Incidence of Active Tuberculosis and Cohort Retention Among Adolescents in Western Kenya

SETTING: Siaya County, with the highest tuberculosis notification rates in Kenya. OBJECTIVES: To determine the incidence of active tuberculosis and 1-year cohort retention in 12-18-year-old adolescents, in preparation for phase III tuberculosis vaccine trials. METHODS: Adolescents were enrolled and followed up for 1-2 years to determine tuberculosis incidence. Adolescents with a positive tuberculin skin test, history of cohabitation with a tuberculosis case or at least 1 tuberculosis symptom received clinical and sputum examination and a chest radiograph. Definite tuberculosis cases were bacteriologically confirmed and clinical cases diagnosed by a clinician based on a suggestive chest radiograph and having clinical symptoms. Risk factors were explored using Poisson regression. RESULTS: Among 4934 adolescents without tuberculosis at baseline, 26 tuberculosis cases were identified during follow-up with a corresponding incidence density of 4.4 [95% confidence interval (CI): 3.0-6.4] events per 1000 person-years of observation, 12 definite tuberculosis cases; incidence density of 2.0 (95% CI: 0.9-3.1). Having previous tuberculosis (rate ratio: 12.5; CI: 1.8-100) and presence of tuberculin skin test conversion (rate ratio: 3.4; CI: 1.5-7.7) were significantly associated with higher risk of tuberculosis. Overall (4086/4925), 83.0% of adolescents were retained in the study after 1 year of follow-up. Being female, older, out of school and being orphaned were significant risk factors for loss to follow-up. CONCLUSION: The tuberculosis incidence in adolescents will help inform future tuberculosis vaccine trial sample size calculations for this setting. The predictive factors for tuberculosis and retention can be further explored in future trials.