Social Behaviour Risk Factors for Drug Resistant Tuberculosis in Mainland China: a Meta-analysis

The Journal of International Medical Research 2012; 40: 436 445 Social Behaviour Risk Factors for Drug Resistant Tuberculosis in Mainland China: a Meta-analysis P ZHAO 1,a, XJ LI 2,a, SF ZHANG 1, XS WANG 3 AND CY LIU 4 1 Chaoyang Centre for Disease Control and Prevention, Beijing, China; 2 Department of Geriatrics, Southwest Hospital, and 3 Department of Respiratory Medicine, Daping Hospital, Third Military Medical University, Chongqing, China; 4 Department of Respiratory Medicine, First Hospital, Jilin University, Changchun, China OBJECTIVE: To determine risk factors associated with drug resistant tuberculosis (TB) in mainland China. METHODS: PubMed and Chinese BioMedical databases were searched. Cohort, case control and cross-sectional studies providing effect estimates of risk factors for any-drug resistant or multidrug resistant (MDR) TB were included. RESULTS: The meta-analysis included 16 studies. Any-drug resistant TB was significantly associated with poor quality directly observed treatment, shortcourse (DOTS) (odds ratio [OR] 2.65, 95% confidence interval [CI] 1.22, 5.79), long term illness > 1 year (OR 2.71, 1.34, 5.48), poor treatment adherence (OR 2.00, 1.17, 3.40), previous treatment (OR 4.54, 2.71, 7.61) and age 40 60 years (OR 1.62, 1.10, 2.38). MDR-TB was significantly associated with poor quality DOTS (OR 1.84, 1.36, 2.49), poor treatment adherence (OR 4.39, 2.97, 6.50), previous treatment (OR 3.83, 2.12, 6.89) and poverty (OR 1.87, 1.38, 2.52). CONCLUSIONS: Previous treatment, poor quality DOTS, poor treatment adherence, long term illness, age 40 60 years and poverty are associated with a greater risk of drug resistant TB in mainland China. KEY WORDS: TUBERCULOSIS; DRUG RESISTANCE; MULTIDRUG RESISTANT TUBERCULOSIS; RISK FACTORS; META-ANALYSIS Introduction Tuberculosis (TB) remains one of the greatest health problems for people living in the developing world. 1 China has the second highest prevalence of TB worldwide 2 and one of the highest rates of drug resistant TB. 3 Based on a recent national anti-tb drug resistance survey, it was estimated that approximately 120 000 new multidrug a P Zhao and XJ Li contributed equally to this work. resistant (MDR) TB cases emerge annually in China, including 9000 extensively drug resistant TB (XDR-TB) cases, accounting for approximately 24% of the global MDR-TB burden. 4 The emergence of drug resistant TB, especially MDR-TB 5 and XDR-TB, 6,7 poses a substantial threat to TB control programs worldwide. In addition to the direct transmission of drug resistant strains from one individual to another, 8 drug resistant TB 436

cases may be caused by the social behavioural risk factors of patients themselves. Previous studies have identified several risk factors for drug resistant TB in mainland China, 9 18 but there was little consensus between different study populations. For example, floating populations were strongly associated with drug resistant TB in some studies, 9,11,17 but this association was not significant in another study. 10 A nationwide survey on the risk factors associated with drug resistant TB in mainland China has not been performed. The present meta-analysis aimed to weigh the strength and quality of the evidence of a causal association between risk factors and drug resistant TB and MDR-TB in mainland China. Materials and methods LITERATURE SEARCH The PubMed and Chinese BioMedical databases were searched by two researchers (X.J.L., S.F.Z.) to identify relevant publications indexed before October 2011. The following search terms were used as a combination of free text and thesaurus terms: tuberculosis ; drug resistance ; multi-drug resistance ; China ; risk factors ; and epidemiologic determinants. The Chinese Journal of Tuberculosis and Respiratory Diseases and the Journal of the Chinese Antituberculosis Association were selected as the key journals for hand searching. Each article chosen for inclusion was reviewed for data extraction