(Plummer et al., 1991; Nyange et al., 1994). Although the magnitude of the increased risk of transmission attributable to genital ulcers in these studies was expressed as risk ratios ranging from 3.3 to 4.7, Hayes and others (1995) have estimated that genital ulcer disease may increase the per exposure risk of transmission by a factor of 10 to 50 for male-to-female transmission and by a factor of 50 to 300 for female-to-male transmission. Boily and Anderson (1996) also have recently demonstrated that published cohort studies are very likely to underestimate the true magnitude of the increased risk of sexual transmission of HIV conferred by other STDs. Some of these studies have also found that gonorrhea of the cervix (Plummer et al., 1991; Laga et al., 1993; Nyange et al., 1994), chlamydial infection of the cervix (Laga et al., 1993) and vaginal trichomoniasis (Laga et al., 1993), all increase the risk of subsequent HIV infection. Although Laga and others initially reported no increase in HIV infection following genital ulcer disease among female sex workers (Laga et al., 1993), a later study involving the same study population did find such an association (Laga et al., 1994).
In the United States, the multi-center AIDS cohort study found no association between herpes simplex virus type 2 antibody and subsequent HIV infection among men who have sex with men (Kingsley et al., 1990). However, an earlier retrospective cohort study in gay men found that herpes simplex virus type 2 infection was a risk factor for HIV infection (Holmberg et al., 1988). In addition, a study of heterosexual men attending an STD clinic in New York showed that men presenting with chancroid were more likely to become infected with HIV than were men presenting with other STDs (Telzak et al., 1993).
The above prospective studies establish the temporal relationship of exposure to STD prior to HIV infection. The observed associations between STDs and HIV infection are compatible with three possible explanations, other than the confounding factors addressed by the study designs: