INT J TUBERC LUNG DIS 19(2):151–156 Q 2015 The Union http://dx.doi.org/10.5588/ijtld.14.0200 E-published ahead of print 16 December 2014
Anti-tuberculosis drug resistance in Bangladesh: reflections from the first nationwide survey S. M. M. Kamal,* Md. A. Hossain,† S. Sultana,‡ V. Begum,§ N. Haque,† Md. J. Ahmed,¶ T. M. A. Rahman,# Md. K. A. Hyder,** Md. S. Hossain,* M. Rahman,¶ R. A. Chowdhury,†† K. J. M. Aung,‡‡ Md. A. Islam,§§ Md. R. Hasan,¶¶ A. Van Deun## *National Tuberculosis Reference Laboratory, National Institute of Diseases of the Chest and Hospital (NIDCH), Dhaka, †National TB Programme, Directorate General of Health Services, Ministry of Health & Family Welfare, Dhaka, ‡National Professional Offices-World Health Organization (NPO-WHO), Dhaka, §NPO-WHO, TB CARE-II, Dhaka, ¶Tuberculosis CARE-II, University Research Company, Dhaka, #Bangladesh Unnayan Parishad (BUP), Dhaka, Bangladesh; **WHO South East Asia Regional Office, New Delhi, India; ††Department of Microbiology, University of Dhaka, Dhaka, ‡‡Damien Foundation, Dhaka, §§BRAC, Dhaka, ¶¶NIDCH, Dhaka, Bangladesh; ##Institute of Tropical Medicine, Antwerp, Belgium SUMMARY O B J E C T I V E : To determine the prevalence of tuberculosis (TB) drug resistance in Bangladesh. D E S I G N : Weighted cluster sampling among smearpositive cases, and standard culture and drug susceptibility testing on solid medium were used. R E S U LT S : Of 1480 patients enrolled during 2011, 12 falsified multidrug-resistant TB (MDR-TB) patients were excluded. Analysis included 1340 cases (90.5% of those enrolled) with valid results and known treatment antecedents. Of 1049 new cases, 12.3% (95%CI 9.3–16.1) had strains resistant to any of the first-line drugs tested, and 1.4% (95%CI 0.7–2.5) were MDR-TB. Among the 291 previously treated cases, this was respectively 43.2% (95%CI 37.1–49.5) and 28.5% (95%CI 23.5–34.1). History of previous anti-tuberculosis treatment was the only predictive factor for first-
line drug resistance (OR 34.9). Among the MDR-TB patients, 19.2% (95%CI 11.3–30.5; exclusively previously treated) also showed resistance to ofloxacin. Resistance to kanamycin was not detected. C O N C L U S I O N : Although MDR-TB prevalence was relatively low, transmission of MDR-TB may be increasing in Bangladesh. MDR-TB with fluoroquinolone resistance is rapidly rising. Integrating the private sector should be made high priority given the excessive proportion of MDR-TB retreatment cases in large cities. TB control programmes and donors should avoid applying undue pressure towards meeting global targets, which can lead to corruption of data even in national surveys. K E Y W O R D S : prevalence; drug resistance; tuberculosis; survey; Bangladesh
BANGLADESH, a developing country in South-East Asia, has a population of over 155 million, making it one of the most densely populated countries in the world. Bangladesh ranks sixth among the high tuberculosis (TB) burden countries. In 2011, the estimated TB burden was as follows: annual incidence of all cases 225 per 100 000 population, incidence of new smear-positive cases 100/100 000, prevalence of all cases 411/100 000 and TB mortality 45/100 000. These rates correspond to 340 000 incident TB cases (all forms), 148 000 new smearpositive cases and 68 000 deaths due to TB.1 In 2011, 148 198 patients (all forms) were notified, of whom 98 948 (67%) were new smear-positive cases. The male:female ratio for new smear-positive cases was 2:1. On a national scale, the treatment success rate in
new sputum smear-positive cases registered during 2010 was 92%, indicating that multidrug-resistant TB (MDR-TB) should not be a major problem in new cases. As expected, the treatment success rate reported among previously treated cases registered in the same year was lower (80%). The full extent of the anti-tuberculosis drug resistance problem in Bangladesh is not known, as no national survey has been conducted before. Damien Foundation (DF), one of the partners implementing the National Tuberculosis Programme (NTP) in mainly rural areas, has reported low and decreasing levels of drug resistance in Greater Mymensingh. In 1995, the prevalence of MDR-TB among new smear-positive cases was 0.7%, while among all cases (new plus previously treated, or
Correspondence to: Armand Van Deun, Institute of Tropical Medicine, Antwerp, Belgium. Fax: (þ32) 3 247 6633. e-mail: [email protected] Article submitted 6 April 2014. Final version accepted 9 September 2014. [A version in French of this article is available from the Editorial Office in Paris and from the Union website www.theunion.org]
152
The International Journal of Tuberculosis and Lung Disease
Figure Flow chart of the study. NTM ¼ non-tuberculous mycobacteria; DST ¼ drug susceptibility testing.
‘combined MDR-TB’) it was 2.5%. In 2001, prevalence had dropped to 0.4% among new cases and 0.7% combined.2 A survey conducted from 2001 to 2003 at the Shyamoli Chest Disease Clinic, a referral centre in the capital city of Dhaka, found a MDR-TB prevalence of 5.5% (combined).3 A small unpublished study conducted in 2006 by the Bangladesh NTP and the National Tuberculosis Reference Laboratory (NTRL) in collaboration with the supranational reference laboratory (SRL) in Antwerp, Belgium, showed that of 96 first-line retreatment regimen (Category 2) failures, 88% had MDR-TB; 14% of the MDR-TB cases were ofloxacin (OFX) resistant. Bangladesh has now completed its first National Tuberculosis Drug Resistance Survey with the goal of improving the efficiency of TB control by strengthening the detection and monitoring of anti-tuberculosis drug resistance levels among TB patients. The objectives were as follows: 1) to determine the prevalence and patterns of first-line anti-tuberculosis drug resistance among newly diagnosed and previously treated sputum-positive cases; and 2) to determine the prevalence and patterns of second-line anti-tuberculosis drug resistance in strains with confirmed resistance to isoniazid (INH, H) and rifampicin (RMP, R), i.e., MDR-TB.
STUDY SUBJECTS AND METHODS The study was cross-sectional, targeting consecutively registered new and previously treated smearpositive TB patients.
