619    

Background:  Recent research suggests that African American patients may metabolize efavirenz (EFV) differently than whites, resulting in prolonged drug clearance and possibly a higher incidence of adverse effects. We sought to determine whether there was a difference between African Americans and whites in durability of EFV use in clinical practice.

Methods:  African American and non-Hispanic white patients from the Johns Hopkins HIV Cohort who received EFV as part of their first HAART regimen started after January 1, 2000 were compared for duration of EFV use. Race was self-identified. The Kaplan-Meier approach was used to compare time to discontinuation of EFV. Cox proportional hazards regression assessed discontinuation by race adjusting for other factors. Time to suppression of HIV-1 RNA (< 400/mL) was assessed. A control group of patients receiving a boosted protease inhibitor (PI) was also compared. Loss to follow-up was low (< 2%).

Results:  Of 218 African American and 65 non-Hispanic white patients who received EFV, 92 (42%) African American and 15 (23%) non-Hispanic white patients discontinued therapy. Probability of discontinuation of EFV by 1 year was 32% in African Americans, and 16% in non-Hispanic whites (p = 0.002). Probability of HIV-1 RNA suppression was 59% at 6 months and 66% at 1-year in African American, and 72% at 6 months and 82% at 1 year in non-Hispanic whites (p = 0.01). By Cox regression, African American had a relative hazard of 2.06 (95% CI 1.16 to 3.68) compared with non-Hispanic whites for discontinuing EFV, adjusting for age, sex, risk group (injecting drug users [IDU] vs non-IDU), associated nucleoside reverse transcriptase inhibitor (NRTI) backbone, CD4, and HIV-1 RNA at start of therapy. Associated with discontinuation of therapy were:  reported adverse effect (African Americans = 30, non-Hispanic whites = 8), failure to suppress HIV-1 RNA (African American = 63, non-Hispanic whites = 9). By genotype, a primary non-NRTI (NNRTI) mutation was found in 26 (28% of d/c and 12% total) of African Americans and 4 (27% of d/c and 6% total) of non-Hispanic whites who discontinued EFV. In contrast, for 69 African Americans and 16 non-Hispanic whites receiving a boosted-PI, the 1-year probability of discontinuation of boosted-PI was 41% in African American and 36% in non-Hispanic whites (p = 0.84), and the probability of HIV-1 RNA suppression was 69% in African Americans and 61% in non-Hispanic whites at 6 months (p = 0.91).

Conclusions:  African Americans were more likely than non-Hispanic whites to discontinue an EFV-based HAART regimen in our clinical practice. There was no difference between the 2 groups when a boosted PI was used. Our results support racial differences in EFV durability that appear to be associated with differences in viral suppression and NNRTI resistance.

Keywords: Race/ethnicity; Efavirenz; Therapeutic response