by three independent reviewers (P.Z., X.S.W., C.Y.L.). INCLUSION AND EXCLUSION CRITERIA Original research reports from mainland China published in Chinese and English were included in the review, and comments, editorials and reviews were excluded. Cohort, case control or cross-sectional studies providing effect estimates of the association between any-drug resistant TB or MDR-TB and risk factors were selected for inclusion. The following types of article were excluded: (i) those with < 10 cases; (ii) those involving drug efficacy testing in vitro or in clinical trials; and (iii) those studying indications for management or the treatment of drug resistant TB. Selected articles were reviewed for information on sample size, population source, age and gender. STUDY DEFINITIONS The meta-analysis used the following definitions: MDR-TB, TB resistant to at least both isoniazid and rifampicin; any-drug resistant TB, TB resistant to at least one of four first-line anti-tb drugs (isoniazid, rifampicin, ethambutol and streptomycin) except for MDR-TB cases; poor quality directly observed treatment, short-course (DOTS), not strictly complying with the World Health Organization s (WHO)- recommended DOTS strategy; 19 poor treatment adherence, not following the treatment plan, including the increasing or decreasing of drug dosages by patients, antibiotic abuse or short-term interruption of treatment (< 60 days); interruption of treatment, the interruption of treatment for 60 days in a course of treatment; delay in diagnosis and treatment, > 30 days from the discovery of symptoms to diagnosis. QUALITY ASSESSMENT Each article received three independent reviews and was assessed for quality according to the following criteria: (i) cases were defined adequately; (ii) cases were consecutive or obviously representative; (iii) cases and controls were drawn from the same population; (iv) controls were defined 437

adequately; (v) secure record or structured interview blind to case/control status; (vi) same method of diagnosis for cases and controls; (vii) similar nonresponse rate between groups; (viii) study design adequate to measure an association; (ix) adequate definition of risk factors; (x) use of univariate and multivariate analyses. The articles received one point for each of the criteria listed. A score of < 5 was considered to be low quality, 5 7 was considered medium quality, and > 7 was considered high quality. STATISTICAL ANALYSES Data were collected and analysed using Review Manager (RevMan) software, version 4.2 (The Cochrane Collaboration, Oxford, UK). Data heterogeneity was analysed using the χ 2 -test (α = 0.05), with the extent of heterogeneity determined using the I 2 method (I 2 values of 25%, 50% and 75% were indicative of low, medium and high heterogeneity, respectively). Statistically homogenous data were assessed using a fixed effects model. A random effects model was used to interpret heterogeneity further and improve test performance. The fail-safe number for P = 0.05 (N fs0.05 ) was used statistically to assess publication bias (N fs0.05 < 10 was considered indicative of statistically significant publication bias). Sensitivity analysis was performed by two methods (http://www.cochrane-net.org/openlearning/ html/mod14-2.htm): one to exclude those studies that included univariate analyses alone in order to eliminate bias, and the other to exchange analysis models. If the conclusion for some risk factors changed slightly or not at all after using these two methods, the results of the meta-analysis were deemed reliable. Results The literature search identified a total of 365 articles published in Chinese or English, 349 of which were excluded from the analysis based on the inclusion and exclusion criteria for the meta-analysis. The final metaanalysis, therefore, included 16 studies from 16 different regions or provinces (the studies being from 16 different regions or provinces was a coincidence). 