Sample size and sampling A cluster sampling strategy was employed. Based on the number of patients recorded by the NTP in 2009 and an estimated 1.2% point prevalence of RMP resistance among new cases, a sample size of 1080 new sputum smear-positive patients was calculated, with a desired precision of 1% at 95% confidence level, taking into account a design effect of 2 and an expected 20% losses. As is usual in these surveys, there was no separate sample size for previously treated smear-positive cases.4 Each cluster included a total of 27 new and all retreatment smear-positive cases registered during the study period (Figure). Enrolled patients were classified as new vs. previously treated cases (treatment for at least 1 month). The latter included failures, relapses and cases who returned after default, as defined in the World Health Organization (WHO) guidelines and adopted by the NTP.5 Enrolment of patients and laboratory procedure All sputum smear-positive patients were eligible for enrolment, including children aged ,15 years. The initial patient interview was followed by the collection and rapid transport of two sputum specimens for culture. Samples were cultured at the NTRL according to standard procedures.6 Sputum samples were concentrated by centrifugation at 3000 x g and washed with distilled water twice. The concentrated deposit was cultured on two slopes of Lowenstein¨ Jensen (LJ) medium with glycerol at 378C for up to 8 weeks. Mycobacterium tuberculosis complex was identified by growth and morphological characteris-
TB drug resistance survey, Bangladesh
Table 1
153
Demographic characteristics and treatment antecedents of enrolled patients New cases (n ¼ 1124) n (%)
Previously treated cases (n ¼ 341) n (%)
Total (n ¼ 1465) n (%)
Sex Male Female
778 (69.2) 346 (30.8)
259 (75.9) 82 (24.1)
1037 (70.8) 428 (29.2)
Age group, years 614 15–24 25–34 35–44 45–54 55–64 765 Unknown
12 238 258 165 164 142 142 3
Place of residence Non-metropolitan area Metropolitan area
(1.1) (21.2) (23.0) (14.7) (14.6) (12.6) (12.6) (0.3)
1031 (91.7) 93 (8.3)
tics, in addition to para-nitrobenzoic acid sensitivity. All strains suspected of being non-tuberculous mycobacteria were forwarded to the Antwerp SRL for confirmation and further identification. Drug susceptibility testing (DST) was performed using the proportion method on LJ medium, using standard critical concentrations and 6 week incubation (0.2 lg/ml INH, 40 lg/ml RMP, 4 lg/ml dihydrostreptomycin (SM) and 2 lg/ml ethambutol [E, EMB]).7 Site monitoring visits were performed before the start and towards the middle of the enrolment period by a team from the NTP, NTRL and partners. Three external technical teams also visited the NTP, NTRL and selected clusters for monitoring and validation of the survey data. All strains found resistant to any drug were forwarded to the SRL for confirmation, together with the next pansusceptible strain. Second-line DST (OFX, kanamycin [KM]) was performed at the SRL for MDR-TB strains only. The SRL also performed genotyping using the mycobacterial interspersed repetitive units-variable number of tandem repeat technique on 12 suspected duplicate strains. Ethical clearance for the survey protocol was obtained from the Bangladesh Medical Research Council, Dhaka, Bangladesh. Data processing and analysis All patient forms and laboratory records were entered, validated and analysed using WHO SDRTB4 software (WHO, Geneva, Switzerland). Prevalence of drug resistance was calculated among cases with available DST results. Only SRL results were considered in case of discordance with the NTRL.
RESULTS Enrolment and specimen collection started in December 2010 and was completed for 1480 patients in
5 52 80 69 49 51 34 1
(1.5) (15.3) (23.5) (20.2) (14.4) (15.0) (10) (0.3)
227 (66.6) 114 (33.4)
17 290 338 234 213 193 176 4
(1.2) (19.8) (23.0) (16.0) (14.5) (13.2) (12.1) (0.3)
1258 (85.9) 207 (14.1)
November 2011. Due to culture/DST failure or lack of information on treatment antecedents, 128 patients were excluded from the analysis. Twelve other new MDR-TB cases with the same resistance profile and belonging to the same two clusters were excluded after confirmation of their multiple inclusion as different patients. Deliberate registration of these falsified cases was confirmed during a site visit. Of the 1343 patients remaining for analysis, 78.1% (n ¼ 1049) were new smear-positive cases and 21.7% (n ¼ 291) had been treated previously. Table 1 summarises the demographic and treatment history data of all 1465 patients with this information recorded after excluding duplicate cases. Among both new and previously treated cases, about 70% of cases were male and 60% were aged ,45 years. Whereas the proportion of total patients from metropolitan areas was determined by the random selection of clusters, it was of note that the proportion of new vs. retreatment cases was reversed, with one third of all retreatment cases enrolled from metropolitan and 90% of new cases enrolled from nonmetropolitan areas. Proficiency testing of NTRL DST The quality of the local DST for INH and RMP was found to be adequate after rechecking 258 TB and 11 non-tuberculous strains at the SRL (sensitivity 98% and 94% and specificity 97% and 95%, respectively, for INH and RMP). Non-tuberculous strains were confirmed to be mostly M. intracellulare or M. chelonae/fortuitum, as pure isolates or mixtures (n = 2) with M. tuberculosis. Prevalence of first-line drug resistance As shown in Table 2, the combined prevalence of any first-line drug resistance was 18.7% (95% confidence interval [CI] 15.2–22.7): 12.3% (95%CI 9.3–16.1)
154
The International Journal of Tuberculosis and Lung Disease
Table 2 Patterns of resistance to first-line anti-tuberculosis drugs (weighted)
Drug resistance pattern
New (n ¼ 1049) % (95%CI)
Previously treated (n ¼ 291) % (95%CI)
Total (n ¼ 1340) % (95%CI)
Susceptible to all drugs
87.7 (84.0–90.7)
56.8 (50.5–62.9)
81.3 (77.3–84.8)
Any drug resistance Any resistance to H Any resistance to R Any resistance to E Any resistance to S Total
5.3 1.6 0.9 9.9 12.3
(3.9–7.0) (0.9–2.9) (0.4–2.1) (7.4–13.0) (9.3–16.1)
35.8 28.9 17.8 33.1 43.2
11.6 7.3 4.4 14.7 18.7
(9.0–14.8) (5.2–10.1) (3.2–6.0) (11.9–18.0) (15.2–22.7)
1.4 0.2 0.2 6.6 8.4
(0.8–2.6) (0.0–1.0) (0.0– 0.8) (4.4–9.8) (5.9–11.9)
2.5 (1.2–5.2) 0.4 (0.0–2.7) 0 7.1 (4.8–10.3) 10.0 (7.3–13.5)
1.6 0.3 0.2 6.7 8.7
(1.0–2.6) (0.0–1.1) (0.0–0.6) (4.8–9.3) (6.6–11.5)
Multidrug resistance HR HRE HRS HRES Total
0.4 0.0 0.4 0.5 1.4
(0.1–1.2) (0.0–0.7) (0.2–1.1) (0.2–1.3) (0.7–2.5)
4.3 3.0 7.0 14.1 28.5
1.2 0.7 1.8 3.3 7.0
(0.6–2.4) (0.4–1.3) (1.1–2.8) (2.3–4.9) (5.0–9.8)
Poly drug resistance* HE HS ES HES RE RS RES Total
0.1 (0.0–0.8) 2.4 (1.5–3.7) 0 0 0 0 0 2.5 (1.6–3.9)
Monoresistance Monoresistance Monoresistance Monoresistance Monoresistance Total
to to to to
H R E S
(29.9–42.0) (23.9–34.4) (13.5–23.3) (27.1–40.0) (37.1–49.5)
(2.2–8.2) (1.8–5.0) (5.1–10.0) (9.9–19.7) (23.5–34.1)
0 4.0 (2.2–7.2) 0.0 (–) 0.7 (0.2–2.9) 0 0 0 4.7 (2.6–8.5)
0.0 (0.0–0.5) 2.7 (1.9–3.9) 0 0.2 (0.0–0.6) 0 0 0 3.0 (2.1–4.2)
* Defined as resistance to more than one drug, except for multidrug resistance. CI ¼ confidence interval; H ¼ isoniazid; R ¼ rifampicin; E ¼ ethambutol; S ¼ streptomycin.
