Trends in the annual risk of tuberculous infection in India

INT J TUBERC LUNG DIS 17(3):312–319 © 2013 The Union http://dx.doi.org/10.5588/ijtld.12.0330 E-published ahead of print 14 January 2013 Trends in the annual risk of tuberculous infection in India V. K. Chadha,* R. Sarin,† P. Narang,‡ K. R. John,§ K. K. Chopra,¶ R. Jitendra,# D. K. Mendiratta,‡ V. Vohra,† A. N. Shashidhara,* G. Muniraj,§ P. G. Gopi,** P. Kumar†† * Epidemiology and Research Division, National Tuberculosis Institute, Bangalore, † Lala Ram Sarup Institute of TB and Allied Respiratory Diseases, New Delhi, ‡ Department of Microbiology, Mahatma Gandhi Institute of Medical Sciences, Sevagram, Wardha, § Department of Community Medicine, SRM Medical College, Kanchipura, ¶ New Delhi TB Centre, New Delhi, # Department of Statistics, National Tuberculosis Institute, Bangalore, ** National Institute of Research in Tuberculosis, Chennai, †† National Tuberculosis Institute, Bangalore, India SUMMARY Estimated ARTI rates in different zones varied between 1.1% and 1.9% in Survey I and 0.6% and 1.2% in Survey II. The ARTI declined by respectively 6.1% and 11.7% per year in the north and west zones; no decline was observed in the south and east zones. National level estimates were respectively 1.5% and 1.0%, with a decline of 4.5% per year in the intervening period. C O N C L U S I O N : Although a decline in ARTI was observed in two of the four zones and at national level, the current ARTI of about 1% in three zones suggests that further intensification of TB control activities is required. K E Y W O R D S : tuberculosis; infection; risk; trends; India Twenty-four districts in India. To evaluate trends in annual risk of tuberculous infection (ARTI) in each of four geographically defined zones in the country. S T U D Y D E S I G N : Two rounds of house-based tuberculin surveys were conducted 8–9 years apart among children aged 1–9 years in statistically selected clusters during 2000–2003 and 2009–2010 (Surveys I and II). Altogether, 184 992 children were tested with 1 tuberculin unit (TU) of purified protein derivative (PPD) RT23 with Tween 80 in Survey I and 69 496 children with 2TU dose of PPD in Survey II. The maximum transverse diameter of induration was measured about 72 h after test administration. ARTI was computed from the prevalence of infection estimated using the mirror-image method. R E S U LT S : INDIA CONTINUES to have the highest tuberculosis (TB) burden in the world.1 To address the problem, a National TB Programme (NTP) was implemented in 1962.2 Following a review of the NTP, the Revised National TB Control Programme (RNTCP), adopting the DOTS strategy, was launched in 1997 and extended in phases to achieve complete population coverage by 2006.3,4 To evaluate the impact of the RNTCP, two rounds of tuberculin surveys were carried out during 2000– 2003 and 2009–2010 (Surveys I and II). For this purpose, the country was divided into four geographic zones, north, south, east and west, with the objective of estimating the annual risk of tuberculous infection (ARTI) and its trends in each zone. ARTI denotes the proportion of persons in a community who become (re-) infected within 1 year.5 It depends not only on the incidence of infectious TB cases but also on the efficiency of the programme.6 In all the zones, high rates of treatment success (⩾80%) among diagnosed TB patients have been achieved since the introduc- tion of the RNTCP, and notification rates among new smear-positive TB cases have shown consistent improvement over the years, reaching 50 per 100 000 population in the south, east and west zones and 60/100 000 in the north zone in 2010.4 SETTING: OBJECTIVES: MATERIAL AND METHODS Study population House-based surveys were carried out among children aged 1–9 years. Survey I was planned to estimate the ARTI primarily among children without bacille Calmette-Guérin (BCG) vaccination (referred to as non-vaccinated children).7 However, children with BCG scar (referred to as vaccinated children) encountered during the process of registering