COPDexacerbationem by Asociación Latinoamericana de Tórax ALAT

Chronic obstructive pulmonary disease exacerbations in emergency departments: predictors of outcome Nicolas Rochea, Antoine Rabbata, Mahmoud Zureikb and Ge´rard Huchona a

Respiratory and Intensive Care Medicine, Hoˆtel-Dieu, Paris Descartes University and bINSERM U700, Faculte´ de Me´decine Xavier Bichat, Paris, France Correspondence to Nicolas Roche, MD, PhD, Service de Pneumologie et Re´animation, Hoˆpital de l’Hoˆtel Dieu, 1, place du Parvis de Notre-Dame, 75004 Paris, France Tel: +33 1 42 34 84 82; fax: +33 1 42 34 84 48; e-mail: nicolas.roche@htd.aphp.fr Current Opinion in Pulmonary Medicine 2010, 16:112–117

Purpose of review Exacerbations have a major short-term and long-term impact on patients with chronic obstructive pulmonary disease (COPD). Risk stratification of patients presenting to the emergency department for an exacerbation of COPD is of utmost importance to help in deciding patients’ orientation and treatment and in improving outcomes. Recent findings Studies on predictors of outcomes of COPD exacerbations are markedly heterogeneous in terms of assessed variables, outcomes of interest and timeframe (short-term vs. long-term outcomes). Age, severity of underlying disease, clinical signs of immediate severity, and comorbidities are among the most frequently identified prognostic factors of in-hospital outcome. In the most severe patients, ICU scores such as the Acute Physiology and Chronic Health Evaluation II (APACHE II) also have a prognostic value. Several biomarkers have also been studied, but their additional value for risk stratification is not clarified. Summary Scores predicting the risk of poor outcome could prove useful in the management of COPD exacerbations. Some have been suggested but remain to be further validated before their use can be generalized. Keywords chronic obstructive pulmonary disease, emergency department, exacerbations, prediction, prognosis Curr Opin Pulm Med 16:112–117 ß 2010 Wolters Kluwer Health | Lippincott Williams & Wilkins 1070-5287

Introduction Exacerbations play a major role in the burden of chronic obstructive pulmonary disease (COPD). They induce significant mortality and morbidity as shown, for example, by the prolonged decrease in daily activity and the marked impairment in health-related quality of life which they induce [1,2]. They contribute to 25–50% of COPD-related healthcare expenses, mainly through hospitalizations [3,4]. From a long-term perspective, repeated exacerbations of COPD are associated with a more pejorative course of the disease, characterized by an accelerated decline in FEV1 (forced expiratory volume in 1 second) and health status and an increased mortality [5,6]. Studies on exacerbations are complicated by difficulties in agreeing on a consensual operational definition [7] and by their under-recognition by patients themselves [2], leading to undertreatment or delayed treatment, which are deleterious in terms of both short-term and long-term prognosis [8,9]. Several studies showed that home-based care for exacerbations can represent a valuable alternative to hospital care, allowing avoiding or shortening hospital stays [10]. 1070-5287 ß 2010 Wolters Kluwer Health | Lippincott Williams & Wilkins

Decreasing the impact of COPD exacerbations on individuals requires implementing early identification and treatment in the appropriate setting. Patients’ orientation relies on several parameters: the severity of the underlying disease, the severity of the acute episode and the social and familial context. Appropriate orientation is one of the major tasks in emergency departments (EDs); in that context, knowledge of prognostic factors is of major importance to determine the most cost-effective setting for patient care. Several short-term outcomes of COPD exacerbations can be considered when assessing the prognostic value of a given parameter: mortality is certainly the most robust, although it varies markedly between studies depending on the characteristics of the population: patients requiring ICU-hospitalization are obviously at greater risk than others. The need for noninvasive mechanical ventilation (NIV), length of hospital stay, requirement of posthospital care and short-term recurrence can also be considered as important outcomes. Patient-reported outcomes include symptoms (not only dyspnea but also cough and sputum production or fatigue), health status DOI:10.1097/MCP.0b013e328335f039