9 18,20 25 All were case control studies published between 2004 and 2010. Of these articles, 11 were extracted from journals, 9,10,12,15,16,20 25 two were master s degree in medicine (MM) theses, 11,13 and three were doctoral degree in medicine (MD) dissertations. 14,17,18 There were 12 studies regarding risk factors in any-drug resistant TB and 11 in MDR-TB. The quality of the articles was variable, with three of low quality, 21,23,24 eight of medium quality, 12 16,18,22,25 and five of high quality. 9 11,17,20 Meta-analysis revealed a significant association between any-drug resistant TB and poor quality DOTS (odds ratio [OR] 2.65, 95% confidence interval [CI] 1.22, 5.79), duration of illness > 1 year (OR 2.71, 1.34, 5.48), poor treatment adherence (OR 2.00, 1.17, 3.40), previous treatment (OR 4.54, 2.71, 7.61) and age 40 60 years (OR 1.62, 1.10, 2.38) (Fig. 1). There was a significant association between MDR-TB and poor quality DOTS (OR 1.84, 1.36, 2.49), poor treatment adherence (OR 4.39, 2.97, 6.50), previous treatment (OR 3.83, 2.12, 6.89), poverty (OR 1.87, 1.38, 2.52), floating population (OR 1.44, 1.08, 1.91) and smoking (OR 1.37, 1.08, 1.76) (Fig. 2). Although MDR-TB was associated with a floating population using the fixed effects model approach for sensitivity analysis, this conclusion was reversed in the random model approach (Fig. 3). There were no other differences in the findings between the two sensitivity analyses. There were no 438

A B Wei et al. 10 Yang 14 Zheng 13 Zhou 11 3.86 [1.98, 7.51] 3.12 [1.64, 5.92] 2.86 [1.15, 7.09] 12.65 [1.43, 111.79] 0.89 [0.56, 1.41] Total () 2.65 [1.22, 5.79] Test for heterogeneity: χ 2 = 20.87, df = 4 (P = 0.0003), I 2 = 80.8% Test for overall effect: Z = 2.46 (P = 0.01) Wei et al. 10 Yang et al. 20 Zheng 13 Zhou 11 3.16 [1.79, 5.58] 0.94 [0.59, 1.51] 3.70 [1.96, 6.98] 2.27 [1.05, 4.91] 6.00 [3.57, 10.10] Total () 2.71 [1.34, 5.48] Test for heterogeneity: χ 2 = 29.42, df = 4 (P < 0.00001), I 2 = 86.4% Test for overall effect: Z = 2.78 (P = 0.005) 0.1 0.2 0.5 1 2 5 10 0.01 0.1 1 10 100 C D E Li et al. 21 Wei et al. 10 Zhang 15 Zhou 11 2.87 [1.51, 5.44] 2.26 [1.19, 4.27] 1.93 [1.26, 2.96] 5.80 [1.47, 22.96] 0.90 [0.60, 1.35] Total () 2.00 [1.17, 3.40] Test for heterogeneity: χ 2 = 16.11, df = 4 (P = 0.003), I 2 = 75.2% Test for overall effect: Z = 2.54 (P = 0.01) 0.01 0.1 1 10 100 Lin et al. 16 Xu et al. 22 Yang 14 Yang et al. 20 Zhang et al. 15 1.79 [0.51, 6.33] 1.31 [0.78, 2.20] 2.10 [1.65, 2.68] 3.30 [1.17, 9.30] 0.71 [0.21, 2.41] 1.04 [1.01, 1.07] 3.10 [0.89, 10.84] Zheng 13 1.40 [0.26, 7.52] Zhou 11 2.03 [1.04, 3.94] Total () 1.62 [1.10, 2.38] Test for heterogeneity: χ 2 = 44.82, df = 8 (P = 0.00001), I 2 = 82.2% Test for overall effect: Z = 2.46 (P = 0.01) Lin et al. 16 Wei et al. 10 Xu et al. 22 Xu et al. 24 Yang 14 Yang et al. 20 Zhang et al. 15 Zheng 13 Zhou 11 Zhu et al. 12 Total () 4.54 [2.71, 7.61] Test for heterogeneity: χ 2 = 196.30, df = 11 (P < 0.00001), I 2 = 94.4% Test for overall effect: Z = 5.74 (P < 0.00001) 0.01 0.1 1 10 100 2.48 [1.41, 4.36] 4.20 [2.97, 5.94] 2.30 [2.02, 2.62] 2.12 [1.45, 3.09] 2.16 [1.14, 4.09] 1.82 [0.97, 3.39] 3.31 [1.40, 7.83] 16.79 [10.14, 27.80] 4.90 [2.58, 9.32] 8.00 [5.13, 12.48] 35.87 [22.30, 57.68] 4.64 [1.92, 11.23] 0.2 0.5 1 2 5 FIGURE 1: Forest plots for meta-analysis outcomes of risk factors for any-drug resistant tuberculosis (TB), showing odds ratios (OR) and 95% confidence intervals (CI). Anydrug resistant TB was defined as TB resistant to at least one of four first-line anti-tb drugs (isoniazid, rifampicin, ethambutol and streptomycin), but not both isoniazid and rifampicin. Plots were generated using combined analysis. (A) Poor-quality directly observed treatment, short-course (DOTS), defined as not strictly complying with the World Health Organization s. 19 recommended DOTS strategy. (B) Long term illness (> 1 year). (C) Poor treatment adherence (not following the treatment plan, including the increasing or decreasing of drug dosages by patients, antibiotics abuse or short-term [< 60 days] interruption of treatment). (D) Previous TB treatment. (E) Aged 40 60 years multivariate analyses regarding smoking in the studies included in the meta-analysis, and there was evidence of publication bias in MDR-TB studies on smoking (N fs.0.05 = 2; Table 1). N fs.0.05 values for all other risk factors assessed in the meta-analysis were > 10, indicating little publication bias (Table 1). Discussion Studies have identified several risk factors in drug resistant TB worldwide. 8,26 29 Differences in the social system and living habits of people in mainland China compared with the rest of the world suggest that the social behaviour risk factors for drug 439

