Elsevier

The Lancet

Volume 348, Issue 9023, 3 August 1996, Pages 283-291
The Lancet

Articles
Delta: a randomised double-blind controlled trial comparing combinations of zidovudine plus didanosine or zalcitabine with zidovudine alone in HIV-infected individuals

https://doi.org/10.1016/S0140-6736(96)05387-1Get rights and content

Summary

Background

Because the benefits of zidovudine (AZT) in HIV-infected individuals are small and do not last long the Delta trial was designed to test whether combinations of zidovudine with didanosine (ddl) or zalcitabine (ddC) were more effective than AZT alone in extending survival and delaying disease progression.

Methods

The trial was randomised, double blind, and international. 3207 participants were allocated to either AZT (600 mg per day) alone (1055), AZT plus ddI (400 mg per day) (1080), or AZT plus ddC (2·25 mg per day) (1072). Participants either had symptoms of HIV disease (if AIDS, with a CD4cell count of >50×106/L) or a CD4 count of less than 350×106/L; 2124 had not had zidovudine before (Delta 1) and 1083 had for at least 3 months (Delta 2).

Findings

Over a median follow-up of 30 months, 699 participants died, and 936 of the 2765 without AIDS at entry developed AIDS or died. In participants who had not had AZT before, both combination regimens had substantial benefits in terms of survival (regardless of disease stage at entry); a relative reduction in mortality of 42%, compared to AZT alone (95% CI 25% to 55%), for AZT plus ddl and of 32% (95% CI 22% to 47%) for AZT plus ddC. In participants who had had AZT before, the addition of ddI improved survival (p=0·05; relative reduction 23% [95% CI 0% to 41%]) but there was no direct evidence of benefit from the addition of ddC (p=0·47; relative reduction 9% [95% CI—17% to 29%]). The overall difference in survival between the treatment groups was significant (p<0·0001; a relative reduction in mortality, compared to AZT alone, of 33% (95% CI 20% to 44%) for AZT plus ddI and 21% (95% CI 6% to 34%) for AZT plus ddC). Benefit in terms of disease progression was seen mainly in participants not previously treated with AZT and overall. There was no unexpected toxicity from the combination treatments.

Interpretation

Initiation of treatment with combinations of AZT plus ddI or ddC prolongs life and delays disease progression compared with AZT alone. The addition of ddI to participants already treated with AZT also improves survival, although the benefit appears less.

Introduction

Zidovudine (AZT), is an established treatment for HIV-infected individuals with symptoms. The first placebo-controlled trial was stopped early because of a significant survival benefit and improvement in morbidity after an average 4 months follow-up.1 Subsequent data from this study showed that AZT delayed death only for a short time.2 Although there is evidence that treatment with AZT in HIV-infected individuals with no symptoms delays the development of disease symptoms,3 5 there was no delay in the onset of AIDS or death from immediate rather than deferred treatment in trials with the longest follow-ups.6, 7 To improve on these benefits, combinations of AZT with dideoxynucleoside-reverse-transcriptase inhibitors were investigated in the hope that they might be more effective in inhibiting viral replication and delay onset of drug resistance, which is thought to be one of the causes of the failure of AZT given alone.8, 9 Such combinations have been shown to have an additive or synergistic effect in vitro, to be free of serious pharmacokinetic interactions, and to have greater effects than monotherapy on markers associated with progression of HIV infection.10, 11

At the start of this trial, the only drugs available to test in combination with AZT were didanosine (ddI, dideoxyinosine, an adenosine analogue) and zalcitabine (ddC, dideoxycytidine, a cytosine analogue). Combinations of ddI or ddC plus AZT were compared with AZT alone in a double-blind controlled trial. Individuals who had not received antiretroviral treatment before, as well as those who had already been treated with AZT were included.

HIV-antibody positive individuals aged 15 or more who had either not received AZT before or had received AZT for at least 3 months and could tolerate a dose of 600 mg/day were eligible unless they had exclusion criteria shown in table 1. Participants gave written informed consent and the appropriate ethics committee approval was obtained for each centre.