COPD exacerbations in emergency departments Roche et al. 113

and daily activity. Functional outcomes are dominated by markers of gas exchange and airflow obstruction (FEV1) but could also include indicators of exercise tolerance, such as the 6-min walking distance. Similarly, there is a great variety of potential predictors of outcomes, including clinical characteristics, lung function data, general biology parameters such as those used in ICU scores and newer biomarkers, which are increasingly studied. Administrative variables regarding organization of care or patient situations can also be of interest. Altogether, studies are highly heterogeneous in terms of measured variables, which make it difficult to determine which prognostic factors are the most important.

In a French prospective cohort study of 794 patients recruited in 104 EDs, only 10.5% were discharged home and the in-hospital mortality rate in the remaining was 7.4% [12 ]. A number of patients’ characteristics were associated with the risk of death on univariate analysis (Table 1). Surprisingly, some classical risk factors such as heart and respiratory rates at entry or body mass index (BMI) were not significantly different between patients who died and those who survived. This may relate to the lack of sufficient heterogeneity of patients regarding these variables. Independent predictors of death were age, clinical signs of severity (especially neurological impairment and use of inspiratory accessory muscles) and baseline dyspnea grade at steady state.

Many studies investigated factors independently associated with outcomes of patients with COPD exacerbations. However, only a few were performed in EDs: in many cases, included patients are recruited while they are already hospitalized in a medical ward or an ICU. In addition, several studies focused on long-term outcomes rather than in-hospital and short-term outcomes, which are the most important factors for the clinician who has to decide where a patient should be cared for. This review will focus only on predictors of robust in-hospital outcomes, that is, the need for mechanical ventilation, mortality, length of stay (LOS) and requirement for posthospital support at discharge. Clinical factors will be addressed before recently studied biomarkers.

Another study in 825 hospitalized patients also identified age and respiratory failure as predictors of in-hospital death; the number of prior admissions for COPD exacerbations also exhibited an independent prognostic value, as did several comorbidities including congestive heart failure, cerebrovascular disease and chronic liver disease [13 ]. In that study, the identification of these comorbid illnesses as prognostic factors was certainly permitted by their relatively high frequency in the studied population (37%). This also explains that another factor was found to have a protective effect, that is, the use of b-blockers. This is in accordance with studies showing that these treatments are beneficial in terms of mortality in patients with COPD and ischemic heart disease [14].

Clinical predictors of outcome Outcomes of interest addressed in the following section are mortality, need for invasive mechanical ventilation, LOS and need for posthospital support. Predictors of mortality

Some clinical factors have a prognostic value and can be used to build prognostic scores, which could be of interest for the practitioner in the ED. Administrative data are more likely to be of help for those who organize the care for COPD exacerbations. Clinical factors

Clinical characteristics of patients are easy to collect rapidly on arrival and have an important prognostic value. For instance, in a retrospective study of 972 patients hospitalized through an ED, a 6.4% in-hospital mortality was observed and independent prognostic factors were the number of Anthonisen criteria (increased dyspnea, sputum volume or purulence), age, severity of underlying COPD, blood gas values and the occurrence of in-hospital complications [11]. The strongest association was found with the number of Anthonisen criteria, with an odds ratio of 7.3 [95% confidence interval (95% CI): 3.6–17.7] for two or three vs. one criteria.

One may find it ‘disappointing’ that persistent smoking was found to have some prognostic value in only a minority of studies [15 ]. However, this may be related to a ‘healthy survivor effect’: persistent smokers who are included in studies are obviously those who already Table 1 factors associated with the risk of death and need for posthospital support in a study of 794 patients recruited in emergency departments: results of univariate analysis [12 ]

Increasing age Referral by a doctor Previous home support or institutionalization Low body mass index High dyspnea grade at steady state Cyanosis Neurological impairment Lower limb edema Asterixis Use of inspiratory accessory muscles Expiratory use of abdominal muscles Clinical severity index High respiratory rate High heart rate Hospitalization during the past 6 months Maintenance oral corticosteroids Long-term oxygen therapy

Death

Need for posthospital support

Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes No No Yes No Yes

Yes Yes Yes No Yes Yes Yes Yes No Yes Yes Yes Yes No Yes No Yes

Yes: significant association. No: nonsignificant.