among new and 43.2% (95%CI 37.1–49.5) among previously treated cases. Any resistance to INH was respectively 11.6% (95%CI 9.0–14.8), 5.3% (95%CI 3.9–7.0) and 35.8% (95%CI 29.9–42.0), and to RMP it was 7.3% (95%CI 5.2–10.1), 1.6% (95%CI 0.9–2.9) and 28.9% (95%CI 23.9–34.4). Any resistance to EMB was the least prevalent and to SM the most prevalent (details shown in Table 2). The combined prevalence of MDR-TB was 7% (95%CI 5.0–9.8): it was 1.4% (95%CI 0.7–2.5) among new and 28.5% (95%CI 25.3–34.1) among previously treated cases. At respectively 0.3% (95%CI 0.0–1.1), 0.2% (95%CI 0.0–1.0) and 0.4% (95%CI 0.0–2.7), RMP monoresistance was uncommon; INH monoresistance was also infrequent, at 1.6% (95%CI 1.0–2.6), 1.4% (95%CI 0.8–2.6) and 2.5% (95%CI 1.2–5.2) respectively. Among the combined resistance profiles excluding MDR-TB (i.e., polydrug resistance), only combined INHþSM was more frequent (2.7% combined, 95%CI 1.9–3.9; 2.4%, 95%CI 1.5–3.7 among new and 4.0%, 95%CI 2.2–7.2 among previously treated). Further details can be found in Table 2. Table 3 examines MDR-TB prevalence by sex, age group, treatment antecedents and place of residence of all 1340 cases included in the resistance status analysis (new and previously treated). The odds ratio
(OR) was exactly 1 when comparing males with females. For age, the only significant difference observed was between those aged ,25 years with the highest (9.8%) MDR-TB prevalence, and the next age group 25–34 years (P ¼ 0.024 in univariate analysis); however, the lowest prevalence was recorded among the elderly (4.7%). Comparing prevalences among those aged ,25 years with all other patients, the difference was borderline significant (P ¼ 0.05, data not shown). In univariate analysis, MDR-TB prevalence was also significantly higher among previously treated cases and in patients from metropolitan areas (defined as the divisional capital cities). When testing parameters found to be significant using a multivariate logistic regression model, the youngest age group still had a significantly higher OR. Furthermore, the association with previous treatment, but not metropolitan residence, remained highly significant, at OR 34.9. Pattern of resistance to second-line drugs No KM resistance or extensively drug-resistant TB (XDR-TB) was detected among the 74 MDR-TB strains tested. Resistance to OFX was observed among 19.2% (95%CI CI 11.3–30.5) of MDR-TB cases overall. These were all previously treated cases,
155
TB drug resistance survey, Bangladesh
Table 3
Risk factors for MDR-TB Univariate
Multivariate
Tested n
MDR-TB %
Sex Male Female
945 395
6.9 7.2
Reference 1.0 (0.6–1.9)
Age group, years 624 25–34 35–44 45–54 55–64 765 Unknown
290 318 215 189 173 151 4
9.8 5.9 6.4 7.9 6.6 4.7 0
Reference 0.6 (0.4–0.9) 0.6 (0.3–1.4) 0.8 (0.4–1.6) 0.6 (0.3–1.2) 0.5 (0.2–1.1) 1
History of treatment New cases Previously treated cases
1049 291
1.4 28.5
Reference 29.1 (15.9–53.0)
0
Reference 34.9 (18.5–65.8)
0
Place of residence Non-metropolitan area Metropolitan area
1156 184
5.6 13.8
Reference 2.5 (1.4–4.6)
0.003
Reference 0.7 (0.4–1.2)
0.193
OR (95%CI)
P value
OR (95%CI)
P value
0.87
0.024 0.252 0.517 0.168 0.072
0.4 0.3 0.7 0.4 0.5 1
(0.2–0.8) (0.1–0.8) (0.3–1.4 (0.2–0.8) (0.2–1.4)
0.012 0.018 0.258 0.015 0.178 —
MDR-TB ¼ multidrug-resistant tuberculosis; OR ¼ odds ratio; CI ¼ confidence interval.
with prevalence of OFX resistance in this subset reaching 23.8% (95%CI 13.0–39.7). Subsets of previously treated cases MDR-TB was observed among 21.1% of the 191 relapses, 63.3% of the 55 failures, 13.8% of the 32 defaulters and 40.8% of the 13 other retreatment cases. Using relapses as reference, failure cases were observed to be at significantly higher risk of MDR-TB (OR 6.4, 95%CI 3.1–13.3), while the risk was lower (but not significantly) among defaulters (OR 0.6, 95%CI 0.2–2.0) and other retreatment cases (OR 2.6, 95%CI 0.5–13.2).
DISCUSSION We report the results of the first nationwide survey of TB drug resistance conducted in Bangladesh following standard WHO guidelines. Compared to other countries in the region, such as India, Nepal or Myanmar, the prevalence of any resistance, any RMP resistance (1.6%) or MDR-TB (1.4%) among new cases was slightly lower.8 Globally, Bangladesh can also still be classified as a low MDR-TB prevalence country. In analogy with the situation in some African countries, this might be due to the late introduction of a regimen with RMP throughout, which was adopted by the NTP only in 2004 after 10 years of programmatically well controlled use of the 8-month thioacetazonecontaining regimen (2EHRZ/6HT)*.9,10 A significantly higher prevalence (9.8%) of combined MDR-TB was detected in the youngest age group (age ,25 years) compared to most other groups * Z ¼ pyrazinamide; T ¼ thioacetazone. Numbers before the letters indicate the duration in months of the phase of treatment.
or overall (6.3%). It has been shown that, in good TB control programmes, most of the resistance observed in previously treated cases represents primary resistance that was already there initially, mixed with infrequent resistance acquired during treatment.4,11,12 The two-fold higher prevalence of combined MDR-TB among the youngest compared to the oldest age group should thus be interpreted as a sign of recent spread of MDR-TB, as TB due to recent transmission compared to disease reactivation is more frequent among the young. This correlates with anecdotal reports from the DF Bangladesh-implemented NTP areas, where new cases included as supposedly drug-susceptible controls in rapid DST runs have too often turned out to be MDR-TB in recent years, considering the very low MDR-TB prevalence documented in these areas.2 All such patients were young adults, returning to rural homes from industrial areas surrounding the capital following anti-tuberculosis treatment. These industries have seen enormous expansion over the last 10–15 years, and the living conditions of their workers are characterised by extreme overcrowding. The NTP has taken steps to organise detection and treatment in collaboration with the factories, but given the number this is a gigantic challenge. However, proper TB control in these areas is essential to stop MDR-TB from spreading all over the country. It is also not inconceivable that these industrial area workers would be more likely to self-medicate. Private pharmacies are abundant, drugs are cheap, and while RMP is rarely found in private pharmacies in Bangladesh, fluoroquinolones are widely used. With RMP resistance and MDR-TB spreading or acquired due to irregular self-medication, fluoroquinolone resistance would start to appear due to the same mechanisms. This might explain the already high
156
The International Journal of Tuberculosis and Lung Disease
prevalence of resistance to OFX observed in our survey among previously treated MDR-TB cases (24%). This is an alarmingly rapid increase compared to the 14% detected among 96 chronic cases tested only a few years earlier at the SRL. Since the start of the revised NTP in 1993, treatment in the private sector has greatly reduced. A significant reduction in combined drug resistance due to the reduction in the number of cases in need of retreatment with proper NTP coverage has therefore been reported in DF-assisted districts in Bangladesh.2 However, the NTP does not seem to have had a positive impact as yet in metropolitan areas, where it was introduced late and still has poor coverage. This is suggested by our observation of a higher prevalence overall of MDR-TB in the metropolitan areas included in the survey, associated with the enrolment of mainly retreatment cases. The survey indicates that there is an urgent need for Bangladesh to scale up its DOTS policy, with higher coverage particularly in metropolitan areas, through better integration with the private sector. This also indirectly suggests that, although sampling did not include private-for-profit health facilities, the impact of the private sector may nevertheless be visible, particularly looking at the combined resistance rates among both new and previously treated cases. Our study has one major limitation. The prevalence of MDR-TB documented in our survey may have been underestimated, as almost half of the new MDR-TB cases (all from two clusters), had to be excluded after being proven to represent false cases by fingerprinting. However, fingerprinting could not be performed for all cases, and we cannot exclude the possibility that some new susceptible cases included in the analysis may also have been falsified. During a site visit, excessive pressure from the authorities on reaching detection targets was found to be at the origin of the falsified cases. The ultimate cause, however, is the obligation felt by NTPs and principal recipients of support from The Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund; Geneva, Switzerland) to obtain high ratings by meeting WHO and Global Fund targets, which are not always realistic. Our study shows that this can corrupt TB control to the point where we no longer know what to believe, even when the data are generated by an expensive, otherwise meticulously conducted survey.
