Bacterial vaginosis, a common gynecologic syndrome characterized by alterations in the vaginal microflora, has been associated with a variety of adverse obstetric and gynecologic outcomes including preterm delivery, pelvic inflammatory disease, endometritis, postpartum fever, post-hysterectomy cuff cellulitis, and postabortal infection.1 Bacterial vaginosis describes a gynecologic condition characterized by an overgrowth of several species of facultative and obligate anaerobic bacteria and genital mycoplasmas and a reduction in the hydrogen peroxidase–producing lactobacilli that predominate in a normal vaginal ecosystem.2
Several cross-sectional studies have demonstrated an association between bacterial vaginosis and HIV infection.3–6 There is also some longitudinal evidence suggesting that bacterial vaginosis may be a risk factor for human immunodeficiency virus (HIV) seroconversion.7 However, little is known about the natural history of bacterial vaginosis in HIV-infected women. The purpose of this study was to describe the prevalence, incidence, persistence, and severity of bacterial vaginosis among HIV-infected and -uninfected women in a large cohort of high-risk women participating in the HIV Epidemiology Research Study (HERS). A previous report from HERS described the baseline prevalence of bacterial vaginosis at participants' enrollment in the study.8 Our study prospectively followed this cohort and includes 9458 patient visits over a 5-year period. The longitudinal nature of this data set allows for estimates of incidence and persistence, in addition to longitudinal prevalence.
MATERIALS AND METHODS
The study methodology for HERS has been described in detail elsewhere.9 Briefly, between April 1993 and January 1995, 871 HIV-infected women and 439 HIV-uninfected, at-risk women were recruited from four sites (Bronx, NY; Baltimore, MD; Detroit, MI; Providence, RI) to participate in HERS. Women aged 16–55 years who had either a history of injection drug use since 1985 or high-risk sexual behaviors (eg, over five sexual partners in the past 5 years, trading sex for drugs or money, or sex with a male injection drug user or a man known or suspected to have HIV) were eligible for the study. HIV-infected women were excluded if they had an acquired immunodeficiency syndrome (AIDS)-defining clinical diagnosis.10 Women were recruited from health care settings, drug treatment programs, social service organizations, and by word of mouth. This analysis is restricted to the 1288 women who had complete gynecologic evaluations at enrollment, and excludes 12 women who HIV-seroconverted during the study period and 10 women who did not have a gynecologic examination at enrollment. The study was approved by the institutional review boards of each of the participating medical centers and the Centers for Disease Control and Prevention (CDC); informed consent was obtained from all participants.
Participants were scheduled for study visits at enrollment and then every 6 months; this analysis includes up to ten study visits. Study visits were scheduled in advance and were not performed for the evaluation or treatment of symptoms. At each study visit, participants were administered a standardized face-to-face interview to obtain medical, obstetric, and gynecologic histories, past and current drug use and sexual behaviors, and information on selected demographic and psychosocial factors. Women were asked what medications, including antibiotics, they were currently taking and what medications they had taken in the 6 months since their last visit. Antiretroviral regimens were classified as highly active antiretroviral therapy (HAART), sub-HAART, or no antiretroviral therapy. The definition of HAART was adapted from the Department of Health and Human Services (DHHS) guidelines11 and includes regimens with clinical evidence of effective and sustained viral suppression. This includes use of a protease inhibitor or non-nucleoside reverse transcriptase inhibitor with two nucleoside reverse transcriptase inhibitors. Antiretroviral regimens that did not qualify as highly active were classified as sub-HAART.