nonvaccinated children were also tuberculin-tested. Analyses revealed that the ARTI rates estimated by including vaccinated children were not different from those based exclusively on non-vaccinated children.8,9 Survey II was therefore carried out to estimate ARTI Correspondence to: V K Chadha, Epidemiology and Research Division, National Tuberculosis Institute, 8 Bellary Road, Bangalore 560 003, India. Tel: (+91) 99164 93109. Fax: (+91) 802 344 0952. e-mail: [email protected]; vineet2.chadha@ gmail.com Article submitted 7 May 2012. Final version accepted 17 September 2012. ARTI trends in India among ‘all children’ (including non-vaccinated, vaccinated and those with a doubtful BCG scar). Sampling The sample size for both surveys in each zone was calculated to estimate the prevalence of tuberculosis infection within 10% of true value at a 5% level of significance and with a design effect of 2 to account for cluster sampling. In Survey I, the sample size for non-vaccinated children in each zone was estimated at 11 045, arbitrarily considering the expected prevalence of infection at 10%. In Survey II, the expected prevalence of infection among all children in each zone was estimated considering a decline of 5% per year from the prevalence of infection estimated in Survey I. The sample size thus estimated for each zone was as follows: north 13 600, west 16 800, east 20 000 and south 25 000. In each zone, the estimated sample size was equally allocated to 600 clusters in Survey I and 300 clusters in Survey II. Numbers of clusters were allocated in ratio of population size to six districts selected by population proportion to size (PPS) sampling. Numbers of clusters allocated to individual districts were further allocated to rural and urban areas in proportion to population; clusters were selected by PPS sampling. Field procedures Three visits were made to each cluster for planning, registration and testing, and reading. During planning, community leaders were apprised of the purpose of the surveys. A sketch of the cluster, including hamlets and lanes, was made and the lane to start fieldwork was selected using a random number table. On the day of testing, children were registered on individual data cards, starting from the first house of the selected lane. Contiguous houses were visited in a clockwise direction until the cluster size was achieved. Tuberculin skin tests (TSTs) were performed at a temporary centre set up by trained testers after obtaining informed written consent from the parent/guardian. Each child was injected intradermally on the anterior aspect of the left forearm with 0.1 ml of ready-to-use dilution containing PPD RT23 with Tween 80 (Statens Serum Institute, Copenhagen, Denmark). 1TU dose was used in Survey I and 2TU in Survey II. Each test was recorded as ‘satisfactory’ if it raised a flat pale wheal with clearly visible pits and well demarcated borders, and ‘unsatisfactory’ in case of leakage or subcutaneous injection. Presence/absence of BCG scar was recorded after examining both shoulders; scars not characteristic of BCG were recorded as ‘doubtful’. Children with fever, history of skin rash or antituberculosis treatment during the previous 6 months were excluded from testing. Reactions were read approximately 72 h after test administration. Trained readers identified the margins 313 of induration by careful palpation and measured the maximum transverse diameter of induration in mm. Most reactions in a given zone were read by a single reader; only in the event of his/her absence did a second trained reader perform the reading. Children suspected to have TB (based on symptoms, contact history and TST result) were further investigated and treated at the nearest government health centre. In Survey II, 150 new smear-positive TB patients also underwent a TST in each zone to facilitate a comparison with tuberculin sensitivity patterns among the truly infected children at the time of analysis. Statistical methods Data were double-entered into Epi Info™ 3.5.3 (Centers for Disease Control and