Heterosexual Transmission of HIV
To the Editor
De Vincenzi (Aug. 11 issue)1 concludes that consistent use of condoms is highly effective in preventing heterosexual transmission of the human immunodeficiency virus (HIV). However, the data analysis is flawed. First, the use of person-years as the unit of comparison for seroconversion rates does not take into account the difference between consistent and inconsistent users of condoms in the frequency of sexual contact. The author reports that the median frequency of sexual contact among the latter was twice that among the former. Once this is accounted for,2 the seroconversion rate for inconsistent users is 2.5 per 100 person-years (95 percent confidence interval, 1.2 to 5.8), which is not significantly different from the rate of 0 for consistent users (95 percent confidence interval, 0 to 1.5 per 100 person-years).
Second, the author does not adequately define the temporal pattern of seroconversion among inconsistent users of condoms. Those who seroconverted within the first three months of the study were most likely infected before enrollment. Thus, the seroconversion rate for the group of inconsistent condom users may be even lower.
This report may offer a false sense of comfort to HIV-discordant couples, but even more important is the issue of whether using condoms with a high-risk partner is as safe as or safer than not using condoms with a low-risk partner. The answer is a resounding no. Choosing a partner who is not in any high-risk group provides a degree of protection that is at least three orders of magnitude higher than that afforded by condom use.3 Sound judgment is a far better defense than the use of any mechanical device.
Jayakrishna Ambati, M.D.
North Shore University Hospital, Manhasset, NY 11030
Balamurali K. Ambati, B.A.
Mount Sinai School of Medicine, New York, NY 10029
Ambati Muralimohan Rao, D.H.Ed.
New York City Department of Corrections, Elmhurst, NY 11370
1. de Vincenzi I. A longitudinal study of human immunodeficiency virus transmission by heterosexual partners. N Engl J Med 1994;331:341-346
2. Kaplan EH. Modeling HIV infectivity: must sex acts be counted? J Acquir Immune Defic Syndr 1990;3:55-61
3. Hearst N, Hulley SB. Preventing the heterosexual spread of AIDS: are we giving our patients the best advice? JAMA 1988;259:2428-2432
To the Editor
The report by de Vincenzi adds important information about heterosexual transmission of HIV, but the medical community should not generalize the low rates of seroconversion in the study to apply to all couples discordant for HIV infection. The use of counseling may have decreased the rates of HIV transmission, and enrollment on the basis of discordant HIV status selects for couples with biologic or behavioral variables that impart a low risk of HIV transmission. Couples at higher risk for HIV transmission would probably already have seroconverted and would not have been eligible.
The author wisely concludes that the risk is not constant from one sexual encounter to another. However, it is important to realize that the enrollment strategy used in this study strongly selected against the enrollment of high-risk sexual contacts. The seroconversion rates observed in this study are underestimates of those in the general population of sexually active couples discordant for HIV infection.
Ellen Morrison, M.D.
Columbia-Presbyterian Medical Center, New York, NY 10032
To the Editor
The problem of subjects' lying (often euphemistically termed “social desirability responding”) about engaging in anal intercourse and intravenous drug use plagues most studies of behavioral risk factors for the transmission of HIV, and the study by de Vincenzi and colleagues is no exception. How was the absence of homosexual contact verified? How was the absence of anal intercourse among the women verified? If only 4 men and 6 women among the 121 couples inconsistently using condoms lied when they denied engaging in anal intercourse (or misreported the facts for other reasons), there would be no cases attributable to vaginal intercourse without a condom. At least this much lying should be expected.1,2
Before vaginal and anal intercourse are assigned comparable degrees of risk and condoms given the credit for saving lives, the alternative explanation that the disease is spread almost exclusively by anal and intravenous transmission must be more rigorously examined. Other investigators found that HIV infection in women was related to anal intercourse (especially among partners of bisexual men) and the number of exposures to the index patient, but not to condom use or the total number of sexual partners.3 The inconsistent condom users in the study by de Vincenzi and colleagues had more than twice the prevalence of anal intercourse at follow-up and also had more sexual contacts. Regular condom use has been associated with reduced sexual activity in surveys of sexual behavior of persons before4 and after5 the identification of AIDS.
Stuart Brody, Ph.D.
University of Tubingen, 72074 Tubingen, Germany
1. Ross MW. Psychosocial factors in admitting to homosexuality in sexually transmitted disease clinics. Sex Transm Dis 1985;12:83-87
2. Brody S. Lack of evidence for transmission of human immunodeficiency virus (HIV) through vaginal intercourse. Arch Sex Behav (in press).
