Multi-drug resistant facultative pathogenic bacteria colonizing the vagina of pregnant women with premature rupture of membrane, Tanzania

Background: Premature rupture of membrane (PROM) contributes to approximately one-third of premature birth and 10% perinatal mortality worldwide. Here, we report the patterns of facultative pathogenic bacteria colonizing the vagina of pregnant women to guide prophylactic antibiotic treatment in the management of PROM. Methods: This comparative cross-sectional study was conducted between August 2015 and March 2016. High vaginal swabs were collected and processed to detect the presence of facultative pathogenic bacteria. Isolate identification and antibiotic susceptibility testing was conducted using MALDI-TOF MS and VITEK-2 system, respectively. Data were analyzed using STATA version 13. Results: A total of 175 pregnant women with PROM and 175 without PROM were investigated. The median age of the pregnant women with PROM was significantly higher than that of pregnant women without PROM: 27 [21-32] vs. 25[21-29], p=0.026. Pregnant women with PROM were significantly more likely to be colonized with facultative pathogenic bacteria 59/175 (33.7%), 95% CI; 26.7-40.7 than pregnant women without PROM; 27/175 (15.4%), 95% CI; 10.1-20.7, P<0.001. Escherichia coli were significantly more isolated from pregnant women with PROM than those without PROM: 36 (73.5%) vs. 13(26.5%), p<0.001. The proportion of resistance among pathogenic isolates from women with PROM to ampicillin, trimethoprim/sulfamethoxazole and cefotaxime were 100%, 66.7% and 40%, respectively. Conclusions: The vagina of pregnant women with PROM is significantly more colonized by multi-resistant facultative pathogenic bacteria than that of pregnant women without PROM. Further studies should be done to elucidate the impact of these bacteria in relation to PROM and the pregnancy outcome.


INTRODUCTION
orldwide, premature rupture of membranes (PROM) among pregnant women has been found to range from 3.3% to 10%, with 80% of them occurring at term 1,2 . PROM leads to the loss of the natural protection of the fetus hence posing a threat to bacterial infections 3 . In addition, PROM is highly associated with the increased pregnancy complications such as preterm labor, fetal demise, respiratory distress syndrome, neonatal sepsis, umbilical cord prolapse, postpartum endometritis, disseminated intravascular coagulopathy (DIC) and chorioamnionitis 1,4 . The pathogenesis of PROM has been linked to the isolation of facultative W Multi-drug resistant facultative pathogenic bacteria colonizing vaginal of pregnant women www.eahealth.org pathogenic bacteria in the vagina [5][6][7][8] , there is a strong association between pathogens colonizing the vagina and subsequent chorioamnionitis. Bacteria and protozoan parasites (Trichomonas vaginalis) secrete proteases and other factors that degrade the collagen and weaken the fetal membrane 9,10 . Furthermore, host inflammatory responses due to pathogenic bacteria can induce the production of prostaglandin which can lead to uterine irritability and membrane collagen degradation hence increasing the risk of PROM 8,11 .
In developing countries, Escherichia coli, Klebsiella pneumoniae, Group B Streptococcus (GBS), Staphylococcus aureus and Streptococcus pyogenes have been found to be the commonest facultative pathogenic bacteria colonizing the vagina and implicated in PROM 5-7,12. These pathogenic bacteria have also been implicated in chorioamnionitis 13,14 . Despite 12% contribution of the PROM to antenatal hospital admissions in East Africa 15 and black women being reported to have the highest risk of being colonized by the potential pathogenic bacteria 16 , the spectrum of the respective bacteria colonizing the vagina of pregnant women with and without PROM in East Africa is not well understood.
A previous study 17 in Uganda among pregnant women with PROM, noted the resistance to most commonly used antibiotics such as ceftriaxone, ampicillin, trimethoprim/sulfamethoxazole and erythromycin was high with good susceptibility to expensive antibiotics such as vancomycin and meropenem. The World Health Organization (WHO) recommends the use of antibiotic prophylaxis among pregnant women with PROM 18 . Different regimen involving penicillin and erythromycin have been recommended. However, number of factors should be assessed such as population of women to be offered antenatal prophylactic antibiotics and types of antibiotics to be used. With increase antibiotic resistance among the pathogens involved in PROM local susceptibility data are crucial in establish empiric treatment protocol with emphasis on the individual tailored treatment. This comparative cross-sectional study was designed to investigate the patterns of facultative pathogenic bacteria colonizing pregnant women with and without PROM and their antimicrobial susceptibility.

