Time-dependent association between cancer and risk of tuberculosis: A population-based cohort study

Abstract

Background: We aimed to investigate the time-dependent association between cancer and the risk of tuberculosis (TB) before and after cancer diagnosis.

Methods: This population-based cohort study incorporated the National Health Insurance Research Database and the National Health Interview Survey in Taiwan to estimate TB risk in cancer and noncancer populations. We estimated the period-specific incidence rate ratio (IRR) between cancer and risk of TB and used Cox proportional hazards models to estimate the average hazard ratio between cancer and TB during the peridiagnostic period.

Results: From 2001 to 2015, 457 673 cancer and 3 738 122 noncancer individuals were enrolled. After stratifying the IRR of TB by year relative to the date of cancer diagnosis, the peak IRRs clustered in the year before and after the index date. In the peridiagnostic period of cancer, the adjusted hazard ratio was 2.29 (95% CI, 2.22-2.35) using the Cox model and 2.20 (95% CI, 2.09-2.32) after adjustment for missing confounders. Patients with cancers in the respiratory tract, upper digestive tract, and hematologic system were at the highest risk for TB.

Conclusions: Cancer is an independent risk factor for TB, with the highest risk observed around the time of cancer diagnosis.

Keywords: Cancer; Missing confounders; Risk; Temporality; Tuberculosis.