tbc_surveyinChina_2012

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

original article

National Survey of Drug-Resistant Tuberculosis in China Yanlin Zhao, Ph.D., Shaofa Xu, M.D., Lixia Wang, M.S., Daniel P. Chin, M.D., Shengfen Wang, Ph.D., Guanglu Jiang, B.S., Hui Xia, M.S., Yang Zhou, M.S., Qiang Li, M.S., Xichao Ou, M.S., Yu Pang, Ph.D., Yuanyuan Song, B.S., Bing Zhao, B.S., Hongtao Zhang, Ph.D., Guangxue He, B.S., Jing Guo, Ph.D., and Yu Wang, M.D.

A bs t r ac t Background

The available information on the epidemic of drug-resistant tuberculosis in China is based on local or regional surveys. In 2007, we carried out a national survey of drug-resistant tuberculosis in China.

Methods

We estimated the proportion of tuberculosis cases in China that were resistant to drugs by means of cluster-randomized sampling of tuberculosis cases in the public health system and testing for resistance to the first-line antituberculosis drugs isoniazid, rifampin, ethambutol, and streptomycin and the second-line drugs ofloxacin and kanamycin. We used the results from this survey and published estimates of the incidence of tuberculosis to estimate the incidence of drug-resistant tuberculosis. Information from patient interviews was used to identify factors linked to drug resistance.

Results

Among 3037 patients with new cases of tuberculosis and 892 with previously treated cases, 5.7% (95% confidence interval [CI], 4.5 to 7.0) and 25.6% (95% CI, 21.5 to 29.8), respectively, had multidrug-resistant (MDR) tuberculosis (defined as disease that was resistant to at least isoniazid and rifampin). Among all patients with tuberculosis, approximately 1 of 4 had disease that was resistant to isoniazid, rifampin, or both, and 1 of 10 had MDR tuberculosis. Approximately 8% of the patients with MDR tuberculosis had extensively drug-resistant (XDR) tuberculosis (defined as disease that was resistant to at least isoniazid, rifampin, ofloxacin, and kanamycin). In 2007, there were 110,000 incident cases (95% CI, 97,000 to 130,000) of MDR tuberculosis and 8200 incident cases (95% CI, 7200 to 9700) of XDR tuberculosis. Most cases of MDR and XDR tuberculosis resulted from primary transmission. Patients with multiple previous treatments who had received their last treatment in a tuberculosis hospital had the highest risk of MDR tuberculosis (adjusted odds ratio, 13.3; 95% CI, 3.9 to 46.0). Among 226 previously treated patients with MDR tuberculosis, 43.8% had not completed their last treatment; most had been treated in the hospital system. Among those who had completed treatment, tuberculosis developed again in most of the patients after their treatment in the public health system.

From the Chinese Center for Disease Control and Prevention (Y. Zhao, L.W., S.W., H.X., Y. Zhou, Q.L., X.O., Y.P., Y.S., B.Z., G.H., Y.W.), the Beijing Tuberculosis and Thoracic Tumor Research Institute (Y. Zhao, S.X., G.J., H.Z.), the Bill and Melinda Gates Foundation, China Office (D.P.C.), and People’s University (J.G.) — all in Beijing.

Address reprint requests to Dr. Y. Wang at the Chinese Center for Disease Control and Prevention, Beijing 102206, China, or at wangyu@chinacdc.cn; or to Dr. Chin at the Bill and Melinda Gates Foundation, China Office, Beijing 100027, People’s Republic of China, or at daniel.chin@ gatesfoundation.org.

Drs. Y. Zhao and Xu and Ms. L. Wang contributed equally to this study.

N Engl J Med 2012;366:2161-70.

Copyright © 2012 Massachusetts Medical Society.

Conclusions

China has a serious epidemic of drug-resistant tuberculosis. MDR tuberculosis is linked to inadequate treatment in both the public health system and the hospital system, especially tuberculosis hospitals; however, primary transmission accounts for most cases. (Funded by the Chinese Ministry of Health.) n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

2161

The

n e w e ng l a n d j o u r na l

D

rug-resistant tuberculosis, especially multidrug-resistant (MDR) and extensively drug-resistant (XDR) tuberculosis, is a major threat to the control of tuberculosis worldwide.1-3 To date, estimates of drug-resistant tuberculosis in countries with the highest burden of tuberculosis, including China, have relied on local or regional surveys. China has carried out surveys of antituberculosis-drug resistance in 10 of its 31 provinces. All the provinces had levels of MDR tuberculosis above the global median, and some have been classified as global “hotspots” of MDR tuberculosis.4-6 Studies have consistently identified prior treatment for tuberculosis as a risk factor for drugresistant tuberculosis.7-12 Although other risk factors have been identified, they have generally not been useful in developing effective responses to this public health threat.8,9,11-16 To help China develop a national response to drug-resistant tuberculosis, the Chinese Center for Disease Control and Prevention (CDC) conducted a national surveillance survey of drug-resistant tuberculosis. The objectives of the survey were to determine the proportion of tuberculosis cases that are resistant to drugs, estimate the incidence of drug-resistant tuberculosis, and identify the factors that are linked to drug-resistant tuberculosis, especially to MDR tuberculosis.