CONCLUSION While still at a relatively low level, overall drug resistance and MDR-TB prevalence may be increasing in Bangladesh due to recent transmission;
fluoroquinolone resistance among MDR-TB cases has also reached alarming rates over the last few years. Metropolitan areas are not well controlled by the NTP, as shown by the excessive proportion of retreatment cases with MDR-TB. NTPs, their partners and major donors should avoid putting excessive pressure on staff towards meeting case detection and treatment success targets. Acknowledgements The authors thank The Global Fund (Geneva, Switzerland), United States Agency for International Development (Washington DC, USA) and the Government of Bangladesh (Dhaka, Bangladesh) for the financial support obtained to conduct the study. We are also grateful to M Zignol, World Health Organization Geneva, for his crucial support with the development of the survey protocol and data analysis. Conflicts of interest: none declared.
References 1 World Health Organization. Global tuberculosis report 2012. WHO/HTM/TB/2012.6. Geneva, Switzerland: WHO, 2012. 2 Van Deun A, Salim A H, Daru P, et al. Drug resistance monitoring: combined rates may be the best indicator of programme performance. Int J Tuberc Lung Dis 2004; 8: 23– 30. 3 Zaman K, Rahim Z, Yunus M, et al. Drug resistance of Mycobacterium tuberculosis in selected urban and rural areas of Bangladesh. Scand J Infect Dis 2005; 37: 21–26. 4 World Health Organization. Guidelines for surveillance of drug resistance in tuberculosis. 4th ed. WHO/HTM/TB/2009.422. Geneva, Switzerland: WHO, 2009. 5 World Health Organization. Treatment of tuberculosis: guidelines for national programmes. 4th ed. WHO/HTM/TB/ 2009.420. Geneva, Switzerland: WHO, 2010. 6 Petroff S A. A new and rapid method for the isolation and cultivation of tubercle bacilli directly from the sputum and feces. J Exp Med 1915; 21: 38–42. 7 Canetti G, Fox W, Khomenko A, Mitchison D A, Rist N, Smelev N A. Advances in techniques of testing mycobacterial drug sensitivity, and the use of sensitivity tests in tuberculosis control programmes. Bull World Health Organ 1969; 41: 21–43. 8 World Health Organization. The WHO/IUATLD Global Project on Anti-Tuberculosis Drug Resistance Surveillance. Anti-tuberculosis drug resistance in the world. Report No. 4. WHO/HTM/TB/2008.394. Geneva, Switzerland: WHO, 2008; 9 Affolabi D, Adjagba O A, Tanimomo-Kledjo B, Gninafon M, Anagonou S Y, Portaels F. Anti-tuberculosis drug resistance among new and previously treated pulmonary tuberculosis patients in Cotonou, Benin. Int J Tuberc Lung Dis 2007; 11: 1221–1224. 10 Chonde T M, Basra D, Mfinanga S G M, et al. National antituberculosis drug resistance study in Tanzania. Int J Tuberc Lung Dis 2010; 14: 967–972. 11 Van Deun A, Hamid Salim A, Rigouts L, Rahman M, Fissette K, Portaels F. Evaluation of tuberculosis control by periodic or routine susceptibility testing in previously treated cases. Int J Tuberc Lung Dis 2001; 5: 329–338. 12 van Rie A, Warren R, Richardson M, et al. Classification of drug-resistant tuberculosis in an epidemic area. Lancet 2000; 356: 22–25.
TB drug resistance survey, Bangladesh
i
RESUME O B J E C T I F : D e´ terminer la pr e´ valence de la pharmacor e´ sistance de la tuberculose (TB) au Bangladesh. S C H E´ M A : Echantillonnage en grappes pond´er´e parmi des cas de TB a` frottis positif. Culture standard et test de pharmacosensibilit´e en milieu solide. R E´ S U L T A T S : Sur 1480 patients enr ol ˆ e´ s pendant l’ann e´ e 2011, 12 faux patients avec une TB multir´esistante (TB-MDR) ont e´ t´e exclus. L’analyse a inclus 1340 cas (90,5% des cas enrol´ ˆ es) avec des re´ sultats valides et des ante´ ce´ dents de traitement connus. Sur 1049 nouveaux cas, 12,3% (IC95% 9,3– 16,1) avaient des souches r´esistant a` au moins un des m´edicaments de premi`ere ligne test´es et 1,4% (IC95% 0,7–2,5) se sont av´er´es porteurs d’une TB-MDR. Pour les 291 cas pr´ec´edemment trait´es, les r´esultats e´ taient respectivement de 43,2% (IC95% 37,1–49,5) et 28,5%
(IC95% 23,5–34,1). Un ante´ ce´ dent de traitement antituberculeux pr´ealable e´ tait le seul facteur pr´edictif de r´esistance aux m´edicaments de premi`ere ligne (OR 34,9). Parmi les TB-MDR, 19,2% (IC95% 11,3–30,5 ; exclusivement trait´es pr´ealablement) ont e´ galement d e´ montr e´ une r e´ sistance a` l’ofloxacine. Aucune r´esistance a` la kanamycine n’a e´ t´e d´etect´ee. C O N C L U S I O N : Bien que la pr´evalence de la TB-MDR ait e´ t´e relativement faible, sa transmission pourrait augmenter au Bangladesh. La TB-MDR avec r´esistance a` la fluoroquinolone s’accroit rapidement. Int´egrer le secteur priv´e doit eˆ tre une priorit´e, e´ tant donn´e la proportion e´ lev´ee de retraitement des TB-MDR dans les grandes villes. Les programmes de lutte contre la TB et les donateurs devraient e´ viter d’exercer une pression excessive sur l’atteinte des objectifs mondiaux qui sont susceptibles de fausser mˆeme les enquˆetes. RESUMEN
O B J E T I V O: Determinar la prevalencia de tuberculosis (TB) farmacorresistente en Bangladesh. M E´ T O D O: Se practico ´ un muestreo por conglomerados ponderado de los casos de TB con baciloscopia positiva. Las muestras se sometieron a un cultivo corriente y a las pruebas de sensibilidad a los medicamentos en medio solido. ´ R E S U LT A D O S: De los 1480 pacientes registrados en el 2011, se excluyeron 12 casos con diagnostico ´ errado de TB multidrogorresistente (TB-MDR). Se incluyeron en el ana´lisis 1340 casos (90,5% de los inscritos) cuyos resultados eran va´lidos y de los cuales se conoc´ıan los antecedentes de tratamiento. De los 1049 casos nuevos, el 12,3% presento´ cepas resistentes a alguno de los medicamentos de primera l´ınea estudiados (IC95% 9,3– 16,1) y se observo´ TB-MDR en el 1,4% de los casos (IC95% 0,7–2,5). La frecuencia de alguna resistencia en los 291 casos con antecedente de tratamiento fue 43,2% (IC95% 37,1–49,5) y la frecuencia de TB-MDR fue 28,5% (IC95% 23,5–34,1). El unico ´ factor pronostico ´
observado de resistencia a los medicamentos de primera l´ınea fue el haber recibido un tratamiento antituberculoso previo (OR 34,9). De los casos que presentaron TB-MDR, el 19,2% exhibio´ adema´s resistencia a ofloxacino (IC95% 11,3–30,5; exclusivamente en casos con tratamiento previo). No se detecto´ ninguna cepa resistente a kanamicina. ´ N: Si bien el estudio puso en evidencia una CONCLUSIO prevalencia relativamente baja de TB-MDR, su transmision ´ puede estar progresando en Bangladesh. Los casos de TB-MDR con resistencia a las fluoroquinolonas aumentan de manera ra´pida. Es necesario atribuir una alta prioridad a la integracion ´ del sector privado, dada la enorme proporcion ´ de casos de TB-MDR en retratamiento que se diagnostican en las grandes ciudades. Los programas contra la TB y los donantes deben evitar ejercer una presion ´ excesiva para obtener el cumplimiento de las metas mundiales, pues esta coercion ´ puede incluso viciar los resultados de las encuestas.
