17. Sexually Transmitted Infections

Effectiveness of the Principal Interventions

Unlike HIV interventions, STI interventions benefit from a large body of rigorous evaluations. STI interventions that have been rigorously evaluated for effectiveness can be organized by intervention level (that is, individual, group, or community); by the outcomes measured; and by the intervention modality used (for example, behavior change, vaccination, topical microbicide use, screening, or treatment). Prevention outcomes may measure the prevention of acquisition, of transmission, and of complications of STIs (Manhart and Holmes 2005). This section reviews the interventions for which the strongest evidence exists.

Individual-Level Interventions

A large number of STI interventions that have been rigorously evaluated are individual-level interventions.

Preventing Acquisition

The following have been the main means of preventing STI acquisition:

• Behavior change. Counseling on risk reduction was the most frequently used behavior-change approach. Most studies showed a reduction in risk behaviors as a result of counseling, and some showed decreases in STI outcomes (Kamb and others 1998).

• Antimicrobial prophylaxis. Two studies (Harrison and others 1979; Kaul and others 2004) showed reductions in the incidence of gonococcal, chlamydial, or trichomonal infections following antimicrobial prophylaxis.

• Vaccines and passive immunization. A yeast-derived HPV type 16 vaccine was 100 percent efficacious in preventing persistent HPV-16 infection in young college women (Koutsky and others 2002), and a bivalent HPV type 16 and type 18 vaccine was also highly efficacious in preventing those infections. An HSV-2 glycoprotein D-adjuvant vaccine among those with no serological evidence of prior HSV-1 infection partially protected women, but not men, from experiencing genital herpes disease, with 73 percent efficacy for such women in one trial and 74 percent in another (Stanberry and others 2002).

• Microbicides. To date, studies have not identified any efficacious topical microbicides.

• Male circumcision. Even though cross-sectional evidence suggesting that male circumcision decreases the risk of acquiring chancroid and HIV is strong, outcome data are not yet available from ongoing randomized trials in Kenya, South Africa, and Uganda.

Preventing Transmission

All individual-level interventions aimed at preventing transmission have involved curative or suppressive therapy. Giving tinidazole to male partners of females treated for vaginal trichomoniasis infections significantly reduced recurrences in the females; administering valacyclovir to positive members of HSV-2 serodiscordant couples reduced the incidence of symptomatic genital herpes and HSV-2 seroconversion in the uninfected partners; patient-delivered therapy to partners of women with chlamydial infection demonstrated a nonsignificant trend toward reduced risk of reinfection with C. trachomatis; and expedited partner therapy (usually patient delivered) significantly reduced persistent or recurrent gonococcal or chlamydial infection in the index patient (Golden and others 2005).

Preventing Complications

Risk-based screening for C. trachomatis infection resulted in a 56 percent reduction in the subsequent risk of incident pelvic inflammatory disease (Scholes and others 1996). Several trials have shown that antiviral suppression decreases clinical and virological recurrences of genital herpes (Corey and Handsfield 2000).

Group-Level Interventions

Studies of behavior-change methods in small-group settings to reduce the acquisition of STIs had mixed outcomes. Behavior-change approaches resulted in significant reduction in incident STIs; antimicrobial prophylaxis and provision of female condoms did not.

Community-Level Interventions

Four community-level randomized trials have sought to reduce the prevalence and transmission of STIs by shortening the duration of infectiousness within the general population (Manhart and Holmes 2005).

The "Mema Kwa Vigara" study in Mwanza, Tanzania, randomized 20 communities to intervention and control communities. The intervention consisted of school-based sexual and reproductive health education, enhanced reproductive health services for youths, condom distribution, and community activities. Knowledge and reported behaviors improved; however, no differences were apparent between the intervention and control communities in relation to HIV or HSV-2 seroincidence, incidence of other STIs, or pregnancy outcomes (Hayes and others 2003).

A second study in Mwanza, Tanzania, randomized communities to intervention and control conditions. The intervention consisted of syndromic treatment of STIs. The results showed a 40 percent reduction in HIV incidence and reductions in symptomatic urethritis in men and prevalence of syphilis seroreactivity; the prevalence of gonorrheal or chlamydial infection in prenatal women did not change (Grosskurth and others 1995; Mayaud and others 1997).

In a community randomized trial in Masaka, Uganda, one community received information, education, and communication; a second community received information, education, and communication plus syndromic management of STIs; and the control received community development assistance. The results showed no differences in HIV-1 incidence. The incidence of HSV-2 seroconversion declined in the community receiving information, education, and communication only; the incidence of syphilis and of gonorrhea decreased in the community receiving information, education, and communication plus STI syndromic management; and condom use increased in all three communities (Kamali and others 2003).

In Rakai, Uganda, a community randomized trial evaluated the efficacy of repeated mass treatment of STIs. Relative to control communities, in intervention communities the prevalence of T. vaginalis in women was reduced significantly, but no significant reduction was apparent in prevalence of gonorrhea, chlamydial infection, new syphilis seroreactivity, and bacterial vaginosis; in HIV incidence; or in history of urethral or vaginal discharge or genital ulcer disease (Wawer and others 1999). A subanalysis of pregnant participants showed a reduction in the prevalence of several STIs in women tested near delivery and in potentially STI-related pregnancy, puerperal, and neonatal morbidity (Gray and others 2001).

Conclusions on Interventions

The review of STI intervention research suggests several, perhaps counterintuitive, insights:

• First, most evidence is on individual-level interventions aimed at reducing STI acquisition, even though individual-level interventions may be costly and difficult to sustain.

• Second, behavior change is the most commonly evaluated modality, followed by treatment.

• Third, theory-based behavioral interventions failed to show an effect as often as behavioral interventions not based on theory.

• Fourth, behavioral interventions delivered in small group settings were as effective as those delivered to individuals (Manhart and Holmes 2005).

• Fifth, the effect of a particular behavior change on STI risk depended on the type of STI; however, the number of partners may be more predictive of risk for highly infectious STIs than for HIV, and unprotected sex acts may be more predictive of risk for HIV than for highly infectious STIs (Semaan and others 2002). Thus, behavioral interventions may have different effects on STIs of differing infectiousness.

• Finally, the number of intervention trials that demonstrate declines in risk behaviors combined with either no effect on STIs or increases in STIs is increasing. This observation calls into question the use of behavioral outcome measures as indicators of biomedical outcomes (Aral and Peterman 2002).