This study was undertaken to assess whether the prevalence of lower genital tract infections among human immunodeficiency virus (HIV)—seropositive women was higher than among high-risk HIV-seronegative women at their baseline visit for the HIV Epidemiology Research Study. Results were available for 851 HIV-seropositive and 434 HIV-seronegative women. Human papilloma virus (HPV) infection was more prevalent among HIV-seropositive women (64% vs. 28%). Bacterial vaginosis was common (35% vs. 33%), followed by trichomoniasis (12% vs. 10%), syphilis (8% vs. 6%), Chlamydia trachomatis infection (4% vs. 5%), candidal vaginitis (3% vs. 2%), and Neisseria gonorrhoeae infection (0.8% vs. 0.3%). Alcohol use (odds ratio [OR], 1.8; 95% confidence interval [CI], 1.3–2.4) and smoking (OR, 1.8; 95% CI, 1.3–2.5) were associated with bacterial vaginosis. Bacterial vaginosis (OR, 2.3; 95% CI, 1.5–3.4), trichomoniasis (OR, 2.3; 95% CI, 1.1–4.7), and syphilis (OR, 3.1; 95% CI, 1.3–7.4) were found to be more prevalent among black women. Our study showed no statistically significant difference in the prevalence of lower genital tract infections except for HPV between HIV-infected and demographically and behaviorally similar HIV-uninfected high-risk women.
Reported AIDS cases in women have increased substantially, from 3% of all cases in the United States in 1981 to 16% in 1998 . Lower genital tract infections, including vulvovaginitis (candidal vaginitis, trichomoniasis, bacterial vaginosis), are common among women. There have been few published data on the relationship between HIV infection and lower genital tract infections, particularly regarding prevalence, severity, and response to treatment.
There have been suggestions of increased prevalence or severity of some lower genital tract infections in women with HIV infection. Numerous reports have shown a higher prevalence of human papilloma virus (HPV) infection among HIV infected women [2–5]. Chronic, recurrent candidal vaginitis has been reported to be frequent among HIV-seropositive women [6, 7]. However, because these were case reports, the relationship between HIV infection and candidal vaginitis has been disputed . Some authors have reported higher prevalences of syphilis and Neisseria gonorrhoeae among HIV-infected women [9, 10]. Others have noted a higher prevalence of asymptomatic genitourinary chlamydial infections among HIV-seropositive women .
There is a complex interaction between HIV infection and many of the lower genital tract infections, since the latter often are sexually transmitted. Sexually transmitted diseases (STDs) that produce ulcerative lesions, such as syphilis, genital herpes, and chancroid, have been associated with higher rates of HIV transmission and acquisition [12, 13]. Nonulcerative lower genital tract infections have also been associated with HIV transmission . In addition, the immunodeficiency associated with HIV infection might affect the clinical course of some lower genital tract infections and their susceptibility to treatment.
To better understand the potential role of lower genital tract infections in the transmission and acquisition of HIV infection and the effect of HIV infection on the risk for and clinical course of lower genital tract infections, it is critical to assess their prevalence and evaluate the factors associated with lower genital tract infections among HIV-seropositive women, compared with a demographically and behaviorally similar HIV-seronegative cohort.
The present study was undertaken to assess the prevalence of lower genital tract infections among HIV-seropositive and high-risk HIV-seronegative women participating in a longitudinal cohort study and to determine the relationship between selected lower genital tract infections identified at the baseline visit and HIV serostatus, immunologic status, demographic characteristics, and sexual and behavioral risk factors.
Materials and Methods
The HIV Epidemiology Research Study is an ongoing longitudinal cohort study of determinants of disease progression in HIV-infected women. There are 4 sites: Brown University (Providence, RI), Johns Hopkins School of Hygiene and Public Health (Baltimore, MD), Montefiore Medical Center (Bronx, NY), and Wayne State University School of Medicine (Detroit, MI).
