Tuberculosis among people with HIV infection in the United Kingdom: opportunities for prevention?

OBJECTIVE: To investigate the incidence of, and risk factors for, tuberculosis among HIV clinic attendees in the United Kingdom. DESIGN AND METHODS: Observational cohort study of 27 868 individuals in the United Kingdom Collaborative HIV Cohort collaboration, 1996-2005. RESULTS: Among individuals not taking combination antiretroviral therapy (cART), tuberculosis incidence was considerably higher among individuals of black African vs. white or other ethnicities {incidence rates 9.9 [95% confidence intervals (CIs) 7.2, 12.6], 2.5 [95% CI 1.8, 3.0] and 4.4 [95% CI 2.7, 6.0] episodes per 1000 person-years, respectively}. Tuberculosis incidence decreased with time after starting cART; among black Africans, incidence was consistently higher and remained substantial (5.3 per 1000 person-years) at 24 months and longer after starting cART. The strongest independent risk factors for tuberculosis after cART start were most recent CD4 cell count: adjusted rate ratios (aRR) 10.65 (95% CI 6.11, 18.57), 3.40 (95% CI 2.05, 5.65), 1.77 (95% CI 1.06, 2.96) and 1.84 (95% CI 1.09, 3.12) for individuals with CD4 cell counts less than 50, 50-199, 200-349 and 350-499 cells/microl, respectively, compared with at least 500 cells/microl; and black African vs. white ethnicity [aRR 2.93 (95% CI 1.89, 4.54)]. HIV risk group, shorter time on cART, later calendar period and unsuppressed viral load were also independently associated with incident tuberculosis. CONCLUSIONS: Tuberculosis incidence among people attending UK HIV clinics is substantial, particularly among those with non-white ethnicity and low CD4 cell counts, even after starting cART. Earlier HIV diagnosis is needed in order to implement interventions to prevent tuberculosis; tuberculosis preventive therapy should be considered in addition to cART.