Prevention, Atlanta, GA, USA) and verified. Analyses were performed using SPSS 15.0 (Statistical Package for the Social Sciences Inc, Chicago, IL, USA) and R software 2.5.1 (R Software Computing, Vienna, Austria), in addition to an in-house programme written in Microsoft Foxpro version 9.0 (Microsoft, Redwoods, WA, USA). The reaction sizes of all children aged 1–9 years were arrayed in frequency distributions to identify the mode of tuberculous reactions. In the event of failure to identify a clear mode, frequency distributions were plotted and examined for subsets of study population in the following order sequentially until decisions could be made on mode (and also anti-mode, applicable to non-vaccinated children only): 1 Non-vaccinated children aged 1–9 years 2 Non-vaccinated children aged 5–9 years 3 Non-vaccinated children aged 5–9 years of age from urban areas Non-vaccinated children are more likely to portray bimodality due to absence of contamination by BCG-induced tuberculin sensitivity, and also an antimode that separates tuberculosis reactions from cross-reactions. Each of the above subsets has a greater likelihood of throwing up mode/anti-mode than the previous subset, due to a higher prevalence of true tuberculous infection.7–11 In Survey II, in the event of continued uncertainty, the mode of the tuberculous reactions was identified after fitting normal distribution to the frequencies of tuberculin reaction sizes among new smear-positive patients who had undergone a TST. The mode seen in any of the above distributions was applied to all children (tuberculous reactions were assumed to be normally distributed around the same mode, irrespective of age group, BCG vaccination status and presence/absence of TB disease8,9,12–18). Prevalence of infection was estimated using the mirror image (MI) method, in which the proportion of reactions larger than the mode of TB reactions is doubled and added to the proportion at the mode.19 314 The International Journal of Tuberculosis and Lung Disease To reduce the influence of digit preference, estimations were made using frequencies of tuberculin reaction size at each millimetre, as obtained by moving averages (5-point).20 In each zone, prevalence and standard deviation (SD), which measures the spread of data about the mean infection, with standard error (SE), as SD/√c , where c is the number of clusters, were first estimated separately for urban and rural areas. Weighted prevalence of infection for zone (PZ) was obtained as ∑wsps/ ∑ws, where ps is prevalence in urban/rural area and ws is weight corresponding to proportion of population (all age groups) of the total zonal population. The SE for zonal level prevalence was estimated using the pooled variance method.21 Zonal estimates of infection prevalence were pooled to obtain the weighted national-level estimate; the weight corresponded to the proportion of population in the zone of the total population in the country. The SE for national-level prevalence was estimated as follows: √∑wi2si2/(∑wi)2 where wi and si represent the weight and SE for the ith zone. 95% confidence intervals (CIs) were derived as prevalence of infection ± 1.96*SE. ARTI was computed from the estimated prevalence of infection using the following formula:22 Rb+a/2 = 1 – (1 – Pb+a)1/a where b = the median year of birth of the children test-read, a = the mean age of the children, Rb+a/2 = the ARTI at the mid-point in calendar time between the median year of birth and the period of the survey, and Pb+a = the prevalence of infection at the time of the survey. Estimation was also carried out separately for nonvaccinated and vaccinated children by the MI method Table 1 using the same mode obtained as above. In addition, estimation among non-vaccinated children was also carried out by the anti-mode (if observed) method, in which reactions greater than or equal to the antimode are attributed to tuberculous infection.19 This method was not applied to all children or vaccinated children due to the greater likelihood of overlap between true