A B Yang 14 Zheng 13 1.67 [1.21, 2.31] 3.01 [1.10, 8.23] 5.04 [1.08, 23.59] Total () 1.84 [1.36, 2.49] Test for heterogeneity: χ 2 = 2.90, df = 2 (P = 0.23), I 2 = 31.0% Test for overall effect: Z = 3.94 (P < 0.0001) Xu et al. 24 1.06 [0.73, 1.54] 2.10 [1.35, 3.27] 3.63 [0.48, 27.25] Total () 1.44 [1.08, 1.91] Test for heterogeneity: χ 2 = 6.13, df = 2 (P = 0.05), I 2 = 67.4% Test for overall effect: Z = 2.51 (P = 0.01) 0.1 0.2 0.5 1 2 5 10 0.2 0.5 1 2 5 C D Wang 18 3.42 [1.64, 7.13] 4.85 [3.05, 7.70] Total () 4.39 [2.97, 6.50] Test for heterogeneity: χ 2 = 0.62, df = 1 (P = 0.43), I 2 = 0% Test for overall effect: Z = 7.41 (P < 0.00001) Wang 18 1.21 [0.83, 1.76] 1.51 [1.09, 2.09] Total () 1.37 [1.08, 1.76] Test for heterogeneity: χ 2 = 0.77, df = 1 (P = 0.38), I 2 = 0% Test for overall effect: Z = 2.54 (P = 0.01) 0.1 0.2 0.5 1 2 5 10 0.1 0.2 0.5 1 2 5 10 E F Li et al. 21 Lin et al. 16 Sun et al. 25 Wang 18 Xu et al. 24 Yang 14 Zhang et al. 23 1.95 [1.23, 3.08] 2.47 [0.89, 6.87] 10.60 [6.88, 16.34] 4.10 [3.22, 5.23] 3.94 [1.59, 9.74] 4.85 [3.05, 7.70] 5.31 [2.59, 10.90] 1.42 [0.54, 3.73] 1.24 [1.12, 1.37] Zheng 13 12.69 [7.03, 22.89] Zhu et al. 12 4.98 [2.19, 11.30] Total () 3.83 [2.12, 6.89] Test for heterogeneity: χ 2 = 239.45, df = 10 (P < 0.00001), I 2 = 95.8% Test for overall effect: Z = 4.47 (P < 0.00001) 1.56 [0.80, 3.04] Lin et al. 16 1.63 [1.11, 2.40] Sun et al. 25 6.67 [1.41, 31.44] Wang 18 1.45 [1.13, 1.86] Yang 14 1.77 [0.71, 4.40] Zhang et al. 23 2.64 [1.91, 3.64] Total () 1.87 [1.38, 2.52] Test for heterogeneity: χ 2 = 11.47, df = 5 (P = 0.04), I 2 = 56.4% Test for overall effect: Z = 4.08 (P < 0.0001) 0.1 0.2 0.5 1 2 5 10 0.01 0.1 1 10 100 FIGURE 2: Forest plots for meta-analysis outcomes of risk factors for multidrug resistant tuberculosis (MDR-TB), showing odds ratios (OR) and 95% confidence intervals (CI). MDR-TB was defined as TB resistant to at least both isoniazid and rifampicin. Plots were generated using combined analysis. (A) Poor-quality directly observed treatment, short-course (DOTS), defined as not strictly complying with the World Health Organization s recommended DOTS strategy. 19 (B) Floating population. (C) Poor adherence to treatment (not following the treatment plan, including the increasing or decreasing of drug dosages by patients, antibiotics abuse or short-term [< 60 days] interruption of treatment). (D) Smoking. (E) Previous TB treatment. (F) Poverty A B Xu et al. 24 1.06 [0.73, 1.54] 2.10 [1.35, 3.27] 3.63 [0.48, 27.25] Total () 1.44 [1.08, 1.91] Test for heterogeneity: χ 2 = 6.13, df = 2 (P = 0.05), I 2 = 67.4% Test for overall effect: Z = 2.51 (P = 0.01) Xu et al. 24 1.06 [0.73, 1.54] 2.10 [1.35, 3.27] 3.63 [0.48, 27.25] Total () 1.59 [0.86, 2.94] Test for heterogeneity: χ 2 = 6.13, df = 2 (P = 0.05), I 2 = 67.4% Test for overall effect: Z = 1.47 (P = 0.14) 0.2 0.5 1 2 5 0.2 0.5 1 2 5 FIGURE 3: Forest plots for sensitivity analysis of meta-analysis outcomes, showing odds ratios (OR) and 95% confidence intervals (CI). (A) OR of floating populations using the fixed effects model. (B) OR of floating populations using the random effects model 440

