos exposure only when SMRs for lung cancer exceeded 2 or 3.17 We also found an increased HR for ever versus never highly exposed, which was borderline significant. As the IOM meta‐analysis of occupational cohorts reported a summary RR of 1.15 (1.02‐1.31) for any versus no asbestos exposure,2 only (prolonged) exposure to higher asbestos levels may increase the risk of colorectal cancer as well. For colorectal cancer subtypes, there is some suggestion in the literature that the association with asbestos might be stronger for colon than for rectal cancer,1 which is also based on studies that found a higher RR of right (or proximal) colon than left (or distal) colon cancer,22,23 or found an association with colorectal cancer but not with rectal cancer.17 One study found, however, a higher SMR of rectal cancer versus intestine and rectal cancer together.24 Our study found significantly increased HRs for colon (both total and distal) and rectal cancer. The association with rectal cancer and, given the location, with distal colon cancer could be explained by asbestos’ relation to smoking carcinogenesis and the fact that associations with smoking have been appreciably stronger for rectal cancer than (proximal) colon cancer.25 It is not surprising that increased HRs are most likely to be observed in those with prolonged exposure to high levels of asbestos. In this respect, evidence for an association between asbestos and all three cancers comes mainly from occupational cohort studies with generally higher exposure levels.1,2,17,18 As the NLCS is a population‐

Chapter 4

based study with a wide range in exposure levels, including those at the lower end of the exposure distribution, this might explain the absence of an association with the other asbestos exposure variables. Another explanation may relate to the fact that occupations with known high asbestos exposure are usually better classified (i.e., with higher specificity) in JEMs than occupations with lower and more variable asbestos exposure, which may lead to false‐negative results. This holds in particular for DOMJEM. Furthermore, our JEM‐based exposure assessment possibly entailed non‐ differential exposure misclassification resulting in bias towards the null value. A previous study in the NLCS using DOMJEM was, however, able to corroborate the well‐ known association between asbestos and pleural mesothelioma.7 Finally, several associations with overall gastric cancer and GNCA were reduced and became non‐ significant after adjusting for smoking status. Therefore, the assumption that potential confounders are only weakly correlated with asbestos exposure seems not totally justified. However, HRs were in most instances only marginally significant to begin with. Contrary to lung cancer, there is almost no epidemiological or experimental evidence addressing whether asbestos is a cofactor of tobacco smoking in the development of esophageal, gastric or colorectal cancer.2 A study by Liddell et al.26 showed a modest interaction between cumulative asbestos exposure and smoking in relation to gastric cancer, and Aliyu et al.27 reported a RR of colorectal cancer of 1.36 (0.96–1.93) for asbestos and smoking together as compared to smoking alone. Our study found no statistically significant interaction on an additive or multiplicative scale for any of the cancers. Sample sizes were small, however, especially for esophageal cancer. Strengths of our study included the prospective design, the long, nearly complete follow‐up and large study size, and the possibility to adjust for several lifestyle confounders as alcohol and smoking.

CONCLUSION In conclusion, this study suggests that increased HRs of overall gastric cancer, and possibly also of overall esophageal and colorectal cancer, are observed after (prolonged) exposure to high asbestos levels. This was also true for EAC, GNCA, total colon, distal colon and rectal cancer. Adjustment for smoking may be relevant when studying overall gastric cancer and GNCA, as several HRs were reduced and became non‐significant after adjusting for smoking status. No statistically significant additive or multiplicative interaction between asbestos and smoking was observed for any of the studied cancers.

Occupational asbestos exposure and gastrointestinal tract tumors

REFERENCES 1.

2. 3. 4. 5.

9. 10. 11. 12. 13.

14. 15.

16.

17. 18. 19. 20. 21. 22.

IARC. Asbestos (Chrysotile, Amosite, Crocidolite, Tremolite, Actinolite and Anthophyllite). In: A Review of Human Carcinogens: Arsenic, Metals, Fibres, and Dusts. Lyon, France: International Agency for Research on Cancer (IARC), 2012. 219‐309. NAS. Asbestos: Selected Cancers. Washington, DC: The National Academies Press, 2006. 193‐229. van den Brandt PA, Goldbohm RA, van 't Veer P, Volovics A, Hermus RJ, Sturmans F. A large‐scale prospective cohort study on diet and cancer in The Netherlands. J Clin Epidemiol 1990;43:285‐95. Prentice RL. A case‐cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika 1986;73:1‐11. Peters S, Vermeulen R, Cassidy A, Mannetje A, van Tongeren M, Boffetta P, Straif K, Kromhout H; INCO Group. Comparison of exposure assessment methods for occupational carcinogens in a multi‐centre lung cancer case‐control study. Occup Environ Med 2011;68:148‐53. Offermans NS, Vermeulen R, Burdorf A, Peters S, Goldbohm RA, Koeman T, van Tongeren M, Kauppinen T, Kant I, Kromhout H, van den Brandt PA. Comparison of expert and job‐exposure matrix‐based retrospective exposure assessment of occupational carcinogens in the Netherlands Cohort Study. Occup Environ Med 2012;69:745‐51. Offermans NS, Vermeulen R, Burdorf A, Goldbohm RA, Kauppinen T, Kromhout H, van den Brandt PA. Occupational asbestos exposure and risk of pleural mesothelioma, lung cancer, and laryngeal cancer in the prospective Netherlands cohort study. J Occup Environ Med 2014;56:6‐19. Stewart PA, Herrick RF, Blair A, Checkoway H, Droz P, Fine L, Fischer L, Harris R, Kauppinen T, Saracci R. Highlights of the 1990 Leesburg, Virginia, International Workshop on Retrospective Exposure Assessment for Occupational Epidemiology Studies. Scand J Work Environ Health 1991;17:281–5. Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika 1982;69: 239‐41. Knol MJ, VanderWeele TJ. Recommendations for presenting analyses of effect modification and interaction. Int J Epidemiol 2012;41:514‐20. Kauppinen T, Toikkanen J, Pukkala E. From cross‐tabulations to multipurpose exposure information systems: a new job‐exposure matrix. Am J Ind Med 1998;33:409‐17. Morgan RW, Foliart DE, Wong O. Asbestos and gastrointestinal cancer. A review of the literature. West J Med 1985;143:60‐5. Gustavsson P, Jakobsson R, Johansson H, Lewin F, Norell S, Rutkvist LE. Occupational exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus: a case‐control study in Sweden. Occup Environ Med 1998;55:393‐400. Parent ME, Siemiatycki J, Fritschi L. Workplace exposures and oesophageal cancer. Occup Environ Med 2000;57:325‐34. Jansson C, Johansson AL, Bergdahl IA, Dickman PW, Plato N, Adami J, Boffetta P, Lagergren J. Occupational exposures and risk of esophageal and gastric cardia cancers among male Swedish construction workers. Cancer Causes Control 2005;16:755‐64. Santibañez M, Vioque J, Alguacil J, Barber X, García de la Hera M, Kauppinen T; PANESOES Study Group. Occupational exposures and risk of oesophageal cancer by histological type: a case‐control study in eastern Spain. Occup Environ Med 2008;65:774‐81. Gamble J. Risk of gastrointestinal cancers from inhalation and ingestion of asbestos. Regul Toxicol Pharmacol 2008;52:S124‐53. Frumkin H, Berlin J. Asbestos exposure and gastrointestinal malignancy review and meta‐analysis. Am J Ind Med 1988;14:79‐95. Jansson C, Plato N, Johansson AL, Nyrén O, Lagergren J. Airborne occupational exposures and risk of oesophageal and cardia adenocarcinoma. Occup Environ Med 2006;63:107‐12. Crew KD, Neugut AI. Epidemiology of gastric cancer. World J Gastroenterol 2006;12:354‐62. Sjödahl K, Jansson C, Bergdahl IA, Adami J, Boffetta P, Lagergren J. Airborne exposures and risk of gastric cancer: a prospective cohort study. Int J Cancer 2007;120:2013‐8. Jakobsson K, Albin M, Hagmar L. Asbestos, cement, and cancer in the right part of the colon. Occup Environ Med 1994;51:95‐101.

Chapter 4

23. Gerhardsson de Verdier M, Plato N, Steineck G, Peters JM. Occupational exposures and cancer of the colon and rectum. Am J Ind Med 1992;22:291‐303. 24. Ferrante D, Bertolotti M, Todesco A, Mirabelli D, Terracini B, Magnani C. Cancer mortality and incidence of mesothelioma in a cohort of wives of asbestos workers in Casale Monferrato, Italy. Environ Health Perspect 2007;115:1401‐5. 25. Liang PS, Chen TY, Giovannucci E. Cigarette smoking and colorectal cancer incidence and mortality: systematic review and meta‐analysis. Int J Cancer 2009;124:2406‐15. 26. Liddell FD, McDonald AD, McDonald JC. The 1891‐1920 birth cohort of Quebec chrysotile miners and millers: development from 1904 and mortality to 1992. Ann Occup Hyg 1997;41:13‐36. 27. Aliyu OA, Cullen MR, Barnett MJ, Balmes JR, Cartmel B, Redlich CA, Brodkin CA, Barnhart S, Rosenstock L, Israel L, Goodman GE, Thornquist MD, Omenn GS. Evidence for excess colorectal cancer incidence among asbestos‐exposed men in the Beta‐Carotene and Retinol Efficacy Trial. Am J Epidemiol 2005;162:868‐78.

Occupational asbestos exposure and gastrointestinal tract tumors

ONLINE SUPPLEMENTS *

Table S4.1 Esophageal, gastric and colorectal cancer cases classified by anatomic site or histological type. Esophageal cancer

Anatomic site† Squamous cell carcinoma (C15) Adenocarcinomas (C15)

Gastric cancer

Cardia adenocarcinomas (C16.0) Non‐cardia adenocarcinomas C16.1‐C16.9, including overlapping (C16.8) and not otherwise specified (C16.9) tumors Proximal colon (C153.0, 153.1, 153.4, 153.5, and 153.6) Distal colon (C153.2, 153.3, and 153.7) Colon not classifiable as either proximal or distal (C153.8 and 153.9)‡ Rectosigmoid (C154.0)‡ Rectum (C154.1)

Colorectal cancer

Histological type† 8050–8076 8140, 8141, 8190–8231, 8260–8263, 8310, 8430, 8480–8490, 8560, and 8570–8572

Incident cancer is monitored by annual record linkage to the Netherlands Cancer Registry and the Dutch 1,2 † Pathology Registry (PALGA). Following the International Classification of Diseases for Oncology, Third ‡ Edition. Cases that were diagnosed with rectosigmoid cancer or colon cancer not classifiable as either proximal or distal were included only when analyzing overall colorectal cancer risk. Table S4.2

Distribution of potential confounders and asbestos exposure among male subcohort members and cancer cases in the NLCS, 1986‐2003.

Age at baseline (yrs) (mean, SD) Family history of esophageal cancer (%) Family history of gastric cancer (%) Family history of colorectal cancer (%) Cigarette smoking (%) Never Former Current Number of cigarettes per day†‡ (mean, SD) Years of smoking†‡ (yrs) (mean, SD) Level of education (%) Lower vocational Secondary and medium vocational Higher vocational/university BMI‡ (kg/m2) (mean, SD) Alcohol consumption‡ (g/d) (mean, SD)

Subcohort Esophageal c ancer c ases Gastric cancer cases Colorectal cancer (n=2101) (n=211) (n=549) cases (n=1949) n Mean/ SD n Mean SD n Mean SD n Mean SD % / % / % / % 2101 61.3 4.2 211 61.9 4.2 549 62.2 4.1 1949 62.0 4.1 17 0.8 2 1.0 5 0.9 16 0.8 141

6.7

8.1

9.7

144

7.4

104

5.0

4.7

5.3

181

9.3

263 1079 759 1717

12.5 51.4 36.1 17.1

10.0 48.3 41.7 20.0

11.8

52 271 226 459

9.5 49.4 41.1 18.1

10.6

21 102 88 178

11.0

231 1128 590 1595

11.8 57.9 30.3 17.2 11.0

1802 33.6

11.9

187

36.0

11.2

487

35.8

11.3

1679

33.1 11.9

106 65 40 204 207

50.2 30.8 19.0 25.4 19.9

301 176 72 533 544

54.8 32.1 13.1 25.0 15.0

903 686 360 1883 1921

46.3 35.2 18.5 25.3 2.7 16.1 17.1

973 733 395 2035 2060

46.3 34.9 18.8 24.9 15.1

2.6 16.9

3.1 21.5

2.8 16.7

Chapter 4

Table S4.2 (continued)

Subcohort (n=2101) n Mean SD / % DOMJEM § Never exposed (%) 1496 71.2 Ever exposed§ (%) 605 28.8 Duration of exposure¶ (yrs) T1 (median:4) 212 35.0 T2 (median:18) (%) 199 32.9 T3 (median:37) (%) 194 32.1 Cumulative probability  intensity of exposure (unit‐years) T1 (median:4) (%) 215 35.5 T2 (median:20) (%) 191 31.6 T3 (median:38) (%) 199 32.9 ¶ Ever highly exposed (%) 49 2.3 ¶ Duration of high exposure (yrs) T1 (median:4) (%) 17 34.6 T2 (median:10.5) (%) 16 32.7 T3 (median:30.5) (%) 16 32.7 FINJEM Never exposed§ (%) 1559 74.2 Ever exposed§ (%) 542 25.8 ¶ Duration of exposure (yrs) T1 (median:7) (%) 184 33.9 T2 (median:25) (%) 176 32.5 T3 (median:37) (%) 182 33.6 Cumulative probability  intensity of exposure (f‐y/ml) T1 (median:0.20) (%) 181 33.4 T2 (median:1.59) (%) 180 33.2 T3 (median:6.60) (%) 181 33.4 ¶ Ever highly exposed (%) 290 13.8 Duration of high exposure¶ (yrs) T1 (median:6) (%) 100 34.4 T2 (median:20) (%) 95 32.8 T3 (median:35) (%) 95 32.8 *

Esophageal cancer cases (n=211) n Mean SD / % 139 65.9 72 34.1 29 40.3 18 25.0 25 34.7 27 37.5 20 27.8 25 34.7 8 3.8 3 37.5 2 25.0 3 37.5 153 72.5 58 27.5 24 41.4 16 27.6 18 31.0 20 34.5 21 36.2 17 29.3 31 14.7 11 35.5 9 29.0 11 35.5

Gastric cancer cases Colorectal cancer (n=549) cases (n=1949) n Mean SD n Mean SD / % / % 389 70.9 1415 72.6 160 29.1 534 27.4 35 21.9 163 30.5 64 40.0 177 33.2 61 38.1 194 36.3 40 25.0 171 32.0 54 33.8 157 29.4 66 41.2 206 38.6 19 3.5 61 3.1 2 10.5 22 36.1 7 36.9 15 24.6 10 52.6 24 39.3 399 72.7 1458 74.8 150 27.3 491 25.2 50 33.4 157 32.0 50 33.3 162 33.0 50 33.3 172 35.0 49 32.7 157 32.0 42 28.0 170 34.6 59 39.3 164 33.4 87 15.9 270 13.9 22 25.3 83 30.8 34 39.1 93 34.4 31 35.6 94 34.8

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. ‡ Among former and current smokers only. Sum of categories deviates from total number because of § missing values. Exposure based on the cumulative probabilityintensity of exposure (unit‐years or f‐y/ml). ¶ Exposure based on the probabilityintensity of exposure (unit‐years or f‐y/ml) per job. f‐y/ml, fiber‐ years/ml; NLCS, Netherlands Cohort Study. †

11 8 37 13 164 50

3151 781 3113 773

Never No. of cases

3151 781

Person‐ years

1 (ref)§ 1.31 (0.83‐2.06) 1.31 (0.83‐2.06)

1 (ref)† 1.44 (0.70‐2.94) 1.44 (0.70‐2.94)

1 (ref)* 3.25 (1.21‐8.77) 3.25 (1.21‐8.77)

HR (95%CI)

738 281

173 78

64 28

1.47 (1.15‐1.88) 1.30 (0.98‐1.72) 0.88 (0.71‐1.08)

1.61 (1.07‐2.41) 1.57 (1.00‐2.47) 1.01 (0.74‐1.39)

1.83 (0.93‐3.60) 1.86 (0.88‐3.93) 1.04 (0.65‐1.68)

6127 2555

6142 2555

356 135

131 54

47 29

Current Person‐ No. of years cases

HR (95%CI) for current cigarette smoking within strata of asbestos exposure 1.99 (0.95‐4.16) 0.95 (0.38‐2.37)

0.09 0.61 0.29

P for multiplicative interaction

1.15 (0.88‐1.52) 1.47 (1.15‐1.88) 1.15 (0.88‐1.52) 1.13 (0.81‐1.57) 1.27 (0.78‐2.06) 1.00 (0.60‐1.69) 0.97 (0.72‐1.29)

1.70 (1.10‐2.62) 1.61 (1.07‐2.41) 1.70 (1.10‐2.62) 1.68 (1.01‐2.79) 1.09 (0.54‐2.19) 1.09 (0.49‐2.39) 0.98 (0.66‐1.47)

HR (95%CI) for former cigarette smoking within strata of asbestos exposure 1.99 (0.95‐4.16) 1.83 (0.93‐3.60) 3.06 (1.38‐6.80) 0.56 (0.22‐1.40) 1.56 (0.93‐2.62) HR (95%CI)

Measure of interaction on additive scale: Relative excess risk due to interaction, RERI (95%CI) Former cigarette smoking Current cigarette smoking ‐2.23 (‐8.55‐0.09) ‐1.18 (‐7.30‐1.21) ‐0.47 (‐1.99‐0.49) ‐0.46 (‐2.12‐0.58) ‐0.48 (‐1.38‐0.10) ‐0.33 (‐1.21‐0.27)

9782 4179

9923 4229

Cigarette smoking Former Person‐ No. of HR (95%CI) years cases

* Adjusted for age (yrs), family history of esophageal cancer (yes/no), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower vocational, secondary and medium vocational, and higher vocational/university), BMI (kg/m2), and alcohol consumption (g/d). † Adjusted for age (yrs), family history of gastric cancer (yes/no), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower vocational, secondary and medium vocational, and higher vocational/university), BMI (kg/m2), and alcohol consumption (g/d). ‡ Adjustment for number of cigarettes smoked per day and years of smoking cigarettes not possible due to collinearity. § Adjusted for age (yrs), family history of colorectal cancer (yes/no), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower 2 vocational, secondary and medium vocational, and higher vocational/university), BMI (kg/m ), and alcohol consumption (g/d). CI, confidence interval; f‐y/ml, fiber‐ years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Esophageal cancer Gastric cancer Colorectal cancer

Esophageal cancer Never exposed Ever exposed HR (95%CI) for asbestos exposure within strata of cigarette smoking Gastric cancer Never exposed Ever exposed HR (95%CI) for asbestos exposure within strata of cigarette smoking‡ Colorectal cancer Never exposed Ever exposed HR (95%CI) for asbestos exposure within strata of cigarette smoking

Asbestos Exposure

Hazard ratios (HRs) and 95% CIs for esophageal, gastric, and colorectal cancer for asbestos exposure (yes/no), by smoking; estimated with DOMJEM in the NLCS, 1986‐2003.

Table S4.3

Occupational asbestos exposure and gastrointestinal tract tumors

Chapter 4

REFERENCES 1.

Casparie M, Tiebosch AT, Burger G, Blauwgeers H, van de Pol A, van Krieken JH, Meijer GA. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol 2007;29:19‐24. Van den Brandt PA, Schouten LJ, Goldbohm RA, Dorant E, Hunen PM. Development of a record linkage protocol for use in the Dutch Cancer Registry for Epidemiological Research. Int J Epidemiol 1990;19: 553‐8.

Occupational asbestos exposure and gastrointestinal tract tumors

Sensitivity analysis In order to provide insight into the methodological uncertainty associated with the choice of JEM, occupational asbestos exposure was also estimated by linkage to a Finnish job‐exposure matrix (FINJEM)1 as described previously.2 Results for FINJEM were essentially similar to those for DOMJEM, except for the (prolonged) highly exposed subjects (see supporting Tables S4.4‐4.6). In previous analyses in the NLCS, we already examined both JEMs with respect to their capacity for retrospective occupational exposure assessment and risk prediction. Both JEMs showed rather similar agreement with case‐by‐case expert assessment and moderate agreement amongst each other.2 When studying the relation between asbestos and risk of mesothelioma, lung and laryngeal cancer, both JEMs seemed to have their particular strengths. FINJEM may be better in discriminating between ever and never‐exposed while DOMJEM appeared better in singling out the prolonged highly exposed subjects.3 The latter is supported by the present study, in which DOMJEM was the only one to show significantly increased HRs for the (prolonged) highly exposed subjects. It has to be noted, however, that ‐ though HRs were sometimes markedly increased ‐ associations were in most instances only marginally significant. Nevertheless, small differences in the ability of both JEMs to discriminate between never and ever exposed or low and high exposure may lead to slightly different results and judgment of a substance as possibly related with the endpoint of interest. Some caution seems thus appropriate when interpreting results from JEM‐based exposure assessment studies. Incorporation of multiple JEMs in epidemiological studies should be encouraged as it reveals some of the uncertainty associated with the choice of exposure method and therefore aids in the interpretation of data.

