The possible risk

Berrey MM et al. Oral sex and HIV transmission (Letter) J AIDS Human Retrovirology. 14: 475-477, 1997.  Gay man seroconverted after tonsillectomy and several episodes of pharyngeal bleeding; he reported 20 ROI partners without ejaculation, 2 of whom were HIV-positive, and one instance of protected RAI.

Chen W et al: Allergy, oral sex, and HIV (Letters), Lancet. Mar 7. 339(8793): 627-628,1992. The authors report two cases of gay men who seroconverted following negative tests on previous occasions and no other risks apart from receptive oral intercourse with ejaculation. Both had a history of allergies which may have caused inflammation of the throat, facilitating HIV infection through this inflamed tissue. The authors speculate that such inflammation may have facilitated HIV infection, but do not rule out the possibility of false negative antibody test results or failure to report higher risk activity.

Dillon B et al. Primary HIV infections associated with oral transmission. Abstract 473, Seventh Conference on Retroviruses, San Francisco, 2000. The study involved 122 infections identified in San Francisco between 1996 and 1999, of which 20 were attributed to oral sex by the infected individuals. When doctors questioned patients more closely, 12 cases were excluded because HIV exposure through unprotected anal intercourse could not be ruled out. However, eight men had no other potential risk factors, and all said that they viewed oral sex as carrying little or no risk for HIV infection.

Edwards SK et al: HIV seroconversion after orogenital contact with successful contact tracing, Int J STD & AIDS 6: 50-51, 1995. A report of a man who developed an acute seroconversion illness three months after orogenital sex. Ejaculation did not occur. The index case subsequently tested positive. Other sexual partners of the infected man in the previous seven months had all tested HIV-negative. The infected man developed increasing antibody responses during the period of hospitalisation for seroconversion illness. The authors report that the infected man had received immunosuppressive chemotherapy treatment for non-Hodgkins lymphoma during the previous few months, and they speculate that this treatment may have reduced the production of HIV inhibitory factors in his saliva. The infected man had no history of oral ulceration.

Gill SK et al: Transmission of HIV -1 infection by oroanal intercourse, Genitourinary Medicine 68(4), pp 254-256, 1992. A report of a case in which HIV is believed to have been transmitted through rimming, and in which the index case was insertive with his tongue.

Goldberg DJ et al: HIV and orogenital transmission, Lancet 2(8624) p1363, 1988. 'A 29 year old homosexual man was diagnosed as HIV seropositive in October 1986. He practised orogenital sex only, and had never had anal intercourse. His only apparent risk exposure was unprotected orogenital sex with an HIV-seropositive man while visiting Canada in 1985.'

Keet IP et al: Short communication: Orogenital sex and the transmission of HIV among homosexual men, AIDS 6:223-226, 1992.

Landor M and Cregler LL: What is safe sex? American Journal of Medicine, 84:1 p175, 1988. A letter reporting a case of a man's penis being cut by his female partner's orthodontic brace during oral sex. However the article is only speculative of possible risk of transmission through saliva, based only upon the earliest report of the isolation of HIV in saliva. No actual association is reported between these orthodontic braces and transmission.

Lane HC et al: HIV Seroconversion and Oral Intercourse, AJPH 81:5 p658, 1991. A report of a 35 year old gay man who seroconverted following receptive oral sex without ejaculation; he reported no sexual contact with another man for a month prior to the incident to which infection was attributed. The authors note that anal intercourse with a condom took place, and that this can be confirmed by a third party.

Lifson A et al: HIV seroconversion in two homosexual men after receptive oral intercourse with ejaculation: implication for counselling concerning safe sexual practices, AJPH 1990 80; 12 pp 1509 - 1511, 1990. Two gay men in the SF cohort who reported no anal intercourse in the previous five years seroconverted after receptive oral intercourse with ejaculation. One of the men had a history of gum disease and had reported ten occasions of ejaculation after receptive oral intercourse in the 11 months preceding a positive test. The other also had milder gum disease, and reported 900 occasions of receptive oral intercourse with ejaculation. The researchers note that gingival recession was common in study participants over 35.