TABLE 1: Assessment of publication bias in the current meta-analysis of risk factors associated with any-drug resistant tuberculosis (TB) and multidrug resistant TB in mainland China a Any drug-resistant TB Multidrug resistant TB Risk factor No. of studies N fs.0.05 No. of studies N fs.0.05 Poor quality DOTS 5 38 3 21 Duration of illness > 1 year 5 110 Poor adherence to treatment 5 38 2 37 Previous TB treatment 12 2205 11 1307 Poverty 6 91 Smoking 2 2 Aged 40 60 years 9 95 a Publication bias was assessed using the fail-safe number for P = 0.05 (N fs0.05 ). Any-drug resistant TB, TB resistant to at least one of four first-line anti-tb drugs (isoniazid, rifampicin, ethambutol and streptomycin), but not both isoniazid and rifampicin; multidrug resistant TB, TB resistant to at least both isoniazid and rifampicin; DOTS, directly observed treatment, short-course; poor-quality DOTS, not strictly complying with the World Health Organization s recommended DOTS strategy; 19 poor treatment adherence, not following the treatment plan, including the increasing or decreasing of drug dosages by patients, antibiotics abuse or short-term (< 60 days) interruption of treatment. resistant TB may differ from those in other countries. The aim of the present metaanalysis was to determine the risk factors for drug resistant TB in mainland China. Several independent risk factors for anydrug resistance and MDR were identified among TB patients in China. The risk factors for any-drug resistant TB included poor quality DOTS, duration of illness > 1 year, poor treatment adherence, previous treatment and being aged 40 60 years. Risk factors for MDR-TB included poor quality DOTS, poor treatment adherence, previous treatment and poverty. A previous history of treatment for TB was the most widely reported risk factor for MDR- TB 8,26 29 and was the strongest risk factor for both any-drug resistant TB and MDR-TB in the present meta-analysis. Drug resistant Mycobacterium tuberculosis strains arise after repeated cycles of killing (during treatment) and regrowth (after treatment). 30 When patients are medicated repeatedly, the mutant strains become dominant. Resistance to a single drug is followed by the emergence of mutant strains that are resistant to multiple drugs, and the occurrence of MDR- TB. 28 In our opinion, it is likely that inadequate treatment is the main reason for drug resistance in previously treated cases. Delays in disease diagnosis or recognition of drug resistance, inappropriate chemotherapy regimens, inadequate or irregular drug supply, and poor adherence by both patients and clinicians are reported reasons for inadequate treatment. 31,32 The DOTS strategy for TB control was launched by the WHO in 1995. 19 The strategy is based around short course treatment regimens for a minimum of 6 months, but it also requires political commitment, good management practices, sputum smear microscopy for diagnosis, and the direct observation of doses to ensure adherence. 33 To date, DOTS remains the cornerstone of global efforts at TB control. 34 The present meta-analysis showed that poor quality DOTS was associated with a 2.65-fold increased risk of any-drug resistant TB, and a 1.84-fold 441

increased risk of MDR-TB in mainland China. DOTS has been shown to reduce transmission of TB, 35,36 reduce the incidence of drug resistance 37 and prevent the acquisition of drug resistance. 36,38 A study of patients with TB in Beijing between 1978 and 1996 found that DOTS could effectively reduce the prevalence of initial drug resistance when the number of patients in floating populations was low, 39 suggesting that good quality DOTS may prevent the development of drug resistance. In accordance with the findings of others, 26 the present meta-analysis revealed that any-drug resistant TB and MDR-TB were both associated with poor treatment adherence. Improving the treatment compliance of patients with TB may be an effective way to reduce drug resistant TB and increase the cure rate. The reasons for poor treatment adherence in mainland China may include the side-effects of treatment, the advanced age of the population, low education level, poverty and distractions of a busy working life or education. 17,40 Poor quality DOTS may be related to this poor treatment adherence. Patients with long term illness (> 1 year) had a 2.71-fold increased risk of any-drug resistant TB in the present analysis. Some drug resistance mutations are spontaneous, and the frequency of mutation is stable. 