Mortality in Patients with HIV-1 Infection Starting Antiretroviral Therapy in South Africa, Europe, or North America: A Collaborative Analysis of Prospective Studies

• Authors: Andrew Boulle, University of Cape Town; Michael Schomaker, University of Cape Town; Margaret T. May, University of Bristol; Robert S. Hogg, British Columbia Centre for Excellence in HIV/AIDS; Bryan E. Shepherd, Vanderbilt University; Susana Monge, Instituto de Salud Carlos III; Olivia Keiser, University of Bern; Fiona C. Lampe, University College London; Janet Giddy, McCord Hospital; James Ndirangu, University of Kwazulu Natal; Daniela Garone, Medecins Sans Frontieres; Matthew Fox, Boston University; Suzanne M. Ingle, University of Bristol; Peter Reiss, Stichting HIV Monitoring; Francois Dabis, Centre INSERM U897; Dominique Costagliola, University of Paris 06; Antonella Castagna, San Raffaele Scientific Institute; Kathrin Ehren, University Hospital of Cologne; Colin Campbell, Generalitat de Catalunya; M. John Gill, University of Calgary; Michael Saag, University of Alabama; Amy C. Justice, Yale University; Jodie L. Guest, Emory University; Heidi M. Crane, University of Washington; Matthias Egger, University of Cape Town; Jonathan A. C. Sterne, University of Bristol
• Journal Title: PLoS Medicine
• Volume: Volume 11, Number 9
• Publisher: Public Library of Science | 2014-09-01, Pages e1001718-e1001718
• Type of Work: Article | Final Publisher PDF
• Publisher DOI: http://doi.org/10.1371/journal.pmed.1001718
• Subjects: Health Sciences, Public Health; Biology, Biostatistics
• Keywords: Science & Technology; Life Sciences & Biomedicine; Medicine, General & Internal; General & Internal Medicine; SUB-SAHARAN AFRICA; TREATMENT PROGRAM; FOLLOW-UP; HIV-1-INFECTED PATIENTS; DISEASE PROGRESSION; SCALE-UP; COHORT; OUTCOMES; ADULTS; 1ST

Abstract

After accounting for under-ascertainment of mortality, with increasing duration on ART, the mortality rate on HIV treatment in South Africa declines to levels comparable to or below those described in participating North American cohorts, while substantially narrowing the differential with the European cohorts.Please see later in the article for the Editors' Summary.High early mortality in patients with HIV-1 starting antiretroviral therapy (ART) in sub-Saharan Africa, compared to Europe and North America, is well documented. Longer-term comparisons between settings have been limited by poor ascertainment of mortality in high burden African settings. This study aimed to compare mortality up to four years on ART between South Africa, Europe, and North America.Data from four South African cohorts in which patients lost to follow-up (LTF) could be linked to the national population register to determine vital status were combined with data from Europe and North America. Cumulative mortality, crude and adjusted (for characteristics at ART initiation) mortality rate ratios (relative to South Africa), and predicted mortality rates were described by region at 0–3, 3–6, 6–12, 12–24, and 24–48 months on ART for the period 2001–2010. Of the adults included (30,467 [South Africa], 29,727 [Europe], and 7,160 [North America]), 20,306 (67%), 9,961 (34%), and 824 (12%) were women. Patients began treatment with markedly more advanced disease in South Africa (median CD4 count 102, 213, and 172 cells/µl in South Africa, Europe, and North America, respectively). High early mortality after starting ART in South Africa occurred mainly in patients starting ART with CD4 count <50 cells/µl. Cumulative mortality at 4 years was 16.6%, 4.7%, and 15.3% in South Africa, Europe, and North America, respectively. Mortality was initially much lower in Europe and North America than South Africa, but the differences were reduced or reversed (North America) at longer durations on ART (adjusted rate ratios 0.46, 95% CI 0.37–0.58, and 1.62, 95% CI 1.27–2.05 between 24 and 48 months on ART comparing Europe and North America to South Africa). While bias due to under-ascertainment of mortality was minimised through death registry linkage, residual bias could still be present due to differing approaches to and frequency of linkage.