Between April 1993 and January 1995, 871 HIV-infected women and 439 HIV-uninfected women with similar demographic and behavioral characteristics were enrolled. At each site, similar proportions of women by race and HIV risk behavior were recruited from HIV-seropositive and -seronegative women. Eligibility criteria for enrollment in the study included HIV status documented within the previous 60 days or consent to HIV testing as part of eligibility screening; fluency in either English or Spanish; age 16–55 years; and report of ⩾1 HIV risk behaviors. Drug use risk was defined as having injected drugs at least once since 1985. Sexual risk was defined as either having sex with ⩾5 partners in the previous 5 years or ever having sex with a male iv drug user, exchanging sex for money or drugs, or having sex with a man suspected or known to be infected with HIV. Exclusion criteria were no identified HIV risk behavior; risk only by transfusion history or birth to an HIV-infected woman; not born female; or not consenting to the full protocol, including pelvic examination, phlebotomy, and repeated HIV counseling and testing (for seronegative women). Detailed information on the selection of participants, study design, and purpose have been published elsewhere . Women consented to an interview, physical examination including a pelvic examination, and collection of blood, urine, and cervicovaginal specimens, which were to be done every 6 months. The study visitswere not symptomdriven. The majority of the women had the baseline examination on the same day of signing the consent form. All physical examinations were done within 2 weeks of enrollment.
Before data collection began, interviewers received uniform, centralized training. Retraining and problem solving were done on a regular basis. Interviewers were observed during annual site visits. The quality control staff at the data management center randomly assigned specific participant visits for each interviewer at which interviews were audiotaped for central evaluation of interviewer proficiency. Physical examiners also received common training to standardize sample collection procedures and diagnosis of clinical conditions.
The core interview covered the previous 6-month history of reproductive events; contraceptive intentions and use; and gynecologic symptoms (vaginal discharge), illnesses (particularly yeast infections), and procedures. Record abstraction was obtained for hospitalizations, outpatient diagnostic procedures, or outpatient or emergency room visits that generated HIV-related diagnoses as reported by any study visit. This did not cover symptomatic lower genital tract infections.
During pelvic examination, samples were collected to test for the presence of lower genital tract infections. For HPV assays, cervicovaginal lavage fluid was aliquoted and frozen at –70°C until testing by PCR, with use of the MY09/MY11/HMBO1 L1 consensus primers . Neisseria gonorrhoeae was cultured on Thayer-Martin medium. Specimens for culture were planted onsite and transported with a CO2 cartridge to the laboratory at room temperature. Average time from collection to the laboratory was 3–4 h. Neisseria gonorrhea was identified by colony morphology. Chlamydia trachomatis was identified by staining with fluorescein-conjugated monoclonal antibodies. Trichomoniasis was diagnosed by the presence of trichomonads on saline mount. The clinical diagnosis of bacterial vaginosis was made if ⩾3 of the following objective criteria were present: abnormal vaginal discharge as reported by the clinician; vaginal pH of >4.5; presence of clue cells; and/or positive amine test with release of fishy odor on addition of 10% KOH to vaginal secretions . Candidal vaginitis was defined by a culture positive for Candida species with abnormal discharge (as reported by the clinician) and either erythema or edema of the vulva or vagina on pelvic examination. Serological testing for syphilis was done with the use of rapid plasma reagin (RPR) test for screening and the fluorescent treponema antibody absorption test (FTA-ABS) for confirmation.
Basic summary statistics, such as proportions, means, and medians, were used to characterize demographic attributes of the HIV seropositive and HIV-seronegative subgroups. We compared the prevalence of lower genital tract infections by HIV status using an OR; we adjusted for site-to-site variability using indicator variables in a logistic regression analysis, which estimates a common OR across sites. To assess the effect of several risk factors simultaneously, logistic regression models were used (see Results). The fit of each model was checked by use of the deciles of risk statistic of Hosmer and Lemeshow . Data analyses were carried out with SAS version 6.12 (SAS Institute, Cary, NC); logistic regressions were fit by use of Proc Genmond in SAS.