Me thods Sampling Method

The National Tuberculosis Reference Laboratory (NTRL) of the Chinese CDC conducted the survey. Cluster-randomized sampling was used to obtain a representative sample of patients with tuberculosis in China.17 The sample size for patients with new cases of smear-positive tuberculosis was set at 3542 patients, with the assumption that 6% of the tuberculosis cases would be resistant to rifampin and that 15% of the culture samples would be lost owing to failure to recover the culture or to growth of nontuberculous mycobacteria, with a precision of ±1.2% for the 95% confidence interval and a design effect of 2 (which translates into a doubling of the sample size). The design effect was set at 2 (by convention) to account for heterogeneity in resistance between clusters. The sample size for patients with previously treated tuberculosis was set at

2162

of

m e dic i n e

1189 patients, with the same assumptions, except that the expected proportion of rifampin resistance was 16% and the desired precision for the 95% confidence interval was ±3.2%.

On the basis of the requirements for sample size and the desire to complete the study in 9 months, the decision was made to select 70 clusters nationwide. The number of clusters assigned to each province was proportional to the number of new smear-positive cases reported by that province relative to the total number of cases nationwide in 2004 and 2005; in addition, it was determined that all provinces should have at least one cluster. The recruitment target per cluster was 51 patients with new cases of tuberculosis and 17 with previously treated cases.

For the cluster sampling, the primary sampling unit was the local tuberculosis clinic at the county or district level; the tuberculosis clinic is the public health center that provides outpatient tuberculosis services throughout China and is usually located at the local site of the CDC. In 2005, there were 2635 local tuberculosis clinics in China. The list of tuberculosis clinics in each province, with the average number of patients with new cases of smear-positive tuberculosis who were notified in 2004 and 2005, was used as the sampling frame. Cluster sampling with a probability proportional to the number of patients with new cases of smear-positive tuberculosis was used to select the local tuberculosis clinics in each province.

Eligible patients had smear-positive tuberculosis that was newly registered during the survey period at tuberculosis clinics selected as sample sites. Patients with new cases were those who had tuberculosis that had never been treated with tuberculosis drugs or that had been treated for less than 1 month. Patients with previously treated cases were those who had been treated for tuberculosis for 1 month or longer.18 Consecutive patients with new or previously treated cases of smear-positive tuberculosis were enrolled in each cluster.

Study Oversight

The study was approved by the Tuberculosis Research Ethics Review Committee of the China CDC. Written informed consent was obtained from each participant. All the authors vouch for the completeness and accuracy of the data presented.

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

National Survey of Drug-Resistant TB in China

Collection of Patient Information

Two trained interviewers separately interviewed each enrolled patient, with the use of a standard questionnaire, to collect sociodemographic data and information on previous treatment for tuberculosis. No information on status with respect to human immunodeficiency virus (HIV) infection was collected. Discrepant interview data were resolved by a third interviewer.

Bacteriologic Examinations

Two sputum samples for culturing were obtained from each eligible patient before the initiation of treatment. For isolation of the culture, each specimen was treated with one volume of 4% sodium hydroxide per one volume of sputum and was then homogenized by vigorous stirring. An aliquot of 0.1 ml of the resulting specimen was inoculated into two tubes of acidified Löwenstein– Jensen medium and incubated at 37°C. The culture was assessed during week 1 for rapidly growing bacteria and again every week thereafter for slower growing bacteria; if no bacteria grew by 8 weeks, the result was recorded as negative.

Cultures with growing colonies were sent to the NTRL for identification and drug-susceptibility testing. Growth characteristics, morphologic characteristics of the colony, and the test for inhibition

by P-nitrobenzoic acid were used to differentiate Mycobacterium tuberculosis from other mycobacteria.

Drug-susceptibility testing was performed after subculturing with the use of Löwenstein–Jensen medium. Drug susceptibility was determined by means of the proportion method, with the following concentrations for the six tuberculosis drugs in the survey: 0.2 μg per milliliter for isoniazid, 4 μg per milliliter for streptomycin, 40 μg per milliliter for rifampicin, 2 μg per milliliter for ethambutol, 2 μg per milliliter for ofloxacin, and 40 μg per milliliter for kanamycin.19 The critical growth proportion for resistance was 1% for all drugs. The NTRL participates in the annual proficiency review of drug-susceptibility testing organized by the Hong Kong Supranational Tuberculosis Reference Laboratory and has passed each review since 2003.

in the statistical analyses. The standard error was adjusted for clustering on the primary sampling unit. The sampling design requires an equal number of new cases in all the clusters. Because of variation among the clusters in the number of cases with drug-susceptibility testing, the sampling weight was given as the reciprocal of the numbers of patients with results of drug-susceptibility testing in each cluster.

Methods published by the World Health Organization (WHO) were used to calculate the incidence of drug-resistant tuberculosis in China.4 We estimated the incidence of drug-resistant cases among patients with new or relapsed cases of tuberculosis by multiplying the 2007 WHO estimate of the incidence of tuberculosis in China (which includes both new and relapsed cases) by the proportion of new and relapsed cases that were drug-resistant.4,20 A patient was considered to have had a relapsed case of tuberculosis if he or she was given drugs for the treatment of tuberculosis and completed a course of treatment or was cured, and tuberculosis developed again.

We estimated the incidence of acquired drugresistant tuberculosis by multiplying the estimated incidence of previously treated but nonrelapsed cases of tuberculosis by the proportion of those cases that were drug-resistant.