Marmor M et al: Possible female to female transmission of HIV. AIM 105:969, 1986. The authors report the case of a 26 year-old woman who developed lymphadenopathy one month after orogenital contact with a lesbian parental drug user with KS.

Mayer KH and DeGruttola V: Human immunodeficiency virus and oral intercourse. AIM 107:3 p428-9, 1987. A letter reporting that `1 of the 8 seroconversions that occurred among 290 initially seronegative men followed at the Fenway Community Health Center in Boston appears to have resulted from oral exposure to semen. The person who seroconverted reported no anal contact, receptive or insertive, since 1982 and had had three negative test results before the first positive test result in January 1987. He did report frequent oral intercourse with a partner known to be seropositive.

Monzon O et al: Female-to-female transmission of HIV, Lancet, ii, 40-1, 1987.

Murray AB et al: Coincident acquisition of Neisseria gonorrhoeae and HIV from fellatio (Letters), Lancet 338 (8770). p 830, 1991. The authors report a case of a gay man who had receptive oral intercourse three weeks and eight weeks before seroconversion illness, with no other reported risks apart from ano-receptive intercourse at least eight years before. He was diagnosed with pharyngeal gonorrhoea and HIV seroconversion illness. The authors speculate that undiagnosed gonorrhoea may have facilitated HIV infection. No information about the serostatus of the casual partner(s?) is available. This is the only case ever reported where pharyngeal gonorrhoea is argued to have played a role in HIV infection, despite many hundreds of reported cases of pharyngeal gonorrhoea evidently acquired through oral sex each year.

Page-Shafer K et al. Sexual behaviour and risk factors for HIV-1 seroconversion in homosexual men participating in the tricontinental seroconverters study, 1982-1994. Am J Epidemiology 146(7): 531-542, 1997. The authors combined data from five cohort studies and looked at risk factors for seroconversion in three discrete periods: 1982-84, 1985-87, 1988-94. 345 seroconverters were matched with 345 controls, and a significant association was shown between the number of receptive oral partners (0, 1,2-9, greater than 10) and risk of seroconversion, and a trend towards comparable relative risk was noted when risk of seroconversion by ROI was analysed by number of RAI partners (0 or 1).

Page-Shafer K et al. 2001 National HIV Prevention Conference

Perry, S: Orogenital transmission of HIV, AIM 111 (ii) pp 951-952, 1989.

Puro V et al: Male-to-female transmission of human immunodeficiency virus infection by oro-genital sex, J Clinical Microbiology and Inf Dis 10 (1) p47, 1991. Two women are reported to have been infected through receptive oral intercourse; no other risk factors were identified. Both male partners had herpes and one was reported to have had bleeding lesions on his penis; the other was p24 antigen positive and in the late stage of HIV disease.

Quarto M et al: HIV transmission by fellatio (Letter), Eur J Epidemiol 6(3) pp 339-40, 1990. A man was reported to have seroconverted after a single episode of fellatio with a casual female partner; his regular female partner remained seronegative.

Rozenbaum W et al: HIV transmission by oral sex, Lancet p1395, 1988. A letter reporting five cases of HIV transmission apparently through oral sex (with the claim that two must have been infected through insertive rather than receptive oral sex). This report has a number of methodological problems. Firstly it claims to have been part of a cohort analysis which has nowhere been published, i.e. its methodology has not been subject to scrutiny by peer review. Additionally the information given is rather skimpy. The reader is simply asked to accept that `We assessed the likelihood of these patients denying anal sex for psychological reasons, and concluded that this was unlikely: 3 patients had freely described participating in anal sex previously...' There is, in other words no evidence that the methodology was sufficiently adapted to account for lack of patient recall, or for a `shame factor' that could easily prevent someone `confessing' to unsafe sex. The fact that someone could happily admit to having had unsafe sex before joining a safer sex study by no means precludes this kind of guilt at having unsafe sex after joining a study. Additionally the mean time of retesting following a negative HIV antibody test was only 5.4 months even though 3 of the patients `reported a mononucleosis-like syndrome' (which would suggest recent infection if this represented the seroconversion illness). However, no clinical confirmation is suggested. Finally we are not even told the size or duration of the cohort study. A response by Dassey et al was published in The Lancet, Oct 29 1988, p1023.