Results were available for 851 HIV-seropositive and 434 HIV-seronegative women who underwent baseline interview and pelvic examination. The mean age of the women was 35 years. Seropositive women were less likely to have completed high school (55% vs. 63%; P = .005). Because history of iv drug use was a matching factor, the proportions of former drug users were comparable by HIV status (52% vs. 50%). The median CD4 cell count among the HIV-seropositive cohort was 377/ μL; 17% had CD4 cell counts of <200/μL, 50% had counts of 200–500/μL, and 32% had counts of 1500/μL.
Self-reported risk factors (within the previous 6 months) are presented in table 1. More black women were enrolled in the HIV-infected group. HIV-seronegative women were more likely to have been sexually active in the previous 6 months, to have had multiple sexual partners, to have reported partners with less consistent condom use, and to have used crack cocaine than were HIV-infected women. Most of the women self-identified as heterosexual.
Seventy-five percent were smokers, with no difference by serostatus. Sixty-one percent reported alcohol intake, with slightly fewer drinkers among HIV-positive women (59% vs. 65%). Eight percent of the women had had a hysterectomy, and this did not differ by HIV serostatus. Twenty-five women (2%) did not have a baseline gynecologic examination and were excluded from the analysis.
Table 2 shows the prevalences of several lower genital tract infections at the baseline visit, stratified by HIV serostatus. HPV infection was more prevalent among HIV-seropositive women (64% vs. 28%). Bacterial vaginosis was common (35% among HIV-positive subjects, 33% among HIV-negative subjects), followed by trichomoniasis (12% vs. 10%), syphilis (8% vs. 6%), C. trachomatis infection (4% vs. 5%), candidal vaginitis (3% vs. 2%), and N. gonorrhoeae infection (0.8% vs. 0.3%). None of the crude ORs comparing the frequencies of lower genital tract infections by HIV status showed significant differences in prevalence except for HPV infection.
Syphilis status was based on serology. Among HIV-seropositive women, 95 (11.2%) of 841 had positive results from RPR testing and 71 (8.4%) had positive results from both RPR and FTA-ABS testing. Among HIV-seronegative women, 35 (8.1%) of 431 had positive results of RPR testing and 24 (5.6%) had positive results from both RPR and FTA-ABS testing.
Among HIV-seropositive women, we found no statistically significant difference in the prevalence of lower genital tract infections by CD4 cell count, except for a higher prevalence of trichomoniasis among women with CD4 cell counts of >500/ μL (OR, 1.9; 95% CI, 1.1–3.0).
We calculated adjusted ORs using logistic regression to control for several behavioral and demographic factors for HPV infection, bacterial vaginosis, trichomoniasis, and syphilis. The other lower genital tract infections were too few in number to be studied with a model. We adjusted for race, use of crack cocaine, iv drug use, alcohol intake, cigarette smoking in the past 6 months, and sexual behavior in the past 6 months (including whether sexually active, number of partners, and frequency of condom use). The conditional OR estimates from the logistic regression model appear in table 3. HPV infection was significantly more prevalent among HIV-seropositive women. Estimates from each model indicate that bacterial vaginosis, trichomoniasis, and syphilis were more prevalent among HIV-positive women, even after adjusting for the factors listed above; however, the effect was not statistically significant in any of the models. For trichomoniasis and syphilis, relatively low prevalence (and hence reduced power) may explain lack of statistical significance. Following are summaries from each model.
HPV prevalence was higher among HIV-infected women (OR, 4.7; 95% CI, 3.5–6.1). Sex for drugs or money (OR, 1.3; 95% CI, 1.9–1.8) was associated with HPV infection.
Bacterial vaginosis was somewhat more prevalent among HIV-positive women (OR, 1.2; 95% CI, 0.9–1.6). Alcohol use (OR, 1.8; 95% CI, 1.3–2.4) and smoking (OR, 1.8; 95% CI, 1.3–2.5) were significantly associated with the presence of bacterial vaginosis. Women with multiple partners had higher prevalence (OR, 1.4; 95% CI, 1.0–1.9), although the effect was not statistically significant. The odds of bacterial vaginosis among black women were found to be more than twice that of white women.
