The discussion here is an attempt to revise that list in the light of a body of epidemiological data that have been accumulating in the past 15 years. Before getting to the list of practices, it is important to state clearly at the outset that, although these recommendations are probably better than those drawn up in the 1980’s, they too have their limitations. The only real way to determine, for example, the probability of transmission of HIV during anal intercourse would be to carry out a controlled experiment with sufficient numbers of HIV-infected and uninfected individuals. That is clearly impossible. The second drawback of these data is that they are heavily reliant on self-reporting of sexual activities. Thus, the information as to whether an individual has or has not had unprotected anal intercourse is determined by asking the individual and trusting his or her answer. Given the fact that sex is a sensitive issue in our culture, the reliability of the answers cannot always be guaranteed with assurance.
Moreover, the data here lump together groups that should be separated into subgroups. For example, it is clear that an HIV positive individual is much more infectious at certain times during his or her infection than at other times. However, there are few data to prove that statement and there is no way to separate those stages based solely on the presence of HIV antibodies in the blood. So although it is possible to make general statements about the average risk of a given sexual practice, and to make qualitative lists of which practices are more risky than others on average, it is vital to remember that whatever the average risk, the result of any exposure to HIV is not the average, it is a certainty. Either the individual becomes infected or not.
With those warnings said, the list presented here represents an update of the 1980’s list based on some specific research studies reported in primary literature (e.g. one by Vittinghof), some compendia of many disparate studies gathered in the secondary or tertiary literature (e.g. Stine’s annually updated book), and even some journalistic summaries from the popular press (e.g. TIME magazine). Since the vast majority of the sexually transmitted cases of HIV infection in the US are the result of sex between two men, wherever appropriate, the data are drawn from surveys of homosexual and bisexual men. It is generally assumed that (for example) receptive anal or oral intercourse carry the same risk for women as for men, but there are no good data to support of refute that assumption.
To give the list some relevance, numbers have been attached. For all the reasons listed above, these must be considered as illustrations rather than firm fact. If they are off by a factor of two or five, that would be no surprise. But they are not likely to be wrong by a factor of 50 or 100. This list below begins with the most unsafe practices and ends with the practices that carry so little risk as to be considered “safe.” But remember, the only truly safe sex is no sex at all.
THE RISKIEST ACTIVITY:
1. Receptive Anal Intercourse (unprotected). This remains the most unsafe practice in terms of HIV transmission. The average likelihood that an HIV infected man will infect his partner by having insertive anal intercourse is probably about 1 in 50. There are several reports that show that the probability of seroconverting increases with number of sexual partners, but the probability seems higher for those who claim to have had only one encounter than would be predicted from those with more. That may be the result of more abrasion and tearing of tissue when the anus and rectum are penetrated for the first time, it may be the fact that the partner, though forced to admit some sexual activity, claims it was only once, or it may be people who are newly active sexually may be more likely to have sex with others who are newly active. In that case their insertive partner is more likely to be recently infected with HIV and may be more infectious. In any event, unprotected receptive anal intercourse is a very risky behavior.
This is confounded by the fact that it seems likely that the risk is highest when the viral titer of the infected partner is highest. The viral titer is high during late stage HIV disease (“full blown AIDS”) but it is also high during the “window period” early after infection, the time when an infected individual has not yet raised an immune response against HIV and thus scores as negative in the standard blood test. Thus unprotected anal intercourse may actually be riskier with some “negative” partners than with many who test positive. In short, unprotected anal intercourse is extremely risky under virtually all conditions.
OTHER HIGH RISK ACTIVITIES:
2. Receptive Vaginal Intercourse (unprotected). It is clear that HIV is transmitted during heterosexual intercourse. The strain of HIV prevalent in the US appears to be less infectious by this route than the strains prevalent in Africa and Asia, or it may be a combination of differing sexual practices or the effects of other sexually transmitted diseases. In any event, it is still transmitted well enough to be a serious risk. The estimates are about a 1 in 300 risk for woman to become infected during a single act of unprotected vaginal intercourse with an HIV infected man. In some surveys of couples where the man was HIV positive and the woman had no risk factors other than her husband, about 20% of the women eventually become HIV positive.
3. Receptive Anal Intercourse (with a condom). The use of a condom reduces the probability of HIV infection. That is clear. What is not clear is just how much. The estimates vary from about a 50% reduction in risk to an 80 or 90% reduction in risk. In other words, the “streetwise” dictum that “sex with a condom is safe sex” is simply incorrect. Proper use of a condom during anal intercourse reduces the risk to the receptive partner by 2-10 fold, with a corresponding saving of many lives. Moreover, this reduction may be enough to block reduce overall transmission rates to a level low enough to cause the epidemic to dies out in time, though this is of little value to those who become infected in the meantime. The risk reduction connected with condom use is much smaller than was expected. This fact has been reproduced in many studies and seems to be quite reliable. The reason for it is less clear.