tuberculous reactions and cross-reactions around the anti-mode. For comparisons, the χ2 test with continuity correction was used for test of significance; P < 0.05 was considered significant. Using R software, changes in ARTI rates and CIs were computed by the beta distribution function, with input as ARTI rates, the SEs in Surveys I and II and the intervening time period in years, by simulating the number of response vectors to 50 000 iterations. Analysis was also undertaken by the mixture model using R software and scripts available at www. tbrieder.org.23 As the model did not generate a good fit in either zone, the results of mixture analysis are not presented here. RESULTS The numbers of children registered and satisfactorily test-read in each zone in Surveys I and II are given in Table 1. The number of children for whom data were available for analysis was much larger than the estimated sample size for Survey I, as explained under ‘Study population’. Zone-wise frequency distributions of reaction sizes for ‘all children’ aged 1–9 years did not reveal clear bi-modality in either zone (Figure 1). Distributions for subsets of children that revealed maximum clarity of mode/anti-mode during Survey I are presented in Figure 2. Distributions of reaction sizes among TB patients in Survey II are presented in Figure 3. In the north zone, a mode at the same reaction size as in TB Numbers of children investigated by zone, Surveys I and II North zone Survey I Survey II South zone Survey I Survey II East zone Survey I Survey II West zone Survey I Survey II Children tested No. read of those satisfactorily tested Children registered n Excluded from testing n (%)* Satisfactory n (%)† Unsatisfactory n (%) BCG– n BCG+ n All‡ n (%)§ 55 433 15 175 3728 (6.7) 1524 (10.0) 51 380 (87.7) 13 309 (94.2) 325 342 25 816 4 286 21 869 7 684 48 323 (94.1) 12 535 (94.2) 52 951 25 704 327 (0.6) 2137 (8.3) 52 300 (98.8) 23 343 (90.8) 324 224 17 811 6 201 32 549 15 577 50 533 (96.6) 22 059 (85.8) 44 165 20 969 883 (2.0) 1326 (6.3) 42 836 (97.0) 19 563 (93.3) 446 80 17 861 6 058 19 488 12 678 37 854 (88.4) 19 159 (97.9) 55 366 16 800 3434 (6.2) 540 (3.2) 51 733 (93.4) 16 065 (95.6) 199 195 22 258 3 770 25 114 11 868 48 282 (93.3) 15 743 (98.0) * Proportion excluded from testing of those registered. † Proportion satisfactorily tested of those registered. ‡ Includes non-vaccinated and vaccinated children and children with doubtful BCG scar. § Proportion of all children read of those satisfactorily tested. BCG = bacille Calmette-Guérin; – = negative ; + = positive. Figure 1 Frequency distributions of reaction sizes for all children aged 1–9 years, irrespective of BCG status, rural and urban populations combined. BCG = bacille Calmette-Guérin. Figure 2 Frequency distribution of tuberculin reaction sizes for selected subgroups, Survey I. BCG = bacille Calmette-Guérin. 316 The International Journal of Tuberculosis and Lung Disease Figure 3 Frequency distribution of reaction sizes among smear-positive TB patients by zone, Survey II. TB = tuberculosis. patients and an anti-mode at 13 mm was seen among non-vaccinated children aged 1–9 years (data not shown); none of the subgroups of children in the other zones revealed a mode/anti-mode in Survey II. Based on these distributions, the modes/anti-modes used for estimations are given in Table 2. The estimated prevalence of infection with computed ARTI rates among all children in Surveys I and II and the average per year change in ARTI are preTable 2 Mode and anti-mode used for estimating prevalence of infection by zone, Surveys I and II Mode Anti-mode Zone Survey I Survey II Survey I Survey II North South East West 20 19 20 20 18 20 19 19 14 14 16 15 13 — — — Table 3 sented in Table 3. Considering the mean age of the test-read children and mid-points of surveys in individual zones, the ARTI rates in Survey I corresponded to the year 1998 for the north, south and west zones and 1999 for the east zone, and 2007 for all zones in Survey II. ARTI rates declined by respectively 6% and 11.7% per year in the north