The odds ratio was used to evaluate univariate risk factors associated with MDR tuberculosis and with drug-resistant (but not MDR) tuberculosis. Variables with a P value of less than 0.15 in the univariate analysis and other relevant variables were used in logistic-regression modeling designed for survey data. All variables were initially included in the model, and the backward method was used to select the final variables; the models were compared with the use of the maximum rescaled R2 statistic. The statistical interaction between relevant variables was explored. The statistical analyses were performed with the use of SAS software, version 9.1 (SAS).

R e sult s Patients

Patient recruitment began on April 1, 2007, and Data were entered independently by two persons ended on December 31, 2007. A total of 4606 with the use of Epi Info software, version 3.5 patients were enrolled in the study; an additional (CDC). The study design was taken into account 17 patients were invited to participate but deStatistical Analysis

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

2163

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

clined. Cultures of specimens from 537 patients (11.7%) did not grow mycobacteria or were contaminated. Of the 4069 patients with a positive mycobacterial culture, 140 patients (3.4%) had nontuberculous mycobacteria. Results of drugsusceptibility testing were available for 3929 patients (85.3% of all enrolled patients): 3037 patients had new cases of tuberculosis and 892 had previously treated tuberculosis. A total of 730 patients with previously treated tuberculosis were recruited during the period in which the 3037 patients with new cases of tuberculosis were recruited.

Therefore, the actual proportion of patients with previously treated cases among all patients seen at the local tuberculosis clinics was 19.4%. Subsequently, another 162 patients with previously treated cases were recruited, for a total of 892.

and 54.5% of the previously treated cases (95% CI, 49.6 to 59.4) were resistant to at least one of the four first-line antituberculosis drugs (isoniazid, rifampin, ethambutol, and streptomycin).

Overall, 5.7% of new cases (95% CI, 4.5 to 7.0) and 25.6% of previously treated cases (95% CI, 21.5 to 29.8) were MDR tuberculosis, which was defined as disease with resistance to at least isoniazid and rifampin (Table 1, and Table S1 in the Supplementary Appendix, available with the full text of this article at NEJM.org).

Another 11.2% (95% CI, 8.4 to 14.2) of patients with new cases and 16.1% (95% CI, 8.9 to 24.7) of those with previously treated cases had resistance to either isoniazid or rifampin (but not both). Taking into account the actual proportion of previously treated patients among all patients, approximately 1 of 4 patients with tuberDrug-Resistant Tuberculosis culosis had disease that was resistant to isoniaIn this survey, 34.2% of the new cases of tuber- zid, rifampin, or both, and 1 of 10 had MDR tuculosis (95% confidence interval [CI], 30.9 to 37.6) berculosis.

Table 1. Drug Susceptibility and Resistance to First-Line and Second-Line Antituberculosis Drugs.*

Susceptibility or Resistance

Patients with New Cases of Tuberculosis (N = 3037)

Patients with Previously Treated Tuberculosis (N = 892)

no.

% (95% CI)†

no.

% (95% CI)†

Susceptibility to all four first-line drugs‡

2009

65.8 (62.4–69.1)

417

45.5 (40.6–50.4)

Resistance to first-line drugs 1028

34.2 (30.9–37.6)

475

54.5 (49.6–59.4)

Isoniazid

Any first-line drug

486

16.0 (13.9–18.1)

335

38.5 (33.7–43.3)

Rifampin

202

6.7 (5.3–8.0)

258

29.4 (24.8–34.0)

Ethambutol

153

4.9 (3.7–6.1)

157

17.2 (13.7–20.7)

Streptomycin

834

27.7 (24.4–31.0)

320

37.2 (32.5–42.0)

Isoniazid or rifampin (but not both)

338

11.2 (8.4–14.2)

141

16.1 (8.9–24.7)

Multidrug resistance§ Susceptibility to ofloxacin and kanamycin

175

5.7 (4.5–7.0)

226

25.6 (21.5–29.8)

2906

95.8 (94.7–96.9)

797

88.6 (85.3–91.8) 11.4 (8.2–14.7)

Resistance to ofloxacin and kanamycin 131

4.2 (3.1–5.3)

95

Ofloxacin

Ofloxacin or kanamycin

88

2.7 (1.8–3.6)

76

8.7 (6.1–11.2)

Kanamycin

59

2.0 (1.4–2.6)

33

4.8 (2.5–7.2)

Multidrug resistance plus resistance to ofloxacin or kanamycin

58

1.8 (1.0–2.6)

73

8.5 (6.4–10.6)

Extensive drug resistance¶

15

0.5 (0.2–0.8)

14

2.1 (0.6–3.5)

* Additional data are available in Table 1 in the Supplementary Appendix. † The calculation of the percentages and their 95% confidence intervals took into account the clustered design; therefore, the reported percentages may not match the percentages obtained by dividing the number of susceptible or resistant cases by the total number of patients. ‡ First-line antituberculosis drugs include isoniazid, rifampin, ethambutol, and streptomycin; second-line antituberculosis drugs include ofloxacin and kanamycin. § Multidrug resistance is defined as resistance to at least isoniazid and rifampin. ¶ Extensive drug resistance is defined as resistance to at least isoniazid, rifampin, ofloxacin, and kanamycin.