Anti-tuberculosis drug resistance in Bangladesh: reflections from the first nationwide survey S. M. M. Kamal,* Md. A. Hossain,† S. Sultana,‡ V. Begum,§ N. Haque,† Md. J. Ahmed,¶ T. M. A. Rahman,# Md. K. A. Hyder,** Md. S. Hossain,* M. Rahman,¶ R. A. Chowdhury,†† K. J. M. Aung,‡‡ Md. A. Islam,§§ Md. R. Hasan,¶¶ A. Van Deun## *National Tuberculosis Reference Laboratory, National Institute of Diseases of the Chest and Hospital (NIDCH), Dhaka, †National TB Programme, Directorate General of Health Services, Ministry of Health & Family Welfare, Dhaka, ‡National Professional Offices-World Health Organization (NPO-WHO), Dhaka, §NPO-WHO, TB CARE-II, Dhaka, ¶Tuberculosis CARE-II, University Research Company, Dhaka, #Bangladesh Unnayan Parishad (BUP), Dhaka, Bangladesh; **WHO South East Asia Regional Office, New Delhi, India; ††Department of Microbiology, University of Dhaka, Dhaka, ‡‡Damien Foundation, Dhaka, §§BRAC, Dhaka, ¶¶NIDCH, Dhaka, Bangladesh; ##Institute of Tropical Medicine, Antwerp, Belgium SUMMARY O B J E C T I V E : To determine the prevalence of tuberculosis (TB) drug resistance in Bangladesh. D E S I G N : Weighted cluster sampling among smearpositive cases, and standard culture and drug susceptibility testing on solid medium were used. R E S U LT S : Of 1480 patients enrolled during 2011, 12 falsified multidrug-resistant TB (MDR-TB) patients were excluded. Analysis included 1340 cases (90.5% of those enrolled) with valid results and known treatment antecedents. Of 1049 new cases, 12.3% (95%CI 9.3–16.1) had strains resistant to any of the first-line drugs tested, and 1.4% (95%CI 0.7–2.5) were MDR-TB. Among the 291 previously treated cases, this was respectively 43.2% (95%CI 37.1–49.5) and 28.5% (95%CI 23.5–34.1). History of previous anti-tuberculosis treatment was the only predictive factor for first-
line drug resistance (OR 34.9). Among the MDR-TB patients, 19.2% (95%CI 11.3–30.5; exclusively previously treated) also showed resistance to ofloxacin. Resistance to kanamycin was not detected. C O N C L U S I O N : Although MDR-TB prevalence was relatively low, transmission of MDR-TB may be increasing in Bangladesh. MDR-TB with fluoroquinolone resistance is rapidly rising. Integrating the private sector should be made high priority given the excessive proportion of MDR-TB retreatment cases in large cities. TB control programmes and donors should avoid applying undue pressure towards meeting global targets, which can lead to corruption of data even in national surveys. K E Y W O R D S : prevalence; drug resistance; tuberculosis; survey; Bangladesh
BANGLADESH, a developing country in South-East Asia, has a population of over 155 million, making it one of the most densely populated countries in the world. Bangladesh ranks sixth among the high tuberculosis (TB) burden countries. In 2011, the estimated TB burden was as follows: annual incidence of all cases 225 per 100 000 population, incidence of new smear-positive cases 100/100 000, prevalence of all cases 411/100 000 and TB mortality 45/100 000. These rates correspond to 340 000 incident TB cases (all forms), 148 000 new smearpositive cases and 68 000 deaths due to TB.1 In 2011, 148 198 patients (all forms) were notified, of whom 98 948 (67%) were new smear-positive cases. The male:female ratio for new smear-positive cases was 2:1. On a national scale, the treatment success rate in
new sputum smear-positive cases registered during 2010 was 92%, indicating that multidrug-resistant TB (MDR-TB) should not be a major problem in new cases. As expected, the treatment success rate reported among previously treated cases registered in the same year was lower (80%). The full extent of the anti-tuberculosis drug resistance problem in Bangladesh is not known, as no national survey has been conducted before. Damien Foundation (DF), one of the partners implementing the National Tuberculosis Programme (NTP) in mainly rural areas, has reported low and decreasing levels of drug resistance in Greater Mymensingh. In 1995, the prevalence of MDR-TB among new smear-positive cases was 0.7%, while among all cases (new plus previously treated, or
Correspondence to: Armand Van Deun, Institute of Tropical Medicine, Antwerp, Belgium. Fax: (þ32) 3 247 6633. e-mail: [email protected] Article submitted 6 April 2014. Final version accepted 9 September 2014. [A version in French of this article is available from the Editorial Office in Paris and from the Union website www.theunion.org]
152
The International Journal of Tuberculosis and Lung Disease
Figure Flow chart of the study. NTM ¼ non-tuberculous mycobacteria; DST ¼ drug susceptibility testing.
‘combined MDR-TB’) it was 2.5%. In 2001, prevalence had dropped to 0.4% among new cases and 0.7% combined.2 A survey conducted from 2001 to 2003 at the Shyamoli Chest Disease Clinic, a referral centre in the capital city of Dhaka, found a MDR-TB prevalence of 5.5% (combined).3 A small unpublished study conducted in 2006 by the Bangladesh NTP and the National Tuberculosis Reference Laboratory (NTRL) in collaboration with the supranational reference laboratory (SRL) in Antwerp, Belgium, showed that of 96 first-line retreatment regimen (Category 2) failures, 88% had MDR-TB; 14% of the MDR-TB cases were ofloxacin (OFX) resistant. Bangladesh has now completed its first National Tuberculosis Drug Resistance Survey with the goal of improving the efficiency of TB control by strengthening the detection and monitoring of anti-tuberculosis drug resistance levels among TB patients. The objectives were as follows: 1) to determine the prevalence and patterns of first-line anti-tuberculosis drug resistance among newly diagnosed and previously treated sputum-positive cases; and 2) to determine the prevalence and patterns of second-line anti-tuberculosis drug resistance in strains with confirmed resistance to isoniazid (INH, H) and rifampicin (RMP, R), i.e., MDR-TB.
STUDY SUBJECTS AND METHODS The study was cross-sectional, targeting consecutively registered new and previously treated smearpositive TB patients.