86 134 53 23 14 16 187 20 18 15 187 159 28 10 9 9 187

No. of cases 1 (ref) 1.28 (0.91‐1.80) 1.59 (0.99‐2.57) 1.14 (0.64‐2.04) 1.09 (0.63‐1.89) 0.427 1.06 (0.94‐1.19) 1.42 (0.86‐2.35) 1.30 (0.77‐2.20) 1.12 (0.64‐1.96) 0.304 1.02 (0.97‐1.07) 1 (ref) 1.30 (0.83‐2.05) 1.23 (0.62‐2.44) 1.25 (0.61‐2.56) 1.17 (0.57‐2.39) 0.433 1.06 (0.90‐1.26) †

1 (ref) 1.19 (0.84‐1.70) 1.53 (0.94‐2.49) 0.97 (0.53‐1.78) 1.05 (0.59‐1.84) 0.682 1.03 (0.91‐1.17) 1.45 (0.87‐2.43) 1.17 (0.68‐2.00) 0.98 (0.55‐1.75) 0.642 1.00 (0.95‐1.06) 1 (ref) 1.11 (0.70‐1.74) 1.12 (0.56‐2.25) 1.00 (0.47‐2.13) 1.03 (0.50‐2.13) 0.857 1.00 (0.84‐1.20)

Esophageal cancer HR† (95%CI) HR‡ (95%CI) 45 16 11 3 2 61 3 8 5 61 48 13 6 3 4 61

No. of cases 1 (ref) 1.18 (0.62‐2.22) 2.57 (1.22‐5.38) 0.67 (0.19‐2.29) 0.39 (0.09‐1.63) 0.316 0.88 (0.70‐1.10) 0.71 (0.21‐2.36) 1.88 (0.79‐4.43) 1.01 (0.39‐2.64) 0.467 0.99 (0.92‐1.07) 1 (ref) 1.70 (0.86‐3.36) 2.81 (1.12‐7.07) 1.38 (0.39‐4.86) 1.50 (0.51‐4.37) 0.178 1.10 (0.87‐1.40)

ESCC HR‡ (95%CI)

89 37 12 11 14 126 17 10 10 126 111 15 4 6 5 126

No. of cases

1 (ref) 1.21 (0.80‐1.84) 1.13 (0.60‐2.12) 1.11 (0.56‐2.18) 1.41 (0.76‐2.61) 0.275 1.09 (0.94‐1.27) 1.80 (1.03‐3.16) 0.90 (0.45‐1.79) 0.97 (0.48‐1.94) 0.912 1.01 (0.94‐1.08) 1 (ref) 0.84 (0.46‐1.52) 0.59 (0.21‐1.68) 0.89 (0.36‐2.19) 0.78 (0.30‐2.04) 0.491 0.94 (0.74‐1.20)

EAC HR‡ (95%CI)

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. Adjusted for age (yrs) and family history of esophageal cancer ‡ (yes/no). Adjusted for age (yrs), family history of esophageal cancer (yes/no), smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower vocational, secondary and medium vocational, and higher 2 § ¶ vocational/university), BMI (kg/m ), and alcohol consumption (g/d). Exposure based on the cumulative probabilityintensity of exposure (f‐y/ml). Exposure based on the probabilityintensity of exposure (f‐y/ml) per job. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Never exposed§ Ever exposed§ Duration of exposure¶ (yrs) T1 (median:7) T2 (median:25) T3 (median:37) P for trend Continuous, per 10 yrs Cumulative probabilityintensity of exposure (f‐y/ml) T1 (median:0.20) T2 (median:1.59) T3 (median:6.60) P for trend Continuous, per 1 fiber‐year Never highly exposed Ever highly exposed (median:3.79 f‐y/ml) Duration of high exposure¶ (yrs) T1 (median:6) T2 (median:20) T3 (median:35) P for trend Continuous, per 10 yrs

Person years in subcohort 20226 6555 2308 1983 2264 26781 2199 2254 2102 26781 23203 3578 1330 1116 1131 26781

Hazard ratios (HRs) and 95% CIs for overall esophageal cancer and subtypes for categories of asbestos exposure*, estimated with FINJEM in the NLCS, 1986‐2003.

Table S4.4

Chapter 4

Person years in subcohort 20226 6555 2308 1983 2264 26781 2199 2254 2102 26781 23203 3578 1330 1116 1131 26781 351 135 46 41 48 486 47 36 52 486 408 78 20 29 29 486

No. of cases 1 (ref) 1.25 (1.00‐1.58) 1.22 (0.86‐1.74) 1.29 (0.88‐1.87) 1.26 (0.89‐1.77) 0.068 1.08 (1.00‐1.16) 1.28 (0.89‐1.82) 1.01 (0.69‐1.48) 1.48 (1.06‐2.07) <0.05 1.03 (1.00‐1.06) 1 (ref) 1.36 (1.03‐1.81) 0.99 (0.60‐1.64) 1.59 (1.02‐2.48) 1.47 (0.95‐2.28) <0.05 1.13 (1.01‐1.26) †

1 (ref) 1.11 (0.87‐1.42) 1.14 (0.80‐1.63) 1.09 (0.74‐1.61) 1.09 (0.76‐1.57) 0.471 1.03 (0.95‐1.12) 1.19 (0.83‐1.72) 0.85 (0.57‐1.27) 1.28 (0.91‐1.81) 0.348 1.02 (0.98‐1.05) 1 (ref) 1.17 (0.87‐1.58) 0.87 (0.52‐1.45) 1.37 (0.86‐2.16) 1.28 (0.82‐1.99) 0.159 1.07 (0.96‐1.20)

Gastric cancer HR† (95%CI) HR‡ (95%CI) 105 38 15 6 17 143 15 10 13 143 122 21 7 7 7 143

No. of cases 1 (ref) 1.07 (0.70‐1.64) 1.18 (0.66‐2.11) 0.54 (0.22‐1.31) 1.44 (0.80‐2.59) 0.554 1.04 (0.89‐1.22) 1.32 (0.72‐2.40) 0.80 (0.39‐1.62) 1.10 (0.60‐2.04) 0.873 0.98 (0.91‐1.05) 1 (ref) 1.06 (0.63‐1.79) 1.00 (0.45‐2.26) 1.03 (0.44‐2.37) 1.10 (0.50‐2.45) 0.790 1.02 (0.83‐1.26)

GCA HR‡ (95%CI)

246 97 31 35 31 343 32 26 39 343 286 57 13 22 22 343

No. of cases

1 (ref) 1.12 (0.85‐1.48) 1.11 (0.73‐1.69) 1.31 (0.86‐2.00) 0.97 (0.63‐1.49) 0.626 1.02 (0.93‐1.13) 1.13 (0.74‐1.74) 0.88 (0.55‐1.38) 1.35 (0.91‐2.01) 0.315 1.03 (0.99‐1.06) 1 (ref) 1.21 (0.86‐1.70) 0.81 (0.44‐1.50) 1.50 (0.90‐2.51) 1.33 (0.80‐2.21) 0.145 1.09 (0.96‐1.24)

GNCA HR‡ (95%CI)

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. Adjusted for age (yrs) and family history of gastric cancer ‡ (yes/no). Adjusted for age (yrs), family history of gastric cancer (yes/no), smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower vocational, secondary and medium vocational, and higher vocational/university), 2 § ¶ BMI (kg/m ), and alcohol consumption (g/d). Exposure based on the cumulative probabilityintensity of exposure (f‐y/ml). Exposure based on the probabilityintensity of exposure (f‐y/ml) per job. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Table S4.5 Hazard ratios (HRs) and 95% CIs for overall gastric cancer and subtypes for categories of asbestos exposure , estimated with FINJEM in the NLCS, 1986‐ 2003.

Occupational asbestos exposure and gastrointestinal tract tumors

88 1 (ref) 1.09 (0.93‐1.27) 1.02 (0.80‐1.31) 1.20 (0.93‐1.55) 1.06 (0.82‐1.36) 0.322 1.04 (0.98‐1.10) 1.04 (0.81‐1.34) 1.14 (0.89‐1.46) 1.08 (0.84‐1.40) 0.305 1.01 (0.98‐1.03) 1 (ref) 1.10 (0.90‐1.35) 0.96 (0.69‐1.33) 1.25 (0.90‐1.75) 1.09 (0.78‐1.52) 0.327 1.03 (0.95‐1.12)

72 81 80 1724

0.92 (0.66‐1.28) 1.16 (0.82‐1.63) 1.06 (0.75‐1.50) 0.565 1.02 (0.93‐1.11)

1 (ref) 1.05 (0.89‐1.25) 0.98 (0.76‐1.27) 1.16 (0.89‐1.51) 1.04 (0.81‐1.35) 0.443 1.03 (0.97‐1.09) 1.03 (0.80‐1.34) 1.08 (0.84‐1.40) 1.05 (0.81‐1.36) 0.534 1.00 (0.98‐1.03) 1 (ref) 1.05 (0.85‐1.30)

Total colorectal cancer HR† (95%CI) HR‡ (95%CI)

1295 429 139 141 149 1724 139 152 138 1724 1491 233

No. of cases

41 49 39 1113

862 251 84 84 83 1113 88 89 74 1113 984 129

No. of cases

0.86 (0.58‐1.28) 1.16 (0.78‐1.71) 0.83 (0.54‐1.26) 0.585 0.98 (0.89‐1.08)

1 (ref) 0.99 (0.82‐1.20) 0.94 (0.70‐1.26) 1.12 (0.83‐1.52) 0.94 (0.69‐1.27) 0.943 1.00 (0.93‐1.07) 1.03 (0.77‐1.38) 1.04 (0.78‐1.40) 0.90 (0.66‐1.23) 0.695 1.00 (0.97‐1.03) 1 (ref) 0.95 (0.74‐1.21)

Colon HR‡ (95%CI)

22 28 16 503

377 126 43 42 41 503 41 49 36 503 437 66 1.04 (0.63‐1.70) 1.49 (0.94‐2.38) 0.74 (0.42‐1.31) 0.968 0.99 (0.88‐1.12)

1 (ref) 1.13 (0.89‐1.44) 1.12 (0.77‐1.61) 1.26 (0.86‐1.85) 1.04 (0.71‐1.52) 0.504 1.03 (0.95‐1.13) 1.10 (0.75‐1.62) 1.31 (0.91‐1.88) 0.99 (0.66‐1.47) 0.532 1.02 (0.98‐1.06) 1 (ref) 1.10 (0.81‐1.50)

Proximal colon No. of HR‡ (95%CI) cases

19 20 20 568

448 120 41 40 39 568 47 39 34 568 509 59 0.79 (0.47‐1.33) 0.91 (0.54‐1.54) 0.84 (0.50‐1.43) 0.394 0.96 (0.84‐1.09)

1 (ref) 0.92 (0.73‐1.18) 0.87 (0.60‐1.27) 1.06 (0.72‐1.55) 0.87 (0.59‐1.29) 0.642 0.98 (0.90‐1.07) 1.05 (0.74‐1.51) 0.90 (0.61‐1.33) 0.81 (0.54‐1.21) 0.325 0.97 (0.93‐1.01) 1 (ref) 0.85 (0.62‐1.17)

Distal colon No. of HR‡ (95%CI) cases

25 24 24 425

294 131 44 40 47 425 38 47 46 425 352 73

1.17 (0.73‐1.87) 1.25 (0.76‐2.05) 1.23 (0.77‐1.99) 0.204 1.05 (0.93‐1.18)

1 (ref) 1.26 (0.98‐1.63) 1.24 (0.86‐1.80) 1.26 (0.85‐1.88) 1.29 (0.89‐1.86) 0.079 1.08 (0.99‐1.18) 1.14 (0.77‐1.69) 1.27 (0.87‐1.85) 1.39 (0.96‐2.00) <0.05 (0.394) 1.01 (0.98‐1.04) 1 (ref) 1.27 (0.93‐1.73)

Rectum No. of HR‡ (95%CI) cases

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. † Adjusted for age (yrs) and family history of colorectal cancer (yes/no). ‡ Adjusted for age (yrs), family history of colorectal cancer (yes/no), smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (lower vocational, secondary and medium vocational, and higher vocational/university), BMI (kg/m2), and alcohol consumption (g/d). § Exposure based on the cumulative probabilityintensity of exposure (f‐y/ml). ¶ Exposure based on the probabilityintensity of exposure (f‐y/ml) per job. # Trends over the exposed subjects were added only if trends over all subjects were statistically significant in the full covariate model. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Person years in subcohort § Never exposed 20038 Ever exposed§ 6492 Duration of exposure¶ (yrs) T1 (median:7) 2288 T2 (median:25) 1975 T3 (median:37) 2229 P for trend Continuous, per 10 yrs 26530 Cumulative probabilityintensity of exposure (f‐y/ml) T1 (median:0.20) 2183 T2 (median:1.59) 2249 T3 (median:6.60) 2060 P for trend (over the exposed only)# Continuous, per 1 fiber‐year 26530 Never highly exposed 22980 Ever highly exposed 3550 (median:3.79 f‐y/ml) Duration of high exposure¶ (yrs) T1 (median:6) 1337 T2 (median:20) 1089 T3 (median:35) 1124 P for trend Continuous, per 10 yrs 26530

Hazard ratios (HRs) and 95% CIs for overall colorectal cancer and subtypes for categories of asbestos exposure*, estimated with FINJEM in the NLCS, 1986‐2003.

Table S4.6

Chapter 4

Occupational asbestos exposure and gastrointestinal tract tumors

REFERENCES 1. 2.

Kauppinen T, Toikkanen J, Pukkala E. From cross‐tabulations to multipurpose exposure information systems: a new job‐exposure matrix. Am J Ind Med 1998;33:409‐17. Offermans NS, Vermeulen R, Burdorf A, Peters S, Goldbohm RA, Koeman T, van Tongeren M, Kauppinen T, Kant I, Kromhout H, van den Brandt PA. Comparison of expert and job‐exposure matrix‐based retrospective exposure assessment of occupational carcinogens in the Netherlands Cohort Study. Occup Environ Med 2012;69:745‐51. Offermans NS, Vermeulen R, Burdorf A, Goldbohm RA, Kauppinen T, Kromhout H, van den Brandt PA. Occupational asbestos exposure and risk of pleural mesothelioma, lung cancer, and laryngeal cancer in the prospective Netherlands cohort study. J Occup Environ Med 2014;56:6‐19.

Chapter 4

Asbestos exposure and tumors of the oral cavity and pharynx

Chapter 5

Occupational asbestos exposure and risk of oral cavity and pharyngeal cancer in the prospective Netherlands Cohort Study Nadine S.M. Offermans Roel Vermeulen Alex Burdorf R. Alexandra Goldbohm András P. Keszei Susan Peters Timo Kauppinen Hans Kromhout Piet A. van den Brandt Scandinavian Journal of Work, Environment & Health. 2014;40:420‐7

Chapter 5

ABSTRACT Objectives The evidence for an association between occupational asbestos exposure and pharyngeal cancer (PhC) is limited, while for oral cavity cancer (OCC) the literature is even sparser. We studied OCC and PhC risk both separately and combined (OCPC) in relation to occupational asbestos exposure, specifically addressing the influence of potential confounders, the existence of an exposure–response relation, and the presence of interaction between asbestos and smoking. Methods Using the prospective Netherlands Cohort Study (N=58 279 men, aged 55–69 years), we estimated asbestos exposure by linkage to a general population job‐exposure matrix (DOMJEM) and a Finnish job‐exposure matrix (FINJEM). After 17.3 years of follow‐up, 58 OCC and 53 PhC cases were available for analysis. Results No association between asbestos and risk of OCC was observed for either JEM. Hazard ratios (HR) of PhC and OCPC increased after adjusting for confounders, particularly alcohol consumption and socioeconomic status. For PhC, a multivariable‐adjusted increased HR was observed for “ever” versus “never” exposed to asbestos [HR 2.20, 95% confidence interval (95% CI) 1.08–4.49] when using FINJEM, but a trend of increased risks with higher cumulative exposure could not be demonstrated for either JEM. Results for OCPC showed patterns similar to those observed for PhC. None of the cancers showed a significant interaction between asbestos and smoking. Conclusions This prospective population‐based study showed no convincing evidence of an association between asbestos and risk of OCC, PhC, and OCPC as an exposure–response relation was lacking, and results were not robust against the use of different JEM. However, the potentially increased HR of PhC and OCPC observed in this and previous studies warrant further research.

Asbestos exposure and tumors of the oral cavity and pharynx

INTRODUCTION Together, oral cavity cancer (OCC) and pharyngeal cancer (PhC) are the eighth most frequently occurring neoplasms worldwide, giving rise to 482 000 new cases in 2008.1 The main risk factors for OCC and PhC are (excessive) alcohol consumption and all forms of tobacco use, while fruits and vegetables have been associated with a decreased risk.2 Occupational carcinogens, such as polycyclic aromatic hydrocarbons3 and solvents,4 may also be involved in the etiology of both cancers, but the strongest suggestion of a possible association focuses on asbestos exposure.3,5,6 There are, however, no supportive data from animal studies and a clear exposure–response relation for asbestos exposure is lacking.5 As such, there is still debate whether asbestos exposure increases the risk of developing OCC and PhC, and if so, whether risk increases with increasing level of cumulative exposure. Besides questions on the strength of a possible association, there are additional questions relating to (i) the association with OCC and PhC separately, as they have often been combined in previous research; (ii) possible confounding by other risk factors, such as smoking and alcohol consumption, as most previous studies were performed in an industry‐setting and were unable to adjust for these factors; and (iii) the presence of an interaction between exposure to asbestos and smoking in relation to the risk of OCC and PhC, as has been suggested for other asbestos‐related cancers.5,6 Population‐based studies are well‐suited to address these questions given their overall wide range in exposure levels, possibility to control for potential confounders, and large size. The prospective Netherlands Cohort Study (NLCS) is a population‐based study, which started in 1986 among 120 852 men and women of the general population.7 Within the framework of the current investigation, we had the following objectives: 1. to investigate the association between occupational asbestos exposure and risk of OCC and PhC both separately and combined, paying special attention to the existence of an exposure–response relation and the influence of potential confounding on risk estimates; 2. to study the presence of an interaction between asbestos and smoking in relation to the risk of OCC and PhC, both separately and combined. As the proportion of long‐term employed women was rather low (resulting in <1% being occupationally exposed to asbestos), this study was conducted only among men.

METHODS Study population and cancer follow‐up The study design and data collection strategies for the NLCS have been described in detail previously.7 In brief, the NLCS started in September 1986 when 58 279 men and

Chapter 5

62 573 women aged 55–69 years, originating from 204 municipalities in the Netherlands with computerized population registries, were enrolled in the cohort (response rate 35.5%; 34.5% among men and 36.6% among women). At baseline, participants completed a self‐administered questionnaire on dietary habits and lifestyle, occupational history, and other potential risk factors for cancer.7 For reasons of efficiency in questionnaire processing and follow‐up, the case‐cohort approach was used.8 Incident cases were enumerated from the entire cohort, whereas the accumulated person‐years at risk in the entire cohort were estimated from a random subcohort of 5000 subjects (2411 men and 2589 women), selected immediately after baseline. This subcohort is being followed‐up for vital status information while the entire cohort is being monitored for incident cancer by annual record linkage to the Netherlands Cancer Registry and the Dutch Pathology Registry (PALGA).9,10 For current analyses, a total of 17.3 years of follow‐up (baseline to December 2003) was available. Completeness of incident cancer coverage was estimated to be almost 100%11 as was the completeness of follow‐up of the subcohort. The institutional review boards of the Netherlands Organization for Applied Scientific Research TNO (Zeist) and Maastricht University approved the NLCS. End points for this study were incident, microscopically confirmed OCC and PhC cases classified by anatomic site or histological type, as defined by the International Classification of Diseases for Oncology, Third Edition. Tumor assignment followed the INHANCE (International Head and Neck Cancer Epidemiology Consortium) collaboration classification12 and consisted of the following categories: (i) oral cavity (includes lip, tongue, gum, floor of mouth, and hard palate): codes C00.3–C00.9, C02.0–C02.3, C03.0, C03.1, C03.9, C04.0, C04.1, C04.8, C04.9, C05.0, C06.0–C06.2, C06.8, and C06.9; (ii) pharynx, consisting of (a) oropharynx (includes base of tongue, lingual tonsil, soft palate, uvula, tonsil, and oropharynx): codes C01.9, C02.4, C05.1, C05.2, C09.0, C09.1, C09.8, C09.9, C10.0–C10.4, C10.8, and C10.9; and (b) hypopharynx (includes pyriform sinus and hypopharynx): codes C12.9, C13.0–C13.2, C13.8, and C13.9; 3) oral cavity, pharynx unspecified or overlapping: codes C02.8, C02.9, C05.8, C05.9, C14.0, C14.2, and C14.8. As most tumors in our study originated from squamous cell tissue (89.6%), analyses were restricted to cases with squamous cell carcinomas. Other than skin cancer, all prevalent cases at baseline were excluded, leaving 2336 male subcohort members, 71 OCC and 74 PhC cases (50 oropharynx and 24 hypopharynx). Cases that could not be classified as oropharynx, hypopharynx, or oral cavity (n=2) were included only when analyzing OCC and PhC together (OCPC; n=147). Due to the low number of oro‐ and hypopharyngeal cancer cases, no subtype analyses were performed. Subjects without any, or only uncodable, information on occupational history or who never worked professionally were omitted from the analyses. As a

Asbestos exposure and tumors of the oral cavity and pharynx

result, 2101 male subcohort members, 63 OCC, 67 PhC, and 132 OCPC cases were available for analysis after 17.3 years of follow‐up.

Occupational exposure assessment Information on lifetime occupational history until 1986 was obtained from the questionnaire completed on study enrolment. Questions concerned the job title, name and type of the company, products made in the department, and period of employment. For all subjects, the job code was assessed for each of the maximally five occupations subjects could enter between starting work and 1986. Job‐exposure matrix We applied two job‐exposure matrices (JEM), a general population JEM from the Netherlands (DOMJEM) and a Finnish JEM (FINJEM) as described previously,13 in order to provide insight into the methodological uncertainty associated with choice of JEM with respect to asbestos exposure. DOMJEM and FINJEM showed moderate agreement amongst each other and rather similar agreement with case‐by‐case expert assessment. Briefly, occupational exposure experts in the Netherlands developed DOMJEM for application in general population studies. It contains a combined measure of the probability×intensity of exposure, which is ordinal (no, low, or high exposure) with a weighting of respectively 0, 1, or 4.14 FINJEM was constructed for exposure assessment in large register‐based studies and is based on both expert assessment and exposure measurements. It contains continuous estimates of the prevalence and intensity of exposure both separately and combined, and contains a time axis.15 Although FINJEM was constructed for Finland, exposure estimates were not adapted to Dutch occupational circumstances before application in the NLCS. Asbestos exposure variables Several exposure variables were defined, which we have described in more detail elsewhere: ever versus never occupationally exposed to asbestos (yes/no), cumulative exposure [CE; a combined measure of the probability (P), intensity (I), and duration (years) of exposure, measured in fiber‐years/ml (f‐y/ml) (FINJEM) or unit‐years (DOMJEM)], ever versus never highly exposed to asbestos (yes/no), and duration of high exposure (years).16 Participants were classified into never‐exposed subjects and tertiles of those exposed to asbestos based on the distribution among the subcohort for the CE (never exposed= reference group) and for the duration of high exposure (never highly exposed= reference group). Continuous variables were also used; for the CE, an increment of 1 unit‐year (DOMJEM) or 1 f‐y/ml (FINJEM) was used.

Chapter 5

For the percentage of the population for whom some information on occupational history could not be coded, exposure to asbestos was set to zero for the period with unclear exposure.

