Sex, Lies, and HIV Transmission
New Book by Medical Psychologist Stuart Brody Concludes That Rectal Intercourse and
Unsterile Needles - Not Vagina Intercourse - Are Real Heterosexual HIV Risk Factors
"The risk of transmitting HIV through vaginal intercourse is near zero
among healthy adults," medical psychologist Stuart Brody writes in his
new book, Sex at Risk: Lifetime Number of Partners, Frequency of Intercourse,
and the Low AIDS Risk of Vaginal Intercourse (Transaction Publishers,
1997), which the Wall Street Journal favorably reviewed on page
A22 of its December 8 edition.
"I'm not saying that it is impossible for unprotected vaginal intercourse
to transmit HIV from a positive to a healthy adult negative partner," Brody
told RA in a telephone interview from his Los Angeles-area home.
"Anything's possible. It's possible to be struck by lightning. But the
two risks share an analogous probability, effectively zero. If healthy,
HIV-negative Americans want to worry about unprotected vaginal intercourse,
they should worry about the drive over to their encounters. If their partners
have never injected drugs or received rectal intercourse or blood therapy,
they are more likely to be killed in an automobile accident on the ride
over than they are to become HIV-positive."
Sex at Risk represents an expansion of several articles Brody
has published in academic journals, including his famous 1995 paper, "Lack
of evidence for transmission of HIV through vaginal intercourse" (Archives
of Sexual Behavior 24:383-393).
Brody does not outright reject a causal role for HIV in AIDS. "I stand
on middle ground on that issue," he told RA. "Although I certainly
recognize that there are several documented cases of AIDS without HIV,
and the near universal presence of other pathogens in AIDS cases, I am
a bit more conservative about HIV than say, [UC-Berkeley molecular biologist
Peter] Duesberg, who considers HIV to be harmless." He also says that he
has not found a single example of AIDS in a person who has been appropriately
cleared of risks, and that the characterization of AIDS patients as "previously
healthy is absolute nonsense."
Many scientists who do outright reject the HIV-AIDS model also
admire Brody. Rather than accept coital transmission without question,
he has applied scientific scrutiny to the published studies claiming to
document it. In a rigorous review of the literature, Brody found no evidence
that a vaginal transmission capacity for HIV had ever been conclusively
established, and is the first authority to stress this point.
He did, however, find documentation that he felt established HIV transmission
via receptive anal intercourse and drug injections. "The data show that
frequency of receptive anal intercourse with an HIV-positive man and frequency
of drug-injecting correlates with seroconversion," Brody says. "And that
makes those activities HIV risk factors. But frequency of unprotected vaginal
intercourse with an HIV-positive person does not correlate with seroconversion,
so that activity does not qualify as a risk factor. Everybody thinks that
unprotected vaginal intercourse with an HIV-positive person will put you
at risk for becoming HIV positive yourself. But this just isn't the case."
Brody is an expert at risk factor analysis, and has worked with physicians
and medical scientists to identify factors which qualify as medical risks,
and to identify potential study subjects who have been exposed to risks
that might confound the results. He says that it is easy to see which risks
must be considered in vaginal transmission studies of HIV. Since anal intercourse
and drug injecting correlate with HIV seroconversion, investigators seeking
to document coital transmission must be very careful to identify and exclude
subjects who are exposed to those factors, and to account for dissimulation
, which Brody defines as "the phenomenon whereby respondents do not reveal
their risk exposures, because they have forgotten about their exposures,
they have misunderstood the question, or -- as is usually the case -- they
have lied to interviewers."
"Research and experience have shown us that people lie often and for
many reasons, and that the content of these lies includes the IV [drug]
and anal intercourse risk factors for HIV transmission," he writes. "Such
lying is one of the factors contributing to an inflated estimate of vaginal
Studies published by the time of his 1995 paper put the coital transmission
frequency for HIV at one per 500 unprotected contacts (male-to-female)
and one per 1,500 (female-to-male) (Blattner, FASEB 5:2340-2348,
July 1991); the latest study revises these figures to one per thousand
and one per 8,000, respectively (Padian N, Am J Epidemiol 146:350-357,
These minute transmission frequencies are "ludicrous over estimates,"
Brody says. He is convinced that -- small as they are -- they would be
even lower if the investigators had taken the proper steps to identify
risk exposures and had accounted for dissimulation.
Since HIV prevalence is high among anal sex recipients and drug injectors,
but low among the general American population, Brody says that researchers
who seek risk-free HIV-positive subjects should expect to attract some
people from the risk categories who deny their risks. "This is a one of
the basic problems in risk factor epidemiology," he says.
And since frequency of vaginal intercourse does not correlate with HIV
seroconversion, studies claiming to document vaginal transmission should
be accepted only if they demonstrate that the study subjects had been properly
screened for factors that do correlate with seroconversion, and
account for dissimulation.
