IN THE DECADE SINCE 1985, many gay and bisexual men (men who have sex with men [MSM]) changed their sexual behaviors in response to the epidemic of human immunodeficiency virus (HIV) infection and acquired immunodeficiency syndrome. Several reports have documented decreased numbers of sex partners, reduced frequencies of unprotected intercourse, or substantially lower rates of sexually transmitted diseases (STDs) in MSM.1–3 Nevertheless, the incidence of STD in this population may remain higher than in exclusively heterosexual men and women. For example, in 1995 in King County, Washington, which includes the city of Seattle, 81 cases of gonorrhea were reported in men who attended public clinics and acknowledged same-sex partners, or in men with rectal gonococcal infections. Based on an estimate that 38,000 self-identified MSM reside in King County,4,5 the rate of reported gonorrhea in MSM approximated 213 cases (95% confidence interval [CI95] 167–260) per 100,000, compared with 80 reported cases (CI95 75–86) per 100,000 in the population as a whole. This calculation underestimates the true rate of gonorrhea in MSM because it excludes urethral and pharyngeal infections diagnosed outside public clinics.
Men who have sex with men who attend urban STD clinics probably represent a subgroup that remains at high risk for acquiring STDs, including HIV infection. Improved knowledge of the demographic and behavioral determinants of STDs in such men may help to design and target prevention efforts. We therefore undertook a cross-sectional study of MSM attending an urban STD clinic to calculate the prevalence of selected STDs and evaluate risk factors for their acquisition.
Our STD clinic uses standardized data collection instruments for clinical encounters, and all data (approximately 380 variables per visit) have been computerized since 1993. Historical data, including behavioral information, are collected by interview. The current study was based on a review of the standardized medical records of men who attended the Seattle-King County Department of Public Health STD Clinic at Harborview Medical Center, Seattle, Washington, during calendar years 1993 and 1994. During this time, there were 14,527 visits by 11,407 men for STD screening or for evaluation of new clinical problems. The gender of the patients' sex partners was recorded for 10,996 (96.4%) of these men. Visits solely for HIV counseling and testing were excluded, as were all follow-up visits, because genital examinations were not consistently performed and, in the case of follow-up visits, behavioral information is not systematically collected. We further limited the analysis to the first clinic visit for each man during the study period. The final study population consisted of 1,253 unduplicated male patients who met these criteria and acknowledged having sex with other men.
Our clinic's routine procedures in evaluating MSM include collection of diagnostic specimens for Neisseria gonorrhoeae from the urethra, rectum and pharynx, and for Chlamydia trachomatis from the urethra and rectum. However, sometimes specimens are not collected from anatomic sites that the patient denies were sexually exposed in the preceding 2 months, and some patients refuse testing of some sites. N. gonorrhoeae was isolated by culture and identified by carbohydrate utilization reactions, and C. trachomatis was isolated by cell culture and identified by fluorescent monoclonal antibody, as previously described.6,7 Standard methods were used for serologic testing for HIV antibody, serologic tests for syphilis, and other routine diagnostic tests.
For the current analysis, gonorrhea and chlamydial infection were defined solely by the culture results, regardless of symptoms, signs, or other clinical or laboratory findings. Nongonococcal urethritis (NGU) was diagnosed if gonorrhea was excluded and at least two of the following three findings were present: history of urethral discharge or dysuria; purulent or mucopurulent urethral discharge on examination; and at least five polymorphonuclear leukocytes per ×1,000 microscopic field of a Gram-stained urethral smear.8 The diagnosis of HIV infection was based on the patient's history or the result of the HIV antibody test. For purposes of this study, all other diagnoses were based on either laboratory documentation or the diagnosis recorded by the examining clinician, without independent analysis of specific (albeit standardized) clinical criteria.
Data Management and Statistical Methods
Sex partner gender is recorded according to the patient's response to the question “Do you have sex with women, men, or both?”; Race is recorded according to the patient's self-description; more than 90% of those who classified themselves as Hispanic were white. Participation in anal insertive, anal receptive, oral insertive, oral receptive, and vaginal insertive intercourse is recorded as yes-or-no responses based on sexual exposures during the preceding 2 months. The reported frequency of condom use in the prior 2 months is recorded categorically as “never,” “sometimes” (1%–49% of exposures), “most” (50%–99%), or “always” (100%); these responses are not stratified according to specific sexual practices. Because patients' understanding of “sometimes” or “most” may be more subjective than for the other categories, condom use was analyzed dichotomously, comparing “always” versus all other patterns of use.