Sabatini MT et al: Kaposi's sarcoma and T-cell lymphoma in an immunodeficient woman: a case report, AIDS Research 1(2) pp 135-157, 1983/4. The case of a 37 year old lesbian who died of an AIDS related illness; both she and her female partner denied sexual contact with men or intravenous drug use; no explanation has been discovered for how either partner acquired HIV infection, but the means of transmission from one partner to another is assumed to have been orogenital contact.

Salahuddin SZ et al: HTLV-III in symptom-free seronegative persons,. Lancet ii p1418-20, 1984. A case report of the wife of a man who was infected by transfusion and who was impotent. The authors speculated that HIV may have been acquired through deep kissing.

Spitzer P et al: Transmission of HIV infection from a woman to a man by oral sex, NEJM 320; 4, p251, 1989. The case of a 60 year-old man who claimed only to have engaged in mutual genital contact with a woman sex worker. Infection was subsequently established to have been through injecting drug use.

Studies showing no detectable risk of oral transmission

Coates RA et al: Risk of HIV seropositivity in relation to specific sexual activities of sexual contacts of men with AIDS or ARC. Presented at 3rd International Conference on AIDS, Washington, June 1987.

Darrow WW et al: Risk factors for human immunodeficiency virus (HIV) infections in homosexual men, Am J Public Health 77:479-83, 1987. A study of 492 men who had participated in a hepatitis B cohort, 1978-1980. Their stored blood samples were compared with new samples. The 240 who had seroconverted to HIV were compared with 119 who had remained seronegative. Of these 240 who seroconverted there were only four men who denied any rectal exposure. Three of these engaged in insertive anal sex. Only one had seroconverted who practised no anal sex whatsoever. Oddly, the authors conclude, very conservatively that `no sexual activities involving exposure to the semen, blood, or excretions of infected persons have been shown to be safe.'

Detels R et al: No HIV seroconversion among men refraining from anal-genital intercourse. Paper presented at the Third International Conference on AIDS, Washington, June 1987.

B Evans BA et al: Sexual lifestyle and clinical findings related to HTLV-III/LAV status in homosexual men, Genitourinary Medicine 62: 384-389, 1986. A study of 304 gay male patients attending a West London clinic. `There was no evidence that HTLV-III/LAV infection measured by seropositivity, was transmitted by oro-anal or oro-genital routes...'

Groopman JE et al: Seroepidemiology of human T-lymphotropic virus type III among homosexual men with the acquired immune deficiency syndrome or generalised lymphadenopathy and among asymptomatic controls in Boston. AIM 102:334-7, 1985.

Grulich AE, Prestage G et al. Oral sex as a risk factor for HIV: a review of Australian data, HARD Conference, Sydney, 2000. This prospective cohort study of over 700 men has found that 95% of men who had casual sexual partners reported having oral sex without a condom. Although receptive oral sex with ejaculation was reported by 26% of men, it was not associated with an increased risk of seroconversion.

Kingsley LA, Detels R, Kaslow R et al: Risk factors for seroconversion to human immunodeficiency virus among male homosexuals, Lancet 1:345-9, 1987. Results from the Multicenter AIDS Cohort Study (MAC) in the USA of 4955 men. 3.8% (95 men) seroconverted and this could be clearly linked to receptive anal intercourse. `No HIV seroconversions occurred in 220 men who did not practice receptive or insertive anal intercourse within twelve months before the follow-up visit. On multivariate analysis receptive anal intercourse was the only significant risk factor for seroconversion to HIV, the risk ration increasing from 3-fold for one partner to 18-fold for five or more partners.'