2164

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

National Survey of Drug-Resistant TB in China

Among the patients with MDR tuberculosis, 8.3% with new cases of tuberculosis (95% CI, 2.9 to 13.6) and 8.0% with previously treated tuberculosis (95% CI, 2.2 to 13.9) had XDR tuberculosis, which was defined as resistance to at least isoniazid, rifampin, ofloxacin, and kanamycin; in addition, 31.4% of the patients with new cases (95% CI, 18.3 to 44.5) and 33.3% of those with previously treated cases (95% CI, 25.6 to 41.0) had disease that was resistant to either ofloxacin or kanamycin. Among all patients with new or previously treated cases of tuberculosis, 0.5% (95% CI, 0.2 to 0.8) and 2.1% (95% CI, 0.6 to 3.5), respectively, had XDR tuberculosis.

Overall, 2.7% of patients with new cases of tuberculosis (95% CI, 1.8 to 3.6) and 8.7% of patients with previously treated tuberculosis (95% CI, 6.1 to 11.2) had resistance to ofloxacin. However, among patients with MDR tuberculosis, 24.9% of those with new cases of tuberculosis (95% CI, 13.1 to 36.8) and 27.5% of those with previously treated tuberculosis (95% CI, 20.8 to 34.3) had resistance to ofloxacin.

In 2007, an estimated 110,000 incident cases of MDR tuberculosis (95% CI, 97,000 to 130,000) developed: 86,000 cases (95% CI, 76,000 to 101,000) of new or relapsed tuberculosis and 24,000 cases (95% CI, 21,000 to 29,000) of acquired MDR tuberculosis. There were an estimated 8200 incident cases (95% CI, 7200 to 9700) of XDR tuberculosis: 7000 cases (95% CI, 6200 to 8300) of new or relapsed tuberculosis and 1200 cases (95% CI, 1000 to 1400) of acquired XDR tuberculosis. Primary transmission of drug-resistant disease accounted for 78% and 86% of incident cases of MDR and XDR tuberculosis, respectively.

Factors Linked to Drug-Resistant Tuberculosis

Tables 2 and 3 show the results of analyses of univariate risk factors for MDR and drug-resistant (but not MDR) tuberculosis in patients with new cases and in those with previously treated cases, respectively, and Table 4 shows the results of an analysis of independent risk factors. In the analysis of risk factors for MDR tuberculosis in previously treated patients, there was an interaction between the number of prior treatment episodes and the health facility at which the last treatment had been received; the highest risk of drug resistance was observed among patients with multiple previous treatment episodes whose last treatment had been received in a tuberculosis hos-

pital. Tuberculosis hospitals are hospitals that specialize in the treatment of tuberculosis, especially difficult-to-treat cases.

Among 226 patients with MDR tuberculosis who had received previous treatment, 99 (43.8%) had not completed their last tuberculosis treatment, whereas 127 (56.2%) had completed their last treatment (Fig. 1). Among those who had completed their last treatment, 115 (90.6%) had relapsed tuberculosis; 71 of the patients with relapsed tuberculosis (61.7%) had received their last treatment in the CDC system.

Discussion The results of this nationwide survey in China confirm that the country has a serious epidemic of drug-resistant tuberculosis. In 2007, one third of the patients with new cases of tuberculosis and one half of the patients with previously treated tuberculosis had drug-resistant disease. At 5.7%, the prevalence of MDR tuberculosis among new cases was 3.5 times the global median and nearly twice the global average.4,5 In addition, one quarter of the patients with previously treated tuberculosis had MDR tuberculosis. XDR tuberculosis is also widespread, as evidenced by the fact that 8% of patients with MDR tuberculosis also had XDR tuberculosis. An estimated 110,000 cases of MDR tuberculosis and 8200 cases of XDR tuberculosis developed in 2007. With the use of the WHO estimate of MDR tuberculosis in various countries as a reference,4 China has the highest annual number of cases of MDR tuberculosis in the world — a quarter of the cases worldwide.

Although China already has a serious epidemic of drug-resistant tuberculosis, the situation could easily be much worse. This 2007 survey showed that, in addition to patients with MDR tuberculosis, 11% of the patients with new cases and 16% of those with previously treated cases had disease that was resistant to either isoniazid or rifampin, placing them just one step away from having MDR tuberculosis. Similarly, among patients with MDR but not XDR tuberculosis, more than one third had disease that was resistant to either ofloxacin or kanamycin, placing them only one step away from having XDR tuberculosis.

Prevention of drug-resistant tuberculosis, especially MDR tuberculosis, is an essential part of tuberculosis-control programs.1 In this survey, more than 40% of patients with MDR tuberculo-

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

2165

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Table 2. Univariate Analysis of Risk Factors for Drug-Resistant Tuberculosis (TB) in Patients with New Cases of Tuberculosis.*

Variable

Pansensitive TB†

Drug-Resistant, but Not MultidrugResistant, TB‡

Drug-Resistant, but Not Multidrug-Resistant, TB vs. Pansensitive TB Odds Ratio (95% CI)¶

MultidrugResistant TB§

P Value

no./total no. (%)¶

Multidrug-Resistant TB vs. Pansensitive TB Odds Ratio (95% CI)¶

P Value

no./total no. (%)¶

Sex Male

1408/2009 (70.1)

604/853 (71.5)

reference

601/2009 (29.9)

249/853 (28.5)

0.9 (0.8–1.1)

Female

103/175 (62.0)

reference

0.46

72/175 (38.0)

1.4 (1.0–2.0)

0.03

Age <20 yr

178/2009 (8.9)

69/853 (7.9)

reference

21/175 (11.6)

reference

20–39 yr

708/2009 (35.0)

309/853 (36.3)

1.2 (0.8–1.6)

0.36

60/175 (33.2)

0.7 (0.4–1.3)

0.55

40–59 yr

578/2009 (29.0)

273/853 (32.0)

1.2 (0.9–1.7)

0.08

58/175 (34.7)

0.9 (0.5–1.6)

0.27

≥60 yr

545/2009 (27.1)

202/853 (23.7)

1.0 (0.7–1.4)

0.19

36/175 (20.5)

0.6 (0.3–1.1)

0.05

Farmer

1310/2009 (66.8)

540/853 (63.4)

reference

0.09

107/175 (59.8)

reference

Other

699/2009 (33.2)

313/853 (36.6)

1.2 (1.0–1.4)

68/175 (40.2)

1.4 (1.0–1.9)

Implemented before 2000

1105/2009 (52.5)

482/853 (54.5)

reference

88/175 (46.8)

reference

Implemented in 2000 or later

904/2009 (47.5)

371/853 (45.5)

0.9 (0.8–1.1)

87/175 (53.2)

1.3 (0.9–1.7)

No treatment, with or without a prior diagnosis of TB

1449/1749 (82.4)

583/740 (78.2)

reference

104/146 (71.5)

reference

Treatment for <1 mo and a prior diag­ nosis of TB

123/1749 (7.3)

78/740 (10.9)

1.6 (1.2–2.1)

0.02

12/146 (7.6)

1.2 (0.6–2.3)

0.46

Treatment of any duration and no prior diagnosis of TB

177/1749 (10.2)

79/740 (10.9)

1.1 (0.8–1.5)

0.50

30/146 (20.9)

2.4 (1.5–3.7)

0.002

Occupation 0.08

Residence in area with DOTS‖

0.34

0.16

History of treatment with TB drugs and prior diagnosis of TB

* Only results for variables included in the final multivariate model are shown here; additional results are available in Table 2 in the Supple­ mentary Appendix. † Pansensitive TB was defined as TB that was susceptible to the four first-line antituberculosis drugs in this survey (isoniazid, rifampin, ethambutol, and streptomycin). ‡ Drug-resistant (but not multidrug-resistant) TB was defined as TB with resistance to any of the four first-line antituberculosis drugs in the survey, with the exception of resistance to both isoniazid and rifampin. § Multidrug-resistant TB was defined as TB that was resistant to both isoniazid and rifampin. ¶ The calculation of percentages and odds ratios took into account the clustered design; therefore, the percentages and odds ratios presented in the table may not be reproducible simply from data in the table. ‖ Directly Observed Treatment Short Course (DOTS) is an internationally recommended strategy for control of tuberculosis.

sis who had received previous treatment for tuberculosis had not completed their last treatment course. This finding points to the need for interventions that will increase continuity of treatment and reduce the rate of treatment default, especially among patients treated within the hospital system.21 In China, patients with 2166

tuberculosis receive treatment either at the local site of the CDC or in the hospital system. Hospitals provide limited outpatient follow-up of patients with tuberculosis. Therefore, the Ministry of Health has strengthened the reporting, referral, and follow-up of patients with tuberculosis who are seen in the hospital system.22 However,

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

National Survey of Drug-Resistant TB in China

Table 3. Univariate Analysis of Risk Factors for Drug-Resistant TB in Patients with Previously Treated TB.* Drug-Resistant, but Not MultidrugPansensitive TB Resistant, TB

Variable

Drug-Resistant, but Not Multidrug-Resistant, TB vs. Pansensitive TB Odds Ratio (95% CI)†

MultidrugResistant TB

P Value

no./total no.(%)†

Multidrug-Resistant TB vs. Pansensitive TB Odds Ratio (95% CI)†

P Value

no./total no. (%)†

Sex Male

350/417 (83.7)

194/249 (75.1)

reference

67/417 (16.3)

55/249 (24.9)

1.7 (1.1–2.7)

<40 yr

106/417 (25.1)

71/249 (10.4)

reference

40–59 yr

154/417 (38.7)

101/249 (39.1)

0.8 (0.5–1.3)

0.99

88/226 (40.2)

0.6 (0.4–1.0)

0.52

≥60 yr

157/417 (36.2)

77/249 (30.5)

0.7 (0.4–1.1)

0.15

43/226 (18.3)

0.3 (0.2–0.5)

<0.001

Implemented before 2000

249/417 (62.7)

133/249 (55.3)

reference

106/226 (50.0)

reference

Implemented in 2000 or after

168/417 (37.3)

116/249 (44.7)

1.4 (1.0–1.9)

120/226 (50.0)

1.7 (1.2–2.4)

1

349/416 (82.3)

185/248 (73.9)

reference

123/225 (53.6)

reference

2

47/416 (12.5)

33/248 (13.7)

1.2 (0.7–2.1)

0.34

53/225 (22.9)

2.8 (1.7–4.6)

≥3

20/416 (5.2)

30/248 (12.4)

2.7 (1.4–5.2)

0.01

49/225 (23.5)

7.0 (3.7–13.1) <0.001

TB clinic system‡

190/417 (46.3)

107/249 (38.9)

reference

General hospital

Female

0.02

159/226 (69.2)

reference

67/226 (30.8)

