Changes in vaginal Ureaplasma and Lactobacillus due to antibiotic regimen for premature rupture of membranes

TL;DR

This study found that a standard antibiotic regimen for PPROM did not effectively reduce Ureaplasma levels and decreased Lactobacillus in the vagina.

Contribution

The study evaluates the effectiveness of azithromycin and ampicillin in reducing Ureaplasma and Mycoplasma in PPROM patients.

Findings

75% of patients who completed the antibiotic regimen had increased Ureaplasma levels.

The antibiotic regimen significantly decreased Lactobacillus spp. in most cases.

No association was found between Ureaplasma changes and neonatal sepsis or bronchopulmonary dysplasia.

Abstract

Preterm premature rupture of membranes (PPROM) is associated with preterm delivery and neonatal complications. PPROM is often complicated by intra-amniotic inflammation and/or microbial invasion of the amniotic cavity with Ureaplasma or Mycoplasma. Various prophylactic antibiotic therapies have been proposed to prolong latency between PPROM and delivery, reduce the risk of clinical chorioamnionitis, and improve neonatal complications. However, information on the potential of azithromycin administration to reduce the microbial load of vaginal Ureaplasma and Mycoplasma remains lacking. This prospective cohort study included singleton pregnancies managed with prophylactic antibiotics for PPROM at less than 36 weeks of gestation. All patients received the standard antibiotic regimen for PPROM, which consisted of a single oral azithromycin and intravenous ampicillin every for 2 days followed by 5 days of oral amoxicillin. Vaginal swabs samples were collected when PPROM was confirmed and after the antibiotic regimen administration. The main outcome measures were to investigate the changes in vaginal Ureaplasma, Mycoplasma, and Lactobacillus spp. due to the antibiotic regimen. In addition, the association between the presence and changes in vaginal Ureaplasma and Mycoplasma, pregnancy outcomes, and neonatal complications were examined. Out of 82 eligible PPROM, 51 had positive vaginal Ureaplasma. Thirty-six patients (52.2%) completed the antibiotic regimen. Among those with positive vaginal Ureaplasma who completed the antibiotic regimen, 75% experienced an increase in vaginal Ureaplasma levels. For those who delivered before completing all antibiotic doses, 40% had increased vaginal Ureaplasma levels. Furthermore, the antibiotic regimen resulted in decreased Lactobacillus spp. in almost all cases. It was suggested that azithromycin and ampicillin may not be effective when targeting Ureaplasma or Mycoplasma. Since this study did not search for resistance genes, it cannot be determined that azithromycin resistance in Ureaplasma or Mycoplasma is responsible for the present results. In addition, vaginal Ureaplasma changes were not found to be associated with neonatal sepsis or bronchopulmonary dysplasia. Future studies are needed to revalidate current antibiotic therapy for PPROM.

Figures

Flowchart of the study population.A total of 82 patients with PPROM were eligible for this study. Vaginal Ureaplasma was positive in 51 patients and negative in 31 patients. Of the patients who had positive vaginal Ureaplasma and completed the standard antibiotic regimen, 18 (75%) had increased vaginal Ureaplasma after treatment. Among patients with a positive vaginal Ureaplasma delivered before completion of all antibiotic doses, vaginal Ureaplasma increased in eight patients (40%). PPROM, preterm premature rupture of membranes; NEG, negative. *intravenous 2 g ampicillin every 6 hours for 2 days, a single oral dose of 1 g azithromycin, and 5 days of oral 500 mg amoxicillin every 8 hours.

The change of microbial load of vaginal Ureaplasma DNA before and after antibiotic regimen.There were many cases of increased vaginal Ureaplasma, especially in patients who had completed the antibiotic regimen. *intravenous ampicillin 2 g every 6 hours for 2 days and a single oral dose of azithromycin 1 g followed by 5 days of oral amoxicillin 500 mg every 8 hours.

The change of microbial load of vaginal Lactobacillus spp.DNA before and after antibiotic regimen. Almost all Lactobacillus spp. were decreased or disappeared after the antibiotic regimen, especially in the cases that completed the antibiotic regimen. *intravenous 2 g ampicillin every 6 hours for 2 days, a single oral dose of 1 g azithromycin, and 5 days of oral 500 mg amoxicillin every 8 hours.