114 Obstructive, occupational and environmental diseases

managed to survive despite continuing to smoke, meaning that they may be more ‘resistant’ to tobacco smoking. Age as a prognostic factor: is mortality a marker of quality of care in the elderly?

The prognostic value of age seems quite logical, but what may be more intriguing is that it persists after adjustment for performance status and several markers of organ failure (blood urea, serum albumin, arterial pH and oxygen saturation). One hypothesis to explain this observation is that older patients may be less likely to receive appropriate treatment: for instance, in the national UK audit patients aged more than 75 years were indeed less likely to be seen by a respiratory nurse or physician during admission, and have arterial blood gases documented and FEV1 measured before or during the 90 days following the exacerbation [16]. In addition, patients admitted under a care of the elderly physician were less likely to have ventilatory support when pH value was below 7.35 [16]. Scoring the risk of mortality

In the French study mentioned in the previous section, a prediction rule shown in Table 2 was derived from factors independently predicting mortality in a randomly selected sample of the population; it was found to have interesting prediction characteristics in the randomly selected validation subcohort of the population (c-statistic in the validation cohort: 0.83; sensitivity and specificity: 74 and 68%, respectively) [12 ]. The external validity of this score will be further studied using data from another prospective cohort of 1817 patients hospitalized in pulmonology departments of 68 French general hospitals [17]. This study already allowed to evaluate inhospital mortality at 2.5%, independent prognostic factors being again age, dyspnea grade at steady state and the number of clinical signs of severity. An additional factor was the presence of a previously known cardiovascular comorbidity. Administrative data

Administrative data may also be of some value to help assessing the risk of in-hospital death: in a cross-sectional study using a database of non-federal US hospitals, hospitalizations for COPD exacerbations were associated with a 2.5% in-hospital mortality [18]. Risk factors for death included older age, male sex, comorbid illnesses,

nonroutine admission source and, surprisingly, higher income. A possible explanation is that patients with higher income can receive appropriate care in an ambulatory setting more easily, so that they are admitted to hospitals only when their condition becomes severe. Although this analysis suffers from the lack of medical data, it can contribute to the identification of factors useful in a ‘triage’ perspective. Predictors of the need for invasive mechanical ventilation

In patients with severe COPD exacerbations, the need for invasive mechanical ventilation is a major predictor of poor in-hospital outcome [19]. Before NIV became widely used, requirement for intubation in ICU patients could be predicted by high APACHE II score, low Glasgow Coma Score (GCS) and poor nutritional status [20]. Then, NIV was shown to improve dramatically the prognosis of COPD exacerbations with acute respiratory failure (ARF), especially in patients with respiratory acidosis and pH more than 7.35 [21]: in that context, NIV definitely reduces mortality and LOS through a decrease in the need for intubation and thereby in the risk of ventilator-acquired pneumonia. Nowadays, in patients meeting criteria for NIV, factors associated with failure and subsequent need for endotracheal intubation and invasive ventilation are of interest as predictors of prognosis. They included the value of pH, GCS and APACHE II score in a study of 122 COPD patients treated by NIV for ARF [22]. Phua et al. [23] found a similar predictive value of the APACHE II score in their 43 patients with ARF due to COPD exacerbation. Length of stay and need for posthospital support

