Background:  Fixed cycle scheduled treatment interruptions (F-STI) have been evaluated in patients with low pre-HAART CD4 count in East Africa and high pre-HAART CD4 count in other continents. We report the results of a fixed cycle 2-months-off/4-months-on HAART trial in patients with high pre-HAART CD4 count in West Africa.
Methods:  We randomized 651 adults with a median pre-HAART CD4 count of 272/μL, current on-HAART CD4 count >350 CD4/μL, and undetectable viral load into a non-inferiority trial including 3 arms:  continuous HAART (CT); CD4-guided HAART; and 2-months-off/4-months-on HAART. There were 3 primary endpoints:   mortality (non-inferiority bound –5), severe HIV-morbidity (WHO stage 3- or 4-events or death) ( non-inferiority bound –15), and percentage of patients with >350 CD4/μL at 24 months (non-inferiority bound +15). Secondary endpoints were cost, resistance, adherence, and toxicity. The CD4-guided arm was prematurely interrupted because of a high rate of morbidity. The continuous heart and 2-months-off/4-months-on arms were continued until trial termination (the last patient reaching 24 months of follow-up, March 2007). We report the results of the final analysis.
Results:  We randomized 435 adults (zidovudine-lamivudine-efavirenz [ZDV-3TC-EFV], 88%) into the continuous and 2-months-off/4-months-on arms. For all primary endpoints, the 2-months-off/4-months-on arm was not inferior to the continuous arm: The incidence or mortality was 0.45/100 person-years in the continuous therapy are and 0.46/100 person-years in the2-months-off/4-months-on arm (lower bound of  the unilateral 95%CI for difference –0.87: > –5 NIB); the incidence of severe HIV morbidity was 6.8/100 person-years in continuous and 9.1/100 person-years in the2-months-off/4-months-on arms (lower bound of 95%CI –6.0: > –15 NIB); the 24-month percentage of patients with >350 CD4/μL was 94% in continuous and 85% in 2-months-off/4-months-on arm (upper bound of 95%CI +14: <+15 NIB). As regards secondary endpoints, the overall cost of care was higher in continuous than in the 2-months-off/4-months-on arm ($960/person-year vs $785/person-year, p <0.0001) and the 24-month percentage of patients with at least one ARV drug resistance was lower in continuous than in 2-months-off/4-months-on arms (9% vs 21%, p = 0.007). Most resistance mutations were to NNRTI (62%) and lamivudine/emtricitabine (3TC/FTC; 23%). There was no difference between arms in terms of adherence, grade-3 to -4 toxicity, severe non-HIV morbidity, hypertension, dyslipidemia, hyperglycemia, and lipodystrophy.
Conclusions:  In these patients, predominantly EFV-treated and with high CD4 nadir and >350 CD4/μL at first interruption, the 2-months-off/4-months-on HAART was clinically not inferior to continuous therapy, but led to unacceptable resistance rates.