The Statistical Package for the Social Sciences (SPSS) program9 was used to perform statistical analysis. The chi-square test was used for all univariate comparisons. Multivariate logistic regression was used to evaluate the behavioral and demographic predictors of these outcomes while adjusting for other factors, using a backward stepwise procedure. All variables found to be associated (P ≤ 0.1) with the outcome of interest by univariate analysis were entered in the multivariate model. All multivariate models exclude several variables (nonwhite race, vaginal intercourse, three or more sex partners, and condom use) that were not significantly associated with any of the outcomes of interest by univariate analysis.
Patient Population and Clinical Presentation
Among the 1,253 MSM in the study group, 856 (68.3%) presented with symptomatic complaints, 245 (19.6%) sought STD screening without symptoms, and 152 (12.1%) had other reasons for presentation, such as referral by a sex partner or another health care provider. The study population is characterized in Table 1. Three hundred forty-eight men (27.8%) had a new STD that could be reasonably assumed to have been acquired within the preceding 12 months, including gonorrhea, chlamydial infection, nonchlamydial NGU, and first-episode genital or anorectal herpes. Among 162 men (12.9%) with HIV infection, 50 were diagnosed at the study visit in patients who had previously had negative tests or who had not been tested previously. One thousand sixty men (84.6%) were sexually active in the preceding 2 months, and condom use history was recorded for 912 of these patients (86.0%); in this group, 249 men (27.3%) stated they always used condoms. Among 586 men who acknowledged insertive or receptive anal intercourse in the prior 2 months, 184 (31.4%) said they always used condoms, compared with 59 (19.1%) of 309 men who had only oral sex (odds ratio [OR] 1.6, CI95 1.3–2.1, P < 0.001).
Of 958 men tested for N. gonorrhoeae, 105 (11.0%) were infected. These included 66 men with urethral gonorrhea (among 899 in whom urethral culture was done), 26 with rectal infection (among 382 with rectal cultures), and 24 with pharyngeal infection (of 666 tested). Eleven men were infected at two or more anatomic sites. All but 1 of the 66 men with urethral gonorrhea had symptomatic urethritis, with history of urethral discharge or dysuria, or abnormal urethral exudate on examination; the remaining patient had neither symptoms nor signs of urethritis. Twelve (46%) of the 26 men with rectal infection complained of anal discharge, anal pruritus, or other symptoms of proctitis, but only 4 (16%) of the 24 men with positive throat cultures had symptoms or signs of pharyngitis.
Table 2 shows the associations between urethral and rectal gonorrhea and the major demographic and behavioral variables. By univariate analysis, urethral infection was associated with both anal insertive and oral insertive intercourse, and there was a borderline association with known HIV infection. Seventeen (26%) of the urethral gonorrhea cases occurred in men who acknowledged oral insertive sex in the preceding 2 months but denied anal insertive intercourse. All these associations were strengthened in the multivariate model, in which both oral insertive intercourse (OR 4.4, CI95 1.4–13.4) and anal insertive intercourse (OR 3.2, CI95 1.7–6.1) remained the strongest predictors of urethral gonorrhea, and known HIV infection became an independent predictor. The multivariate analysis also showed negative associations of urethral gonorrhea with anal receptive sex and oral receptive intercourse.
In the analysis of rectal gonorrhea, a weak association was observed for known HIV infection in both the univariate and multivariate analyses. There were no statistically significant associations of rectal infection with specific sex practices. Pharyngeal gonococcal infection was weakly associated with nonwhite race in both the univariate (OR 2.1, CI95 0.9–5.0) and multivariate (OR 2.5, CI95 1.0–6.3) analyses, but no other significant associations were observed (data not shown).
Chlamydial Infection and Nongonococcal Urethritis
Chlamydia trachomatis was isolated from 31 (3.5%) of 896 men tested, including 27 with urethral infection (among 976 in whom urethral chlamydial culture was done) and 6 with rectal infection (of 305 tested); 2 men were infected at both anatomic sites. Among the men with urethral infection, seven (26%) were asymptomatic and had no abnormal urethral exudate on examination. Among the six men with rectal infection, four were asymptomatic. As shown in Table 3, urethral chlamydial infection was significantly associated with young age (multivariate OR 2.4, CI95 1.1–5.5) and with anal insertive sex (OR 3.2, CI95 1.3–7.6), and there was a weak negative association with anal receptive intercourse. Vaginal intercourse was not associated with urethral chlamydial infection. There were too few rectal infections for meaningful analysis.
