Cost effectiveness of testing HIV infected individuals for TB in a low TB/HIV setting

•Testing all UK HIV clinic attendees for latent and active TB is not cost-effective.•Modelling shows only testing highest-risk groups for latent TB is cost effective.•More comprehensive testing may become cost-effective if the cost of an IGRA falls. Guidelines recommend routine testing for latent TB infection (LTBI) in people living with HIV. However there are few cost-effectiveness studies to justify this in contemporary high resource, low TB/HIV incidence settings. We sought to assess the uptake, yield and cost-effectiveness of testing for latent and active TB. Adults attending an ambulatory HIV clinic in London, UK were prospectively recruited by stratified selection and tested for TB infection using symptom questionnaires, chest radiograph (CXR), tuberculin skin test (TST), T-Spot.TB and induced sputum. From this, 30 testing strategies were compared in a cost-effectiveness model including probabilistic sensitivity analysis using Monte Carlo simulation. 219 subjects were assessed; 95% were using antiretroviral therapy (ART). Smear negative, culture positive TB was present in 0.9% asymptomatic subjects, LTBI in 9%. Only strategies testing those from subSaharan Africa with a TST or interferon gamma release assay (IGRA) with or without CXR, or testing those from countries with a TB incidence of >40/100,000 with TST alone were cost-effective using a £30,000/QALY threshold. Cost-effectiveness analysis in an adult HIV cohort with high ART usage suggests there is limited benefit beyond routine testing for latent TB in people from high and possibly medium TB incidence settings.