Abstract

HIV-controllers with a known date of seroconversion - what happens before and during HIV-control, results from CASCADE

Y. Madec¹, F. Boufassa², M. Prins³, C. Sabin⁴, A. d'Arminio Monforte⁵, P. Amornkul⁶, A. Venet⁷, O. Lambotte⁷, K. Porter⁸, L. Meyer², CASCADE Collaboration

¹Institut Pasteur, Unité d'Epidémiologie des Maladies Emergentes, Paris, France, ²INSERM 1018-CESP, Kremlin-Bicêtre, France, ³Amsterdam Public Health Service, Amsterdam, Netherlands, ⁴Royal Free & University College Medical School, London, United Kingdom, ⁵University of Milan, Milan, Italy, ⁶International AIDS Vaccine Initiative, New York, United States, ⁷INSERM 802, Kremlin-Bicêtre, France, ⁸Medical Research Council, London, United Kingdom

Background HIV-controllers spontaneously maintain their level of viraemia at an undetectable level. Little is known about the time taken to achieve virological control after HIV-infection or about CD4 and viral load (VL) dynamics before and during control.
Methods We identified HIV-controllers from a pooled dataset of 24 seroconverter cohorts (N=19,615) from Europe, Australia and North-America (CASCADE). HIV-controllers were patients with ≥5 consecutive VL below detection limit, while ART-naive, the last of these VL being measured ≥5 years after seroconversion. End of HIV-control was defined as two consecutive VL >2000 copies/mL.
Delay from seroconversion to control and duration of control were described using Kaplan-Meier estimates, and factors associated with duration of HIV-control identified using a Cox model. Mixed linear models were used to describe HIV-RNA and CD4 progression before and during HIV-control.
Results We identified 125 HIV-controllers with a known date of seroconversion. At seroconversion, their mean CD4 level was estimated at 753 cells/mm³ (95% CI: 645-860) which remained stable until HIV-control, and mean HIV-RNA was 3.6 log copies/mL (95% CI: 3.3-3.9).
The median (IQR) delay to HIV-control was 1.4 (0.6-3.9) years in the 60 HIV-controllers with 1^st VL measured within 24 months from seroconversion. Median duration of control was 14.8 years, and end of control occurred in 20 controllers. Presence of blips during control was the only factor associated with an increased risk of escape (HR 8.9 (95% CI: 3.1-25.5)). Moreover, during control, CD4 count decreased at a higher rate in those with blips than those without (-44 (95% CI: -61;-26) versus -15 (95% CI: -29;-2) cells/mm³/year; p=0.001).
Conclusions HIV-control may occur late after seroconversion. HIV-control lasted more than 15 years in 50% of the HIV-controllers, duration of HIV-control being reduced by occurrence of blips but neither by delay to control nor CD4 level at seroconversion.
Download the e-Poster