2.3 (1.5–3.6)

95/226 (41.6)

reference

<0.001

Age

Residence in area with DOTS 0.09

0.005

No. of prior courses of TB drug treatment 0.83

Medical facility providing last TB treatment 106/226 (43.7)

reference

161/417 (38.2)

80/249 (33.0)

1.0 (0.7–1.6)

0.02

59/226 (26.2)

0.7 (0.5–1.1)

0.005

TB hospital

28/417 (6.2)

25/249 (11.2)

2.2 (1.1–4.2)

0.22

28/226 (15.1)

2.6 (1.4–4.8)

0.02

Private hospital or clinic

17/417 (3.9)

15/249 (5.1)

1.5 (0.7–3.5)

0.98

13/226 (5.1)

1.4 (0.6–3.2)

0.98

Other

21/417 (5.4)

22/249 (11.8)

2.7 (1.3–5.4)

0.07

20/226 (9.8)

1.9 (0.9–4.1)

0.26

* Only results for variables included in the final multivariate model are shown here; additional results are available in Table 3 in the Supplementary Appendix. † The calculation of percentages and odds ratios took into account the clustered design; therefore the percentages and odds ratio presented in the table may not be reproducible simply from data in the table. ‡ The TB clinic is the public health center that provides outpatient TB services throughout China and is usually located at the local site of the Chinese Center for Disease Control and Prevention.

more needs to be done. The CDC system, which is responsible for monitoring patients with tuberculosis in the community, can pilot new approaches to improving adherence to treatment, such as mobile-phone text messaging, and can expand their use if they prove to be effective.23,24 In addition to direct efforts to prevent MDR tuberculosis, effective treatment for MDR and other drug-resistant forms of tuberculosis is needed to decrease the number of prevalent cases and reduce transmission of drug-resistant M. tuberculosis. Transmission of MDR and XDR strains of M. tuberculosis will continue to go unchecked if

there is no effective treatment program. In the coming years, China plans to expand an effective treatment program for MDR tuberculosis that is modeled after international best practice.20 However, the high level of resistance to ofloxacin — a drug that is in an important class of second-line tuberculosis drugs — may reduce the effectiveness of current treatment for MDR tuberculosis and of drug regimens under development.25 Other results of our survey point to the need for additional interventions to prevent MDR tuberculosis in China. First, 11% of patients

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

2167

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Table 4. Multivariate Analysis of Risk Factors for Drug-Resistant TB in New and Previously Treated Cases of TB.* Risk Factor

New Cases of TB Drug-Resistant, but Not MultidrugResistant, TB

Previously Treated TB

MultidrugResistant TB

Drug-Resistant, but Not MultidrugResistant, TB

MultidrugResistant TB

adjusted odds ratio (95% CI) Female sex

0.9 (0.7–1.1)

1.5 (1.0–2.1)

1.7 (1.1–2.7)

2.2 (1.4–3.5)

Age >60 yr Occupation as nonfarmer

0.9 (0.7–1.1)

NA†

0.8 (0.5–1.2)

0.4 (0.3–0.7)

1.2 (1.0–1.4)

1.4 (1.0–2.0)

NA†

Residence in area with DOTS implementation in 2000 or after

0.8 (0.7–1.0)

NA†

1.4 (1.0–2.0)

NA† 1.7 (1.2–2.6)

History of treatment with TB drugs and prior TB diagnosis reference

reference

NA

NA

Treatment of <1 mo and a prior TB diagnosis

No treatment, with or without a prior TB diagnosis

1.6 (1.1–2.1)

1.2 (0.7–2.4)

NA

NA

Treatment of any duration and no prior TB diagnosis

1.2 (0.9–1.5)

2.4 (1.5–3.8)

NA

NA

One prior course of treatment, at medical facility other than TB hospital

NA

NA

reference

reference

One prior course of treatment, at TB hospital

NA

NA

1.6 (0.7–3.5)

1.5 (0.6–3.4)

≥2 prior courses of treatments, with last treatment at medical facility other than TB hospital

NA

NA

1.5 (0.9–2.3)

3.3 (2.1–5.2)

≥2 prior courses of treatments, with last treatment at TB hospital

NA

NA

4.0 (1.2–14.0)

13.3 (3.9–46.0)

No. of prior TB treatment episodes and medical facility ­providing last TB treatment

* NA denotes not applicable. † These variables were not included in the final logistic-regression models.

with new cases of tuberculosis and 16% of those with previously treated tuberculosis in China’s CDC system already had disease that was resistant to either isoniazid or rifampin (but not both) before they received standard first-line short-course treatment. Resistance to these drugs is usually not detected because culturing and drug-susceptibility testing are not routinely performed at the local tuberculosis clinics. The use of standard first-line drugs in the treatment of these patients increases the risk of relapse, treatment failure, and acquired resistance.26-29 To prevent MDR tuberculosis from developing in patients who are being treated at the local tuberculosis clinics, testing for drug resistance should be performed before the initiation of treatment, and the choice of treatment should be based on the results of that testing.

