Læknaneminn - 01.04.1997, Qupperneq 127

The drug-AIDS hypothesis *
sexual men at “highest risk” studied had used nitrites,
in addition to various combinations of 12 other recre-
ational drugs 104.
Because of their complete disregard for the medical
consequences of drug use, most AIDS epidemiologists
do not even look for a drug-free AIDS case although
many acknowledge bewildering drug use (see Tables 4
and. 6). An event at a conference on the role of nitrites
in Kaposi’s sarcoma in 1994 illustrates this bias per-
fectly. Asked whether there was even one AIDS patient
who never used drugs, an investigator of the largest
group of male homosexuals ever studied for “HIV dis-
ease progression,” the MACS cohort, responded, “I
never looked at the data in this way” 96,109. But the
MAC study, which is supported by the NIAID with
several million dollars annually, has repeatedly record-
ed heavy drug use for over 10 years (Table 5) '°3, l04, m.
However, until drug-free controls are available, con-
clusions that HIV rather than drugs cause AIDS are
un-informed speculations. In fact the sheer multiplic-
ity of epidemiological studies describing “HlV-disease
progression” only in drug users from San Francisco S0,
'°2, Vancouver 102,339, Chicago - Los Angeles - Balti-
more - Pittsburgh I03, l04, Sydney 102, Milan 93, Amster-
dam 102, London 106 can hardly be an accident. It sug-
gests that drugs are causing AIDS.
To avoid the pitfalls of confounding variables of
HIV, matched groups rnust be compared that differ
only in one variable 34°. Thus an appropriate statistical
analysis of the role of drugs in AIDS would compare
two groups of HlV-positives (or two groups of HIV-
negatives) matched for all variables but drug use.
Based on Feynman’s standards of science, there are
three contending explanations why so many AIDS-epi-
demiologists have omitted drug-free controls: (a) either
they are ignorant of drug toxicity, or (b) they are igno-
rant of confounding variables in epidemiological stud-
ies, or (c) there are no drug-free AIDS cases, because
drugs cause AIDS.
7.6. Confounding “confounding viariables”. The
Nature commentary also demonstrates the “proper
methods” used by HIV researchers to eliminate “con-
founding variables” such as drug use from the non-con-
founding variable HIV 80.
In view of the “fact” that homosexual men who were
“heavy” nitrite users had twice as much Kaposi’s sarco-
ma as those who were “light” users, the authors argued
as follows: “This crude association is apparently the
basis for Duesberg’s hypothesis. Further analysis of the
data reveals a similar association between drug use and
HIV positivity, and when controlled for HIV serosta-
tus, there is no overall effect of drug use on AIDS. A
similar effect, a marginal association that drops after
controlling for HIV serostatus, is seen in cases which
end in Kaposi’s sarcoma. Thus when proper methods
are used to assess the role of confounding variables,
there is no evidence of a drug effect” 80. With this rea-
soning the article proudly rejected the drug hypothesis
with, “such claims have no basis in fact.” The anti-
drug bias of Nature is so pervasive that the editor open-
ly censored 341 all critics pointing out confounding by
drug use 1H->15.222,342_ However, The Lancet allowed two
critical letters 47,223.
Called to task on the possibility of confounding two
years later in Science, the authors simply restated their
conclusion without lifting the secret of their “proper
methods”: “The standard statistical methods that we
used to differentiate cause from confounding factors
showed, in this case, that HIV was the cause and that
drug-use association was spurious” 337.
In short, Nature has refuted the drug hypothesis by
first commissioning a commentary that relied on AIDS
patients who had all (!) used a multiplicity of recre-
ational drugs in addition to AZT, and then by openly
censoring all objections to its methodological flaws and
unscientific manipulations - a bewildering achieve-
ment coming from the world’s oldest science journal.
7.7. Grouping drug-using with non-drug using HIV-
positives. This manipulation credits the diseases of drug
users to non-drug users within the same study group of
HlV-positive people. For example, HlV-positive babies
who either shared recreational drugs with their moth-
ers or received AZT from their doctors are grouped
with babies who neither received drugs from their
mothers nor AZT, and the diseases of the HlV-positive
“group” as a whole are then compared to those of HIV-
free babies 205’314'322 (see 6.9.). But mothers of HlV-free
babies typically have not used cocaine, nor are HIV-
free babies ever treated with AZT 26.
Likewise, the mortality of groups of HlV-positive
hemophiliacs who on average have received many more
immunosuppressive transfusions than HlV-negatives
LÆKNANEMINN
125 1. tbl. 1997, 50. árg.