and west zones; there was no change in the south and east zones. ARTI rates at the national level declined by 4.5% per year between 1998 and 2007. Infection prevalence estimates were lower in vaccinated children than in non-vaccinated children in most of the surveys (Table 4); the estimated ARTI rates were similar between the two subgroups. DISCUSSION The estimated ARTI rates among all children aged 1– 9 years varied between 1.1% and 1.9% in Survey I Estimated prevalence of infection and ARTI by zone, Surveys I and II Survey I Zone North South East West National level Survey II Prevalence % (95%CI) ARTI % (95%CI) Prevalence % (95%CI) ARTI % (95%CI) Change in ARTI /year % (95%CI) 10.1 (9.1–11.1) 6.1 (5.4–6.7) 6.2 (5.5–7.0) 8.7 (7.7–9.6) 7.8 (7.3–8.2) 1.9 (1.7–2.1) 1.1 (1.0–1.2) 1.2 (1.0–1.3) 1.7 (1.5–1.9) 1.5 (1.4–1.6) 5.9 (4.7–7.0) 5.3 (4.5–6.0) 6.5 (4.8–8.2) 3.0 (2.3–3.7) 5.2 (4.6–5.8) 1.1 (0.8–1.3) 1.0 (0.8–1.1) 1.2 (0.9–1.5) 0.6 (0.4–0.7) 1.0 (0.8–1.1) –6.1 (–8.8 to –3.6 ) –1.0 (–3.0 to +0.8) –0.1 (–3.2 to +3.0) –11.6 (–14.7 to –8.7) – 4.5 (–6.3 to –2.8) ARTI = annual rate of tuberculous infection; CI = confidence interval. ARTI trends in India 317 Table 4 Estimated prevalence of infection and ARTI by BCG scar status Survey I North zone Test read, n MI method Prevalence ARTI AM method Prevalence ARTI South zone Test read, n MI method Prevalence ARTI AM method Prevalence ARTI East zone Test read, n MI method Prevalence ARTI AM method Prevalence ARTI West zone Test read, n MI method Prevalence ARTI AM method Prevalence ARTI National level Test read, n MI method Prevalence ARTI AM method Prevalence ARTI Non-BCGvaccinated % (95%CI) BCGvaccinated % (95%CI) 25 816 21 869 Survey II P value* Non-BCGvaccinated % (95%CI) BCGvaccinated % (95%CI) 4 286 7 684 P value* 10.9 (9.7–12.1) 2.0 (1.8–2.2) 9.5 (8.3–10.7) 1.8 (1.6–2.0) <0.001 6.7 (4.1–9.3) 1.1 (0.7–1.6) 5.4 (4.0–6.8) 1.0 (0.7–1.3) 0.004 11.4 (10.5–12.3) 2.1 (1.9–2.2) —† —† 7.6 (5.8–9.5) 1.3 (1.0–1.6) —† —† 17 811 32 549 6 201 15 577 6.6 (5.6–7.5) 1.2 (1.0–1.3) 5.7 (5.0–6.4) 1.1 (0.9–1.2) <0.001 5.2 (3.9–6.4) 0.9 (0.7–1.2) 5.1 (4.2–5.9) 1.0 (0.8–1.1) 0.67 7.1 (6.4–7.8) 1.2 (1.1–1.4) —† —† —† —† —† 6 058 12 678 17 861 19 488 6.6 (5.7–7.6) 1.2 (1.0–1.4) 6.0 (5.2–6.2) 1.2 (1.1–1.3) 0.02 7.7 (5.0–10.3) 1.3 (0.8–1.8) 5.7 (4.2–7.3) 1.1 (0.8–1.4) <0.001 7.7 (6.8–8.6) 1.4 (1.3–1.6) —† —† —† —† —† 3 770 11 868 22 258 25 114 9.6 (8.5–10.7) 1.8 (1.6–2.0) 7.8 (6.5–9.2) 1.5 (1.2–1.8) <0.001 —† 2.9 (1.7– 4.1) 0.5 (0.3–0.7) 3.0 (2.2–3.7) 0.6 (0.4–0.7) 0.77 9.9 (9.0–10.7) 1.9 (1.7–2.0) —† —† —† —† —† 83 746 99 020 18 400 47 807 5.7 (4.6–6.7) 1.0 (0.8–1.2) 4.8 (4.2–5.4) 0.9 (0.8–1.0) 8.4 (7.9–9.0) 1.5 (1.4–1.6) 7.3 (6.7–7.8) 1.4 (1.3–1.5) 9.0 (8.6–9.5) 1.7 (1.6–1.8) —† <0.001 —† <0.001 —† * P values are for test of significance between the estimated prevalence among children with BCG scar and those without BCG scar. to non-identification of anti-mode, estimates of prevalence of infection and P values were not available. ARTI = annual risk of tuberculosis infection; BCG = bacille Calmette-Guérin; MI = mirror-image; AM = anti-mode method. † Due and 0.6% and 1.2% in Survey II in the different zones. Variable trends in ARTI were observed in the different zones during the intervening period of roughly 8–9 years, which coincided with the expansion phase of the RNTCP.3,4 While significant declines in ARTI rates were estimated for the north and west zones, no such decline was noticed in the south and east zones. This was contrary to expectations, as uniform progress in RNTCP implementation should have reduced transmission rates in all of the zones. It was perhaps easier for the RNTCP to bring down the level of infection from relatively higher levels during Survey I in the west and north zones, where the estimated ARTI rates were respectively about 1.7% and 1.9% compared to respectively 1.1% and 1.2% in the east and south zones. However, this incongruence could also be attributed to limitations of the methods available to segregate true tuberculous infections from cross-reactions and