30 Over time, these mutations build up and drug resistance may occur. 30 The present analysis found no association between MDR- TB and long term illness; this may have been influenced by bias since many of the papers included in the analysis had not studied the association between MDR-TB and long term illness. This association needs to be studied further. The 1.62-fold increase in risk of anydrug resistant TB in patients aged 40 60 years may be related to the previously reported poor adherence to treatment by patients in this age group. 40 There was no association between MDR-TB and patients aged 40 60 years in the present analysis, in spite of the inclusion of one report 9 (covering five relevant articles) 9,12,14,16,17 to the contrary. The Chinese Ministry of Health s Fourth National Epidemiological Sampling Survey of TB, conducted in 2000, 41 found that 77.9% of all patients with TB in China had below average per capita income, and 53.1% of interrupted treatment and positive sputum smear cases were attributable to poverty. Studies in other countries have also found a close association between poverty and TB or MDR-TB. 42,43 Poverty was associated with MDR-TB in the present analysis. This may be due to the increased risk of contracting TB in crowded and poor living conditions, 28,29 as well as the longer infectious period associated with a shortage of medical and healthcare services. 31 There were several limitations to the present analysis. First, the majority of the included studies (14/16) were published in Chinese, 10 18,21 25 and the quality of the reports was not as high as those from English language journals. Some studies did not include all the required information, 12,13 such as the age and gender of the study population. There was heterogeneity in the pooled estimates of any-drug resistant TB and MDR-TB for some risk factors. These factors may be mutually confounding, and the analysis did not include any cohort studies. Secondly, the possibility of publication bias cannot be completely excluded since positive results are more likely to be published, although no major publication bias was observed. Thirdly, meta-analysis has well-known inherent limitations, including publication and citation bias, misclassification bias, selection and inclusion bias, and the 442

combination of heterogeneous data. 44 Metaanalysis reduces random error but does not necessarily reduce (and may even increase) systematic error. In spite of these limitations, meta-analysis is an attractive approach for combining the findings from independent studies. In this way, the necessary number of patients may be reached and relatively small effects can be detected or excluded with confidence. Meta - analysis can also contribute to considerations regarding the generalizability of study results. In conclusion, the present meta-analysis indicated that previous treatment is the strongest determinant of drug resistant TB in mainland China, with poor quality DOTS and poor treatment adherence both significantly associated with any-drug resistant TB and MDR-TB. Long term illness (> 1 year) and being aged 40 60 years were both significantly associated with increased risk of any-drug resistant TB, and poverty was significantly associated with MDR-TB. These social behaviour risk factors deserve special attention in mainland China, and the screening of patients with these risk factors would assist in both the prevention of drug resistance and the control of TB in general. Conflicts of interest The authors had no conflicts of interest to declare in relation to this article. Received for publication 24 November 2011 Accepted subject to revision 30 November 2011 Revised accepted 10 February 2012 Copyright 2012 Field House Publishing LLP References 1 Gie R, International Union against Tuberculosis and Lung Disease (The Union): Diagnostic Atlas of Intrathoracic Tuberculosis in Children: a Guide for Low Income Countries. Paris: International Union Against Tuberculosis and Lung Disease (The Union), 2003. 