In most studies, mean or median length of hospital stay varies between 6 and 11 days. Tsimogianni et al. [24] studied retrospectively data collected prospectively in 81 patients admitted for COPD exacerbations and found that Medical Research Council (MRC) dyspnea grade was the only independent predictor of LOS, although BMI and respiratory rate were also associated with this outcome in univariate analysis. Indeed, BMI and MRC dyspnea both contributed to a model designed to predict 3-year mortality. In the UK first national audit on COPD exacerbations performed in 1997 (n ¼ 1400 admissions in 38 acute hospitals), predictors of LOS were age more than 65 years, poor performance status and low FEV1 at steady

Table 2 Score predicting the risk of death according to a study in 794 patients recruited in emergency departments with a 7.4% overall in-hospital mortality rate [12 ]

Age (years) Clinical signs of severitya Baseline dyspnea MRC grade

0 point

1 point

<70 None 0–1

70 2–3

2 points

3 points

1–2 4–5

Mortality was 0 for a score of 0 or 1, 5% for a score of 2 or 3, 12% for a score > 4 points (maximum score: 6). MRC, Medical Research Council. Among signs significantly associated with the risk of death in univariate analysis (see Table 1).

COPD exacerbations in emergency departments Roche et al. 115

state [25]. In the subsequent 2003 audit (n ¼ 7529 from 234 acute hospitals), hospital resources and organization were also assessed and contributed to predict LOS, which was reduced in units with more respiratory consultants, an early discharge scheme, local COPD guidelines and higher organization scores [26]. Clinical factors associated with LOS were age, performance status, arterial oxygen saturation, respiratory rate and serum albumin. In the study by Stolz et al. [27] on biomarkers detailed in the following section, cardiopathy was the only clinical parameter independently associated with length of hospital stay. No significant association was observed for age, FEV1, PaO2, PaCO2, renal failure, diabetes mellitus and arterial hypertension. In the above-mentioned French study of patients recruited in EDs, 27% of patients required posthospital support (at home or in convalescence or nursing homes). Associated factors were identical to those that predicted the risk of death (age, number of clinical signs of severity, dyspnea grade at steady state) plus female sex, which may be related to the longer life expectancy of women [12 ]. Thus, determinants of prognosis are also associated with dependency.

Biomarkers Biomarkers are increasingly studied in many areas of medicine. Some could prove useful for detection of an individual susceptibility to the deleterious effects of cigarette smoke or other environmental factors on the airways; others could help assessing the prognosis and predict the upcoming course of the disease, or anticipate the response to treatments. All these possible applications are related to patients’ characterization, which is becoming particularly important in COPD: although this disease now has a unique definition, it is also characterized by a multidimensional impact, explaining why patients are markedly heterogeneous in terms of phenotypic characteristics (e.g. emphysema vs. chronic bronchitis etc.). In addition, the use of biomarkers may help fill some persisting gaps in the knowledge of the pathophysiological mechanisms involved in COPD and their relationship with clinical phenotypes. In stable COPD, several studies assessed the prognostic value of markers of systemic inflammation such as Creactive protein (CRP). Others focused on markers of the respiratory component of the disease, such as surfactant protein D or Clara cell protein 16 in the Eclipse study [28,29]. Some studies even focused on serum biomarkers profiles (using proteomics) and their association with clinical predictors of long-term outcome such as lung function, exercise tolerance, BODE (Body mass index,

airflow Obstruction, Dyspnea, Exercise tolerance) index and exacerbation frequency [30]. In exacerbations of COPD, several biomarkers have been the topic of studies assessing their prognostic value. The rationale for such studies is that, although clinical information may be significantly associated with outcomes at a collective level, it may not be able to provide individual risk assessment. Biomarkers of interest have the potential to assess several aspects of exacerbations including the degree of systemic inflammation, the presence of an infectious etiology, hemodynamic consequences and so on. Most studies on biomarkers focused on both shortterm and long-term outcomes. As for clinical factors, only biological predictors of short-term outcomes will be discussed here. Biomarkers studied in chronic obstructive pulmonary disease exacerbations

