IN 1955, GARDNER AND DUKES1 first described “Hemophilus vaginalis vaginitis,” now characterized as bacterial vaginosis (BV). They concluded that it was a sexually transmitted disease (STD), based on the high prevalence of H. vaginalis (renamed Gardnerella vaginalis) in male contacts of women with BV and on the ability to transmit BV to other women by vaginal inoculation of vaginal fluid from women with BV. Later, it was shown that G. vaginalis is also common in women without signs of BV,2 and that other bacteria, such as Mobiluncus spp, other strictly anaerobic bacteria, and Mycoplasma hominis are involved in the vaginal bacterial flora changes seen in BV.
Although clinical symptoms associated with BV are mild, it is also associated with more severe sequelae such as pelvic inflammatory disease and obstetric complications such as premature labor, premature rupture of membranes, preterm birth, amniotic fluid infection, and postpartum endometritis.3–5 Prevention of these consequences calls for a more adequate understanding of the etiology and risk factors associated with BV.
Sexual contact as a mode of transmission for BV remains a controversial matter. Previous studies have studied indirect measures of sexual activity to suggest such transmission: the prevalence of concomitant STDs,6 low prevalence of BV in virgins,7 and its occurrence in sexually abused girls,8 lesbians,9 and contraceptive users.2 However, there is no proven effect of partner therapy,10–13 and several of these studies9,10–13 indicate that BV is not an STD as generally defined.
Some studies have investigated limited sexual history factors with regard to BV,14–17 all of which found a correlation between multiple sexual partners and BV. However, the literature lacks studies that have analyzed detailed sexual histories of women with this type of change in their vaginal flora. Therefore, the aim of this study was to elucidate if women with BV share sexual characteristics (in addition to multiple partners) with women who had acquired an established STD such as chlamydial infection. We included sexual risk factors such as casual sex, sex during travel, infidelity, nonvaginal intercourse, and a history of abuse and rape in a multivariate analysis of BV.
Materials and Methods
Between November, 1989 and January, 1991, three to four women per day were randomly chosen from the out-patient lists of the family planning clinics at Eskilstuna Hospital, Eskilstuna, and Danderyd Hospital, Stockholm, and of the youth clinic in Eskilstuna, to be enrolled in the Swedish Women's Study.18 Eskilstuna is a mediumsized town in Sweden.
Specially trained midwives conducted a personal, structured, in-depth interview that included detailed questions on different aspects of sexual behavior. Confidentiality was guaranteed to ensure frank and complete answers. Identifying information that could link a patient with interview data was not included in any patient record forms or in computer files. Only the authors had access to the patient codes. The 96 questions on sexual behavior included the number of recent and lifetime sexual partners, history and frequency of group sex, oral (male recipient) and anal sex, and homosexuality, as well as questions on casual sex, sex during travels (with previously unknown male partner), and drug use associated with sex. Details about first intercourse, including age and location, frequency of masturbation, orgasm ability in different situations, general sexual satisfaction, use of sexual devices (including shower jet) and toys, infidelity, and casual sex with previously unknown men were also asked for. Finally, there were questions about negative sexual experiences, such as sexual harassment, abuse, and rape.
Medical and Laboratory Data
Bacterial vaginosis was diagnosed according to the criteria of Amsel et al.2 Diagnosis of infection with Chlamydia trachomatis was based on cervical and urethral cultures using cycloheximide-treated McCoy cells19 or by enzyme immunoassay (Chlamydiazyme; Abbott Laboratories, Abbott Park, IL), followed by Abbott's confirmatory test (cervical specimens only). Specimens were obtained with nontoxic, cotton-tipped swabs and transported to the laboratory in a sucrose phosphate buffer with 10% fetal calf serum. Syphilis and human immunodeficiency virus (HIV) serology, and culture for Neisseria gonorrhoeae (cervix, urethra, and rectum) were performed with standard methods. Trichomoniasis was diagnosed by vaginal wet smear microscopy.
Analysis was performed using the Statistical Analysis System JMP statistical program.20 Comparisons were made using chi-square tests (Pearson and likelihood ratio) for nominal and t test for continuous variables. Logistic regression (analysis of log-likelihood) was used for multivariate analyses. A P value less than 0.05 was regarded as statistically significant.
Of the 1,077 women who were asked, 1,011 (93.9%) agreed to participate in the study. Because of incomplete records, 55 women (5.4%) were excluded from the study, which left 956 (88.8%) women to be evaluated.