Delta was intended to be analysed separately for participants who had not had AZT before (Delta 1) and for those who had had at least 3 months (Delta 2) as well as an overall analysis.

All participants were to take AZT 600 mg/day (200 mg three times a day) and were randomly allocated in a 2:1:2:1 ratio to also take ddI, matching ddI placebo; ddC or matching ddC placebo. The daily dose of ddC was 2–25 mg (0·75 mg three times a day) and of ddI 400 mg (200 mg twice a day). AZT or blinded trial tablets could be modified according to protocol. The protocol allowed participants to change treatment after progression to severe AIDS-related complex (ARC) or AIDS for those without symptoms or with only CDC group IVE12, 16 at entry, to AIDS for those with ARC at entry or to severe AIDS for those with AIDS at entry. Participants allocated to AZT plus ddI or its matching placebo were to switch to AZT plus ddC and those allocated to AZT plus ddC or its matching placebo to AZT plus ddI. The trial began in March, 1992.

The protocol was amended in September, 1993, to allow a change of treatment if there was progression of HIV disease not meeting the criteria for a clinical endpoint and/or a 50% decrease in CD4 count from baseline; those allocated to AZT plus ddI or its matching placebo were to change to AZT plus open label ddI, and those allocated to AZT plus ddC or its matching placebo to AZT plus open label ddC.

In March, 1995, another protocol amendment allowed a second randomisation for participants who had reached a clinical endpoint or had evidence of disease progression (as defined above) or had completed 2 years on blinded trial therapy. Those allocated to AZT plus ddI or matching placebo (including those who had received open label ddI but not ddC) were randomised to receive AZT plus either ddC or lamivudine (3TC) and those originally allocated to AZT plus ddC or matching placebo (including those who had received open label ddC but not ddI) to AZT plus either ddI or 3TC. In the analysis all participants were to be included in their original allocated treatment group regardless of whether they were entered in the second randomisation.

Concomitant medication including primary prophylaxis for opportunistic infections, was allowed but it was recommended that blinded trial drug be interrupted during administration of intravenous pentamidine.

Participants were randomised in 175 centres in France (55), the UK (47), Italy (21), Australia (21), the Netherlands (17), Switzerland (7), Germany (4), New Zealand (2), and Ireland (1). Randomisation was undertaken centrally by national Trials Centres in Paris (France, Italy), London (UK, Ireland, Switzerland), Amsterdam (Netherlands, Germany), and Sydney (Australia, New Zealand) by telephone, fascimile, or electronic transfer. Randomisation was stratified by participating centre and in Australia, New Zealand, Germany, and the Netherlands by previous use of AZT. In the UK, Ireland, and Switzerland a minimisation algorithm was used to maintain balance for duration of previous AZT treatment and HIV disease stage.14 Trial drugs were to be prescribed 2 weeks after randomisation, except in those clinical centres where drug supplies were kept locally and treatment was started at randomisation. Median time between randomisation and trial drug prescription was 14 days (Interquartile range [IQR] 12–16).

Between August, 1994, and September, 1995, in all countries except Austrialia and New Zealand, a new formulation was supplied for both active ddI and its matched placebo; this had equivalent bioavailability but an improved flavour.

The planned intake was 1500 in Delta 1 and 1000 in Delta 2 but this was increased to 2250 in Delta 1 and 1200 in Delta 2 in September, 1993, because the proportion of participants discontinuing allocated treatment was higher than anticipated. The aim was to have the ability to detect with 80% power at the 1% significance level (two-tailed test) a 50% relative reduction in deaths for those with AIDS at entry and in progression to AIDS or death for those without AIDS at entry.