Sample size and sampling A cluster sampling strategy was employed. Based on the number of patients recorded by the NTP in 2009 and an estimated 1.2% point prevalence of RMP resistance among new cases, a sample size of 1080 new sputum smear-positive patients was calculated, with a desired precision of 1% at 95% confidence level, taking into account a design effect of 2 and an expected 20% losses. As is usual in these surveys, there was no separate sample size for previously treated smear-positive cases.4 Each cluster included a total of 27 new and all retreatment smear-positive cases registered during the study period (Figure). Enrolled patients were classified as new vs. previously treated cases (treatment for at least 1 month). The latter included failures, relapses and cases who returned after default, as defined in the World Health Organization (WHO) guidelines and adopted by the NTP.5 Enrolment of patients and laboratory procedure All sputum smear-positive patients were eligible for enrolment, including children aged ,15 years. The initial patient interview was followed by the collection and rapid transport of two sputum specimens for culture. Samples were cultured at the NTRL according to standard procedures.6 Sputum samples were concentrated by centrifugation at 3000 x g and washed with distilled water twice. The concentrated deposit was cultured on two slopes of Lowenstein¨ Jensen (LJ) medium with glycerol at 378C for up to 8 weeks. Mycobacterium tuberculosis complex was identified by growth and morphological characteris-
TB drug resistance survey, Bangladesh
Table 1
153
Demographic characteristics and treatment antecedents of enrolled patients New cases (n ¼ 1124) n (%)
Previously treated cases (n ¼ 341) n (%)
Total (n ¼ 1465) n (%)
Sex Male Female
778 (69.2) 346 (30.8)
259 (75.9) 82 (24.1)
1037 (70.8) 428 (29.2)
Age group, years 614 15–24 25–34 35–44 45–54 55–64 765 Unknown
12 238 258 165 164 142 142 3
Place of residence Non-metropolitan area Metropolitan area
(1.1) (21.2) (23.0) (14.7) (14.6) (12.6) (12.6) (0.3)
1031 (91.7) 93 (8.3)
tics, in addition to para-nitrobenzoic acid sensitivity. All strains suspected of being non-tuberculous mycobacteria were forwarded to the Antwerp SRL for confirmation and further identification. Drug susceptibility testing (DST) was performed using the proportion method on LJ medium, using standard critical concentrations and 6 week incubation (0.2 lg/ml INH, 40 lg/ml RMP, 4 lg/ml dihydrostreptomycin (SM) and 2 lg/ml ethambutol [E, EMB]).7 Site monitoring visits were performed before the start and towards the middle of the enrolment period by a team from the NTP, NTRL and partners. Three external technical teams also visited the NTP, NTRL and selected clusters for monitoring and validation of the survey data. All strains found resistant to any drug were forwarded to the SRL for confirmation, together with the next pansusceptible strain. Second-line DST (OFX, kanamycin [KM]) was performed at the SRL for MDR-TB strains only. The SRL also performed genotyping using the mycobacterial interspersed repetitive units-variable number of tandem repeat technique on 12 suspected duplicate strains. Ethical clearance for the survey protocol was obtained from the Bangladesh Medical Research Council, Dhaka, Bangladesh. Data processing and analysis All patient forms and laboratory records were entered, validated and analysed using WHO SDRTB4 software (WHO, Geneva, Switzerland). Prevalence of drug resistance was calculated among cases with available DST results. Only SRL results were considered in case of discordance with the NTRL.
RESULTS Enrolment and specimen collection started in December 2010 and was completed for 1480 patients in
5 52 80 69 49 51 34 1
(1.5) (15.3) (23.5) (20.2) (14.4) (15.0) (10) (0.3)
227 (66.6) 114 (33.4)
17 290 338 234 213 193 176 4
(1.2) (19.8) (23.0) (16.0) (14.5) (13.2) (12.1) (0.3)
1258 (85.9) 207 (14.1)
November 2011. Due to culture/DST failure or lack of information on treatment antecedents, 128 patients were excluded from the analysis. Twelve other new MDR-TB cases with the same resistance profile and belonging to the same two clusters were excluded after confirmation of their multiple inclusion as different patients. Deliberate registration of these falsified cases was confirmed during a site visit. Of the 1343 patients remaining for analysis, 78.1% (n ¼ 1049) were new smear-positive cases and 21.7% (n ¼ 291) had been treated previously. Table 1 summarises the demographic and treatment history data of all 1465 patients with this information recorded after excluding duplicate cases. Among both new and previously treated cases, about 70% of cases were male and 60% were aged ,45 years. Whereas the proportion of total patients from metropolitan areas was determined by the random selection of clusters, it was of note that the proportion of new vs. retreatment cases was reversed, with one third of all retreatment cases enrolled from metropolitan and 90% of new cases enrolled from nonmetropolitan areas. Proficiency testing of NTRL DST The quality of the local DST for INH and RMP was found to be adequate after rechecking 258 TB and 11 non-tuberculous strains at the SRL (sensitivity 98% and 94% and specificity 97% and 95%, respectively, for INH and RMP). Non-tuberculous strains were confirmed to be mostly M. intracellulare or M. chelonae/fortuitum, as pure isolates or mixtures (n = 2) with M. tuberculosis. Prevalence of first-line drug resistance As shown in Table 2, the combined prevalence of any first-line drug resistance was 18.7% (95% confidence interval [CI] 15.2–22.7): 12.3% (95%CI 9.3–16.1)
154
The International Journal of Tuberculosis and Lung Disease
Table 2 Patterns of resistance to first-line anti-tuberculosis drugs (weighted)
Drug resistance pattern
New (n ¼ 1049) % (95%CI)
Previously treated (n ¼ 291) % (95%CI)
Total (n ¼ 1340) % (95%CI)
Susceptible to all drugs
87.7 (84.0–90.7)
56.8 (50.5–62.9)
81.3 (77.3–84.8)
Any drug resistance Any resistance to H Any resistance to R Any resistance to E Any resistance to S Total
5.3 1.6 0.9 9.9 12.3
(3.9–7.0) (0.9–2.9) (0.4–2.1) (7.4–13.0) (9.3–16.1)
35.8 28.9 17.8 33.1 43.2
11.6 7.3 4.4 14.7 18.7
(9.0–14.8) (5.2–10.1) (3.2–6.0) (11.9–18.0) (15.2–22.7)
1.4 0.2 0.2 6.6 8.4
(0.8–2.6) (0.0–1.0) (0.0– 0.8) (4.4–9.8) (5.9–11.9)
2.5 (1.2–5.2) 0.4 (0.0–2.7) 0 7.1 (4.8–10.3) 10.0 (7.3–13.5)
1.6 0.3 0.2 6.7 8.7
(1.0–2.6) (0.0–1.1) (0.0–0.6) (4.8–9.3) (6.6–11.5)
Multidrug resistance HR HRE HRS HRES Total
0.4 0.0 0.4 0.5 1.4
(0.1–1.2) (0.0–0.7) (0.2–1.1) (0.2–1.3) (0.7–2.5)
4.3 3.0 7.0 14.1 28.5
1.2 0.7 1.8 3.3 7.0
(0.6–2.4) (0.4–1.3) (1.1–2.8) (2.3–4.9) (5.0–9.8)
Poly drug resistance* HE HS ES HES RE RS RES Total
0.1 (0.0–0.8) 2.4 (1.5–3.7) 0 0 0 0 0 2.5 (1.6–3.9)
Monoresistance Monoresistance Monoresistance Monoresistance Monoresistance Total
to to to to
H R E S
(29.9–42.0) (23.9–34.4) (13.5–23.3) (27.1–40.0) (37.1–49.5)
(2.2–8.2) (1.8–5.0) (5.1–10.0) (9.9–19.7) (23.5–34.1)
0 4.0 (2.2–7.2) 0.0 (–) 0.7 (0.2–2.9) 0 0 0 4.7 (2.6–8.5)
0.0 (0.0–0.5) 2.7 (1.9–3.9) 0 0.2 (0.0–0.6) 0 0 0 3.0 (2.1–4.2)
* Defined as resistance to more than one drug, except for multidrug resistance. CI ¼ confidence interval; H ¼ isoniazid; R ¼ rifampicin; E ¼ ethambutol; S ¼ streptomycin.
