Investigation of household contacts of pulmonary tuberculosis patients increases case detection in Mwanza City, Tanzania

BACKGROUND Tuberculosis (TB) contact tracing is a key strategy for containing TB and provides addition to the passive case finding approach. However, this practice has not been implemented in Tanzania, where there is unacceptably high treatment gap of 62.1% between cases estimated and cases detected. Therefore calls for more aggressive case finding for TB to close this gap. We aimed to determine the magnitude and predictors of bacteriologically-confirmed pulmonary TB among household contacts of bacteriologically-confirmed pulmonary TB index cases in the city of Mwanza, Tanzania. METHODS This study was carried out from August to December 2016 in Mwanza city at the TB outpatient clinics of Tertiary Hospital of the Bugando Medical Centre, Sekou-Toure Regional Hospital, and Nyamagana District Hospital. Bacteriologically-confirmed TB index cases diagnosed between May and July 2016 were identified from the laboratory registers book. Contacts were traced by home visits by study TB nurses, and data were collected using a standardized TB screening questionnaire. To detect the bacterioriologically-confirmed pulmonary TB, two sputum samples per household contact were collected under supervision for all household contacts following standard operating procedures. Samples were transported to the Bugando Medical Centre TB laboratory for investigation for TB using fluorescent smear microscopy, GeneXpert MTB/RIF and Löwenstein-Jensen (LJ) culture. Logistic regression was used to determine predictors of bacteriologically-confirmed pulmonary TB among household contacts. RESULTS During the study period, 456 household contacts from 93 TB index cases were identified. Among these 456 household contacts, 13 (2.9%) were GeneXpert MTB/RIF positive, 18 (3.9%) were MTB-culture positive and four (0.9%) were AFB-smear positive. Overall, 29 (6.4%) of contacts had bacteriologically-confirmed pulmonary TB. Predictors of bacteriologically-confirmed pulmonary TB among household contacts were7being married (Odds ratio [OR], 3.3; 95% confidence interval [CI], 1.4-8.0; p = 0.012) and consuming less than three meals a day (OR, 3.7; 95% CI, 1.6-8.7; p = 0.009). CONCLUSIONS Our data suggest that in Mwanza, Tanzania, seven in 100 contacts living in the same house with a TB patient develop bacteriologically-confirmed pulmonary TB. These results therefore underscore the need to implement routine TB contact tracing to control tuberculosis in high TB burden countries such as Tanzania.