All heterosexual HIV transmission studies to date have relied on questionnaires
or interviews to identify risk exposure, and the investigators have almost
always accepted all denials as accurate. Brody considers this to be sloppy
risk factor analysis. "It is not so easy to get all people who inject drugs
or who participate in rectal intercourse to admit to these activities,"
he says. "Lots of research has been done on the topic of accurate responses
to questions about taboo subjects. It's very embarrassing for both the
interviewer and the respondent. Appreciable numbers of smokers, for example,
will not admit to smoking. It is unfortunate that this knowledge has not
been applied to HIV risk factor assessment."
The studies involve HIV-negative test subjects who engage in regular
heterosexual contact with HIV-positive partners. Participants identify
themselves as monogamous heterosexuals who don't inject drugs. The women
deny or are not asked about receptive rectal intercourse. Participation
requires frequent HIV testing, at which time subjects estimate their number
of unprotected copulations since the previous test. Some studies, like
those of Nancy Padian, compile the number of reported contacts preceding
negative tests in order to calculate transmission frequencies.
A 1993 report of 79 women having regular sex with HIV-positive men is
a typical example (Saracco A, J AIDS 6:497-502). Although each of
the women said they never used condoms, only eight seroconverted after
an average of 1.76 years of observation. Most denied having practiced rectal
intercourse, and all denied having injected drugs. The authors of this
study assumed all risk denials were accurate, and concluded that they had
documented instances of coital transmission.
Yet they stated that frequency of rectal intercourse, but not coitus,
correlated with seroconversion. In light of this, plus what is generally
known about dissimulation, Brody concludes that the seroconverting women
in such studies who denied both activities are lying.
Brody demonstrates that about one million Americans inject drugs, and
that about 10% of all American females, as well as some unknown fraction
of men, have experienced receptive anal intercourse. "A total liar rate
of 5% is more than adequate to account for all the cases of HIV transmission
and AIDS which are classified as heterosexual."
Brody concludes that the rare seroconversion observed in HIV coitus
transmission studies represents participants who have not been forthright
in their self-appraisal of risk exposure, or who have not been asked the
specific questions that would have identified their risks. For example,
he says that men who have been raped in prison might not consider themselves
to have ever had sex with a man. The question, "Have you ever had sex with
a man?" would not identify all these recipients of anal intercourse.
Brody says he's not found any properly constructed studies on the topic
of heterosexual HIV transmission. He suspects that political and social
pressures compel researchers to accommodate as best they can the politically
correct assertion that "everyone is at risk for HIV and AIDS."
"HIV researchers who publish these papers do not seem to be serious
about accurately accounting for anal intercourse and drug injecting," he
says. "The very studies that claim to document vaginal transmission show
that coitus frequency does not correlate with seroconversion, but that
frequency of receptive anal intercourse does." A properly constructed study
would recognize that an apparent case of coital transmission might really
be a case of anal transmission in disguise. But the heterosexual HIV transmission
studies that Brody has found do not do this.
Brody's analysis has a significant effect on the HIV risk typically
advanced by critics, which is based on official transmission rates (1-per-1,000
man-to-woman and 1-per-8,000 woman-to-man) and HIV prevalence in the general
US population (1-per-5,000 men and 1-per-10,000 women). Using these figures,
healthy, drug-free Americans having regular unprotected coitus with positive
partners have barely any chance of becoming positive. And single acts of
unprotected coitus with random, drug-free partners have preposterously
low risks of seroconversion: 1-in-10 million for women (1/1,000 x 1/5,000),
and 1-in-80 million for men (1/8,000 x 1/10,000). These risks are less
than that of being struck by lighting during the course of a year, which
is one in a million (250 of 250 million Americans are struck by lightning
each year; Discover , May 1996, p. 82).
Brody predicts that if researchers took care to compile accurate data,
they would find that the actual unprotected coital transmission frequencies
would each be on the order of the annual American lightning risk, around
once every million contacts.
And so would be the chance of finding an HIV-positive person from among
the general American population. The best data available for the HIV prevalence
in the general American public are those tracked by the CDC for first-time
blood donors, Gays and drug injectors are officially discouraged from donating,
and potential blood donors are given questionnaires that ask about histories
of drug injecting and male homosexual sex. Positive respondents are rejected.
which show an HIV rate of one-per-7,500 for the most recent year, 1994
(one-per-5,000 men, and one-per-10,000 women). "But women are not asked
about rectal intercourse, there is the same problem with dissimulation,
blood donors are not a representative sample of the population, and many
people from the risk groups use blood donations as a way to get free HIV
tests," Brody says. "So I expect the true HIV rate for Americans free of
injected drugs and receptive anal intercourse to be far less than one in
7,500." He expects that accurate evaluation would identify no more than
one HIV-positive risk-free American per million.
Combining those figures, Brody says that a risk-free American who has
a single act of unprotected coitus with a random risk-free partner is about
as likely to become HIV-positive as "be struck multiple times by lightning
in one year, or win several state lotteries."
Scientists and physicians ignore these facts, and instead promote the
politically correct idea that "everyone is at risk." "Ideological knowledge
about AIDS is far more likely to filter through society than scientific
knowledge," he writes. The idea that HIV is vaginally transmitted was developed
before any data were generated. Now that data have been generated, they
do not support this concept, though most researchers refuse to reconsider
For instance, in some African regions, researchers find relatively
high HIV-positive rates in the general population (about 5%), and assume
this is due to coital transmission combined with rampant promiscuity. According
to Brody, vaginal transmission is just too inefficient to explain high
HIV rates in a heterosexual population. He offers a better explanation
for high African HIV rates: the widespread use in underdeveloped regions
of unsterile hypodermic equipment to provide vaccinations and blood transfusions.
