Background:  In South Africa, second-line ART consists of lopinavir/ritonavir (LPV/r) + 2 NRTI. We hypothesized that second-line LPV/r-based regimens would be highly suppressive at 6-month follow-up and well tolerated.

Methods:  We analyzed data from a pilot study of protease inhibitor (PI) -naive patients who received second-line LPV/r-based ART at McCord Hospital in Durban, South Africa. Patients with ≥6 months follow-up were included. Characteristics at initiation of LPV/r-based treatment, adverse effects and outcomes were recorded with a standardized instrument. Virologic suppression (<50 copies/mL) at 6 months was the primary outcome.

Results:  Of 3365 patients initiating ART since 2004, 192 (6%) have required second-line ART. We report outcomes for 135 patients. Indications for second-line ART were virologic failure (72%), prior adverse effect (25%), and other (3%). Patients' characteristics were:  64% women; median age 38 years; number of prior regimens, 1 (59%), ≥2 (41%)) had, at the initiation of second-line ART, a median viral load of 24,898 copies/mL (IQR 6500 to 92,000 copies/mL), and a median CD4 count of 143/mm³ (IQR 78 to 218/mm³). NRTI backbones used with LPV/r were:  zidovudine (AZT)/didanosine (ddI) (47%), AZT/lamivudine (3TC) (29%), stavudine (d4T)/3TC (15%), and other (9%). After 6 months, 82% achieved virologic suppression to <50 copies/mL. We did not find a large difference in the percentage achieving suppression by NRTI backbone used with LPV/r (AZT/ddI [83%] vs non-ddI-containing backbones [82%], p = 0.9), indication for second-line ART (virologic failure [79%] vs adverse reaction /other indication [89%], p = 0.2), number of prior first-line regimens (1 [78%] vs ≥2 [89%], p = 0.08), or concurrent use of tuberculosis (TB) therapy during second-line ART (concurrent TB therapy [82%] vs no concurrent TB therapy [82%], p = 0.9). A significant difference was seen in rate of suppression by gender (women [89%] vs men [71%], p = 0.01). Combined grade 3/4 adverse effects occurred in 7% in the first 6 months; 2% required regimen withdrawal. Hypercholesterolemia (>215 mg/dl) was noted in 25% and hypertriglyceridemia (>200 mg/dL) in 33%.

Conclusions:  These pilot data suggest that LPV/r-based second-line regimens are effective in PI-naive patients in South Africa. Although no large differences in early outcome were noted in preselected subgroups, future studies must confirm these findings. Overall, few patients required early regimen change, but dyslipidemia was observed in greater than one quarter.