Abstract
Background/Aim: This retrospective cohort study enrolled hospitalized women with 24+0 to 33+6 gestational weeks with conditions associated with preterm birth. We evaluated the ability of vaginal swab isolates to guide antibiotic management decisions in the setting of threatened preterm towards a clinical advantage, i.e., longer delay between diagnosis and birth, better neonatal outcomes. Patients and Methods: Vaginal swabs were obtained from all patients and antibiotic resistance profiles determined in case of growth. The cohort was divided into two groups: the antibiogram-noncongruently managed Group 1 and the antibiogram-congruently managed Group 2. These groups were compared in regard to multiple maternal and neonatal endpoints. Results: In total, 698 cases were analyzed – 224 in Group 1 and 474 in Group 2. Antibiotics were ordered/continued by the treating physician in 138 cases (138/698; 19.8%) upon review of vaginal swab cultures results. Forty-five among them (32.6%) received antibiotics inactive against the isolated bacteria. 335 (25.4%) patients had only normal vaginal flora, and 95.6% of them had not received antibiotics. Facultatively pathogenic microorganisms were isolated in 52% patients. Only 5% of the neonates had bacterial isolates identical to those of their mothers. There were no significant differences in outcomes between Group 1 and Group 2. Conclusion: No association was found between a swab-result-guided antibiotic management protocol and maternal or fetal outcome in the setting of preterm birth risk between 24 and 34 gestational weeks. These findings underline the importance of critical rethinking the frequency of vaginal smears and fine-tuning the indications for antibiotic treatment.
A child born after less than 37 weeks of gestation is considered preterm (1). Preterm birth significantly influences neonatal morbidity and mortality, but can have long-term health repercussions as well (2). The global preterm birth rate is estimated to be between 5% and 18% (amounting to approximately 15 million cases annually) and shows a rising trend (1). Germany has a preterm birth rate of approximately 8.7% - one of the highest in Europe (3). Developing countries are in an even more precarious position in this regard (4). Finally, preterm birth is estimated to be an economically taxing pathologic entity with associated medical and educational expenditures and loss of productivity incurring a cost upwards of 26.2 billion dollars in 2005 in the US alone (5).
Spontaneous preterm labor without any other discernible coinciding pathologic processes accounts for up to a half of all preterm births (6). Some 30% of preterm births are associated with a preterm prelabor rupture of membranes (PPROM), while approximately 20-30% are due to myriad other causes leading to medically indicated preterm birth (6-8). A small number of infants are born preterm due to cervical insufficiency; however, this has proven difficult to evaluate, since cervical shortening is only occasionally a feature of preterm labor (6, 9, 10). Preterm labor is usually defined as the specific contraction pattern and cervical effacement consistent with labor and occurs before 37 weeks of gestation (11). Spontaneous preterm labor occurs due to a number of reasons (infection/colonization by various microorganisms is one of the most commonly cited) and the pathophysiology remains to be fully elucidated (12, 13).
Based on comprehensive studies, most obstetric authorities agree that late preterm labor after 34 weeks of gestation can be managed without labor inhibition due to the fact that lung surfactant levels are significant at that point and beyond in most cases (11, 14). However, as previously stated, it has long been known that preterm labor (especially when associated with PPROM) coincides with infections of the woman’s reproductive tract (7, 15). Also, genital tract colonization with any one of a menagerie of microorganisms, with Group B Streptococcus (GBS) being the most clearly demonstrated to be pathogenic, can predispose to intra-amniotic infection, endometritis, puerperal sepsis, neonatal sepsis and meningitis, etc. (16-18). However, this association is not straightforward when it comes to some bacteria like E. coli which is the most common cause of early onset sepsis and meningitis in preterm newborns (19). Even more intriguing is the fact that vaginal microbiome perturbations seem to be associated with preterm labor (20). Nevertheless, antibiotic therapy was not proven to be useful in managing preterm labor with intact membranes outside the GBS colonization context (especially in the late preterm setting) (21). Hence, antibiotics are not routinely recommended in the context of preterm labor with intact membranes.
Antibiotic use in cervical insufficiency is less well studied and there is no consensus about their usefulness (except periprocedurally for cerclage placement) (22, 23). PPROM is an indication for antibiotic prophylaxis of intrauterine infection (24). Vaginal bleeding is not considered an indication for antibiotic therapy unless it is associated with PPROM, signs of intrauterine infection or GBS colonization (25).