This "effectively places the African patient in a situation not unlike
that of IV-drug users anywhere sharing unsterile hypodermic equipment."
He also notes that "starvation, pathogen-laden drinking water, and rampant
tropical infections such as malaria make for different immune functions
than those of healthy Westerners."
In industrial nations, like the United States, where HIV-positive rates
are low in the general population, health officials, professors, and researchers
attribute this to relative sexual moderation among American heterosexuals.
The data do not support this view, either, Brody says. Since frequency
of coitus does not correlate with seroconversion, it doesn't matter how
promiscuous a population is, so long as anal intercourse is not a central
activity. For HIV-positive Americans who have neither been transfused nor
treated for hemophilia, Brody considers receptive anal intercourse and
unsterile needles to account for effectively all cases officially listed
as "heterosexual HIV transmission."
Brody considers a positive HIV test (within the bounds of its accuracy)
to effectively indicate receptive anal intercourse or injections with unsterile
needles in cases where in-utero transmission and transfusions can be excluded.
This is not to say that everyone who injects drugs or receives anal intercourse
will test HIV-positive. In some cities, the rates can get as high as 30%
for gay men and 50% for drug users, but never near 100%. How-ever, the
data support the following conclusion: effectively 100% of Americans who
test HIV-positive have received anal sex, injected drugs, had a blood transfusion,
been treated for hemophilia, or been prenatally exposed, he says. The problem
is that far less than 100% of these people will admit to the activities
that involve sex or drugs.
None of the HIV coital transmission studies so far have addressed Brody's
main arguments. Moreover, no studies have compared the tiny proposed HIV
transmission frequencies to those of other microbes known to be coitally
contagious, such as chlamydia and gonorrhea, which transmit about half
the time (one per two).
Brody blames "politically correct thought" which "impedes scientific
progress when taboo themes, regardless of their validity, cannot be pursued."
He views the "everyone is at risk for HIV and AIDS" campaign as an attempt
by the American government to "micromanage everyone's lives," and fears
the consequences of subjecting the public to mass screenings with HIV tests.
"When you test people from a population with an extremely low base rate,
a relatively high fraction of positive results are false, due to testing
inaccuracies. In some special cases, and this may be true for HIV, most
of the positive results you get are false. So you run the risk of creating
more heart attacks from false positives than identifying people who really
The following example illustrates Brody's point. The CDC calls the HIV
antibody testing battery (two ELISAs and one Western blot) "greater than
99% accurate" (CDC Fax Information Service, Document 320310, Jan. 1993;
phone 800-458-5231). Putting aside all the criticisms about the validity
of these tests, let's assume an accuracy of 99.9% and consider the implications
of testing a population of 7,500 in which only one person is truly positive.
When we test the 7,499 truly negative people, we should expect 7,499 x
0.999 = 7,493 accurate results: that would mean 7,493 true negatives, and
6 false positives. But when we test that one truly positive person, we
should expect 1 x 0.999 = 0.999, or one, accurate result: a true positive.
So, in addition to the expense and trouble of testing 7,500 people just
to identify a single truly positive person, we should expect to obtain
six times more false positives than true positives.
This inefficiency does not hold true in high-risk populations, though.
In a population where 30% of the people are truly positive, out of 7,500
people, that comes out to 2,250 truly positive people, and 5,000 truly
negative people. If we test the 2,250 truly positive people, we should
expect 2,250 x 0.999 = 2,248 accurate results: that would mean 2,248 true
positives and two false negatives. If we test the 5,000 truly negative
people, we should expect 5,000 x 0.999 = 4,995 accurate tests: 4,995 true
negatives, and five false positives. So for 2,250 true positives, we get
five false positives, and for 4,995 true negatives, we get two false negatives.
"For this reason," Brody says, "HIV tests should be offered only to
people with the risks: gay recipients of anal intercourse, drug injectors,
hemophiliacs, transfusion recipients, and women who have had receptive
anal intercourse with a man from one of the other risk groups, or who is
known to be HIV-positive."
Normally, testing for diseases that cluster in risk groups is limited
to the risk groups. Men and young girls aren't screened for breast cancer,
for example, though some men and young girls have developed breast cancer.
But, as is so often the case, special regard is afforded HIV and AIDS.
In his book, Brody concludes that the "gross exaggeration of AIDS risk
to healthy, non-IVDU heterosexuals is not only psychologically damaging,
but also constitutes unethical behavior on the part of many public health
officials, journalists, and others."
-- Paul Philpott
Sex at Risk is available for $32.95 (232 pages, hardcover,
alkaline paper) in book stores, by phone (888-999-6778), and on the internet
(www.amazon.com). Brody's web page (www.usinter.net/~tcibr/) contains references
to his fifty-plus academic papers.