One hundred thirty-one women (13.7%) met the criteria for BV. The remaining 825 (86.3%) women served as a comparison group. There was no difference in the mean age of women with BV (26.6 ± 6.9 years) and the comparison group (25.7 ± 6.9 years, P = 0.15). However, all further comparisons were adjusted for age.
C. trachomatis infection (CT) was diagnosed in 85 women (8.9%). Of these, 13 also had BV and were excluded from the analyses. Comparisons of women with BV and those with CT were thus based on 118 cases of BV and 72 cases of CT. Women with CT were significantly younger (23.5 ± 5.3 years, P = 0.0003) than those with BV. Therefore, all comparisons between women with BV and those with CT were also adjusted for age.
One case each of trichomoniasis and gonorrhea was found. No cases of HIV infection or syphilis were diagnosed.
In the crude analyses, BV was associated with number of partners during the last month and lifetime partners. CT was associated with number of partners during the last 6 months and history of casual sexual contacts. Thus, multivariate analysis of BV with the CT comparison group adjusted for these sexual variables (as well as age) that were associated with BV and CT, respectively (Tables 1–4). In the multivariate analysis of the BV and the comparison group, adjustment was made for age, number of partners during the last month, and lifetime partners.
There was no significant difference in age at first intercourse between women with BV and the non-BV comparison group (Table 1). However, the BV group was more likely than the comparison group to report having known the partner for less than 1 month before debut and to have ended the relationship less than 1 month after debut. After adjustment for the previously mentioned sexual variables, the probability values decreased (P = 0.08 and 0.16, respectively). There were no significant differences between the BV and CT groups with respect to sexual debut factors.
There was a significant trend toward higher numbers of lifetime sexual partners in the BV group (Table 2). More than 10 lifetime sexual partners was reported in 30% of women with BV, compared with 15% in the comparison group. The mean number of lifetime sexual partners in women with BV was 16.6 and in the comparison group 10.0 (P < 0.0001). The BV group reported more sexual partners than those in the CT group (n = 10.9, P = 0.02), but this was a reflection of the BV group being older; after adjustment for age, the association was no longer statistically significant (P = 0.11).
Women with BV were more likely to report more than one sexual partner during the preceding month, but not the last 6 months, compared with the non-BV group (Table 3). The difference remained significant (P = 0.0003) after adjustment for number of partners during the last month and lifetime partners. Numerically, women with BV were more likely to initiate intercourse with the latest partner after less than 1 month, and reported more frequent experience of infidelity (for both the woman and her partner) than the non-BV group. However, none of these differences was statistically significant. There were also no such differences between the BV and the CT groups.
The BV group reported greater satisfaction with their sex life (intercourse frequency) than the comparison group: 87% versus 78% (P = 0.01; not shown in table). The difference between groups increased when adjustment for number of sexual partners was made (P = 0.005).
The reporting of sexual behaviors other than vaginal intercourse, which included history of oral or anal intercourse and lesbian sex, did not differ between the BV group and the non-BV group, with the exception of history of group sex (Table 4). More frequent history of group sex in the BV group remained significant after adjustment for lifetime sexual partners (P = 0.04), but this association vanished when number of sexual partners during the previous month was added to the multivariate model (P = 0.10). There were no significant differences in these variables when the BV and CT groups were compared. Use of sexual toys was equally distributed in the three groups.
Women with BV were more likely to report a history of sexual abuse and rape (in most cases, by previously known men) than the comparison group. A history of having been raped remained significantly associated with BV versus the non-BV comparison group after adjustment for number of lifetime sexual partners and partners during the last month (P = 0.02). There were no significant differences between the women with BV and CT either in the crude analysis or in the multivariate analyses (P = 0.20) with respect to history of abuse.
With the exception of ability to attain orgasm during petting, orgasm ability did not differ between the three groups. Orgasm with the last intercourse was reported by 76.4% of the BV group, 70.0% of the CT group, and 75.5% in the comparison group (not shown in the tables). In contrast, orgasm during petting was significantly more common in women with BV than in the comparison group (89.3% vs. 78.7%, P = 0.008). This association remained after adjustment for sexual partners (P = 0.007).
History of casual sex during travel was equally distributed between the three groups. Women in the CT group reported more frequent history of other types of casual sex than the two other groups (of borderline significance, Table 4). This difference was statistically significant after adjustment for number of sexual partners (P = 0.04 compared with the comparison group; P = 0.03 compared with the women with BV).
