Genitourin Med 1997;73:23-28 23 What is normal vaginal flora? Original J C F Priestley, B M Jones, J Dhar, Linda Goodwin Genitourin Med: first published as 10.1136/sti.73.1.23 on 1 February 1997. Downloaded from Article Objective: To observe the composition of the vaginal flora of healthy women over time, and in relation to hormonal changes, sexual activity, and hygiene habits. Design: A longitudinal surveillance of the vaginal flora over an eight week period. Subjects: 26 female health care workers in local genitourinary medicine clinics. Methods: The participants were anonymised. They filled in diary cards daily. Blind vaginal swabs were self-taken two-seven times weekly. A smear was air-dried for later Gram staining. The swabs were also cultured for Candida spp, Gardnerella vaginalis, anaerobes, Mycoplasma hominis and Ureaplasma urealyticum. Results: Of 26 subjects, only four had normal vaginal microbiology throughout. One woman, who was not sexually active, had bacterial vaginosis (BV) throughout and nine (35%) had inter- mittent BV. Candidiasis was found intermittently in eight women (31%), and eight had normal microscopy. U urealyticum was isolated intermittently in 40% of women with BV, 25% with can- dida, and 50% with normal microscopy. Many women were symptomatic, but symptoms corre- lated poorly with microbiological findings. All but two women were sexually active; however, more women with BV were exposed to semen. BV seemed to be related to frequent use of scented soap, and there appeared to be an additive effect of clothing and hygiene factors. Conclusions: Our study raises doubts about what should be regarded as normal vaginal flora. It calls into question the significance of finding BV or U urealyticum on a single occasion in asymp- tomatic women, or of finding normal flora in symptomatic women. The effect of external factors on the vaginal flora deserve further study. (Genitourin Med 1997;73:23-28) Keywords: vaginal flora; bacterial vaginosis; candidiasis; ureaplasma Introduction Methods Bacterial vaginosis was first reported in 1955 Volunteer women http://sti.bmj.com/ by Gardner and Dukes,' who described the The study took place over an eight week unique clinical signs and symptoms and the period to cover two menstrual cycles. Subjects distinctive nature of the vaginal discharge were recruited from health care staff of geni- associated with it. They also described a tourinary medicine departments in Sheffield, "new") causative organism, which they named Rotherham, and Barnsley. They were "Haemophilis vaginalis", subsequently anonymised to encourage participation and on September 30, 2021 by guest. Protected copyright. renamed Gardnerella vaginalis. increase the accuracy of recorded personal Forty years on, while the complexity of the data. Application was made to the local Ethics microbiological picture has been well Committee; however it was felt that formal described,23 the aetiology and pathogenesis of approval for the study was not required as par- the condition remain poorly understood, and ticipation was both voluntary and anonymous. the reasons as to why the normal, healthy lacto- The participants randomly chose numbered bacillus flora is superseded by an overwhelming packs containing swabs, slides and Amies "cocktail" of bacteria are still not known. transport media, and were asked to complete Department of Fortunately, laboratory diagnosis can be readily diary forms to record age, contraceptive Genitourinary Medicine, The Royal made by simple, Gram stain microscopy method used, menstruation, tampon use, Hallamshire Hospital, where "clue cells" and the typical, abnormal coitus, apparel (tights and trousers), bath Sheffield bacterial flora of BV are easily recognised.4 additives, and any symptoms of discharge, irri- C J F Priestley Dhar The experience of some workers suggests tation, or odour. The diary forms were J that factors such as hormonal status, coitus, retained separately until the end of the study. Department of Medical Microbiology, and exposure to semen may play an important The women took blind vaginal swabs from University of Sheffield role in the pathogenesis of BV. We were also themselves two-seven times per week, prior to Medical School interested in the role of clothing and hygiene bathing. They prepared air-dried smears, for B M Jones L Goodwin habits such as the use of certain bath additives, subsequent Gram staining, before placing the Address correspondence to: which are said to precipitate vulvovaginal can- swab into the transport medium for culture. Dr C J F Priestley, didiasis in some women. As there are few doc- These swabs were left at room temperature Department of Genitourinary Medicine, Weymouth and umented studies on these aspects, we designed until the participant could deliver them to the District Hospital, Melcombe a longitudinal study to observe the composi- clinic, and swabs taken on a Friday or Avenue, Weymouth, Dorset DT4 7TB, UK. tion