Stolz et al. assessed the prognostic value of several biomarkers in a cohort of 167–202 patients presenting to the ED [31–33]. First, CRP, procalcitonin and copeptin were measured in these individuals [31]. Copeptin is the C-terminal part of the vasopressin precursor and has the advantage of being more stable, and therefore easier to measure reliably. CRP and procalcitonin are well known markers of inflammation and bacterial infection, respectively. It has been showed that procalcitonin guidance allowed to reduce the number of antibiotic courses for COPD exacerbations without impairing outcomes [32]. Plasma pro-adrenomedullin and pro-endothelin were also studied in that cohort [33]. Adrenomedullin has immunomodulating, metabolic and vascular actions: it exerts effects on leukocyte migration and has bactericidal activity, and is involved in the control of electrolyte balance and pulmonary blood flow. Endothelin-1 has vasoconstrictive and proinflammatory effects and is induced by infection and hypoxia. Finally, in a third analysis of their cohort, the authors focused on B-type natriuretic peptide (BNP), which is released by cardiac ventricles in response to cardiac stress (i.e. increased volume or pressure load and myocardial stretch) to induce natriuresis, diuresis and vasodilation. Predictive value of studied biomarkers

In these studies, the main outcomes of interest were LOS and short-term and long-term (2 years) mortality. Predictors of short-term mortality were copeptin and pro-adrenomedullin but not CRP, procalcitonin, proendothelin and BNP [27,32,33]. Factors associated with LOS were pro-adrenomedullin, pro-endothelin, copeptin

116 Obstructive, occupational and environmental diseases

and procalcitonin but not CRP [27,32]. BNP was only associated with ICU requirement [33]. Copeptin levels were associated with a history of cardiopathy and levels of procalcitonin. Pro-adrenomedullin and pro-endothelin levels correlated with age, number of comorbidities, CRP and procalcitonin levels [32].

Table 3 Variables independently associated with the risk of poor outcome (death, need for invasive mechanical ventilation, prolonged hospitalization) in at least one study using multivariate analysis to identify prognostic factors in chronic obstructive pulmonary disease exacerbations Clinical variables Age Male (sex) Severity of underlying COPD (baseline FEV1 or MRC dyspnea grade at steady state) Number of previous admissions for COPD exacerbations Persistent smoking Higher income (US administrative study) Number of Anthonisen criteria (increased dyspnea, sputum volume or purulence) Respiratory rate Number of clinical signs of severityb Comorbid illnesses (cardiopathy including congestive heart failure, cerebrovascular disease, chronic liver disease) Poor performance status SpO2 APACHE II score Blood gas values pH Serum albumin level Hospital organization (including number of respiratory consultants, early discharge scheme and local COPD guidelines)

Discussing the yield of biomarkers as predictors of outcome in chronic obstructive pulmonary disease exacerbations

When interpreting these data, one major question is: what do biomarkers add to clinical variables in the prediction of in-hospital outcomes? To address this issue, multivariate models including and excluding biomarkers of interest would be useful, for instance in assessing the additional percentage of variation in outcomes explained when biomarkers are considered together with clinical variables. In a study of 147 patients with 160 admissions, a composite criterion was used to define an adverse outcome; it included death in the hospital or within 15 days of discharge, ICU requirement and development of acute heart failure during hospital stay [15 ]. Multivariate analysis found that a CRP level of 50 mg/l or more was associated with a nearly five-fold increase in the risk of adverse outcome. Other independent predictors were confusion at admission, at least two comorbidities and current smoking. The multivariate model was used to develop a score for prediction of an adverse outcome. The area under the receiver operating characteristic curve was 0.80, which is interesting but requires further external validation before generalization. In addition, the specific contribution of CRP to the model was not explicitly provided.