Latex condoms appear to be an effective barrier that HIV cannot cross, though this may vary with brand, storage, and use. The failure rate of latex condoms during normal use (tears and breaks) is thought to be about 1-2%. Thus one would expect the risk of HIV infection during protected anal intercourse should be about 50-100 fold lower than unprotected anal intercourse. The reasons for the discrepancies between the many surveys and the expected reduction are unknown, but not unlike similar discrepancies when the effectiveness of latex condoms in preventing pregnancy is assessed. Biased reporting as to how consistent and careful the insertive partner was in using the condom may contribute. So too may a very aggressive form of anal intercourse that occurs in certain settings, leading to higher than normal failure rate. Thus the risk to the receptive partner is probably about 1 in 500, though it may actually be somewhat higher, perhaps as high as 1 in 200.
4. Accidental needle-sticks in a health care setting. This is not a sexual activity but is presented here as a benchmark. Doctors and nurses are often accidentally jabbed with needles that are contaminated with blood from an HIV infected patient. The experience over time suggests that the risk for HIV transmission is also about 1 in 300.
5. Insertive Anal Intercourse (unprotected). HIV is transmitted about ten times more efficiently from the insertive to the receptive partner than from the receptive to the insertive. Nevertheless, there is good evidence that HIV can be transmitted to the insertive partner, presumable through the mucous membranes of the penis and urethra. The number here are not good, but a best guess puts the risk at about 1 in 500.
6. Insertive Vaginal Intercourse (unprotected). HIV can be transmitted from women to men during vaginal intercourse, but again, the transmission is less efficient than from men to women. For couples where the man’s only risk factor was his HIV infected partner, the eventual number of men who seroconvert is about 1%. The numbers for risk are even worse here than for insertive anal intercourse but the best guess is that it may be slightly less, if only because the sex may be less aggressive and better lubricated in many cases. Still, a conservative guess would put the risk at about 1 in 1000.
LESS RISKY ACTIVITIES:
7a. Receptive Oral Intercourse (unprotected fellatio). This is the most controversial placement in this list. Since HIV-containing semen is transferred to a cavity lined with mucous membrane, it had been assumed that his was a very high risk behavior. However, several studies found no evidence for efficient HIV infection by this route, though anecdotal instances persist. There was a problem of small numbers, since few people who engaged in frequent oral sex could be shown to abstain from all other forms of sexual activity. In the few documented cased where oral sex appeared to be the only risk factor, further examination revealed that at least some of the respondents later (and more privately) admitted to unprotected anal sex as well. However a recent study of men who have recently seroconverted suggests that as many as 8% of them claim no risk factor other than receptive oral sex. This number seems high in light of other studies and it has not yet been replicated, but so long as it stands, unprotected receptive oral intercourse cannot be put into the “essentially safe” category.
It is important to note that for Receptive Oral Intercourse (unprotected), other risk factors must be considered. Broken skin, bleeding gums, herpes sores, gonorrheal and syphilitic lesions, and even the abrasions caused by toothbrushing and eating potato chips would be expected to affect transmission by this route. Thus, no risk numbers are assigned here. Because of the newest study, this activity is placed at the boundary between very risky and less risky. It must be noted however, that many other diseases are transmitted very efficiently by this route (including hepatitis, gonorrhea, syphilis, herpes, warts, etc.)
7b. Cunnilingus (unprotected). It is assumed that the transmission from vaginal secretions to the oral cavity would be somewhat less than from semen, but there simply are no data. Again the risk of other STD’s must be considered.
LOW RISK ACTIVITIES:
8. Receptive Oral Intercourse (protected). There are no data to support transmission of HIV by this route, but one can extrapolate the 3 to 5-fold reduction in HIV transmission with condom use to expect that oral sex with a barrier (condom or dental dam) between the mouth of the HIV-negative recipient and the penis or vagina of an HIV-positive donor would carry a similar reduction in risk.
9. “Insertive” Oral Intercourse. It is quite unlikely that HIV will be transmitted from the mouth of an HIV-positive individual to the penis or vagina or an uninfected individual unless there are bleeding lesions in the mouth of the HIV-positive person. The risk of transmission by this route, whether with or without protection, appears to be small, though smaller with a latex barrier than without. In fact for most of the activities listed below this point, the major risk factor derives from transmission of infectious fluids (blood, semen, and vaginal secretions) to and from open lesions on the skin.
