AZT Trial

Published

Fischl MA et al. "The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex". The New England Journal of Medicine. 1987. 317(4):185-191.
PubMed • Full text • PDF

Clinical Question

Among patients with HIV/AIDS, does therapy with zidovudine reduce mortality when compared to placebo?

Bottom Line

Zidovudine reduces mortality in patients with HIV/AIDS when compared to placebo in this trial with an average follow-up of 4 months.

Major Points

Given the high mortality with HIV/AIDS and lack of an effective therapy, much effort went into identifying a potentially effective antiretroviral therapy in the 1980s. The nucleoside analog reverse transcriptase inhibitor (NRTI) zidovudine (ZDV; formerly azidothymidine or AZT) halts HIV replication through termination of viral DNA-chain elongation. It showed promise as the first antiretroviral medication for use in HIV/AIDS. A clinical trial demonstrating its effectiveness was lacking.

The 1987 AZT Trial included 282 mostly male HIV-positive patients with AIDS, defined by a single event of PCP pneumonia or AIDS-related complex -- an antiquated term for weight loss or oral candidiasis plus at least one other finding seen with HIV/AIDS. (The vast majority of patients included in this trial would meet the current CDC definition of AIDS, as the two groups had an average CD4 count of 71 and 187 cells/mm², respectively.) Participants were randomized to zidovudine 250 mg PO q4h or placebo. The trial was stopped early because of efficacy of the medication. With an average follow-up of 4 months, the zidovudine group had a higher rate of survival (99.3% vs. 86.1%) as well as reduction in opportunistic infections (OIs). A concurrent publication from the same group described a higher rate of nausea, anemia, and neutropenia with zidovudine. A reduced-dose regimen was effective and had fewer toxicities than this regimen in a follow-up trial.

Zidovudine gained FDA approval through a new expedited pathway shortly after publication of the AZT trial, though long-term follow-up data had not been collected. A mortality benefit was suggested with prolonged use of the medication in a follow-up publication by the same group. However, use as a monotherapy is limited given the development of viral resistance, first reported in 1989 by Larder and colleagues. The ACTG 229 study (1996) demonstrated improved viral suppression when the protease inhibitor saquinavir was combined with the nucleoside analog reverse transcriptase inhibitor zalcitabine and zidovudine, beginning the revolution of highly-active antiretroviral therapy (HAART; now simply ART) -- multiple drug combinations that greatly suppressed viral activity and greatly reduced mortality from HIV/AIDS.

Guidelines

NIH ART Panel - OARAC (2014, adapted)

• Initiate ART for HIV-infected patients to reduce the risk of disease progression:
  • CD4 <350 (AI)
  • CD4 350-500 (AII)
  • CD4 >500 (BIII)

WHO ART (2013, adapted)

• Start ART in if one of the following (strong recommendation, moderate-quality evidence):
  • Severe or advanced HIV clinical disease (WHO clinical stage 3 or 4)
  • Individuals with CD4 count <350 cells/mm3
• Start ART if HIV and a CD4 count >350 cells and <500 cells/mm3 regardless of WHO clinical stage (strong recommendation, moderate-quality evidence)
• Start ART if HIV, regardless of stage if one of the following:
  • Active TB (strong recommendation, low-quality evidence)
  • Coinfection of HIV and HBV and severe chronic liver disease (strong recommendation, low-quality evidence)
  • Serodiscordant couples to reduce risk of transmission to seronegative partner (strong recommmendation, high-quality evidence)

Design

• Multicenter, randomized, placebo controlled trial
• N=282
  • Zidovudine (n=145)
  • Placebo (n=137)
• Setting: 12 centers in the US
• Enrollment: February-June 1986
• Mean follow-up: 4 months (terminated early)
• Analysis: Intention-to-treat
• Primary outcome: Survival

Population

Inclusion Criteria

• HIV antibody positive
• AIDS, defined by first episode of PCP pneumonia in prior 120 days
• AIDS-related complex - unexplained weight loss >10% of TBW in prior 90 days or oral candidiasis with ≥1 of the following:
  • Unexplained fever
  • Extrainguinal lymphadenopathy
  • Oral hairy leukoplakia
  • Unexplained night sweats
  • Herpes zoster
  • Unexplained diarrhea
• Hgb ≥9.5 g/dL, granulocyte count ≥1000/mm³, platelet count ≥75,000/mm³, AST <3x ULN, creatinine <2 mg/dL, ANC <500/mm³, skin-test anergy

Exclusion Criteria

• Multiple episodes of PCP pneumonia
• Other OI
• Any neoplasm
• Recent treatment with antiretroviral medications, immunomodulators, or systemic antimicrobials

Baseline Characteristics

Limited baseline characteristics were presented by the authors. Presented as an average of both groups except where stated. Comparisons are zidovudine vs. placebo.

