CONDOMS AND HIV: WHAT EVERYONE KNOWS IS ONCE AGAIN WRONG

It’s a universally believed shibboleth that using condoms cuts the risk of becoming HIV-positive. For example, a recent exchange in the British Medical Journal (BMJ, 26 January 2008, 336: 184-5) was quoted to the effect that “Consistent condom use can reduce the spread of HIV” (ScienceDaily [Jan. 25, 2008]). That exchange, as well as an earlier editorial in the BMJ (24 July 2004, 329: 185-6), cite for this a workshop held by the National Institute of Allergy and Infectious Diseases in June 2000.

The actual findings are less than unambiguous. According to the Executive Summary in the workshop report: “The published data documenting effectiveness of the male condom were strongest for HIV. The Panel concluded that, based on a meta-analysis of published studies ‘always’ users of the male condom significantly reduced the risk of HIV infection in men and women. These data provided strong evidence for the effectiveness of condoms in preventing HIV transmission in both men and women who engage in vaginal intercourse.”

This is notably less than a definitive claim: “strong evidence” is not the same as “proof”. The report itself acknowledges that nothing new is being revealed: “HIV infection is the only STD for which formal meta-analyses have been published . . . . The most recent analysis by Davis and Weller . . . was evaluated.”

The latter, a meta-analysis (quantitative assessment) of a dozen studies had concluded, “Among those who reported never using condoms, the summary estimate of HIV/AIDS incidence from the seven studies was 6.7 seroconversions per 100 person years”. Now: if there are 67 acts of intercourse per year, that corresponds to a transmission rate of a mere 1 per 1000 unprotected acts; at more than 67 acts per year, it corresponds to an even lower rate of apparent transmission. As often pointed out, such rates are far too low to bring about an epidemic (see, for instance, James Chin, “The AIDS Pandemic”; or Chakraborty et al., AIDS 15 [2001] 621-6).

One can accept without qualms the results of studies that find apparent transmission of HIV not to exceed a few parts per thousand, for higher rates would be more readily observed. But one could well have qualms about how accurately one could specify such very low apparent rates. Consider how many people must be observed over how long a period of time in order to make such a measurement reliably. If you study 100 discordant couples (one HIV-positive, the other not) for a year, you might observe a few seroconversions on average, but owing to the usual chance variations, to get a quantitatively meaningful rate one would need to observe many hundreds of discordant couples. That this is far from easily done is illustrated by the most cited such study (Padian et al., American Journal of Epidemiology, 146 [1997] 350-7), which managed to enroll fewer than 450 such couples over a ten-year period.

For a good measurement of yet lower rates than a few per thousand, even larger numbers of discordant couples would have to be enrolled; which makes one wonder how meaningful could be the finding of the meta-analysis that always using condoms reduced the rate to 0.9 seroconversions per 100 person years, in other words a few per 10,000. Nevertheless, these numbers have been cited to claim that condom use can reduce transmission of HIV by 85%.

On the other hand, several studies have not supported the claim that condom use decreases apparent transmission of the “HIV-positive” condition at all (see pp. 44, 109, 115 in The Origins, Persistence and Failings of HIV/AIDS Theory).

To those earlier-cited data one can now add remarkably detailed official figures from Rwanda’s Demographic and Health Survey (2005 edition, published July 2006).

The suggested reason for this overwhelming contradiction of the conventional wisdom seems a little thin: “It is difficult to establish the exact relationship between condom use and HIV. Condoms could be used by those who are HIV negative to protect themselves from the disease, but they could also be used by those who are seropositive to protect their partners” (p. 236); in other words, it’s being suggested that those who are already HIV-positive might be more likely to use condoms.

One might also speculate that condom use could be more frequent among those who have sex often or with many partners, and whose greater consequent risk is not balanced by the protective effect of the condoms. But there seems to be no correlation between rate of testing HIV-positive and sexual behavior, for example, “Paradoxically, it is not only women who have engaged in higher-risk sex, but also those who had no intercourse in the past 12 months that prevalence rates are the highest (8.2 percent and 8.9 percent, respectively)” (p. 235). While HIV-positive rates are reported to rise somewhat with numbers of sexual partners, on the other hand for men who reported no sex in the past 12 months the HIV-positive rate was 2.9%; for those who had higher-risk sex in the last 12 months, it was a bit less, 2.7%; and for those who had had lower-risk sex, it was highest at 3.5%. In other words, the apparent relationship between sexual behavior and HIV-positive rate is so inconsistent or even itself paradoxical as to exclude this as a plausible explanation for the comparative figures in the tables shown above.

The facts seem quite clear. There is no indication that use of condoms decreases the risk of becoming HIV-positive. What’s more, there is no correlation between becoming HIV-positive and having sex, not having sex, or having high-risk sex.

Here is yet another piece of evidence, from official data, that HIV tests do not track a sexually transmitted condition.

(Nothing here suggests, of course, that one shouldn’t use condoms. They offer some protection—though far from certain—against a number of actual STDs like syphilis or gonorrhea; and they are a fairly effective—though by no means certain—means of contraception.)