Bulletin of the World Health Organization

Surveillance of anti-tuberculosis drug resistance in the world: an updated analysis, 2007–2010

Matteo Zignol a, Wayne van Gemert a, Dennis Falzon a, Charalambos Sismanidis a, Philippe Glaziou a, Katherine Floyd a & Mario Raviglione a

a. STOP TB Department, World Health Organization, 20 avenue Appia, 1211 Geneva 27, Switzerland.

Correspondence to Matteo Zignol (e-mail: zignolm@who.int).

(Submitted: 24 June 2011 – Revised version received: 07 October 2011 – Accepted: 10 October 2011 – Published online: 07 November 2011.)

Bulletin of the World Health Organization 2012;90:111-119D. doi: 10.2471/BLT.11.092585

Introduction

Surveillance of resistance to drugs against tuberculosis (TB) is a cornerstone of any TB control programme. Surveillance data on drug resistance are needed to track the effectiveness of TB prevention and control activities; accurately forecast the need for patient treatments and plan accordingly; design standardized regimens for the treatment of drug-resistant TB; assess epidemiological trends; and promptly identify and respond to outbreaks of drug-resistant TB.1 Since 1994 the Global Project on Anti-Tuberculosis Drug Resistance Surveillance of the World Health Organization (WHO) has supported national TB control programmes worldwide in implementing drug resistance surveillance activities. Country data are routinely collected, analysed and disseminated to describe the global problem of drug-resistant TB.211

Patients whose mycobacteria are resistant to rifampicin, isoniazid and other anti-TB drugs require longer, expensive and more toxic treatment regimens and are less likely to be cured. This presents a formidable challenge to programmes, particularly in low-resource settings.12 Policy guidance on the programmatic management of drug-resistant TB1315 and on how to control the transmission of resistant strains16 has been developed by WHO, and access to quality-assured second-line anti-TB drugs for the treatment of multidrug-resistant TB (MDR-TB) is facilitated through the Green Light Committee mechanism.17 The number of TB patients diagnosed and treated for MDR-TB, which is defined as TB caused by strains of Mycobacterium tuberculosis that are resistant to at least isoniazid and rifampicin,13 is increasing worldwide, but much remains to be done. In 2010, only 16% of the TB patients estimated to have MDR-TB were diagnosed and given appropriate treatment.11,12, Routine surveillance of drug resistance must be linked to patient care.

Over the past three years, WHO has been actively encouraging countries to establish continuous TB drug resistance surveillance systems based on routine drug susceptibility testing of all patients, with priority given to patients previously treated, who are at highest risk of developing drug resistance.1820 Although limited laboratory capacity for drug susceptibility testing still represents a major obstacle to the establishment of surveillance systems in low-resource settings, new diagnostic tools such as line probe assays21 and Xpert MTB/RIF,22 combined with greater resources for laboratory strengthening, offer an unprecedented opportunity to scale up surveillance systems worldwide.

In this paper we evaluate the existing information on anti-TB drug resistance surveillance, with an emphasis on data reported in 2007–2010, after the publication of WHO’s fourth global report on anti-TB drug resistance surveillance.8,9 We present a global overview of the extent of the problem of MDR-TB, explore associations between MDR-TB and human immunodeficiency virus (HIV) infection and sex, discuss time trends in drug resistance, and present available data on extensively drug-resistant TB (XDR-TB) – the latter defined as MDR-TB plus resistance to a fluoroquinolone and at least one second-line injectable agent (amikacin, kanamycin or capreomycin).13

Methods

Definitions and data collection

Drug resistance surveillance data are gathered following three main principles: (i) the data are representative of TB cases in the country or geographical setting under study; (ii) drug resistance among new TB cases is examined separately from drug resistance among previously treated TB cases; and (iii) laboratory methods for drug susceptibility testing are selected from among those that are recommended by WHO, with quality assurance for all laboratory processes conducted in cooperation with a partner supranational reference laboratory from the global network of 29 such laboratories.18,19,23,24

