Safety in Numbers

Chapter 3

Promoting Safer Sex

Oral sex

One of the most controversial areas in safer sex advice is the issue of the degree of risk involved in oral sex. It is widely believed by AIDS educators that it is impossible to give firm advice about HIV transmission by this route. However, there is reliable epidemiological evidence to show that the risk of HIV transmission through oral sex is very small, and that the practice can, happily, be described as a form of safer sex. Objections to this classification stem largely from misconceptions or discomfort with the notion of risk reduction – as opposed to risk elimination – which is implicit in the concept of safer sex.

Some of the disagreements about the riskiness of oral sex are derived from misunderstandings of the basis for our knowledge about the ways in which HIV is transmitted in practice. As described in Chapter 2, safer sex guidelines were originally formulated before the cause of AIDS was established. At that time the most prominent three theories focused either on factors relating to gay men’s ‘lifestyle’ such as drug use; or on the effects of repeated infection with known pathogens such as CMV; or on a new transmissible agent, either alone or in conjunction with co-factors. Safer sex advice was based on the latter two, and assumed that the cause or causes of AIDS was sexually transmitted, and that transmission could be stopped by preventing one partner’s body fluids from entering the other’s body. While these guidelines have proved to be essentially correct, accumulating epidemiological evidence has provided increasingly clear indications of the relative risk of particular sexual acts, and allowed educators to become more sophisticated in their recommendations for safer sex.

It is not the case that our understanding of the ways in which HIV can be transmitted is based on laboratory research. Although test-tube studies are able to identify which body fluids contain HIV at infectious levels, they provide little or no information on the risk of transmission through particular acts. Discussing the possibility of oral transmission of HIV, Canadian researchers pointed out that:


The isolation of HTLV-III in a body fluid or tissue does not prove that the fluid or tissue represents a mode of transmission. Factors which determine whether infection is transmitted include the concentration and viability of the agent within the fluid or tissue, access to a port of entry for the fluid or medium, the presence of receptors at the site of entry, and natural host defences near the site of entry. (41)

The physical mechanisms by which HIV enters someone’s body and infects him or her are still not fully understood, and there is still no entirely convincing non-human animal model for researchers to study; laboratory scientists therefore tend to rely on epidemiology to indicate the ways in which infection occurs in practice, and then adjust their theories accordingly.(42) The assertion that more research is needed to understand particular issues in HIV transmission, which is most commonly made in relation to woman-to-woman transmission, is thus highly misleading if it is taken to refer to bench science as opposed to epidemiological surveillance.

Gerald Oppenheimer has described how epidemiologists identified the routes by which HIV is transmitted among gay men, essentially by identifying those men in cohort studies who were HIV-positive, and then analysing how their sexual behaviour had differed from men who were not infected:


[Risk factors] included sexual contact with a person known to have AIDS and participation as the receptive partner in anal intercourse, a risk that increased with the number of persons with whom one acted as the anal receptive partner. These behaviours heightened the chances of viral transmission. Implicated as well was a history of anal douche use. In the population studied, therefore, HIV infection is an STD in which anal mucosa appears to be an inefficient barrier to infection, especially when traumatised by frequent contact. These results [were] consistent over many epidemiological studies.(43)

The potential difficulty is that of distinguishing between a high risk activity, such as receptive anal sex, and a lower risk activity such as oral sex. In these studies, the vast majority of the men who have had unsafe anal sex have also had oral sex, which may seem to make it hard to know which activity is actually responsible for transmitting HIV. However, in virtually all such studies, a large number of uninfected men also report having had oral sex. It is the fact that HIV transmission is typically not shown to occur unless someone has unprotected anal sex, regardless of whether or not they also have oral sex, which proves that anal sex carries by the greatest risk, and that any risk in oral sex is very low indeed.

