Tuberculosis and osteoporotic fracture risk: development of individualized fracture risk estimation prediction model using a nationwide cohort study

Abstract

Purpose: Tuberculosis (TB) is linked to sustained inflammation even after treatment, and fracture risk is higher in TB survivors than in the general population. However, no individualized fracture risk prediction model exists for TB survivors. We aimed to estimate fracture risk, identify fracture-related factors, and develop an individualized risk prediction model for TB survivors. Methods: TB survivors (n = 44,453) between 2010 and 2017 and 1:1 age- and sex-matched controls were enrolled. One year after TB diagnosis, the participants were followed-up until the date of fracture, death, or end of the study period (December 2018). Cox proportional hazard regression analyses were performed to compare the fracture risk between TB survivors and controls and to identify fracture-related factors among TB survivors. Results: During median 3.4 (interquartile range, 1.6–5.3) follow-up years, the incident fracture rate was significantly higher in TB survivors than in the matched controls (19.3 vs. 14.6 per 1,000 person-years, p < 0.001). Even after adjusting for potential confounders, TB survivors had a higher risk for all fractures (adjusted hazard ratio 1.27 [95% confidence interval 1.20–1.34]), including hip (1.65 [1.39–1.96]) and vertebral (1.35 [1.25–1.46]) fractures, than matched controls. Fracture-related factors included pulmonary TB, female sex, older age, heavy alcohol consumption, reduced exercise, and a higher Charlson Comorbidity Index (p < 0.05). The individualized fracture risk model showed good discrimination (concordance statistic = 0.678). Conclusion: TB survivors have a higher fracture risk than matched controls. An individualized prediction model may help prevent fractures in TB survivors, especially in high-risk groups.