• Male: 95.3%
• HIV severity: AIDS 56.7%, AIDS-related complex 43.2%
• Days since PCP diagnosis in patients with AIDS 77.5 vs. 86.6 days (P<0.04)
• Karnofsky performance score: 90
• Weight: 68.7 kg
• CD4 count:
  • AIDS: 71/mm³
  • AIDS-related complex: 187/mm³
• Isolation of HIV virus: 58%

Interventions

• Randomization to a group:
  • Zidovudine - zidovudine 250 mg PO q4h, though frequency reduced to q8h if any severe adverse reactions documented
  • Placebo
• Up to 72 hours of ASA or APAP therapy for analgesia or fever
• Withdrawal from the study if development of toxic effects or neoplasm requiring therapy
• OI prophylaxis not allowed

Outcomes

Presented as a zidovudine vs. placebo.

Primary Outcome

Survival

99.3% vs. 86.1% (P<0.001)

At 24 weeks, probability: 0.98 vs. 0.78 (P<0.001)

Secondary Outcomes

OI

Any: 16.5% vs. 32.8%

>1: 1 vs. 5 participants

Of note, the authors state that no difference in the rate of OIs existed until after 6 weeks of therapy. At that point, individuals in the zidovudine group no longer developed OIs. >50% of all OIs were PCP pneumonia.

Kaposi's sarcoma

4.13% vs. 7.29% (P>0.20)

Karnofsky performance score

Change at 4 weeks: -0.3 vs. -2.2 (P<0.05)

Change at 12 weeks: +1.3 vs. -5.2 (P<0.001)

Change at 16 weeks: No difference (P>0.05)

Weight

Change at 4 weeks: +0.5 vs. -0.1 kg (P<0.05)

Change at 16 weeks: +2.0 vs. -1.3 kg (P<0.001)

CD4 count change from baseline

AIDS:

4 weeks: +131% vs. -11% (P<0.001)

8 weeks: +88% vs. -16% (P<0.001)

12 weeks: +61% vs. -27% (P<0.001)

16 weeks: +23% vs. -22% (P=0.008)

20 weeks: -1% vs. -39% (P=0.034)

24 weeks: -44% vs. -56% (P=0.024)

AIDS-related complex:

4 weeks: +29% vs. +4% (P<0.001)

8 weeks: +12% vs. -8% (P=0.003)

12 weeks: +27% vs. -6% (P<0.001)

16 weeks: +35% vs. +2% (P<0.001)

20 weeks: +49% vs. -10% (P<0.001)

24 weeks: +32% vs. +33% (P=0.737)

The authors comment that patients were more likely taken off the study if they had very low CD4 counts, likely influencing the outcome of the AIDS-related complex findings at 24 weeks.

Reactivity to a skin antigen

Trichophyton, tetanus toxoid, candida, and PPD.

29% vs. 9% (P<0.001)

Negative HIV culture at week 20

51% vs. 26% (P=0.52)

Subgroup Analysis

Probability of survival at 24 weeks

AIDS: 0.96 vs. 0.76 (P<0.001)

AIDS-related complex: 1.00 vs. 0.81 (P<0.016)

CD4 count ≤100: 0.96 vs. 0.70 (P<0.001)

CD4 count 101-499: 1.00 vs. 0.91 (P=0.28)

Adverse Events

Withdrawal for adverse events

21 vs. 40 individuals

Detailed adverse events were described in a concurrent publication from the same group:

Any

145 vs. 137 individuals

Anorexia: 11% vs. 8%

Asthenia: 19% vs. 18%

Diarrhea: 12% vs. 18%

Dizziness: 6% vs. 4%

Fever: 16% vs. 12%

Headache: 42% vs. 37%

Insomnia: 5% vs. 1% (P=0.043)

Malaise: 8% vs. 7%

Myalgia: 8% vs. 2%

Nausea: 46% vs. 18% (P=0.001)

Abdominal pain: 17% vs. 15%

Photosensitivity: 0% vs. 3% (P=0.041)

Rash: 17% vs. 15%

Somnolence: 8% vs. 9%

Taste Perversion: 5% vs. 8%

Hematologic abnormalities

Hgb <7.5 g/dL: 24.5% vs. 12.6% (P<0.001)

MCV >100 um³: 69% vs. 0%

MCV >110 um³: 41% vs. 0%

ANC <500: 16% vs. 2% (P<0.001)

Platelets >50% reduction: 12% vs. 5%

Criticisms

• Early termination limits interpretation of results for those with a trend towards better survival (like those with AIDS-related complex and CD4 count >100)
• Unclear if this would benefit patients with HIV not meeting the trial's definitions of AIDS or AIDS-related complex
• Unknown long-term outcomes
• Higher baseline performance status in the zidovudine group
• Disallowance of PCP prophylaxis

Funding

Not explicitly stated though the zidovudine patent was held by the Burroughs Wellcome Company (which was subsequently purchased by GlaskoSmithKline, which joined with Pfizer to become ViiV Healthcare) participated and presumably helped fund the trial. Zidovudine was subsequently sold under the trade name Retrovir by these groups.