Full text

Introduction

Preterm premature rupture of membranes (PPROM) occurs in 2% − 3% of pregnancies and is associated with preterm delivery, neonatal complications, and maternal sepsis.[1–4] PPROM is often complicated by intra-amniotic inflammation and/or microbial invasion of the amniotic cavity with Ureaplasma or Mycoplasma. [5] Ureaplasma is most frequently isolated from the amniotic fluid and placenta in preterm birth. [6,7] Ureaplasma or Mycoplasma colonization is associated with infertility, stillbirth, chorioamnionitis, neonatal bronchopulmonary dysplasia, meningitis, and perinatal death.[8–11] The vaginal microbiome shows a lower alpha diversity than the intestinal tract. Although there are differences in bacterial species among ethnicities, Lactobacillus species are often dominant in healthy people. [12]Lactic acid bacteria suppress the growth of pathogenic microorganisms through the production of lactic acid, hydrogen peroxide, etc. If the normal lactobacillus-dominant environment is disrupted, bacterial vaginosis, viral infections, and candidiasis can develop. Lactic acid bacteria preparations are being investigated as probiotics. [13]Ureaplasma spp. are typical urease-producing bacteria that are often detected in the human vagina. If Ureaplasma infection occurs during pregnancy, Ureaplasma causes chorioamnionitis and leads to miscarriage or premature birth. The urease degrades urea into ammonia and carbon dioxide, leading to an elevation in intravaginal pH. [14]The existence of Ureaplasma is also one of the indicators of an increase in intravaginal pH. We previously reported that the amount of ureaplasma DNA in the placenta of preterm infants is associated with small airway obstruction (SAO) in school-age children. [15]Interestingly, some premature infants who developed SAO did not show any respiratory problems during the neonatal period. Mycoplasma hominis is also a bacterium that is often detected from the vaginal swab, and not all but some of the M. hominis have arginine deiminase, which metabolizes arginine to citrulline and ammonia.[16]

Various prophylactic antibiotic regimens were proposed to prolong the latency between PPROM and delivery, reduce the risk of clinical chorioamnionitis, and improve neonatal outcomes. [17–20] The American College of Obstetricians and Gynecologists recommends intravenous erythromycin and ampicillin for two days, followed by oral erythromycin and amoxicillin for five days, as the most common regimen used in PPROM. [21]Ampicillin has been the first-line prophylactic antibiotic for PPROM for a long time. According to a systematic review of the efficacy of prophylactic antibiotics for PPROM, cephalosporins with or without macrolides were considered the primary prophylactic regimen. In addition, penicillin alone also prevented maternal chorioamnionitis. [22]The national Institute for Health and Care Excellence recommends oral erythromycin 250 mg four times a day for a maximum of ten days or until the woman is in established labor.[23] A regimen with azithromycin has been suggested as an alternative to erythromycin. [18,24,25]

At Osaka Women’s and Children’s Hospital, patients with PPROM receive a standard antibiotic regimen consisting of azithromycin and ampicillin. However, it is unclear whether administering azithromycin can reduce the microbial load of vaginal Ureaplasma and Mycoplasma DNA. Therefore, this prospective cohort study aimed to investigate the changes in vaginal Ureaplasma, Mycoplasma, and Lactobacillus, as well as maternal blood Ureaplasma, resulting from the antibiotic regimen used for PPROM cases. In addition, the association between the presence and changes in vaginal Ureaplasma and Mycoplasma, pregnancy outcomes, and neonatal complications were examined, and that aim was achieved.

Materials and Methods

This single-center, prospective cohort study included singleton pregnancies managed with prophylactic antibiotics for PPROM at less than 36 weeks of gestation between October 2019 and December 2021. Patients under the age of 20, those who received other antibiotics within 2 weeks of PPROM, and those allergic to ampicillin and azithromycin were excluded. All relevant national regulations and institutional policies were compiled, and the study was conducted under the tenets of the Helsinki Declaration. It was approved by the ethical review board of Osaka Women’s and Children’s Hospital (approval number, 1211; date of approval, June 3, 2019). Written informed consent was obtained from all study subjects. Recruitment of subjects for this study began on October 1, 2019 and ended on December 31, 2021. The standard deviation of the amount of vaginal Ureaplasma colonization was 10^2.7^. To have a 90% power to detect changes above 10^2^ colony-forming units/mL, the study aimed to enroll 80 patients.