10. Anilingus (“Rimming”). The practice of an individual using the tongue to stimulate the anus of a partner is generally regarded as very low risk for HIV transmission, whichever partner is HIV-positive. However the practice is very high risk for the transmission of numerous other parasitic, viral, and bacterial diseases (including many incurable diseases like hepatitis). Thus “rimming” should be avoided if possible; if practiced, it should always be done with a latex barrier (dental dam) between the anus and the tongue.
11. Brachioproctal or Brachiovaginal Manipulation (Anal or Vaginal “Fisting”). Early in the AIDS epidemic, there was a strong correlation between the men who developed AIDS and the men who had engaged in the practice of Fisting. It seemed prudent to consider Fisting a high risk behavior and a series of guidelines was developed to try to limit this risk, including use of latex gloves. It now appears that there is little risk of HIV transmission by this route so long and the hand of the insertive partner has not breaks in the skin. There are certainly infected macrophages present in the lower bowel and on its linings, but these are not likely to penetrate unbroken skin. Similarly, when the insertive partner is infected, there is little likelihood that there will be transmission of any infectious fluids to the recipient unless the hand of the insertive partner has bleeding or oozing lesions.
The use of latex gloves is still recommended when such lesions are present. It is also prudent to wash the hands of the insertive partner whenever a new receptive partner presents. (This practice was and still is practiced in a social setting and may involved multiple partners in a single evening.) Lubricants, gloves, and hands should not be used on successive partners if cross contamination is likely.
SAFE ACTIVITIES:
12. Masturbation. Mutual masturbation onto unbroken skin is thought to be quite safe. There are no documented cases of transmission by this route. The key word is “unbroken” skin, otherwise the risk is not from masturbation but from contact with the bloodstream.
13. Kissing. Dry kissing (as you would kiss your mother) is certainly safe. There are no documented transmissions of HIV by this route. Deep or “wet” kissing (“French kissing”) is also generally thought to be safe. The few cases of documented transmission by aggressive wet kissing have also involved bleeding gums, ulceration of the mouth and tongue, and other bloody sores in the mouth. Thus wet kissing was not the primary route, but rather blood-blood contact. HIV has been detected in saliva, but the numbers are low and the HIV appears to be rendered non-infectious by a mechanism that is unknown. Thus saliva does not appear to be an “infectious bodily fluid.”
Urine is another bodily fluid that appears to be non-infectious even though HIV has been detected in urine. HIV transmission does not occur when an HIV-infected individual urinates on the unbroken skin of an HIV-negative person. Even when the urine is swallowed or inserted into the rectum, the risk appears to be low.
14. Intimate, non-sexual contact. Intimate contact such as touching, shaking hands, sharing living space (including bedding, toiletries, toothbrushes, etc.), or dishes does not pose a risk of HIV transmission.
15 Sex involving special equipment. The use of vibrators, dildoes, or other “sex toys” is safe whenever the toys are not shared. If they are shared, they need to sterilized between uses. The only truly safe mode of sterilization is exposure to hot steam under pressure at 250 degrees for fifteen minutes or longer. HIV is quite fragile and does not survive drying, detergents, or strong bleaching agents very well. Thus a toy that has been washed in detergent, soaked in bleach, and dried thoroughly is probably safe to reuse, at least in terms of HIV transmission. However some other viruses and bacteria are much more resistant to these treatments than is HIV, so there is still a risk with shared toys.
16. Casual contact. Casual contact with HIV-infected individuals is not a route of transmission. HIV cannot be transmitted on toilet seats, public telephones, or doorknobs, in swimming pools or bowling shoes, by coughing, sneezing, or breathing, from the communion cup or the drinking fountain.
This discussion is summarized in the following list, that groups activites into risk categories. It is important to remember that the risk here is only of HIV transmission. (For example, anilingus is extremely dangerous for the transmission of parasites, and viruses like hepatitis) Moreover, all of these activities assume that other transmission routes (blood to blood contact) is not an issue.
EXTREMELY HIGH RISK
- Unprotected anal intercourse (receptive partner)
HIGH RISK
- Unprotected vaginal intercourse (receptive partner)
- Protected anal intercourse (receptive partner)
- Unprotected anal intercourse (insertive partner)
- Unprotected vaginal intercourse (insertive partner)
LESS RISKY
- Unprotected oral intercourse (felatio) (receptive partner)
- Unprotected oral intercourse (cunilingus) (receptive partner)
LOW RISK
- Receptive oral intercourse with a barrier (condom or dental dam)
- Insertive oral intercourse
- Anilingus (“Rimming”)
- Brachioproctal/Brachiovaginal manipulation (“Fisting”)
SAFE ACTIVITIES
- Masturbation
- Kissing (dry or wet)
- Intimate but non-sexual contact
- Use of sex toys (single person use)
- Prolonged casual contact