Lyman D et al.: Minimal risk of transmission of AIDS associated retrovirus infection by oral-genital contact. JAMA 255:13:1703, 1986. A study of a cohort (SFMHS) of 1034 men in San Francisco selected by multistage probability sampling (rather than being based on clinic or volunteer samples). `The adjusted relative risk for oral/genital contact with ejaculation of 1.01 (P  = O.97) indicates no significantly increased risk of HIV infection by this sexual practice ... [however] the sizes of the subsamples were insufficient to provide the necessary statistical power to reject the possibility of even the small relative risks demonstrated for these practices.

Melbye M et al: Seroepidemiology of HTLV-III antibody in Danish homosexual men: prevalence, transmission, and disease outcome, BMJ 289:573-5, 1987. A study concluding that sexual exposure to a man from a high incidence area and anal receptive intercourse were the only variables associated with HIV infection in Denmark. `Oral receptive intercourse appeared to be slightly protective [against HIV infection].'

Moss AR et al: Risk factors for AIDS and HIV seropositivity in homosexual men, American Journal of Epidemiology 125:1035-47, 1987. A case-control study in which two groups of gay men with AIDS were compared with two groups of controls: randomly selected gay men chosen as neighbourhood controls and gay men chosen as STD clinic controls. Comparison was made between the group of men with AIDS and the seronegative men from the two controls. `There was no consistent evidence in the current study for oral-genital, oral-anal, or other sexual transmission of HIV. This study cannot rule out such transmission, but it suggests that the risk associated with non-rectal transmission must be of far lower magnitude than that associated with rectal sex.

Page-Shafer K et al. Risk of infection attributable to oral sex among men who have sex with men and in the population of men who have sex with men. AIDS 16, 17, 2350 – 2352, 2002. Investigators recruited 239 gay men seeking anonymous HIV testing in San Francisco between December 1999 and 2001. The men were asked to complete a questionnaire about the type of sex they had had in the previous six months prior to having an HIV test. Although over 10,000 men attended for testing and completed the survey, only 239 reported having oral sex exclusively and were eligible for the study.

The San Francisco researchers were also keen to identify the population-attributable risk percentage of oral sex, as even a very small individual risk of HIV transmission from oral sex for an individual could result in a substantial number of infections in the population as a whole.

On average, the men in the study had had receptive oral sex with three different men in the past six months (range 0 – 400). The overwhelming majority, 98%, of oral sex was without a condom and 35% of men reported getting semen in their mouths, 70% of whom swallowed.

None of the men in the study tested HIV-positive, meaning that the individual risk of being infected with HIV by oral sex was zero. As the average number of oral sex partners in the past six months was three, the investigators also calculated the population attributable risk percentage for men with one, two and three partners. Although the population attributable risk percentage increased with the number of partners, it remained extremely low, at 0.18% for one partner, 0.25% for two partners and 0.31% for three partners.

“The absence of HIV infection detected in this sample confirms …that orally acquired HIV infection is rare” commented the investigators, noting that the prevalence of HIV amongst gay men reporting unprotected anal sex was 5.1% showing “the striking difference in the risk of HIV between those who report exclusively fellatio and those who report higher-risk sexual behaviors.”

The investigators believe that a particular strength of their study was the questioning of men about their sexual practices before HIV-testing, as men with positive diagnoses “may inaccurately report higher risk exposures for reasons including social desirability and recall.”

The researchers also suggest that improper condom usage could result in more cases of HIV transmission amongst gay men than oral sex. However, they acknowledge that their study sample was relatively small and “therefore, we cannot rule out the possibility that the probability of infection is indeed greater than zero.”

This study dramatically contradicts a much reported investigation, also conducted in San Francisco, which suggested that as many as 8% of all HIV infections amongst gay men were due to oral sex. The conclusions of this study were subsequently undermined, and a recently published Spanish study involving heterosexual couples where one partner was HIV-positive and the other negative, concluded that not a single case of HIV transmission could be attributed to oral sex in a ten year period.

The authors of the new San Francisco study conclude that “it is important that health professionals…have valid information to impart to their sexually active clients ... acquiring HIV through fellatio is significantly less risky than from anal sex, and therefore one’s choice of sexual practices do matter.”