Drug resistance surveillance data are collected separately for new (previously untreated) and previously treated TB cases25 via special surveys or continuous surveillance. Special surveys measure drug resistance among a representative sample of notified cases of smear-positive pulmonary TB; continuous surveillance systems are based on routine diagnostic drug susceptibility testing in all bacteriologically-confirmed TB patients. Aggregated data from special surveys are collected through a standard data collection form, whereas continuous surveillance data are captured through "WHO[’s ] global TB data collection system”.26 WHO ascertains whether survey and continuous surveillance data meet quality and representativeness standards through criteria detailed elsewhere.10 The main indicator reported to estimate the frequency of MDR-TB is the proportion of confirmed TB cases with resistance to rifampicin and isoniazid. Data on resistance to any fluoroquinolone and second-line injectable agent among confirmed cases of MDR-TB are used to estimate the frequency of XDR-TB.

Data description, analysis and trends

The proportions of new and previously treated TB cases with MDR-TB and the proportion of MDR-TB cases with XDR-TB were calculated using the latest available national and subnational data. To derive global estimates for these indicators and to investigate the association between MDR-TB and HIV infection and sex, individual-level analyses were conducted using random-effects or robust standard errors logistic regression models to account for the clustering effect at the level of a country or territory. We used the I2 index27 to assess heterogeneity in country-level odds ratios (OR) before we combined these to obtain a pooled estimate. STATA version 11 (StataCorp. LP, College Station, United States of America) was used for all analyses.

Time trends in MDR-TB rates (annual number of new cases per 100 000 population)28 between 1994 and 2010 were calculated by multiplying the new TB case notification rates reported annually to WHO11 by the reported frequency of MDR-TB among new TB cases in the same setting and year. Exponential lines were fitted and the annual percentage change in the rate of MDR-TB was calculated for settings where anti-TB drug resistance had been measured in at least three different years.

Results

Since the launch of the Global Project on Anti-tuberculosis Drug Resistance Surveillance in 1994, drug resistance data have been systematically collected and analysed from 127 countries, or 66% of WHO’s 193 Member States. This includes 64 countries that have continuous surveillance systems based on routine diagnostic drug susceptibility testing of all patients. The remaining 63 countries have relied on special surveys of representative samples of patients. Of the 127 countries with surveillance information, 56 have data from one year only, 20 from two years, and 51 from three or more years (Fig. 1).

Fig. 1. Number of country–year data points for drug resistance surveillance, 1994–2010
Fig. 1. Number of country–year data points for drug resistance surveillance, 1994–2010

Most recent data, 2007–2010

Between 2007 and 2010, resistance to first-line anti-TB drugs was reported from 80 countries and 8 territories, 72 of which provided data from continuous surveillance and 16 from special surveys (Table 1, available at: http://www.who.int/bulletin/volumes/90/2/11-092585). Almost all countries (82/88, or 93%) reported nationwide data. Bangladesh (14 districts covering a population of 30 million), the Plurinational State of Bolivia, Chile, Colombia, El Salvador, Fiji, Kazakhstan, Lebanon, Mongolia, the Republic of Moldova, Rwanda and Sri Lanka provided continuous surveillance data on previously treated but not new TB cases. Subnational data were reported from Bangladesh, the Central African Republic, Indonesia, the Russian Federation (12 oblasts [administrative regions] and republics), Tajikistan and Uganda.

The proportion of new TB cases reported as showing multidrug resistance in these years ranged from 0% to 28.9%. Proportions exceeding 12% (in countries reporting more than 10 MDR-TB cases) were documented in Belarus (25.7%), Estonia (18.3%), several oblasts of the Russian Federation (with Murmansk having the highest level, 28.9%) and Tajikistan (Dushanbe city and Rudaki district, 16.5%).