Statistical analysis Cox proportional hazards models were used to estimate age‐adjusted as well as multivariable‐adjusted hazard ratios (HR) and corresponding 95% confidence intervals (95% CI). The total person‐years at risk were estimated from the subcohort,17 and we estimated standard errors using a robust covariance matrix estimator to account for increased variance due to sampling the subcohort from the entire cohort.18 The covariates included in the multivariable‐adjusted models were either a priori‐ selected risk factors based on the literature or variables that changed the age‐adjusted regression coefficients by >10% (using a backwards stepwise procedure). Smokeless tobacco has not been included in the multivariable‐adjusted model as the total number of cases that used this form of tobacco was too low (n=4) to be of any influence as a confounder in the NLCS. Polycyclic aromatic hydrocarbons and solvents have also not been included in the final model as these exposures hardly changed the age‐adjusted regression coefficients. For all endpoints, the full covariate model consisted of smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), socioeconomic status (by level of education; lower vocational, secondary and medium vocational, and higher vocational/university), alcohol consumption, and consumption of vegetables. Information on consumption of vegetables was gathered in the dietary section of the baseline questionnaire. This section consisted of a 150‐item semi‐quantitative food‐ frequency questionnaire, which concentrated on habitual consumption during the year preceding the start of the study. All covariates were entered into the models as continuous variables, except for smoking status and socioeconomic status. To enable comparison, the models adjusted for age and family history of cancer were restricted to subjects included in the multivariable‐ adjusted analyses (ie, with no missing values on confounding variables), which left 1858 subcohort members, 58 OCC, 53 PhC, and 113 OCPC cases for analyses. For each analysis, scaled Schoenfeld residuals were used to test the proportional hazards assumption. 19 Trends for all subjects were evaluated with the Wald test by assigning subjects the median value for each level of the categorical variable among the subcohort members, and this variable was entered as a continuous term in the Cox regression model. Furthermore, we tested for a possible interaction between occupational exposure to asbestos (yes/no) and smoking status (never/former/current) in relation to OCC, PhC, and OCPC. As testing for departure from additivity was not possible due to low numbers, we only studied statistically significant departure from multiplicativity by including an interaction term in the Cox regression model. All tests (2‐tailed) were

Asbestos exposure and tumors of the oral cavity and pharynx

performed using Stata, version 10 (Stata Corp, College Station, TX, USA), and differences were regarded as statistically significant at P<0.05.

RESULTS The distribution of occupational exposure to asbestos and potential confounders among male subcohort members and cancer cases in the NLCS is presented in Table 5.1. Overall, OCC and PhC cases more often smoked cigarettes, smoked more cigarettes per day and for a longer period of time, and consumed more alcohol per day than subcohort members. OCC and PhC cases generally had a higher socioeconomic status than subcohort members. According to both JEM, OCC cases were overall less often exposed to asbestos than subcohort members, while PhC cases were more often exposed. As the number of highly exposed subjects was very low, certainly for DOMJEM, no further results will be presented for ever versus never highly exposed and the duration of high exposure. Overall, most asbestos exposure variables showed no association with OCC, PhC, and OCPC (Table 5.2). Adjusting for potential confounders was generally of minor influence, except for alcohol consumption and socioeconomic status, which increased the HR of PhC and OCPC. Therefore, only multivariable‐adjusted results are presented below. For OCC, no associations were observed when using DOMJEM or FINJEM for ever versus never exposed (HR 0.86 (95% CI 0.41–1.82) and HR 1.18 (95% CI 0.53–2.61), respectively). For PhC, an elevated risk was observed for ever versus never exposed when using FINJEM (HR 2.20 (95% CI 1.08–4.49)), but not when using DOMJEM (HR 1.16 (95% CI 0.56–2.38)). No trends of increased risks with higher cumulative exposure could be demonstrated when using DOMJEM or FINJEM. Certainly when using FINJEM, risk in tertile 3 of cumulative exposure was lower (HR 0.30 (95% CI 0.04–2.46)) than in tertiles 1 and 2 (HR 4.04 (95% CI 1.69–9.64)) and (HR 3.01 (95% CI 1.14–7.96)). For OCPC, results showed patterns similar to those observed for PhC, with tertiles 1 and/or 2 of the cumulative exposure showing an association (tertile 2 when using DOMJEM: HR 2.56 (95% CI 1.28–5.09) and tertiles 1 and 2 when using FINJEM: HR 2.64 (95% CI 1.34–5.19) and HR 2.15 (95% CI 1.02–4.52), respectively). The P for trend was non‐significant for any of the cancer endpoints. For ease of presentation, we will only present interaction results for FINJEM (Table 5.3). Although the stratum‐specific HR may be suggestive of a negative interaction between asbestos and smoking for OCC, none of the cancers showed a statistically significant interaction. Analyses with DOMJEM revealed no interactions (data not shown).

Age at baseline (yrs) (mean, SD) Cigarette smoking (%) Never Former Current Number of cigarettes per day†‡ (mean, SD) Years of smoking†‡ (yrs) (mean, SD) Level of education (%) Lower vocational Secondary and medium vocational Higher vocational/university Alcohol consumption‡ (g/d) (mean, SD) Consumption of vegetables (g/d) (mean, SD) DOMJEM Never exposed§ (%) Ever exposed§ (%) Cumulative probabilityintensity of exposure (unit‐years) T1 (median:4) (%) T2 (median:20) (%) T3 (median:38) (%) Ever highly exposed¶ (%) Duration of high exposure¶ (yrs) T1 (median:4) (%) T2 (median:10.5) (%) T3 (median:30.5) (%)

n 2101 263 1079 759 1717 1802 973 733 395 2060 2101 1496 605 215 191 199 49 17 16 16

Subcohort (n=2101) Mean/% 61.3 12.5 51.4 36.1 17.1 33.6 46.3 34.9 18.8 15.1 191.2 71.2 28.8 35.5 31.6 32.9 2.3 34.6 32.7 32.7

16.9 86.4

10.6 11.9

SD 4.2

n 63 11 14 38 47 50 20 26 17 62 63 50 13 5 7 1 0 0 0 0

Oral cavity cancer cases (n=63) Mean/% 61.7 17.5 22.2 60.3 21.2 39.5 31.7 41.3 27.0 33.3 189.3 79.4 20.6 38.5 53.8 7.7 0.0 0.0 0.0 0.0.0

27.7 77.8

11.5 9.0

SD 4.2

n 67 3 22 42 55 61 30 17 20 65 67 47 20 5 9 6 1 0 0 1

Pharyngeal cancer cases (n=67) Mean/% SD 62.1 4.2 4.5 32.8 62.7 22.6 13.4 39.8 9.8 44.8 25.4 29.8 38.4 32.7 181.5 86.4 70.2 29.8 25.0 45.0 30.0 1.5 0.0 0.0 100.0

Distribution of potential confounders and asbestos exposure* among male subcohort members and cancer cases in the NLCS, 1986‐2003.

Table 5.1

Chapter 5

n 1559 542 181 180 181 290 100 95 95

Subcohort (n=2101) Mean/% 74.2 25.8 33.4 33.2 33.4 13.8 34.4 32.8 32.8

n 49 14 5 5 4 7 1 5 1 †

Oral cavity cancer cases (n=63) Mean/% SD 77.8 22.2 35.7 35.7 28.6 11.1 14.3 71.4 14.3

n 47 20 9 8 3 6 2 3 1

‡

Pharyngeal cancer cases (n=67) Mean/% SD 70.2 29.8 45.0 40.0 15.0 9.0 33.3 50.0 16.7

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. Among former and current smokers only. Sum of categories § ¶ deviates from total number because of missing values. Exposure based on the cumulative probabilityintensity of exposure (unit‐years or f‐y/ml). Exposure based on the probabilityintensity of exposure (unit‐years or f‐y/ml) per job. f‐y/ml, fiber‐years/ml; NLCS, Netherlands Cohort Study.

FINJEM Never exposed§ (%) § Ever exposed (%) Cumulative probabilityintensity of exposure (f‐y/ml) T1 (median:0.20) (%) T2 (median:1.59) (%) T3 (median:6.60) (%) ¶ Ever highly exposed (%) ¶ Duration of high exposure (yrs) T1 (median:6) (%) T2 (median:20) (%) T3 (median:35) (%)

(continued)

Table 5.1

Asbestos exposure and tumors of the oral cavity and pharynx

100

DOMJEM Never exposed§ Ever exposed§ Cumulative probabilityintensity of exposure (unit‐years) T1 (median:4) T2 (median:20) T3 (median:38) P for trend Continuous, per 1 unit‐year FINJEM Never exposed§ Ever exposed§ Cumulative probabilityintensity of exposure (f‐y/ml) T1 (median:0.20) T2 (median:1.59) T3 (median:6.60) P for trend Continuous, per 1 fiber‐year

19246 7421 2891 1997 2533 26667 20204 6463 2215 2187 2061 26667

46 12 5 6 1 58 47 11 5 4 2 58

Person‐years No. of in subcohort cases 1 (ref) 0.70 (0.37‐1.34) 0.74 (0.29‐1.89) 1.31 (0.55‐3.16) 0.17 (0.02‐1.26) 0.133 0.98 (0.95‐1.00) 1 (ref) 0.76 (0.39‐1.50) 1.00 (0.39‐2.57) 0.85 (0.30‐2.38) 0.43 (0.10‐1.79) 0.257 0.87 (0.73‐1.03)

1 (ref) 0.86 (0.41‐1.82) 0.72 (0.22‐2.32) 2.17 (0.83‐5.70) 0.24 (0.03‐1.78) 0.415 0.98 (0.96‐1.01) 1 (ref) 1.18 (0.53‐2.61) 1.71 (0.61‐4.82) 1.51 (0.49‐4.65) 0.54 (0.12‐2.47) 0.753 0.90 (0.79‐1.03)

Oral cavity cancer HR† (95%CI) HR‡ (95%CI) 37 16 4 8 4 53 36 17 9 7 1 53

No. of cases

†

1 (ref) 1.15 (0.63‐2.10) 0.73 (0.26‐2.08) 2.17 (0.98‐4.81) 0.85 (0.30‐2.41) 0.578 1.01 (0.99‐1.02) 1 (ref) 1.54 (0.86‐2.75) 2.35 (1.11‐4.96) 1.92 (0.84‐4.39) 0.28 (0.04‐2.05) 0.816 0.84 (0.69‐1.02)

1 (ref) 1.16 (0.56‐2.38) 0.44 (0.10‐1.86) 3.18 (1.24‐8.15) 1.09 (0.37‐3.20) 0.289 1.01 (0.99‐1.03) 1 (ref) 2.20 (1.08‐4.49) 4.04 (1.69‐9.64) 3.01 (1.14‐7.96) 0.30 (0.04‐2.46) 0.360 0.84 (0.71‐0.98)

Pharyngeal cancer HR† (95%CI) HR‡ (95%CI)

‡

Oral cavity and Pharyngeal cancer combined No. of HR† (95%CI) HR‡ (95%CI) cases 85 1 (ref) 1 (ref) 28 0.88 (0.57‐1.37) 0.99 (0.58‐1.69) 9 0.72 (0.35‐1.46) 0.57 (0.22‐1.51) 14 1.66 (0.91‐3.02) 2.56 (1.28‐5.09) 5 0.46 (0.19‐1.16) 0.62 (0.24‐1.58) 0.450 0.918 113 1.00 (0.98‐1.01) 1.00 (0.99‐1.02) 85 1 (ref) 1 (ref) 28 1.07 (0.69‐1.67) 1.59 (0.93‐2.73) 14 1.55 (0.86‐2.80) 2.64 (1.34‐5.19) 11 1.28 (0.67‐2.46) 2.15 (1.02‐4.52) 3 0.36 (0.11‐1.14) 0.42 (0.12‐1.50) 0.412 0.740 113 0.85 (0.75‐0.97) 0.87 (0.78‐0.97)

Exposure dichotomized or categorized in never‐exposed and tertiles (T) of exposed in the subcohort. Age‐adjusted model. Adjusted for age (yrs), smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (three categories), alcohol consumption (g/d), and consumption of vegetables (g/d). § Exposure based on the cumulative probabilityintensity of exposure (unit‐years or f‐ ¶ y/ml). Exposure based on the probabilityintensity of exposure (unit‐years or f‐y/ml) per job. CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Hazard ratios (HRs) and 95% CIs for oral cavity and pharyngeal cancer both separately and combined, for categories of asbestos exposure estimated with DOMJEM and FINJEM in the NLCS, 1986‐2003.

Table 5.2

Chapter 5

Never No. of cases 7 4 1 1 8 5

Person‐ years 3124 758 3124 758 3124 758

1 (ref)* 4.60 (1.40‐15.04) 10903 3402

Person‐ years * 1 (ref) 10903 4.02 (1.09‐14.85) 3402 1 (ref)* 10903 7.81 (0.49‐123.45) 3402

HR (95%CI)

24 6

0.81 (0.35‐1.88) 0.79 (0.26‐2.43)

Cigarette smoking Former No. of HR (95%CI) cases 11 0.42 (0.15‐1.13) 3 0.46 (0.11‐1.96) 13 3.45 (0.44‐27.0) 3 3.06 (0.30‐30.90) 6177 2303

Person‐ years 6177 2303 6177 2303 53 17

1.88 (0.78‐4.54) 3.05 (1.18‐7.92)

0.13

P for multiplicative Current interaction No. of HR (95%CI) cases 29 1.43 (0.51‐3.97) 4 0.96 (0.25‐3.63) 0.13 22 5.14 (0.60‐44.13) 13 16.37 (1.92‐139.38) 0.18

* Adjusted for age (yrs), smoking status (never/former/current), number of cigarettes smoked per day (centered variable), years of smoking cigarettes (centered variable), level of education (three categories), alcohol consumption (g/d), and consumption of vegetables (g/d). CI, confidence interval; f‐y/ml, fiber‐years/ml; HR, hazard ratio; NLCS, Netherlands Cohort Study; ref, reference.

Oral cavity cancer Never exposed Ever exposed Pharyngeal cancer Never exposed Ever exposed Oral cavity and Pharyngeal cancer combined Never exposed Ever exposed

Asbestos Exposure

Hazard ratios (HRs) and 95% CIs for oral cavity and pharyngeal cancer both separately and combined, for asbestos exposure (yes/no) by smoking; estimated with FINJEM in the NLCS, 1986‐2003.

Table 5.3

Asbestos exposure and tumors of the oral cavity and pharynx

101

Chapter 5

DISCUSSION This study showed no convincing evidence of an association between asbestos and risk of OCC, PhC, and OCPC, as an exposure–response relation was lacking and the ever versus never exposed estimates were inconsistent across both JEM. Nevertheless, some HR of PhC and OCPC were increased. Adjustment for especially alcohol consumption and socioeconomic status further increased HR of PhC and OCPC. Although the strata‐ specific HR may be suggestive of a negative interaction between asbestos and smoking for OCC, none of the cancers showed a significant interaction.

Oral cavity cancer There are only few studies that have investigated the asbestos‐related risk of OCC. These studies showed non‐significantly increased risks20,21 as well as decreased risks22,23, with a recent meta‐analysis revealing relative risk (RR) estimates of HR 1.13 (95% CI 0.81–1.57) and HR 1.15 (95% CI 0.84–1.57) for low and high exposure, respectively, based on five studies.3 Risk estimates in our study are in line with the results of the meta‐analysis.

Pharyngeal cancer For PhC, risk estimates of around one21,22 as well as increased risks have been reported after asbestos exposure in both overall PhC23 and hypoPhC20,24. Our study reported multivariable‐adjusted increased HR of PhC for ever versus never exposed to asbestos (HR 2.20, 95% CI 1.08–4.49) when using FINJEM, but a trend of increased risks with higher cumulative exposure could not be demonstrated. Furthermore, given that results were not robust against the use of different JEM, one could conclude that our study showed no convincing evidence of an association between asbestos and PhC. However, the number of cases in our study was small. A meta‐analysis also showed increased risks of PhC without evidence of an exposure–response relation (relative risk (RR) of 1.26 (95% CI 0.96–1.66) and 1.27 (95% CI 0.98–1.66) for low and high exposure, respectively),3 whereas a recent large case–control study observed an exposure– response relation.23 Therefore, the rather consistent observation of a possible association between asbestos and PhC could be more than a mere chance finding and warrants further research in studies with a larger number of cases.

Oral cavity and pharyngeal cancer combined Because other studies mostly combined OCC and PhC, we also studied both cancers together (OCPC), notwithstanding the possibility that the asbestos‐related risk may differ for both cancers. We found increased risks for ever versus never exposed and the CE, similar to but less strong than for PhC. Results of previous cohort studies are rather

102

Asbestos exposure and tumors of the oral cavity and pharynx

consistent and show modestly increased risks with a meta‐RR of 1.44 (95% CI 1.04– 2.00), while case–control studies are rather limited in number and show inconsistent results.5 Data on exposure–response patterns from both types of study are limited and tend towards lower risks for the more extreme exposures,5 as was the case in the present study. Another study using FINJEM found no exposure–response association either,4 but a meta‐analysis stratifying results on exposure circumstance showed a RR of 1.63 (95% CI 1.27–2.09) for asbestos miners and millers with the highest exposures.3 Due to the small number of cases, we were not able to run the analyses for the (prolonged) highly exposed subjects and test this result. Moreover, exposure levels of the miners and millers were probably much higher than in the NLCS. Contrary to lung cancer, there is almost no epidemiological or experimental evidence addressing whether interaction may be present between asbestos and tobacco smoking in the development of OCC, PhC, and OCPC.5 Previous studies found no significant interaction for overall PhC23 and hypoPhC.24 Our study also found no significant interaction for any of the cancers. The number of cases was small, however, and non‐ differential asbestos exposure misclassification, due to using JEM, may have hampered finding significant results. The present study suffers from some of the same limitations as previous studies, ie, a small number of cases and suboptimal characterization of asbestos exposure due to using JEM and the fact that information on occupational history was gathered at baseline in 1986 while study subjects were followed‐up to December 2003. Nevertheless, our results are of importance for future meta‐analyses given the limited number of studies on this subject. Moreover, the strengths of our study include the prospective design, the long, nearly complete follow‐up, and the possibility to adjust for several lifestyle confounders. Evidence comes mainly from occupational cohorts that do often not allow for adjustment for potential confounders. Since HR of PhC and OCPC increased after adjustment for especially alcohol consumption and socioeconomic status, taking lifestyle factors into account may be important when studying these cancers. Furthermore, while our JEM‐based exposure assessment possibly entailed non‐differential exposure misclassification resulting most likely in bias towards the null value, a previous study in the NLCS using both DOMJEM and FINJEM was able to demonstrate the well‐known associations between asbestos and cancers of the pleura and lungs.16 In addition, as study subjects were between 55–69 years of age at the start of the study in 1986, the amount of exposure misclassification resulting from the fact that we had no information on occupational history from 1986–2003 will probably be limited.

103

Chapter 5

CONCLUSION In conclusion, this study showed no convincing evidence of an association between asbestos and OCC, PhC, and OCPC risk as an exposure–response relation was lacking and results were not robust against the use of different JEM. However, increased HR of PhC and OCPC were observed in this study as well as in previous studies and warrant further research.

104

Asbestos exposure and tumors of the oral cavity and pharynx

REFERENCES 1. 2. 3.

4. 5. 6. 7. 8. 9.

10.

11. 12.

13.

14.

15. 16.

17. 18. 19.

Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer. 2010;127:2893–917. Schottenfeld D, Fraumeni J. Cancer Epidemiology and Prevention. New York: Oxford University Press, 2006. 674–96. Paget‐Bailly S, Cyr D, Luce D. Occupational exposures to asbestos, polycyclic aromatic hydrocarbons and solvents, and cancers of the oral cavity and pharynx: a quantitative literature review. Int Arch Occup Environ Health. 2012;85:341–51. Tarvainen L, Kyyrönen P, Kauppinen T, Pukkala E. Cancer of the mouth and pharynx, occupation and exposure to chemical agents in Finland [in 1971‐95]. Int J Cancer. 2008;123:653–9. NAS. Asbestos: Selected Cancers. Washington, DC: The National Academies Press, 2006. 159‐72. IARC. Asbestos (Chrysotile, Amosite, Crocidolite, Tremolite, Actinolite and Anthophyllite). In: A Review of Human Carcinogens: Arsenic, Metals, Fibres, and Dusts. Lyon, France: International Agency for Research on Cancer (IARC), 2012. 219‐309. van den Brandt PA, Goldbohm RA, van 't Veer P, Volovics A, Hermus RJ, Sturmans F. A large‐scale prospective cohort study on diet and cancer in The Netherlands. J Clin Epidemiol. 1990;43:285–95. Prentice RL. A case‐cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika. 1986;73:1–11. Casparie M, Tiebosch AT, Burger G, Blauwgeers H, van de Pol A, van Krieken JH, Meijer GA. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol. 2007;29:19–24. Van den Brandt PA, Schouten LJ, Goldbohm RA, Dorant E, Hunen PM. Development of a record linkage protocol for use in the Dutch Cancer Registry for Epidemiological Research. Int J Epidemiol. 1990;19:553–8. Goldbohm RA, van den Brandt PA, Dorant E. Estimation of the coverage of Dutch municipalities by cancer registries and PALGA based on hospital discharge data. Tijdschr Soc Gezonheidsz. 1994;72:80–4. Hashibe M, Brennan P, Benhamou S, Castellsague X, Chen C, Curado MP, Dal Maso L, Daudt AW, Fabianova E, Fernandez L, Wünsch‐Filho V, Franceschi S, Hayes RB, Herrero R, Koifman S, La Vecchia C, Lazarus P, Levi F, Mates D, Matos E, Menezes A, Muscat J, Eluf‐Neto J, Olshan AF, Rudnai P, Schwartz SM, Smith E, Sturgis EM, Szeszenia‐Dabrowska N, Talamini R, Wei Q, Winn DM, Zaridze D, Zatonski W, Zhang ZF, Berthiller J, Boffetta P. Alcohol drinking in never users of tobacco, cigarette smoking in never drinkers, and the risk of head and neck cancer: pooled analysis in the International Head and Neck Cancer Epidemiology Consortium. J Natl Cancer Inst. 2007;99:777–89. Offermans NS, Vermeulen R, Burdorf A, Peters S, Goldbohm RA, Koeman T, van Tongeren M, Kauppinen T, Kant I, Kromhout H, van den Brandt PA. Comparison of expert and job‐exposure matrix‐based retrospective exposure assessment of occupational carcinogens in the Netherlands Cohort Study. Occup Environ Med. 2012;69:745–51. Peters S, Vermeulen R, Cassidy A, Mannetje A', van Tongeren M, Boffetta P, Straif K, Kromhout H; INCO Group. Comparison of exposure assessment methods for occupational carcinogens in a multi‐centre lung cancer case‐control study. Occup Environ Med. 2011;68:148–53. Kauppinen T, Toikkanen J, Pukkala E. From cross‐tabulations to multipurpose exposure information systems: a new job‐exposure matrix. Am J Ind Med. 1998;33:409–17. Offermans NS, Vermeulen R, Burdorf A, Goldbohm RA, Kauppinen T, Kromhout H, van den Brandt PA. Occupational asbestos exposure and risk of pleural mesothelioma, lung cancer, and laryngeal cancer in the prospective Netherlands Cohort Study. J Occup Environ Med 2014;56:6–19. Volovics A, van den Brandt PA. Methods for the analysis of case‐cohort studies. Biomed J. 1997;39:195– 214. Barlow WE, Ichikawa L, Rosner D, Izumi S. Analysis of case‐cohort designs. J Clin Epidemiol. 1999;52:1165–72. Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69:239–41.