subjectivity in measuring TST reaction size. It may not be naïve to say that while significant declines in the transmission of infection were observed in two zones, similar declines were not observed in the other zones due to limitations of the survey tool. It would also be improbable to segregate the impact of control efforts from improvements in socio-economic conditions, as seen in the significant increase in gross domestic product (GDP) in all the Indian states. Although the surveys were not designed (with respect to sample size) to estimate ARTI and trends by area (rural, urban) or by age group, the ARTI was higher in urban than in rural areas in all the zones in both rounds of surveys. Estimated ARTI rates were similar among the 1–4 year and 5–9 year age groups, 318 The International Journal of Tuberculosis and Lung Disease which may be due to the fact that estimates in the latter group included the exposure experienced in the younger age group, thus masking the actual size of the difference and making it undetectable. The pooled national-level estimates of ARTI were respectively 1.5% and 1.0% in Surveys I and II, with a decline in the intervening period of 4.5% per year. It has been estimated that if at least 70% of incidence cases of smear-positive pulmonary TB are detected every year and 85% of them are cured, ARTI rates would decline by 7–11% per year.23 Declines of 10– 15% per year have been observed under good programme conditions in some Western countries.22,24–26 An important difference between the two surveys was that while 1TU PPD was used during Survey I, its non-availability during Survey II prompted the use of 2TU, which is now recommended as the standard dose.12 Differences between these two doses usually show up in the intermediate range of reaction sizes and are therefore not likely to influence the estimates, particularly those made using the MI method.27 To evaluate the effect of BCG vaccination, if any, estimations were also carried out separately for nonvaccinated and vaccinated children using the MI method. Estimated ARTI rates were found to be similar between vaccinated and non-vaccinated children in most of these surveys. However, the prevalence of infection was generally lower among the former due to their lower mean age. This was due to a decline in the proportions of vaccinated children with age, supporting the phenomenon of scar disappearance with age in a proportion of children.28,29 In addition, estimations among children without BCG scar were also carried out using the anti-mode method to ascertain any gross variation between estimates using the MI method in all children when compared to the estimates using the anti-mode method in non-vaccinated children. These differences were not found to be pronounced. None of the data sets among all children showed clear bimodal patterns. On examining the data subsets by BCG status and age group, only mode or antimode could be discerned in some of them. Such identified modes of tuberculous reactions were not far askew from the normal pattern one would expect in a group of infected individuals, and were generally in line with tuberculin distributions seen among TB patients. Nevertheless, challenges faced in segregating true tuberculous reactions from cross-reactions suggest that future tuberculin surveys should be restricted to high-risk populations. Acknowledgements The authors are grateful to the field and supervisory staff for timely and quality work, the district officials for support and P Glaziou for help in calculating changes in annual risk of tuberculous infection rates. Funds were provided by the Ministry of Health and Family Welfare, Government of India. Conflict of interest: none declared. References 1 World Health Organization Global tuberculosis control: WHO report 2011. WHO/HTM/TB/2011.16. Geneva, Switzerland: WHO, 2011. 2 National Tuberculosis Institute, Bangalore. Annals of the National Tuberculosis Institute. Bangalore, India: Government of India, 2000. http://ntiindia.kar.nic.in./docs/nti_annals/annals/ index.htm Accessed January 2013. 