As stated, GBS colonization of the female genital tract has been unequivocally associated with adverse outcomes during labor thus necessitating a preemptive management strategy (16). Other bacteria commonly isolated from vaginal swab cultures (especially E. coli, and S. aureus) are considered capable of causing infections in the mother and/or child in the perinatal period (26, 27). However, the association has not thus far been deemed strong enough to necessitate active screening and subsequent preemptive treatment to that end. Many other bacteria that can colonize the female genital tract have been proven to cause infection immediately after labor, sometimes with serious consequences (27).
Interestingly, no studies that we are aware of have comprehensively scrutinized vaginal swabs analyzed by modern microbiologic techniques as a means of predicting adverse outcomes in women in the context of spontaneous premature labor. Given current GBS prevention practices most obstetricians routinely obtain vaginal swabs at presentation if preterm birth of any etiology is a possibility providing an ideal opportunity for the expansion of microbiologic screening. If this approach is proven useful in identifying women with a higher risk of complications and an earlier delivery, then an expansion of the indications for antimicrobial therapy in this setting might be prudent. Furthermore, many neonatologists assume that the knowledge of the maternal vaginal microbiome at the time of preterm delivery might be helpful to select the most suitable antibiotic therapy for neonates with signs of infection (19, 28, 29).
In this study we evaluated the ability of standard vaginal swab specimen cultures to guide antibiotic management decisions and correlate adequate antibiotic therapy with maternal and neonatal outcomes in the setting of threatened preterm birth before 34 weeks of gestation.
Patients and Methods
This retrospective cohort study enrolled all obstetric patients hospitalized due any condition associated with preterm birth at the Department of Gynecology and Obstetrics, University Medical Center Mannheim, Germany from January 1, 2010, to December 31, 2014. The patients were diagnosed according to the current local guidelines (14, 30). The four pathologic entities associated with suspected preterm birth our study focused on were preterm premature rupture of membranes, preterm labor, vaginal bleeding in late pregnancy (i.e., women with threatened abortions were not included in our analysis), and cervical insufficiency. Women diagnosed with late suspected spontaneous preterm birth (from 34+0 to 36+6 weeks) and those whose fetuses were considered periviable at presentation (gestational age less than 24+0 weeks) were excluded from the study due to specific considerations that would add undue complexity to our research question. Patients with fungi isolated from their vaginal swabs were excluded from the study for the same reason. Furthermore, to be eligible for enrolment, the patients had to have been at least 18 years of age at time of presentation. If a patient had been demonstrated to have had bacterial vaginosis at presentation she also was excluded from the study in an attempt to streamline our research since this entity is of unclear significance in the context of preterm birth, yet some have demonstrated its association with preterm birth and recommend treatment (17, 31-33). Finally, patients with GBS isolated from their vaginal and/or rectal swabs and/or urine were excluded from the study given that such patients were previously definitively demonstrated to be at risk of maternal and/or fatal infection and as such undoubtedly require targeted antibiotic therapy.
A vaginal swab was obtained from all patients at presentation. Subsequently all samples were cultured on four different media: Columbia Agar with 5% sheep blood, McConkey Agar, Gardnerella Agar, and Group B Streptococcal Selective Agar. The only exception were samples from patients who had not been registered as having had been admitted to the delivery room or an obstetric ward and whose diagnoses in the microbiology laboratory request form had not been pregnancy related, i.e., the microbiologist had no way of knowing that the sample had been obtained from a patient who was pregnant. Swabs from these patients were not analyzed for the presence of GBS. Columbia sheep blood 5% agar, McConkey agar and an aerobic broth were used to culture blood samples and samples from infected sites of the neonate (the throat, ear or eye, gastric secretions, and other potentially infected areas) as indicated by pediatricians. The final identification of a bacterium species was made in pure culture utilizing matrix-assisted laser desorption/ionization time of flight spectrometry (MALDI-TOF) or VITEK®-2, while VITEK®-2/Advanced Expert System (VITEK-2/AES), agar disc diffusion or the Epsilometer-Test (E-Test) were used for antibiotic resistance testing. Microbiologists reviewed the results and made comments regarding the pathogenicity of the isolated bacteria in the clinical context of every individual case. Atypical bacteria like genital Mycoplasma spp. were not specifically sought, because they are part of the routine early pregnancy screening (up until 28+0 weeks) and would be treated early on if diagnosed. All microbiologic investigations were conducted at the Institute for Medical Microbiology and Hygiene of the University Medical Center Mannheim, Heidelberg University, Mannheim, Germany.