Second, patients who had received multiple treatments for tuberculosis and who had received their last treatment in a tuberculosis hospital were 13 times as likely to have MDR 2168

tuberculosis as those who had received one prior treatment elsewhere. There are several possible explanations for this finding: the patients may already have had MDR tuberculosis when they were admitted to the tuberculosis hospital and then did not receive effective treatment for MDR tuberculosis; they may have been admitted without MDR tuberculosis and received improper treatment, which led to the development of MDR tuberculosis; or they may have acquired MDR tuberculosis from nosocomial transmission. Regardless of the cause, a major effort to improve treatment in tuberculosis hospitals is needed.

Third, 11% of patients with new cases in this survey reported that they had been treated previously with tuberculosis drugs even though they had never received a diagnosis of tuberculosis.

These patients had a higher risk of MDR tuberculosis. One possible explanation is that these patients were suspected to have tuberculosis and were started on tuberculosis drugs but did not actually receive a definitive diagnosis of tuberculosis. Without a tuberculosis diagnosis, many

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

National Survey of Drug-Resistant TB in China

Treatment complete, 127 (56%)

Private facilities, Other 2 (2%) facilities, 9 (7%)

Treatment incomplete, 99 (44%)

Private facilities, 11 (11%)

TB hospital, 13 (10%) General hospital, 23 (18%)

Other facilities, 11 (11%) TB hospital, 15 (15%)

CDC, 80 (63%)

CDC, 26 (26%)

General hospital, 36 (36%)

Figure 1. Patients with Previously Treated Tuberculosis in Whom Multidrug-Resistant Tuberculosis Was Detected, According to Completion or Noncompletion of Their Last Treatment Course and Location of Their Last Treatment.

In China, most patients with tuberculosis receive treatment either at the local tuberculosis clinic of the Chinese Center for Disease Control and Prevention (CDC) or in the hospital system. Patients receiving care in hospitals that specialize in the treatment of tuberculosis (TB hospitals) are more likely to be those with cases that are difficult to treat.

were not followed up for this disease. MDR tuberculosis developed in some of these patients when they took their drugs improperly. Recently, the WHO recommended the use of the GeneXpert system (Cepheid) to evaluate patients who are suspected to have MDR tuberculosis and those with negative sputum smears for acid-fast bacilli.30,31 The use of this diagnostic system can provide a definitive diagnosis for nearly all patients with pulmonary tuberculosis. Thus, it is possible to accurately determine who needs treatment and follow-up for tuberculosis. Appropriate follow-up of these patients during treatment could help prevent MDR tuberculosis.32 There are several limitations of this survey.

First, the burden of drug-resistant tuberculosis was probably underestimated because patients in the hospital system — who are more likely to have drug-resistant disease — were not included in the survey. Second, the burden of XDR tuberculosis was underestimated because only resistance to ofloxacin and kanamycin — not resistance to capreomycin or other aminoglycosides — was used in the identification of XDR tuberculosis. Third, this survey did not collect infor-

mation on HIV infection status because patients with tuberculosis in China are not routinely tested for HIV. Fourth, information on previous treatment was based only on interviews with patients; this may have resulted in inaccurate information regarding treatment.

This survey presents a comprehensive view of the epidemic of drug-resistant tuberculosis in China. The country has the world’s largest number of patients with MDR tuberculosis, and primary transmission is responsible for most cases.

MDR tuberculosis is also linked to inadequate treatment in both the public health system and the hospital system. One in four patients with tuberculosis treated in the public health system are receiving inadequate treatment because they have disease that is resistant to isoniazid or rifampin (or both) but are still given these firstline tuberculosis drugs as part of standard treatment. Therefore addressing MDR tuberculosis in China will require selection of treatment regimens on the basis of testing for initial drug resistance.

It is also important to improve treatment in tuberculosis hospitals and enhance the continuity of treatment after patients leave the hospitals. Ulti-

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.

2169

National Survey of Drug-Resistant TB in China

mately, by treating patients with drug-resistant tuberculosis, China can prevent the development of more drug-resistant forms of tuberculosis and reduce the transmission of drug-resistant tuberculosis.

No potential conflict of interest relevant to this article was reported.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank the provincial and local CDC staffs for their help in carrying out this survey and Dr. Shuigao Jin for assistance with data analysis.

References 1. Nathanson E, Nunn P, Uplekar M, et al.

MDR tuberculosis — critical steps for prevention and control. N Engl J Med 2010;363:1050-8.

2. Prevention and control of multidrugresistant tuberculosis and extensively drug-resistant tuberculosis: World Health Assembly resolution 62.15, May 2009 (http://apps.who.int/gb/ebwha/pdf_files/ A62/A62_R15-en.pdf).

3. Gandhi NR, Nunn P, Dheda K, et al.

Multidrug-resistant and extensively drugresistant tuberculosis: a threat to global control of tuberculosis. Lancet 2010;375: 1830-43.

4. Multidrug and extensively drug-resistant TB (M/XDR-TB): 2010 global report on surveillance and response. Geneva: World Health Organization, 2010 (WHO/ HTM/TB/2010.3).

5. Wright A, Zignol M, Van Deun A, et al.

Epidemiology of antituberculosis drug resistance 2002-07: an updated analysis of the Global Project on Anti-Tuberculosis Drug Resistance Surveillance. Lancet 2009;373:1861-73.

6. Espinal MA, Laszlo A, Simonsen L, et al. Global trends in resistance to antituberculosis drugs. N Engl J Med 2001; 344:1294-303.

7. Espinal MA, Laserson K, Camacho M, et al. Determinants of drug-resistant tuberculosis: analysis of 11 countries. Int J Tuberc Lung Dis 2001;5:887-93.