of the vaginal flora in healthy women, in Saturday, in particular, could be left for two to Accepted for publication relation to hormonal changes and external three days before being plated out. For this 8 November 1996 factors. reason we placed greater emphasis on the 24 24Priestley, Jones, Dhar, Goodwin Gram stained slides than cultures, in particu- further three only completed six or seven lar for the diagnosis of BV. The samples, iden- weeks of the study, but we felt that sufficient tified only by number, were tested "blind" in data were obtained to include them in the the laboratory. We also asked the participants analysis. Overall, the 26 women included in to measure the pH of the secretions, using the analysis took a mean of 3.2 swabs per week Genitourin Med: first published as 10.1136/sti.73.1.23 on 1 February 1997. Downloaded from standard paper strips, prior to preparing the (range 1.7-7.0) for a six to eight week period. slides; however, we became concerned about the accuracy of a reading from a blind vaginal Microbiology swab because of potential contamination from One woman had BV throughout the period of the vulva or cervical mucosa, and pH was not study. G vaginalis, anaerobes, Ureaplasma ure- measured by all the participants. alyticum and Mycoplasma hominis were isolated from the majority of her swabs. She com- Microbiological tests plained of some irritation but not of vaginal Microscopy Gram stained smears were exam- discharge or odour, the typical symptoms of ined for the presence of lactobacilli, "clue BV. BV was found intermittently in nine cells", "gram-variable" organisms, curved rods women (34-6%) in between 10 and 62% and yeasts. The results were recorded as "nor- (mean 27.5%) of swabs taken. G vaginalis was mal", where the flora comprised lactobacilli isolated from 64% and anaerobes from 28% of only; "normal, but with yeasts"; "intermedi- swabs showing BV. ate", with an abnormal BV-type flora but Candida spp were found intermittently in some lactobacilli still present; or "bacterial eight women (308%) in between 14 and 70% vaginosis", where the smears showed "clue (mean 35 5%) of swabs taken. Only eight cells" and masses of gram-variable bacteria, women (30.8%) had normal microscopy and a complete absence of lactobacilli.4 throughout. Ureaplasma urealyticum was isolated inter- Cultures The swabs were cultured for G vagi- mittently in 10 (38.5%) women; four (40%) of nalis, anaerobic organisms, Mycoplasma those with BV, two (25%) of those with can- hominis, Ureaplasma urealyticum, and yeasts. didiasis, and four (50%) of those with normal Plates of Columbia agar (Unipath Ltd, microscopy. Table 1 shows the characteristics Basingstoke, UK) containing 7% horse blood of these women. The presence of U ure- were inoculated and incubated, at 37°C in an alyticum did not appear to be related to sexual atmosphere of 5% CO, in air for 48 hours. G activity or exposure to semen. vaginalis isolates were identified by Gram stain Mycoplasma hominis was only isolated in two and hippurate hydrolysis tests. Yeasts were women-the one with constant BV, in whom confirmed by microscopy only but were it was isolated in 96% of swabs, and one other assumed to be Candida spp. For the isolation with intermittent BV, in whom it was found in of anaerobic organisms, plates of Columbia 3% of swabs taken. ,B-haemolytic streptococci Blood agar, containing kanamycin at 75 ,ug per were isolated from two women, both of whom ml were inoculated and incubated for 48 had Candida spp; in one it was found only on a http://sti.bmj.com/ hours, at 37°C in an anaerobic cabinet (Don single occasion, and in the other it was isolated Whitley Scientific Ltd, UK) containing nitro- from four consecutive swabs (22% of total) gen 80%, hydrogen 10% and carbon dioxide during the second week of the study. 10%. A 50 ,ug metronidazole disc was added Overall, only four out of 26, or 15% of to the inoculated streaks and, if a zone of women, had apparently completely normal colony inhibition was seen after incubation, flora throughout. on September 30, 2021 by guest. Protected copyright. the presence of "anaerobes" was recorded. There were no significant differences in the Cultures for the isolation of mycoplasma and number of swabs taken per week between the Ureaplasma spp were made by inoculating women with BV, Candida spp, normal plates of "MUD" medium (Unipath Ltd, microscopy, or U urealyticum, and in addition, Basingstoke, UK) directly with the swab, and there were no significant differences in the incubating for two days in the anaerobic cabi- proportions of positive swabs taken on a net. Cultures were examined and species iden- Friday or Saturday with those taken the rest of tified using a x 10 microscope objective. the week. This suggests that the intermittent nature of the microbiological findings was not Analysis Statistical analysis was performed due to a significant loss of sensitivity arising using Epi info, version 6.