Prevalence of trichomoniasis was also higher among HIV seropositive women (OR, 1.2; 95% CI, 0.8–1.9) and among iv drug users (OR, 1.5; 95% CI, 1.0–2.4), although the differences were not statistically significant. Trichomoniasis was found to be more prevalent among black women than among white women (OR, 2.3; 95% CI, 1.1–4.7).
Syphilis was also found to be more prevalent among HIV seropositive women conditional on the behavioral and demographic factors (OR, 1.4; 95% CI, 0.8–2.3). Among sexually active women, those with history of engaging in sex for drugs or money were 4.3× more likely to have syphilis than were those who did not. Syphilis was more prevalent among black women than among white women (OR, 3.1; 95% CI, 1.3–7.4).
In this multisite study, HPV infection was more prevalent among HIV-infected women, but we found no statistically significant difference in the prevalence of N. gonorrhoeae or C. trachomatis infection, syphilis, bacterial vaginosis, trichomoniasis, and candidal vaginitis by HIV serostatus. This study included demographically and behaviorally similar cohorts, unlike previous studies.
Several studies have shown increased prevalence and incidence of genital tract infections among HIV-infected women. These studies have reported rates for gonorrhea at 7%–17%, chlamydia at 12%–18%, syphilis at 22%–29%, bacterial vaginosis at 42%–55%, trichomoniasis at 27%, and candidal colonization or vaginitis at 10%–35%. Some of the studies had no control groups; others involved select populations, such as prostitutes or pregnant women [19–24].
Our findings are similar to those of previous studies showing that HPV infection is more prevalent among HIV-infected women than among HIV-uninfected women [2–5]. Among the HIV-infected women, there was no difference in the prevalence of HPV infection by CD4 cell count. It should be noted that only 17% of the women had CD4 cell counts of <200/¼L. A cohort with more advanced immunosuppression may show different results. These findings are important because of the role of HIV and HPV infection in the development of cervical dysplasia and cancer. In 1992, the Centers for Disease Control and Prevention added cervical cancer as an AIDS-defining diagnosis .
The prevalence of several STDs (e.g., N. gonorrhoeae and C. trachomatis infection) was lower in this study than in some previous studies, particularly those conducted in STD clinics. There are several explanations for this: pelvic examination and tests for lower genital tract infections were done on a scheduled study visit and were not symptom-driven; bias may have occurred if HIV-infected women had been treated recently for lower genital tract infections before enrollment in the study or if HIV-infected women modified their behaviors due to illness or AIDS education efforts; gonorrhea and chlamydial infections were diagnosed by culture rather than by amplification techniques; and the mean age of the women enrolled in the study was 35 years. Over the last decade, gonorrhea rates have declined among both men and women, except among those aged 15–19 years. In 1991, the gonorrhea rate was 1043 per 100,000 among women aged 15–19 years, which was the highest rate among women .
The number of diagnosed cases of candidal vaginitis was very low (2%–3%). There may be several reasons for this. We used a very strict definition for candidal vaginitis (culture positive for Candida with abnormal discharge, and either erythema or edema of the vaginal or vulva on pelvic examination); the baseline questionnaire failed to capture vulvar pruritus or itching, which is the most common symptom associated with candidal vaginitis; and 83% of the HIV-infected women had CD4 cell counts of >200/¼L, which is significant because acute or recurrent candidal vaginitis may occur more in those with CD4 cell counts of >200/¼L or who are receiving antibiotics either for treatment or prophylaxis. Our results, however, are similar to those of a New York study in which the rate of candidal vulvovaginitis was 3.5% and was similar among immunocompetent HIV-seropositive and HIV-seronegative women. The study showed that vaginal colonization and symptomatic vaginitis were increased only among women with CD4 counts of <200/¼L .