Trial entry was defined as the date the trial drug was prescribed. Assessment was carried out 2 weeks before, at randomisation (at trial entry if after randomisation), at 2 and 4 weeks, then 4-weekly until 24 weeks after entry and subsequently every 8 weeks. Adverse events and concomitant medications were recorded and case-record forms were sent by post, facsimile, or electronic transfer to national Trials Centres. If participants were lost to follow-up, national registers, when available, were used to identify first AIDS events or death. The Data and Safety Monitoring Committee (DSMC) reviewed summary data every 4 to 6 months, and a full interim analysis was planned in July 1994. The DSMC's guidelines included the statistical criterion that a difference of at least three standard deviations in an interim analysis of a major endpoint may be needed to justify halting or modifying the study prematurely. This has the advantage that the number of interim analyses has little impact on the interpretation of the results.15 Two further interim analyses were undertaken at the request of the DSMC and after the second, their recommendation that the trial be terminated 3 months before the planned date was endorsed by the International Coordinating Committee (ICC) on 15th September, 1995. The blinded phase of the trial ended on 25th September, 1995, and follow-up continues 6-monthly.

Eligibility and baseline data, serious adverse events, clinical endpoints, and deaths were validated against original clinic records, as were samples of other data. For all participants, status at entry, endpoints (including death), other AIDS events and serious adverse events were reviewed by the relevant trial physicians. All clinical endpoints not clearly fulfilling protocol definitions, all deaths where the cause was uncertain, and all cases of pancreatitis were reviewed by an Endpoint Review Committee consisting of the principal investigators, heads of national Trials Centres, and trial physicians, all of whom remained blinded to the treatment allocated.

Full blood count and immunology subsets were measured at all visits (except weeks 2 and 20) up to week 64 and then every 16 weeks. Potassium, urea or creatinine, alkaline phosphatase, aspartate or alanine transaminase, and total amylase were assayed and serum stored at all visits (except weeks 2, 20, 24, 40, and 56) up to week 64 and then every 16-weeks.

Efficacy Primary endpoints were death, and AIDS or death in those without AIDS at entry. AIDS events were defined by the 1987 CDC classification.12 In September, 1993, microsporidiosis with persistent diarrhoea lasting more than one month, leishmaniasis causing visceral disease, and invasive cervical carcinoma (but not pulmonary tuberculosis nor recurrent bacterial pneumonia) were included as AIDS-defining conditions13 and in February, 1995, T-cell lymphoma was added by the Endpoint Review Committee; all such events were included from trial entry.

Toxicity Primary endpoints for toxicity were any serious adverse event (expected or unexpected), or adverse events leading to discontinuation of trial treatment. Serious adverse events were defined as those not related to HIV infection and either fatal, life threatening, causing or prolonging hospital admission, causing permanent disability; cancer; congenital anomaly; clinical pancreatitis; peripheral neuropathy of grade 2 (moderate) or worse; clinical and laboratory events of grade 3 (severe) or worse; and overdose of trial tablets.

Severe ARC or AIDS or death in those without symptoms (or only CDC IVE) at entry. Severe ARC was defined as the development of at least one new condition in CDC group IVA16 (constitutional symptoms) with at least one new condition in CDC group IVC2 (opportunistic infection) with two consecutive CD4 counts of 100×l06/L or less at least two weeks apart. As only 25 (11 in Delta 1 and 14 in Delta 2) participants developed severe ARC, this analysis is not reported here.

Advanced AIDS or death in those with AIDS but not advanced AIDS at entry. The original endpoint was severe AIDS, (development of at least one new [non-recurrent] AIDS condition with two consecutive CD4 counts of 50×l06/L at least 2 weeks apart) but this was modified because of concerns about including CD4 counts as part of the definition in the light of the Concorde trial.7 In May, 1995, a definition of advanced AIDS was defined as a first diagnosis of progressive multifocal leucoencephalopathy (PML), cerebral or non-Hodgkin's lymphoma, HIV encephalopathy, cytomegolovirus disease (retinitis or other viscera), disseminated Mycobacterium avium-intracellulare disease or intracerebral lesion(s) of indeterminate cause (including cases considered to be possible PML, cerebral lymphoma, or HIV encephalopathy but not toxoplasmosis) —diseases associated with a worse prognosis.17

The development of a new (non-recurrent) AIDS event or death in all participants.

All adverse events leading to permanent dose reduction.

