Background: Women with cervicitis frequently have bacterial vaginosis (BV). Prior studies have suggested that BV is involved in the pathogenesis of cervicitis.
Goals: To delineate the association of BV and cervicitis and to determine whether treatment of BV results in increased resolution of cervicitis.
Study Design: Women with clinically diagnosed cervicitis and BV received doxycycline and ofloxacin and were randomized to treatment with metronidazole gel or placebo. Resolution of BV and cervicitis was assessed.
Results: Metronidazole gel was associated with resolution of BV. Resolution of cervicitis was associated with use of metronidazole gel versus placebo (24/27 [88.9%] versus 15/24 [62.5%];P = 0.03). When further stratified by resolution of BV, those whose BV resolved were more likely to have resolution of cervicitis than those with persistent BV. Although these trends persisted in multivariate analyses, they did not achieve statistical significance.
Conclusion: Univariate analyses suggest an association between BV and cervicitis. An association between the use of metronidazole gel and resolution of cervicitis was demonstrated. Hypotheses regarding the latter include nonspecific antiinflammatory effects of metronidazole, its effect on BV flora, and its effect on a specific unrecognized pathogen.
There may be a specific component of the largely anaerobic microflora associated with bacterial vaginosis that is pathogenic in cervical tissue and is eradicated by metronidazole.
From the *Department of Medicine and †Biostatistics Center, University of Alabama at Birmingham, Birmingham, Alabama
The authors thank Moira Venglarik, CRNP, and Bari Cotton, RN, for their efforts in enrolling and observing patients, and Heather Cox for assistance with manuscript preparation.
This project was supported by the Sexually Transmitted Diseases Cooperative Research Center (NIH/NIAID U19 AI38514). 3M Pharmaceuticals provided the metronidazole gel and placebo.
Correspondence: Jane R. Schwebke, MD, 703 19th Street South, Zeigler Research Building, No. 239, Birmingham, AL 35294-0007. E-mail: firstname.lastname@example.org
Received for publication March 13, 2001,
revised June 7, 2001, and accepted June 14, 2001.
CERVICITIS, or inflammation of the uterine cervix, is a syndromic diagnosis derived from the observation of a mucopurulent discharge from the cervical os, as well as other signs of inflammation such as edema and easily induced endocervical bleeding. Its presence suggests either gonorrhea or chlamydial infection, both of which may cause this syndrome. However, gonorrhea and chlamydial infection are identified in only half of all such cases, and the causes of the remaining cases are unknown. 1,2
Bacterial vaginosis (BV), which is a disturbance in the vaginal flora, frequently coexists with cervicitis. 2–4 In one study, 50% of women attending a clinic for treatment of sexually transmitted diseases (STDs) who had cervicitis also had BV. 2 In a pilot study designed to evaluate the appropriate medical management of concurrent BV and cervicitis, an association between treatment of BV and resolution of cervical inflammation was noted. 5 Studies such as these have led to the hypothesis that BV may play a role in the occurrence of cervicitis. The current study seeks to further examine these relationships.