3. Padian N, Marquis L, Francis DP, et al. Male-to-female transmission of human immunodeficiency virus. JAMA 1987;258:788-790
4. Giese H, Schmidt G. Studenten Sexualitaet. Hamburg, Germany: Rowohlt, 1968.
5. Glor JE, Severy LJ. Frequency of intercourse and contraceptive choice. J Biosoc Sci 1990;22:231-237
To the Editor
In a prospective study of 256 couples discordant for HIV infection who continued to be sexually active during 20 months of follow-up, de Vincenzi1 reports that fewer than 50 percent consistently used condoms. The overall seroconversion rate was 2.3 per 100 person-years. Johnson rightly despairs in an editorial (Aug. 11 issue)2 that in spite of repeated counseling, so few of the couples engaged in safe sex. By way of contrast, we offer updated results of our longitudinal study with a similar design.3
We followed 158 couples discordant for HIV infection for a total of 243 couple-years. These couples were sexually active at entry into the study and mutually monogamous. During follow-up, we observed no new seroconversions. At base line the rate of consistent condom use was 30 percent, and the rate of anal intercourse was 39 percent. At the final follow-up visit the respective rates were 100 percent and 8 percent (both P<0.001 by McNemar's test for matched pairs). The rate of abstinence at follow-up was 12 percent.
Although not all couples reported using condoms at each follow-up visit, close to 90 percent did. We attribute our success in part to the fact that we counseled subjects in couples as well as individually. In addition, study staff members were available (and were called on) as needed, not only during the scheduled visits every six months. Although these procedures are costly, the savings in terms of prevention of infection are well worth it.
Nancy S. Padian, Ph.D.
Eric Vittinghoff, Ph.D.
Stephen Shiboski, Ph.D.
University of California, San Francisco, San Francisco, CA 94110
1. de Vincenzi I. A longitudinal study of human immunodeficiency virus transmission by heterosexual partners. N Engl J Med 1994;331:341-346
2. Johnson AM. Condoms and HIV transmission. N Engl J Med 1994;331:391-392
3. Padian NS, O'Brien TR, Chang Y, Glass S, Francis DP. Prevention of heterosexual transmission of human immunodeficiency virus through couple counseling. J Acquir Immune Defic Syndr 1993;6:1043-1048
Response
Dr. de Vincenzi replies:
To the Editor: We agree with Dr. Morrison that since my colleagues, I enrolled only couples discordant for HIV infection, we may have selected for characteristics associated with a low risk of transmission. As we discussed, no mean rate of transmission can be generalized from any study, since the risk is highly variable from one contact to another and the distribution of risk factors varies greatly from one study population to another. The concept of “the general population of sexually active couples discordant for HIV infection” is also hardly operational, since there is no such population and the characteristics of the couples depend on multiple factors, such as whether they live in the United States, Europe, Africa, or Asia. We therefore estimated seroconversion rates according to the characteristics of the couples in our study.
The variability between study samples is illustrated by the study reported by Padian et al., in which 100 percent of couples used condoms and no partner seroconverted. Padian et al. state that the successful adoption of condom use is due to counseling. However, many factors, including social, cultural, and economic ones, influence condom use. In the absence of a control group that did not receive counseling, such a conclusion may be an overinterpretation. Furthermore, a specific counseling technique may be successful in California among bisexual men and their female partners, but not among injection-drug users in France or African immigrants living in Belgium.
Dr. Ambati and colleagues suggest a high risk of heterosexual transmission, even through protected contacts, whereas Dr. Brody denies the existence of a risk of transmission through unprotected vaginal intercourse. Neither position is based on scientific evidence.
Dr. Ambati et al. refer to a report by Kaplan,1 in which the author answered “no” to the question “Must sex acts be counted?” Moreover, that study mixed the issues of the efficacy of the mechanical barrier and its acceptability. Even if its efficacy is no longer in doubt, its acceptability varies between populations. We agree that alternative strategies (such as no penetrative sex or monogamous relationships between HIV-negative partners) should be used when condom use is not acceptable. However, it is hazardous to promote the idea that unprotected sex with low-risk people is safer than protected sex with high-risk people. How do Dr. Ambati et al. define high-risk people? By epidemiologic criteria, perhaps. But how does an epidemiologic classification help to determine individual attitudes in the choice of a sexual partner?
We agree with Dr. Brody that our prospective analysis lacks statistical power to show an increased risk associated with anal intercourse. Indeed, we found such an association in the cross-sectional analysis.2 However, from a public health point of view, no one should state that there is no risk of HIV transmission through vaginal sex, since the vast majority of cases of AIDS throughout the world are acquired in this manner.
I. de Vincenzi, M.D., M.P.H.
Hopital National de Saint-Maurice, 94410 Saint-Maurice, France
for the European Study Group on Heterosexual Transmission of HIV
1. Kaplan BH. Modeling HIV infectivity: must sex acts be counted? J Acquir Immune Defic Syndr 1990;3:55-61
2. European Study Group on Heterosexual Transmission of HIV. Comparison of female to male and male to female transmission of HIV in 563 stable couples. BMJ 1992;304:809-813