After the interview, all women received a standardized physical exam and gave blood for HIV testing (enzyme-linked immunosorbent assay [ELISA] with confirmatory Western blot), immunophenotyping, syphilis serology, and other laboratory tests. As part of the physical exam, all women underwent a standardized pelvic examination, which included the following: 1) a description of vaginal discharge; 2) Papanicolaou smear; 3) a vaginal culture for Candida; 4) wet mounts (saline and potassium hydroxide) for Trichomonas vaginalis, clue cells, and yeast; 5) vaginal pH assessment using indicator strips (Baxter S/P, Glendale, CA); 6) a whiff-amine test performed by adding a drop of 10% potassium hydroxide to vaginal fluids and noting the release of a fishy odor; 7) vaginal Trichomonas vaginalis culture; 8) a swab of the posterior vaginal fornix used to prepare a Gram-stained slide; 9) an endocervical culture for Neisseria gonorrhoeae; and 10) an endocervical culture and direct fluorescent antibody test for Chlamydia trachomatis. Due to the low frequency, routine screening for gonorrhea and chlamydia were discontinued after the third visit. Vaginal colonization with Candida was defined as having a positive Candida culture at the current study visit. Vaginal colonization with Trichomonas was defined as either a positive culture or wet mount.
As in other studies,12 bacterial vaginosis was defined using a Gram-staining scoring system described by Nugent et al:13 a score of 7–10 was defined as bacterial vaginosis, a score of 4–6 was defined as abnormal vaginal flora, and a score of 0–3 was considered normal. To prepare the Gram-stained slide, the smear of the posterior vaginal fornix was air dried and fixed in methanol, then shipped to a central laboratory (Detroit Medical Center University Laboratories, Detroit, MI) for Gram staining. Gram-stained slides were read by a single technologist who was unaware of the clinical or HIV status of participants. Under oil immersion (×1000 magnification), slides were examined for the Lactobacillus morphotypes, Gardnerella and Bacteroides species morphotypes (small gram-negative to gram-variable rods), and Mobiluncus species morphotypes. Each of the three morphotypes was quantitated and scored as described by Nugent et al.12 For the Lactobacillus, Gardnerella, and Bacteroides species morphotypes, the scores ranged from 0–4; for the Mobiluncus species morphotype the score ranged from 0–2. For the Gardnerella, Bacteroides, and Mobiluncus morphotypes, the highest score was given to patients having the highest number of organisms. For the Lactobacillus morphotypes, the scoring was inversely related to the number of organisms present; patients with the highest numbers of organisms received the lowest score. The scores from each of the three morphotypes were added together to get a range of possible scores from 0–10. A random sample of Gram-stain slides (n = 10) was read by an independent investigator (Carol Spiegel, University of Wisconsin, Madison, WI) with 100% agreement on the final Nugent score between the two readers.
All models in this analysis use methods to account for within-subject correlation across multiple observations on individuals. We used generalized estimating equation methods with m-dependent working correlation structure. Confidence intervals (CIs) and P values are based on standard errors estimated by robust methods in SAS (SAS Institute Inc., Cary, NC). Unadjusted odds ratios (ORs) were calculated using models that included three study design variables—site, risk cohort, and HIV status—because these variables were used for frequency matching upon enrollment in the study. Unadjusted ORs were used to screen for inclusion in an initial multivariate model; variables that exhibited at least moderate association (P < .20) with the outcome in the presence of these design variables were considered for inclusion in the final models. Variables were added and removed from the multivariate models in a stepwise fashion. Two-way interaction terms composed of HIV status and each variable retained in the final model were subsequently examined.
For calculating incidence and persistence, we included only those visits for which the previous consecutive visit was available. For prevalence and incidence, normal vaginal flora (Gram-staining score of 0–3) was used as the referent group. Prevalence was defined as the number of visits with bacterial vaginosis or abnormal vaginal flora divided by the total number of visits, with an adjustment made to account for multiple visits for each woman. Whereas a cross-sectional design might yield an estimate of point prevalence, the longitudinal design of this study allows us to estimate the average prevalence over the course of the study. Incidence was defined as a visit in which a patient had bacterial vaginosis or abnormal vaginal flora present, when in the previous consecutive visit the patient had normal vaginal flora. Persistence was defined as two consecutive visits with bacterial vaginosis or two consecutive visits with abnormal flora. The referent group consisted of those patients with normal flora who had, at the previous visit, abnormal flora or bacterial vaginosis.