Patients were recruited from among attendees of the Jefferson County Department of Health STD Clinic (Birmingham) who presented for screening for STDs. The institutional review boards of the University of Alabama at Birmingham and the Jefferson County Department of Health approved the study. Women who met the clinical criteria for the diagnosis of BV and cervicitis were eligible for the study. Women were excluded from enrollment in the study if they were pregnant, had received antibiotic therapy within the past 2 weeks, or required therapy for trichomoniasis. After informed consent was obtained, a standardized questionnaire about medical, sexual, and behavioral history was administered by one of two research clinicians. A pelvic examination was performed and physical findings were recorded. Subjects were screened for STDs and vaginal infections. Diagnostic testing for the latter included determination of the vaginal pH with indicator strips (ColorpHast; EM Science, Gibbstown, NJ), a “whiff” test, wet-preparation microscopy, culture for Trichomonas vaginalis, and vaginal smear for Gram stain. After collection of the vaginal specimens, the endocervix was examined for signs of inflammation, and the presence or absence of mucopurulent discharge from the cervical os was specifically noted. Endocervical swab specimens were obtained for cultures for gonorrhea and chlamydial infection. A clinical diagnosis of BV was based on the criteria of Amsel et al 6 and required the presence of three of the following findings: homogenous vaginal discharge, pH > 4.5, positive whiff test, and clue cells. Cervicitis was defined as the presence of mucopurulent discharge from the endocervix. 1
All subjects received doxycycline (100 mg orally twice a day for 7 days) and ofloxacin (400 mg orally), per the protocol of the clinic and Centers for Disease Control and Prevention for treatment of cervicitis. 7 Subjects were randomized in a double-masked fashion to treatment with metronidazole intravaginal gel (once a day at bedtime for 5 days) 7 or placebo gel. The placebo gel has the same ingredients as the treatment gel, with the exception of metronidazole, and it has a higher pH (5.0–5.5, versus 3.6–4.2 for metronidazole gel). It contains methyl paraben, propyl paraben, propylene glycol, carbomer-934P, edetate disodium, sodium hydroxide, and water. The clinician instructed each subject about correct use of the vaginal gel. If a patient's culture for T vaginalis became positive, the patient was appropriately treated and withdrawn from the study. Patients were reexamined 2 and 4 weeks after enrollment, at which times the clinician repeated the pelvic examination and the diagnostic testing and carefully recorded the physical findings.
Vaginal smears were gram-stained and classified according to the criteria of Nugent et al. 8 This method is based on the quantity of three different bacterial morphotypes: large gram-positive rods, which represent lactobacilli; small gram-variable rods, which represent Gardnerella and Bacteroides/ Prevotella species; and curved rods, which represent Mobiluncus species. On the basis of these results the specimen is assigned a score from 0 to 10, with 0 to 3 as normal, 4 to 6 as intermediate, and 7 to 10 as BV. T vaginalis was sought by direct wet-mount examination as well as culture (InPouch TV test; BioMed Diagnostics, San Jose, CA). 9 Cultures for gonorrhea and chlamydial infection (on microtiter plates) were processed according to standard methods. 10,11
All women who were randomized to receive the study drug were included in an intent-to-treat analysis. The proportion of women whose BV resolved was compared between the metronidazole and placebo groups by means of the chi-square or Fisher exact test. The univariate association of clinical and sexual behavior variables was also evaluated between the treatment groups, with use of the chi-square or two-sample t test. Multivariate logistic regression models were used to determine the association between treatment and resolution of BV, with adjustment for other significant variables from the univariate analysis. Odds ratios and 95% confidence intervals were calculated for the variables in the multivariate logistic model. Similar analyses were performed to evaluate resolution of cervicitis.
Seventy-five women were enrolled in the study. Four patients did not meet eligibility criteria for the study (three because of positive screening cultures for Trichomonas and one because of pregnancy). Nine patients were completely lost to follow-up before visit 2, whereas one patient returned only for the final study visit, at week 4 (thus, 61 patients were available for the visit 2 analysis). Eleven additional patients were lost to follow-up between visits 2 and 3 (51 patients were available for the visit 3 analysis).
Table 1 shows the demographic, behavioral, and baseline microbiologic characteristics of the study population, stratified by treatment with metronidazole gel versus placebo. There was no significant difference between the two cohorts with regard to age, race, history of BV or STD, or baseline laboratory screening. Cultures for gonorrhea and chlamydial infection were positive at enrollment for 9 of 69 (7.2%) and 8 of 70 (11.4%), respectively, among the entire cohort. Oral contraceptive use was more frequent among women in the treatment group. A greater percentage of women in the placebo group reported douching, and the placebo recipients also had a significantly higher mean number of sex partners within the past month and the past year than did the treatment recipients.
Resolution of BV by treatment arm is shown in Table 2. Treatment with metronidazole gel was highly associated with resolution of BV at both follow-up visits. At the second follow-up, 19 (70.4%) of 27 women treated with metronidazole were cured, versus 5 (20.8%) of 24 in the placebo group. The mean Nugent Gram-stain scores were significantly lower in the metronidazole group than in the placebo group (3.8 versus 7.1;P = 0.003).