Another means by which it may be possible to distinguish between activities with different levels of risk is through multivariate analysis. This is a statistical approach which makes explicit adjustments to its calculations to take account of this type of possible confounding effect. Virtually all of the studies described below used multivariate analysis in demonstrating that there is little or no risk of HIV transmission associated with oral sex.(44)

For example, in 1984, Danish researchers analysed the correlation between specific sexual acts and risk of HIV infection among 250 gay men, and found that “[t]he number of acts as the recipient of anal intercourse correlated with seropositivity whereas being a recipient of oral intercourse ... did not appear to be important”, or if anything was “slightly protective” since men who did not practise sucking proved to be more likely to have passive anal sex.(45)

A study of 304 gay men in London during 1985 found that “the swallowing of semen did not show any significant relationship with anti-HTLV-III/LAV seropositivity”.(46) Researchers in San Francisco reported that there was “no significantly increased risk of HIV infection by ... oral/genital contact with ejaculation”. To check this conclusion, they collected data on 64 men who reported no anal sex or only oral sex since June 1982, and found that the 14 who were HIV-positive all reported having had receptive anal sex before June 1982. They concluded that “[t]he data from the San Francisco Men’s Health Study confirm that receptive anal/genital contact is the major mode of transmission of HIV infection. In fact, there was no evidence of epidemic spread due to any other sexual mode of transmission”.(47)

Very similar findings were reported from the Canadian cohort of 700 gay men known as the Vancouver Lymphadenopathy-AIDS Study, in which, “in multivariate analyses, no risk associated with oral sexual contact was detected”. As confirmation, researchers studied 21 uninfected gay men who reported no receptive anal sex during the year prior to enrolment in the study or during subsequent follow-up: all still practiced oral sex. They concluded that:


The sexual practices of the 21 men we studied, the number of their partners and the prevalence of HTLV-III seropositivity in homosexual men in Vancouver (at least 35%) combine to suggest that during the observation period these men received frequent oral exposure to HTLV-III. Yet only 1 man seroconverted and this probably happened through insertive anal intercourse, a known mode of transmission. Our findings corroborate the lack of oral transmission of HTLV-III.(48)

No significant risk of HIV infection was found to be associated with oral sex in the Multicenter AIDS Cohort Study in the USA. Of 147 men who performed oral sex with at least one partner during a six-month period, but who reported no anal sex during the preceding year, not a single seroconversion occurred. All but 3 of the 95 men who did become infected during this period had had unprotected passive anal sex during the twelve months before they tested HIV-positive, and the remaining three had been the insertive partner during unprotected anal sex.(49) In 1989 an update on the seronegative gay men in the Multicenter AIDS Cohort Study reported that:


an increasing number of men over the course of follow-up in the MACS have given up anal-genital intercourse while continuing to practise other sexual activities. If these practices were associated with an appreciable risk of infection over time, we should have observed seroconversions due to these other exposures (. . .) There was only one man in this study who appears to have been infected by sexual activities other than anal-genital intercourse. The accuracy of this report, is of course, dependent upon the accuracy of recall of the individual. The motivation for denial of anal-genital intercourse increases as it becomes more and more apparent to the homosexual community that this is the major route of infection.(50)

The researchers came to the conclusion that, on the basis on this individual’s report, “such transmission is possible, albeit rare”.

This assessment is also confirmed by the patterns of HIV transmission seen since large numbers of gay men started practising safer sex. Surveys show that although many men have avoided unprotected anal sex, whether by using condoms or by giving it up altogether, considerable numbers have continued to practise oral sex. As early as 1984, researchers in Vancouver reported that over a three-month period, “[s]pecific sexual practices were more likely to have decreased than increased, except for oral-genital contact”.(51) Although the proportion who avoid getting semen in their mouth has increased(52), only a tiny proportion have started to use condoms during sucking. In a sample of 229 sexually active British gay men who responded to a questionnaire in 1988, 180 (79%) had practised oral sex during the previous year, 117 to ejaculation. Of these, 103 (88%) never used a condom, the commonest reason being that they did not feel that it was necessary.(53) In some studies oral sex has become much more common in recent years: in Project SIGMA, the proportion of men who had oral sex in the month before they were interviewed rose from about 60% in 1988 to about 75% in 1991.(54)

But nevertheless, HIV transmission rates have fallen dramatically, in line with the reduction in unprotected anal sex. As Adam Carr has pointed out:


oral sex is now a far more common practice among gay men than unprotected anal sex. In cities like San Francisco and Sydney, sexually-active gay men are statistically likely to have frequent sexual encounters with HIV-infected partners. Seroconversions due to unprotected anal sex with HIV-infected partners continue to be documented; the Sydney study documented 55, or 8.5% of their total cohort, within four years. This makes the lack of documented seroconversions among the large majority of gay men who do not practise unprotected receptive anal sex with casual partners, but who do practise receptive oral sex, all the more striking.(55)