The perinatal management protocol for PPROM at Osaka Women’s and Children’s Hospital is as follows. The diagnosis of PPROM is confirmed by observing amniotic fluid leakage from the cervix or pooling in the vagina, along with positive diagnostic tests measuring vaginal pH. In cases where the diagnosis is not clearly confirmed, an immunochromatographic strip test with insulin-like growth factor-binding protein-1 (ActimProma) is used to detect amniotic fluid in the vagina or ultrasound evaluation is used to confirm the presence of oligohydramnios. All patients were admitted to the hospital and received the standard antibiotic regimen for PPROM before 36 weeks of gestation with intravenous ampicillin (2 g dose every 6 hours) for 48 hours and single oral azithromycin (1 g dose) followed by oral amoxicillin (500 mg dose every 8 hours) for 5 days. Betamethasone was administrated for patients with gestational age less than 34 weeks and imminent delivery. Intravenous magnesium sulfate and/or ritodrine hydrochloride and/or oral nifedipine as tocolytic agents were administered when patients have short cervical length and frequent uterine contraction. Indications for delivery included clinical chorioamnionitis, onset of labor, nonreassuring fetal status, or gestational age beyond 36 weeks. Clinical chorioamnionitis was diagnosed based on the presence of maternal fever (≥38.0°C) and two or more clinical signs: maternal leukocytosis (white blood cell count > 15,000/mm^3^), maternal tachycardia (heart rate >  100 beats/min), fetal tachycardia (heart rate >  160 beats/min), uterine tenderness, and purulent or foul-smelling amniotic fluid or vaginal discharge. [26]

Vaginal swabs and maternal blood samples were collected from women when the diagnosis of PPROM was confirmed. The quantity of vaginal secretions was measured by weighing the swab tube before and after the collection. Vaginal swabs were taken from the lateral vaginal walls and the anterior and posterior fornix, and they were transferred to a cobas® polymerase chain reaction (PCR) Media Dual Swab Sample Kit from Roche Diagnostics GmbH, Germany. Vaginal swabs and maternal blood samples were retaken after the completion of the antibiotic regimen. Scrapings of the placental surface were also performed. If delivery occurred before the completion of the regimen, vaginal swabs and maternal blood samples were collected before the delivery. These results were not available to managing clinicians. The amount and species of Ureaplasma collected from the vaginal secretions, along with the amount of Lactobacillus spp. and Mycoplasma (M.) hominis and the presence of inerolysin (pore-forming toxin produced by Lactobacillus (L.) iners that can cause preterm labor) were determined using real-time PCR. The amount of Ureaplasma was also measured from maternal blood and placental surface. An increase or decrease of 10^2^ or more in DNA copy numbers was considered significant. To quantify the copy numbers in the samples, we first PCR amplified the interested gene, and then subcloned the products were into the pT7Blue T-vector (Novagen). The DNA amounts of the standard controls were calculated by Quantus fluorometer (Promega) using QuantiFluor® ONE dsDNA System. After calculating the number of molecules based on the molecular weight of the standards, the serial dilutions of the control DNA were prepared and used.

This study investigated neonatal and infant mortality and any of the following adverse perinatal outcomes for up to 6 months of age: neonatal intensive care unit admission, mechanical ventilation, respiratory distress syndrome, retinopathy of prematurity, intraventricular hemorrhage 3 or 4 grade, periventricular leukomalacia, necrotizing enterocolitis, neonatal sepsis, bronchopulmonary dysplasia (BPD), and congenital anomaly. BPD is a chronic lung disease, which is most commonly seen in premature infants requiring respiratory support, including oxygen supplementation at 36 weeks corrected gestational age. [27] Neonatal sepsis is defined as the presence of a positive blood culture. [28]

In this study, changes in vaginal Ureaplasma, Mycoplasma, and Lactobacillus are divided into two groups: complete and incomplete antibiotic protocols. Cases of complete protocol are defined as cases in which all antibiotic regimens had been completed. Cases of incomplete protocol are those who delivered in the middle of their antibiotic regimen, but all of them had received at least one intravenous ampicillin and oral azithromycin administration.

DNA extraction and quantitative real-time PCR of Ureaplasma

DNA was extracted from vaginal swab and maternal blood samples using a Maxwell® RSC Blood DNA Kit (AS1400, Promega, Japan). The DNA copy number of Ureaplasma was measured using QuantStudio 5Real-Time PCR System (Thermo Fisher Scientific) targeted for the 16S ribosomal RNA gene. The amplification reaction mixtures comprised 5 mL of TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific), primers (0.9 μM each), probe (Thermo Fisher Scientific, 0.25 μM), and 1 μL of DNA sample; this was adjusted to 10 μL with DDW. Nuclease-free water was used instead of template in the negative control. The primers and probe used were as follows: 16S-Urea F1(AGGCATGCGTCTAGGGTAGGA), 16S-Urea R1(ACGTTCTCGTAGGGATACCTTGTTA), and 16S-Urea-FAM-MGB probe (FAM-CGGTGACTGGAGTTAA-MGB).

