Socio-behavioural determinants of mortality in HIV-infected patients receiving combined antiretroviral treatment (cART): results from the ANRS CO8 APROCO-COPILOTE cohort

Background: HIV-related mortality has been dramatically reduced following the advent of combined antiretroviral treatment (cART). Identifying the changing pattern of predictors of mortality in HIV-infected patients is becoming a major concern. Few studies have addressed this issue over the long term. This study aimed to disentangle the impact of clinical and socio-behavioral predictors of mortality in HIV-infected individuals started on cART 11 years ago in the French ANRS CO8 APROCO-COPILOTE cohort.
Methods: The APROCO-COPILOTE cohort enrolled 1,281 patients initiating a protease-inhibitor-containing cART between 1997 and 1999. Clinical and laboratory data were collected every 4 months. Self-administered questionnaires collected psycho-social and behavioral characteristics at enrolment (M0), and every 8-12 months thereafter. Deaths from all causes except for accidents, overdoses and suicides, occurring after M4 were included in the analysis. Predictors of mortality were studied using a shared-frailty Cox model to take into account informative dropouts.
Results: During follow-up (M4-M144), 100 deaths occurred among the 1,100 eligible patients who completed a self-administered questionnaire at M0 and who had >4 months of follow-up (incidence rate [95% CI]=1.3 [1.0-1.6] per 100 person-years). The main causes of death were related to cancer (n=26), AIDS (n=23), cardiovascular diseases (n=12) and hepatitis (n=10). After adjustment for age, baseline CDC clinical stage C and plasma HIV RNA ≥105 copies/ml, both antiretroviral naivety at enrolment (only in patients having a time since diagnosis >8 years) and limited early immune recovery (CD4+ cell count/mm3 < 200 at M4 for patients with a level < 200 at M0 or < 500 at M4 for patients with a level between 200 and 350 at M0) were associated with higher mortality risk. Moreover, low social status (no high-school education and/or unemployment) (HR [95% CI]=1.97 [1.12-3.46]) and moderate alcohol consumption (≤1 alcohol unit/day) (0.48 [0.27-0.87]) were independently associated with mortality. The risk of death for patients with higher alcohol consumption was not significantly different from that for abstainers.
Conclusion: Higher social status and moderate alcohol consumption seem to play a protective role on mortality in HIV-infected patients receiving ART. The potential benefits of moderate alcohol consumption need greater exploration in different socio-cultural contexts.

C. Protopopescu1,2,3, M.P. Carrieri1,2,3, V. Le Moing4, P. Reboud5, L. Piroth6, L. Cuzin7, F. Couturier5, F. Raffi8, B. Spire1,2,3, C. Leport9, the ANRS CO 8 APROCO-COPILOTE Study Group
1INSERM, UMR 912 (SESSTIM), Marseille, France, 2Aix Marseille Université, UMR_S912, IRD, Marseille, France, 3ORS PACA, Observatoire Régional de la Santé Provence-Alpes-Côte d'Azur, Marseille, France, 4CHU Montpellier, UMI 233 TransVIHMI, IRD, Université Montpellier 1, Montpellier, France, 5INSERM, U 897, Bordeaux, France, 6CHU Dijon, Dijon, France, 7CHU Purpan, Toulouse, France, 8CHRU Hôtel Dieu, Nantes, France, 9Université Paris Diderot, Sorbonne Paris Cité, INSERM, UMR 738, Paris, France