8. Lomtadze N, Aspindzelashvili R, Janj­ gava M, et al. Prevalence and risk factors for multidrug-resistant tuberculosis in the Republic of Georgia: a population-based study. Int J Tuberc Lung Dis 2009;13:6873.

9. Aguiar F, Vieira MA, Staviack A, et al.

Prevalence of anti-tuberculosis drug resistance in an HIV/AIDS reference hospital in Rio de Janeiro, Brazil. Int J Tuberc Lung Dis 2009;13:54-61.

10. Clark CM, Li J, Driver CR, Munsiff SS. Risk factors for drug-resistant tuberculosis among non-US-born persons in New York City. Int J Tuberc Lung Dis 2005;9: 964-9.

11. Boonsarngsuk V, Tansirichaiya K,

­ iatboonsri S. Thai drug-resistant tuberK culosis predictive scores. Singapore Med J 2009;50:378-84.

12. Kliiman K, Altraja A. Predictors of extensively drug-resistant pulmonary tuberculosis. Ann Intern Med 2009;150:766-75.

13. Shen X, DeRiemer K, Yuan ZA, et al.

Drug-resistant tuberculosis in Shanghai, China, 2000-2006: prevalence, trends and risk factors. Int J Tuberc Lung Dis 2009; 13:253-9.

14. Telles MAS, Ferrazoli L, Waldman EA, et al. A population-based study of drug resistance and transmission of tuberculosis in an urban community. Int J Tuberc Lung Dis 2005;9:970-6.

15. Moniruzzaman A, Elwood RK, Schulzer M, FitzGerald JM. A population-based study of risk factors for drug-resistant TB in British Columbia. Int J Tuberc Lung Dis 2006;10:631-8.

16. Caminero JA. Multidrug-resistant tuberculosis: epidemiology, risk factors and case finding. Int J Tuberc Lung Dis 2010; 14:382-90.

17. Guidelines for surveillance of drug resistance in tuberculosis. Geneva: World Health Organization, 2009 (WHO/HTM/ TB/2009.422).

18. Treatment of tuberculosis: guidelines for national programmes. Geneva: World Health Organization, 2003 (WHO/CDS/ TB/2003.313).

19. Canetti G, Froman S, Grosset J, et al.

Mycobacteria: laboratory methods for testing drug sensitivity and resistance. Bull World Health Organ 1963;29:565-78.

20. World Health Organization. Tuberculosis (TB) (http://www.who.int/tb/country/ data/download/en/index.html).

21. Engaging all health care providers in TB control: guidance on implementing public-private mix approaches. Geneva: World Health Organization, 2006 (WHO/ HTM/TB/2006.360).

22. Wang L, Liu J, Chin DP. Progress in tuberculosis control and the evolving public-health system in China. Lancet 2007;369:691-6.

23. Barclay E. Text messages could hasten

tuberculosis drug compliance. Lancet 2009;373:15-6. [Erratum, Lancet 2009;373: 632.]

24. Lester RT, Ritvo P, Mills EJ, et al. ­Effects of a mobile phone short message service on antiretroviral treatment adherence in Kenya (WelTel Kenya1): a randomised trial. Lancet 2010;376:1838-45.

25. Ma Z, Lienhardt C, McIlleron H, Nunn AJ, Wang X. Global tuberculosis drug development pipeline: the need and the reality. Lancet 2010;375:2100-9.

26. Mak A, Thomas A, Del Granado M, Zaleskis R, Mouzafarova N, Menzies D. Influence of multidrug resistance on tuberculosis treatment outcomes with standardized regimens. Am J Respir Crit Care Med 2008;178:306-12.

27. Lew W, Pai M, Oxlade O, Martin D, Menzies D. Initial drug resistance and tuberculosis treatment outcomes: systematic review and meta-analysis. Ann Intern Med 2008;149:123-34.

28. Menzies D, Benedetti A, Paydar A, et al. Standardized treatment of active tuberculosis in patients with previous treatment and/or with mono-resistance to isoniazid: a systematic review and metaanalysis. PLoS Med 2009;6(9):e1000150.

29. Yoshiyama T, Yanai H, Rhiengtong D, et al. Development of acquired drug resistance in recurrent tuberculosis patients with various previous treatment outcomes.

Int J Tuberc Lung Dis 2004;8:31-8.

30. Boehme CC, Nabeta P, Hillemann D, et al. Rapid molecular detection of tuberculosis and rifampin resistance. N Engl J Med 2010;363:1005-15.

31. Policy statement: automated real-time nucleic acid amplification technology for rapid and simultaneous detection of ­t uberculosis and rifampicin resistance: Xpert MTB/RIF system. Geneva: World Health Organization, 2011 (WHO/HTM/ TB/2011.4).

32. Small PM, Pai M. Tuberculosis diagnosis — time for a game change. N Engl J Med 2010;363:1070-1.

Copyright © 2012 Massachusetts Medical Society.

receive immediate notification when an article is published online first

To be notified by e-mail when Journal articles are published Online First, sign up at NEJM.org.

2170

n engl j med 366;23 nejm.org june 7, 2012

The New England Journal of Medicine Downloaded from nejm.org on June 8, 2012. For personal use only. No other uses without permission.

Copyright © 2012 Massachusetts Medical Society. All rights reserved.