This study confirmed the results of several previous studies14–17 that women with BV had reported more sexual partners than control groups. A high number of lifetime sexual partners is considered to be a major indicator of high-risk sexual behavior and is probably the most studied marker in measuring risk of acquiring STDs. Our study also showed that there was no difference in number of sexual partners between women with BV and those with an established STD, in this case genital CT.
The study evaluated a number of other sexual behavior risk factors, most of which indicated that the women with BV do report high-risk sexual behaviors in common with women with STDs as generally defined. In fact, the sexual behaviors of women with BV and those with genital CT show great similarities in this study, a comparison that has not been made before in the literature.
Young age at first intercourse is considered one important marker of high-risk sexual behavior21; in our study, however, this was similar in all three groups. Amsel and colleagues2 found age at first intercourse to be significantly lower in women with BV than in control subjects (by 0.8 years), a finding that Larsson and colleagues15 did not confirm. In our study, the women with BV practiced sexual risk-taking early in their sexual life. They made their sexual debut after a short acquaintance, and they were more likely to report that the relationship ended within a month after the first intercourse. In this aspect, the BV group did not differ from the women with genital CT. These circumstances related to the sexual debut emerged as better indicators of high-risk sexual behavior than age at first intercourse. This supports our previous finding22 that negative circumstances at sexual debut can predict future sexual risk-taking behavior in women.
A higher frequency of anal sex could support the theory of the rectum being a reservoir for the microorganisms found in the vagina of women with BV,3 but the practice of anal sex and the frequency of anal intercourse during the previous month were similar in the three groups studied (unpublished results). However, practice of anal and oral sex was more frequently reported by women with BV in the study by Livengood et al.23
The strongest association in our study was between BV and more than one sexual partner during the last month, when using the non-BV group for comparison. This has also been reported in a study by Barbone and colleagues.14 This difference remained significant after adjustment for number of sexual partners during the last month and lifetime partners. On the other hand, Amsel and colleagues2 found no difference in the frequency of sexual abstinence during the last month between women with BV and control subjects. However, only 13 women in the latter study were sexually abstinent, and the actual length of the abstinence was not recorded.
Our results indicate that some women acquired BV during the previous month after the introduction of a new sexual partner. This supports the common clinical observation that BV symptoms often appear coincidentally with the initiation of a new sexual relationship.4
Another indication that partner change might be of importance in BV is the finding of more frequent history of group sex. This association was significant after adjustment for lifetime sexual partners, but diminished when number of partners during the last month was added to the analyses.
We have no explanation for the significant difference in orgasm ability with petting among BV and comparison groups. If the difference was due to sexual experience, the significance would be expected to decrease when adjusted for lifetime sexual partners. However, this was not the case. The association could be caused by chance, but it was notably strong.
We found a predominance of women with BV, compared with the non-BV group, who reported a history of sexual abuse or rape. In our study, these factors were also associated with the occurrence of other STDs apart from genital CT (unpublished results).
Few previously published results indicate sexual transmission of microorganisms occurring in the flora changes associated with BV. Gardner and Dukes1 inoculated 13 women with a pure culture of G. vaginalis, but only 1 of them became symptomatic. BV has been diagnosed in virgins.7 The extremely high frequency (over 50%) of BV in lesbians9 may suggest sexual transmission through other sexual behaviors (i.e., nonvaginal intercourse), which are prominent in this group of women. A high colonization rate of BV-associated bacteria is found in the sulcus of the glans penis of male partners of women with BV5; with the exception of M. hominis, these disappear after 2 weeks of condom use.22 The results of the Swedish Women's Health Study have recently shown that occurrence of M. hominis in vaginal fluid samples, regardless of concomitant BV, is associated with sexual activity risk factors.24 This might indicate that the occurrence of BV is related to sexual activity, because growth of M. hominis is a common finding in women with BV. However, partner treatment has not been shown to be effective in lowering recurrence rates.10–13
After measuring a large number of different aspects of sexual behavior, our overall conclusion is that women with BV have many sexual activity characteristics in common with women with genital CT, which, in this study, served as a model for an established STD. This is consistent with the clinical impression that the occurrence of BV often coincides with a change of sexual partner. One possible mechanism in the pathogenesis might be reinforcement of BV by certain men and women by some yet unknown exogenic factors present in semen or saliva (lesbians).