among new and 43.2% (95%CI 37.1–49.5) among previously treated cases. Any resistance to INH was respectively 11.6% (95%CI 9.0–14.8), 5.3% (95%CI 3.9–7.0) and 35.8% (95%CI 29.9–42.0), and to RMP it was 7.3% (95%CI 5.2–10.1), 1.6% (95%CI 0.9–2.9) and 28.9% (95%CI 23.9–34.4). Any resistance to EMB was the least prevalent and to SM the most prevalent (details shown in Table 2). The combined prevalence of MDR-TB was 7% (95%CI 5.0–9.8): it was 1.4% (95%CI 0.7–2.5) among new and 28.5% (95%CI 25.3–34.1) among previously treated cases. At respectively 0.3% (95%CI 0.0–1.1), 0.2% (95%CI 0.0–1.0) and 0.4% (95%CI 0.0–2.7), RMP monoresistance was uncommon; INH monoresistance was also infrequent, at 1.6% (95%CI 1.0–2.6), 1.4% (95%CI 0.8–2.6) and 2.5% (95%CI 1.2–5.2) respectively. Among the combined resistance profiles excluding MDR-TB (i.e., polydrug resistance), only combined INHþSM was more frequent (2.7% combined, 95%CI 1.9–3.9; 2.4%, 95%CI 1.5–3.7 among new and 4.0%, 95%CI 2.2–7.2 among previously treated). Further details can be found in Table 2. Table 3 examines MDR-TB prevalence by sex, age group, treatment antecedents and place of residence of all 1340 cases included in the resistance status analysis (new and previously treated). The odds ratio
(OR) was exactly 1 when comparing males with females. For age, the only significant difference observed was between those aged ,25 years with the highest (9.8%) MDR-TB prevalence, and the next age group 25–34 years (P ¼ 0.024 in univariate analysis); however, the lowest prevalence was recorded among the elderly (4.7%). Comparing prevalences among those aged ,25 years with all other patients, the difference was borderline significant (P ¼ 0.05, data not shown). In univariate analysis, MDR-TB prevalence was also significantly higher among previously treated cases and in patients from metropolitan areas (defined as the divisional capital cities). When testing parameters found to be significant using a multivariate logistic regression model, the youngest age group still had a significantly higher OR. Furthermore, the association with previous treatment, but not metropolitan residence, remained highly significant, at OR 34.9. Pattern of resistance to second-line drugs No KM resistance or extensively drug-resistant TB (XDR-TB) was detected among the 74 MDR-TB strains tested. Resistance to OFX was observed among 19.2% (95%CI CI 11.3–30.5) of MDR-TB cases overall. These were all previously treated cases,
155
TB drug resistance survey, Bangladesh
Table 3
Risk factors for MDR-TB Univariate
Multivariate
Tested n
MDR-TB %
Sex Male Female
945 395
6.9 7.2
Reference 1.0 (0.6–1.9)
Age group, years 624 25–34 35–44 45–54 55–64 765 Unknown
290 318 215 189 173 151 4
9.8 5.9 6.4 7.9 6.6 4.7 0
Reference 0.6 (0.4–0.9) 0.6 (0.3–1.4) 0.8 (0.4–1.6) 0.6 (0.3–1.2) 0.5 (0.2–1.1) 1
History of treatment New cases Previously treated cases
1049 291
1.4 28.5
Reference 29.1 (15.9–53.0)
0
Reference 34.9 (18.5–65.8)
0
Place of residence Non-metropolitan area Metropolitan area
1156 184
5.6 13.8
Reference 2.5 (1.4–4.6)
0.003
Reference 0.7 (0.4–1.2)
0.193
OR (95%CI)
P value
OR (95%CI)
P value
0.87
0.024 0.252 0.517 0.168 0.072
0.4 0.3 0.7 0.4 0.5 1
(0.2–0.8) (0.1–0.8) (0.3–1.4 (0.2–0.8) (0.2–1.4)
0.012 0.018 0.258 0.015 0.178 —
MDR-TB ¼ multidrug-resistant tuberculosis; OR ¼ odds ratio; CI ¼ confidence interval.
with prevalence of OFX resistance in this subset reaching 23.8% (95%CI 13.0–39.7). Subsets of previously treated cases MDR-TB was observed among 21.1% of the 191 relapses, 63.3% of the 55 failures, 13.8% of the 32 defaulters and 40.8% of the 13 other retreatment cases. Using relapses as reference, failure cases were observed to be at significantly higher risk of MDR-TB (OR 6.4, 95%CI 3.1–13.3), while the risk was lower (but not significantly) among defaulters (OR 0.6, 95%CI 0.2–2.0) and other retreatment cases (OR 2.6, 95%CI 0.5–13.2).
DISCUSSION We report the results of the first nationwide survey of TB drug resistance conducted in Bangladesh following standard WHO guidelines. Compared to other countries in the region, such as India, Nepal or Myanmar, the prevalence of any resistance, any RMP resistance (1.6%) or MDR-TB (1.4%) among new cases was slightly lower.8 Globally, Bangladesh can also still be classified as a low MDR-TB prevalence country. In analogy with the situation in some African countries, this might be due to the late introduction of a regimen with RMP throughout, which was adopted by the NTP only in 2004 after 10 years of programmatically well controlled use of the 8-month thioacetazonecontaining regimen (2EHRZ/6HT)*.9,10 A significantly higher prevalence (9.8%) of combined MDR-TB was detected in the youngest age group (age ,25 years) compared to most other groups * Z ¼ pyrazinamide; T ¼ thioacetazone. Numbers before the letters indicate the duration in months of the phase of treatment.
or overall (6.3%). It has been shown that, in good TB control programmes, most of the resistance observed in previously treated cases represents primary resistance that was already there initially, mixed with infrequent resistance acquired during treatment.4,11,12 The two-fold higher prevalence of combined MDR-TB among the youngest compared to the oldest age group should thus be interpreted as a sign of recent spread of MDR-TB, as TB due to recent transmission compared to disease reactivation is more frequent among the young. This correlates with anecdotal reports from the DF Bangladesh-implemented NTP areas, where new cases included as supposedly drug-susceptible controls in rapid DST runs have too often turned out to be MDR-TB in recent years, considering the very low MDR-TB prevalence documented in these areas.2 All such patients were young adults, returning to rural homes from industrial areas surrounding the capital following anti-tuberculosis treatment. These industries have seen enormous expansion over the last 10–15 years, and the living conditions of their workers are characterised by extreme overcrowding. The NTP has taken steps to organise detection and treatment in collaboration with the factories, but given the number this is a gigantic challenge. However, proper TB control in these areas is essential to stop MDR-TB from spreading all over the country. It is also not inconceivable that these industrial area workers would be more likely to self-medicate. Private pharmacies are abundant, drugs are cheap, and while RMP is rarely found in private pharmacies in Bangladesh, fluoroquinolones are widely used. With RMP resistance and MDR-TB spreading or acquired due to irregular self-medication, fluoroquinolone resistance would start to appear due to the same mechanisms. This might explain the already high
156
The International Journal of Tuberculosis and Lung Disease
prevalence of resistance to OFX observed in our survey among previously treated MDR-TB cases (24%). This is an alarmingly rapid increase compared to the 14% detected among 96 chronic cases tested only a few years earlier at the SRL. Since the start of the revised NTP in 1993, treatment in the private sector has greatly reduced. A significant reduction in combined drug resistance due to the reduction in the number of cases in need of retreatment with proper NTP coverage has therefore been reported in DF-assisted districts in Bangladesh.2 However, the NTP does not seem to have had a positive impact as yet in metropolitan areas, where it was introduced late and still has poor coverage. This is suggested by our observation of a higher prevalence overall of MDR-TB in the metropolitan areas included in the survey, associated with the enrolment of mainly retreatment cases. The survey indicates that there is an urgent need for Bangladesh to scale up its DOTS policy, with higher coverage particularly in metropolitan areas, through better integration with the private sector. This also indirectly suggests that, although sampling did not include private-for-profit health facilities, the impact of the private sector may nevertheless be visible, particularly looking at the combined resistance rates among both new and previously treated cases. Our study has one major limitation. The prevalence of MDR-TB documented in our survey may have been underestimated, as almost half of the new MDR-TB cases (all from two clusters), had to be excluded after being proven to represent false cases by fingerprinting. However, fingerprinting could not be performed for all cases, and we cannot exclude the possibility that some new susceptible cases included in the analysis may also have been falsified. During a site visit, excessive pressure from the authorities on reaching detection targets was found to be at the origin of the falsified cases. The ultimate cause, however, is the obligation felt by NTPs and principal recipients of support from The Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund; Geneva, Switzerland) to obtain high ratings by meeting WHO and Global Fund targets, which are not always realistic. Our study shows that this can corrupt TB control to the point where we no longer know what to believe, even when the data are generated by an expensive, otherwise meticulously conducted survey.
