Abstract
National and sub-national population-based surveys, when performed at intervals, may assess important changes in TB prevalence. In 2022 we re-surveyed nine Nairobi County neighbourhoods that were previously surveyed in 2015. We aimed to determine pulmonary TB prevalence, compare prevalence to 2015 estimates, and evaluate changes in risk groups. Participants who reported cough of any duration and/or whose chest x-ray suggested TB submitted sputum for smear microscopy, Xpert Ultra, and liquid culture. We defined prevalent TB as Mycobacterium tuberculosis detection by sputum Xpert or culture, excepting individuals who were only trace positive. Our methods differed from 2015, which used solid media, Xpert MTB/RIF, and cough duration >2 weeks. We calculated TB prevalence using random-effects logistic regression models with missing value imputations and inverse probability weighting. In 2022 among 6369 participants, 1582 submitted ≥1 sputum sample, among whom 42 (2.7%) had TB, a weighted TB prevalence of 806/100,000 (95% confidence interval (CI), 518-1096). An additional 31 (2.0%) participants tested Ultra trace-positive/culture-negative. For comparison to 2015, we excluded 2022 participants (n = 2) whose only criterion for sputum was cough <2 weeks. There was no evidence for a decline in overall TB prevalence from 2015 to 2022. TB prevalence among men was high (1301/100,000) and remained high compared to 2015 (p-value <0.05). The age group with the highest estimated prevalence remained people ages 45-54 years. Among people with prevalent TB who reported cough, 76% had not sought health care. Dissimilar from other serial surveys that showed declines in TB prevalence, we found persistently high TB prevalence over a 7-year period in Nairobi County. Limitations of this study include changes in methodology between the two surveys and complex effects of the COVID-19 pandemic.