Nongonococcal urethritis was the most frequent clinical problem observed. Two hundred twenty-seven men (18.1%) initially were diagnosed by the clinical criteria described earlier, but C. trachomatis or N. gonorrhoeae was isolated from the urethras of 16 and 4 of these men, respectively. In 11 other men, urethral specimens were not collected for isolation of one or both of these pathogens. Thus, 196 patients were definitively classified as having nonchlamydial NGU. As shown in Table 3, this diagnosis was independently associated with oral insertive sex (multivariate OR 2.2, CI95 1.3–3.7), but not with anal insertive intercourse. Being uncircumcised was negatively associated with nonchlamydial NGU (OR 0.5, CI95 0.3–0.9), as was a history of anal receptive intercourse (OR 0.6, CI95 0.4–0.9). For 81 (40%) of the men with nonchlamydial NGU, oral insertive intercourse was the only acknowledged sexual practice in the preceding 2 months.
Human Immunodeficiency Virus Infection
Serologic tests for HIV infection were done in 597 men (47.6%), of whom 54 (9.0%) were seropositive for HIV-1; no HIV-2 infections were diagnosed. Fifty of the seropositive men were not previously known to be infected; 27 of these had tested negative within the preceding 5 years and 11 within the past year. The same variables shown in Tables 2 and 3 were analyzed as possible predictors of the 50 newly documented HIV infections. In the multivariate model, new HIV infection was independently associated with anal receptive intercourse (OR 2.6, CI95 1.3–4.9), nonwhite race/ethnicity (OR 3.1, CI95 1.6–5.9), and age greater than 25 years (OR 2.5, CI95 1.5–5.0), but not with condom use or any other variable.
Despite overall trends toward safer sex practices in gay and bisexual men,1–3 some MSM remain at high risk for STD acquisition, and many of these men still seek care in STD clinics. For example, N. gonorrhoeae was isolated in 11% of MSM in this study, compared with 6% of exclusively heterosexual men who attended the STD clinics during the same time period. We found that a history of consistent condom use offered no apparent protection from either rectal or urethral STDs. Zenilman and colleagues made a similar observation in a predominantly heterosexual STD clinic population.10 Undoubtedly, the explanation for the apparent absence of protection is related to behavioral issues, sample selection, and the accuracy of patients' claims of consistent condom use, and not to the inherent efficacy of condoms. MSM who use condoms correctly and consistently may be less likely than our patients to attend STD clinics. We also suspect that the tendency of people to give socially desirable responses to questions about health-related behaviors resulted in inflated estimates of condom use. These observations indicate that a history of consistent condom use in MSM may not be reliable in assessing patients' risk for HIV infection and other STDs, and should not dissuade clinicians from conducting screening tests for common STDs. The number of sex partners in the preceding 2 months also seems to be an inadequate basis for screening decisions in MSM attending STD clinics and perhaps in other settings.
We found urethral gonococcal infection to be independently associated not only with anal insertive intercourse, as expected, but with oral insertive sex. In fact, the independent risk was higher for fellatio (OR 4.4) than for anal insertive sex (OR 3.2). Although scattered reports indicate that urethral gonorrhea can be acquired by fellatio,11,12 the risk is usually considered to be low. However, other data suggest that there has been a shift in the relative frequencies of oral and anal intercourse among MSM and that condoms are not commonly used during fellatio.13 Thus, the population attributable risk of fellatio for urethral gonorrhea in MSM may have risen in the past decade. The negative association of urethral gonorrhea with anal receptive sex (OR 0.5) and oral receptive intercourse (OR 0.4) may be related to variations in sexual behavior. For example, some MSM who participate in insertive sex practices may be less likely to have receptive intercourse. Unlike gonorrhea, urethral chlamydial infection was associated only with anal insertive intercourse and not fellatio, a finding consistent with the infrequency of pharyngeal C. trachomatis infection.14,15 Chlamydial infection was not associated with vaginal intercourse.