105

Chapter 5

20. Reid A, Ambrosini G, de Klerk N, Fritschi L, Musk B. Aerodigestive and gastrointestinal tract cancers and exposure to crocidolite (blue asbestos): incidence and mortality among former crocidolite workers. Int J Cancer. 2004;111:757–61. 21. Purdue MP, Järvholm B, Bergdahl IA, Hayes RB, Baris D. Occupational exposures and head and neck cancers among Swedish construction workers. Scand J Work Environ Health. 2006;32:270–5. 22. Gustavsson P, Jakobsson R, Johansson H, Lewin F, Norell S, Rutkvist LE. Occupational exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus: a case‐control study in Sweden. Occup Environ Med. 1998;55:393–400. 23. Langevin SM, O'Sullivan MH, Valerio JL, Pawlita M, Applebaum KM, Eliot M, McClean MD, Kelsey KT. Occupational asbestos exposure is associated with pharyngeal squamous cell carcinoma in men from the greater Boston area. Occup Environ Med. 2013;70:858–63. 24. Marchand JL, Luce D, Leclerc A, Goldberg P, Orlowski E, Bugel I, Brugère J. Laryngeal and hypopharyngeal cancer and occupational exposure to asbestos and man‐made vitreous fibers: results of a case‐control study. Am J Ind Med. 2000;37:581‐9.

106

General discussion

Chapter 6

General discussion

107

General discussion

This chapter discusses the most important findings on the association between occupational asbestos exposure and the risk of cancer in the population‐based Netherlands Cohort Study (NLCS). First, we aimed to evaluate three candidate job‐ exposure matrices (JEMs; Asbestos JEM, DOMJEM and FINJEM) in terms of reliability by comparing the exposure estimates of these JEMs with case‐by‐case expert assessment. Second, we studied the performance of two of the JEMs (DOMJEM, FINJEM) by linking their exposure estimates to two established cancer endpoints (pleural mesothelioma, and lung cancer). Thirdly, we related the exposure estimates to additional cancers of interest being respiratory (laryngeal cancer), gastrointestinal (esophageal, gastric and colorectal cancer), and oral cavity (OCC) and pharyngeal cancer (PhC), both separately and combined (OCPC). 1. We studied the overall association between occupational asbestos exposure and the risk of these cancers, with special attention to: i. risk associated with the lower end of the exposure distribution ii. risk differences between relatively low and high exposure iii. the existence of an exposure‐response relation iv. potential confounding 2. We studied the association between occupational asbestos exposure and subtypes of the cancers of interest 3. We studied the possible additive or multiplicative interaction between smoking and asbestos in relation to the cancers of interest Furthermore, we will discuss some important strengths and limitations of the population‐based NLCS for studying the association between occupational asbestos exposure and cancer, as compared to industry‐based studies. In addition, we will address the implications of our findings for policymakers and the society as a whole in terms of the burden of disease, both in the Netherlands and in low‐ to middle‐income countries. Finally, we will give some recommendations for future research.

1. MAIN FINDINGS We scored three JEMs on reliability in the NLCS by assessing agreement between these JEMs and case‐by‐case expert assessment (available for the subcohort) by means of Cohen’s Kappa and the prevalence of exposure. Case‐by‐case expert assessment was available for exposure to asbestos, PAHs, and welding fumes. The expert assessment resulted in the lowest exposure prevalence for all three substances in the NLCS. DOMJEM1 and FINJEM2 showed rather similar agreement with the case‐by‐case expert assessment and moderate agreement amongst each other, while the Asbestos JEM3 appeared to be less appropriate for use in the NLCS as it showed a lower agreement

109

Chapter 6

with the expert assessment and the other two JEMs. This is not completely unexpected as the Asbestos JEM was originally designed to aid subjects with asbestos‐related diseases in pursuing compensation in legal trials. It only includes those 123 job codes with definite historical exposure to asbestos, among which administrative employees (i.e., white‐collar workers) in companies manufacturing asbestos products. Consequently, this partly disease‐oriented JEM might be more sensitive than DOMJEM and FINJEM, which are more aimed at specificity than sensitivity because of the low exposure levels in the general population. For this reason, the Asbestos JEM will not be further discussed here.

In order to provide insight into the methodological uncertainty associated with the choice of JEM, we studied the asbestos‐cancer associations using both DOMJEM and FINJEM in the NLCS. Below, a short recapitulation of the most important results for the three cancer‐related aims is given, organised per cancer group (i.e., respiratory, gastrointestinal, and oral cavity and pharyngeal) (see also Table 6.1).

1.1 Respiratory cancers Aim 1: overall association For both pleural mesothelioma and lung cancer, HRs were significantly elevated in our study when using both DOMJEM and FINJEM. However, only FINJEM was able to detect a significantly increased risk for the lowest tertile of cumulative exposure (CE) (median: 0.20 fiber‐years/ml (f‐y/ml) (HR(95%CI)=2.69(1.60‐4.53) and 1.44(1.12‐1.86), respectively). Adjusting for potential confounders had no influence on the association with mesothelioma, and little to no effect on the association with lung cancer. For mesothelioma, results from previous studies were inconsistent, with some studies observing no threshold level for asbestos‐related mesothelioma,4,5 while others found evidence of a threshold level.6,7 Unfortunately, we cannot ‐based on our data‐ subscribe to the (non‐) existence of a threshold level for asbestos, because i) there is no uniform definition what low level exposure entails and ii) exposure assessment in our study is JEM‐based possibly entailing non‐differential exposure misclassification. For lung cancer, the FINJEM risk estimate of 1.44(1.12‐1.86) for the lowest tertile of exposure (median: 0.20 f‐y/ml) was comparable to or higher than the relative risks presented in the meta‐analysis by van der Bij and colleagues for a higher exposure level of 4 f‐y/ml, depending on the model they used.8 Therefore, results of our study are in line with this study and other studies that showed that relative risks at the lower end of the exposure distribution might be higher than predicted by downward linear extrapolation from highly exposed occupational cohorts.8‐10

110

0.86 1.16 0.99

2.62 1.19 1.19 1.29 1.29 0.96 1.20 0.94 2.14 1.35 1.08 1.54 1.02 1.01 1.02 0.95 0.89 0.94 0.87 1.05 0.72 0.44 0.57

∙ ∙ ∙

1.18 2.20 1.59

1.71 4.04 2.64

∙ ∙ ∙

A + +

A A A

0.13 0.18 0.13

DOMJEM FINJEM Both JEMs Cumulative† Ever‡ Duration‡ of Ever† Cumulative† Ever‡ Duration‡ of Confounding Exposure‐ P¶ for exposure highly high exposure exposed exposure highly high exposure response§ interaction T1 (median: exposed T3 T1 (median: exposed T3 4 unit‐years) 0.20 f‐ y/ml) 0.71 ∙ 13.66 3.02 2.69 ∙ 3.28 A + 0.41 0.91 ∙ 2.99 1.50 1.44 ∙ 1.74 A B 0.50 0.94 ∙ 6.38 1.56 0.81 ∙ 2.52 A +/‐ ∙ 0.77 ∙ 1.51 1.62 1.41 ∙ 2.00 A +/‐ ∙ 0.95 ∙ 2.87 1.68 1.94 ∙ 1.54 A A ∙ 0.84 ∙ 2.36 1.09 1.08 ∙ 1.37 A A ∙ 0.76 ∙ 6.36 1.42 1.59 ∙ 1.49 A B 0.85 0.61 ∙ 7.09 1.09 1.41 ∙ 1.23 A A ∙ 1.23 ∙ 4.24 2.66 2.31 ∙ 2.39 A A ∙ 1.31 2.22 2.50 1.19 1.45 1.11 1.03 A A 0.09 1.11 1.64 2.64 1.18 0.71 1.70 1.50 A A ∙ 1.50 2.52 2.27 1.21 1.80 0.84 0.78 A A ∙ 0.69 1.72 2.67 1.11 1.19 1.17 1.28 − A 0.61 0.53 1.63 1.05 1.07 1.32 1.06 1.10 A A ∙ 0.76 1.79 3.35 1.12 1.13 1.21 1.33 − A ∙ 0.78 1.53 1.87 1.05 1.03 1.05 1.06 A A 0.29 0.74 1.25 2.19 0.99 1.03 0.95 0.83 A A ∙ 0.68 1.28 1.97 1.13 1.10 1.10 0.74 A A ∙ 0.87 1.26 2.54 0.92 1.05 0.85 0.84 A A ∙ 0.78 2.15 1.31 1.26 1.14 1.27 1.23 A A ∙

Multivariable‐adjusted hazard ratios, with the particular confounders depending on the cancer endpoint; results highlighted in bold are statistically significant (P<0.05), results printed in italic are borderline significant. † Reference group are the never exposed. ‡ Reference group are the never highly exposed. § Based on the P for trend over the exposed for the cumulative exposure measure. ¶ Calculated for both JEMs but, according to the choices made in the respective chapters, results presented here are based on FINJEM for the respiratory cancers and oral cavity and pharyngeal cancer, and based on DOMJEM for the gastrointestinal cancers. A = absent, + = statistically significant, positive trend for both JEMs, B = borderline significant trend for one JEM, +/‐ = statistically significant, positive trend for one JEM, − = lower hazard ratios (HRs) after adjustment for confounders, + = higher HRs after adjustment for confounders. f‐y/ml, fiber‐years/ml; T, tertile.

Respiratory cancers (Chapter 3) Pleural mesothelioma Lung cancer Small cell Large cell Squamous cell carcinoma Adenocarcinoma Laryngeal cancer Glottis Supraglottis Gastrointestinal cancers (Chapter 4) Esophageal cancer Squamous cell carcinoma Adenocarcinoma Gastric cancer Cardia adenocarcinoma Non‐cardia adenocarcinoma Colorectal cancer Colon Proximal colon Distal colon Rectum Oral cavity and pharyngeal cancer (Chapter 5) Oral cavity Pharynx Oral cavity and pharynx

Ever† exposed

Summary of the main findings* of this thesis, with respect to the cancer‐related aims.

Table 6.1

General discussion

111

Chapter 6

For laryngeal cancer, the FINJEM result of 1.42(0.99‐2.03) for ever versus never exposed to asbestos was comparable to the summary relative risk (RR) of 1.43 in a meta‐analysis of case‐control studies,11 while for DOMJEM the HR for ever versus never exposed was slightly lower (1.20(0.84‐1.72)). Adjusting for alcohol consumption and smoking did not alter these results, which is in line with some,11 though not all previous studies.12 Aim 2: association with subtypes Contrary to the other lung cancer subtypes, adenocarcinoma showed a weak association only after prolonged higher asbestos exposure. The weaker association between asbestos and adenocarcinoma is of interest as previous studies showed asbestos to be associated preferentially with adenocarcinoma,13,14 though this association has not been reported consistently throughout the literature.15,16 As asbestos and smoking may have similar exposure routes and related pathophysiological pathways,17 and because smoking is only weakly associated with adenocarcinoma,15,16 this might explain why only after prolonged higher asbestos exposure a positive association with adenocarcinoma was observed. Another reason might be that adenocarcinoma is more frequently observed among non‐smokers15 and since most subjects exposed to asbestos were smokers, the true association between asbestos and adenocarcinoma may have been stronger. When stratifying by smoking, the association with adenocarcinoma for never‐smokers ever exposed to asbestos compared to never‐ smokers never exposed to asbestos was indeed higher (2.43(0.83‐7.11)), though numbers were low (FINJEM results, not shown). Relative risk by sublocalisation of laryngeal cancer in the NLCS is comparable to some, though not all studies.11 We showed overall stronger associations for supraglottis than glottis cancer. As the location of the supraglottis seems more readily exposed to tobacco,18 and because exposure routes and pathophysiological pathways of tobacco and asbestos are believed to be related,17 the higher HR for cancer of the supraglottis might have been expected. Unfortunately, we could not check if and to what degree this association between asbestos and cancer of the supraglottis was driven by smoking as all asbestos‐exposed men were smokers. In addition, numbers were low, certainly for supraglottis cancer. Aim 3: possible interaction As expected, our study showed that the risk of mesothelioma was not influenced by smoking as opposed to the risk of lung and laryngeal cancer.19 A synergistic effect between asbestos and smoking in relation to lung cancer is supported by a number of systematic reviews,20‐22 though the degree of synergism remains uncertain as results range from additivity to supramultiplicativity.9,23 Our study

112

General discussion

found no statistically significant interaction on an additive or multiplicative scale for lung cancer, and joint effects were between additivity and multiplicativity. For laryngeal cancer, there was also no statistically significant interaction on an additive or multiplicative scale. Joint effects were closer to additivity than multiplicativity which is in line with previous studies, though one study found a supramultiplicative joint effect of asbestos and smoking.11 Results have to be interpreted carefully, though, as numbers were low most notably among the reference category of the never smokers not exposed to asbestos.

1.2 Gastrointestinal cancers Aim 1: overall association When using DOMJEM, we found increased HRs for all three cancers mainly for the (prolonged) highly exposed subjects; borderline significant for esophageal and colorectal cancer. This is in line with previous studies that observed increased risks and exposure‐response relations mainly in occupational cohorts with higher asbestos exposure.11,17,24‐26 Therefore, we hypothesize that for gastrointestinal cancers mainly (prolonged) exposure to higher asbestos levels may be associated, though FINJEM was not able to replicate the findings for the (prolonged) highly exposed as HRs were remarkably lower and non‐significant. FINJEM results for the other asbestos exposure variables were essentially similar to those for DOMJEM showing no significantly increased risk of gastrointestinal cancers. Several associations with overall gastric cancer and GNCA were reduced and became non‐significant after adjusting for smoking status. Therefore, taking into account lifestyle factors may be important when studying these cancers. However, HRs were in most instances only marginally significant to begin with. Aim 2: association with subtypes Our study showed significantly increased HRs for EAC, not only for those highly exposed to asbestos but also for those ever exposed to asbestos and for the duration of asbestos exposure. As esophageal cancer is a relatively rare cancer, only few studies looked at subtypes. For both histological types, results are mixed, with a suggestion of an exposure‐response relation only for EAC.27‐30 As such, the analyses reported here provide some further support for an association with EAC. We observed an increased risk of GNCA but not of GCA, which is comparable to previous studies that showed no association with GCA risk.29,31 As diverging trends and marked geographic variation suggest GCA and GNCA to be separate disease entities

113

Chapter 6

with different etiologies,32 these diverging results for asbestos might therefore be a reflection of this difference in etiology. An industry‐based study, however, did not observe an association with GNCA after either moderate or high asbestos exposure.33 For colorectal cancer subtypes, there is some suggestion in the literature that the association with asbestos might be stronger for colon than for rectal cancer,17 which is also based on studies that found a higher RR of right (or proximal) colon than left (or distal) colon cancer,34,35 or found an association with colorectal cancer but not with rectal cancer.24 One study found, however, a higher SMR of rectal cancer versus intestine and rectal cancer together.36 Our study found significantly increased HRs for colon (both total and distal) and rectal cancer. The association with rectal cancer and, given the location, with distal colon cancer could be explained by asbestos’ relation to smoking carcinogenesis and the fact that associations with smoking have been appreciably stronger for rectal cancer than (proximal) colon cancer.37 Aim 3: possible interaction Contrary to lung cancer, there is almost no epidemiological or experimental evidence addressing whether asbestos is a cofactor of tobacco smoking in the development of gastrointestinal cancers.11 A study by Liddell et al. showed a modest interaction between cumulative asbestos exposure and smoking in relation to gastric cancer,38 and Aliyu et al. reported a RR of colorectal cancer of 1.36(0.96‐1.93) for asbestos and smoking together as compared to smoking alone.39 Our study found no statistically significant interaction on an additive or multiplicative scale for any of the cancers. Sample sizes were small, however, especially for esophageal cancer.

1.3 Oral cavity and pharyngeal cancer Aim 1: overall association For OCC, no associations were observed when using DOMJEM or FINJEM, which is in line with previous studies on the asbestos‐related risk of OCC.27,40‐43 This suggests that asbestos may not be associated with an increased risk of OCC, though the number of cases was rather low. For PhC, our study reported multivariable‐adjusted increased HRs for ever versus never exposed to asbestos (HR 2.20, 95% CI:1.08‐4.49) when using FINJEM, but a clear trend of increased risks with higher cumulative exposure could not be demonstrated. Furthermore, given that results were not robust against the use of different JEMs, one could conclude that our study showed no convincing evidence of an association between asbestos and PhC. However, the number of cases in our study was small. Furthermore, previous studies also showed increased risks without evidence of an

114

General discussion

exposure‐response pattern,40,41,44 while a recent large case‐control study observed an exposure‐response relation.43 Therefore, the rather consistent observation of a possible association between asbestos and PhC could be more than a mere chance finding and warrants further research in studies with a larger number of cases. For OCPC, risk estimates were similar to but less strong than for PhC and higher than the reported meta‐RR of 1.44(1.04‐2.00) for ever versus never exposed,11 at least when using FINJEM. Although data on exposure‐response patterns in the literature are limited, they tend towards lower risks for the more extreme exposures,11 as was the case in our study. Another study using FINJEM found no exposure‐response pattern either,45 but a meta‐analysis stratifying results on exposure circumstance showed a RR of 1.63(1.27‐2.09) for asbestos miners and millers with the highest exposures.40 HRs of PhC and OCPC increased after adjustment for especially alcohol consumption and socioeconomic status. Therefore, taking into account lifestyle factors may be important when studying these cancers. Aim 3: possible interaction Also in the development of oral cavity and pharyngeal cancer, there is almost no epidemiological or experimental evidence addressing whether interaction may be present between asbestos and tobacco smoking.11 Previous studies found no significant interaction for overall pharyngeal cancer43 and hypopharyngeal cancer.44 Our study found no significant interaction either, for none of the cancers, though the number of cases was small.

2. METHODOLOGICAL CONSIDERATIONS Most studies on occupational asbestos exposure and possible cancer risks have been performed in an industry‐based setting. Industry‐based studies are a powerful tool for studying exposure‐response relationships. Their primary advantage is the ability for a more precise exposure assessment based on measurements in the particular facilities and the ability to address the morphology of asbestos fibers. Potential limitations of many of the industry‐based studies include i) the inclusion of mostly heavily exposed subjects which hinders the extrapolation to low asbestos exposure levels which are nowadays present in most industrialized countries, ii) the absence or sparseness of data on potential confounders, iii) the relatively small number of cases in most industrial cohorts, particularly if the risk of cancers other than lung cancer was studied, and iv) the limited information on disease status as most cohort studies had to rely on mortality data due to the absence of a national cancer registry, such as in the US where many industrial cohort studies have been performed. This may be problematic when cancers with a fairly good prognosis are being studied.

115

Chapter 6

The NLCS is a population‐based study which can overcome some of the limitations mentioned above due to i) the presence of a wide range in exposure levels including those at the lower end of the exposure distribution, ii) the possibility to control for potential confounders such as alcohol and smoking, iii) the large size which enables studying relatively rare cancers, cancer subtypes, and a possible interaction between asbestos exposure and smoking in relation to cancer risk, and iv) the availability of cancer incidence data. However, as in most population‐based studies, we had to rely on more tenuous methods for assessing occupational exposure, because no workplace specific exposure measurements were available. In the following section, these differences will be discussed, whenever applicable.

2.1 Selection bias Selection bias due to selection of the study population is probably not of major importance in cohort studies as the population is disease‐free at the start of the study. Selection bias due to study subjects being lost to follow‐up could arise if loss to follow‐ up is correlated with both exposure and disease.46 However, in the NLCS, completeness of incident cancer coverage was estimated to be almost 100%,47 and completeness of follow‐up of the subcohort was almost 100% (only two male subcohort members (and later on only one) were lost to follow‐up). Consequently, it is unlikely that selection bias due to follow‐up loss has distorted our findings.

2.2 Information bias 2.2.1 Exposure misclassification Differential misclassification of asbestos exposure, i.e. dependent on disease status, is unlikely in the NLCS because the population under study was still disease‐free at the time of reporting their occupational history. Non‐differential (or random) misclassification of asbestos exposure may have occurred in the NLCS, through a number of sources: 1. We had to rely on a self‐administered questionnaire on occupational history. Subjects could enter a maximum of five occupations held between starting work and baseline at 1986. Although this results in a large time‐span (1929‐1986), subjects will probably have been able to recall in which jobs they were engaged as cohort subjects held on average 1.9 job codes during their working life up to 1986. Furthermore, it has been shown that even after a considerable number of years, recall of job history appears to be rather reliable.48 In addition, not all subjects were retired at baseline. For these subjects, data on occupational asbestos exposure are missing for a part of their occupational history, which may have resulted in an underestimation of their true occupational exposure to asbestos. As these subjects are the younger part of the study population who have a lower risk

116

General discussion

of developing cancer, this may have led to biased results. However, as this only pertains to 2.5% of the subcohort, the resulting bias is probably negligible; also, because occupational asbestos exposure levels decreased again during the 1980s due to increasingly stringent legal standards and reductions in the threshold limit values. 2. Coding of the occupational history ‐ which occurred blinded ‐ may also be problematic, especially when coding on occupational level is required instead of on industry level.48,49 As the occupational coding used in the NLCS is rather detailed (i.e., occupations were coded according to the Standard Occupational Classification of 1984 of the Dutch Central Bureau of Statistics (CBS‐84, four‐digit), supplemented by a three‐digit code assigned within the NLCS (CBS‐NLCS) based on the job title), some amount of misclassification may be present in the original coding. It has been shown, however, that different occupational codes often lead to comparable exposure estimates when the occupational codes are related.49 3. As actual asbestos exposure measurements are not available in the NLCS, we had to rely on JEMs for the exposure assessment. In order to assign exposure, a JEM has to be linked to a job title (or code). As the occupational classifications used in DOMJEM and FINJEM deviate from the one used in the NLCS, the CBS‐NLCS codes had to be recoded to the job classification used in the job axis of the JEMs (ISCO‐68 for DOMJEM, and Finnish codes for FINJEM) in order to be able to link the JEMs. For this purpose, an algorithm recoding from CBS‐NLCS to ISCO‐68 has been developed for the NLCS. 49 An algorithm recoding ISCO‐68 to the Finnish job codes had already been developed within the framework of the INTEROCC Study.50 Without doubt, these algorithms contain some amount of non‐differential exposure misclassification on the occupational coding level. Certainly, as both algorithms contain recodings translating one occupation of classification X into many occupations of classification Y (scenario A) or the other way around (scenario B). Scenario A may result in different exposure estimates where in reality exposure may be identical, while scenario B results in identical exposure estimates where in reality exposure may differ.50,51 Nevertheless, occupational exposure estimates obtained by recoding job codes using an algorithm have been shown to be rather similar to those obtained by direct manual coding of the occupational histories by occupational experts.49 Probably, this result only holds provided that the algorithm recodes to a related and less detailed occupational classification,49 which applies to both algorithms used in our study. 4. A final source of possible exposure misclassification considers the JEMs themselves. JEMs allocate the same exposure estimate to all workers within a job code as opposed to case‐by‐case expert assessment, thereby disregarding possible inter‐individual variability within job codes. This is a major drawback since there may be large differences in exposure levels between individuals with the same job in the same company and/or different companies.

