Estimating the risk per exposure

Published: 07 April 2009
  • The difficulty of gathering evidence means that scientists do not agree on the proportion of sexually transmitted HIV infections that come from oral sex. The UK’s Health Protection Agency estimates that between 1 and 3% of HIV cases result from oral transmission of the virus.
  • The most plausible minimum and maximum figures are 0.1% (one infection per 1000 HIV infections due to oral sex) and 5% (one infection per 20), although one study found a much higher estimate.

A satisfactory answer to the question, 'How high is the risk of HIV transmission through oral sex?' has been notoriously elusive. The following factors, which have often been cited as challenges, were recently summarised in a systematic review of studies:1

  • Very few people report oral sex as their sole risk.
  • Self-reported data on sexual behaviour are hard to collect accurately, and unprotected anal or vaginal sex may be underreported. (Studies have shown that inconsistent reporting of anal sex can be fairly widespread.)
  • If a person practises any other form of unprotected intercourse (anal or vaginal) in addition to unprotected oral sex, any resulting HIV infection is usually attributed to the higher risk behaviour.
  • Studies have frequently grouped all oral sex practices together, often not distinguishing receptive from insertive roles, whether ejaculation occurred in the mouth, etc.
  • Studies of serodiscordant couples do not often include HIV-positive partners during primary infection, when they are much more infectious.

Many reports of oral transmission are in the form of isolated and anecdotal reports, rather than in the more controlled context of cohort studies (in other words, people questioned about their sexual practices on a regular basis and tested for HIV antibodies) or other studies with more rigorous patient follow-up.

Most reliable cohort studies have tended to show very low levels of risk, in many cases approaching zero. However, a few carefully conducted studies have given considerably higher estimates which are difficult to reconcile with the others.

What do the numbers compare?

The question of oral risk is further complicated by the way risk is measured and compared. Risk can be evaluated in at least three ways:

  • The overall number of HIV infections attributable to oral sex (i.e., of a total number of people infected, what proportion were infected through oral sex?).
  • Per-partner risk: what is the overall, cumulative risk of long-term oral sex with one partner?
  • Per-contact risk: what is the risk of HIV infection through one act of oral sex with an HIV-positive partner?

The per-contact risk is probably the most relevant measure for sexually active people who are seeking counselling. However, most of the studies discussed below are compared on the basis of the first category – the percentage of overall HIV infections attributable to oral sex.

The lower limit: zero to 0.7%

Five of the studies included in the systematic review provided risk estimates based on cumulative, multiple oral sex acts during the span of a serodiscordant relationship. Three of these studies gave that estimate as zero – no transmissions were reported.2 3 4 The fourth study provided a figure of 1% for receptive fellatio.5 (The fifth is discussed in the next section, The upper limit.)

Three further studies followed HIV-negative gay and bisexual men who reported unprotected oral sex as their sole risk factor.6 7 8 These studies gave overall risk of infection estimates of between zero and 0.5% over the course of the research study.

A long-standing study (1990 to 2002) of heterosexual serodiscordant couples in Madrid4 found that out of 135 couples who reported oral sex, with 19,316 individual acts of exposure over 210 person-years of study, not one HIV seroconversion happened. In the heterosexual couples studied, this would set a lower limit of zero on the proportion of HIV infections due to oral sex.

Even if zero transmissions occur within a study of a finite group of people, there is a statistical chance that some infections might have been seen in a larger group. (In somewhat the same way that a small patch of the pavement could remain dry in a very light rain.) In the study above, this statistical upper limit would be 0.74%.

A study in San Francisco7 identified 239 gay men seeking HIV testing between 1999 and 2001 who reported exclusively having oral sex. None tested HIV-positive. (The number of participants subsequently grew to 363, still with no seroconversions.9) The authors calculated that, out of all HIV infections observed in the real world, there was a maximum possible per-contact risk of 0.04% - that is, HIV might be passed on in one in 2500 acts of oral sex between serodiscordant people.

The upper limit: 21% in Swedish study

One study of gay men in Stockholm between 1990 and 1992 stands in dramatic contrast to these lower estimates. This study found that, out of 28 men who seroconverted, six (21%) were believed to have been infected through oral sex.10 On the one hand, these findings were considered reliable due to the careful analysis of sexual partners, serostatus and specific sexual behaviours involved. On the other hand, this is a single study of a small number of gay men, conducted in the early 1990s.

