Background: Ritonavir (RTV) -boosted protease inhibitors (PI) are a critical component of many successful antiretroviral (ARV) regimens. Selective non-adherence to or staggering of RTV doses (i.e. dosing >4 hours from the other PI) may reduce exposure to the boosted PI. Adherence studies employing electronic monitoring (MEMS) have generally been limited to the boosted PI; little is known about concordant adherence to and staggering of RTV.

Methods: HIV⁺ subjects receiving RTV-boosted atazanavir (ATV) or fosamprenavir (fAPV) enrolled in a 24-week, prospective adherence study. Subjects used MEMS monitors on both their RTV and ATV or fAPV bottles. Subjects and staff were blinded to adherence data. The adherence rate for each medication was defined as 100x (# bottle openings/# prescribed doses). Subjects with RTV adherence >5% less than ATV/fAPV adherence were considered selectively non-adherent to RTV. The percentage of RTV doses that occurred >4 hours apart from the corresponding ATV/fAPV dose (staggered doses) was calculated. HIV-1 viral load was measured at 24 weeks.

Results: A total of 35 subjects completed the study:  mean age 45.2; 57% male, 40% female, 3% transgender; 46% Hispanic, 49% black non-Hispanic; HIV risk factors were: 40% heterosexual, 34% intravenous drug users (IDU), 3% men who have sex with men (MSM)/non-IDU, 23% unknown. At enrollment 74% had CDC-defined AIDS; 71% had >5 years of ARV experience; mean entry CD4 477 cells/µL; 74% with viral load <75 copies/mL. ARV regimens were:  26 ATV/RTV, 9 fAPV/RTV; 80% 2 NRTI, 17% 3 NRTI, 3% 1 NRTI; 97% received their PI once daily. Mean adherence rates to ATV/fAPV and RTV were both 77%.Of 35 patients, 3 (9%) were selectively non-adherent to RTV. Over the course of the study, 212 of 4710 (4.5%) doses of RTV were staggered, and 7 of 35 (20%) subjects staggered >5% of RTV doses. Mean adherence to PI (both ATV/fAPV and RTV) in those (n = 27) with viral <75 copies/mL was significantly higher than those (n = 8) with viral ≥75 copies/mL (86% vs 46%, p = 0.005 Wilcoxon). The 3 subjects with selective non-adherence to RTV had a mean adherence to ATV/fAPV of 93%, and all achieved viral load <75 copies/mL. There was no difference in proportion of subjects achieving viral load <75 copies/mL among subjects who staggered doses compared to those who did not.

Conclusions: Of the 35 subjects receiving RTV-boosted ATV or fAPV, 9% were selectively non-adherent to RTV. Dose-staggering, i.e. taking RTV >4 hours apart from the boosted PI, occurred in 20%. Efforts to maintain and improve adherence in recipients of RTV-boosted PI should consider these relatively common variations in adherence behaviors.