117

Chapter 6

In general, non‐differential exposure misclassification reduces the likelihood of detecting a true association between an exposure and a disease.52‐54 Consequently, non‐differential misclassification will likely result in false negative associations and could exert an influential role in studies that only find a weak association between exposure and disease.55 In addition, using a JEM is more likely to result in Berkson error than classical error by allocating the same exposure estimate to all workers within a job code, thereby disregarding i) an individual's deviation from the risk‐weighted average personal exposure and, when the JEM is partly based on measurements, ii) the difference between the measured and the true ambient levels, which includes spatial variation and instrument error. As such, using a JEM may lead to a loss of power.56 Although it was not possible to determine the extent of misclassification from all the above‐mentioned sources, we used two JEMs to study the asbestos‐related cancer risks in the NLCS in order to reveal some of the uncertainty associated with the choice of exposure method. In the following paragraphs, a short overview of the robustness of results against the use of DOMJEM and FINJEM is given as well as a comprehensive comparison of both JEMs, in order to shed some light on the observed method uncertainty. Robustness of results For the respiratory cancers, there was generally a high degree of similarity in results between both JEMs, which reinforced our confidence in the results presented. Though results were generally robust against the use of both JEMs, DOMJEM showed higher HRs for the duration of high exposure and FINJEM revealed somewhat higher HRs for ever versus never exposed. For the gastrointestinal cancers, results for FINJEM were essentially similar to those for DOMJEM, except for the (prolonged) highly exposed subjects for which DOMJEM was the only one to show (borderline) significantly increased HRs. It has to be noted, however, that ‐ though HRs were sometimes markedly increased ‐ associations were in most instances only marginally significant. For oral cavity and pharyngeal cancer, results were not robust against the use of both JEMs as only the use of FINJEM resulted in finding an association with PhC and OCPC, though an exposure‐response relation was lacking. Therefore, it seems that both JEMs have their particular strengths. DOMJEM appeared better in singling out the highly exposed subjects while FINJEM may be better in discriminating between ever and never‐exposed. This is also reflected by a lower exposure prevalence of the ever highly exposed when using DOMJEM and a somewhat lower exposure prevalence of the ever exposed when using FINJEM. As the prevalence of exposure and the strength of the association both determine the population attributable fraction (PAF) ‐ which we calculated for pleural mesothelioma, lung and

118

General discussion

laryngeal cancer ‐ PAFs for DOMJEM were (slightly) different from those for FINJEM, with the latter generally revealing somewhat higher PAFs. The question that remains is what underlies these different strengths. JEM‐JEM comparison DOMJEM and FINJEM differ on several aspects which might explain the discrepancy in performance between both JEMs: i) the country of origin, ii) the inclusion of a time axis, iii) specificity and sensitivity, and iv) the measurement scale of the exposure estimates. Country of origin Contrary to DOMJEM, which has been developed by occupational exposure experts from the Netherlands for use in general population studies,1 FINJEM has been specifically constructed for use in Finland.2 We did not adapt exposure estimates to Dutch occupational circumstances before application in the NLCS as the inter‐country differences are not easy to predict without an extensive comparative study of the economic, industrial, and legal settings in both countries. A comparison of FINJEM exposure estimates with those from a JEM derived from expert assessments in Montreal showed that agreement on asbestos exposure was at least moderate, depending on the sensitivity analysis performed.51 In the Nordic Occupational Cancer Study, a comparison of FINJEM with local exposure estimates has been carried out for the Nordic countries involved in the study, and despite their largely homogeneous economic structures inter‐country differences have been noticed for the levels and prevalence of occupational asbestos exposure. This was partly due to the mining of asbestos in Finland from 1945‐1972 which did not take place in the other Nordic countries.57 Also in the Netherlands, there were no asbestos mines. Therefore, using FINJEM may have resulted in assignment of higher levels of asbestos exposure to miners where in reality exposure levels were lower for miners in the Netherlands. In DOMJEM, though being designed by occupational exposure experts from the Netherlands, miners are also assigned an exposure to asbestos, quite similar in magnitude to FINJEM. Therefore, this intercountry difference was probably not influential in explaining the difference in performance between both JEMs. Other intercountry differences may have been of importance, such as a more stringent occupational policy in Finland as opposed to the Netherlands. Time axis As opposed to FINJEM, DOMJEM contains no time axis as exposure estimates in this JEM are an estimation of the average exposure over time. Lack of a time axis could introduce bias as exposure levels generally decrease over time, because of better understanding of occupational hazards and subsequent regulations,58,59 resulting in a high rate of false positive exposure estimates. However, the decline mainly concerns the absolute levels of exposure. The relative ranking of jobs in terms of exposure intensity is not expected to change over time or differ between countries. Furthermore, time trends in exposure levels may be modest compared to the difference between low

119

Chapter 6

and high intensity, making inclusion of a time axis of less importance; certainly, as asbestos exposure levels assigned pertain to the period before the ban (i.e., occupations until 1986 were entered by the study subjects, while the asbestos ban in The Netherlands was only enforced in 1993). Specificity versus sensitivity Given that specific occupational exposures are relatively rare, the slightest deviation of perfect specificity will lead to an (marked) underestimation of the degree of association.60 Therefore, it is important to avoid assigning exposure to workers unlikely to be exposed or whose exposure is trivial. The experts who developed the DOMJEM scored occupations, based on their knowledge and experience, with an emphasis on specificity rather than sensitivity by taking the probability of exposure explicitly into account.1 In FINJEM, the probability and intensity of asbestos exposure were assessed on continuous scales. The resulting quantitative exposure estimates enable the user of FINJEM to include or exclude low exposures, which is equivalent to maximizing sensitivity or specificity, respectively.2 We chose not to adapt these estimates, as HRs increased only slightly when asbestos exposure was no longer assigned to those occupations in FINJEM with a prevalence of exposure of <20%, which was the cut‐off point in DOMJEM for assigning low asbestos exposure. Besides, the use of FINJEM even resulted in a slightly lower prevalence of subjects ever exposed to asbestos than the use of DOMJEM (25.8% and 28.8%, respectively). The prevalence of those highly exposed to asbestos was remarkably higher when using FINJEM (13.9%) than when using DOMJEM (2.3%) and may explain the lower HRs found for ever versus never highly exposed and the duration of high exposure when using FINJEM compared to those obtained when using DOMJEM. However, when modeling FINJEM according to DOMJEM by making it more specific for the highly exposed subjects (i.e., a higher prevalence and intensity of exposure is required in order to define a subject as being highly exposed), HRs for mesothelioma increased only slightly. We performed this sensitivity analysis on mesothelioma, because for this cancer it is safe to assume that the higher the HR, the better the performance of the JEM. Ordinal versus continuous exposure estimates DOMJEM contains a combined measure of the probabilityintensity of exposure, which is ordinal (no, low, or high exposure). 1 FINJEM contains continuous estimates of the prevalence and intensity of exposure.2 In theory, quantitative data are to be preferred over qualitative data in assessing exposure. However, this depends on the source (or nature) and the amount of quantitative data. If quantitative data are based on actual exposure measurements, the assumption may hold. As FINJEM is based on both expert assessment and exposure measurements,2 it is not truly quantitative by nature. Furthermore, in DOMJEM a weighting of 0, 1 and 4 was assigned to the exposure intensity scores of no, low and high exposure, respectively, to mirror the log‐normal (multiplicative) nature of occupational exposure levels.1 The weighting was based on

120

General discussion

reported levels for semi‐quantitatively scored exposure, thereby assuring a balanced weighting between intensity and duration in the calculation of cumulative exposure.61 Consequently, although arbitrary units were used in DOMJEM, they reflect an appropriate spacing of the exposure groups and are not per definition inferior to the continuous FINJEM estimates. Nevertheless, using DOMJEM results in unit‐years that are more difficult to interpret and compare than the often reported fiber‐years/ml that FINJEM provides with the downside that we do not know how accurate the estimates are. It is difficult to conclude which of the abovementioned differences between DOMJEM and FINJEM explains the observed discrepancy in performance between both JEMs; especially the markedly better performance of DOMJEM for those highly exposed to asbestos. As specific occupational exposures are relatively rare, a higher specificity of exposure seems to be influential, but increasing specificity for FINJEM by increasing the cut‐off points for defining a subject as being highly exposed hardly increased the HRs of mesothelioma. As the time axis and the measurement level are probably of minor importance, the explanation may involve intercountry differences in asbestos exposure levels or intrinsic differences between both JEMs. Both JEMs have been used in a number of studies on asbestos‐related cancer risks, 1,10,30,62,63 and when interpreting their results, these differences should be kept in mind. 2.2.2 Disease misclassification Disease ascertainment In the literature, misclassification of disease is a serious problem, especially for mesothelioma which did not have a diagnostic category in the ICD system until 1999.17 The NLCS is being monitored for incident cancer by annual record linkage to the Netherlands Cancer Registry and the Dutch Pathology Registry (PALGA).64,65 Data from the Netherlands Cancer Registry are of high quality as specially trained registration clerks register directly from the medical records in the hospital, collecting both clinical and pathological information.66 The record linkage enabled us to restrict our analyses to microscopically confirmed cancer cases classified by anatomic site or histological type, defined by the International Classification of Diseases for Oncology, Third Edition. Furthermore, since completeness of incident cancer coverage was estimated to be almost 100%,47 we do not expect that misclassification of disease has distorted our analyses. In most of the industry‐based (cohort) studies, the main outcome measure was cause‐ specific mortality.11 For diseases with poor survival, such as mesothelioma and lung cancer, mortality is likely suitable for assessing cancer incidence. This does not hold, however, for diseases with a fairly good prognosis such as colorectal cancer. In addition, the use of mortality data for assessing cancer incidence also depends on the

121

Chapter 6

accuracy of both identification and coding of the cause of death. As clinicians became aware of the asbestos‐related cancer risks, they may have diagnosed diseases such as lung cancer more frequently among workers in known asbestos‐related industries than in the general population, resulting in some amount of disease misclassification and possibly an overestimation of the observed HRs.11 Therefore, it is important that population‐based studies with information on cancer incidence have confirmed the causal relation between asbestos exposure and the risk of various cancers. Latency (i.e. time since first occupational exposure to asbestos) The risk of cancer in asbestos workers is very much dependent on latency period, with most of the increased risk occurring after at least 25 years from the onset of occupational asbestos exposure.67 As a result of this long latency, short term follow‐up may impede finding a positive association if such an association were truly present. Previous studies have dealt with latency in various ways. Some studies, especially earlier reports, accumulated person‐years at risk upon first exposure, which may dilute the observed risk as many years of low risk are included. Others have only accumulated person‐years at risk after a certain period of time upon first exposure, usually 20 years. Moreover, the number of years of follow‐up since first occupational exposure to asbestos varied widely between studies.17 As a result of the case‐cohort design in the NLCS, we were not able to include a lag period to avoid a variable latency time resulting from truncating occupational histories at 1986. As participants were aged 55‐69 years at baseline, when they reported their occupational history, and were followed for 17.3 years, time since first exposure was for most participants long enough for the selected cancers to develop. Furthermore, stratifying analyses on follow‐up period (cut‐off point 8 years) even showed higher HRs for the shorter follow‐up period, though differences between both periods were small. Therefore, although we were not able to avoid a variable latency time, we do not believe that our inability to include a lag‐period has biased our results.

2.3 Confounding Confounding is a potential problem in observational studies for which has to be controlled, either in the design phase of the study by restriction of the selected study population or in the analysis phase by stratification or by multivariable‐adjusted modelling. If left unattended, bias may be introduced resulting in an under‐ or overestimation of the risk estimates. Certainly in studies with only modestly increased risks, concerns are often raised that the observed associations may be explained by differences in smoking, alcohol consumption, or other lifestyle factors between exposure groups.68‐70

122

General discussion

In most industry‐based studies, there is generally only little information available to adjust for potential confounders, in contrast to most population‐based studies.55 Therefore, in many occupational reports the subject of uncontrolled confounding is extensively discussed. As the NLCS was designed to study diet and cancer, we were able to address potential confounding by nonoccupational risk factors (such as smoking, alcohol consumption, and other lifestyle factors). We adjusted for these factors by including them as covariates in the multivariable‐adjusted models. The covariates included were either a priori‐selected risk factors based on the literature or variables that changed the age‐adjusted regression coefficients by at least 10 percent (using a backwards stepwise procedure). Several associations with overall gastric cancer and GNCA were reduced and became non‐significant after adjusting for smoking status. Furthermore, HRs of PhC and OCPC increased after adjustment for especially alcohol consumption and socioeconomic status. For the other cancers, our results do not point at an influential role of potential confounders when studying the association with occupational asbestos exposure. Consequently, in several instances taking into account lifestyle factors was important in the NLCS and may be necessary when studying the respective cancers. Of course, among others this depends on the comparability of the study groups (i.e., blue collar‐blue collar or blue collar‐general population). In industry‐ based studies in which different groups of blue collar workers are being compared, the question remains if one should be concerned that unmeasured confounding either increases or decreases the risk estimates substantially, as it seems unlikely that different groups of blue collar workers within the same company, or industry for that matter, have systematic differences in for example smoking and drinking habits as large as those observed between blue and white collar workers. In many industry‐based studies, however, disease rates of a worker cohort and the general population are being compared and confounding may bias the risk estimates. As is true for occupational asbestos exposure, confounders are often measured imperfectly. This entails either exposure misclassification or too wide exposure categories, which both could lead to residual confounding.71 The amount of confounding that remains after adjustment is probably proportional to the amount removed.72 Therefore, it is important to investigate the influence of adjustment for confounders on the asbestos‐related cancer risks by presenting both the age‐adjusted (and, if applicable, the for family history of cancer adjusted) and the multivariable‐ adjusted results. As stated above, the confounders included in the multivariable‐ adjusted models did not substantially alter the results, except for smoking status in the case of overall gastric cancer and GNCA, and alcohol consumption and socioeconomic status in the case of PhC and OCPC. Hence, for these cancers residual confounding may have been present.

123

Chapter 6

3. IMPLICATIONS FOR PUBLIC HEALTH The results of our study are of importance to public health, both with respect to i) the risk assessment of low asbestos exposure levels in order to set acceptable exposure limits and to ii) estimating the impact of asbestos‐related diseases on society, in the Netherlands as well as in low‐ and middle‐income countries.

3.1 Risk assessment of low asbestos levels: qualitative versus quantitative A brief inspection of some of the more recent meta‐analyses conducted on occupational asbestos exposure and cancer‐related risk reveals that most studies included are not very recent and frequently involved heavily exposed subjects.8,11,73‐75 Currently, at least in developed countries, most individuals are no longer exposed to these high levels due to increased regulation on the use of asbestos. Due to the uncertainty of extrapolating risks associated with high levels of asbestos exposure to low levels of exposure, the focus has shifted towards estimating excess risk accurately for levels encountered at the lower end of the exposure distribution; a topic of debate particularly for mesothelioma and lung cancer.4‐10 The NLCS is a population‐based study with a wide range in exposure levels, including those at the lower end of the exposure distribution, (i.e. exposure levels in jobs outside asbestos mining, insulation, cement and textile manufacturing, and other more highly exposed jobs). This enabled us to study if cancer risks are not only increased for higher asbestos exposure levels but also for levels encountered at the lower end of the exposure distribution. When using FINJEM, we observed significantly increased risks of pleural mesothelioma (2.69(1.60‐ 4.53)) and lung cancer (1.44(1.12‐1.86)) for the lowest tertile of cumulative asbestos exposure (median: 0.20 f‐y/ml), but not when using DOMJEM. As not all results were robust against the use of two JEMs, we refrained in chapter 2 from comparing our JEM‐based results quantitatively to the large body of literature available on the subject; also because this was not our initial goal. In addition, we have no information on fiber type and size in the NLCS, both of which are presumed to have a (rather) large influence on the strength of the associated cancer risks,73,76 though recent investigations have challenged this assumption.75,77 As we can see that population‐based studies, such as the NLCS, are often the only direct source available on cancer risks associated with the lower end of the exposure distribution and therefore potentially interesting for setting acceptable exposure limits, we calculated the excess risk of mesothelioma and lung cancer per unit increase (unit‐years or fiber‐ years; for DOMJEM and FINJEM, respectively) of asbestos exposure (expressed as the potency or the so‐called KM value (for mesothelioma) and KL value (for lung cancer)). Based on these potency factors and a life table analysis it is possible to determine the excess risk in order to assess if the prevailing standards are sufficiently strict. In 2010,

124

General discussion

the Health Council of the Netherlands has updated the occupational exposure limits for asbestos based on a number of studies that demonstrated that environmental exposure to asbestos at levels much lower than typically seen in occupational situations may cause mesothelioma.78‐82 We calculated the excess risk using a model with a fixed intercept of α=1 since no confounding was present in our study for mesothelioma (as expected) and lung cancer. For mesothelioma, a Km of 9.06 and 50.79 was observed when using DOMJEM and FINJEM, respectively. For lung cancer, a Kl of 1.02 and 13.06 was observed when using DOMJEM and FINJEM, respectively. As previous studies calculated these potency factors per fiber type (i.e., chrysotile and amphibole, with those exposed to both being categorized as having experienced “mixed” exposure; if possible, amphiboles were subdivided into crocidolite, amosite, and tremolite) in contrast to what was possible in our study, we compared our results to those for a mixed exposure type, though we know that for most branches of industry in the Netherlands, chrysotile will have constituted more than 95% of all exposure to asbestos fibres,83 which is presumed to be (far) less carcinogenic than amphibole asbestos. When comparing our results, despite methodological disparities, to recent meta‐analyses performed under the authority of the Health Council of the Netherlands, showing a meta‐Km of 1.3 and a meta‐Kl of 0.48,75,77,84 Km and Kl values in our study are rather high. Gustavsson et al. and van der Bij et al. also observed that relative risks at the lower end of the exposure distribution might be higher than predicted by downward linear extrapolation from highly exposed occupational cohorts, with a Kl of 15.5 for the Gustavsson study.8‐10 Together with the fact that we observed increased risks for the lowest tertile of exposure, with the acceptable risks being even lower, our results may be of importance for a subsequent update of the Health Council of the Netherlands to assess if the current standard is sufficiently strict to protect the general public and those occupationally involved with asbestos.

3.2 Burden of asbestos‐related disease in the Netherlands In order to estimate the burden of disease due to occupational asbestos exposure, we calculated the population attributable fraction (PAF) for pleural mesothelioma, lung and laryngeal cancer in chapter 2. The PAF is the proportion of the total burden of a disease that would not have occurred had the factor been nonexistent in the population in question.46 Population‐based studies are well‐suited to estimate the impact at a population level, as opposed to industry‐based studies which are seldom used to estimate the PAF. In the latter, the prevalence and intensity of exposure are mostly higher than in the general population and are, as a consequence, not representative for that population.

125

Chapter 6

Though suitable, mostly no measurements are available in population‐based studies and some of the difficulties surrounding retrospective exposure assessment also play a part in the calculation of PAFs. In chapter 2, we noticed that PAFs were affected by the choice of JEM in that they were overall higher (up to threefold) for FINJEM than for DOMJEM. Furthermore, PAFs for mesothelioma (31.9% for DOMJEM and 34.3% for FINJEM) are lower than reported in the literature.85‐88 As virtually all mesothelioma is due to (occupational) asbestos, higher PAFs might have been expected. Therefore, some caution seems appropriate when judging PAFs which are JEM‐based. An advantage of the NLCS is that we were able to incorporate lower level exposures, which contribute to a substantial proportion of the excess number of cases as this concerns a large number of workers.85,89 Despite methodological disparities between studies, PAFs for lung cancer (5.2% for DOMJEM and 11.5% for FINJEM) are well within the range reported in the literature (5.4%‐35.9%).10,85‐88,90 For laryngeal cancer, PAFs in the range of 8.1%‐13.2% have been reported in previous studies. The PAF of 9.8% for laryngeal cancer estimated with FINJEM fits well into this range. The PAF derived by using DOMJEM was somewhat lower (5.5%).87,90 As difficulties with the management of asbestos exposure arising during asbestos removal and site cleaning in the general environment have been reported, exposure to asbestos still occurs.91 In addition, we were not able to include environmental asbestos exposure or exposure of women through their occupationally exposed husbands. Also, we did not calculate the PAFs for the other cancers as the body of evidence is still too small for judging the associations as causal. If these sources also contribute to cancer incidence, the burden of asbestos‐related diseases is even higher than already expected, with a forecast of at least 13,000 mesothelioma deaths and another 13,000 deaths due to asbestos‐related lung cancer in the period 2000‐2028 in the Netherlands.92

3.3 Asbestos burden in low‐ and middle‐income countries As already mentioned in the introduction, an estimated 125 million people worldwide were still occupationally exposed to asbestos in 2006,93 with a large number of countries still using, importing, and exporting asbestos and asbestos‐containing products. These are almost all countries in Asia, Eastern Europe, Latin America, and Africa.94 Predicting the future course of the asbestos epidemic in low‐ and middle‐ income countries is severely hampered by the paucity of data on local exposure and disease occurrence. Although little is known about the level of asbestos exposure in low‐ and middle‐income countries, what has been reported suggests that the levels may be quite high in India and China.95 The mere fact that asbestos exposure does not seem to be monitored, or at least is not being reported, in many of the countries that are currently producing or using asbestos is worrying. It suggests that these countries do not have the technology and expertise ‐ or political willingness ‐ to cope with this

126

General discussion

hazardous material.95 The increasing trend in asbestos use and the probably poor control of asbestos exposure will translate into a future rise in the incidence of asbestos‐related diseases in these countries. This may have far‐reaching consequences as health services in these countries are less available than in industrialized countries.96 On top of this, other factors are involved, including a poor socio‐economic status,97 which may intensify the asbestos‐related health risks. As even the best workplace control measures cannot prevent exposure to asbestos once in use or when joining the waste stream, we can only strongly support the call for a worldwide ban on all types of asbestos. Safer substitutes for asbestos exist and are feasible for use in low‐ and middle‐income countries.98

4. POPULATION‐BASED VERSUS INDUSTRY‐BASED STUDIES: NLCS As most population‐based studies, the NLCS contains a wide range in exposure levels including those at the lower end of the exposure distribution which are nowadays present in most industrialized countries. Results indicated that relative risks at the lower end of the exposure distribution might be higher than predicted by downward linear extrapolation from highly exposed occupational cohorts. As such, studies as the NLCS can potentially be useful for setting acceptable exposure limits. Furthermore, we were able to adjust for potential confounders and in several instances this made a difference. In addition, the large study size enabled us to study relatively rare cancers, cancer subtypes, and a possible interaction between asbestos exposure and smoking in relation to cancer risk, with some novel observations. As in most population‐based studies, we had no information on fiber type as no measurements were available. Nevertheless, we could dispose of different JEMs in order to provide insight into the methodological uncertainty associated with the choice of JEM. DOMJEM and FINJEM appeared to have their own strengths and though not all results were robust against the use of DOMJEM and FINJEM, both JEMs were able to confirm the well‐established associations between asbestos and mesothelioma, lung and laryngeal cancer. As DOMJEM and FINJEM contain exposures other than asbestos, both JEMs could be useful in future occupational studies in the NLCS. In conclusion, the population‐based NLCS has some important strengths when it comes to studying the association between asbestos ‐ and possibly other occupational exposures ‐ and the risk of cancer. The weakness of not having any exposure measurements has been partly overcome by relying on more than one JEM. It depends on the research question which type of study (population‐based such as the NLCS or industry‐based) to prefer. However, in order to decide whether a substance is

127

Chapter 6

carcinogenic or not, both types of study are needed, so preferably results from one type of study are replicated in the other.