Romero J et al. Evaluating the risk of HIV transmission through unprotected orogential sex. AIDS 16:9:1269-97, 2002 Over 19,000 instances of unprotected oral sex did not lead to a single case of HIV transmission amongst a cohort of 135 HIV-negative Spanish heterosexuals in a sexual relationship with a person with HIV.

A research team at the sexually transmitted diseases and HIV testing clinic in Madrid monitored a cohort of HIV-negative heterosexuals in a “steady” sexual relationship with a person with HIV. Both members of the couple were monitored at the clinic every six months and the HIV-negative partner was tested for HIV. Both members of the couple were asked to complete a structured questionnaire about their sexual practices and other possible exposures to HIV. The questionnaire explicitly asked people to specify each type of sexual practice since their last clinic visit and an approximate weekly and monthly frequency of each type of sexual activity was calculated by the researchers from their answers. A distinction was made between unprotected and protected sex and people were also asked whether ejaculation occurred.

In total, 135 HIV-negative people (110 women and 25 men) were recruited to the study. Of the women, 96 had performed fellatio on their HIV-positive partner, giving an estimated total of 8,965 instances of unprotected fellatio, with ejaculation occurring in the mouth on an estimated 3,060 occasions (34%). Ninety-eight HIV-positive men carried out unprotected cunnilingus on their HIV-negative partner.

Amongst the 25 HIV-negative men with a positive partner, 12 had unprotected cunnilingus, with an estimated 614 total number of episodes. Twenty-four of the 25 men had passive fellatio, with a total of 1,081 instances of fellatio without a condom performed by the HIV-positive partner.

It is thought that certain factors increase the chances of passing on HIV through oral sex. These include the HIV-positive person having a high viral load. In the Madrid study, viral load results were only available for 60 members of the cohort. Of these 10% (6) had a viral load above 10,000 copies/ml. CD4 counts below 200 cells/mm3 were recorded in 15.6% of the HIV-positive people in the study.

Other risk factors include the HIV-positive person ejaculating into the mouth of their partner; the presence of an sexually transmitted infection; and poor oral health. Amongst the HIV-positive men, 34% ejaculated into the mouths of their partners and vaginal infections were detected in two of the HIV-positive women who had cunnilingus performed on them by a HIV-negative partner.

In total almost 19,000 instances of unprotected oral sex were estimated to have occurred involving the 135 couples over the ten years of the study, but not a single case of HIV transmission was detected. The study authors conclude that: “this seems to point to a very low probability of HIV transmission related to this practice.”

Schechter MT et al: Can HTLV-III be transmitted orally? Lancet 1:379, 1986.

Schechter M et al: HIV seroconversion in a cohort of homosexual men, Canadian Medical Association Journal,  135:12:1355-60, 1986. A cohort study of 345 gay men running since November 1982. `Multivariate analysis failed to reveal any role of oral sexual activity in the transmission of HIV. Oral ingestion of semen was not associated with seroconversion in either univariate or multivariate analysis.'

Schechter M et al: HIV seroconversion and sexual behaviour in a cohort of homosexual men, Canadian Medical Association Journal, 1987. 137:6:474-80. The Vancouver study results challenged, reassessed and reaffirmed.

Van Griensven GJP, Tielman RAP, Goudsmit J et al: Risk factors and prevalence of HIV antibodies in homosexual men in the Netherlands, American Journal of Epidemiology 125:1048-57, 1987. A study of 741 gay men with multiple sex partners in Amsterdam. `In this study [oral sex was] not found to be significantly related to anti-HIV. This result however, does not permit the conclusion that these sexual techniques are without risk. For a number of reasons, it may be possible that HIV is transmitted by these sexual techniques, although with lower efficiency, hence not permitting their role as risk factors to be detectable in the present study. Therefore, more profound research and statistical methods are needed to determine whether, under which conditions, and to what degree these techniques are independent risk factors for seroconversion.

Winkelstein W et al: Sexual practices and risk of infection by the human immunodeficiency virus. San Francisco Men's Health Study. JAMA, 257:3:321-5, 1987.