3 Khatri G R, Frieden T R. The status and prospects of tuberculosis control in India. Int J Tuberc Lung Dis 2000; 4: 193– 200. 4 Central TB Division, Directorate General Health Services, Ministry of Health and Family Welfare, Government of India. TB India 2010, RNTCP status report. New Delhi, India: Government of India, 2011. http://www.tbcinda.nic.in/documents. html Accessed October 2012. 5 Styblo K. The relationship between the risk of tuberculous infection and the risk of developing infectious tuberculosis. Bull Int Union Tuberc Lung Dis 1985; 60: 117–199. 6 World Health Organization. Generic guidelines for the estimation of the annual risk of tuberculous infection (ARTI) 2006. New Delhi, India: Regional Office For South East Asia, World Health Organization. http: //www.who.int/tb/advisory_bodies/ impact_measurement_taskforce/meetings/ARTI_generic_ protocol. pdf Accessed October 2012. 7 Chadha V K, Agarwal S P, Kumar P, et al. Annual risk of tuberculous infection in four defined zones of India: a comparative picture. Int J Tuberc Lung Dis 2005; 9: 569–575. 8 Chadha V K, Jagannatha P S, Kumar P. Can BCG-vaccinated children be included for tuberculin surveys to estimate the annual risk of tuberculous infection in India? Int J Tuberc Lung Dis 2004; 8: 1437–1442. 9 Gopi P G, Subramani R, Venkatesh Prasad V, Narayanan P R. Estimation of annual risk of tuberculous infection among children irrespective of BCG scar in South India. Indian J Tuberc 2006; 53: 7–11. 10 Chadha V K, Vaidyanathan P S, Jagannatha P S, Unnikrishnan K P, Savanur S J, Mini P A. Annual risk of tuberculosis infection in the western zone of India. Int J Tuberc Lung Dis 2003; 7: 536–542. 11 Chadha V K, Vaidyanathan P S, Jagannatha P S, Unnikrishnan K P, Mini P A. Annual risk of tuberculosis infection in north zone of India. Bull World Health Organ 2003; 81: 573–581. 12 Arnadottir T, Rieder H L, Trébucq A, Waaler H T. Guidelines for conducting tuberculin skin test surveys in high-prevalence countries. Tubercle Lung Dis 1996; 77 (Suppl 1): 1–20. 13 Rieder H L. Methodological issues in the estimation of the tuberculosis problem from tuberculin surveys. Tubercle Lung Dis 1995; 76: 114–120. 14 El Ibiary S, de Coster E J M, Tolba F M, et al. Trend in the annual risk of tuberculous infection in Egypt, 1950–1996. Int J Tuberc Lung Dis 1999; 3: 294–299. 15 Nyboe J. The efficacy of the tuberculin test. An analysis based on the results from 33 countries. Bull World Health Organ 1960; 22: 5–37. 16 Narain R, Nair S S, Rao R, Chandrasekhar P, Lal P. Enhancing of tuberculin allergy by previous tuberculin testing. Bull World Health Organ 1966; 34: 623–638. 17 Rust P, Thomas J. A method for estimating the prevalence of tuberculosis infection. Am J Epidemiol 1975; 101: 311–322. 18 Palmer C E, Edwards L B, Hopwood L, Edwards P Q. Experimental and epidemiologic basis for the interpretation of tuberculin sensitivity. J Pediatr 1959; 55: 413– 429. 19 Bleiker M A, Sutherland I, Styblo K, Ten Dam H G, Misljenovic O. Guidelines for estimating the risk of tuberculous infection from tuberculin test results in a representative sample of children. Bull Int Union Tuberc Lung Dis 1989; 64 (2): 7–12. 20 Rieder H L, Chadha V K, Nakelkerke N J D, van Leth F, van ARTI trends in India 21 22 23 24 25 der Werf M J. Guidelines for conducting tuberculin skin test surveys in high-prevalence countries. Int J Tuberc Lung Dis 2011; 15 (Suppl 1): S1–S25. Lemeshow S, Hosmer D W Jr, Klar J, Lwanga S K. Adequacy of sample size in health studies. Chichester, UK: Wiley, 1990. Cauthen G M, Pio A, ten Dam H G. Annual risk of tuberculous infection. Bull World Health Organ 2002; 80: 503–511. Neuenschwander B E. Bayesian mixture analysis for tuberculin induration data. Paris, France: International Union Against Tuberculosis and Lung Disease, 2010. http://www.tbrieder.org/ research/mixture/mixture.html Accessed October 2012. Raviglione M C, Sudre P, Reider H L, Spinaci S, Kochi A. Secular trends of tuberculosis in western Europe. Bull World Health Organ 1993; 71: 297–306. Styblo