Natural history of HIV-control since seroconversion : AIDS (original) (raw)

Epidemiology and Social

Madec, Yoanna; Boufassa, Faroudyb,c; Porter, Kholoude; Prins, Mariaf; Sabin, Carolineg; Monforte, Antonella d’Arminioh; Amornkul, Paulii; Bartmeyer, Barbaraj; Sannes, Mettek; Venet, Alainl; Lambotte, Olivierl,m,n; Meyer, Laurenceb,c,d,n

aInstitut Pasteur, Emerging Diseases Epidemiology Unit, Paris

bINSERM U1018, Centre de Recherche en Epidémiologie et Santé des Populations, Le Kremlin-Bicêtre

cFaculté de Médecine Paris Sud, Université Paris Sud, Paris

dDepartment of Public Health, AP-HP, Bicêtre Hospital, Le Kremlin-Bicêtre, France

eMedical Research Council, Clinical Trials Unit, London, UK

fCluster Infectious Diseases, Department of Research, Center for Infection and Immunity Amsterdam (CINIMA), Public Health Service, Amsterdam, The Netherlands

gResearch Department of Infection and Population Health, UCL Medical School, London, UK

hSan Paolo Hospital, Milan, Italy

iInternational AIDS Vaccine Initiative, San Francisco, USA

jRobert Koch Institut, Berlin, Germany

kUlleval Hospital, Oslo, Norway

lINSERM U1012, Le Kremlin-Bicêtre

mDepartment of Internal Medicine, AP-HP, Bicêtre Hospital, Le Kremlin-Bicêtre

nUniversité Paris-Sud, Paris, France.

Correspondence to Yoann Madec, Unité d’Epidemiologie des Maladies Emergentes, Institut Pasteur, 25–28, rue du Docteur Roux, 75015 Paris, France. Tel: +33 1 40 61 3828; fax: +33 1 45 68 8876; e-mail: [email protected]

Abstract

Objectives:

HIV-controllers spontaneously maintain HIV viremia at an undetectable level. We aimed to describe the delay to control from seroconversion, the duration of control, and risk factors for losing control.

Methods:

HIV-controllers were identified from a pooled dataset of 24 seroconverter cohorts from Europe, Australia, and Canada (CASCADE). HIV-controllers had at least five consecutive viral loads less than 400/500 copies/ml, while antiretroviral therapy naive, for at least 5 years after seroconversion. End of control was defined as two consecutive viral loads above 2000 copies/ml. Duration of control was described using Kaplan–Meier estimates; factors associated with duration of control were identified using a Cox model. CD4+ cell count evolution during control was described using a mixed model.

Results:

Of 9896 eligible seroconverters, we identified 140 (1.4%) HIV-controllers, the largest database of HIV-controllers followed from seroconversion. For 64 with viral load measured within 24 months from seroconversion, median delay to control was 16.7 (interquartile range: 7.8–37.9) months. Probability of maintaining control 20 years after seroconversion was 0.74 [95% confidence interval (CI): 0.64–0.85]. Occurrence of blips followed by return to undetectability did not increase the risk of loss of control [hazard ratio: 0.81 (95% CI: 0.10–6.70)]. However, CD4+ cell loss during control was significantly accelerated in individuals with blips.

Conclusion:

In most individuals, control occurred rapidly after seroconversion; however, more than 3 years were required to achieve control in 25% of HIV-controllers. Control may be sustained even when CD4+ cell levels are below 500 cells/μl, opening important new perspectives to understand the physiopathology underlying control.