Bacterial isolates from patients with preterm labor with and without preterm rupture of the fetal membranes.

OBJECTIVE
The aim of this study is to describe the bacterial flora of women in preterm labor with or without premature rupture of membranes.


METHODS
Retrospective studies of 239 patients with preterm labor were performed.


RESULTS
One hundred and twenty-three of 239 patients with preterm labor (51.5%) had bacterial vaginosis. Seventy of the 239 patients with preterm labor (29.3%) developed premature rupture of the membranes (preterm PROM). Of the 70 patients with preterm PROM, 51 (72.9%) had bacterial vaginosis. Therefore, 51 of the 123 patients with bacterial vaginosis (41.5%) developed preterm PROM. An increased number of organisms detected from the vaginal discharge in patients with preterm labor was associated with preterm PROM by Cochran-Armitage test. An increased number of organisms detected from the vaginal discharge in patients with preterm labor complicated with bacterial vaginosis was significantly associated with preterm PROM by Cochran-Armitage test.


CONCLUSIONS
In preterm labor, the number of different species detected in the vagina provide sensitive and specific prediction of preterm PROM in patients with preterm labor.

KEY WORDS preterm labor; preterm PROM; prediction; bacterial vaginosis reterm birth is the leading perinatal complication. [1][2][3] Since intrauterine infection from organisms found in the genital tract has been implicated in both the etiology and adverse sequelae of preterm premature rupture of the membranes (preterm PROM), -6 culturing the vagina has not been an essential part of clinical management. Although there are many studies that evaluate such cultures in the prediction of intrauterine pregnancies, they have been focused on the identification of bacteria. [4][5][6] The aim of this study is to investigate the significance of bacterial culture in predicting preterm PROM in patients with preterm labor.

MATERIALS AND METHODS
Subjects From 1992 to 1996, we sampled the vaginal discharge of 239 patients with preterm labor that were referred to the hospital of the School of Medicine of Gifu University for further assessment. The patients were Japanese married women of middleclass socioeconomic status who had had no sexual intercourse in the past 10 days, had no history of sexual transmitted diseases, and had taken no antibiotic therapy during pregnancy. Patients with chlamydia, trichomoniasis, or candidiasis infection were excluded from the study. Consent to the study was obtained from the patients and the committee of the institution.

Diagnosis of Bacterial Vaginosis
Bacterial vaginosis was diagnosed with intact membranes in 239 patients by the presence of the following four criteria: 7-1 thin and homogenous vaginal discharge, vaginal pH above 4.5, positive amine "whiff' test, and the presence of clue cells in wetmount preparations of vaginal discharges.

Sampling Method
Vaginal discharge was collected from the posterior vaginal fornix with a sterilized cotton wool swab, which gathers 0.05 mL of sample per swab.

Culture Methods
Immediately after collection, the swab. specimen (0.05 mL) was suspended in 5 mL of the anaerobic diluent for culture. [8][9][10][11] The composition of the anaerobic buffer was as follows: 4 g of KHzPO4, 6 g of NazHPO4, g of L-cysteine-HC1-H20, g of Twecn 80 (Sigma, St. Louis, MO), and agar, and 1,000 mL of distilled water with a pH of 7.2. All the components were mixed in a solution and heated at 80C for 30 minutes. All tubes were sterilized in an autoclave at 115C for 20 minutes. A 9-mL aliquot of the buffer was transferred to each test tube, air in the tube was immediately replaced with CO z, and the tube was sealed with a butyl rubber stopper.
After samples were suspended in the buffer, the tubes were resealed under a continuous stream of CO z gas of commercial grade to drive air out. Exposure of samples to atmospheric oxygen was restricted to 5 minutes or less. Incubation commenced immediately after a sample was suspended in a solution. A small aliquot for the subsequent culture was aspirated with a syringe via a butyl rubber stopper.
For quantitative bacterial determination, serial dilutions were made with the anaerobic buffer. The time lag from the specimen collection to the quantitative determination was within hour.
Staphylococcus-selective agar (Nissui Pharmaceutical Co. Ltd., Tokyo, Japan) and MacConkey agar (Becton Dickinson and Co., Cockeysville, MD) were used for aerobic culture. Sheep blood agar (Becton Dickinson) and chocolate agar (Becton Dickinson) were used for CO z culture.
As for Gardnerella vaginalis, human bloodbilayer-Tween (HBT) medium, developed by Totten et al., lz was used for CO z culture. It consists of a bottom layer of Columbia colistin-nalidixic acid agar (Becton Dickinson) supplemented with 1% Proteose Peptone no. 3 (Difco Laboratories, Detroit, MI), amphotericin B (2.0 pg/mL), and 0.0075% Tween 80 (Becton Dickinson) and a top layer of the same composition with 5% human blood added.
Aerobic culture was performed at 37C for 2 days, CO z culture in 5% CO z in air at 37C for 3 days, anaerobic culture in a GasPak Pouch at 37C for 7 days, and fungal culture at 37C for 7 days.