5. RECOMMENDATIONS FOR FUTURE RESEARCH As our goals were twofold, i) studying the reliability of three JEMs and the risk prediction ability for asbestos exposure of two of them in the NLCS, providing insight into the methodological uncertainty associated with the choice of JEM and ii) shedding light on several remaining questions on the asbestos‐cancer relations, our final recommendations are also twofold. First, our study showed that (small) differences in the ability of both JEMs to discriminate between never and ever exposed and low and high exposure may lead to (slightly) different results and judgment if asbestos is related to the endpoint of interest or not. Therefore, though JEMs are an established means of retrospectively assessing occupational exposure, they need to be elaborated in order to reduce the amount of exposure misclassification as far as possible. Until this has been achieved, we suggest that incorporation of multiple JEMs in epidemiological studies should be encouraged as it reveals some of the uncertainty associated with the choice of exposure method and therefore aids in the interpretation of data. Second, though it might still be interesting from a scientific point of view to study if asbestos is also associated with other (more rare) tumors, no more research on asbestos is needed to determine if asbestos is carcinogenic and should be globally banned for that reason. All asbestos fiber types are carcinogenic and no more lives should be sacrificed for economical and/or political purposes. Therefore, I appeal to the moral sense of those in the position to end the everlasting story asbestos has become, with the asbestos industry just simply moving to those places where it has not been banned yet. Safer alternatives are available and feasible and should be used instead of asbestos.

128

General discussion

REFERENCES 1.

2. 3.

5. 6. 7. 8.

10.

11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22. 23. 24.

Peters S, Vermeulen R, Cassidy A, Mannetje A', van Tongeren M, Boffetta P, Straif K, Kromhout H; INCO Group. Comparison of exposure assessment methods for occupational carcinogens in a multi‐centre lung cancer case‐control study. Occup Environ Med. 2011;68:148‐53. Kauppinen T, Toikkanen J, Pukkala E. From cross‐tabulations to multipurpose exposure information systems: a new job‐exposure matrix. Am J Ind Med. 1998;33:409‐17. Swuste P, Dahhan M, Burdorf A. Linking expert judgement and trends in occupational exposure into a job‐exposure matrix for historical exposure to asbestos in the Netherlands. Ann Occup Hyg. 2008;52:397‐403. Iwatsubo Y, Pairon JC, Boutin C, Ménard O, Massin N, Caillaud D, Orlowski E, Galateau‐Salle F, Bignon J, Brochard P. Pleural mesothelioma: dose‐response relation at low levels of asbestos exposure in a French population‐based case‐control study. Am J Epidemiol. 1998;148:133‐42. Hillerdal G. Mesothelioma: cases associated with non‐occupational and low dose exposures. Occup Environ Med. 1999;56:505‐13. Ilgren EB, Browne K. Asbestos‐related mesothelioma: evidence for a threshold in animals and humans. Regul Toxicol Pharmacol. 1991;13:116‐32. Gibbs GW, Berry G. Mesothelioma and asbestos. Regul Toxicol Pharmacol. 2008;52:S223‐31. van der Bij S, Koffijberg H, Lenters V, Portengen L, Moons KG, Heederik D, Vermeulen RC. Lung cancer risk at low cumulative asbestos exposure: meta‐regression of the exposure‐response relationship. Cancer Causes Control. 2013;24:1‐12 Gustavsson P, Nyberg F, Pershagen G, Schéele P, Jakobsson R, Plato N. Low‐dose exposure to asbestos and lung cancer: dose‐response relations and interaction with smoking in a population‐based case‐ referent study in Stockholm, Sweden. Am J Epidemiol. 2002;155:1016‐22. De Matteis S, Consonni D, Lubin JH, Tucker M, Peters S, Vermeulen RCh, Kromhout H, Bertazzi PA, Caporaso NE, Pesatori AC, Wacholder S, Landi MT. Impact of occupational carcinogens on lung cancer risk in a general population. Int J Epidemiol. 2012;41:711‐21. NAS. Asbestos: Selected Cancers. Washington, DC: The National Academies Press, 2006. Ramroth H, Ahrens W, Dietz A, Becher H. Occupational asbestos exposure as a risk factor for laryngeal carcinoma in a population‐based case‐control study from Germany. Am J Ind Med. 2011;54:510‐4. Johansson L, Albin M, Jakobsson K, Mikoczy Z. Histological type of lung carcinoma in asbestos cement workers and matched controls. Br J Ind Med. 1992;49:626‐30. Raffn E, Lynge E, Korsgaard B. Incidence of lung cancer by histological type among asbestos cement workers in Denmark. Br J Ind Med. 1993;50:85‐9. Lee BW, Wain JC, Kelsey KT, Wiencke JK, Christiani DC. Association of cigarette smoking and asbestos exposure with location and histology of lung cancer. Am J Respir Crit Care Med. 1998;157:748‐55. Paris C, Benichou J, Saunier F, Metayer J, Brochard P, Thiberville L, Nouvet G. Smoking status, occupational asbestos exposure and bronchial location of lung cancer. Lung Cancer. 2003;40:17‐24. IARC. Asbestos (Chrysotile, Amosite, Crocidolite, Tremolite, Actinolite and Anthophyllite). In: A Review of Human Carcinogens: Arsenic, Metals, Fibres, and Dusts. Lyon, France: International Agency for Research on Cancer (IARC), 2012. 219‐309. Schottenfeld D, Fraumeni J. Cancer Epidemiology and Prevention. New York: Oxford University Press, 2006. Kanarek MS. Mesothelioma from chrysotile asbestos: update. Ann Epidemiol. 2011;21:688‐97. Erren TC, Jacobsen M, Piekarski C. Synergy between asbestos and smoking on lung cancer risks. Epidemiology. 1999;10:405‐11. Lee PN. Relation between exposure to asbestos and smoking jointly and the risk of lung cancer. Occup Environ Med. 2001;58:145‐53. Liddell FD. The interaction of asbestos and smoking in lung cancer. Ann Occup Hyg. 2001;45:341‐56. IARC. Tobacco smoke. In: A Review of Human Carcinogens: Tobacco Smoke and Involuntary Smoking. Lyon, France: International Agency for Research on Cancer (IARC), 2004. 917–9. Gamble J. Risk of gastrointestinal cancers from inhalation and ingestion of asbestos. Regul Toxicol Pharmacol. 2008;52:S124‐53.

129

Chapter 6

25. Frumkin H, Berlin J. Asbestos exposure and gastrointestinal malignancy review and meta‐analysis. Am J Ind Med. 1988;14:79‐95. 26. Morgan RW, Foliart DE, Wong O. Asbestos and gastrointestinal cancer. A review of the literature. West J Med. 1985;143:60‐5. 27. Gustavsson P, Jakobsson R, Johansson H, Lewin F, Norell S, Rutkvist LE. Occupational exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus: a case‐control study in Sweden. Occup Environ Med. 1998;55:393‐400. 28. Parent ME, Siemiatycki J, Fritschi L. Workplace exposures and oesophageal cancer. Occup Environ Med. 2000;57:325‐34. 29. Jansson C, Johansson AL, Bergdahl IA, Dickman PW, Plato N, Adami J, Boffetta P, Lagergren J. Occupational exposures and risk of esophageal and gastric cardia cancers among male Swedish construction workers. Cancer Causes Control. 2005;16:755‐64. 30. Santibañez M, Vioque J, Alguacil J, Barber X, García de la Hera M, Kauppinen T; PANESOES Study Group. Occupational exposures and risk of oesophageal cancer by histological type: a case‐control study in eastern Spain. Occup Environ Med. 2008;65:774‐81. 31. Jansson C, Plato N, Johansson AL, Nyrén O, Lagergren J. Airborne occupational exposures and risk of oesophageal and cardia adenocarcinoma. Occup Environ Med. 2006;63:107‐12. 32. Crew KD, Neugut AI. Epidemiology of gastric cancer. World J Gastroenterol 2006;12:354‐62. 33. Sjödahl K, Jansson C, Bergdahl IA, Adami J, Boffetta P, Lagergren J. Airborne exposures and risk of gastric cancer: a prospective cohort study. Int J Cancer 2007;120:2013‐8. 34. Jakobsson K, Albin M, Hagmar L. Asbestos, cement, and cancer in the right part of the colon. Occup Environ Med. 1994;51:95‐101. 35. Gerhardsson de Verdier M, Plato N, Steineck G, Peters JM. Occupational exposures and cancer of the colon and rectum. Am J Ind Med. 1992;22:291‐303. 36. Ferrante D, Bertolotti M, Todesco A, Mirabelli D, Terracini B, Magnani C. Cancer mortality and incidence of mesothelioma in a cohort of wives of asbestos workers in Casale Monferrato, Italy. Environ Health Perspect. 2007;115:1401‐5. 37. Liang PS, Chen TY, Giovannucci E. Cigarette smoking and colorectal cancer incidence and mortality: systematic review and meta‐analysis. Int J Cancer. 2009;124:2406‐15. 38. Liddell FD, McDonald AD, McDonald JC. The 1891‐1920 birth cohort of Quebec chrysotile miners and millers: development from 1904 and mortality to 1992. Ann Occup Hyg. 1997;41:13‐36. 39. Aliyu OA, Cullen MR, Barnett MJ, Balmes JR, Cartmel B, Redlich CA, Brodkin CA, Barnhart S, Rosenstock L, Israel L, Goodman GE, Thornquist MD, Omenn GS. Evidence for excess colorectal cancer incidence among asbestos‐exposed men in the Beta‐Carotene and Retinol Efficacy Trial. Am J Epidemiol. 2005;162:868‐78. 40. Paget‐Bailly S, Cyr D, Luce D. Occupational exposures to asbestos, polycyclic aromatic hydrocarbons and solvents, and cancers of the oral cavity and pharynx: a quantitative literature review. Int Arch Occup Environ Health. 2012;85:341‐51. 41. Reid A, Ambrosini G, de Klerk N, Fritschi L, Musk B. Aerodigestive and gastrointestinal tract cancers and exposure to crocidolite (blue asbestos): incidence and mortality among former crocidolite workers. Int J Cancer. 2004;111:757‐61. 42. Purdue MP, Järvholm B, Bergdahl IA, Hayes RB, Baris D. Occupational exposures and head and neck cancers among Swedish construction workers. Scand J Work Environ Health. 2006;32:270‐5. 43. Langevin SM, O'Sullivan MH, Valerio JL, Pawlita M, Applebaum KM, Eliot M, McClean MD, Kelsey KT. Occupational asbestos exposure is associated with pharyngeal squamous cell carcinoma in men from the greater Boston area. Occup Environ Med. 2013;70:858‐63. 44. Marchand JL, Luce D, Leclerc A, Goldberg P, Orlowski E, Bugel I, Brugère J. Laryngeal and hypopharyngeal cancer and occupational exposure to asbestos and man‐made vitreous fibers: results of a case‐control study. Am J Ind Med. 2000;37: 581‐9. 45. Tarvainen L, Kyyrönen P, Kauppinen T, Pukkala E. Cancer of the mouth and pharynx, occupation and exposure to chemical agents in Finland [in 1971‐95]. Int J Cancer. 2008;123:653‐9. 46. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. Philadelphia: Lippincott Williams & Wilkins, 2008.

130

General discussion

47. Goldbohm RA, van den Brandt PA, Dorant E. Estimation of the coverage of Dutch municipalities by cancer registries and PALGA based on hospital discharge data. Tijdschr Soc Gezonheidsz. 1994;72:80‐4. 48. Kromhout H, Vermeulen R. Application of job‐exposure matrices in studies of the general population: some clues to their performance. Eur Respir Rev. 2001;11:80‐90. 49. Koeman T, Offermans NS, Christopher‐de Vries Y, Slottje P, Van Den Brandt PA, Goldbohm RA, Kromhout H, Vermeulen R. JEMs and incompatible occupational coding systems: effect of manual and automatic recoding of job codes on exposure assignment. Ann Occup Hyg. 2013;57:107‐14. 50. van Tongeren M, Kincl L, Richardson L, Benke G, Figuerola J, Kauppinen T, Lakhani R, Lavoué J, McLean D, Plato N, Cardis E; INTEROCC STUDY GROUP. Assessing occupational exposure to chemicals in an international epidemiological study of brain tumours. Ann Occup Hyg. 2013;57:610‐26. 51. Lavoué J, Pintos J, Van Tongeren M, Kincl L, Richardson L, Kauppinen T, Cardis E, Siemiatycki J. Comparison of exposure estimates in the Finnish job‐exposure matrix FINJEM with a JEM derived from expert assessments performed in Montreal. Occup Environ Med. 2012;69:465‐71. 52. Copeland KT, Checkoway H, McMichael AJ, Holbrook RH. Bias due to misclassification in the estimation of relative risk. Am J Epidemiol. 1977;105:488‐95. 53. Dosemeci M, Wacholder S, Lubin JH. Does nondifferential misclassification of exposure always bias a true effect toward the null value? Am J Epidemiol. 1990;132:746‐8. 54. Loomis D, Salvan A, Kromhout H, Kriebel D. Selecting indices of occupational exposure for epidemiological studies. Occup Eviron Med. 1999;5:73‐91. 55. Pearce N, Checkoway H, Kriebel D. Bias in occupational epidemiology studies. Occup Environ Med. 2007;64:562‐8. 56. Zeger SL, Thomas D, Dominici F, Samet JM, Schwartz J, Dockery D, Cohen A. Exposure measurement error in time‐series studies of air pollution: concepts and consequences. Environ Health Perspect. 2000;108:419‐26. 57. Kauppinen T, Heikkilä P, Plato N, Woldbaek T, Lenvik K, Hansen J, Kristjansson V, Pukkala E. Construction of job‐exposure matrices for the Nordic Occupational Cancer Study (NOCCA). Acta Oncol. 2009;48:791‐800. 58. Siemiatycki J, Richardson L, Straif K, Latreille B, Lakhani R, Campbell S, Rousseau MC, Boffetta P. Listing occupational carcinogens. Environ Health Perspect. 2004;112:1447‐59. 59. Creely KS, Cowie H, Van Tongeren M, Kromhout H, Tickner J, Cherrie JW. Trends in inhalation exposure‐ ‐a review of the data in the published scientific literature. Ann Occup Hyg. 2007;51:665‐78. 60. Kromhout H, Heederik D, Dalderup LM, Kromhout D. Performance of two general job‐exposure matrices in a study of lung cancer morbidity in the Zutphen cohort. Am J Epidemiol. 1992;136:698‐711. 61. Stewart PA, Herrick RF, Blair A, Checkoway H, Droz P, Fine L, Fischer L, Harris R, Kauppinen T, Saracci R. Highlights of the 1990 Leesburg, Virginia, International Workshop on Retrospective Exposure Assessment for Occupational Epidemiology Studies. Scand J Work Environ Health. 1991;17:281‐5. 62. Hardt JS, Vermeulen R, Peters S, Kromhout H, McLaughlin JR, Demers PA. A comparison of exposure assessment approaches: lung cancer and occupational asbestos exposure in a population‐based case‐ control study. Occup Environ Med. 2014;71:282‐8. 63. Santibañez M, Alguacil J, de la Hera MG, Navarrete‐Muñoz EM, Llorca J, Aragonés N, Kauppinen T, Vioque J; PANESOES Study Group. Occupational exposures and risk of stomach cancer by histological type. Occup Environ Med. 2012;69:268‐75. 64. Casparie M, Tiebosch AT, Burger G, Blauwgeers H, van de Pol A, van Krieken JH, Meijer GA. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol. 2007;29:19‐24. 65. Van den Brandt PA, Schouten LJ, Goldbohm RA, Dorant E, Hunen PM. Development of a record linkage protocol for use in the Dutch Cancer Registry for Epidemiological Research. Int J Epidemiol. 1990;19:553‐8. 66. Schouten LJ, Jager JJ, van den Brandt PA. Quality of cancer registry data: a comparison of data provided by clinicians with those of registration personnel. Br J Cancer. 1993;68:974‐7. 67. Selikoff IJ, Seidman H. Asbestos‐associated deaths among insulation workers in the United States and Canada, 1967‐1987. Ann N Y Acad Sci. 1991;643:1‐14. 68. Axelson O. Confounding from smoking in occupational epidemiology. Br J Ind Med. 1989;46:505‐7.

131

Chapter 6

69. Blair A, Steenland K, Shy C, O'Berg M, Halperin W, Thomas T. Control of smoking in occupational epidemiologic studies: methods and needs. Am J Ind Med. 1988;13:3‐4. 70. Blair A, Stewart WF, Stewart PA, Sandler DP, Axelson O, Vineis P, Checkoway H, Savitz D, Pearce N, Rice C. A philosophy for dealing with hypothesized uncontrolled confounding in epidemiological investigations. Med Lav. 1995;86:106‐10. 71. Szklo M, Nieto FJ. Epidemiology, Beyond the Basics. Sudbury, Massachusetts: Jones and Bartlett Publishers, 2007. 72. Savitz DA, Baron AE. Estimating and correcting for confounder misclassification. Am J Epidemiol. 1989;129:1062‐71. 73. Berman DW, Crump KS. A meta‐analysis of asbestos‐related cancer risk that addresses fiber size and mineral type. Crit Rev Toxicol. 2008;38 Suppl 1:49‐73. 74. Hodgson JT, Darnton A. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure. Ann Occup Hyg. 2000;44:565‐601. 75. Lenters V, Vermeulen R, Dogger S, Stayner L, Portengen L, Burdorf A, Heederik D. A meta‐analysis of asbestos and lung cancer: is better quality exposure assessment associated with steeper slopes of the exposure‐response relationships? Environ Health Perspect. 2011;119:1547‐55. 76. Berman DW, Crump KS. Update of potency factors for asbestos‐related lung cancer and mesothelioma. Crit Rev Toxicol. 2008;38 Suppl 1:1‐47. 77. Burdorf A, Heederik D. Applying quality criteria to exposure in asbestos epidemiology increases the estimated risk. Ann Occup Hyg. 2011;55:565‐8. 78. Burdorf A, Dahhan M, Swuste PHJJ. Milieublootstelling aan asbest en het optreden van pleura mesothelioom. Ned Tijschr Geneesk. 2004;148:1727‐31. 79. Sinninghe Damsté HEJ, Siesling S, Burdorf A. Milieublootstelling aan asbest in de regio Goor vastgesteld als oorzaak van maligne pleuramesothelioom bij vrouwen. Ned Tijdschr Geneeskd. 2007;151:2453‐9. 80. Driece HA, Siesling S, Swuste PH, Burdorf A. Assessment of cancer risks due to environmental exposure to asbestos. J Expo Sci Environ Epidemiol. 2010;20:478‐85. 81. Magnani C, Dalmasso P, Biggeri A, Ivaldi C, Mirabelli D, Terracini B. Increased risk of malignant mesothelioma of the pleura after residential or domestic exposure to asbestos: a case‐control study in Casale Monferrato, Italy. Environ Health Perspect. 2001;109:915‐9. 82. Kurumatani N, Kumagai S. Mapping the risk of mesothelioma due to neighborhood asbestos exposure. Am J Respir Crit Care Med. 2008;178:624‐9. 83. Burdorf A, Dahhan M, Swuste P. Occupational characteristics of cases with asbestos‐related diseases in The Netherlands. Ann Occup Hyg. 2003;47:485‐92. 84. Gezondheidsraad. Asbestos: risks of environmental and occupational exposure. The Hague, the Netherlands: Health Council of the Netherlands, report 2010/10E. http://www.gezondheidsraad.nl/sites/default/files/201010E.pdf. Accessed February 4, 2015. 85. Albin M, Magnani C, Krstev S, Rapiti E, Shefer I. Asbestos and cancer: An overview of current trends in Europe. Environ Health Perspect. 1999;107 Suppl 2:289‐98. 86. Driscoll T, Nelson DI, Steenland K, Leigh J, Concha‐Barrientos M, Fingerhut M, Prüss‐Ustün A. The global burden of disease due to occupational carcinogens. Am J Ind Med. 2005;48:419‐31. 87. Nurminen M, Karjalainen A. Epidemiologic estimate of the proportion of fatalities related to occupational factors in Finland. Scand J Work Environ Health. 2001;27:161‐213. 88. Boffetta P, Autier P, Boniol M, Boyle P, Hill C, Aurengo A, Masse R, Thé Gd, Valleron AJ, Monier R, Tubiana M. An estimate of cancers attributable to occupational exposures in France. J Occup Environ Med. 2010;52:399‐406. 89. Rushton L, Bagga S, Bevan R, Brown TP, Cherrie JW, Holmes P, Fortunato L, Slack R, Van Tongeren M, Young C, Hutchings SJ. Occupation and cancer in Britain. Br J Cancer. 2010;102:1428‐37. 90. Serrier H, Sultan‐Taieb H, Luce D, Bejean S. Estimating the social cost of respiratory cancer cases attributable to occupational exposures in France. Eur J Health Econ. 2014;15:661‐73. 91. Provincie Noord‐Holland. Asbestketen ontrafeld. Eindrapport ketenonderzoek asbest provincie Noord‐ Holland. 2010. 92. Segura O, Burdorf A, Looman C. Update of predictions of mortality from pleural mesothelioma in the Netherlands. Occup Environ Med. 2003;60:50‐5.

132

General discussion

93. WHO. Elimination of asbestos‐related diseases. http://whqlibdoc.who.int/hq/2006/WHO_SDE_OEH_ 06.03_eng.pdf. Accessed February 4, 2015. 94. Ramazzini C. Asbestos is still with us: repeat call for a universal ban. J Occup Environ Med. 2010;52: 469‐72. 95. Stayner L, Welch LS, Lemen R. The worldwide pandemic of asbestos‐related diseases. Annu Rev Public Health. 2013;34:205‐16. 96. Levy BS. Global occupational health issues: Working in partnership to prevent illness and injury. AAOHN J. 1996;44:244‐7; discussion 247. 97. Kromhout H. Occupational hygiene in developing countries: something to talk about? Ann Occup Hyg. 1999;43:501‐3. 98. World Bank Group (WBG). Good Practice Note: Asbestos: Occupational and Community Health Issues. http://www.bwint.org/pdfs/WB‐AsbestosGuidanceNote.pdf. Accessed February 4, 2015.