Severity of bacterial vaginosis was assessed in two ways. First, clinical signs were assessed using modified Amsel's criteria.14 Modified Amsel's criteria for the clinical diagnosis of bacterial vaginosis requires that three of four conditions be present: vaginal pH ≥ 4.7, abnormal vaginal discharge, positive whiff-amine test, or any clue cells on microscopy. A severe case was defined as meeting all four criteria. Second, Gram-stain criteria were used to assess severity; a score of 9 or 10 was considered severe. For the incidence, persistence, and severity models, two separate models were constructed: 1) all women were included in a model that contained an HIV status variable, and 2) HIV-infected women were included in a model that contained a CD4+ cell-count variable. Incidence and persistence were derived from the same models; severity was derived from separate models.
If a woman was diagnosed with clinical bacterial vaginosis on the basis of Amsel's criteria, at two of the sites (RI, MI) she was offered standard antibiotic therapy15 at the time of the examination. At the other two sites (NY, MD) she was referred to her primary physician for treatment.
Baseline characteristics of the 854 HIV-infected and 434 HIV-uninfected women included in this analysis are described in Table 1. Most of the women were between the ages of 25 and 44 (84%), and more than half were black. Among the 984 sexually active women, 57% (564) reported inconsistent condom use, defined as not always using a condom during sexual intercourse. Twenty-five percent of women reported using injection drugs in the past 6 months. More than half (52%) of the women had nine or 10 study visits; 9.2% (120) of enrolled women had only the baseline evaluation and no follow-up visits. At enrollment, 17% (146 of 848) of the HIV-infected women with available CD4+ cell counts had measurements of less than 200 cells/μL. At the time of enrollment, almost two-thirds (65%) of the women had not received any antiretroviral therapy, and none were taking highly active antiretroviral therapy (HAART) because HAART was not available until after enrollment was complete. By the end of the study, about 25% of women were taking HAART. Women reported metronidazole or topical clindamycin use in less than 1% of study visits.
At most of the study visits, a slightly higher percent of HIV-infected women than HIV-uninfected women had abnormal vaginal flora and bacterial vaginosis (Figure 1). In adjusted analyses, HIV-infected women were more likely than HIV-uninfected women to have bacterial vaginosis (adjusted OR 1.29; 95% CI 1.08, 1.55) and abnormal vaginal flora (OR 1.26; 95% CI 1.05,1.52) (Table 2). Other significant predictors of both abnormal vaginal flora and bacterial vaginosis in adjusted analyses included black race, less than a high school education, and a positive Trichomonas by either culture or wet mount. Vaginal Candida colonization was inversely associated with both abnormal vaginal flora and bacterial vaginosis. Significant interaction (P < .001) was detected between HIV seropositivity and vaginal Candida colonization. In women with a positive Candida culture, HIV-seropositivity was a risk factor for bacterial vaginosis (OR 1.55; 95% CI 1.18, 2.02). However, among women with a negative Candida culture, HIV-seropositivity was not a significant risk factor for bacterial vaginosis (OR 1.01; 95% CI 0.85, 1.21). Using a variable that combined frequency of sexual intercourse with regularity of condom use, more frequent sexual intercourse combined with inconsistent condom use was a risk factor for bacterial vaginosis. In addition, frequent sexual intercourse was also a risk factor in those with consistent condom use (Table 2).
Among HIV-infected women, low CD4+ cell count (less than 200 cells/μL) was a risk factor for bacterial vaginosis in the adjusted analysis (adjusted OR 1.29; 95% CI 1.03, 1.60) (Table 2). Use of HAART was inversely associated with bacterial vaginosis (adjusted OR 0.64; 95% CI 0.49, 0.82); use of sub-HAART was also inversely associated with bacterial vaginosis, but with a point estimate suggesting somewhat less protection than with use of HAART (adjusted OR 0.74; 95% CI 0.63, 0.86). There was no increased prevalence of abnormal vaginal flora among HIV-infected women with CD4+ cell count of less than 200 cells/μL.