Study design, area, and population
The comparative cross-sectional study was conducted between August 2015 and March 2016 at Bugando Medical Centre (BMC), Sekou Toure Regional Hospital, Nyamagana District Hospital and Buzuruga Health Center in the city of Mwanza, Tanzania. The selected hospital are the highly populated faith base and government hospital which cover large percentage of the Mwanza population. Bugando Medical Centre is the tertiary consultant teaching hospital of Catholic University of Health and Allied Sciences with bed capacity of 900. Sekou Toure Regional Referral hospital is located in Ilemela district with bed capacity of 375 serving the all referrals from six district hospitals of Mwanza region. While Nyamagana District hospital has maternity bed capacity of 30 and Buzuruga health center has maternity ward with a bed capacity of 15. The study involved all pregnant women with gestation age of 36 weeks and above with and without PROM during the study period.

Sampling, inclusion, and exclusion criteria
Using Kirkwood formula for comparative studies 19 and assumption of the 10% effect size by the prevalence obtained in the previous study 12, the minimum sample size obtained was 350 pregnant women (175 pregnant women with and 175 pregnant women without PROM). The pregnant women with and without PROM were recruited serially until the sample size was reached. The study excluded all pregnant women with cervical incompetency, polyhydramnios, mal-presentation, multiple pregnancies, fever, abdominal pain, foul smelling per vaginal leakage and history of antibiotic therapy in the past two weeks prior to the study to minimize the bias in relation to the colonization of the facultative bacteria. In the current study, pathogenic bacteria were defined as bacteria which are potentially capable of causing clinical infections in the genital urinary tract 20 .

Sample collection and laboratory procedures:
By the use of sterile Cusco speculum, cervix was exposed, and high vaginal swab was taken using a sterile swab. The swabs were transported to the microbiology laboratory using Stuart transport media (HiMedia, India) within 2 hours of collection. All swabs were cultured on the 5% sheep blood agar (BA) and MacConkey agar (MCA) (Oxoid, UK) and aerobically incubated at 37 o C for 24-48 hours. Identification to species level was done by matrixassisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (Bruker Daltonics, Germany) on extracted cells as previous described 21,22. Only facultative pathogenic bacteria were included for statistical analysis and subsequent antimicrobial susceptibility testing. Potential contaminants, e.g. S. epidermidis, Bacillus spp., were excluded. The current study mainly concentrated on the antibiotics that are recommended for PROM management. The tested antibiotics included: ampicillin (AMP), trimethoprim/sulfamethoxazole, (SXT), ciprofloxacin (CIP), Gentamicin (CN), ceftazidime (CAZ), ceftriaxone Multi-drug resistant facultative pathogenic bacteria colonizing vaginal of pregnant women www.eahealth.org (CRO) and meropenem (MRP) (Oxoid, UK). Antimicrobial susceptibility testing was done using VITEK-2 system (bioMérieux, France) and interpreted as per EUCAST (http://www.eucast.org/clinical_breakpoints/) guidelines. Using excel sheet, data were double entered, cleaned and transferred to STATA version 11 for analysis. Categorical variables such as residence, marital status, education, occupation, gravidity, history of PROM, number of antenatal care (ANC) visit, presence or absence of PROM and positive or negative bacterial growth were summarized as proportions. Continuous data (age, gestation age and parity) were summarized using median and inter quartile range. The statistical significance was set at a p value of less than 0.05. Two-sample test of proportions was used to compare the pattern of facultative pathogenic bacteria colonization among pregnant women with PROM and those without PROM.