Being black was significantly associated with the diagnosis of bacterial vaginosis, trichomoniasis, and syphilis in our study. In a large study of pregnant women, the prevalence of bacterial vaginosis varied widely by ethnicity, from 8.8% of white women to 22.7% of black women . In 1991, the rate of primary and secondary syphilis among black women was ∼10× higher than the rate among Hispanics and >60× higher than the rate among whites .
Our study found that, except for HPV infection, there was no significant difference in the prevalence of lower genital tract infections between HIV-infected women and demographically and behaviorally similar HIV-uninfected high-risk women. However, there was a paucity of gonorrhea and chlamydial infection in this group of women, whose mean age was 35 years. Further studies are necessary to determine whether findings would be similar in a younger group of women with a higher prevalence of STDs. We found that, with the exception of chlamydial infection, the estimated OR is >1 for each lower genital tract infection, which indicates that each is more prevalent among HIV-infected than among uninfected women. The increase in prevalence among HIV-seropositive women ranged from 20% to >400% (in the case of HPV). Lack of statistical significance in many of the infection-specific tests for the effect of HIV may give the impression that lower genital tract infection is not associated with HIV; however, strict reliance on statistical significance for drawing conclusions may be misleading here. From a clinical and public health point of view, we believe that it is appropriate to consider the OR estimates in the aggregate; taken together, they suggest that HIV-seropositive women have higher prevalence of lower genital tract infections.
Several lower genital tract infections considered here have low prevalence in our cohort, which leads to reduced power. Further studies of these lower genital tract infections is warranted and should be targeted to cohorts in which the prevalence of lower genital tract infection is expected to be higher.
Bacterial vaginosis and trichomoniasis were common infections. These nonulcerative lower genital tract infections, which are often undiagnosed and untreated, may have significant impact on sexual transmission of HIV. Although the number of studies implicating nonulcerative lower genital tract infections in HIV transmission and acquisition have been few, it is possible that even a slight increase in the risk of HIV transmission caused by nonulcerative lower genital tract infections might lead to a large number of HIV infections because of the high prevalence of these infections in the general population. Bacterial vaginosis causes a change in vaginal microflora and pH, and trichomoniasis induces an inflammatory response producing more cells in the genital mucosa capable of transmitting or acquiring HIV. Studies in Thailand and Uganda have found an increased frequency of HIV-1 infection associated with bacterial vaginosis [21, 22]. Gardnerella vaginalis has been found to activate HIV production in vitro .
Our findings have several public health implications. Lower genital tract infections, including STDs, are thought to facilitate HIV transmission by disrupting normal epithelial barriers and by recruiting of HIV-infected and HIV-susceptible cells to the infected area, which results increased infectiousness or susceptibility of the individual . Some data indicate that HIV shedding is increased in women with cervicitis . Proper diagnosis and treatment of lower genital tract infections may decrease sexual transmission of HIV. A study done in Malawi showed that HIV-1—seropositive men with urethritis had HIV-1 RNA concentrations in seminal plasma 83 higher than those of seropositive men without urethritis (P = .035). Antimicrobial therapy directed against STDs significantly decreased HIV-1 concentrations in the semen . Improved management of STDs in rural health units in the Mwanza region, Tanzania, reduced the incidence of HIV-1 infection in the general population by ∼40% . HIV-infected and high-risk HIV-uninfected women need regular gynecologicmonitoring beyond testing for and treatment of the classic STDs. Pelvic examination should include assessment for vulvovaginitis.
Future studies are needed to further assess the impact of lower genital tract infections on HIV shedding in the genital tracts of women and to determine the response of lower genital tract infections to standard therapy among HIV-infected women and the effect of such therapy on the amount of HIV genital shedding. There is also a need to evaluate the effect of immunosuppression on the frequency and severity of lower genital tract infections. Additional studies should assess whether HIV transmission and acquisition can be reduced by preventing or treating lower genital tract infections.
We would like to thank the study staff and especially the women who participated in the study.