Bacterial Identification
Each colony with an individual appearance on an agar plate was subcultured three times on a nonselective medium to obtain a pure culture for bacterial identification. 8-1 Among aerobes, grampositive and catalase-positive cocci were identified by Api STAPH identification system (bioMerieux SA, Marcy l'Etoile, France). Gram-positive and catalase-negative cocci were identified by Api STREP identification system (bioMerieux SA), and gram-negative rods were identified by Enterotube II (Becton Dickinson) or Oxi/Ferm Tube II (Becton Dickinson). Haemophilus species and G. vaginalis were identified by Rap ID NH (Innovative Diagnostic System, Inc., Atlanta, GA). Anaerobic bacteria were identified by Rap ID ANA System II (Innovative Diagnostic System). For identification of Mobiluncus spp., colonial characteristics were observed and a Gram stain was performed. Gram-negative or gram-variable curved rods were identified by Rap ID ANA System II.
Nitrate reduction and hippuric acid hydrolysis tests were used for identification at the subspecies level.    Table 2 shows the number of different species detected from the vaginal discharge in the 123 patients with bacterial vaginosis and the 51 patients with both preterm PROM following preterm labor and bacterial vaginosis. An increase in the number of organisms detected in the vaginal discharge of patients with preterm labor complicated with bacterial vaginosis was significantly associated with preterm PROM by Cochran-Armitage test (P 0.0030).
The organisms isolated from the vaginal discharge in patients with preterm labor are listed in

DISCUSSION
Recent studies have reported the relation between preterm PROM and bacterial vaginosis; 3-s,ll this condition is characterized by depletion of vaginal Lactobacilli with increased G. vaginalis in association with such organisms as Bacteroides species, Prevotella species, Peptostreptococcus species, and Mycoplasma species. 7-aa This study suggests that the increased number of organisms detected from the vaginal discharge might be predictive of the prevalence of preterm labor and preterm PROM. This means that the number of organisms detected in the vagina provides sensitive and specific prediction of intrauterine infection with aerobic or anaerobic organisms in pregnancies with preterm PROM. Most studies associated with vaginal flora in pregnancies have found that it is the relative quantity of organisms. 3-s' However, great efforts are needed in the quantitative assays. Therefore, we gave attention to the numbers of different species detected.   tures, the same organisms were recovered from vaginal swabs. 11 This finding is compatible with the hypothesis that the lower genital tract is the source of the offending organisms in intrauterine infection associated with preterm labor.
Most bacteria responsible for ascending infections are derived from the indigenous genital flora. 11 Anaerobic bacteria are of major importance. 11 Many of these anaerobes are considered potentially pathogenic, as they have been frequently isolated in patients with genital infections. [5][6]11 Genera such as the anaerobic cocci, Bacteroides spp., Prevotella spp., and Fusobacterium spp. have been implicated in maternal or neonatal infections accompanying premature labor or PROM.S-6,11 In a matched case-control study of 54 consecutive women in preterm labor, Gravett et:al. found a significant association between bacterial vaginosis and preterm labor (43% of study patients versus 14% of control patients, P 0.02, relative risk ratio of 3.80). Martius et al. 14 performed a similar study at the same center using Gram stain to diagnose bacterial vaginosis and had similar findings. Bacterial vaginosis was significantly associated with preterm labor (odds ratio of 2.3), as was the presence of Chlamydia trachomatis and the absence of Lactobacilli species. In this study, 70% of women with preterm labor had PROM.
The results of our study add to the existing evidence that bacterial vaginosis is an independent risk factor for preterm birth and suggest that the timing of this infection in gestation significantly affects this risk. It remains to be shown whether this association is causative or whether bacterial vaginosis is associated with some as yet unidentified factor that initiates preterm labor and birth.
Although bacterial vaginosis is commonly found in pregnant women, preterm birth usually does not occur in these pregnancies. In addition, it remains unclear whether treatment of pregnant women with bacterial vaginosis decreases their risk of preterm birth. A large, well-controlled, randomized trial of treatment for bacterial vaginosis in pregnancy is needed to answer this question.