The Francis Crick Institute
jia2.25772 (1).pdf (1.32 MB)

What is the optimum time to start antiretroviral therapy in people with HIV and tuberculosis coinfection? A systematic review and meta-analysis.

Download (1.32 MB)
journal contribution
posted on 2021-07-28, 09:53 authored by Rachael M Burke, Hannah M Rickman, Vindi Singh, Elizabeth L Corbett, Helen Ayles, Andreas Jahn, Mina C Hosseinipour, Robert J Wilkinson, Peter MacPherson
BACKGROUND: HIV and tuberculosis are frequently diagnosed concurrently. In March 2021, World Health Organization recommended that antiretroviral therapy (ART) should be started within two weeks of tuberculosis treatment start, at any CD4 count. We assessed whether earlier ART improved outcomes in people with newly diagnosed HIV and tuberculosis. METHODS: We did a systematic review by searching nine databases for trials that compared earlier ART to later ART initiation in people with HIV and tuberculosis. We included studies published from database inception to 12 March 2021. We compared ART within four weeks versus ART more than four weeks after TB treatment, and ART within two weeks versus ART between two and eight weeks, and stratified analysis by CD4 count. The main outcome was death; secondary outcomes included IRIS and AIDS-defining events. We pooled effect estimates using random effects meta-analysis. RESULTS AND DISCUSSION: We screened 2468 abstracts, and identified nine trials. Among people with all CD4 counts, there was no difference in mortality by earlier ART (≤4 week) versus later ART (>4 week) (risk difference [RD] 0%, 95% confidence interval [CI] -2% to +1%). Among people with CD4 count ≤50 cells/mm3 , earlier ART (≤4 weeks) reduced risk of death (RD -6%, -10% to -1%). Among people with all CD4 counts earlier ART (≤4 weeks) increased the risk of IRIS (RD +6%, 95% CI +2% to +10%) and reduced the incidence of AIDS-defining events (RD -2%, 95% CI -4% to 0%). Results were similar when trials were restricted to the four trials which permitted comparison of ART within two weeks to ART between two and eight weeks. Trials were conducted between 2004 and 2014, before recommendations to treat HIV at any CD4 count or to rapidly start ART in people without TB. No trials included children or pregnant women. No trials included integrase inhibitors in ART regimens. DISCUSSION: Earlier ART did not alter risk of death overall among people living with HIV who had TB disease. For logistical and patient preference reasons, earlier ART initiation for everyone with TB and HIV may be preferred to later ART.


Crick (Grant ID: 10218, Grant title: Wilkinson, R FC001218)