The strongest independent predictor of nonchlamydial NGU was oral insertive intercourse (OR 2.2), and 40% of these infections were diagnosed in men who said they had no other sexual exposures in the previous 2 months. Hernandez-Aguado et al.16 reported a similar association of Chlamydia-negative, Ureaplasma-negative NGU with fellatio in both heterosexual and homosexual men. McGowan et al.17 also described NGU in MSM in whom fellatio was the probable source of exposure. These observations suggest that oral bacteria or other oral factors can cause urethritis, although McGowan et al. found no increased prevalence of “oral” flora in the urethras of MSM with NGU compared with a control group without urethritis.17 It is interesting that we found no association of nonchlamydial NGU with anal insertive sex, because MSM are at increased risk of urethritis and urinary tract infection due to coliform bacteria, presumably acquired during anal intercourse.18 Additional research is indicated, including prospective epidemiologic studies in both heterosexual men and MSM, and comparison of the urethral microbial flora in people with NGU with the oral flora of their sexual partners.
The apparent protective effect of being uncircumcised on nonchlamydial NGU (or, conversely, an increased risk in circumcised men) may be an artifact related to other behavioral characteristics or partner selection. An earlier study in our clinic found no association between NGU and circumcision status.19 Perhaps circumcision is associated with undefined differences in partner selection or other sexual behaviors.
The association of both urethral gonorrhea and nonchlamydial NGU with oral sex might suggest that HIV may be acquired through such exposure. However, saliva has not been documented as a source of HIV infection, and our results should not be interpreted as implying a high risk for transmission of HIV to the insertive partner by fellatio. On the other hand, pharyngeal gonorrhea is readily acquired by fellatio, and HIV can be transmitted to the receptive partner in oral intercourse.20 We believe MSM should be counseled that fellatio carries a risk of acquiring HIV and other STDs for the receptive partner and gonorrhea and NGU by the insertive partner.
We found an independent association of known HIV infection with urethral gonorrhea (OR 2.3) and a similar but statistically nonsignificant association with rectal gonococcal infection (OR 2.4), suggesting that some HIV-infected men continue unprotected sex with high-risk partners. This disconcerting observation replicates those of other investigators in studies of hemophiliacs, injection drug users, and MSM.21–24 Although some HIV-infected people undoubtedly warn or otherwise protect their sex partners, and some may select partners who also are HIV positive, it is likely that many do not take these precautions.22–24 Clinicians who provide health care for HIV-infected people have a responsibility continually to reassess their patients' sexual (and substance-using) behaviors, to counsel repeatedly those who may be continuing risky practices, and to screen periodically their sexually active patients for new STDs.
Some of our findings were neither unexpected nor new. For example, our results are consistent with other data documenting receptive anal intercourse as the highest-risk sexual practice for HIV acquisition in MSM.25 We also found young age to be an independent predictor of urethral chlamydial infection, a well documented association in women26 and heterosexual men.27
This study has several limitations. People attending STD clinics are a highly selected group, and our patients undoubtedly are not representative of most MSM. The study was cross-sectional and the database was not specifically designed to explore the associations we found. The data were collected by several providers in a clinical care setting, with no attempt to standardize the wording of questions about sexual behavior. Even so, studies such as ours highlight the utility of systematic data collection in the course of providing routine STD care. Although the quality of some of the data may be inferior to that achievable in research environments, this approach has the advantage of avoiding biases that may result from the requirement for informed consent and other issues that can affect study participation.
In summary, several of our results have implications for STD/HIV prevention in MSM. Our data suggest that urethral gonorrhea and nonchlamydial NGU can be acquired by fellatio and that oral intercourse accounts for up to half of all urethral infections in MSM. We learned that a history of consistent use of condoms or of having few or no recent sexual exposures is not associated with a reduced risk of sexually acquired urethritis or rectal gonococcal infection, at least when such information is collected using brief questions in nonresearch clinical settings. Finally, we confirmed other investigators' observations that infection with HIV is a risk factor for future STD acquisition, which has important implications for continued risk assessment and counseling of HIV-infected people.
1. Doll LS, Judson FN, Ostrow DG, et al. Sexual behavior before AIDS: The hepatitis B studies of homosexual and bisexual men. AIDS 1990; 4:1067–1073.
2. Steiner S, Lemke AL, Roffman RA. Risk behavior for HIV transmission among gay men surveyed in Seattle bars. Public Health Rep 1994; 109:563–566.
3. Handsfield HH, Schwebke JR. Trends in sexually transmitted diseases in homosexually active men in King County, Washington, 1980–1990. Sex Transm Dis 1990; 17:211–215.
4. Bibus D, Wood R, Hartfield K, Hanrahan M, Wood C. A model for distributing HIV-prevention resources. AIDS and Public Policy Journal 1995; 9:197–207.