The proportion of previously treated cases having MDR-TB ranged from 0% to 65.1%. Countries or subnational areas with proportions exceeding 50% included Belarus (60.2%), Lithuania (51.5%), the Republic of Moldova (65.1%), five oblasts of the Russian Federation, and Tajikistan (Dushanbe city and Rudaki district, 61.6%) (Table 1).

The largest country that conducted a nationwide survey in the reporting period was China, where 5.7% of new TB cases and 25.6% of previously treated cases were found to have multidrug resistance (Table 1).

Surveillance data on XDR-TB were reported from 38 countries and 3 territories, 34 of which routinely test all patients with MDR-TB for second-line anti-TB drug resistance. Only 6 out of 41 (15%) countries and territories reported more than 10 cases of XDR-TB; the proportion of MDR-TB cases that were extensively drug-resistant exceeded 10% in Estonia (19.7%), Latvia (15.1%), South Africa (10.5%) and Tajikistan (Dushanbe city and Rudaki district, 21.0%) (Table 2, available at: http://www.who.int/bulletin/volumes/90/2/11-092585).

Overall data, 1994–2010

The proportions of new and previously treated TB cases in the world that were multidrug resistant are shown in Fig. 2 and Fig. 3, respectively. Overall, when data from all countries and territories were combined, the global proportions of new and previously treated TB cases showing multidrug resistance were 3.4% (95% CI: 1.9–5.0) and 19.8% (95% CI: 14.4–25.1), respectively. Regional level estimates of the proportion of cases with MDR-TB are shown in Table 3.

Fig. 2. Distribution of percentage of new tuberculosis cases with multidrug-resistant tuberculosis (MDR-TB)
Fig. 2. Distribution of percentage of new tuberculosis cases with multidrug-resistant tuberculosis (MDR-TB)
Note: Showing latest available data, 1994–2010. Reported data for the Democratic Republic of the Congo, Luxembourg, New Caledonia and the Solomon Islands are not disaggregated by new and previously treated tuberculosis cases and are therefore not shown.
Fig. 3. Percentage of previously treated tuberculosis patients with multidrug-resistant tuberculosis (MDR-TB)
Fig. 3. Percentage of previously treated tuberculosis patients with multidrug-resistant tuberculosis (MDR-TB)
Note: Showing latest available data, 1994–2010. Reported data for the Democratic Republic of the Congo, Luxembourg, New Caledonia and the Solomon Islands are not disaggregated by new and previously treated tuberculosis cases and are therefore not shown.

XDR-TB has been identified in 77 countries globally, and 57 countries and 3 territories were able to report representative data from continuous surveillance or special surveys on the proportion of XDR-TB cases among MDR-TB cases. Combined data from all countries and territories showed that the proportion of MDR-TB cases with extensive drug resistance was 9.4% (95% CI: 7.4–11.6).

Risk factors

When data from 17 countries and 1 territory that reported drug resistance data stratified by HIV status were combined, the odds of having MDR-TB among HIV-positive cases were found to be 40% higher than among HIV-negative cases (pooled odds ratio, OR: 1.4; 95% CI: 0.7–3.0; OR consistent across countries, I2 = 23.2%; P-value = 0.19), but the difference was not significant. Thus, no association was noted between the presence of MDR-TB and HIV status.

A total of 58 countries and 2 special territories reported drug resistance surveillance data disaggregated by sex. Overall, when data from these settings were combined, the odds of having MDR-TB were found to be 10% higher among females than males (OR: 1.1; 95% CI: 0.8–1.4; OR heterogeneous across countries, I2 = 32.9%; P-value = 0.009), but the difference was not significant. Thus, no association was noted between the presence of MDR-TB and the sex of the patient.