Histories and examination findings, as well as data on therapeutic measures and disease courses were retrospectively extracted from both the digital patient databases (SAP® and Gedowin®) and the archived paper-based patient histories. The patients were considered underweight if their BMI was lower than 18.5 kg/m2 and overweight if their BMI was higher than 24.9 kg/m2. Laboratory values and smear results came directly from the laboratory server of the University Medical Center Mannheim (UMM). The data on the neonates also came from the aforementioned digital databases from the UMM Department of Neonatology. Apgar scores after 1 min were used.
Patients were discharged when they showed signs of clinical improvement and their condition was deemed stable. Conversely, if signs and symptoms of the pathology for which the patient had been admitted were not abating with therapy, the women remained hospitalized until delivery. Immediately after birth, the newborn was examined and in the case of malformation, peripartum infection or prematurity, he/she was transferred to the Department of Neonatology, where further diagnostics and therapy were carried out by pediatricians. Data collection began in September 2015 and a Microsoft Excel database was compiled by mid-2017.
The study cohort was divided into two groups: antibiogram-non concordantly treated group (Group 1) and the antibiogram-concordantly treated group (Group 2). Group 2 comprised patients who A) either received no antibiotics (empirically or after receiving culture results) and were demonstrated to have a normal vaginal flora; or B) were treated with antibiotics that were active against the isolated bacterium/bacteria. In Group 2 antibiotics were given either empirically (before swabs culture results became available) or as causal therapy (after a bacterium was cultured and antimicrobial sensitivity was determined) but only patients receiving antibiogram-concordant treatment (or no therapy if the flora was shown to be normal) after the isolate became known were delegated to Group 2.
All patients that did not fulfill the criteria to be included in Group 2 were designated to Group 1. This includes patients who were managed with antibiotics that were not active against the isolate/s and patients treated with antibiotics even though they were demonstrated to have a normal vaginal flora. To reiterate, patients were considered treated with antibiogram- concordant therapy only if they received antibiotics active against the isolated bacterium, i.e., their empiric antibiotic therapy was proven active against the isolate and was, therefore, continued or they were switched to antibiotics active against the isolate. Logically, patients that were switched from active (based on the isolate’s antibiotic susceptibility) empiric antibiotic therapy to an antibiotic which was not effective against the isolated bacterium were assigned to Group 1. All patients diagnosed with PPROM were assigned to Group 2 if they had received antibiotics and to Group 1 if they had not received antibiotics.
Our premise was that if vaginal swab cultures were used (other than to diagnose GBS), the group of patients managed with antibiotics active against the isolates should be in an advantage (longer delay between diagnosis and birth, better neonatal outcomes, etc.).
The characteristics of a given patient and her child were presented as proportions, percentages and medians and ranges (the data were non-normally distributed). The primary end points were the deferral of preterm labor to attain the 34 and 37 week of gestation milestones. Neonatal mortality, duration of in-patient treatment, early onset neonatal infection, NICU (neonatal intensive care unit) admission, mechanical ventilation of the neonate, Apgar score, birth weight, puerperal infection, labor delay of at least 48 h and delivery at term were also evaluated. The neonates were followed-up until hospital discharge or death.
Anonymized patient data were incorporated into retrospective statistical analyses, which were conducted using IBM’s SPSS Statistics software v23 (IBM Corp., Armonk, NY, USA). A p-value of <0.05 was utilized throughout our study. Categorical variables were analyzed using the Chi square and Fisher tests as appropriate. Between group comparisons were carried out using the Mann–Whitney-U-test. We also used multivariate logistic regression analysis to control for confounders and determine predictors of birth before 34 weeks of gestation. We included all the factors we deemed important for this purpose regardless of their significance in univariant regression (the Enter method in SPSS); these included maternal age, week of gestation on presentation, number of fetuses, presence of cervical insufficiency, presence of vaginal bleeding, presence of PPROM, premature labor, antibiotic administration at any time point of the hospitalization, vaginal swab isolate-guided antibiotic management, the presence of normal vaginal flora and use of tocolysis.
