AZT: New Survival Data

A statistical study of over 2,500 HIV-positive men,
published April 16 in the NEW ENGLAND JOURNAL OF MEDICINE, found
that early AZT use and early pneumocystis prophylaxis each
independently increased the chance of survival. ("Early"
treatment in this study was defined as treatment beginning before
the first AIDS-defining event.) According to an April 15 press
release from the U. S. National Institute of Allergy and
Infectious Diseases, "Early treatment with AZT reduced the risk
of death at six months by 57 percent, whether or not it was
combined with a drug that prevented Pneumocystis carinii
pneumonia (PCP). Moreover, investigators found that, after two
years, early treatment reduced the risk of death by 33 percent,
when compared to those who did not take AZT before a diagnosis of
AIDS."

The benefits were seen especially in men who started AZT
with T-helper counts between 200 and 350. In those with T-helper
counts above 350 there were not enough deaths in this study to
collect reliable survival data.

The MACS Study

This survival analysis was done with data from the
Multicenter AIDS Cohort Study (MACS). The MACS cohort consists
of over 5,000 gay and bisexual men in four metropolitan areas:
Chicago, Pittsburgh, Los Angeles, and Baltimore/Washington; most
of them were enrolled in the study between April 1984 and March
1985, and followed since at twice-yearly visits. About half of
the volunteers in the MACS study are HIV-negative; only those who
are HIV-positive were included in the survival analysis. Deaths
and AIDS diagnoses through April 1, 1991, were included.

This study is not a randomized clinical trial, as the
treatments used were chosen by the patients and their physicians,
not assigned at random. Therefore there could be self-selection
bias, and precautions were taken in the analysis to minimize it.
The study is a prospective one, however, meaning that the
volunteers were selected first and followed forward in time,
whatever outcome occurred. This kind of study, which the MACS
cohort makes possible, is more reliable than a retrospective
study of medical records.

Discussion

The recent paper discusses earlier studies of survival with
and without AZT treatment. One observational study in Maryland,
published in 1991, found a median survival of 690 days for AZT
users, compared with 280 days for those who never used AZT. An
Australian study showed an even larger difference. But these
studies were in patients with AIDS; they did not compare the
effect of AZT as early treatment.

The paper also discusses the controversial Veterans Affairs
Cooperative Zidovudine Study, published February 13, 1991 in the
NEW ENGLAND JOURNAL OF MEDICINE, which found that early AZT use
greatly reduced the progression to AIDS, but did not find any
survival difference:

"In the Veterans Affairs trial, 338 patients with CD4+
lymphocyte [T-helper cell] counts of 0.2 to 0.5 cells x 10(9) per
liter (200 to 500 per cubic millimeter) were randomly assigned to
early zidovudine therapy (at the beginning of the study) or later
treatment (when the CD4+ lymphocyte count fell below 0.2 cells x
10(9) per liter). After two years of followup, the early-
treatment group had significantly fewer AIDS-defining events than
the late-treatment group (25 vs. 44, p = 0.02) but no significant
differences in numbers of deaths (23 vs. 19). The small numbers
of outcomes [deaths], the use of high-dose treatment regimens
(1500 mg per day), and alterations to the protocol during the
study suggest that these mortality data may not reflect the true
treatment effect in persons taking zidovudine at currently
recommended doses (500 mg per day)." [Brackets indicate
explanatory notes added by AIDS TREATMENT NEWS.]

The paper concludes "both antiviral [treatment] and anti-PCP
prophylaxis are necessary to maximize the survival benefit."

For more information, see "The Effects on Survival of Early
Treatment of Human Immunodeficiency Virus Infection," by NM
Graham, SL Zeger, LP Park, and others, The NEW ENGLAND JOURNAL OF
MEDICINE, April 16, 1992, pages 1037-1042.