1. Gardner HL, Dukes CD. Haemophilus vaginalis
vaginitis. Am J Obstet Gynecol 1955; 69:962–976.
2. Amsel R, Totten PA, Spiegel CA, Chen KCS, Eschenbach D, Holmes KK. Nonspecific vaginitis: Diagnostic criteria and microbial and epidemiologic associations. Am J Med 1983; 74:14–22.
3. Holst E. Reservoir of four organisms associated with bacterial vaginosis suggests lack of sexual transmission. J Clin Microbiol 1990; 28:2035–2039.
4. Spiegel CA. Bacterial vaginosis. Clin Microbiol Rev 1991; 4:485–502.
5. Mårdh P-A. The vaginal ecosystem. Am J Obstet Gynecol 1991; 165:1163–1168.
6. Moi H. Prevalence of bacterial vaginosis and its association with genital infections, inflammation, and contraceptive methods in women attending sexually transmitted disease and primary health clinics. Int J STD AIDS 1990; 1:86–94.
7. Bump RC, Buesching WJ. Bacterial vaginosis in virginal and sexually active adolescent females: Evidence against exclusive sexual transmission. Am J Obstet Gynecol 1988; 158:935–939.
8. Gardner JJ. Comparison of the vaginal flora in sexually abused and non-abused girls. J Pediatr 1992; 120:872–877.
9. Varney P, McCaffrey M, Taylor-Robinson D, Evans B. Abnormal vaginal colonization in lesbian women. Presented at the Third International Symposium on Vaginitis/Vaginosis, Oxford Clinical, Oxford, United Kingdom, January 25–29, 1994.
10. Eschenbach DA, Critchlow CW, Watkins H, et al. A dose-duration study of metronidazole for the treatment of nonspecific vaginosis. Scand J Infect Dis Suppl 1983; 40:73–80.
11. Swedberg J, Steiner JF, Deiss F, Steiner S, Driggers DA. Comparison of single-dose vs. one-week course of metronidazole for symptomatic bacterial vaginosis. JAMA 1985; 254:1046–1049.
12. Vejtorp M, Bollerup AC, Vejtorp L, et al. Bacterial vaginosis: A double-blind randomized trial of the effect of treatment of the sexual partner. Br J Obstet Gynaecol 1988; 95:920–926.
13. Moi H, Erkkola R, Jerve et al. Should male consorts of women with bacterial vaginosis be treated? Genitourin Med 1989; 65:263–268.
14. Barbone F, Austin H, Louv WC, Alexander WJ. A follow-up study of methods of contraception, sexual activity, and rates of trichomoniasis, candidiasis, and bacterial vaginosis. Am J Obstet Gynecol 1990; 163:510–514.
15. Larsson P-G, Platz-Christensen J-J, Sundström E. Is bacterial vaginosis a sexually transmitted disease? Int J STD AIDS 1991; 2:362–436.
16. Hart G. Factors associated with trichomoniasis, candidiasis and bacterial vaginosis. Int J STD AIDS 1993; 4:21–25.
17. Hart G. Risk profiles and epidemiologic interrelationships of sexually transmitted diseases. Sex Transm Dis 1993; 20:126–136.
18. Hellberg D, Borendal N, Sikström B, Nilsson S, Mårdh P-A. Comparison of women with cervical human papilloma virus infection and genital warts: Some behavioral factors and clinical findings. Genitourin Med 1995; 71:88–91.
19. Ripa KT, Mårdh P-A. Cultivation of genital Chlamydia trachomatis
in cycloheximide-treated McCoy cells. J. Clin Microbiol 1977; 6:328–331.
20. SAS Institute, Inc. JMP User's Guide. Cary, NC: SAS Institute, Inc., 1994.
21. Greenberg J, Magder L, Aral S. Age at first coitus: A marker for risky sexual behavior in women. Sex Transm Dis 1992; 19:331–334.
22. Arvidson M, Hellberg D, Mårdh P-A. Sexual risk behavior in Swedish women with experience of casual travel sex abroad. Int J Gynaecol Obstet 1996; 2:73–76.
23. Livengood CH, Thomason JL, Hill GB. Bacterial vaginosis: Diagnostic and pathogenetic findings during topical clindamycin therapy. Am J Obstet Gynecol 1990; 163:515–520.
24. Elshibly S, Hellberg D, Nilsson S, Mårdh P-A. Sexual risk behavior in women with Mycoplasma hominis
found on vaginal culture. Br J Obstet Gynaecol 1996;103:1124–1128.