INT J TUBERC LUNG DIS 13(9):1100–1105 © 2009 The Union The decline of tuberculosis in Yemen: evaluation based on two nationwide tuberculin surveys A. Al-Absi,* A. Bassili, † H. Abdul Bary,* A. Barker, ‡ M. Daniels, ‡ A. Munim, § A. Seita, † B. Williams, ¶ C. Dye ¶ * National Tuberculosis Control Programme, Ministry of Public Health and Population, Sana’a, Yemen; † Stop TB, World Health Organization, Regional Office for the Eastern Mediterranean, Cairo, Egypt; ‡ Desmond Tutu TB Centre, Cape Town, South Africa; § World Health Organization, Khartoum, Sudan; ¶ Stop TB, World Health Organization, Geneva, Switzerland Correspondence to: Amal Bassili, Tropical Disease Research, Stop TB/World Health Organization/Regional Office for the Eastern Mediterranean, Abdul Razzak Al Sanhouri Street, P O Box 7608, Nasr City, Cairo 11371, Egypt. Tel: (+202) 276 5275. Fax: (+202) 670 2492/94. e-mail: bassilia@emro.who.int Article submitted 25 March 2009. Final version accepted 17 May 2009. SETTING: Yemen. OBJECTIVE: To review the epidemiological situation of tuberculosis (TB) in Yemen by conducting a tuberculin survey and by comparing the results obtained with those of a previous tuberculin survey from 1991. DESIGN: A nationwide tuberculin survey enrolling 31 276 schoolchildren aged between 7 and 12 years. RESULTS: Skin indurations were recorded for 28 499 schoolchildren, of whom 16 927 (59.4%) had no bacille Calmette-Guérin (BCG) scar. Analysis of the distribu- tion of indurations was difficult as it did not show any bimodal pattern. Prevalence of infection and annual risk of tuberculosis infection (ARTI) were thus estimated us- ing the mirror image and mixture methods, and not the cut-off point method. The two methods indicated simi- lar results: respectively 0.45% and 0.51% for prevalence of infection, and 0.05% and 0.05% for ARTI. In com- parison with the 1991 tuberculin survey, the average an- nual decline of ARTI was 9.0% by the mirror method and 5.5% by the mixture method. CONCLUSION: Yemen seems to have a low ARTI (0.05%), and TB infection seems to be declining consid- erably. Analysis of the survey results highlighted the lim- itations of tuberculin surveys in countries with compa- rable epidemiological situations. KEY WORDS: tuberculosis; tuberculin survey; preva- lence; annual risk of TB infection YEMEN adopted the DOTS strategy as national pol- icy in 1995 and has since expanded it considerably. In the past few years, the national DOTS population coverage has been almost 100%, except for in limited areas where the security situation has been volatile. The national tuberculosis control programme (NTP) has consistently reported more than 10 000 cases each year, in a country with a population of 22 mil- lion. The treatment success rate for the cohort 2001– 2005 was close to the 85% target. Evaluation of pro- gramme activities showed improved performance, and the analysis of surveillance data reported a better epidemiological situation. 1 The smear-positive case notification rate, which had shown a steady annual increase, reaching a peak in 2000, declined from 27 to 15 per 100 000 popula- tion between 2001 and 2006, with an average annual decline of 11.4%. The smear-positive case detection rate was 46% in 2007. 2 However, there are uncertain- ties about the estimated tuberculosis (TB) incidence. The NTP therefore decided to conduct a nationwide tuberculin survey to determine the prevalence of TB infection and annual risk of TB infection (ARTI) and to assess the trend, comparing the results with those of a previous nationwide tuberculin survey conducted in 1991. MATERIALS AND METHODS The study methods followed internationally recom- mended guidelines for conducting tuberculin surveys. 3 Study design, area and participants A cross-sectional survey was conducted during a 7- week period, whereby schoolchildren aged between 7 and 12 years with consenting parents or guardians were enrolled. The tuberculin skin test (TST) was also applied in a sample of 150 newly diagnosed spu- tum smear-positive cases. Sample size estimation and sampling technique The sample size was estimated as follows: N = 10.5 [P1(1 − P1) + P2(1 − P2)]/(P1 − P2) 2 SUMMARY