The univariate analysis of factors associated with persistence of BV, other than drug therapy, is presented in Table 3. At the first follow-up visit, persistence of BV was significantly associated with a greater number of sex partners in the 30 days before study enrollment, as well as a self-reported history of gonorrhea. At the second follow-up visit, persistence of BV was again significantly associated with the number of sex partners in the 30 days before enrollment in the study, as well as a history of douching. With regard to the latter, women who reported more frequent douching (i.e., 2 to 4 times per month) were more commonly among those women with persistent BV. Also noted was a trend toward a significant association of persistent BV with continued cervical inflammation. Among those women with persistent BV, 9 (33.3%) of 27 had cervicitis, versus 3 (12.5%) of the 24 women whose BV resolved (P = 0.08).
Multivariate analysis revealed that having received metronidazole gel instead of placebo was highly associated with cure of BV at both follow-up visits (odds ratio [OR] 5.6, 95% CI 1.7–18.8, P = 0.005 for first follow-up; OR 7.5, 95% CI 1.7–32.5, P = 0.007 for second follow-up). Women without a history of gonorrhea were also significantly more likely to have resolution of BV at the first follow-up visit (OR, 3.6; 95% CI, 1.1–11.7;P = 0.03).
Resolution of the mucopurulent discharge from the cervical os (cervicitis), stratified by receipt of metronidazole gel versus placebo, is shown in Table 4. In analysis of all subjects at the second follow-up visit, resolution of cervicitis was more frequent among those women who received metronidazole gel than those who received placebo (24/27 [88.9%] versus 15/24 [62.5%];P = 0.03). Removal of the patients positive for gonorrhea or chlamydial infection from the analysis did not appreciably alter these results. When stratified by resolution of BV, all 20 women whose BV resolved who were treated with metronidazole gel also had resolution of cervicitis noted at the first follow-up visit, as compared with 4 of 7 (57%) who had spontaneous resolution of BV (placebo group) (P = 0.01). Among those women with continued evidence of BV at follow-up visits, there was no significant difference between the metronidazole and placebo groups. Among those women treated with metronidazole gel, those whose BV resolved were significantly more likely to have cervicitis resolve than were those with persistent BV (20/20 [100%] versus 9/12 [75%];P = 0.04). Univariate analysis of factors other than drug therapy associated with resolution or persistence of cervicitis failed to show any significant associations. Multivariate analysis showed that women who received metronidazole were three times as likely to be cured of cervicitis (OR, 3.4;P = 0.13) and that women who had persistent BV at follow-up were less likely to be cured of cervicitis (OR, 0.5;P = 0.4). However, these associations did not reach statistical significance.
Cervical inflammation, characterized by mucopurulent discharge from the cervical os at the time of physical examination, is a condition frequently affecting young, sexually active women. 1,2 Although some cases are linked to chlamydial infection or gonorrhea, the causes of at least half are unknown. 2 Nonetheless, treatment with doxycycline has been shown to be effective in many cases. 2,12 Mucopurulent cervicitis has many similarities with nongonococcal urethritis in males; for example, in many cases the cause is unknown, yet they clinically respond to antibiotic therapy. 1
Bacterial vaginosis is a disturbance of the vaginal flora characterized by a decrease in the concentration of hydrogen peroxide–producing lactobacilli. The cause of BV is unknown, although its epidemiology is that of a sexual activity–associated infection. 3,13,14 The apparent lack of a host inflammatory response to BV, as suggested by the absence of polymorphonuclear leukocytes in the vaginal fluid of women with BV, led to use of the term vaginosis, as opposed to vaginitis. However, BV may be associated with less visible inflammatory markers such as lactoferrins or various cytokines. 10,15 Bacterial vaginosis has been associated with many STDs, including gonorrhea, chlamydial infection, trichomoniasis, pelvic inflammatory disease, HIV disease, and cervicitis. 16–21 The leading hypothesis concerning these associations is that the absence of protective lactobacilli increases the biologic risk for acquiring an STD upon exposure. 22
The association of BV with cervicitis has been previously noted. Paavonen et al 2 found that 50% of women attending an STD clinic who had mucopurulent cervicitis also had BV. The authors further hypothesized that cervical inflammation might predispose to altered vaginal flora. In a pilot study examining appropriate treatment for patients with BV and cervicitis, it was found that specific therapy for BV (metronidazole), along with standard antimicrobial therapy for cervicitis, was necessary for cure of BV. It was further observed that resolution of BV was associated with the resolution of cervical inflammation, but the numbers were small and the study was not randomized. 5
In the current double-blind randomized study of metronidazole gel versus placebo in addition to doxycycline/ofloxacin for patients with BV and cervicitis, therapy with metronidazole was highly associated with cure of BV. Cure rates for BV in this study (70% at 4 weeks) were similar to those reported in the literature. 23 In approximately 20% of women who received placebo, BV also resolved, but other studies have also shown spontaneous cures. 20 In univariate analyses, behavioral and historical factors significantly associated with persistence of BV included number of sex partners, history of gonorrhea, and douching, variables that have been previously reported to be associated with BV. 3,6,13,24 Other factors that were associated with BV—although not significantly so—were a history of trichomoniasis and cervical inflammation. It is of interest that these are variables frequently associated with women who are at higher risk for STDs. It is not known whether the finding of BV at follow-up visits in this study represented failure of metronidazole to effect a cure, or relapse or reinfection.
With regard to the associations between BV and cervical inflammation, this study is supportive of data from pilot investigations. 5 Among women whose BV resolved and who received metronidazole, 100% had cervical inflammation resolve, whereas cervical inflammation resolved in 75% of metronidazole recipients whose BV did not resolve. This finding suggests that BV may be linked to cervical inflammation, at least in some patients. This study has also provided new data regarding treatment of cervical inflammation with metronidazole gel. Among all subjects, those randomized to treatment with metronidazole gel along with doxycycline/ofloxacin were significantly more likely to have cervicitis resolve by 3 weeks post-therapy (follow-up visit 2) than those women randomized to receive the placebo gel. Although similar trends were seen in the multivariate analysis, statistical significance did not persist, perhaps because of an insufficient number of observations.
The potential effects of metronidazole on cervical inflammation could be due to nonspecific antiinflammatory properties of the antibiotic or specific interactions with BV or an unrecognized pathogen. Although antiinflammatory properties of metronidazole have not been demonstrated in vitro, 25 this antibiotic is widely used as adjunctive therapy along with oral antiinflammatory agents for Crohn's disease. 26 Alternatively, there may be a specific component of the largely anaerobic microflora associated with BV that is pathogenic in cervical tissue and is eradicated by metronidazole.
Of interest in addition to the BV/cervicitis data are those concerning factors other than drug therapy that were associated with persistence of BV, an infection whose cause is unclear. Persistence of BV was associated with a greater number of sex partners in the prior 30 days and a self-reported history of gonorrhea and douching. These factors have been previously shown to be associated with BV and now appear to be risk factors for treatment failure or perhaps reinfection.
The major limitation of the current study relates to statistical power. Although the study was originally statistically powered to answer the question of whether successful treatment of BV would increase cure rates for cervicitis, particularly cases not caused by gonorrhea or chlamydial infection, when these new data are considered the power is only 55%. Thus, a statistically significant association between BV and cervicitis could not be demonstrated. The significant association of treatment with metronidazole gel and resolution of cervicitis that was found in univariate analysis was unanticipated. The current statistical power of the study to address this is only 22%. A larger study will be needed to answer these questions fully.
1. Brunham RC, Paavonen J, Stevens CE, et al. Mucopurulent cervicitis—the ignored counterpart in women of urethritis in men. N Engl J Med 1984; 311: 1–6.
2. Paavonen J, Critchlow CW, DeRouen T, et al. Etiology of cervical inflammation. Am J Obstet Gynecol 1986; 154: 556–564.
3. 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.
4. Paavonen J, Roberts PL, Stevens CE, et al. Randomized treatment of mucopurulent cervicitis with doxycycline or amoxicillin. Am J Obstet Gynecol 1989; 161: 128–135.
5. Schwebke JR, Schulien MB, Zajackowski M. Pilot study to evaluate the appropriate management of patients with coexistent bacterial vaginosis and cervicitis. Infect Dis Obstet Gynecol 1995; 3: 199–122.
6. Amsel R, Totten PA, Spiegel CA, Chen KCS, Eschenbach D, Holmes KK. Non-specific vaginitis: diagnostic and microbial and epidemiological associations. Am J Med 1983; 74: 14–22.
7. Centers for Disease Control and Prevention. Guidelines for treatment of sexually transmitted diseases. MMWR Morb Mortal Wkly Rep 1998; 47. (RR-1)
8. 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: 297–301.
9. Draper D, Parker R, Patterson E, et al. Detection of Trichomonas vaginalis
in pregnant women with the InPouch TV system. J Clin Microbiol 1993; 31: 1016–1018.
10. Reichart C, Gaydos C, Brady W, Quinn T, Hook E III. Evaluation of Abbott Testpack Chlamydia for detection of Chlamydia trachomatis
in patients attending sexually transmitted disease clinics. Sex Transm Dis 1990; 17: 147–151.
11. Smeltzer M, Curran J, Brown S, Pass J. Accuracy of presumptive criteria for culture diagnosis of Neisseria gonorrhoeae
in low-prevalence populations of women. J Clin Microbiol 1980; 11: 485–487.
12. Romanowski B, Talbot H, Stadnyk M, Kowalchuk P, Bowie W. Minocycline compared with doxycycline in the treatment of nongonococcal urethritis and mucopurulent cervicitis. Ann Intern Med 1993; 119: 16–22.
13. 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.
14. Lossick JG. The descriptive epidemiology of vaginal trichomoniasis and bacterial vaginosis. In: Horowitz BJ, Mardh PA, eds. Vaginitis and Vaginosis. New York: Wiley-Liss, 1991: 77–84.
15. Sturm-Ramirez K, Gaye-Diallo A, Eisen G, Mboup S, Kanki P. High levels of tumor necrosis factor and interleukin-1 in bacterial vagi- nosis may increase susceptibility to human immunodeficiency virus. J Infect Dis 2000; 182: 467–73.
16. Thomason JL, Gelbart SM, Sobun JF, Schulien MB, Hamilton PR. Comparison of four methods to detect Trichomonas vaginalis
. J Clin Microbiol 1988; 26: 1869–1870.
17. Cohen CR, Duerr A, Pruithithada N, et al. Bacterial vaginosis and HIV seroprevalence among female commercial sex workers in Chiang Mai, Thailand. AIDS 1995; 9: 1093–1097.
18. Taha TE, Hoover DR, Dallabetta GA, et al. Bacterial vaginosis and disturbances of vaginal flora: association with increased acquisition of HIV. AIDS 1998; 12: 1699–1706.
19. Westrom L, Mardh PA. Acute pelvic inflammatory disease (PID). Sex Transm Dis 1990; 17: 593–613.
20. Hillier SL, Krohn MA, Nugent RP, Gibbs RS. Characteristics of three vaginal flora patterns assessed by gram stain among pregnant women. Am J Obstet Gynecol 1992; 166: 938–944.
21. Martin HL Jr, Richardson B, Nyange P, et al. Vaginal lactobacilli, microbial flora, and risk of human immunodeficiency virus type 1 and sexually transmitted disease acquisition. J Infect Dis 1999; 180: 1863–1868.
22. Hillier S. The vaginal microbial ecosystem and resistance to HIV. AIDS 1998; 14: 17–21.
23. Joesoef MR, Schmid GP, Hillier SL. Bacterial vaginosis: review of treatment options and potential clinical indications for therapy. Clin Infect Dis 1999; 28: S57–S65.
24. Hawes S, Hillier S, Benedetti J, et al. Hydrogen peroxide–producing lactobacilli and acquisition of vaginal infections. J Infect Dis 1996; 174: 1058–1063.
25. Ganrot-Norlin K, Stalhandske T, Karlstrom R. Lack of antiinflammatory activity of metronidazole. Acta Pharmacol Toxicol 1981; 49: 130–133.
26. Pallone F, Boirivant M, Fais S, et al. Antibacterial drugs in Crohn's disease. Ital J Gastroenterol 1992; 24: 17–18.