PMID- 27495971 OWN - NLM STAT- MEDLINE DCOM- 20170602 LR - 20170602 IS - 1365-3156 (Electronic) IS - 1360-2276 (Linking) VI - 21 IP - 10 DP - 2016 Oct TI - Risk of self-reported symptoms or diagnosis of active tuberculosis in relationship to low body mass index, diabetes and their co-occurrence. PG - 1272-1281 LID - 10.1111/tmi.12763 [doi] AB - OBJECTIVE: Globally, tuberculosis prevalence has declined, but its risk factors have varied across place and time - low body mass index (BMI) has persisted while diabetes has increased. Using India's National Family Health Survey (NFHS), wave 3 and World Health Survey (WHS) data, we examined their relationships to support projection of future trends and targeted control efforts. METHODS: Multivariate logistic regressions at the individual level with and without diabetes/BMI interactions assessed the relationship between tuberculosis, diabetes and low BMI and the importance of risk factor co-occurrence. Population-level analyses examined how tuberculosis incidence and prevalence varied with diabetes/low BMI co-occurrence. RESULTS: In NFHS, diabetic individuals had higher predicted tuberculosis risks (diabetic vs. non-diabetic: 2.50% vs. 0.63% at low BMI; 0.81% vs. 0.20% at normal BMI; 0.37% vs. 0.09% at high BMI), which were not significantly different when modelled independently or allowing for risk modification with diabetes/low BMI co-occurrence. WHS findings were generally consistent. Population-level analysis found that diabetes/low BMI co-occurrence may be associated with elevated tuberculosis risk, although its predicted effect on tuberculosis incidence/prevalence was generally