Time trends

Data on time trends in drug resistance were available from 71 countries and 751 country-year data points. Selected data to illustrate the diversity of trends in TB and MDR-TB worldwide are presented in Fig. 4. In a first group of countries, composed of Botswana, Peru and the Republic of Korea, the estimated notification rate of MDR-TB is increasing (+10.9%, +19.4% and +4.3% per year, respectively). In these countries, trends in notifications of new TB cases vary, with a clear increase in the Republic of Korea (+7.4% per year), a rather stable trend in Botswana (+0.3% per year) and a clear decline in Peru (−3.3% per year). A second group is composed of three Russian oblasts where TB notification rates are stable or decreasing. Although in these oblasts MDR-TB rates were on the rise until around 2005–2006, they have subsequently been falling in all three settings. In a third group of countries, composed of Estonia, Latvia and the United States, surveillance data suggest that both TB and MDR-TB rates have been falling for more than a decade. In the United States the rate of MDR-TB is falling even more quickly than the TB case notification rate.

Fig. 4. Time trends in population rates of new cases of tuberculosis (TB) and new cases of multidrug-resistant tuberculosis (MDR-TB)
Fig. 4. Time trends in population rates of new cases of tuberculosis (TB) and new cases of multidrug-resistant tuberculosis (MDR-TB)
Note: Data are for new TB cases, except for those of the United States of America, whose data are for combined new and previously treated TB cases.

Discussion

In 2007–2010, several countries provided drug resistance surveillance data generated from continuous surveillance systems rather than special surveys, a change from previous reports.8,9 Of particular interest is a group of 12 countries that have succeeded in establishing continuous surveillance systems for previously treated TB cases: Bangladesh (14 districts covering a population of 30 million), the Plurinational State of Bolivia, Chile, Colombia, El Salvador, Fiji, Kazakhstan, Lebanon, Mongolia, the Republic of Moldova, Rwanda and Sri Lanka. This is the first step towards routine drug susceptibility testing for all TB cases and allows early identification of drug resistance in the population at greatest risk.1820

Available data confirm that eastern Europe and central Asia continue to have the world’s highest proportion of MDR-TB among TB cases. In 2007–2010 the highest proportions ever reported globally were documented in areas of the former Soviet Union; MDR-TB was reported among nearly 30% of new TB cases in the oblast of Murmansk in the Russian Federation and among 65% of previously treated TB cases in the Republic of Moldova. In a few other oblasts of the Russian Federation in the same region, levels of MDR-TB appear to be stabilizing or even decreasing, which confirms that addressing MDR-TB is feasible even in high-burden areas. Unfortunately, large parts of eastern Europe and central Asia still lack representative data. This applies to the whole of Kyrgyzstan and most of Azerbaijan, the Russian Federation, Tajikistan, Turkmenistan, Ukraine and Uzbekistan. With planned and ongoing surveys and improvements in continuous surveillance in these countries, major strides towards improving our understanding of the true burden of drug-resistant TB are expected in the near future.

China conducted its first nationwide survey in 2007. The survey, which confirmed previously published estimates8,9 based on extrapolation from subnational level data, represents a critical step towards addressing MDR-TB in one of the largest TB control programmes in the world. Whereas China has been able to conduct a nationwide survey, India and the Russian Federation – the other two large countries that, with China, contribute to more than 50% of the estimated global burden of MDR-TB – have only produced reliable subnational level data to date. To understand the magnitude of the MDR-TB problem and address it, nationwide surveillance systems should be established in all countries, with greater urgency in the highest burden settings.

Only 34 countries and settings have a system in place to routinely test all patients with MDR-TB for second-line anti-TB drug resistance. These are generally countries with established or emerging economies, as laboratory capacity for second-line drug susceptibility testing in resource-limited settings is still scarce.

