Background
High-risk alcohol use is associated with increased risk for active TB, the leading cause of death among persons with HIV (PWH). We hypothesized that the incidence rate of active TB after receipt of TB Preventive Therapy (TPT) would differ by level of alcohol use among PWH.
Methods
We conducted a review of HIV/TB clinic charts to find active TB diagnosis among 988 PWH with prior positive tuberculin skin tests in a sample enriched for persons with high-risk alcohol use in South-western Uganda. These persons engaged in our prior research and received 6 months of TPT, with electronic pill-cap monitoring. We used person-time methods and defined time at risk as months from 6-months after TPT start to date (t0) until TB diagnosis, last known TB-free clinic or study visit, or death not known to be due to TB. We conducted Cox modeling to obtain the hazards ratio (HR) and 95% confidence intervals (CI) of high-risk alcohol use at t0, defined as Alcohol Use Disorders Identification Test – Consumption (AUDIT-C, modified to cover 3 months) ≥6 or the alcohol biomarker, phosphatidylethanol (PEth) ≥200 ng/mL, versus not high-risk, adjusting for gender and age. We further adjusted for the following variables if they were individually associated (in models with alcohol use, gender, and age) with active TB (p<0.10): HIV viral non-suppression (≥200 copies/ml), ART regimen, CD4 count, smoking, TPT non-completion (<90% pills taken), body mass index, and number of people/household room.
Results
968 (98%) persons had total follow-up of 3229 person-years (py). There were 21 cases of TB, confirmed by Gene Expert (n=13) or lipoarabinomannan (LAM, n=8), for a TB incidence rate of 0.65% (95% CI: 0.42-1.00%) per py (ppy). 410 (42.4%) had high-risk alcohol use, 634 (65.5%) completed TPT, and 32 (3.3%) were HIV non-suppressed. The TB incidence rate was 1.28% (95% CI: 0.79-2.09%) ppy among those with high-risk alcohol use, and 0.25% (95% CI: 0.11-0.61%) ppy among those without high-risk alcohol use (Figure). The adjusted hazard ratio (aHR) for high-risk alcohol use yes versus no was 3.74 (95% CI: 1.29-10.79). Cigarette smoking (aHR: 2.42, 95% CI: 0.99-5.92) and HIV non-suppression (aHR: 5.58, 95% CI: 1.64-18.98) were also independent predictors of active TB. There was no interaction by gender.
Conclusions
These results suggest ongoing TB acquisition and/or reactivation despite high rates of viral suppression and receipt of TPT among PWH, particularly among those with high-risk alcohol use and cigarette smoking.
Table or Figure
![]()