Intermediate findings from other studies

A study presented at the Seventh Conference on Retroviruses and Opportunistic Infections in San Francisco in January 2000 postulated that 6.6% of new infections were due to oral sex. The study examined 122 infections identified in San Francisco between 1996 and 1999. 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.11

While some of the men did subsequently report also having unprotected anal intercourse, at least one was confirmed as being infected by his partner, with whom he had only had oral sex. This appears to indicate that something between 1 and 8% of sexual HIV infections in the study might come from oral sex. Half of the seroconverters also had significant gum disease or ulcers, and six of the eight men thought to have been infected orally had taken ejaculate in their mouth.

Trying to find consensus

In March 2003, HIV InSite convened a panel of experts, including some of the researchers involved in the studies described here, to try to reach some consensus on the risk of HIV transmission via oral sex.9

The panel ultimately could not come to a consensus on the proportion of infections due to oral sex. Different ranges were suggested, from a minimum of one in 1000 to a maximum of one in 20 (5%). Ultimately it appears to come down to whether researchers believe people are telling the truth when they say that oral sex has been their only risk factor.

Several systematic reviews have also attempted to synthesise study findings. The authors of the review cited above1 – which found only ten studies considered comprehensive enough to include in the analysis – concluded that "it would be inappropriate to make summary estimates for the transmission risk through oral sex" from the study data available.

Putting the numbers in context

The overall message from the above numbers is that, compared to the high risk of unprotected anal or vaginal sex, unprotected sex is considered a 'low-risk' activity – less than high risk, but more than negligible. It may be helpful to put these 'low' figures into some context.

Taking the estimate from the San Francisco study mentioned above7 as an example: the upper estimate of per-contact risk calculated in that study was 0.04% – that is, HIV might be passed on in one in 2500 acts of oral sex between serodiscordant people. This might not sound like much – but you only need 100 different HIV-positive partners, or 100 contacts with one HIV-positive partner, for that risk to become one in 25.

This 0.04% level of risk (one in 2500 exposures) is approximately 20 times lower than the estimated risk for receptive anal sex, but is only half the risk estimated for receptive vaginal sex with a partner during chronic infection (0.08%). It is equal to the estimated risk for insertive anal sex between men, with use of a condom, with a partner whose HIV status is unknown.


  1. Baggaley RF et al. Systematic review of orogenital HIV-1 transmission probabilities. International Journal of Epidemiology 37:1255-65, 2008
  2. De Vincenzi I et al. A longitudinal study of human immunodeficiency virus transmission by heterosexual partners. NEJM 331(6): 341-346, 1994
  3. Raiteri R et al. Lesbian sex and the risk of HIV transmission AIDS 12:450-51, 1998
  4. Del Romero J et al. Evaluating the risk of HIV transmission through unprotected orogenital sex. AIDS 16(9): 1296-1297, 2002
  5. Samuel MC et al. Infectivity of HIV by anal and oral intercourse among homosexual men. Estimates from a prospective study in San Francisco. In: Kaplan EH, Brandeau ML (eds). Modeling the AIDS Epidemic: Planning, Policy and Prevention. New York: Raven Press, 1994
  6. Detels R et al. Seroconversion, sexual activity, and condom use among 2915 HIV seronegative men followed for up to 2 years. J Acquir Immune Defic Syndr 2:77–83, 1989
  7. Page-Shafer K et al. Risk of HIV 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
  8. Lavoie E et al. Determinants of HIV seroconversion among men who have sex with men living in a low HIV incidence population in the era of highly active antiretroviral therapies. Sex Transm Dis 35:25–9, 2008
  9. HIV InSite Risk of HIV Infection Through Receptive Oral Sex. Round table discussion. See, 14 March, 2003
  10. Grutzmeier S HIV transmission in gay men in Stockholm, 1990-1992. Ninth International Conference on AIDS, Po C02–2584, Berlin, 1993
  11. Dillon B et al. Primary HIV Infections associated with oral transmission. Seventh Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 473, 2000
This content was checked for accuracy at the time it was written. It may have been superseded by more recent developments. NAM recommends checking whether this is the most current information when making decisions that may affect your health.
Community Consensus Statement on Access to HIV Treatment and its Use for Prevention

Together, we can make it happen

We can end HIV soon if people have equal access to HIV drugs as treatment and as PrEP, and have free choice over whether to take them.

Launched today, the Community Consensus Statement is a basic set of principles aimed at making sure that happens.

The Community Consensus Statement is a joint initiative of AVAC, EATG, MSMGF, GNP+, HIV i-Base, the International HIV/AIDS Alliance, ITPC and NAM/aidsmap

This content was checked for accuracy at the time it was written. It may have been superseded by more recent developments. NAM recommends checking whether this is the most current information when making decisions that may affect your health.

NAM’s information is intended to support, rather than replace, consultation with a healthcare professional. Talk to your doctor or another member of your healthcare team for advice tailored to your situation.