133

Chapter 6

Summary

135

Summary

SUMMARY The term asbestos is applied to several silicate minerals when they occur in a fibrous form known as ‘asbestiform’, characterized by bundles of thin, long, and separable fibers. Asbestos minerals possess a number of properties useful in commercial applications: heat stability, thermal and electrical insulation, wear and friction characteristics, tensile strength, the ability to be woven, and resistance to chemical and biological degradation. It was gradually recognized that using asbestos is not only advantageous but is also associated with the occurrence of serious health consequences, which resulted in increasingly strict measures to reduce exposure to asbestos. A number of adverse health outcomes is now established as causally associated with exposure to asbestos, regardless of fiber type: asbestosis, mesothelioma (i.e., cancers of the pleura and peritoneum), lung, laryngeal, and ovarian cancer. Although the health effects of asbestos are one of the best documented in occupational health, they are also one of the most controversial as there is still considerable debate around asbestos carcinogenicity. Firstly, remaining questions pertain to the risk at the lower end of the exposure distribution for mesothelioma and lung cancer, in order to set acceptable exposure limits and substantiate compensation claims. Secondly, it still needs to be determined if asbestos is causally related to gastrointestinal cancers and oral cavity and pharyngeal cancer for which currently limited evidence exists. Thirdly, most studies have been carried out in an industry setting with limited information on important confounders, resulting in the possibility of uncontrolled confounding due to lifestyle factors as smoking and alcohol for especially laryngeal, pharyngeal, and oral cavity cancer. Fourthly, except for lung cancer the number of cases was generally too small in most industry‐based studies to investigate the association with subtypes of cancer. Finally, the possible interaction between asbestos and smoking in relation to cancer has only been thoroughly examined for lung cancer with considerable variability in the magnitude of joint effects varying between additive and multiplicative. Population‐based studies can be well suited to address (some of) these questions given their overall wide range in exposure levels including those at the lower end of the exposure distribution, the possibility to control for potential confounders, and the often larger number of cases. The research described in this thesis was carried out within the framework of a population‐based study: the Netherlands Cohort Study (NLCS), which included 58 279 men and 62 573 women aged 55‐69 years at baseline in September 1986. Within this prospective cohort study, we tried to answer these remaining questions on asbestos carcinogenicity for those occupationally exposed to asbestos. We only studied men as the proportion of long‐term employed women was low in the NLCS. At baseline, participants completed a self‐administered questionnaire on dietary habits and lifestyle, occupational history and other potential risk factors for cancer. For reasons of efficiency in questionnaire processing and follow‐up, the NLCS

137

uses a case‐cohort approach. Incident cases were enumerated from the entire cohort, whereas the accumulated person‐years at risk in the entire cohort were estimated from a random subcohort, selected immediately after baseline. This subcohort is being followed‐up for vital status information while the entire cohort is being monitored for incident cancer by annual record linkage to the Netherlands Cancer Registry and the Dutch Pathology Registry (PALGA). The research described in this thesis was based on 17.3 years of follow‐up of the cohort (1986‐2003). To estimate occupational asbestos exposure levels in the NLCS we employed two commonly used methodologies: case‐by‐case expert assessment and job‐exposure matrices (JEMs). Though neither methodology is perfect, case‐by‐case expert assessment is generally considered the best possible method for assessing occupational exposure in population‐based studies, depending on the accuracy and detail of the information available to the expert. For a random sample of this cohort, case‐by‐case expert assessment, executed in the framework of a previous study in the NLCS, was available for exposure to asbestos, polycyclic aromatic hydrocarbons and welding fumes. As expert assessment was not possible for the cohort as a whole within the framework of this thesis, we had to employ JEMs. This resulted in an additional aim of this thesis which was to evaluate several candidate JEMs in terms of reliability, using the case‐by‐case expert assessment as the ‘gold standard’ for the selection of the most appropriate JEM for studying the asbestos cancer‐related aims in the NLCS. In Chapter 2 of this thesis, the Asbestos JEM and DOMJEM, both developed in the Netherlands, and the Finnish FINJEM were evaluated on reliability in the NLCS by assessing agreement between these JEMs and the available case‐by‐case expert assessment by means of Cohen’s Kappa and the prevalence of exposure. Results showed case‐by‐case expert assessment to result in the lowest prevalence of occupational asbestos exposure in the NLCS. DOMJEM and FINJEM proved to be rather similar in agreement when compared with the expert assessment and with each other. The Asbestos JEM appeared to be less appropriate for use in the NLCS as it showed a lower agreement with the expert assessment and the other two JEMs. This is not completely unexpected as the Asbestos JEM was originally designed to aid subjects with asbestos‐related diseases in pursuing compensation in legal trials. Consequently, this partly disease‐oriented JEM might be more sensitive than DOMJEM and FINJEM, which are more aimed at specificity than sensitivity because of the low exposure levels in the general population. For this reason, the DOMJEM and FINJEM were preferred for subsequent analyses in the NLCS. In Chapter 3, we examined the association between occupational asbestos exposure and the risk of respiratory cancers (i.e., pleural mesothelioma, lung and laryngeal cancer), using both DOMJEM and FINJEM. In addition, we calculated population‐ attributable fractions (PAFs) for these cancers. We were able to confirm the well‐ established associations between asbestos exposure and mesothelioma, lung and laryngeal cancer, revealing elevated risks at the lower end of the exposure distribution

138

Summary

in the population‐based NLCS. When comparing these risks at the lower end of the exposure distribution to those obtained by downward linear extrapolation from highly exposed occupational cohorts, it seemed that they may be lower than expected based on the occupational cohorts. Associations with lung cancer subtypes were generally comparable to overall lung cancer, except for adenocarcinoma, which showed only a weak positive association after prolonged higher asbestos exposure. For laryngeal cancer, associations were usually stronger for supraglottis than glottis cancer, which showed only a positive association after prolonged higher asbestos exposure. There was no statistically significant interaction on an additive or multiplicative scale between asbestos and smoking in causing the respiratory cancers. PAFs for mesothelioma were lower than reported in the literature. As virtually all mesothelioma is due to (occupational) asbestos, higher PAFs might have been expected. PAFs for lung and laryngeal cancer were within the range reported in the literature albeit at the lower end. Results were robust against the use of different JEMs, with DOMJEM showing somewhat higher hazard ratios (HRs) for the highly exposed subjects and FINJEM revealing somewhat higher HRs for ever versus never exposed. In Chapter 4, we studied the association between occupational asbestos exposure and the risk of gastrointestinal cancers (i.e., esophageal, gastric and colorectal cancer), using both DOMJEM and FINJEM. Mainly after (prolonged) exposure to high levels of asbestos, significantly increased HRs for overall gastric cancer, esophageal adenocarcinoma, gastric non‐cardia adenocarcinoma (GNCA), total and distal colon and rectal cancer were observed, but only when using DOMJEM. For FINJEM, HRs were markedly lower for those (prolonged) highly exposed, though the DOMJEM associations were in most instances only marginally significant. Adjustment for smoking may be relevant when studying overall gastric cancer and GNCA, as several HRs were reduced and became non‐significant after adjusting for smoking status. No statistically significant additive or multiplicative interaction between asbestos and smoking was observed for any of the gastrointestinal cancers. In Chapter 5, we looked into the association between occupational asbestos exposure and the risk of tumors of the oral cavity and pharynx, using both DOMJEM and FINJEM. Results showed no convincing evidence of an association between asbestos and the risk of oral cavity cancer (OCC), pharyngeal cancer (PhC), and oral cavity and pharyngeal cancer combined (OCPC) as an exposure‐response relation was lacking, and results were not robust against the use of DOMJEM and FINJEM. However, the potentially increased HRs of PhC and OCPC observed in this and previous studies warrant further research. HRs of PhC and OCPC increased after adjustment for especially alcohol consumption and socioeconomic status. Therefore, taking into account lifestyle factors may be important when studying these cancers. Although the strata‐specific HRs were suggestive of a negative interaction between asbestos and smoking for OCC, none of the cancers showed a significant interaction.

139

This thesis concludes with a summary of the main findings of occupational asbestos research in the population‐based NLCS, in light of some important strengths and limitations ‐ especially when compared to industry‐based studies ‐ and recommendations for future research (Chapter 6). Our study had a number of strengths. First, the NLCS contains a wide range in exposure levels including those at the lower end of the exposure distribution which are nowadays present in most industrialized countries. Results indicated that relative risks at the lower end of the exposure distribution might be higher than predicted by downward linear extrapolation from highly exposed occupational cohorts. As such, population studies as the NLCS can potentially contribute to asbestos risk assessment. Second, we were able to adjust for potential confounders and in several instances we had indications of confounding of the asbestos cancer association by lifestyle factors. This is important as most industry‐based studies have not been able to correct for these factors leading possibly both in upward and downward biased risk estimates. Third, the large study size enabled us to study relatively rare cancers, cancer subtypes, and a possible interaction between asbestos exposure and smoking in relation to cancer risk, with some novel observations. As in most population‐based studies, we had no information on fiber type. Nevertheless, we could dispose of different JEMs in order to provide insight into the methodological uncertainty associated with the choice of JEM. DOMJEM and FINJEM appeared to have their own strengths: FINJEM may be better in discriminating between ever and never‐exposed while DOMJEM appeared better in singling out the prolonged highly exposed subjects. Small differences in the ability of both JEMs to discriminate between never and ever exposed or low and high exposure may lead to slightly different results and judgment of a substance as possibly related with the endpoint of interest. Some caution seems thus appropriate when interpreting results from JEM‐based exposure assessment studies. Incorporation of multiple JEMs in epidemiological studies should be encouraged as it reveals some of the uncertainty associated with the choice of exposure method and therefore aids in the interpretation of data.

140

Valorisation

141

Valorisation

VALORISATION Knowledge valorisation refers to the ‘process of creating value from knowledge, by making knowledge suitable and/or available for social (and/or economic) use, and by making knowledge suitable for translation into competitive products, services, processes and new commercial activities’ (adapted definition based on the National Valorisation Committee 2011:8). The challenge of bridging our results with public awareness, public policy and community action is best met through knowledge transfer and exchange. The opportunities for knowledge valorisation of the different results presented in this thesis are threefold, but before I discuss what merit these results may have on society, I will give some more information on the mineral asbestos, its health consequences, current production and use, and some facts on asbestos exposure in the Netherlands nowadays. Asbestos is a mineral that is characterized by bundles of thin, long, and separable fibers. 1,2 Exposure mainly occurs through inhalation and ingestion.3 Six minerals (or fiber types) are denoted as asbestos and they are divided into two types on the basis of their fiber morphology. Chrysotile, also known as white asbestos, is the sole representative of the serpentine type; so called because of the twisted appearance of its fibers. The second type, the amphiboles, all have straight, needle‐like fibers, and this type accounts for the remaining five minerals: amosite (brown asbestos), crocidolite (blue asbestos), and the more rarely occurring tremolite, anthophyllite, and actinolite. Chrysotile asbestos is the only type of asbestos still produced today, and accounts for more than 95% of all asbestos ever mined.4 Asbestos fibers possess a number of properties useful in commercial applications, such as their ability to withstand heat, and mechanical and chemical damage,1,2 which resulted in a burgeoning of the applications of asbestos in the previous century.5 It was only gradually recognized that using asbestos is not only advantageous but is also associated with the occurrence of serious health consequences. A number of adverse health outcomes is now established as causally associated with exposure to asbestos, regardless of fiber type: asbestosis, mesothelioma (i.e., cancers of the pleura and peritoneum), lung, laryngeal, and ovarian cancer.6,7 From the 1980’s onwards, the growing body of evidence on asbestos’s carcinogenicity led to the banning of amosite and crocidolite in many developed economies, followed by bans on chrysotile throughout the 1990s and the early part of this century. Now, more than 50 countries ban the use of all forms of asbestos, including the entire European Union, and a de‐facto ban exists in many other countries such as the USA and Canada. Almost all of the worldwide use of asbestos is now concentrated in the newly industrializing countries.(4) According to the United States Geological Survey figures, just over two‐thirds of the world’s asbestos use is being accounted for by the rapidly developing BRIC nations: China (30%), India (15%), Russia (12%), and Brazil (9%).

143

Also in the Netherlands, despite being banned in 1994, asbestos is still a public health concern with respect to asbestos removal and site cleaning in the general environment.8 Occupational asbestos exposure can still take place when homes and other buildings are demolished, when soil purification activities are undertaken, and when ships, drilling platforms and other machines with asbestos insulation are repaired or dismantled. In addition, non‐occupational exposure may also occur in the context of building renovations and if asbestos is present in the environment,8 where is was used to harden dirt tracks, yards, and driveways.9 As mentioned above, the opportunities for knowledge valorisation of the different results presented in this thesis are threefold. The first opportunity concerns asbestos exposure in developed economies, such as the Netherlands, more specifically the risk assessment of low asbestos exposure levels nowadays present in these countries. The second opportunity refers to the expanding evidence base for cancers not yet defined as being causally related to asbestos exposure and in connection with this the possibility of legal compensation. The third opportunity concerns asbestos exposure in the newly industrializing countries, more specifically the scientifically‐based call for a worldwide ban on asbestos.

Risk assessment of low asbestos levels

Although the health effects of asbestos are one of the best documented in occupational health, they are also one of the most controversial as there is still considerable debate around asbestos carcinogenicity. In this thesis, we addressed several remaining questions, of which the risk at low asbestos exposure levels may have potential for valorisation. It is important to know what concentration of asbestos fibers over a certain period of time (i.e., cumulative asbestos exposure) may already suffice to induce cancer, in order to set acceptable exposure limits. A brief inspection of some of the more recent meta‐analyses conducted on occupational asbestos exposure and cancer‐related risk reveals that most studies included are not very recent and frequently involved heavily exposed subjects.1,10‐13 Currently, at least in developed countries, most individuals are no longer exposed to these high levels due to increased regulation on the use of asbestos. Due to the uncertainty of extrapolating risks associated with high levels of asbestos exposure to low levels of exposure, the focus has shifted towards estimating excess risk accurately for levels encountered at the lower end of the exposure distribution (i.e. exposure levels in jobs outside the more highly exposed jobs as asbestos mining, insulation, cement and textile manufacturing); a topic of debate particularly for mesothelioma and lung cancer.13‐19 To address this topic, we conducted a prospective analysis on the relation between occupational asbestos exposure and the risk of cancer in a large population of men

144

Valorisation

from the Netherlands aged 55‐69 years (i.e., the Netherlands Cohort Study, abbreviated as NLCS), among whom a small percentage was occupationally exposed to asbestos. Population‐based studies such as the NLCS are, as opposed to the widely used industry‐ based studies on this topic, often the only direct source available on cancer risks associated with the lower end of the exposure distribution. The range in asbestos exposure levels to which the male study subjects in the NLCS were exposed is wide and includes those at the lower end of the exposure distribution. This enabled us to study if cancer risks are not only increased for higher asbestos exposure levels but also for levels encountered at the lower end of the exposure distribution. As most population‐ based studies, the NLCS contains no information on the actual exposure levels. Therefore, we had to estimate occupational asbestos exposure. Results of our study showed increased risks of pleural mesothelioma and lung cancer for the lowest category of cumulative asbestos exposure. More importantly, we observed that relative risks at the lower end of the exposure distribution might be higher than predicted by downward linear extrapolation from highly exposed occupational cohorts, which is comparable to several previous studies. 13,18,19 This may be a reason for regulatory agencies, such as the Health Council, to evaluate once more the risks induced by asbestos exposure at the lower end of the distribution in order to protect the general public and those occupationally involved with asbestos. As such, our study can be of use for assessing if the prevailing standards for asbestos are strict enough.

Legal compensation of asbestos‐related diseases

Another remaining question on asbestos carcinogenicity we tried to answer concerned the possible association between asbestos exposure and the risk of cancers not yet defined as causally related to asbestos exposure. These cancers are the gastrointestinal cancers (i.e., esophageal, gastric and colorectal cancer) and oral cavity and pharyngeal cancer. Our study showed increased risks for all three gastrointestinal cancers, mainly for those subjects that were (prolonged) highly exposed to asbestos. This is in line with previous studies that observed increased risks and dose‐response relations (i.e., the higher the exposure the higher the risk, which may indicate causality between exposure and outcome) mainly in occupational studies with higher asbestos exposure.1,7,20‐22 Therefore, we hypothesized that for gastrointestinal cancers mainly (prolonged) exposure to higher asbestos exposure levels may be associated. For oral cavity cancer, no associations with asbestos were observed in the NLCS, which is in line with previous studies on the asbestos‐related risk of oral cavity cancer.23‐27 This suggests that asbestos may not be associated with an increased risk of oral cavity cancer, though the number of cancer cases in our study was rather low. For pharyngeal cancer, our study reported increased risks for ever versus never exposed to asbestos, but a clear dose‐ response relation could not be demonstrated. However, also for this cancer the number of cases in our study was rather small. Furthermore, previous studies also

145

showed increased risks without evidence of a dose‐response pattern,23,24,28 while a recent large study observed a dose‐response relation.27 Therefore, the rather consistent observation of a possible association between asbestos and pharyngeal cancer could be more than a mere chance finding and warrants further research in studies with a larger number of cases. These results of our study can be used in subsequent evaluations of asbestos carcinogenicity, among others by the International Agency for Research on Cancer (IARC), and contribute to the existing evidence base. For gastric and colorectal cancer, the evidence is now classified as limited, though the working group involved at IARC was evenly divided as to whether the evidence was strong enough to justify classification as sufficient for colorectal cancer.7 For esophageal cancer, no IARC classification existed and we followed the recent meta‐analysis conducted by the Institute of Medicine (IOM) on which the last IARC evaluation (of 2012) draws. IOM considered the evidence to be inadequate.1 IARC classified the evidence for pharyngeal cancer as limited,7 while no separate conclusion was drawn on oral cavity cancer. When the evidence base for these cancers is large and convincing enough to define the association between asbestos exposure and the risk of (some of) these cancers as causal, which may be in the near future for colorectal cancer, then patients with these cancers may qualify for legal compensation. It has to be noted, however, that despite being causally related to asbestos exposure, lung and laryngeal cancer are not being compensated in the Netherlands; only mesothelioma and asbestosis (lung fibrosis, which can also be induced by asbestos exposure) are. The reason for this is that other factors, besides asbestos exposure, may also contribute to cancer risk ‐ most notably smoking (for lung and laryngeal cancer) and alcohol consumption (for laryngeal cancer) ‐ while in the Netherlands, mesothelioma has no other risk factor besides asbestos exposure. That is why we studied asbestos exposure in a population‐based study as the NLCS. The NLCS contains information on dietary habits and lifestyle of the subjects and as such, we were able to correct for a possible confounding influence of, among others, smoking and alcohol. We found that even in the presence of these factors, there is an increased risk of lung and laryngeal cancer that is due to asbestos exposure. Therefore, our study would support legal compensation for lung and laryngeal cancer, though the compensation may be partial due to the fact that it is not possible for an individual case to determine to what extent exposure to asbestos is responsible for the induced cancer en to what extent other risk factors are to blame. Also for the gastrointestinal cancers and oral cavity and pharyngeal cancer, the increased risks we observed were still present after taking other risk factors into account. Nevertheless, for these cancers, causality of a possible association with asbestos still has to be determined before (partial) compensation may be justifiable.

146

Valorisation

A worldwide ban on asbestos

Another question concerns the carcinogenic potential of the different fiber types as there is still discussion on the potency of the different fibers to cause cancer, especially for chrysotile asbestos which is the only fiber type still being produced in some countries. Some scientists ‐ of whom most are (silently) affiliated with the asbestos industry ‐ still argue that chrysotile is far less potent in causing cancer or even harmless as opposed to the amphibole fiber types. They do so regardless of the fact that i) several studies have proven that exposure to pure chrysotile asbestos was solely responsible for the observed increase in cancer incidence or mortality,29‐34 and ii) asbestos is ‐ in part due to these studies ‐ regarded as a human carcinogen by the World Health Organization and the International Agency for Research on Cancer, regardless of fiber type. Unfortunately, the NLCS contains no information on fiber type. However, asbestos use in the Netherlands mainly consisted of chrysotile asbestos (>90%), with the two most widely applied amphibole fiber types being crocidolite and amosite (<10%).8 The latter were primarily used in the insulation industry and in ship building and maintenance. For most branches of industry, chrysotile will have constituted more than 95% of all exposure to asbestos fibers.35 Therefore, though we studied occupational asbestos exposure of a mixed origin, the incidence of cancer in our study is most likely partly due to exposure to chrysotile asbestos. In addition, results of this thesis also point to a possible association between asbestos exposure and overall gastric cancer, esophageal adenocarcinoma, gastric non‐cardia adenocarcinoma, total and distal colon and rectal cancer. If these endpoints also turn out to be causally related to asbestos, the burden of asbestos‐related diseases is even higher than already expected, with an annual forecast of 100,000–140,000 of asbestos‐related cancer deaths in workers worldwide.36 Until a worldwide ban on asbestos is achieved, it is estimated that the asbestos cancer pandemic may take as many as 10 million lives.37 As mentioned before, almost all of the worldwide use of asbestos is now concentrated in the newly industrializing countries, with an estimated 125 million people still being occupationally exposed to asbestos in 2006.38 Predicting the future course of the asbestos epidemic in these countries is severely hampered by the paucity of data on local exposure and disease occurrence. Little is known about the level of asbestos exposure in these countries,39 and the mere fact that asbestos exposure does not seem to be monitored, or at least is not being reported, in many of the countries that are currently producing or using asbestos is worrying. It suggests that these countries do not have the technology and expertise ‐ or political willingness ‐ to cope with this hazardous material.39 The increasing trend in asbestos use and the probably poor control of asbestos exposure will translate into a future rise in the incidence of asbestos‐related diseases in these countries. This may have far‐reaching consequences as health services in these countries are less available than in industrialized countries.40

147

On top of this, other factors are involved, including a poor socio‐economic status,41 which may intensify the asbestos‐related health risks. As even the best workplace control measures cannot prevent exposure to asbestos once in use or when joining the waste stream, I can only strongly support the call for a worldwide ban on all types of asbestos. Safer substitutes for asbestos exist and are feasible for use in newly industrializing countries.42 Nevertheless, asbestos keeps on being used and produced, with the asbestos industry now residing in the newly industrializing countries. How is this possible as we know that all asbestos fiber types are carcinogenic? Is there something that makes asbestos production and use ‘legitimate’? I can come up with no other answer than an economic one. Yet, there is no in‐depth research and understanding of the current and future health and environmental costs of inaction, which will be very high by now. While industry takes the profits, the costs are put on the shoulders of the public taxpayer. One of the main arguments used by groups promoting the continued mining and use of chrysotile asbestos is that since asbestos has been around for many decades and has been studied extensively for its cancer risks, the methods to control its use are well known and hence it can be safely used. The inadequacy of this argument is readily apparent to anyone with any knowledge of the poorly developed regulatory approach to asbestos and other workplace hazards in many of the newly industrializing countries.43 I think the way to a worldwide ban for all asbestos should involve social media, using effective platforms to actively communicate the truth about asbestos exposure not only to those occupationally involved with asbestos but also to the general public that may live in the vicinity of an asbestos mine or factory. In this way, the boycott of the listing of asbestos on Annex III of the Rotterdam Convention ‐ which is intended to ensure that importing governments are able to assess adequately the risks posed by chrysotile asbestos ‐ by countries still using and producing asbestos, is overcome. There have been many international efforts, for many years now, to eliminate the mining of asbestos, the manufacture of asbestos containing products, and the reduction in exposure to existing asbestos materials, both in the workplace and the general community and with success as more and more countries ban asbestos. Even in Canada ‐ for a long time the world’s largest producer of asbestos and an opponent of the listing of asbestos as a dangerous substance under the earlier mentioned Rotterdam Convention ‐ the considerable international pressure and many years of intransigence regarding its continued mining and export of chrysotile asbestos have led to a de‐facto ban. Therefore, it is my hope that citizens all over the world will come into action, among others via social media, and help the many ‘ban asbestos committees and organizations’ in their struggle to consign asbestos‐related diseases to the past. There is still enough to be done.