Conclusion Table 3 summarizes clinical and biological variables that have been associated with poor short-term outcomes during COPD exacerbations. Some of these variables can be readily available in the ED and may therefore help physicians in the risk stratification of patients. Others (biomarkers) can only be measured in a research context, and little is known on what they add to clinical variables. Risk stratification in the ED is of utmost importance to help physicians deciding where patients should be oriented (discharge, medical ward, ICU). Stratification may also be of help to anticipate the burden of care that a patient will require during his or her hospital stay. Recent studies developed scores for prediction of the risk of death or posthospital support. Further validation

Biomarkers CRP Procalcitonin Copeptina Pro-adrenomedullina Pro-endothelina

APACHE II, Acute Physiology and Chronic Health Evaluation II; COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in 1 second; MRC, Medical Research Council. a Research setting only. b See Table 1.

in other populations is required before their used is generalized.

References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: of special interest of outstanding interest Additional references related to this topic can also be found in the Current World Literature section in this issue (p. 163). 1

Donaldson GC, Wilkinson TM, Hurst JR, et al. Exacerbations, and time spent outdoors in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2005; 171:446–452.

Seemungal TA, Donaldson GC, Paul EA, et al. Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998; 157:1418–1422.

Detournay B, Pribil C, Fournier M, et al. The SCOPE study: health-care consumption related to patients with chronic obstructive pulmonary disease in France. Value Health 2004; 7:168–174.

Andersson F, Borg S, Jansson SA, et al. The costs of exacerbations in chronic obstructive pulmonary disease (COPD). Respir Med 2002; 96:700–708.

Anzueto A, Leimer I, Kesten S. Impact of frequency of COPD exacerbations on pulmonary function, health status and clinical outcomes. Int J Chron Obstruct Pulmon Dis 2009; 4:245–251.

Soler JJ, Martinez-Garcia MA, Roman P, et al. Severe acute exacerbations and mortality in patients with chronic obstructive pulmonary disease. Thorax 2005; 60:925–931.

Celli BR, Barnes PJ. Exacerbations of chronic obstructive pulmonary disease. Eur Respir J 2007; 29:1224–1238.

Bourbeau J, Ford G, Zackon H, et al. Impact on patients’ health status following early identification of a COPD exacerbation. Eur Respir J 2007; 30:907–913.

COPD exacerbations in emergency departments Roche et al. 117 Langsetmo L, Platt RW, Ernst P, et al. Underreporting exacerbation of chronic obstructive pulmonary disease in a longitudinal cohort. Am J Respir Crit Care Med 2008; 177:396–401.

20 Vitacca M, Clini E, Porta R, et al. Acute exacerbations in patients with COPD: predictors of need for mechanical ventilation. Eur Respir J 1996; 9:1487– 1493.

10 Ram FS, Wedzicha JA, Wright J, et al. Hospital at home for patients with acute exacerbations of chronic obstructive pulmonary disease: systematic review of evidence. BMJ 2004; 329:315.

21 Brochard L. Noninvasive ventilation for acute exacerbations of COPD: a new standard of care. Thorax 2000; 55:817–818.

11 Bustamante-Fermosel A, Miguel-Yanes JM, Duffort-Falco M, et al. Mortalityrelated factors after hospitalization for acute exacerbation of chronic obstructive pulmonary disease: the burden of clinical features. Am J Emerg Med 2007; 25:515–522. 12 Roche N, Zureik M, Soussan D, et al. Predictors of outcomes in COPD exacerbation cases presenting to the emergency department. Eur Respir J 2008; 32:953–961. This prospective study developed and provided a first validation of a clinical score predicting the risk of death in patients hospitalized through EDs. 13 Dransfield MT, Rowe SM, Johnson JE, et al. Use of beta blockers and the risk of death in hospitalised patients with acute exacerbations of COPD. Thorax 2008; 63:301–305. Comorbid illnesses are associated with an increased risk of death in patients with COPD exacerbations, whereas beta-blockers exert a protective effect. 14 Gottlieb SS, McCarter RJ, Vogel RA. Effect of beta-blockade on mortality among high-risk and low-risk patients after myocardial infarction. N Engl J Med 1998; 339:489–497. 15 Ruiz-Gonzalez A, Lacasta D, Ibarz M, et al. C-reactive protein and other predictors of poor outcome in patients hospitalized with exacerbations of chronic obstructive pulmonary disease. Respirology 2008; 13:1028– 1033. CRP may add to the prognostic value of clinical features in COPD exacerbations. This study is also one of the very few that identified persistent smoking as an independent risk factor of death in patients hospitalized for COPD exacerbations. 16 Connolly MJ, Lowe D, Anstey K, et al. Admissions to hospital with exacerbations of chronic obstructive pulmonary disease: effect of age related factors and service organisation. Thorax 2006; 61:843–848.

