Influence of HIV and other risk factors on tuberculosis

Abstract

Background. Tuberculosis (TB) notification in South Africa has increased six-fold over the past two decades, mainly because of the HIV epidemic.

Objectives. To describe the sociodemographic and outcome characteristics of TB patients, and to identify risk factors associated with TB treatment outcomes stratified by HIV status.

Methods. A cross-sectional study was used to analyse data from the Cape Town Metro East geographical service area (GSA) electronic TB register (ETR.Net), including adult patients aged ≥15 years who initiated TB treatment between 1 July 2011 and 30 June 2012.

Results. TB case notification in the GSA was 922 per 100 000 population. Of the 12 672 TB patients registered, 50.5% were co-infected with HIV. The death rate in co-infected patients was 5.4% v. 2.8% in HIV-negative patients, the rate of treatment success 66.6% v. 73.5%, and the rate of unfavourable treatment outcome 28.1% v. 23.7%. The Khayelitsha subdistrict had the highest proportion of TB burden (37.0%) and co-infection (47.6%). Fourteen percent of patients had extrapulmonary TB, 65.9% of whom were co-infected with HIV. In the multivariate analysis, HIV infection (relative risk (RR) 1.2), retreatment (RR 1.4) and sputum smear microscopy not done (RR 1.4) were significantly associated with unfavourable treatment outcome. The Eastern (RR 0.9) and Northern (RR 0.7) subdistricts were less likely to have unfavourable outcomes compared with Khayelitsha. In the stratified analysis, retreatment and smear not done were significant risk factors for an unfavourable treatment outcome in both co-infected and HIV-negative patients.

Conclusions. The burdens of both TB and co-infection were high in this community, although HIV prevalence varied. Mortality was higher and treatment completion lower in co-infected patients than in those who were HIV-negative. Co-infection, previous TB treatment and smear not done were significant risk factors for an unfavourable outcome in all patients.