The impact of a change in infant BCG vaccination policy on adolescent TB incidence rates: A South African population-level cohort study

Objective: South Africa's infant Bacille Calmette Guerin (BCG) vaccine policy changed from percutaneous (PC) BCG Japan to intradermal (ID) BCG Denmark in 2000. This study investigated whether this change in infant BCG vaccination had any durable impact on TB incidence rates (IR) into adolescence. Methods: The Cape Town electronic TB register provided data (from 2008 to 2018) on HIV-negative TB patients born in 1991–1999 (BCG Japan cohort) and 2001–2008 (BCG Denmark cohort). Statistics South Africa provided population estimates. Annual TB IR per 100,000 population were calculated stratified by age, gender and birth year. Interrupted time series analysis with a segmented Poisson regression and birth cohort analyses were used to compare incidence between the BCG cohorts and trends over time. Findings: TB IR increased throughout adolescence, with 17-year-olds having 7.34 [95% confidence interval (CI), 6.48–8.32] times higher TB IR than 10-year-olds. Females had 1.22 [95% CI 1.17–1.27] higher IR than males. Overall, adolescents who received ID BCG Denmark had a lower TB IR compared to PC BCG Japan (rate ratio 0.86, [95% CI 0.80–0.94]). No interaction between BCG and age, nor BCG and gender were identified. Birth cohort analyses showed the increase in TB IR started around one year earlier in females than in males. Conclusion: The change in infant BCG policy was associated with a modest decrease in TB incidence in 10- to 17-year-old HIV-negative adolescents. However, TB incidence rapidly increased with age in both adolescent cohorts and remained high despite BCG vaccination at birth.