The average proportions of MDR-TB cases among diagnosed TB cases detected in this study are consistent with previous reports.8,9 The lack of data on drug-resistant TB in most African countries is still a matter of major concern (Fig. 2 and Fig. 3). This situation should be urgently addressed, especially since the African Region accounts for over 80% of the TB cases among people living with HIV and since higher mortality from MDR-TB and XDR-TB has been documented in HIV-positive patients.29 The availability of new molecular technologies for diagnosing TB and detecting rifampicin resistance, including line probe assays and Xpert MTB/RIF, represent an unprecedented opportunity for countries with severely limited laboratory infrastructure to diagnose drug resistance more easily. Line probe assays permit safer transportation of specimens, require a lower workload than conventional culture and drug susceptibility tests, and reduce to two days the time needed for the diagnosis of MDR-TB.30,31 Xpert MTB/RIF is an automated nucleic acid amplification technology that detects rifampicin resistance in less than two hours. It is very simple to use and requires limited training and biosafety measures.22 A few countries have piloted the use of molecular technologies in drug resistance surveys,30,31 but data from surveys using exclusively those techniques are not yet available. Molecular technologies are expected to contribute substantially to surveillance of drug-resistant TB in low-resource settings in the future.

The analysis of risk factors for MDR-TB showed that the overall risk of harbouring MDR-TB strains is not influenced by sex. The sex distribution of patients with MDR-TB does not differ from that of patients with drug-susceptible TB. This finding is not surprising, since MDR-TB is a form of TB and has similar risk factors. Countries where an association is documented should be investigating the possible reasons.

Although an association between HIV infection and MDR-TB has been widely documented in hospital outbreaks of drug-resistant TB among people living with HIV,3234 the population-based data gathered to date suggest that the relationship between multidrug resistance and HIV infection is not consistent across settings (although the available data are limited to a few countries). In addition, HIV status is unknown for large proportions of patients in these cohorts. Countries are still experiencing great difficulties in incorporating HIV testing in drug resistance surveys, as this requires strong collaboration between HIV and TB control programmes. Understanding the relationship between HIV infection and drug-resistant TB at the population level is critical to identify high-risk groups in need of additional support.

Trend analysis suggests that MDR-TB can be controlled once bold policy decisions are put into practice and the correct prevention and control measures are implemented. This is illustrated by recent findings reported from selected oblasts in the Russian Federation, where MDR-TB has been recognized as a serious problem since the time of the dissolution of the Soviet Union. In Arkhangelsk, Tomsk and Orel oblasts, TB case notifications are stable or decreasing, and although MDR-TB rates were increasing until 2005–2006, more recent data show a stabilizing (Arkhangelsk) or even declining trend (Tomsk and Orel). These settings, which have been treating many cases with MDR-TB in recent years, show that MDR-TB can be controlled even in places heavily affected by drug resistance. The same can be said for Estonia and Latvia, where TB and MDR-TB have been declining for more than a decade. In the United States, rates of MDR-TB are falling even more quickly than rates of TB. These last three countries have strong control programmes that have succeeded in reducing both susceptible and resistant forms of TB. In contrast, in the Republic of Korea, TB and MDR-TB notifications are both increasing, the latter more rapidly than the former. The diversity of treatment options and case management in the country, particularly in the large private health sector, may be facilitating the development of drug resistance.35 In Botswana TB notification rates have stabilized, whereas in Peru they have declined, in line with previous assessments.36 However, in both countries MDR-TB notification rates are showing a very marked increase.

Conclusion

Following 15 years of intensive effort, high-quality surveillance data on anti-TB drug resistance are available for two thirds of all countries in the world. These data show where MDR-TB rates are highest and demonstrate that in selected settings a proper response can alleviate the problem. At the same time, global trends in rates of MDR-TB remain unclear, largely because national representative data are lacking in many large countries, including India and several African countries. A better understanding of epidemiological trends in drug resistance at the global and national levels can be achieved only through repeated surveys and, ultimately, by establishing continuous surveillance based on routine drug susceptibility testing of all confirmed TB cases, with priority given to previously treated patients. Special studies are also needed to help us better understand the factors conducive to the development and spread of MDR-TB. If properly and intensively implemented and followed by appropriate treatment of all TB patients, new technologies can accelerate the response to the threat of drug resistance, save lives and reduce the burden TB imposes on individuals, households and communities.


Competing interests:

None declared.

References

Share