There was no difference in the incidence of bacterial vaginosis by HIV status or with respect to CD4+ group (Table 3). HIV-infected women had a higher incidence of abnormal vaginal flora than uninfected women did (adjusted OR 1.30; 95% CI 1.00, 1.70). Among HIV-infected women, those with a CD4+ cell count of less than 200 cells/μL had a higher incidence of abnormal vaginal flora than HIV-infected women with a CD4+ cell count of over 500 cells/μL (adjusted OR 1.92; 95% CI 1.27, 2.90). HIV-infected women were more likely than HIV-uninfected women to have persistent bacterial vaginosis (adjusted OR 1.49; 95% CI 1.18, 1.89) (Table 3). Among HIV-infected women, those with a CD4+ cell count of less than 200 cells/μL were more likely than women with higher CD4+ cell counts (over 500 cells/μL) to have persistent bacterial vaginosis. (adjusted OR 1.38; 95% CI 1.01, 1.91) (Table 3).
Using two different methods to assess severity of bacterial vaginosis, both clinical criteria and Gram-stain scoring, there were no differences in severity by HIV status (Table 3). However, among HIV-infected women, those with a CD4+ cell count of less than 200 cells/μL were more likely than women with higher CD4+ cell counts (over 500 cells/μL) to have severe bacterial vaginosis using either of the two criteria (adjusted OR 1.40; 95% CI 1.08, 1.82 for clinical criteria; adjusted OR 1.50; 95% CI 1.12, 2.00 for Gram-stain scoring).
Previously, several cross-sectional studies reported higher rates of bacterial vaginosis among HIV-infected women compared with women not infected with HIV, in both international3,5,6 and domestic settings.4 There is also recent evidence suggesting that bacterial vaginosis may alter susceptibility of women to HIV and may be a risk factor for HIV acquisition7,16 However, less is known about the natural history of bacterial vaginosis in women once they are infected with HIV, and whether the natural history of bacterial vaginosis differs in HIV-negative and -positive women.
Although most cross-sectional studies have found an association of HIV-seropositivity and bacterial vaginosis, the Women's Interagency HIV Study (WIHS), a large cohort study of HIV-infected women and a high-risk comparison group, did not confirm the findings of other cross-sectional studies.17 A cross-sectional analysis of data from WIHS reported a high prevalence of bacterial vaginosis diagnosed by Gram staining, but found that bacterial vaginosis was more prevalent among HIV-uninfected women when compared with HIV-infected women (48% versus 42%; P = .03). Although it is unclear why the WIHS did not find a positive association of HIV-seropositivity and prevalent bacterial vaginosis, it was limited in its ability to draw conclusions about prevalence because of the cross-sectional design of the analysis. For example, because the prevalence of disease is dependent on the rate of occurrence of new disease (incidence) and the duration of the disease (persistence), a longitudinal analysis is necessary to better understand any potential relationship between HIV-seropositivity and bacterial vaginosis.
This longitudinal report of bacterial vaginosis in a large cohort of HIV-infected women and a high-risk comparison group of HIV-uninfected women describes the prevalence of bacterial vaginosis over a 5-year period and breaks down prevalence into its component parts, incidence and persistence. In this analysis, HIV-infected women have a higher prevalence of bacterial vaginosis compared with HIV-uninfected women. This increased prevalence is due to more persistent infections rather than more incident infections. Similarly, among HIV-infected women, immunocompromised women with a CD4+ cell count of less than 200 cells/μL have a higher prevalence of bacterial vaginosis compared with HIV-infected women with higher CD4+ cell counts. Again, this increased prevalence was the result of increased persistence and not increased incidence of bacterial vaginosis. An increased prevalence of bacterial vaginosis with decreasing CD4+ cell count was also reported in a study of HIV-infected women in Thailand.18
In addition to HIV-seropositivity, another important risk factor for bacterial vaginosis identified in this analysis was sexual behavior, with more frequent sex and inconsistent condom use associated with increased risk for prevalent bacterial vaginosis. The evidence regarding a potential role of sexual transmission in the etiology of bacterial vaginosis is conflicting.14,19 The current report supports the hypothesis that sexual behavior is important in the etiology of bacterial vaginosis and that consistent condom use may be important in preventing bacterial vaginosis.
Somewhat paradoxically, a positive vaginal Candidal culture was associated with a decreased risk of bacterial vaginosis in this and other studies.16 It has been suggested that the overgrowth of anaerobic bacteria and increased pH associated with bacterial vaginosis might inhibit fungal growth.
The severity of some genital tract infections, such as pelvic inflammatory disease,20 may be increased in immunocompromised women compared with HIV-infected women with higher CD4+ cell counts. Therefore, it is not surprising that we found immunocompromised women (with less than 200 CD4+ cells/μL) are at increased risk for having severe bacterial vaginosis, assessed both clinically and using Gram-staining criteria.
The major limitation of this study was the ascertainment of complete treatment information. Because this was a cohort study and not a randomized, controlled clinical trial, information regarding treatment for bacterial vaginosis was limited. However, at each visit women were asked about interval medication use, including the use of antibiotics. In only a very small percentage of study visits (less than 1%), metronidazole or topical clindamycin use was reported. This suggests that widespread and frequent treatment of bacterial vaginosis was not occurring. Additionally, in univariate analyses, metronidazole use was not associated with any of the outcomes of interest.
Early in the HIV/AIDS epidemic, there was widespread concern about the gynecologic manifestations of HIV-associated illness and speculation that the prevalence and severity of genital tract infections in HIV-infected women compared with uninfected women might be increased. For a number of genital tract infections, differences in infection rates and severity between HIV-infected and -uninfected women seem to be due more to the differences in behavior and other confounding factors, rather than true, biologic, clinically significant differences.21 In terms of bacterial vaginosis, we found a modest, but significant difference in the prevalence and persistence of bacterial vaginosis among HIV-infected women, particularly among immunocompromised women. This situation may be somewhat analogous to infection with human papillomavirus (HPV), in which the prevalence and persistence of cervical HPV is increased among HIV-infected women compared with women not infected with HIV.22
The positive association of HIV-seropositivity with prevalent bacterial vaginosis has important implications for the care of HIV-infected women. Bacterial vaginosis has been associated with a variety of adverse gynecologic and obstetric outcomes.1 As increasing numbers of women are living with HIV infection, the magnitude of problems associated with bacterial vaginosis may also increase. It will therefore be important to evaluate treatment of bacterial vaginosis so that adverse sequelae of bacterial vaginosis can be averted in both HIV-infected and -uninfected women.
1. American College of Obstetricians and Gynecologists. Vaginitis. ACOG technical bulletin no. 217. Washington DC: American College of Obstetricians and Gynecologists, 1996.
2. Sobel JD. Vaginitis. N Engl J Med 1997; 337(26):1896–903.
3. Cohen CR, Duerr A, Pruithithada N, Rugpao S, Hillier S, Garcia P, et al. Bacterial vaginosis and HIV seroprevalence among female commercial sex workers in Chiang Mai, Thailand. AIDS 1995;9(9):1093–7.
4. Royce RA, Thorp J, Granados JL, Savitz DA. Bacterial vaginosis associated with HIV infection in pregnant women from North Carolina. J Acquir Immune Defic Syndr Hum Retrovirol 1999;20(4):382–6.
5. Sewankambo N, Gray RH, Wawer MJ, Paxton L, McNaim D, Wabwire-Mangen F, et al. HIV-1 infection associated with abnormal vaginal flora morphology and bacterial vaginosis. Lancet 1997;350(9077):546–50.
