Klebanoff, Mark A. MD, MPH*; Schwebke, Jane R. MD†; Zhang, Jun MD, PhD†; Nansel, Tonja R. PhD*; Yu, Kai-Fun PhD*; Andrews, William W. PhD, MD‡
Bacterial vaginosis is a condition in which the normal vaginal flora changes from a predominance of hydrogen peroxide–producing Lactobacillus species to high concentrations of a variety of anaerobic organisms, Gardnerella vaginalis and Mycoplasma hominis.1 Typical symptoms are said to include vaginal malodor and a thin discharge.2 However, although bacterial vaginosis is the most prevalent cause of vaginal discharge in reproductive-aged women,3 a substantial but highly variable number of women with bacterial vaginosis are alleged to be asymptomatic.4 The fraction of women with bacterial vaginosis who report symptoms has ranged from 10%5 to 66%.6 In most of these studies, the fraction of women without bacterial vaginosis who also reported symptoms has been highly variable as well.
Whether a woman with bacterial vaginosis reports symptoms may depend on the vigor with which she is questioned. Amsel et al2 noted that 25% of symptomatic women presented to clinics for reasons other than suspicion of infection and only admitted to symptoms upon direct questioning. Possible reasons why women with bacterial vaginosis might fail to report symptoms include the reason for the clinic visit (sexually transmitted disease [STD] versus routine care versus research study), embarrassment about discussing these symptoms, or failure to recognize mild degrees of discharge or malodor as abnormal because they are very common in the population. The purpose of this report is to describe the prevalence of lower genital tract symptoms, as assessed by direct, detailed, and explicit questioning of the study subjects, among women with and without bacterial vaginosis defined by both Gram stain and clinical criteria.
MATERIALS AND METHODS
The women in this study were enrolled in the Longitudinal Study of Vaginal Flora. The purpose of the Flora study was to evaluate the natural history of bacterial vaginosis as well as to describe factors associated with the acquisition and loss of this condition in a cohort of women monitored during quarterly visits for a year. Enrollment occurred from August 1999 to February 2002. Nonpregnant women aged 15–44 years were eligible for the study when they presented for a routine health care visit, not specifically because of suspicion of infection, at 1 of 12 clinics in the Birmingham, Alabama, area. Medical/gynecological conditions that caused a woman to be ineligible included immunocompromised status (congenital, acquired, or secondary to medication), postmenopausal, posthysterectomy, postpelvic radiotherapy, receiving chronic (daily for at least 30 days) antibiotics, or participation in a clinical trial in which antibiotics or topical genital microbicides were administered in a blinded manner. Women expected to develop 1 of these conditions during the next 12 months were also ineligible, although women who developed 1 of the conditions unexpectedly continued to be observed. Additional exclusions were if the subject was planning to move from the area in the next 12 months; was nonfluent in English; or had affective disorder, psychosis, or emotional/intellectual limitations that precluded informed consent. The protocol was approved by the Institutional Review Boards of the Jefferson County Department of Health, the University of Alabama at Birmingham, and the National Institute of Child Health and Human Development. All participating women provided written informed consent.
At the initial study visit, women underwent a standardized pelvic examination and assessment of clinical symptoms. Particular attention was given to the amount and characteristics of the vaginal discharge. Examinations were performed in a single research clinic by research nurses who underwent training sessions to achieve standardized results on the examination and the wet prep. Vaginal samples were taken for pH; wet prep for “clue cells,” yeast, and trichomonads; the potassium hydroxide “whiff test”; cultures for Trichomonas vaginalis (InPouch; BioMed Diagnostics, San Jose, CA) and Lactobacillus species; Gram stain for the evaluation of the flora according to the Nugent criteria;7 and gas-liquid chromatography (sterile deionized water vaginal wash). Cervical samples were obtained for Neisseria gonorrhoeae culture and Chlamydia trachomatis ligase chain reaction, and serum was obtained for a Venereal Disease Research Laboratories test. The women also completed a detailed in-person interview given by specifically trained female interviewers that covered demographic factors, obstetric and gynecologic history, dental symptoms and practices, feminine hygienic and health behaviors, sexual history and practices, history of genital tract infections or STDs, alcohol and drug use, and psychosocial status. In addition, a dental hygienist conducted a standardized examination that was focused on periodontal disease; women with significant disease were given an appropriate referral. These data will be reported in future publications. The interviewer was blinded to the results of the physical examination and vice versa; Gram stains were evaluated blind to clinical data.
The purpose of this analysis was to describe the presence of symptoms reported among women with and without bacterial vaginosis who received specific, detailed questioning from our researchers. To develop the questionnaire, we conducted a series of focus groups among women similar to those intended for enrollment in the study.8 At the initial interview, the women were asked “In the past 6 months, have you had (symptom)?” The symptoms included vaginal wetness, vaginal discharge, vaginal odor, vaginal itch that does not easily go away, vaginal irritation, pain when you urinate, and abdominal or pelvic pain. A distinction was drawn between wetness and discharge because the focus group study revealed that “wetness” is considered normal whereas “discharge” may carry a connotation of disease in this population.8 For each symptom reported, the woman was asked “How often does it occur?” with choices of once a month or less, 2 to 3 times a month, about once a week, or several times a week or more. The woman also was asked, “How bad was it?” with choices of “mild, that is, you hardly notice it,” “moderate, that is, you notice it but it doesn't bother you a lot,” and “severe, that is, it really bothers you.” To investigate the possibility that women had genital odor but didn't recognize it as abnormal, all women were then asked “My usual vaginal odor can best be described as …,” and the interviewer coded the woman's spontaneous response.
If a woman had been sexually active in the past 6 months, she was also asked about the presence and frequency of the 7 symptoms during or immediately after sexual intercourse. Finally, women who had been sexually active in the past 6 months were asked “In the past 6 months, have you noticed a ‘fishy’ smell after sex?” with choices of always, most of the time, about half the time, seldom, or never.
Bacterial vaginosis was alternatively defined as a Nugent Gram stain score of 7 or greater7 or as a clinical diagnosis based on at least 3 of the 4 Amsel criteria (vaginal pH higher than 4.5; thin, homogeneous vaginal discharge; “whiff test”; and more than 20% of all vaginal epithelial cells examined on a standardized wet prep were “clue cells”).2 The Nugent Gram stain score evaluates the number of Lactobacillus (large, gram-positive rods), Gardnerella (gram-negative coccobacillary organisms), and Mobiluncus (thin, curved gram-variable rods) morphotypes per oil-immersion microscope field. By convention, a score of 0–3 represents normal flora, 4–6 represents intermediate flora, and 7–10 represents bacterial vaginosis flora. To eliminate potential confounding by the presence of STDs, women with gonorrhea, Chlamydia, or trichomonas, on the basis of the laboratory tests noted above, were excluded from this analysis. Differences in continuous variables were assessed by the Student t test and in categorical variables by the χ2 test. When the categories were ordered (eg, no, mild, moderate, or severe vaginal odor), the Cochran–Armitage test for trend9 was used to assess statistical significance. In all cases a 2-tailed P value of < .05 was considered statistically significant.
The cohort enrolled 3,620 women, 3,614 of whom completed the baseline interview. Exclusion of women with gonorrhea (n = 86), Chlamydia (n = 356), or trichomonas (n = 265) reduced the number for this analysis to 2,907; Gram stain results were available for 2,888, 1,063 (or 37%; 95% confidence interval [CI] 35–39%) of whom had bacterial vaginosis. Basic demographic characteristics of the women and the prevalence of bacterial vaginosis by Gram stain are presented in Table 1. The women were predominantly African American, young, parous, unmarried, and of low income. Bacterial vaginosis was more common in African-American women and women who smoked. It was less common among women who were currently married and among women with more than 12 years of education, higher income, and younger (aged less than 20 years) age. Worsening general health status was associated with increasing prevalence of bacterial vaginosis.
The distributions of symptoms are shown for women with and without Gram stain scores of 7 or greater in Figure 1. As a result of women not knowing or refusing to answer individual questions, the numbers for the symptoms in the figure ranged from 2,836 to 2,881. Abdominal or pelvic pain was very rare, did not differ by bacterial vaginosis status, and therefore is not presented. Reported vaginal discharge did not differ between women with and without bacterial vaginosis, although vaginal wetness was significantly more frequent when bacterial vaginosis was present. Odor was highly significantly more frequent among women with bacterial vaginosis. Interestingly, dysuria and vaginal irritation were significantly less frequent among women with bacterial vaginosis. Although associations with wetness, odor, irritation, and dysuria were statistically significant, the actual magnitude of these associations was not great. For example, 75% (95% CI 72–77%) of women with and 82% (95% CI 80–83%) of women without bacterial vaginosis reported never noticing a vaginal odor; differences in other symptoms were even smaller. Cumulatively, 58% of women with and 57% of women without bacterial vaginosis reported wetness, discharge, and/or odor in the past 6 months (P = .70). Among those who reported a particular symptom, itching was significantly less severe (P = .02) and irritation was marginally less severe (P = .09) in women with bacterial vaginosis. Severity of the remaining symptoms was similar between women with and without bacterial vaginosis (data not shown).
Among sexually active women, only any odor and fishy odor during or immediately after sexual intercourse were significantly different between women with and without bacterial vaginosis. However, the magnitude of the difference was small—81% of women with and 86% of women without bacterial vaginosis never noticed any odor, and 82% and 87%, respectively, never noticed a fishy odor during or immediately after intercourse. When women with Gram stain scores of 4–6 (“intermediate” flora) were excluded, none of the results were substantially changed.
The corresponding results for bacterial vaginosis defined as at least 3 of the 4 Amsel criteria were generally similar to those for bacterial vaginosis defined by Gram stain (data not shown). Vaginal wetness, discharge, and odor occurred significantly more frequently among women with than among women without bacterial vaginosis by Amsel criteria, although the actual magnitude of the differences was small. There was no significant difference among the women regarding irritation, itching, or dysuria by clinical bacterial vaginosis status. Among those reporting a symptom, the discharge was marginally more severe (P = .09) among women with bacterial vaginosis, but none of the other severity scores even approached significance (data not shown).
The distribution of reported “usual vaginal odor” is shown in Table 2. Virtually all of the various descriptions of odor were more common among women with bacterial vaginosis, although there was no predominance of any particular description. Of note, although “fishy” was reported more frequently among women with bacterial vaginosis, only 6% of women with and 4% of women without bacterial vaginosis reported this as their usual odor. Although the difference in distribution of reported odor between women with and without bacterial vaginosis was statistically significant, it was of small magnitude: 35% of women with bacterial vaginosis and 45% of women without bacterial vaginosis reported no vaginal odor of any kind.
Table 3 shows the prevalence of bacterial vaginosis according to combinations of symptoms. For any given combination of vaginal discharge and odor, the addition of irritation reduced the prevalence of bacterial vaginosis. Discharge alone was not associated with an increased prevalence of bacterial vaginosis compared with women without symptoms. Vaginal odor was the only symptom associated with even a moderate increase in the prevalence of bacterial vaginosis, and the addition of discharge to odor did not increase the prevalence of bacterial vaginosis further. Nevertheless, bacterial vaginosis was still prevalent among women with no symptoms (37.1%), among women whose only symptom was vaginal irritation (26.3%), or among women who had both irritation and discharge (19.4%). Results were generally similar when reported vaginal “wetness” was substituted for “discharge.”
Reported symptoms correlated significantly, but not strongly, with the corresponding physical sign. When the woman reported no discharge, the nurse noted a thin, homogeneous discharge 35% of the time compared with 49% when the woman noted severe discharge. The “whiff” test was positive among 36% of woman who reported no odor versus 48% of women reporting severe odor. Frequency of positive signs for women reporting “mild” and “moderate” symptom severity were generally intermediate between those for women reporting “severe” and no symptoms (P for trend = .005 for discharge, and < .001 for odor). Associations between frequency of symptoms and physical signs were of similar magnitude to those for severity of symptoms (data not shown).
Candida was found on wet prep in 5.2% of women. It was less common when bacterial vaginosis was present (3.1%) than when bacterial vaginosis was absent (6.4%). However, the associations between bacterial vaginosis and symptoms did not change substantially by excluding women with Candida.
This study found that vaginal wetness/discharge and odor were more common, and irritation less common, in women with bacterial vaginosis than in women without bacterial vaginosis. Although statistically significant, the actual difference in prevalence of symptoms between women with and without bacterial vaginosis was small. In addition, although having any symptom was common among all women, the vast majority of women with bacterial vaginosis denied individual symptoms, even upon direct, detailed, and explicit questioning. There also were no substantial differences in the description of their “usual” vaginal odor between women with and without bacterial vaginosis. Thus, the failure to report symptoms does not appear to be the result of women with bacterial vaginosis considering some degree of vaginal malodor to be “normal.” We inquired about symptoms during the prior 6 months, and it is possible that the correlation between symptoms and bacterial vaginosis would have been stronger had we asked about a shorter time period, such as the prior month.
Our results are similar to those recently reported among unselected female U.S. Marine Corps recruits,10 among whom 24% with a Gram stain diagnosis of bacterial vaginosis reported vaginal discharge versus 18% of women without bacterial vaginosis. The corresponding percentages for vaginal odor were 10% and 5%, respectively. This reinforces our conclusion that although vaginal discharge and odor are reported more frequently by women with than by women without bacterial vaginosis, the differences are not great, and the majority of women with bacterial vaginosis do not report these symptoms even upon direct questioning.
Our findings are in contrast to those of Amsel et al,2 who reported that the prevalence of symptoms among women with bacterial vaginosis increased substantially upon direct, specific questioning. Amsel also reported that 46% of women with bacterial vaginosis diagnosed by clinical criteria reported symptoms. However, Amsel studied a population of women presenting to a student health clinic for a variety of reasons, including symptoms of vaginitis, and in this population 32% of women without bacterial vaginosis also reported symptoms. Nevertheless, it would appear that the nature of the questions (general versus specific) and the manner in which they are asked (blind to clinical findings or not) could have a substantial impact on the reported prevalence of symptoms.
We also found that although “vaginal wetness” was similarly associated with bacterial vaginosis as diagnosed by either Gram stain or clinical (Amsel) criteria, vaginal discharge was only associated with bacterial vaginosis diagnosed by Amsel criteria. Because Gram stain is a more sensitive method to diagnosing bacterial vaginosis than are the Amsel criteria, it is possible that only the more severe cases of bacterial vaginosis fulfill at least 3 Amsel criteria. This may account for the difference we observed in these associations.
“Inflammatory” symptoms—irritation, itching, and dysuria—were less prevalent among women with bacterial vaginosis; differences in irritation and dysuria were statistically significant. When itching was present, it was less severe among women with bacterial vaginosis. Candida vulvovaginitis is generally considered to feature inflammatory symptoms prominently.11 We found, as have others,12 that women with bacterial vaginosis are less likely than women without bacterial vaginosis to be colonized with Candida species. It is therefore possible that differences in Candida colonization account for the reduced occurrence of inflammatory symptoms among women with bacterial vaginosis. However, when women with Candida noted on wet prep were excluded, the reduced prevalence of inflammatory symptoms among women with bacterial vaginosis remained. Microscopic examination of vaginal fluid is a less sensitive method to detect Candida colonization than is fungal culture. It is therefore possible that, had we performed fungal cultures in this study, we might have found that differences in Candida colonization among women with and without bacterial vaginosis account for the difference we observed in inflammatory symptoms.
In summary, we found that although vaginal discharge and odor are more prevalent among women with compared with women without bacterial vaginosis, the difference in prevalence of these symptoms is not of a great magnitude. The vast majority of women do not report any vaginal discharge, wetness, or odor even upon explicit questioning. Future research should determine factors among women with bacterial vaginosis that result in the reporting of symptoms.