In virtually every instance, gay men who become infected with HIV today report having had unprotected anal sex. For example, in 1991 researchers on Project SIGMA reported that:


no-one in the study who has not engaged in fucking is HIV antibody positive and all the men who are positive have engaged in fucking. In a case study of the 7 sero-conversions which have happened during the course of the Project's first 4 years, 5 of them can be linked to being fucked without a condom and the other 2 to fucking someone else without a condom.(56)

Despite the increase in oral sex among study participants, SIGMA found "no statistical relationship between HIV status and the general number of sexual partners, which if sucking were implicated would be the case".(57) Similarly, investigators with the cohort study at St Mary’s Hospital in London concluded in summer 1992 that “[t]he observation of declining rates of seroconversion despite the fact that a large proportion of seronegative subjects (greater than 75%) have continued to practise oral sex suggests that this practice carries very little or no risk for HIV transmission”.(58)

It is also helpful to reflect on the established modes of transmision of hepatitis B virus (HBV). In a two-and-a-half year prospective study of gay men recruited into the Pittsburgh arm of the Multicenter AIDS Cohort Study, it was established that HBV is transmitted 8.6-fold more efficiently than HIV between gay men, but nevertheless, no association between receptive oral sex and HBV transmission could be detected.(59) This is consistent with earlier research on HBV transmission(60), and suggests that the risk of HBV transmission during oral sex is at best very small, and that the risk of HIV being passed on in this way is likely to be even lower.

It is on these data that safer sex advice should be based. Much of the confusion about oral sex may derive from the varying advice given in safer sex materials published by different organisations in different countries at different times. However, rather than reflecting confusion and uncertainty, the evolution of those messages should actually be seen as an indication of increasing knowledge and growing confidence that oral sex can be considered safer sex.(61)

Early safer sex advice, developed during the years before HIV was discovered, erred on the side of caution by advising against exposure to any body fluid, including saliva during kissing. For example, in How to Have Sex in an Epidemic: One Approach, Callen and Berkowitz warned that “sucking is a moderate risk for CMV ... If you suck, you can reduce your risk for CMV by preventing your partner from coming in your mouth ... Apart from CMV, sucking can of course transmit other diseases such as syphilis and gonorrhea and in certain settings amoebas”.(62) By 1985, it had been established that the recently discovered HIV, then known as HTLV III or LAV, was indeed transmitted in blood or semen, but there was still doubt about the infectivity of other body fluids like saliva and urine. The Terrence Higgins Trust guidelines on safer sex from that time ranked both oral sex and deep kissing as ‘medium risk’, in what was described as an “educated guess”.(63) The precise means by which someone could become infected with HIV was still a complete mystery at this time: did it merely have to come into contact with mucous membranes (and if so, did the mucous membranes of the mouth differ from those of the rectum), or did it actually have to enter the bloodstream through cuts, sores or abrasions? The contradictory opinions on this are reflected even in different Terrence Higgins Trust materials from 1985: the Medical Briefing simply warned against letting HIV “enter the body”(64), while the Sex... leaflet for gay men was less stringent, advising that “Sucking is OK - but cum or pre-cum in the mouth is risky if there are cuts or sores on your cock or gums”.(65)

By 1988, AIDS organisations in Canada, Australia and the UK - but not the USA - were sufficiently persuaded by the consistent data emerging from cohort studies that their advice about oral sex was significantly relaxed. London Lesbian and Gay Switchboard, whose advice on oral sex had classified it as only a possible low risk for some years, now placed full-page advertisements in the gay press encouraging gay men to ‘SUCK’. The Canadian AIDS Society convened a panel of experts which reclassified oral sex as low-risk; in the words of one researcher, the Canadian government “decided to give oral sex back to gay men”.(66) The AIDS Committee of Toronto designed a major campaign with erotic leaflets and posters to reassure gay men about sucking; their director of education maintained that they “could not equate the risk of unprotected anal sex with unprotected oral sex (. . .) Oral sex is closer to no-risk”.(67) In Germany, Deutsche AIDS-H�lfe published a explicit poster illustrating sucking, with the slogan “Nicht in den Mund Spritzen” - “Don’t cum in his mouth”.