6. Taha TE, Gray RH, Kumwenda NI, Hoover DR, Mtimavalye LA, Liomba GN, et al. HIV infection and disturbances of vaginal flora during pregnancy. J Acquir Immune Defic Syndr Hum Retrovirol 1999;20(1):52–9.
7. Taha TE, Hoover DR, Dallabetta GA, Kumwenda NI, Mtimavalye LA, Yang LP, et al. Bacterial vaginosis and disturbances of vaginal flora: Association with increased acquisition of HIV. AIDS 1998;12(13):1699–706.
8. Warren D, Klein RS, Sobel J, Kieke B, Brown W, Schuman P, et al for the HIV Epidemiology Research Study Group. A multicenter study of bacterial vaginosis in women with or at risk for human immunodeficiency virus infection. Infect Dis Obstet Gynecol 2001;9:133–41.
9. Smith DK, Warren DL, Vlahov D, Schuman P, Stein MD, Greenberg BL, et al. Design and baseline participant characteristics of the Human Immunodeficiency Virus Epidemiology Research (HER) Study: A prospective cohort study of human immunodeficiency virus infection in US women. Am J Epidemiol 1997;146(6):459–69.
10. Centers for Disease Control and Prevention. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Morb Mortal Wkly Rep 1992;41(RR-17):1–19.
11. Panel on Clinical Practices for Treatment of HIV Infection. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. Convened by the Department of Health and Human Services and the Henry J. Kaiser Family Foundation. Available at: http://hivatis.org
12. Carey JC, Klebanoff MA, Hauth JC, Hillier SL, Thom EA, Ernest JM, et al. Metronidazole to prevent preterm delivery in pregnant women with asymptomatic bacterial vaginosis. National Institute of Child Health and Human Development Network of Maternal-Fetal Medicine Units. N Engl J Med 2000;342(8):534–40.
13. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 1991;29(2):297–301.
14. Amsel R, Totten PA, Spiegel CA, Chen KC, Eschenbach D, Holmes KK. Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations. Am J Med 1983;74(1):14–22.
15. Centers for Disease Control and Prevention. 1998 guidelines for treatment of sexually transmitted diseases. MMWR Morb Mortal Wkly Rep 1998;47(RR-1):1–111.
16. Martin HL, Richardson BA, Nyange PM, Lavreys L, Hillier SL, Chohan B, et al. Vaginal lactobacilli, microbial flora, and risk of human immunodeficiency virus type 1 and sexually transmitted disease acquisition. J Infect Dis 1999;180(6):1863–8.
17. Greenblatt RM, Bacchetti P, Barkan S, Augenbraun M, Silver S, Delapenha R, et al. Lower genital tract infections among HIV-infected and high-risk uninfected women: Findings of the Women's Interagency HIV Study (WIHS). Sex Transm Dis 1999;26(3):143–51.
18. Rugpao S, Nagachinta T, Wanapirak C, Srisomboon J, Suriyanon V, Sirirojn B, et al. Gynaecological conditions associated with HIV infection in women who are partners of HIV-positive Thai blood donors. Int J STD AIDS 1998;9(11):677–82.
19. Holst E, Wathne B, Hovelius B, Mardh PA. Bacterial vaginosis: Microbiological and clinical findings. Eur J Clin Microbiol 1987;6(5):536–41.
20. Barbosa C, Macasaet M, Brockmann S, Sierra MF, Xia Z, Duerr A. Pelvic inflammatory disease and human immunodeficiency virus infection. Obstet Gynecol 1997;89:65–70.
21. Hader SL, Smith DK, Moore JS, Holmberg SD. HIV infection in women in the United States: Status at the Millennium. JAMA 2001;285(9):1186–92.
© 2001 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
22. Sun XW, Kuhn L, Ellerbrock TV, Chiasson MA, Bush TJ, Wright TC Jr. Human papillomavirus infection in women infected with the human immunodeficiency virus. N Engl J Med 1997;337(19):1343–9.