New paint: no harm, no foul?

Objective: In vitro, Candida albicans has demonstrated the ability to inhibit replication of selected bacteria. Little information exists on the impact of C. albicans on the vaginal bacterial flora in vivo. The purpose of this study is to identify the coexisting bacterial flora when C. albicans is isolated from vaginal cultures submitted to a hospital-based testing facility for reasons other than vulvovaginitis. Methodology: All specimens (240) received from ambulatory care clinics over a six-month period were cultured for aerobic and anaerobic bacteria and Candida species. Those specimens submitted for cervicitis, vaginitis, or vaginal discharge and those from which yeasts other than C. albicans were isolated were eliminated. To control for sample biases, a subgroup composed of all pregnant women for whom cultures were done as screening procedures was similarly studied. Chi-square analyses, comparing the prevalence of individual bacteria isolated with and without the presence of C. albicans, were done for all study populations using SPSS for Windows software (1994). Results: Two hundred and forty consecutive specimens were bacteriologically analyzed. Of the 220 vaginal samples used in the study, C. albicans was isolated in 44 instances (20%). Neither the presence of the lactobacilli nor the presence of Gardnerella vaginalis markedly influenced the iso

presence of the lactobacilli nor the presence of Gardnerella vaginalis markedly influenced the isolation rate of C. albicans. The group B streptococci had a greater probability of coisolation when C. albicans was present (27.3% verses 16%), but this was not statistically significant (P < 0.8). Dissociation between the presence of C. albicans and the coisolation of Peptostreptococcus species and anaerobic gram-positive cocci and/or bacilli was noted (P < 0.0819). while the incidence of gram-positive aerobic bacilli was reduced in the presence of C. albicans (30/176 17.1%] versus 6/44 [13.6%]), this reduced incidence was not statistically significant. Isolation data of the subgroup of pregnant women supported these observations. Conclusion: Within the limitations of the study, statistically, the data suggests that an inverse relationship exists between the presence of C. albicans and recovery of Peptostreptococcus and anaerobic gram-positive cocci and bacilli. Infect. Dis

Patient Population
The patient population was made up of women attending the ambulatory care facilities of Resurrection Medical Center. The clinical indications for the vaginal specimens obtained are listed in Table  1. All cultures submitted for endometritis, cervicitis or vaginal discharge, or vaginal itching/burning were excluded from analysis. All culture data derived from pregnant women was analyzed separately.

Specimen Handling
All specimens received over a six-month period were cultured for aerobic and anaerobic bacteria and Candida species. These specimens were received primarily from ambulatory care clinics. The specimens had been obtained with Culturette II swabs. Gram stains were performed on all cases, and the swabs were cultured for aerobes on colistin-nalidixic acid agar, chocolate blood agar and eosin-methylene blue agar; and for anaerobes on phenylethyl alcohol agar, Centers for Disease Control and Prevention (CDC) anaerobic blood agar, kanamycin-vancomycin agar, and enriched thioglycollate medium (media were obtained from BBL, Becton Dickinson Microbiology Systems, Cock-eysville, MD). Aerobic media were incubated at 35C and analyzed after 24 hours, while anaerobic media were incubated in 70% nitrogen and 30% carbon dioxide at 35C and analyzed after 48 hours. Additional work-up of aerobic cultures involved interpretation of gram stains, hemolysis patterns, and colony characteristics. These findings plus catalase (0.1% hydrogen peroxide) and coagulase (Staphaurex, Murex Diagnostics, Daford, England) test results were used in the Vitek System (Bio-Merieux, Hazelwood, MO) for identification and antibiotic sensitivities.
Additional work-up of anaerobic cultures required confirmation of anaerobic bacteria by inoculating the suspected anaerobic organism onto blood agar and anaerobic CDC media. After overnight incubation in aerobic and anaerobic conditions, respectively, organisms, which grew on the latter medium but not the former, were worked up as anaerobes. Gram stains, colony morphology, and the Rapid ANA II System (Innovative Diagnostic Systems Inc., Norcross, GA) contributed to the identification of these organisms.

STATISTICAL ANALYSES
Chi-square analysis compared the prevalence of individual bacteria in both subgroups of pregnant women. No continuity correction for sample size was necessary. Power analysis showed a 75% and a 57% power to detect differences of sizes observed using a one-tail test at the P < 0.05 significance level. Statistical analysis were done using the statistical software package SPSS.

RESULTS
Two hundred and forty consecutive specimens were bacteriologically analyzed. Thirteen were eliminated because of the diagnosis of cervicitis (two) or vaginitis/vaginal discharge (11  The principal subgroup (N 67) within the study sample was composed of pregnant women. The main reason for taking these cultures was to rule out the presence of group B stretococci and Neisseria gonorrhoeae. Candida glabrata was isolated from two of these specimens, and they were discarded from the study. Seven additional specimens were deleted from the study based on a clinical diagnosis suggesting vaginitis. Of the remaining 58 vaginal specimens derived from pregnant women, C. albicans was identified in 13 (22.4%). The relationships previously described for the nonpregnant women were sustained in the analyses applied to this subgroup. The incidence of isolation of group B streptococci was 12% (7/58) in the absence of C. albicans versus 30.8% (4/13) in the presence of C.
albicans. The incidences of recovery of Peptostreptococcus and anaerobic gram-positive bacteria among the pregnant patients were 20% (9/45) in the absence of C. albicans versus 7.7% (1/13) in the presence of C. albicans. The distribution of bacteria among the pregnant women relative to the presence or absence of C. albicans is listed in Table 3. While numerically superior to most published studies, the current study suffers from a number of defects, namely, the lack of quantitative microbiology for each isolate and the use of prospectively identified study populations constituted by diverging clinical status. The clinical indications received with the specimens might not have adequately characterized the actual reasons for culture submission. Fortunately, the samples derived from pregnant women, which were submitted primarily for prenatal indications, provided a subset against which observation derived from the general population could be compared. An additional shortcoming of this study is the technique used to obtain clinical specimens and the handling of specimens. The specimens were suboptimally handled for recovery of anaerobic bacteria specimens prior to reaching the biological test facility. Those isolates achieved probably represent bacteria present in high numbers in terms of colony-forming units per gram of vaginal fluid.
Microbiological studies of the bacterial flora in cases of vulvovaginal candidiasis have been severely limited by their limited scope and by the lack of quantitative data. The issues of interspecies regulation by C. albicans (whether through bacteriocins, hydrogen peroxide or other bactericidal proteins, or enzymes) have largely been neglected. [6][7][8] Confronted with a paucity of data, the prohibitive costs of obtaining such data, and the small likelihood that the required funding will materialize, one is forced to deal with fragmentary observation.
The overall decrease in the nubmer of coisolates achieved when C. albicans is present observed in this study is consistent with the observations in the literature. When disease due to C. albicans is present, the number of colony-forming units per milliliter of vaginal fluid is probably increased to a point where C. albicans can probably exert a maximal selective inhibitory effect. This study focused on cases where vulvovaginal candidiasis was not clinically present and asked whether at nondisease levels there was any disparity in the vaginal flora when C. albicans was present and when it was not.

INFECTIOUS DISEASES IN OBSTETRICS AND GYNECOLOGY 55
Statistically, the current study suggests that an inverse relationship exists between the presence of C. albicans and recovery of Peptostreptococcus and anaerobic gram-positive cocci and bacilli.