Intensification of antiretroviral therapy through addition of enfuvirtide in naive HIV-1-infected patients with severe immunosuppression does not improve immunological response: results of a randomized multicenter trial (ANRS 130 Apollo).

We studied whether addition of enfuvirtide (ENF) to a background combination antiretroviral therapy (cART) would improve the CD4 cell count response at week 24 in naive patients with advanced HIV disease. ANRS 130 Apollo is a randomized study, conducted in naive HIV-1-infected patients, either asymptomatic with CD4 counts of <100/mm(3) or stage B/C disease with CD4 counts of <200/mm(3). Patients received tenofovir-emtricitabine with lopinavir-ritonavir (LPV/r) or efavirenz and were randomized to receive ENF for 24 weeks (ENF arm) or not (control arm). The primary endpoint was the proportion of patients with CD4 counts of ≥ 200/mm(3) at week 24. A total of 195 patients were randomized: 73% had stage C disease, 78% were male, the mean age was 44 years, the median CD4 count was 30/mm(3), and the median HIV-1 RNA load was 5.4 log(10) copies/ml. Eighty-one percent of patients received LPV/r. One patient was lost to follow-up, and eight discontinued the study (four in each arm). The proportions of patients with CD4 counts of ≥ 200/mm(3) at week 24 were 34% and 38% in the ENF and control arms, respectively (P = 0.53). The proportions of patients with HIV-1 RNA loads of <50 copies/ml were 74% and 58% at week 24 in the ENF and control arms, respectively (P < 0.02), and the proportion reached 79% in both arms at week 48. Twenty (20%) and 12 patients (13%) in the ENF and control arms, respectively, experienced at least one AIDS event during follow-up (P = 0.17). Although inducing a more rapid virological response, addition of ENF to a standard cART does not improve the immunological outcome in naive HIV-infected patients with severe immunosuppression.

About the necessity to manage events coded with MedDRA prior to statistical analysis: proposal of a strategy with application to a randomized clinical trial, ANRS 099 ALIZE.

BACKGROUND: Since 2003, the Medical Dictionary for Regulatory Activities (MedDRA) is the regulatory standard for safety report in clinical trials in the European Community. Yet, we found no published example of a practical experience for a scientifically oriented statistical analysis of events coded with MedDRA. We took advantage of a randomized trial in HIV-infected patients with MedDRA-coded events to explain the difficulties encountered during the events analysis and the strategy developed to report events consistently with trial-specific objectives. METHODS: MedDRA has a rich hierarchical structure, which allows the grouping of coded terms into 5 levels, the highest being "System Organ Class" (SOC). Each coded term may be related to several SOCs, among which one primary SOC is defined. We developed a new general 5-step strategy to select a SOC as trial primary SOC, consistently with trial-specific objectives for this analysis. We applied it to the ANRS 099 ALIZE trial, where all events were coded with MedDRA version 3.0. We compared the MedDRA and the ALIZE primary SOCs. RESULTS: In the ANRS 099 ALIZE trial, 355 patients were recruited, and 3,722 events were reported and documented, among which 35% had multiple SOCs (2 to 4). We applied the proposed 5-step strategy. Altogether, 23% of MedDRA primary SOCs were modified, mainly from MedDRA primary SOCs "Investigations" (69%) and "Ear and labyrinth disorders" (6%), for the ALIZE primary SOCs "Hepatobiliary disorders" (35%), "Musculoskeletal and connective tissue disorders" (21%), and "Gastrointestinal disorders" (15%). CONCLUSIONS: MedDRA largely enhanced in size and complexity with versioning and the development of Standardized MedDRA Queries. Yet, statisticians should not systematically rely on primary SOCs proposed by MedDRA to report events. A simple general 5-step strategy to re-classify events consistently with the trial-specific objectives might be useful in HIV trials as well as in other fields.