The PCR conditions were as follows: predenaturation at 95°C for 20 s, 40 cycles of denaturation at 95°C for 1 s, and annealing at 58 C for 20 s.

The standard curve of Ureaplasma was determined using serial dilutions with copy numbers of 1.0 ×  10^1^ to 1.0 ×  10^6^ copies per reaction mixture by real-time PCR. In each run with clinical specimens, negative control and standards were included. To differentiate Ureaplasma (U.) parvum from U. urealyticum, the following primers and probes were used: ureG F2(CAACATTTAGTCCAGATTTAG), ureG R2(TAGCACCAACATAAGGAG), Up-FAM-MGB probe (FAM-TTGACCACCCTTACGAG-MGB), and Uu-VIC-MGB probe (VIC-TTGTCCGCCTTTACGAG-MGB). The amplification reaction mixtures comprised 5 mL of TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific), primers (0.9 μM each), probes (Thermo Fisher Scientific, 0.2 μM each), and 1 μL of sample DNA; this was adjusted to 10 μL with DDW.

Quantitative real-time PCR analysis of M. hominis

The primers and probe used were as follows:

Mh-yidC-F(ACCCGGTTTAGTGAGTTTGCT), Mh-yidC-R (CCTCAGTTTATTGCATTGCCA), and Mh-yidC-VIC probe (VIC-AACAAGCAACCTGATATT-MGB). The PCR conditions were as follows: predenaturation at 95°C for 20 s, 35 cycles of denaturation at 95°C for 1 s, and annealing at 60°C for 20 s.

Quantitative real-time PCR of Lactobacillus species and inerolysin detection

The primers for Lactobacillus spp. used were Lacto-F (TGGAAACAGRTGCTAATACCG) and Lacto-R (GTCCATTGTGGAAGATTCCC).[29] The primers for inerolysin were; inerolysin-F(CGTGGTCGTACATCAGGCTT) and inerolysin-R(TTCTCCACCATTCCCATGCC).

PCRs were performed in a reaction volume of 10 μL, each containing 0.5 µ M of inerolysin-F and inerolysin-R primers, 5 μL of PowerUP^TM^ SYBR^TM^ Green Master Mix (Thermo Fisher Scientific), and 1.0 μL of genomic DNA. PCR amplification was performed under the following conditions: initial cycle of 95°C for 2 min, 35 cycles of 95°C for 1 s, 58°C for 15 s, and 72°C for 60 s.

Statistical analysis

Categorical and continuous variables were analyzed using chi-square or Fisher’s exact tests and compared using Kruskal–Wallis test. Paired t-tests and Wilcoxon signed-rank sum tests were performed for changes in bacterial DNA levels before and after antibiotic regimen. Multivariate logistic regression analyses were performed to examine the adjusted odds ratios (aOR) and 95% confidence intervals (CI) for changes in Ureaplasma and perinatal outcomes. Statistical significance was set at P < 0.05. Two-sided P-values were reported, and analyses were conducted using a JMP software (version 14; SAS Institute, Cary, NC, USA).

Results

In total, 82 patients with PPROM were included, and 51 were positive for vaginal Ureaplasma. Thirty-six patients (52.2%) completed the standard antibiotic regimen. (Fig 1) Of those who had positive vaginal Ureaplasma and completed the standard antibiotic regimen, vaginal Ureaplasma increased in 18 patients (75%). Among patients with positive vaginal Ureaplasma who delivered before the completion of all antibiotic doses, vaginal Ureaplasma increased in eight patients (40%).

*Flowchart of the study population.A total of 82 patients with PPROM were eligible for this study. Vaginal Ureaplasma was positive in 51 patients and negative in 31 patients. Of the patients who had positive vaginal Ureaplasma and completed the standard antibiotic regimen, 18 (75%) had increased vaginal Ureaplasma after treatment. Among patients with a positive vaginal Ureaplasma delivered before completion of all antibiotic doses, vaginal Ureaplasma increased in eight patients (40%). PPROM, preterm premature rupture of membranes; NEG, negative. intravenous 2 g ampicillin every 6 hours for 2 days, a single oral dose of 1 g azithromycin, and 5 days of oral 500 mg amoxicillin every 8 hours.

