Predictors and impact of suboptimal immune response in HIV-infected patients in South Africa
Background: In most resource-limited settings, viral load monitoring for antiretroviral therapy (ART) is not available and CD4 count is the primary indication of successful response to ART. We aimed to describe predictors of suboptimal immune response by 12 months and its impact on attrition in Johannesburg, South Africa. Methodology: All ART-naïve adults initiating first-line ART between April 2004 and July 2012 at Themba Lethu HIV Clinic with both a baseline and 12 month CD4 count were included. Patients were followed from the date of the 12 month CD4 count until death, transfer, loss to follow-up (LTF), or dataset close (July 2013). Suboptimal immune response was defined as achieving <50 cell/mm3 gain in CD4 count by 12 months on ART. Predictors of suboptimal immune response and the association between immune response and risk of attrition (mortality and LTF) were estimated using log-binomial regression and presented as adjusted risk ratios (aRR). Results: Of the 6639 included patients, 63.7% of were female, median (IQR) age was 36.6 (31.5-43.1) years, and median (IQR) baseline CD4 count was 100 (39-169) cells/mm3. Patients who achieved suboptimal immune response were more likely to be male (43.3% vs. 35.3%), had a higher median baseline CD4 count (146.5 vs. 94), were less likely to be WHO Stage III/IV (31.5% vs. 41.4%) and have TB at ART initiation (8.8% vs. 15.5%) than patients who achieved a greater gain in CD4 count. For the 816 patients who experienced suboptimal immune response, the median (IQR) change was 15 (-10.5-34) cells/mm3 while it was 181 (122-261) cells/mm3 for the remaining 5823 patients. In an adjusted model, men (aRR: 1.39; 95% CI: 1.21-1.59), older patients (≥45 vs. <30: aRR 1.45; 95% CI: 1.16-1.82), and patients with higher baseline CD4 counts (100-199 vs. <50: aRR 2.79; 95% CI: 2.25-3.47) were more likely to experience suboptimal immune response (Table). Patients experiencing suboptimal immune response were 50% more likely to die or become lost as patients achieving a greater response to ART (aRR: 1.47; 95% CI: 1.29-1.67). Conclusions: We found a large proportion of the population is at increased risk of experiencing suboptimal immune response, putting them at higher risk for attrition. Routine CD4 count monitoring in resource-limited settings can be utilized to identify potentially poorly responding patients in order to determine the cause of their non-response to reduce the risk of attrition.
Conference: IWHOD 2014, Sitges, Spain