2 World Health Organization: Global Tuberculosis Control: Epidemiology, Strategy, Financing: WHO Report 2009. Geneva: WHO, 2009; WHO/HTM/TB/2009.411. 3 World Health Organization: Multidrug and Extensively Drug-resistant TB (M/XDR-TB): 2010 Global Report on Surveillance and Response. Geneva: WHO, 2010; WHO/HTM/TB/2010.3. 4 Ministry of Health of the People s Republic of China: Nationwide Anti-tuberculosis Drug Resistant Baseline Surveillance in China (2007 2008). Beijing: People s Public Health Press: 2010 [in Chinese]. 5 Aziz M A, Wright A, Laszlo A, et al: Epidemiology of antituberculosis drug resistance (The Global Project on Anti-tuberculosis Drug Resistance Surveillance): an updated analysis. Lancet 2006; 368: 2142 2154. 6 Wright A, Bai G, Barrera L, et al: Emergence of Mycobacterium tuberculosis with extensive resistance to second-line drugs worldwide, 2000 2004. MMWR 2006; 55: 301 305. 7 Gandhi N R, Moll A, Sturm A W, et al: Extensively drug-resistant tuberculosis as a cause of death in patients coinfected with tuberculosis and HIV in a rural area of South Africa. Lancet 2006; 368: 1575 1580. 8 Faustini A, Hall AJ, Perucci CA: Risk factors for multidrug resistant tuberculosis in Europe: a systematic review. Thorax 2006; 61: 158 163. 9 Shen X, DeRiemer K, Yuan ZA, et al: Drugresistant tuberculosis in Shanghai, China, 2000 2006: prevalence, trends, and risk factors. Int J Tuberc Lung Dis 2009; 13: 253 259. 10 Wei CY, Wang Q, Chen J: Study on risk factors for acquired drug resistant tuberculosis in some province. Mod Prev Med 2007; 34: 3275 3277 [in Chinese, English abstract]. 11 Zhou LX: A Case Control Study on Risk Factor for Acquired Drug Resistant Tuberculosis of Adult. MM Thesis, Guangxi Medical University, China, 2009. 12 Zhu JF, Wang WB, Wang XC, et al: Epidemic pattern of drug-resistant tuberculosis and its risk factors in Deqing County. Zhejiang J Prev Med 2009; 21: 6 8 [in Chinese, English abstract]. 13 Zheng Z: Epidemiological Studies on the Prevalence Rate and the Impact Factors of Drug Resistant Tuberculosis Among Pulmonary Tuberculosis Inpatients. MM Thesis, Xiangya Medical College, Central South University, China, 2004. 14 Yang BF: Determinants and Molecular Epidemiology of Drug-resistant Tuberculosis in Rural Area of North Jiangsu Province. MD Dissertation, Fudan University, China, 2004. 443

15 Zhang H, Zhang A, Zhao P, et al: Analysis of the status and risk factors of drug resistant tuberculosis in Chaoyang District in Beijing. J Chin Antituberc Assoc 2009; 31: 218 222 [in Chinese, English abstract]. 16 Lin H, Liu J, Chen L, et al: Drug resistance of tuberculosis from 2003 to 2006 in Chongqing. Acta Academiae Medicinae Militaris Tertiae 2008; 30: 1183 1185 [in Chinese, English abstract]. 17 Chen J: Preliminary Study on Social Behavior Factors of TB Patients and Genotyping of Mycobacterium tuberculosis. MD Dissertation, Sichuan University, China, 2007. 18 Wang WH: Study of Risk Factors of Multi-drug Resistant Pulmonary Tuberculosis and the Quality of Life of Patients. MD Dissertation, Huazhong Science and Technology University, China, 2005. 19 World Health Organization: WHO Report on the Tuberculosis Epidemic, 1995. Stop TB at the Source. Geneva: WHO, 1995; Report No. WHO/TB/95.183. 20 Yang X, Li Y, Wen X, et al: Risk factors for drug resistance in pulmonary tuberculosis inpatients. J Evid Based Med 2010; 3: 162 167. 21 Li H, Yan CL, Zhu SJ, et al: Risk factors of multidrug-resistant tuberculosis. J Clin Pulm Med 2010; 15: 367 368 [in Chinese, English abstract]. 22 Xu B, Hu Y, Wang WB, et al: Molecularepidemiological study on the transmission of drug resistant tuberculosis and its influencing factors in rural areas of eastern China. Zhonghua Liu Xing Bing Xue Za Zhi 2010; 31: 525 529. 23 Zhang M, Lou P, Liu L, et al: Drug resistance and its influencing factors among patients with multidrug-resistant tuberculosis. Chin J Public Health 2010; 26: 1439 1440 [in Chinese, English abstract]. 