148

Valorisation

REFERENCES 1. 2.

3. 4. 5. 6.

9. 10. 11. 12.

13.

14.

15. 16. 17. 18.

19.

20. 21. 22.

NAS. Asbestos: Selected Cancers. Washington, DC: The National Academies Press, 2006. Case BW, Abraham JL, Meeker G, Pooley FD, Pinkerton KE. Applying definitions of "asbestos" to environmental and "low‐dose" exposure levels and health effects, particularly malignant mesothelioma. J Toxicol Environ Health B Crit Rev. 2011;14:3‐39. National Toxicology Program. Asbestos. Rep Carcinog. 2011;12:53‐6. Holmes D. IARC in the dock over ties with asbestos industry. Lancet. 2013;381(9864):359‐61. Virta RL. Worldwide Asbestos Supply and Consumption Trends From 1900 Through 2003. U.S. Geological Survey Circular 1298, 2006. Straif K, Benbrahim‐Tallaa L, Baan R, Grosse Y, Secretan B, El Ghissassi F, Bouvard V, Guha N, Freeman C, Galichet L, Cogliano V; WHO International Agency for Research on Cancer Monograph Working Group. A review of human carcinogens‐‐part C: metals, arsenic, dusts, and fibres. Lancet Oncol. 2009;10:453‐4. IARC. Asbestos (Chrysotile, Amosite, Crocidolite, Tremolite, Actinolite and Anthophyllite). In: A Review of Human Carcinogens: Arsenic, Metals, Fibres, and Dusts. Lyon, France: International Agency for Research on Cancer (IARC), 2012. 219‐309. Gezondheidsraad. Asbestos: risks of environmental and occupational exposure. The Hague, the Netherlands: Health Council of the Netherlands, report 2010/10E. http://www.gezondheidsraad.nl/ sites/default/files/201010E.pdf. Accessed May 1, 2015. Driece HA, Siesling S, Swuste PH, Burdorf A. Assessment of cancer risks due to environmental exposure to asbestos. J Expo Sci Environ Epidemiol. 2010;20:478‐85. Berman DW, Crump KS. A meta‐analysis of asbestos‐related cancer risk that addresses fiber size and mineral type. Crit Rev Toxicol. 2008;38 Suppl 1:49‐73. Hodgson JT, Darnton A. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure. Ann Occup Hyg. 2000;44:565‐601. Lenters V, Vermeulen R, Dogger S, Stayner L, Portengen L, Burdorf A, Heederik D. A meta‐analysis of asbestos and lung cancer: is better quality exposure assessment associated with steeper slopes of the exposure‐response relationships? Environ Health Perspect. 2011;119:1547‐55. van der Bij S, Koffijberg H, Lenters V, Portengen L, Moons KG, Heederik D, Vermeulen RC. Lung cancer risk at low cumulative asbestos exposure: meta‐regression of the exposure‐response relationship. Cancer Causes Control. 2013;24:1‐12. Iwatsubo Y, Pairon JC, Boutin C, Ménard O, Massin N, Caillaud D, Orlowski E, Galateau‐Salle F, Bignon J, Brochard P. Pleural mesothelioma: dose‐response relation at low levels of asbestos exposure in a French population‐based case‐control study. Am J Epidemiol. 1998;148:133‐42. Hillerdal G. Mesothelioma: cases associated with non‐occupational and low dose exposures. Occup Environ Med. 1999;56:505‐13. Ilgren EB, Browne K. Asbestos‐related mesothelioma: evidence for a threshold in animals and humans. Regul Toxicol Pharmacol. 1991;13:116‐32. Gibbs GW, Berry G. Mesothelioma and asbestos. Regul Toxicol Pharmacol. 2008;52:S223‐31. Gustavsson P, Nyberg F, Pershagen G, Schéele P, Jakobsson R, Plato N. Low‐dose exposure to asbestos and lung cancer: dose‐response relations and interaction with smoking in a population‐based case‐ referent study in Stockholm, Sweden. Am J Epidemiol. 2002;155:1016‐22. De Matteis S, Consonni D, Lubin JH, Tucker M, Peters S, Vermeulen RCh, Kromhout H, Bertazzi PA, Caporaso NE, Pesatori AC, Wacholder S, Landi MT. Impact of occupational carcinogens on lung cancer risk in a general population. Int J Epidemiol. 2012;41:711‐21. Gamble J. Risk of gastrointestinal cancers from inhalation and ingestion of asbestos. Regul Toxicol Pharmacol. 2008;52:S124‐53. Frumkin H, Berlin J. Asbestos exposure and gastrointestinal malignancy review and meta‐analysis. Am J Ind Med. 1988;14:79‐95. Morgan RW, Foliart DE, Wong O. Asbestos and gastrointestinal cancer. A review of the literature. West J Med. 1985;143:60‐5.

149

23. Paget‐Bailly S, Cyr D, Luce D. Occupational exposures to asbestos, polycyclic aromatic hydrocarbons and solvents, and cancers of the oral cavity and pharynx: a quantitative literature review. Int Arch Occup Environ Health. 2012;85:341‐51. 24. Reid A, Ambrosini G, de Klerk N, Fritschi L, Musk B. Aerodigestive and gastrointestinal tract cancers and exposure to crocidolite (blue asbestos): incidence and mortality among former crocidolite workers. International journal of cancer Journal international du cancer. 2004;111:757‐61. 25. Purdue MP, Jarvholm B, Bergdahl IA, Hayes RB, Baris D. Occupational exposures and head and neck cancers among Swedish construction workers. Scandinavian journal of work, environment & health. 2006;32(4):270‐5. 26. Gustavsson P, Jakobsson R, Johansson H, Lewin F, Norell S, Rutkvist LE. Occupational exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus: a case‐control study in Sweden. Occup Environ Med. 1998;55:393‐400. 27. Langevin SM, O'Sullivan MH, Valerio JL, Pawlita M, Applebaum KM, Eliot M, McClean MD, Kelsey KT. Occupational asbestos exposure is associated with pharyngeal squamous cell carcinoma in men from the greater Boston area. Occup Environ Med. 2013;70:858‐63. 28. Marchand JL, Luce D, Leclerc A, Goldberg P, Orlowski E, Bugel I, Brugère J. Laryngeal and hypopharyngeal cancer and occupational exposure to asbestos and man‐made vitreous fibers: results of a case‐control study. Am J Ind Med. 2000;37: 581‐9. 29. Mirabelli D, Calisti R, Barone‐Adesi F, Fornero E, Merletti F, Magnani C. Excess of mesotheliomas after exposure to chrysotile in Balangero, Italy. Occup Environ Med. 2008;65:815‐9. 30. Magnani C, Ferrante D, Barone‐Adesi F, Bertolotti M, Todesco A, Mirabelli D, Terracini B. Cancer risk after cessation of asbestos exposure: a cohort study of Italian asbestos cement workers. Occup Environ Med. 2008;65:164‐70. 31. Wang XR, Yu IT, Qiu H, Wang MZ, Lan YJ, Tse LY, Yano E, Christiani DC. Cancer mortality among Chinese chrysotile asbestos textile workers. Lung Cancer. 2012;75:151‐5. 32. Wang X, Lin S, Yu I, Qiu H, Lan Y, Yano E. Cause‐specific mortality in a Chinese chrysotile textile worker cohort. Cancer Sci. 2013;104:245‐9. 33. Wang X, Yano E, Lin S, Yu IT, Lan Y, Tse LA, Qiu H, Christiani DC. Cancer mortality in Chinese chrysotile asbestos miners: exposure‐response relationships. PLoS One. 2013;8:e71899. 34. Wang X, Lin S, Yano E, Yu IT, Courtice M, Lan Y, Christiani DC. Exposure‐specific lung cancer risks in Chinese chrysotile textile workers and mining workers. Lung Cancer. 2014;85:119‐24. 35. Burdorf L, Swuste PH, Heederik D. A history of awareness of asbestos disease and the control of occupational asbestos exposures in The Netherlands. Am J Ind Med. 1991;20:547‐55. 36. Ramazzini C. Asbestos is still with us: repeat call for a universal ban. J Occup Environ Med. 2010;52: 469‐72. 37. LaDou J. The asbestos cancer epidemic. Environ Health Perspect. 2004;112:285‐90 38. WHO. Elimination of asbestos‐related diseases. http://whqlibdoc.who.int/hq/2006/WHO_SDE_ OEH_06.03_eng.pdf. Accessed May 1, 2015. 39. Stayner L, Welch LS, Lemen R. The worldwide pandemic of asbestos‐related diseases. Annu Rev Public Health. 2013;34:205‐16. 40. Levy BS. Global occupational health issues: Working in partnership to prevent illness and injury. AAOHN J. 1996;44:244‐7; discussion 247. 41. Kromhout H. Occupational hygiene in developing countries: something to talk about? Ann Occup Hyg. 1999;43:501‐3. 42. World Bank Group (WBG). Good Practice Note: Asbestos: Occupational and Community Health Issues. http://www.bwint.org/pdfs/WB‐AsbestosGuidanceNote.pdf. Accessed May 1, 2015. 43. Sim MR. A worldwide ban on asbestos production and use: some recent progress, but more still to be done. Occup Environ Med. 2013;70:1‐2.

150

Dankwoord

151

Dankwoord

DANKWOORD Eindelijk is het dan zo ver, na nog ‘even’ de valorisatieparagraaf, samenvatting en een paar andere zaken ‘gepiept’ te hebben, is het boekje nu echt bijna af! Alleen mijn dankwoord ontbreekt nog. Aangezien ook ik mijn promotie niet in mijn eentje had kunnen doen, zou ik graag de volgende mensen willen bedanken voor hun bijdrage, steun of ‘gewoon’ hun vriendschap de afgelopen jaren. (co)Promotoren: Om te beginnen wil ik natuurlijk Piet bedanken, niet alleen voor de kans die je me geboden hebt om deze promotie te kunnen doen, maar ook voor alle waardevolle inzichten die je me gegeven hebt. Hoewel ik niet altijd stond te springen om weer een van die ‘vervelende’ tijdsplanningen te moeten maken: ze hebben me wel heel erg geholpen om een reële inschatting van de benodigde tijd te leren maken, i.p.v. een ietwat optimistische. Als iemand de tijdslijnen en het overzicht goed in de gaten weet te houden, dan ben jij dat wel! Sandra, ik snap wel dat jij en Piet ooit samen begonnen zijn aan de NLCS, jullie vullen elkaar namelijk uitstekend aan. Hoe fijn was het om naast Piet’s visie ook jouw commentaar op mijn stukken te mogen ontvangen, ondanks de soms beperkte tijd doordat je eigenlijk twee banen naast elkaar had. Heel knap dat je dat zo lang hebt weten vol te houden. Ik hoop voor de huidige/ toekomstige NLCS’ers dat zij daar ook nog eens gebruik van mogen maken, want je bent zo heerlijk direct, maar altijd op een prettige manier. Bedankt daarvoor! En dan Roel, last maar zeker niet least, jij was voor mij de perfecte aanvulling op het team. Ook al leek de afstand tussen Maastricht en Utrecht soms groot, jouw verfrissende blik op het geheel maakt dat ik onze discussies beslist niet had willen missen. Jouw kennis van het werkveld is groot en ik ben blij dat ik daarin heb mogen delen. Co‐auteurs: Veel co‐auteurs hebben een bijdrage geleverd aan de artikelen in dit proefschrift. Ik wil iedereen dan ook hartelijk danken voor het kritisch becommentariëren van mijn stukken. Leden van de beoordeligscommissie: Ik wil alle leden graag bedanken voor het kritisch lezen en beoordelen van mijn proefschrift. Timo Kauppinen: I am grateful that you offered us the opportunity to use FINJEM in the NLCS, thank you for this! Tom: Wat een pret die beroepenhercodering hè?! Nee zonder gein, ik denk dat we er allebei veel van hebben opgestoken en die uurtjes bij de DE waren beslist geen straf . Bedankt voor de prettige samenwerking!

153

Collega’s Epidemiologie: Mede dankzij jullie ben ik altijd met veel plezier naar het DEB gekomen. In het bijzonder wil ik graag Yvonne, Mariëlle, Petra, Jolanda, Sacha, Conny, Jos en Harry bedanken. Grote of kleine vragen, slimme of minder slimme (computervragen)...ik kon altijd bij jullie terecht, dank daarvoor! Tevens wil ik Liesbeth Preller bedanken voor het waardevolle voorwerk omtrent het gebruik van blootstellingsmatrices in de NLCS. Jong (of ietwat minder jong) EPID: Wat hebben we een leuke tijd gehad en nog steeds! Of het komt door onze beruchte PhD‐tour of gewoon, omdat we een ondernemend clubje hebben: er is een klik en dat heb je niet altijd bij zo’n grote groep. Ik denk dan ook met veel plezier terug aan de vele etentjes, borrels en de o zo hilarische Sinterklaasavondjes . Ook nu nog zie ik meerdere van jullie met regelmaat en ik hoop dat dat zo mag blijven! Ow, en Ivette: zodra we allebei onze verdediging erop hebben zitten, stel ik voor dat we het rennen/fietsen gauw weer oppakken! Dan hebben we allebei geen goed excuus meer ;). Anne en Milan: Mijn twee kamergenootjes en wat voor een! Zonder jullie was mijn tijd op het DEB beslist minder leuk geweest. Ik weet nog toen ik begon dat ik alleen op die grote kamer zat...wat was ik blij toen jij de boel kwam versterken, Anne! En Milan, we hadden geen betere ‘derde’ kamergenoot kunnen hebben dan jij: je hield de boel in evenwicht zullen we maar zeggen . Ik vind het super dat jullie nu allebei het buitenlandavontuur zijn aangegaan en hoewel ik vermoed dat qua werk mijn wegen die van jullie niet meer zo snel zullen kruisen, hoop ik dat we ooit nog eens, als jij Milan een van de proffen bent op het DEB (ik bied 100 euro ;)), voor één dagje een kamer kunnen delen, hoe grappig zou dat zijn?! Mijn paranimfen: Chantalle, hoe lang kennen we elkaar inmiddels al niet en wat hebben we al niet samen meegemaakt...leuke, maar ook minder leuke dingen. Wat ben ik trots op hoe jij in het leven staat, ondanks alles wat je overkomen is. Dat is echt wel een heel dikke pluim waard! Jouw vriendschap had ik voor geen goud willen missen, is blijven zitten in de 5e toch ergens goed voor geweest ;). Jij weet als geen ander hoe je iemand aan het lachen krijgt, als jij iets vertelt dan zie ik het gewoon al gebeuren. Ook nu zit ik met een smile op mijn gezicht dit tekstje over jou te typen. Mocht je ooit toch iets anders willen dan het CTCM (geen zorgen, Marijke, dat zie ik nog niet gebeuren): stand‐up comedian zou je beslist liggen . Bedankt dat jij op die grote, spannende dag achter me wilt staan! Colinda, ik voel me vereerd dat je ondanks de thuissituatie toch mijn paranimf wilt zijn. Er zijn misschien mensen die denken dat jij heel rustig bent, maar als je eenmaal begint te kletsen, dan overtref je zelfs mij ;). Bedankt voor alle fijne gesprekken op, maar ook naast het werk. Ik hoop dat er nog vele mogen volgen!

154

Dankwoord

Mijn nieuwe collega’s van het CTCM: Ik werk nog niet zo heel lang bij jullie, maar desondanks wil ik jullie graag bedanken voor de fijne werksfeer die er bij het CTCM heerst. Ik beloof bij dezen ook plechtig dat ik vanaf nu wat vaker mee zal gaan lunchen . Vrienden en familie: Wat moest ik zonder jullie?! Tjandra, wat hebben we al een hoop (buitenland)avonturen samen mogen beleven en wat hebben we heerlijk weten te speculeren over wat we in de toekomst wel niet allemaal zouden gaan doen. Ondanks dat we allebei onze idealen voor ons gevoel nog niet helemaal waar hebben kunnen maken, ben ik ervan overtuigd dat dat nog wel gaat komen, zij het misschien in een iets minder grote vorm dan in eerste instantie gedacht ;). Ik hoop dat we nog lang vriendinnen mogen blijven en van elkaar mee mogen maken wat de toekomst in petto heeft! ‘Uni friends’, wat een grap dat ik de meiden van dit gezelschap pas in het tweede jaar van mijn studie heb leren kennen, ondanks dat we vast vaak genoeg in dezelfde trein gezeten hebben! Die schade hebben we daarna wel ingehaald en ik zeg dan ook dankjewel voor alle gezellige etentjes en weekendjes weg! En Lian: zodra mijn verdediging erop zit, stel ik voor dat we ons meteen inschrijven bij Atletiek Maastricht, want ons kent ons . Linda en Nadina, ondanks dat we niet meer bepaald bij elkaar in de buurt wonen (kleine understatement ;)), ben ik heel blij dat we toch altijd contact hebben weten te houden. Nadina, het lijkt me een goed plan als wij, de twee Nadines, ’t Linda gauw eens gaan opzoeken zodra zij, Massimo, Anna en ... en ... gesetteld zijn in Belgrado! Mirella, ondanks dat we elkaar veel minder vaak zien dan we zouden willen, voelt het meteen weer vertrouwd als ik jou aan de lijn heb. Hopelijk lukt het ons in de toekomst wat vaker om af te spreken en kunnen we ooit samen het Zillertal nog eens onveilig maken! Manon, ik hoor het je al denken ‘waarom moet je mij nou voor dit boekje bedanken?’..., toch wilde ik je even laten weten dat het ravotten met Tristan en Merlin heel verhelderend werkt . Jij en Dion hebben echt twee superleuke kids en ik hoop in de toekomst weer wat vaker bij jullie langs te komen! Lieve papa en mama, jullie zijn zonder twijfel de liefste ouders die je je maar wensen kunt! Hoewel ik lang getwijfeld heb over wat ik wilde worden, hebben jullie altijd in mij geloofd en mij de kans gegeven om mijn eigen weg te vinden. Bedankt daarvoor! Lieve Sander en Ellen, ondanks dat ik tegenwoordig broer en zus moet zeggen, hou ik me daar deze keer niet aan op ;). Broertje van me, jouw kennis en kunde is zo veel groter dan je zelf beseft. Je hoeft je grote zus allang niet meer om advies te vragen, jij bent nu de expert! Ik hoop dat jij en Jill volgend jaar een onvergetelijke trouwdag mogen hebben. Zusje van me, dankjewel dat jij mijn kaft hebt willen maken, heel speciaal en wat is ie mooi geworden!  Wat ben ik trots op waar jij nu staat. Ik weet zeker dat de

155

toekomst nog heel veel moois in petto heeft voor jou, samen met d’r Guyon! En dan Ria, mijn lieve tante, ook op jou ben ik trots, want wie gaat op haar 65e nog op voor het rijbewijs? Ik weet zeker dat het je gaat lukken! Tot slot, Huppie, je mag dit helaas niet meer meemaken, maar als laatste eer draag ik dit boekje op aan jou!

156

Curriculum Vitae

157

Curriculum Vitae

CURRICULUM VITAE Nadine Offermans was born on March 29th, 1979 in Heerlen, the Netherlands. After completing secondary school at ‘Gymnasium Rolduc’ in Kerkrade in 1998, she started with Psychology at Maastricht University, in Maastricht, the Netherlands. After one year she decided that this study did not fit her all that well and that it would be wise to get some working experience first in order to over think what study would fit her most. In 2001, she started with Health Sciences, again at Maastricht University, with a specialisation in Movement Sciences. For a master’s thesis, she performed a 5‐month internship at sports physiotherapy practice Fysiovision, in Geleen, the Netherlands, during which she studied the influence of overreaching on the heart rate variability using the Co2ntrol. She graduated in 2007. In December 2008, she started as a research assistant in a study on the quality of care for patients with Parkinson’s disease at the UMC St. Radboud, in Nijmegen, the Netherlands. This is where the love for research and her motivation to have her own PhD project grew. In November 2009, Nadine began her PhD at the Department of Epidemiology, under supervision of Prof. dr. Piet van den Brandt, dr. Sandra Bausch, and dr. Roel Vermeulen. Her research, as described in this thesis, was performed within the framework of the Netherlands Cohort Study (NLCS). During her PhD, she co‐ organized an eight‐day scientific study tour for PhD students from the Department of Epidemiology to the United Kingdom in 2011. In March 2015, Nadine started with a new challenge and she is now working as a risk counsellor/auditor at the Clinical Trial Center Maastricht (CTCM) in Maastricht, the Netherlands.

159

List of publications

161

List of publications

LIST OF PUBLICATIONS Offermans NS, Vermeulen R, Burdorf A, Goldbohm RA, Keszei AP, Peters S, Kauppinen T, Kant I, Kromhout H, van den Brandt PA. Occupational asbestos exposure and risk of oral cavity and pharyngeal cancer in the prospective Netherlands Cohort Study. Scand J Work Environ Health 2014;40:420‐7. Offermans NS, Vermeulen R, Burdorf A, Goldbohm RA, Keszei AP, Peters S, Kauppinen T, Kant I, Kromhout H, van den Brandt PA. Occupational asbestos exposure and risk of esophageal, gastric and colorectal cancer in the prospective Netherlands Cohort Study. Int J Cancer 2014;135:1970‐7. Offermans NS, Vermeulen R, Burdorf A, Goldbohm RA, Kauppinen T, Kromhout H, van den Brandt PA. Occupational asbestos exposure and risk of pleural mesothelioma, lung and laryngeal cancer in the prospective Netherlands Cohort Study. J Occup Environ Med 2014;56:6‐19. Offermans NS, Vermeulen R, Burdorf A, Peters S, Goldbohm RA, Koeman T, van Tongeren M, Kauppinen T, Kant I, Kromhout H, van den Brandt PA. Comparing JEMs in population‐based studies: what if expert assessment and measurements are not available? Occup Environ Med 2013;70:519. Koeman T, Offermans NS, Christopher‐de Vries Y, Slottje P, van den Brandt PA, Goldbohm RA, Kromhout H, Vermeulen R. JEMs and Incompatible Occupational Coding Systems: Effect of Manual and Automatic Recoding of Job Codes on Exposure Assignment. Ann Occup Hyg 2013;57:107‐14. Offermans NS, Vermeulen R, Burdorf A, Peters S, Goldbohm RA, Koeman T, van Tongeren M, Kauppinen T, Kant I, Kromhout H, van den Brandt PA. Comparison of expert and job‐exposure matrix‐based retrospective exposure assessment of occupational carcinogens in the Netherlands Cohort Study. Occup Environ Med 2012;69:745‐51.

163