K, Meijer J. Recent advances in tuberculosis epidemiol- 26 27 28 29 319 ogy with regard to formulation or readjustment of control programs. Bull Int Union Tuberc 1978; 53 (4): 301–312. Styblo K. Epidemiology of tuberculosis; selected papers. Vol. 24. The Hague, The Netherlands: Royal Netherlands Tuberculosis Association, 1991. Chadha V K, Jagannatha P S, Vaidyanathan P S, Jagota P. PPD RT23 for tuberculin surveys in India. Int J Tuberc Lung Dis 2003; 7: 172–179. Chakraborty A K, Ganapathy K T, Nair S S, Bhushan K. Prevalence of non-specific sensitivity to tuberculin in a South Indian rural population. Indian J Med Res 1976; 64: 639–651. Chanbasavaiah R, Mohan M, Suryanarayana H V, Krishnamurthy M S, Shashidhara A N. Waning of BCG scar. Indian J Tuberc 1993; 40: 137–144. ARTI trends in India i RÉSUMÉ Vingt-quatre districts en Inde. Déterminer les tendances du risque annuel d’infection tuberculeuse (ARTI) dans chacune des quatre régions géographiquement définies du pays. S C H É M A : On a mené deux séries d’enquête tuberculinique basées sur le domicile à 8–9 ans d’intervalle chez les enfants âgés de 1 à 9 ans dans des grappes statistiquement sélectionnées au cours des périodes 2000– 2003 et 2009–2010 (Enquêtes I et II). Au total, 184 992 enfants ont été testés au moyen d’une unité internationale de tuberculine purifiée (PPD) RT23 avec Tween 80 lors de l’Enquête I et 69 496 enfants au moyen d’une dose de 2 unités de PPD lors de l’Enquête II. Le diamètre transversal maximal de l’induration a été mesuré environ 72 h après l’administration du test. On a comparé l’ARTI à partir des prévalences d’infection estimées au moyen de la méthode des images en miroir. CONTEXTE : OBJECTIF : Les taux d’ARTI estimés dans les différentes zones ont varié entre 1,1% et 1,9% lors de l’Enquête I et entre 0,6% et 1,2% lors de l’Enquête II. L’ARTI a diminué de 6,1% et 11,7% par année respectivement dans les zones du nord et de l’ouest ; on n’a observé aucune diminution dans les zones du sud et de l’est. Les estimations nationales du niveau ont été respectivement de 1,5% et 1%, avec une diminution de 4,5% par année dans la période concernée. C O N C L U S I O N : Bien qu’un déclin de l’ARTI ait été observé dans deux des quatre zones ainsi qu’au niveau national, le niveau actuel d’ARTI d’environ 1% dans trois zones suggère la nécessité d’une intensification future des activités de lutte contre la tuberculose. R É S U LTAT S : RESUMEN MARCO DE REFERENCIA: Veinticuatro distritos en la India. Evaluar las tendencias del riesgo anual de infección tuberculosa (ARTI) en cuatro zonas del país definidas geográficamente. M É T O D O S : Se llevaron a cabo dos sesiones de encuestas de tuberculina en los hogares con un intervalo de 8 a 9 años, en niños de 1 a 9 años de edad, en un muestreo por conglomerados del 2000 al 2003 y del 2009 al 2010 (Sesiones I y II). En total, se practicó la prueba tuberculínica a 184 992 niños con 1 unidad tuberculínica purificada (PPD) RT23 y polisorbato 80 en las encuestas de la Sesión I y a 69 496 niños con 2 unidades tuberculínicas (TU) PPD en las encuestas de la Sesión II. El diámetro máximo de induración se midió 72 h después de la administración de la tuberculina. El ARTI se ob- OBJETIVOS: tuvo a partir de la prevalencia de infección, calculada mediante el método de la imagen especular. R E S U LTA D O S : El índice del ARTI en las diferentes zonas osciló entre 1,1% y 1,9% en las encuestas de la Sesión I y entre 0,6% y 1,2% en las encuestas de la Sesión II. El ARTI disminuyó a 6,1% por año en la zona norte y a 11,7% por año en la zona occidental. No se observó ninguna disminución en las zonas sur y oriental. A escala nacional, el ARTI fue 1,5% al norte y 1,0% al occidente, con una disminución de 4,5% por año, durante el período de la intervención. C O N C L U S I Ó N : Si bien se observó una disminución del ARTI en dos de las cuatro zonas estudiadas y a escala nacional, este criterio que fue cercano a 1% en tres zonas justifica una mayor intensificación de las actividades del control de la tuberculosis.