62
Adherence to NNRTI-based Regimens and Virologic Outcomes in HIV-infected South African Adults
Jean Nachega*1, M Hislop2, L Rengensberg2, R Chaisson3, and G Maartens4
1Johns Hopkins Univ, Baltimore, MD, US; 2AID For AIDS, Cape Town, South Africa; 3Johns Hopkins Univ, Sch of Med, Baltimore, MD, US; and 4Univ of Cape Town, South Africa
Background: Non-nucleoside reverse transcriptase inhibitor
(NNRTI)-based ART has emerged as the preferred option for first-line treatment
and is used in ART roll-out programs in sub-Saharan Africa.
High levels (³95%) of adherence are required for virological suppression with unboosted
protease inhibitors. However,
it remains uncertain whether this high level of adherence is required for virologic suppression on NNRTI-based ART.
Methods: We studied 3325 HIV-infected adult South African adults enrolled in a
private-sector HIV/AIDS disease management program who began NNRTI-based ART between January 1998 and March
2003. ART adherence was calculated by dividing the number of months that
patients submitted claims by the number of months since ART initiation. Viral
suppression was measured as achieving a viral load <400 copies/ml. Logistic
regression and c2 statistics were used to identify predictors
of viral suppression.
Results: The mean age (±SD) at ART initiation was 37
(±7.7) years; 2077 patients (62.5%) were female; 3220 (96.8%) were black
Africans. The median (IQR) follow-up was 2.4 (1.9 to 3.0) years, and CD4+
count and log10 HIV viral load at enrolment were 151 (66 to 232)
cells/ mm3 and 5.1 (4.6 to 5.6) log10 copies/mL, respectively. A significant dose-response relationship
was identified between viral load suppression and adherence across all
adherence strata. Viral suppression was achieved by 210 of 952 (22%) patients
with <70% adherence, 160 of 283 (56.5%) with 70 to 79% adherence, 342 of 499
(68.5%) with 80 to 89% adherence, 290 of 381 (76.0%) with 90 to 94% adherence,
and 967 of 1210 (80%) with ≥95% adherence (p for trend <0.001). In a multivariate logistic regression model, variables significantly
associated with viral suppression (OR, 95% confidence interval) were: high ART
adherence (15.6, 12.4 to 19.5, for ≥95% vs
<70% adherence),
high baseline CD4+ count (1.4, 1.1 to 1.7, for >200 vs ≤50 cells/mm3), and baseline viral load
≤105 copies/mL (1.3, 1.1 to 1.5).
Conclusions: High levels of adherence, as assessed by pharmacy claim data in a
private-sector management program, strongly predict viral suppression in
HIV-positive South African adults on NNRTI-based ART. Pharmacy records are a
simple and a valid program-level adherence monitoring tool.
|
