Abstinence Ineffective in HIV, Pregnancy Prevention

OXFORD, England, Aug. 3 -- Abstinence-only programs to prevent HIV and pregnancy in high-income countries are ineffective, according to a systematic review of 13 U.S. trials.

Participants in the trials did not report significant differences either in risky sexual behaviors or biological outcomes compared with controls, according to Kristin Underhill, M.Sc., and colleagues, at the University of Oxford's Centre for Evidence-Based Intervention.

The finding parallels an earlier study, which found that such programs in developing countries had little effect, Underhill and colleagues reported in the Aug. 4 issue of BMJ.

The investigators undertook an international literature search, looking for randomized or quasi-randomized trials of abstinence-only prevention programs aimed at HIV, pregnancy, or both.

Despite its international focus, the search found only 13 randomized trials that met standards for inclusion -- all conducted in the U.S. and all focusing on youths. No quasi-randomized trials met inclusion criteria.

The 13 studies, enrolling almost 16,000 youths -- had such a wide variety of participants, control groups, and outcome measures that a formal meta-analysis wasn't possible, the investigators said.

Interestingly, none of the trials -- even those whose focus was HIV alone -- used incidence of HIV as an outcome measure. Instead, the studies used unprotected vaginal sex as a proxy for HIV risk.

Sexual behavior was not reported in all of the studies and was limited to vaginal sex. None of the studies evaluated oral sex, anal sex, or same-sex behavior, the researchers said.

Control groups included usual care (which was rarely defined), safer sex counseling, no treatment, time-matched abstinence-plus programs, and abstinence programs without enhancements such as parent-child homework.

Outcomes -- all based on self-reported questionnaires -- included diagnosis of sexually transmitted infection, pregnancy, first sexual experience, having sex recently, having unprotected sex recently, and having sex without a condom.

Despite those limitations and others, "the evidence from this systematic review is clear," the investigators said:

"When compared with a variety of control groups, the participants in these 13 abstinence-only program trials did not report differences in risky sexual behaviors or biological outcomes."

Specifically:

• No trial reported a significant benefit in preventing sexually transmitted infections and one reported short- and long-term harm.
• None of the trials reported any benefit in preventing pregnancy and one reported that the program increased the risk, compared with usual care or no treatment.
• Only five trials had enough data to allow an estimate of rates of recent unprotected vaginal sex and none showed a benefit or harm.
• Seven trials reported the incidence of any vaginal sex-six found no effect and one found the program cut the risk by 47% at two months follow-up compared with usual care.
• Four trials assessed frequency of sex and none found any effect.
• Nine trials looked at condom use and none found an effect.
• Ten trials looked at sexual initiation and again none found a significant effect.

The fact that all 13 trials took place in the U.S. could be explained two ways, the researchers said: Either studies in other countries are "inaccessible by search methods" or abstinence-only programs are not popular as HIV prevention strategies in other high-income nations.

The study raises the question of whether such programs are an effective use of public money, said Nancy Kiviat, M.D., of the University of Washington, and colleagues.

In an accompanying editorial, they noted that abstinence has become a "politically charged issue" because a large share of the money allocated for prevention in the President's Emergency Plan for AIDS Relief (PEPFAR) must, by law, be spent on promoting abstinence.

While the review makes clear that such programs don't work in the U.S., Dr. Kiviat and colleagues said, there is abundant evidence that promoting condom use does prevent HIV transmission.

"Therefore," they said, "priority should be given to culturally sensitive, sex-specific, behavioral interventions that target black and Hispanic patients in clinics for sexually transmitted infections, men who have sex with men, and adolescents being treated for drug misuse who are at highest risk of acquiring HIV."