22 Carratu P, Bonfitto P, Dragonieri S, et al. Early and late failure of noninvasive ventilation in chronic obstructive pulmonary disease with acute exacerbation. Eur J Clin Invest 2005; 35:404–409. 23 Phua J, Kong K, Lee KH, et al. Noninvasive ventilation in hypercapnic acute respiratory failure due to chronic obstructive pulmonary disease vs. other conditions: effectiveness and predictors of failure. Intensive Care Med 2005; 31:533–539. 24 Tsimogianni AM, Papiris SA, Stathopoulos GT, et al. Predictors of outcome after exacerbation of chronic obstructive pulmonary disease. J Gen Intern Med 2009; 24:1043–1048. 25 Roberts CM, Lowe D, Bucknall CE, et al. Clinical audit indicators of outcome following admission to hospital with acute exacerbation of chronic obstructive pulmonary disease. Thorax 2002; 57:137–141. 26 Price LC, Lowe D, Hosker H, et al. The UK National COPD Audit 2003: impact of hospital resources and organisation of care on patient outcome following admissions for acute COPD exacerbation. Thorax 2006; 61:837–842. 27 Stolz D, Christ-Crain M, Morgenthaler NG, et al. Copeptin, C-reactive protein, and procalcitonin as prognostic biomarkers in acute exacerbation of COPD. Chest 2007; 131:1058–1067. 28 Lomas DA, Silverman EK, Edwards LD, et al. Evaluation of serum CC-16 as a biomarker for COPD in the ECLIPSE cohort. Thorax 2008; 63:1058–1063. 29 Lomas DA, Silverman EK, Edwards LD, et al. Serum surfactant protein D is steroid sensitive and associated with exacerbations of COPD. Eur Respir J 2009; 34:95–102. 30 Pinto-Plata V, Toso J, Lee K, et al. Profiling serum biomarkers in patients with COPD: associations with clinical parameters. Thorax 2007; 62:595–601.

17 Piquet J, Maurer C, Barbieux H, et al. Acute exacerbations of COPD admitted to hospital: predictive factors of 3 year mortality. Rev Mal Respir 2007; 24:909–916.

31 Stolz D, Christ-Crain M, Bingisser R, et al. Antibiotic treatment of exacerbations of COPD: a randomized, controlled trial comparing procalcitoninguidance with standard therapy. Chest 2007; 131:9–19.

18 Patil SP, Krishnan JA, Lechtzin N, et al. In-hospital mortality following acute exacerbations of chronic obstructive pulmonary disease. Arch Intern Med 2003; 163:1180–1186.

32 Stolz D, Christ-Crain M, Morgenthaler NG, et al. Plasma pro-adrenomedullin but not plasma pro-endothelin predicts survival in exacerbations of COPD. Chest 2008; 134:263–272.

19 Ucgun I, Metintas M, Moral H, et al. Predictors of hospital outcome and intubation in COPD patients admitted to the respiratory ICU for acute hypercapnic respiratory failure. Respir Med 2006; 100:66–74.

33 Stolz D, Breidthardt T, Christ-Crain M, et al. Use of B-type natriuretic peptide in the risk stratification of acute exacerbations of COPD. Chest 2008; 133:1088–1094.