Evidence for prolonged clinical benefit from initial combination antiretroviral therapy: Delta extended follow-up

The findings from therapeutic trials in HIV infection with surrogate endpoints based on laboratory markers are only partially relevant for clinical decisions on treatment. Although the collection of clinical follow-up data from such a trial would be relatively straightforward, this rarely occurs. An important reason for this may be the perception that such data have little value because the number of participants remaining on their original allocated therapy has usually fallen substantially.

Delta was an international, multicentre trial in which 3207 HIV infected individuals were randomly allocated to treatment with zidovudine (ZDV) alone, ZDV combined with didanosine (ddI) or ZDV combined with zalcitabine (ddC). Although the trial closed in September 1995, information on vital status, AIDS events, treatment changes and CD4 counts was still collected every 12 months until at least March 1997. This has allowed analyses of the longer term clinical effect of treatment.

The median follow-up to date of death or last known vital status was 43 months (10th percentile 18 months; 90th percentile 55 months). The proportion of participants remaining on their allocated treatment fell steadily over time; by 4 years after trial entry, 3% remained on ZDV, 20% on ZDV + ddI and 21% on ZDV + ddC. Changes mainly involved the stopping, addition or switching of a nucleoside reverse transcriptase inhibitor (NRTIs). There was little use of protease inhibitors (PIs) or non-nucleoside reverse transcriptase inhibitors (NNRTIs) before the third year of the trial. Between the third and fourth years, regimens included a drug from one of these classes for approximately 17% of person-time in all treatment groups. Relative to ZDV monotherapy, the beneficial effects of combination therapy on mortality and disease progression rates increased significantly with time since randomization. The maximum effects on mortality were observed between 2 and 3 years, with a 48% reduction for ZDV + ddI and a 26% reduction for ZDV + ddC. These rates were observed when the original allocated treatment was received 42% and 47% of the time in the ZDV + ddI and ZDV + ddC groups, respectively. The mean CD4 count remained significantly higher (approximately 50 cells/microL) in the combination therapy groups 4 years after randomization, suggesting a projection of a clinical benefit beyond this time point.

The sustained clinical effect of the initial allocation to combination therapy, particularly ZDV + ddI, was remarkable in light of the convergence of drug regimens actually received across the three treatment groups. Interpretation of this finding is not straightforward. One of the possible explanations is that the effectiveness of ddI and ddC is diminished if first used later in infection or with greater prior exposure to ZDV, although the data do not clearly support either hypothesis. This analysis highlights the value of long-term clinical follow-up of therapeutic trials in HIV infection, which should be considered in the planning of all new studies.