The study was approved by the Ethics Committee of the Mannheim Medical Faculty of the University of Heidelberg, (approval no. 2015-851R-MA, issued October 8, 2015) and was conducted adhering to the principles outlined in the Declaration of Helsinki.
Results
We analyzed 698 hospitalizations in the obstetric unit of our university hospital – 224 were assigned to Group 1 and 474 to Group 2. The median patient age was 30 years (range=18-47 years) and the median week of gestation at presentation was 30 weeks (range=24-33 weeks). Other patient and newborn characteristics are presented in Table I. The study groups were not significantly different regarding baseline patient and newborn characteristics. Fifty-six patients (8.2%) were underweight, and 212 (30.4%) were overweight.
Clinicopathological characteristics by patient group.
Antibiotics were administered in 324 of the cases (this includes both patients given antibiotics empirically and patients first given antibiotics after reviewing bacterial cultures). Two hundred and ninety patients (290/698, 41.6%) were treated with antibiotics empirically; 70 of these patients (70/290, 24.1%) were treated with antibiotics that were proven active against vaginal swab isolates. Only normal vaginal flora was isolated from fifteen patients treated with antibiotics empirically. One hundred twenty-two patients (122/698; 17.5%) were diagnosed with PPROM. However, 31 of these received no antibiotics (31/122, 25.4%).
Antibiotics were ordered/continued by the treating physician in 138 cases (138/698; 19.8%) after vaginal swab culture results were reviewed. Forty-five amongst them (45/138, 32.6%) received antibiotics that were not active against the isolated bacteria. Three hundred thirty-five (335/698; 25.4%) patients had only normal vaginal flora isolated with most (320/335; 95.6%) not given antibiotics after the clinician was informed of the results.
Vaginal bleeding seems to have influenced the decision of the clinicians to prescribe antibiotics that were not indicated in the context of the bacteria cultured from their patient’s vaginal swabs (Table I).
Table II outlines the frequency of isolation of individual bacterial species or groups of bacteria. Three hundred thirty-five patients (325/698, 48%) had a normal vaginal smear, whereas facultatively pathogenic microorganisms were isolated in 363 patients. As could be expected, most of the isolates are common skin and gastrointestinal tract colonizers. Coagulase-negative staphylococci (CNSs) were the single most common isolate. Surprisingly, S. aureus was detected in only 1.1% of patients. In 5% percent of the analyzed neonates (45/900), bacterial isolates were deemed to be identical to those isolated from the mother’s vaginal swab.
Isolates from vaginal swabs.
Table III outlines the maternal and neonatal outcomes by study group and overall. Three hundred ninety-five patients (56.6%) eventually carried their pregnancy to term. Nine newborns died. Three were born to mothers colonized with E. coli, and 2 to mothers with a normal vaginal flora. Antibiotics were administered to the mother on presentation to the hospital in eight of these instances. There were no significant differences in outcomes between Group 1 and Group 2.
Pregnancy outcomes and neonatal management characteristics.
The results of our multivariate logistic regression aimed at finding the predictors of delaying labor until week 34 are presented in Table IV. Even though the model was statistically significant (p<0.001) in predicting birth delay until the fetus is at least 34 weeks old, it was successful in explaining only 47.4% of the variability (Nagelkerke R square 0.474).
Multivariate logistic regression analysis of the predictors of delaying labor until week 34.
Women who presented with more advanced pregnancies were more likely to give birth after 34 weeks of gestation. The opposite is true for older women and women admitted because of PPROM and vaginal bleeding and women with multiple fetuses. Interestingly, tocolysis did not seem to help delay birth in our study. Neither overall antibiotic use, nor antibiotic management guided by vaginal swab results, seemed to make an impact on birth delay until gestational week 34.