5. Holmberg SD. The estimated prevalence and incidence of HIV in 96 large U.S. metropolitan areas. Am J Public Health 1996; 86:642–654.
6. Bonin P, Tanino TT, Handsfield HH. Isolation of Neisseria gonorrhoeae
on selective and nonselective media in a sexually transmitted disease clinic. J Clin Microbiol 1984; 19:218–220.
7. Stamm WE, Tam M, Koester M, Cles L. Detection of Chlamydia trachomatis
inclusions in McCoy cell cultures with fluorescein-conjugated monoclonal antibodies. J Clin Microbiol 1983; 17:666–668.
8. Bowie WR. Urethritis in males. In: Holmes KK, et al, eds. Sexually Transmitted Diseases. 2nd ed. New York: McGraw-Hill, 1990:627–639.
9. SPSS for Windows: Advanced Statistics, Release 6.1. Chicago:SPSS, Inc., 1996.
10. Zenilman JM, Weisman CS, Rompalo AM, et al. Condom use to prevent incident STDs: The validity of self-reported condom use. Sex Transm Dis 1995; 22:15–21.
11. Soendjojo A. Gonococcal urethritis due to fellatio. Sex Transm Dis 1983; 10:41–42.
12. Tice AW, Rodriguez VL.o Pharyngeal gonorrhea. JAMA 1981;246:2717–2719.
13. Schwarcz SK, Kellog TA, Kohn RP, et al. Temporal trends in human immunodeficiency virus seroprevalence and sexual behavior at the San Francisco municipal sexually transmitted disease clinic, 1989–1992. Am J Epidemiol 1995; 142:314–322.
14. Jones RB, Rabinovich RA, Katz BP, et al. Chlamydia trachomatis
in the pharynx and rectum of heterosexual patients at risk for genital infections. Ann Intern Med 1985; 102:757–762.
15. Jebakumar SP, Storey C, Lusher M, et al. Value of screening for oro-pharyngeal Chlamydia trachomatis
infection. J Clin Pathol 1995; 48:658–661.
16. Hernandez-Aquado I, Alvarez-Dardet C, Gili M, Perea EJ, Camacho F. Oral sex as a risk factor for Chlamydia
-negative, ureaplasma-negative nongonococcal urethritis. Sex Transm Dis 1988; 15:100–102.
17. McGowan I, Radcliffe KW, Bingham JS, et al. Non-gonococcal urethritis in men practising “safe” sex. Genitourin Med 1991; 67:70–71.
18. Barnes RC, Daifuku R, Roddy RE, Stamm WE. Urinary tract infection in sexually active homosexual men. Lancet 1986; 1:171–173.
19. Cook LS, Koutsky LA, Holmes KK. Circumcision and sexually transmitted diseases. Am J Public Health 1994; 84:197–201.
20. Schacker T, Collier AC, Hughes J, Shea T, Corey L. The clinical and epidemiological presentation of primary HIV infection. Ann Intern Med 1996; 125:257–264.
21. Dublin S, Rosenberg PS, Goedert JJ. Patterns and predictors of high-risk sexual behavior in female partners of HIV-infected men with hemophilia. AIDS 1992; 6:475–482.
22. Marks G, Ruiz MS, Richardson JL, et al. Anal intercourse and disclosure of HIV infection among seropositive gay and bisexual men. J AIDS 1994:7; 866–869.
23. Centers for Disease Control and Prevention. Continued sexual risk behavior among HIV-seropositive, drug-using men—Atlanta; Washington, DC; and San Juan, Puerto Rico, 1993. MMWR 1996; 45:151–152.
24. Torian LV, Weisfuse IB, Makki HA, et al. Trends in HIV seroprevalence in men who have sex with men: New York City Department of Health sexually transmitted disease clinics, 1988–1993. AIDS 1996; 10:187–192.
25. Elifson KW, Boles J, Sweat M. Risk factors associated with HIV infection among male prostitutes. Am J Public Health 1993; 83:79–83.
26. Handsfield HH, Jasman LL, Roberts PL, et al. Criteria for selective screening for Chlamydia trachomatis
infection in women attending family planning clinics. JAMA 1986; 255:1730–1734.
27. Leonardi GP, Seitz M, Edstrom R, et al. Evaluation of three immunoassays for detection of Chlamydia trachomatis
in urine specimens from asymptomatic males. J Clin Microbiol 1992; 30:2793–2796.