Home Search Abstracts View Session E-mail Abstract Author


Session 131 Poster Abstracts
Treatment Outcomes in Scale-up Programs
Session Day and Time: Monday, 1-4 pm
Room: Hall B


825
Risk Factors for Discordant Responses to HAART at Themba Lethu Clinic, Johannesburg, South Africa
Daniel Westreich*1, B Malope2, P MacPhail2, S Nagar2, A Van Rie1, and I Sanne2
1Univ of North Carolina at Chapel Hill, Sch of Publ Hlth, US and 2Univ of the Witwatersrand, Johannesburg, South Africa

Background:  Little is known about risk factors for discordant immunologic/virologic responses within 6 months of initiation of HAART in developing countries.

Methods:  We analyzed prospectively collected clinical data from the Themba Lethu Clinic in Johannesburg, South Africa. Immunologic response was defined as CD4 gain of at least 50 cells/mm3, and virologic response as viral load <400 copies/mL, both within 6 months of HAART initiation. We defined discordant response as immunologic but not virologic response, or virologic but not immunologic response. Log-binomial regression models were used to identify statistically significant predictors of discordancy. Models controlled for baseline age, CD4 count, body mass index (BMI), hemoglobin, tuberculosis, history of ART prior to HAART initiation, HAART regimen, adherence, gender, and pregnancy, and used backwards elimination (a = 0.10) to achieve parsimony.

Results:  Between April 1, 2004 and March 31, 2007, 3813 adults on HAART were available for analysis, of whom 920 (24.1%) had a discordant response at 6 months:  733 (19.2%) were immunologic but not virologic discordant, and 187 (4.9%) were virologic but not immunologic discordant. Immunologic but not virologic responses were more common among individuals with baseline CD4 >200 cells/mm3 (37.5% vs 17.2%, p <0.0001), and virologic but not immunologic outcomes were more common in those with CD4 ≤200 cells/mm3 (5.2% vs 2.1%, p = 0.0069). Risk of immunologic but not virologic outcomes among those with CD4 ≤200 were lower with use of nevirapine (NVP) (compared with all other drug regimens, RR 0.52, 95%CI 0.34 to 0.79) and in those with CD4 <50 (RR 0.84, 95%CI 0.72 to 0.98); risk increased for male gender (RR 1.33, 95%CI 1.14 to 1.54) and history of ART (RR 1.68, 95%CI 1.17 to 2.42). In those with CD4 >200, baseline pregnancy reduced risk of immunologic but not virologic (RR 0.55, 95%CI 0.30 to 1.00). In both CD4 categories, risk of immunologic but not virologic increased slight with age (per year older, RR 1.01, 95%CI 1.00 to 1.02). Risk of virologic but not immunologic discordancy was higher in those with baseline CD4 <50 (RR 1.47, 95%CI 1.11 to 1.96), and lower in those with baseline CD4 >200 (RR 0.41, 95%CI 0.20 to 0.85), both compared to those with baseline CD4 of 50 to 200 cells/mm3. Individuals with any history of ART had 2.16 times the risk (95%CI 1.17 to 4.01) of virologic but not immunologic outcomes as those without ART history.

Conclusions:  Overall, both immunologic but not virologic, and virologic but not immunologic 6-month discordancy was strongly predicted by CD4 counts and prior history of ART in our cohort. Results suggest both immunologic but not virologic, and virologic but not immunologic discordancy may be in part a consequence of CD4 recovery rates.