Ethics approval and consent to participate
The joint CUHAS/BMC research ethics and review committee granted ethical clearance with certificate number CREC/096/2015. Permission to conduct the study was sought from all hospital administrations. All patients were requested to sign a written informed consent before recruitment was done.
The median gestation age of pregnant women at the time of enrollment was 38 [36][37][38][39][40] weeks for pregnant women with PROM and 38 [37][38][39] for pregnant women without PROM. History of PROM in previous pregnancy was higher among pregnant women with PROM than pregnant women without PROM 20 (11.4%) vs. 6 (3.4%), p=0.004. In addition, history of preterm birth in the previous pregnancy was higher among pregnant women with PROM than those without PROM 16  Multi-drug resistant facultative pathogenic bacteria colonizing vaginal of pregnant women www.eahealth.org

DISCUSSION
Premature rupture of the membrane (PROM) contributes to approximately one third of premature births 2,23,24 , and approximately 10% of perinatal mortality 25 . A significant proportion of pregnant women with PROM in the current study had history of PROM and premature delivery in the previous pregnancy. This has also been reported elsewhere 26,27 and could be due to the possibility of the pregnant woman genetic defect in collagen synthesis that can affect the structure and function of the fibrillar collagens 28 . The weakening of the connective tissue by the enzymatic depolarization of the collagen fibers in fetal membrane can also explain the observed findings 29,30 . Women with PROM in the current study were significantly older than those without PROM pointing to the possibility of age-dependent collagen synthesis 31 . However, there is no documentation of the influence of age and facultative pathogenic bacteria colonizing vagina during the reproductive age. The age has been found to influence normal flora before puberty and after menopause 32, 33. Vaginal colonization with facultative pathogenic bacteria was significantly more often observed in pregnant women with PROM than in pregnant women without PROM. This has also been reported in previous studies in Tanzania and India 12,25 . Bacteria in the vagina have been found to secrete enzymes that can either degrade the fetal membranes or increase production of prostaglandins 4,12,34-39 high concentration of prostaglandins can stimulate the uterine contractions leading to the membrane rupture. This is further supported by the fact that the presence of pathogenic bacteria has been associated with chorion thinning among PROM pregnant women 40. As it was also previously reported in Tanzania and India 12,41 , E. coli and Pseudomonas spp. were the commonest bacteria detected. This could partly be explained by the fact that these pathogens belong to the normal flora of the gastrointestinal tract and might therefore be present in perineum with increased chance to colonize the genital tract. Staphylococcus aureus was the commonest gram-positive bacterial species detected among pregnant women with PROM in this study.
Similar observations were made previously 12 25 . Detection of S. aureus colonizing women with PROM has been linked with other factors like urinary tract infections and bacterial vaginosis 25 , these factors were not investigated in the current study.
Bacteria isolated from pregnant women with PROM were more resistant to ampicillin and trimethoprim/sulfamethoxazole. Similar results have been reported in other studies from pregnant women 42,43 and postdelivery women in similar settings 44 . This could partly be explained by the fact that ampicillin and trimethoprim/sulfamethoxazole are the commonest class of antibiotic in use in the study settings. Despite the fact that this study has clearly demonstrated the significant differences in patterns of bacteria colonizing pregnant women with PROM and those without PROM, neonatal outcomes which could give more evidence on the association of vagina colonization of the facultative pathogenic bacteria and clinical fetal infections especially for the pregnant women with PROM were not recorded. Furthermore, the presence of Candida species were not assessed, this has been recommended for future studies.
In conclusion, vagina of pregnant women with PROM was more colonized by multi-resistant facultative pathogenic bacteria than the one of pregnant women without PROM. Escherichia coli strains were the commonest pathogenic bacteria and were highly resistant to ampicillin and trimethoprim/sulphamethoxazole. Further studies should be done to elucidate the impact of these pathogens in relation to PROM and the pregnancy outcome. There is a need to adjust the empirical prophylaxis treatment of PROM based on the local susceptibility profile.