The impact of the SARS-CoV-2 pandemic on tuberculosis notifications and deaths in the state of São Paulo, Brazil: a cross-sectional study.

Background: The state of São Paulo reports the highest number of tuberculosis cases in Brazil. We aimed to analyze the SARS-CoV-2 pandemic's impact on tuberculosis notifications and identify factors associated with reduced notifications and tuberculosis deaths in 2020-2021. Methods: This retrospective cross-sectional study analyzed data from 126,649 patients with tuberculosis notified in São Paulo from 2016 to 2021. Interrupted time series analysis assessed the pandemic's impact on notifications. Descriptive statistics and logistic regressions identified factors associated with decreased tuberculosis notifications and deaths during the pandemic (2020-2021) compared to the pre-pandemic period (2019). Findings: Tuberculosis notifications decreased by 10% and 8% in 2020 and 2021, respectively, with declines 2-3 times higher among individuals with no education or deprived of liberty. Contrastingly, tuberculosis notifications increased 68% among corrections workers in 2021. Diagnostics and contact tracing were compromised. Individuals with HIV, drug addiction, or deprived of liberty had lower odds of notification during the pandemic. Black and Pardo individuals or those with diabetes, treatment interruption history, or treatment changes post-adverse events had higher odds of notification. However, adverse events and tuberculosis-diabetes cases have been increasing since 2016. During the pandemic, tuberculosis-related deaths rose 5.0%-12.7%. Risk factors for mortality remained similar to 2019, with Pardo ethnicity, drug addiction and re-treatment post-adverse events emerging as risk factors in 2020/2021. Interpretation: The pandemic affected tuberculosis notifications and deaths differently among populations, exacerbating inequalities. Treatment interruption, loss of follow-up, and challenges in accessing healthcare led to increased mortality. Funding: FAPESP, CNPq and CAPES, Brazil.