CONCLUSION While still at a relatively low level, overall drug resistance and MDR-TB prevalence may be increasing in Bangladesh due to recent transmission;
fluoroquinolone resistance among MDR-TB cases has also reached alarming rates over the last few years. Metropolitan areas are not well controlled by the NTP, as shown by the excessive proportion of retreatment cases with MDR-TB. NTPs, their partners and major donors should avoid putting excessive pressure on staff towards meeting case detection and treatment success targets. Acknowledgements The authors thank The Global Fund (Geneva, Switzerland), United States Agency for International Development (Washington DC, USA) and the Government of Bangladesh (Dhaka, Bangladesh) for the financial support obtained to conduct the study. We are also grateful to M Zignol, World Health Organization Geneva, for his crucial support with the development of the survey protocol and data analysis. Conflicts of interest: none declared.
References 1 World Health Organization. Global tuberculosis report 2012. WHO/HTM/TB/2012.6. Geneva, Switzerland: WHO, 2012. 2 Van Deun A, Salim A H, Daru P, et al. Drug resistance monitoring: combined rates may be the best indicator of programme performance. Int J Tuberc Lung Dis 2004; 8: 23– 30. 3 Zaman K, Rahim Z, Yunus M, et al. Drug resistance of Mycobacterium tuberculosis in selected urban and rural areas of Bangladesh. Scand J Infect Dis 2005; 37: 21–26. 4 World Health Organization. Guidelines for surveillance of drug resistance in tuberculosis. 4th ed. WHO/HTM/TB/2009.422. Geneva, Switzerland: WHO, 2009. 5 World Health Organization. Treatment of tuberculosis: guidelines for national programmes. 4th ed. WHO/HTM/TB/ 2009.420. Geneva, Switzerland: WHO, 2010. 6 Petroff S A. A new and rapid method for the isolation and cultivation of tubercle bacilli directly from the sputum and feces. J Exp Med 1915; 21: 38–42. 7 Canetti G, Fox W, Khomenko A, Mitchison D A, Rist N, Smelev N A. Advances in techniques of testing mycobacterial drug sensitivity, and the use of sensitivity tests in tuberculosis control programmes. Bull World Health Organ 1969; 41: 21–43. 8 World Health Organization. The WHO/IUATLD Global Project on Anti-Tuberculosis Drug Resistance Surveillance. Anti-tuberculosis drug resistance in the world. Report No. 4. WHO/HTM/TB/2008.394. Geneva, Switzerland: WHO, 2008; 9 Affolabi D, Adjagba O A, Tanimomo-Kledjo B, Gninafon M, Anagonou S Y, Portaels F. Anti-tuberculosis drug resistance among new and previously treated pulmonary tuberculosis patients in Cotonou, Benin. Int J Tuberc Lung Dis 2007; 11: 1221–1224. 10 Chonde T M, Basra D, Mfinanga S G M, et al. National antituberculosis drug resistance study in Tanzania. Int J Tuberc Lung Dis 2010; 14: 967–972. 11 Van Deun A, Hamid Salim A, Rigouts L, Rahman M, Fissette K, Portaels F. Evaluation of tuberculosis control by periodic or routine susceptibility testing in previously treated cases. Int J Tuberc Lung Dis 2001; 5: 329–338. 12 van Rie A, Warren R, Richardson M, et al. Classification of drug-resistant tuberculosis in an epidemic area. Lancet 2000; 356: 22–25.
TB drug resistance survey, Bangladesh
i
RESUME O B J E C T I F : D e´ terminer la pr e´ valence de la pharmacor e´ sistance de la tuberculose (TB) au Bangladesh. S C H E´ M A : Echantillonnage en grappes pond´er´e parmi des cas de TB a` frottis positif. Culture standard et test de pharmacosensibilit´e en milieu solide. R E´ S U L T A T S : Sur 1480 patients enr ol ˆ e´ s pendant l’ann e´ e 2011, 12 faux patients avec une TB multir´esistante (TB-MDR) ont e´ t´e exclus. L’analyse a inclus 1340 cas (90,5% des cas enrol´ ˆ es) avec des re´ sultats valides et des ante´ ce´ dents de traitement connus. Sur 1049 nouveaux cas, 12,3% (IC95% 9,3– 16,1) avaient des souches r´esistant a` au moins un des m´edicaments de premi`ere ligne test´es et 1,4% (IC95% 0,7–2,5) se sont av´er´es porteurs d’une TB-MDR. Pour les 291 cas pr´ec´edemment trait´es, les r´esultats e´ taient respectivement de 43,2% (IC95% 37,1–49,5) et 28,5%
(IC95% 23,5–34,1). Un ante´ ce´ dent de traitement antituberculeux pr´ealable e´ tait le seul facteur pr´edictif de r´esistance aux m´edicaments de premi`ere ligne (OR 34,9). Parmi les TB-MDR, 19,2% (IC95% 11,3–30,5 ; exclusivement trait´es pr´ealablement) ont e´ galement d e´ montr e´ une r e´ sistance a` l’ofloxacine. Aucune r´esistance a` la kanamycine n’a e´ t´e d´etect´ee. C O N C L U S I O N : Bien que la pr´evalence de la TB-MDR ait e´ t´e relativement faible, sa transmission pourrait augmenter au Bangladesh. La TB-MDR avec r´esistance a` la fluoroquinolone s’accroit rapidement. Int´egrer le secteur priv´e doit eˆ tre une priorit´e, e´ tant donn´e la proportion e´ lev´ee de retraitement des TB-MDR dans les grandes villes. Les programmes de lutte contre la TB et les donateurs devraient e´ viter d’exercer une pression excessive sur l’atteinte des objectifs mondiaux qui sont susceptibles de fausser mˆeme les enquˆetes. RESUMEN
O B J E T I V O: Determinar la prevalencia de tuberculosis (TB) farmacorresistente en Bangladesh. M E´ T O D O: Se practico ´ un muestreo por conglomerados ponderado de los casos de TB con baciloscopia positiva. Las muestras se sometieron a un cultivo corriente y a las pruebas de sensibilidad a los medicamentos en medio solido. ´ R E S U LT A D O S: De los 1480 pacientes registrados en el 2011, se excluyeron 12 casos con diagnostico ´ errado de TB multidrogorresistente (TB-MDR). Se incluyeron en el ana´lisis 1340 casos (90,5% de los inscritos) cuyos resultados eran va´lidos y de los cuales se conoc´ıan los antecedentes de tratamiento. De los 1049 casos nuevos, el 12,3% presento´ cepas resistentes a alguno de los medicamentos de primera l´ınea estudiados (IC95% 9,3– 16,1) y se observo´ TB-MDR en el 1,4% de los casos (IC95% 0,7–2,5). La frecuencia de alguna resistencia en los 291 casos con antecedente de tratamiento fue 43,2% (IC95% 37,1–49,5) y la frecuencia de TB-MDR fue 28,5% (IC95% 23,5–34,1). El unico ´ factor pronostico ´
observado de resistencia a los medicamentos de primera l´ınea fue el haber recibido un tratamiento antituberculoso previo (OR 34,9). De los casos que presentaron TB-MDR, el 19,2% exhibio´ adema´s resistencia a ofloxacino (IC95% 11,3–30,5; exclusivamente en casos con tratamiento previo). No se detecto´ ninguna cepa resistente a kanamicina. ´ N: Si bien el estudio puso en evidencia una CONCLUSIO prevalencia relativamente baja de TB-MDR, su transmision ´ puede estar progresando en Bangladesh. Los casos de TB-MDR con resistencia a las fluoroquinolonas aumentan de manera ra´pida. Es necesario atribuir una alta prioridad a la integracion ´ del sector privado, dada la enorme proporcion ´ de casos de TB-MDR en retratamiento que se diagnostican en las grandes ciudades. Los programas contra la TB y los donantes deben evitar ejercer una presion ´ excesiva para obtener el cumplimiento de las metas mundiales, pues esta coercion ´ puede incluso viciar los resultados de las encuestas.