24 Xu L, Yang Z, Lv DL, et al: Analysis of the situation of anti-tb drug resistance and its risk factors in Shenzhen City. China Trop Med 2010; 10: 1333 1335 [in Chinese, English abstract]. 25 Sun B, Hu Y, Zhu FD, et al: Prevalence and analysis of the multidrug-resistant tuberculosis factors in rural areas of north Jiangsu. Chin Prim Health Care 2008; 22: 66 68 [in Chinese, English abstract]. 26 Pritchard AJ, Hayward AC, Monk PN, et al: Risk factors for drug resistant tuberculosis in Leicestershire poor adherence to treatment remains an important cause of resistance. Epidemiol Infect 2003; 130: 481 483. 27 Moniruzzaman A, Elwood RK, Schulzer M, et al: A population-based study of risk factors for drug resistant TB in British Columbia. Int J Tuberc Lung Dis 2006; 10: 631 638. 28 Espinal MA, Laserson K, Camacho M, et al: Determinants of drug-resistant tuberculosis: analysis of 11 countries. Int J Tuberc Lung Dis 2001; 5: 887 893. 29 Law WS, Yew WW, Chiu Leung C, et al: Risk factors for multidrug-resistant tuberculosis in Hong Kong. Int J Tuberc Lung Dis 2008; 12: 1065 1070. 30 Mitchison DA: How drug resistance emerges as a result of poor compliance during short course chemotherapy for tuberculosis. Int J Tuberc Lung Dis 1998; 2: 10 15. 31 Pablos-Méndez A, Raviglione MC, Laszlo A, et al: Global surveillance for antituberculosis-drug resistance, 1994 1997. World Health Organization International Union against Tuberculosis and Lung Disease Working Group on Anti-tuberculosis Drug Resistance Surveillance. N Engl J Med 1998; 338: 1641 1649. 32 World Health Organization: Anti-tuberculosis Drug Resistance in the World, Report No. 3. Geneva: WHO, 2004; WHO/HTM/TB/2004.343. 33 Cox HS, Morrow M, Deutschmann PW: Long term efficacy of DOTS regimens for tuberculosis: systematic review. BMJ 2008; 336: 484 487. 34 World Health Organization: The Stop TB Strategy: Building on and Enhancing DOTS to Meet the TB-related Millennium Development Goals. Geneva: WHO, 2006; WHO/HTM/STB/2006.37. 35 Frieden TR, Fujiwara PI, Washko RM, et al: Tuberculosis in New York City turning the tide. N Engl J Med 1995; 333: 229 233. 36 Munsiff SS, Li J, Cook SV, et al: Trends in drugresistant Mycobacterium tuberculosis in New York City, 1991 2003. Clin Infect Dis 2006; 42: 1702 1710. 37 DeRiemer K, García-García L, Bobadilla-del- Valle M, et al: Does DOTS work in populations with drug-resistant tuberculosis? Lancet 2005; 365: 1239 1245. 38 Yew WW: Directly observed therapy, shortcourse: the best way to prevent multidrugresistant tuberculosis. Chemotherapy 1999; 45(suppl 2): 26 33. 39 Zhang LX, Tu DH, Enarson DA: The impact of directly-observed treatment on the epidemiology of tuberculosis in Beijing. Int J Tuberc Lung Dis 2000; 4: 904 910. 40 Liu YN, Hou M, Wang W, et al: Investigation of impact factors for chemotherapy compliance of tuberculosis patients in Beijing. J Capital Med Univ 2011; 32: 125 128 [in Chinese, English abstract]. 41 National Technical Steering Group of the Epidemiological Sampling Survey for Tuberculosis: Report on Fourth National Epidemiological Sampling Survey of Tuberculosis. Chin J Tuberc Respir Dis 2002; 25: 3 7 [in Chinese, English abstract]. 42 Marahatta SB: Multi-drug resistant tuberculosis burden and risk factors: an update. Kathmandu Univ Med J (KUMJ) 2010; 8: 116 125. 43 Keshavjee S, Gelmanova IY, Pasechnikov AD, et al: Treating multidrug-resistant tuberculosis in 444

Tomsk, Russia: developing programs that address the linkage between poverty and disease. Ann NY Acad Sci 2008; 1136: 1 11. 44 Tacconelli E, Cataldo MA: Identifying risk factors for infections: the role of meta-analyses. Infect Dis Clin North Am 2009; 23: 211 224. Authors address for correspondence Dr Xingsheng Wang Department of Respiratory Medicine, Daping Hospital, Third Military Medical University, 10 Changjiang zhi Road, Daping, Yuzhong District, Chongqing 400042, China. E-mail: meiyi2433@sohu.com Dr Chaoying Liu Department of Respiratory Medicine, First Hospital, Jilin University, 1 Xinmin Street, Chaoyang District, Changchun, Jilin Province 130021, China. E-mail: liuchaoyingliuchao@126.com 445