The two remaining contentious areas were the significance of cuts or sores around the mouth or gums, and the advice about getting semen or pre-ejaculatory fluid in the mouth. The latter is the most straightforward: it is clear that HIV cannot be transmitted from the skin of the penis, and that there is no risk of infection when there is no contact with cum or pre-cum. However, HIV is present in both fluids, although it is not known whether the levels in pre-cum are sufficiently high to be infectious.(68) Some advice has therefore made a quite reasonable distinction between oral sex in which no cum or pre-cum is taken into the mouth, when HIV transmission is impossible, and sucking to ejaculation, where there at least grounds for suspicion that infection could occur. Most researchers appear not to have performed separate analyses of the risk associated with oral sex both with and without the ingestion of semen. However, it is reasonable to assume that a fair proportion of incidents of oral sex in the studies described earlier will have involved ejaculation into the mouth, yet nevertheless, very little or no risk of infection was observed. Moreover, in some studies the acts of ‘oral receptivity’ and swallowing semen have been differentiated, without revealing evidence of transmission of HIV.(69)

One factor which may play a role in preventing HIV transmission through oral sex is the inhibitory properties of saliva. In May 1988 the Journal of the American Dental Association reported that an unidentified component of saliva prevented HIV infection of lymphocytes, even when greatly diluted.(70) It has also been demonstrated that HIV is rapidly destroyed by bile in the intestines, and researchers assume that the digestive enzymes and changes in pH (acidity/alkalinity) in the stomach which inactivate other viruses would also kill HIV. Surveys have shown that many gay men have based their decision to continue having oral sex on these pieces of evidence.(71)

The remaining question, however, is concerned with whether any risk as a result of getting semen in the mouth is increased if there are cuts or sores in the mouth. First, it should be remembered that oral sex has been shown to be associated with very little or no risk of infection in cohort studies, and that it is reasonable to assume that in terms of their oral hygiene, the men in those cohort studies will not have had significantly more or fewer cuts and sores in their mouths than any other cross-section of the gay population. As described by a Briefing Paper on oral transmission prepared for advisors working on the National AIDS Helpline in Britain:


The mouth is an entry point into the digestive system where enzymes in saliva are secreted to begin the task of digesting food. It is very thick compared with the mucous membrane in the rectum, and fluids stay in contact with it for a very short time because swallowing clears the mouth regularly. Thus the likelihood of semen, blood or vaginal fluid being absorbed through damaged tissue is minimal.(72)

However, in the isolated case reports in which infection is claimed to have taken place through oral sex, poor oral health has been reported on a number of occasions.(73) It may therefore be reasonable to distinguish between the low risk from getting cum or pre-cum in a healthy mouth, and a possibly slightly greater but still low risk if the tissue of the mouth is damaged. Some reports have also suggested that any risk of infection through oral sex may also be somewhat increased if the mouth or throat are compromised in other ways, such as through inflammation due to infections such as pharyngitis(73) or gonorrhoea.(74)

It is those case reports which have caused the greatest amount of confusion and concern about the safety of oral sex. As Nicholas Mulcahy has observed:


A case study is the tale of an individual - in this instance, one who evidently becomes infected with HIV from sucking cock. Because it personalizes an issue, a case study can have a more powerful effect on the imagination than a study with hundreds of participants. However, the sexual histories of ten individuals are not a good basis for determining the riskiness of oral sex. Large-scale epidemiology studies reduce the chance that errant or omitted details in a person’s sexual history will distort a larger reality.(76)

Reports of such case studies are usually made in letters to medical journals, which unlike formal articles are not peer-reviewed. Thus, some have included such eccentricities as the suggestion that HIV transmission may take place from the partner performing the sucking.(77) Moreover, they rely on the testimony of individuals, who may be understandably reluctant to admit to practises about which they feel guilty or embarrassed, especially if they have had unsafe anal sex which is increasingly stigmatised among gay men. One Dutch study assessed the sexual activities of 102 HIV-positive gay men for whom the date of infection was known. In a written questionnaire, 20 men denied having had receptive anal sex during the 6-9 months before seroconversion; however, in face-to-face interviews, 11 subsequently contradicted this. A further five had been the active partner in anal sex. The researchers concluded that “orogenital transmission of HIV does appear to occur, but a psychological barrier in reporting the practice of anogenital receptive intercourse may lead to an overestimation of the transmission rate”.(78)

Thus, these isolated cases often attract attention which is out of all proportion to their significance, while the more reliable conclusions from large cohort studies are overlooked. Doctors have been known to argue that just a single case of alleged oral transmission is sufficient “to show that we can’t assume that it’s safe any more”.(79) This reflects a fundamental misunderstanding of what is meant by the term safer sex, as opposed to 100% safe sex, which is discussed next.