• The Context Study: Efficacy and Safety of GW433908/RTV in PI-Experienced Subjects With Virological Failure (24 Week Results) (Oral 178)
  Authored by E. DeJesus, A. LaMarca, M. Sension, C. Beltran, P. Yeni
  View the original abstract

The Context Study evaluated the efficacy and safety of GW433908 (908)/ritonavir (RTV, Norvir) in protease inhibitor (PI)-experienced subjects with virologic failure. This analysis summarized the 24-week data and was presented by Edwin DeJesus.

Study subjects were randomized 1:1:1 to receive 908, the calcium phosphate ester pro-drug of amprenavir (APV, Agenerase), 1,400 mg/RTV 200 mg QD, 908 700 mg/RTV 100 mg BID, or lopinavir (LPV+RTV, Kaletra) 3 PO BID with two NRTIs based on resistance testing. Eligibility criteria required that patients must have prior experience with one or two PIs with HIV RNA > 1000 copies/ml. Use of NNRTIs was not allowed.

The primary study endpoint was the time-averaged change in viral load from baseline (AAUCMB) at 24 and 48 weeks. This presentation summarized the week 24 data. Approximately 105 subjects were enrolled in each of the three arms. The median baseline viral load (4.14 log) and CD4 count (263 cells/mm³) were similar across the treatment arms.

By ITT analysis, the mean AAUCMB (log10/ml) was -1.47 for the 908 QD arm, -1.5 for the 908 BID arm, and -1.66 for the LPV+RTV arm. By ITT analysis, the percentage < 400 and 50 copies was 58 and 40, 60 and 42, and 69 and 48 for the 908 QD, 908 BID, and LPV+RTV arms respectively. The median CD4 cell change from baseline was 72, 62, and 63 for the 908 Q day, 908 BID, and LPV+RTV arms. Drug-related clinical adverse rates (> grade 2) and treatment limiting rates were 19 percent/3 percent, 35 percent/3 percent, and 34 percent/6 percent for the 908 QD, 908, and LPV+RTV arms respectively.

The main conclusion was that the virologic activity of the LPV+RTV arm was mildly superior (but not statistically significant). From the statistical perspective, the 908 arms were considered non-inferior (the study was powered to show non-inferiority). The lipid changes were relatively minor in all groups (partially due to the abacavir (ABC, Ziagen)/3TC (lamivudine, Epivir) NRTI backbone).

I would have liked to see resistance data to assess the degree of PI resistance at study entry. Also, if there was a fair degree of PI and NRTI resistance it may have been in the patient's best interest to allow use of NNRTIs (which were excluded). Joe Eron, who was in the audience, made an excellent point. He commented that the use of the AAUCMN (average area under the curve minus baseline) may underestimate the effectiveness of treatment when many study subjects achieve high level of suppression (a better indicator for this intermediate salvage population may well be percent < 50 copies). Nonetheless either the QD or BID 908 with RTV boosting appeared roughly comparable to the current gold standard PI (LPV+RTV) used in the setting of initial PI failure. Further data from the 48-week and genotypic analysis will be useful in order to place the results of this study in perspective since the potential for downstream salvageability of virologic failures in the study is a key issue.