Baseline laboratory measurements are those recorded nearest to but before and within 6 weeks of trial entry. When baseline values were used to define changes over time, only those values within 1 week of trial entry were used. Analyses were done separately for Delta 1 and Delta 2 and overall combined analyses were stratified by previous AZT treatment. Time-to-event methods including Kaplan-Meier plots, log-rank test, and Cox proportional-hazards models15, 18 (with and without stratification by country, HIV disease stage, and baseline CD4 count) were used to compare the treatment groups for the primary and secondary clinical endpoints. CD4 counts over time were summarised in each treatment group by the median change from baseline and its standard error for values at fixed assessment times and for the summary measure of area under the curve to 6 months. For adverse events, treatment groups were compared for the frequency and rate of occurrence of all serious adverse events and all adverse events leading to premature discontinuation of trial therapy.

Analyses (except those of adverse events) compared the randomised treatment groups for their entire follow up (intention to treat). All p-values reported are two sided. For the analysis of adverse events, follow-up was censored 30 days after stopping trial treatment.

Section snippets

Results

Between March 2, 1992, and May 10, 1994, 3308 individuals were randomised (France 1408, the UK and Ireland 801, Australia and New Zealand 322, Italy 303, the Netherlands 260, Germany 129, and Switzerland 85); 2188 had never received AZT (Delta 1) and 1120 had previously received AZT (Delta 2).

Of the 3308, 101 (33 allocated to AZT plus ddI, 35 to AZT plus ddC, and 33 to AZT alone) who were never prescribed a trial drug (64 Delta 1; 37 Delta 2) were not included in the main analysis. 32 did not

Discussion

The important result from Delta is the substantial reduction in mortality from the combinations of AZT plus either ddI or ddC compared with AZT alone. The benefits appear to be greater in those patients who had not had AZT before, in whom a survival benefit occurred at all stages of HIV disease. We believe that a relative reduction in mortality of 37% (for the combination groups combined compared to AZT alone) over an average follow up of 2–5 years is clinically relevant. As the difference in

References (23)

  • MA Fischl et al.

    The efficacy and azidothymidine (AZT) in the treatment of participants with AIDS and AIDS-related complex: a double-blind, placebo-cont rolled trial

    N Engl J Med

    (1987)
  • MA Fischl et al.

    Prolonged zidovudine therapy in patients with AIDS and advanced AIDS-related complex

    JAMA

    (1989)
  • PA Volberding et al.

    Zidovudine in asymptomatic human immunodeficiency virus infection. A controlled trial in persons with fewer than 500 CD4-positive cells per cubic millimeter. The AIDS Clinical Group of the National Institute of Allergy and Infectious Diseases

    N Engl J Med

    (1990)
  • DA Cooper et al.

    Zidovudine in persons with asymptomatic HIV infection and CD4+ cell counts greater than 400 per cubic millimeter. The European-Australian Collaborative Group

    N Engl J Med

    (1993)
  • PA Volberding et al.

    A comparison of immediate with deferred zidovudine therapy for asymptomatic HIV- infected adults with CD4 cell count of 500 or more per cubic millimetre. N

    Engl J Med

    (1995)
  • Concorde: MRC/ANRS randomised double-blind controlled trial in immediate and deferred zidovudine in symptom-free HIV infection

    Lancet

    (1994)
  • B Larder et al.

    HIV with reduced sensitivity in zidovudine (AZT) isolated during prolonged therapy

    Science

    (1989)
  • DD Richman et al.

    Effect of stage of disease and drug dose on zidovudine susceptibilities of isolates of human immunodeficiency virus

    J AIDS

    (1990)
  • R Yarchoan et al.

    A randomized pilot study of alternating or simultaneous zidovudine and didanosine therapy in patients with symptomatic human immunodeficiency virus infection

    J Infect Dis

    (1994)
  • TC Meng et al.

    Combination therapy with zidovudine and dideoxycytidine in patients with advanced human immunodeficiency virus infection. A phase I/II study

    Ann Intern Med

    (1992)
  • Revision of the CDC surveillance case definition for Acquired Immunodeficiency Syndrome

    MMWR

    (1987)
  • Cited by (0)

    Listed at end of paper

    View full text