Discussion
Our study demonstrated that many different potentially pathogenic microorganisms can be cultured from the vagina of women with suspected preterm birth that exhibit no clinical signs of infection. It has, however, proven difficult to reliably gauge their presumed pathogenicity. We have isolated E. coli, CNS, S. aureus, Enterococcus spp., and numerous other, mostly Gram-negative, bacteria. It is unclear whether these organisms are involved in the complex pathophysiology of preterm birth (other than in the context of PPROM). The prevailing opinion is that immune activation and direct damage caused by microorganism can indeed cause premature labor, PROM and PPROM (12, 13, 15). Our isolate frequencies were similar to those reported by other authors, although, unlike us, others have reported on the frequency of bacterial vaginosis and atypical bacteria in the vaginal swabs obtained from their patients (19, 34, 35). Enterococcus spp. were less frequently isolated according to data presented by some other authors (19, 34, 36).
E. coli has been known to cause neonatal meningitis and sepsis for decades (19). However, it is unclear if the detection of vaginal colonization by this bacterium and consequential antimicrobial treatment reduces the incidence of neonatal disease caused by E. coli. Moreover, usual contaminants like CNS may also play a role in the pathology of preterm birth (37-40). Numerous cases of neonatal colonization by the same bacterial strains found to inhabit the birth canal and adjacent tissues were reported (28, 41-43).
There were no statistically significant differences in relation to gestational birth age, length of time from presentation to delivery, birth weight, neonatal hospitalization duration, NICU admission and mechanical ventilation, Apgar score, and neonatal death rate between patients treated with antibiotics targeted at bacteria isolated from vaginal swabs and those who were not. Logistic regression modeling was used to control for confounding factors and identify true predictors of delivery deferral until at least to 34 weeks of gestation. Management with antibiotics to which vaginal swab isolates were susceptible and withholding antibiotics for patients with normal vaginal flora was not found to be a significant predictor of delivery postponement. By extension, this means that vaginal swab cultures (other than those aimed at identifying GBS), seem not to be useful for guiding antibiotic therapy in the context of threatened preterm birth. Further, this could mean that vaginal swab cultures (other than GBS cultures) are probably an unnecessary medical expense. Finally, even though vaginal swabs are ordered for most patients with threatened preterm birth, the results mostly do not lead to a switch in management strategies.
To the best of our knowledge, in only 5% (benchmark) of the newborns do the microbiological smears contain the same bacteria as the vaginal smear of the mother.
Others somewhat corroborate our findings, although no other studies dealt with exactly the same questions we attempted to tackle here (19). There is no proven association between neonatal colonization with maternal vaginal flora other than GBS and adverse outcomes (44). Some authors advocate empiric antibiotic therapy when pregnant patients present with cervical insufficiency though this still is a matter of debate and it is unclear whether vaginal swab cultures can help optimize therapy (22, 23). PPROM is considered a clear indication for empiric antibiotics by most authorities (24). However, vaginal swab cultures (other than those aimed at identifying GBS) do not seem to be useful in predicting causative organisms of intraamniotic infection (45). Spontaneous preterm labor is not an indication for antibiotic therapy since multiple studies (including a meta-analysis) have not found a benefit from such management strategies (11).
There are several limitations to our study. We chose a very specific subset of patients that has not (to our knowledge) previously been directly studied in the context proposed here, somewhat compromising the generalizability of the conclusions. Further, an inherent drawback of retrospective cohort studies is the fact that it is an immense task to control for all confounders. The conclusion is made indirectly, which is problematic because there may be confounders that are unaccounted for.
Conclusion
The efficacy of routine testing of the vaginal microbiome in the case of suspected preterm birth seems to be limited. We were able to demonstrate a lack of association between a swab-result guided antibiotic management protocol and an unfavorable maternal or fetal outcome in the setting of preterm birth risk between 24 and 34 gestational weeks. The detection of similar non-GBS bacteria in/on neonatal tissues and the vaginas of their mothers prior to the delivery was uncommon. These findings reaffirm the necessity of more stringent indications for unspecific vaginal microbiological diagnostics in the second half of the pregnancy.
Footnotes
Authors’ Contributions
MS and SS contributed to the study conception and design. Material preparation, data collection and analysis were performed by EW, RH, MG, and UK. Supervision and validation were conducted by MS, BT, and SB. Statistical analysis was performed by CW. The first draft of the manuscript was written by SS and all Authors commented on previous versions of the manuscript. All Authors read and approved the final manuscript.
Conflicts of Interest
The Authors declare that they have no competing interests in relation to this study.
- Received February 24, 2023.
- Revision received March 6, 2023.
- Accepted March 7, 2023.
- Copyright © 2023 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.