Table 1 shows the maternal characteristics and perinatal findings classified into three groups based on vaginal Ureaplasma changes: Ureaplasma-negative before and after the antibiotic regimen, Ureaplasma-positive that decreased after the antibiotic regimen, and Ureaplasma-positive that increased after the antibiotic regimen.

Table 1: Maternal characteristics and perinatal findings, classified into three groups based on vaginal Ureaplasma changes by antibiotic treatment.

The median gestational week at PPROM and delivery did not differ among the three groups. However, latency from PPROM to delivery was significantly prolonged in the group with increased Ureaplasma after the antibiotic regimen. There was no difference in the incidence of histological chorioamnionitis (hCAM), but funisitis was significantly more common in the group with increased Ureaplasma after the antibiotic regimen (p =  0.02).

Fig 2 shows the changes in vaginal Ureaplasma spp. before and after the antibiotic regimen. The median microbial load of vaginal Ureaplasma DNA at the time

*The change of microbial load of vaginal Ureaplasma DNA before and after antibiotic regimen.There were many cases of increased vaginal Ureaplasma, especially in patients who had completed the antibiotic regimen. intravenous ampicillin 2 g every 6 hours for 2 days and a single oral dose of azithromycin 1 g followed by 5 days of oral amoxicillin 500 mg every 8 hours.

of PPROM was 1.9 ×  10^7^ (1.8 ×  10^2^ −  4.6 ×  10^9^). Many cases had increased vaginal Ureaplasma. In patients who completed the antibiotic regimen, Ureaplasma was significantly increased after antibiotic regimen. (P = 0.01) As shown in Fig. 3, Lactobacillus spp. decreased after the antibiotic regimen in almost all cases. Lactobacillus spp. DNA levels were significantly reduced by antimicrobials in both patients who completed and did not complete an antibiotic regimen (P < 0.001 for both groups). Table 2 and Table 3 shows the vaginal and maternal blood Ureaplasma, vaginal M. hominis, Lactobacillus spp., and inerolysin classified by vaginal Ureaplasma changes. All M. hominis-positive cases were positive for vaginal Ureaplasma. Three patients had a positive vaginal M. hominis result at the time of PPROM and did not decrease in all cases after antibiotic regimen, but rather turned positive in three new cases. The loss of Lactobacillus spp. was more common in patients who completed the antibiotic regimen, especially in the group with increased Ureaplasma. hCAM and funisitis were more common in those with completed antibiotic regimen, regardless of vaginal Ureaplasma changes due to the antibiotic regimen and presence of Ureaplasma at the time of PPROM (aOR,7.5; 95% CI; 2.1–26.6; P =  0.002, aOR4.5; 95%CI 1.4–11.4; P =  0.01, respectively). Among incomplete the antibiotic regimen, there were less funisitis in the group with decreased Ureaplasma due to antibiotic therapy than in the group with increased or negative Ureaplasma at the time of PPROM.

*The change of microbial load of vaginal Lactobacillus spp.DNA before and after antibiotic regimen. Almost all Lactobacillus spp. were decreased or disappeared after the antibiotic regimen, especially in the cases that completed the antibiotic regimen. intravenous 2 g ampicillin every 6 hours for 2 days, a single oral dose of 1 g azithromycin, and 5 days of oral 500 mg amoxicillin every 8 hours.

Table 2: Changes of Ureaplasma, Mycoplasma hominis, and Lactobacillus spp. in completed antibiotic regimen group.

Table 3: Changes of Ureaplasma, Mycoplasma hominis, and Lactobacillus spp. in uncompleted antibiotic regimen group.

Table 4 shows the neonatal complications according to vaginal Ureaplasma changes. Vaginal Ureaplasma changes were not associated with neonatal intensive care unit admission, neonatal sepsis, mechanical ventilation, or BPD. However, BPD was more frequent in those who completed protocol than in those who did not complete the protocol, regardless of vaginal Ureaplasma changes due to the antibiotic regimen, presence of Ureaplasma at the time of PPROM, and gestational age at delivery. (aOR, 7.2; 95%CI 1.3–40.1; *P = * 0.02).

Table 4: The neonatal complications according to vaginal Ureaplasma changes due to the antibiotic regimen.

Discussion

Almost all Lactobacillus spp. decreased or disappeared, while most vaginal Ureaplasma and M. hominis increased after the antibiotic regimen, particularly in patients who completed the regimen. Patients with increased vaginal Ureaplasma had longer latency. This result did not indicate that cases of increased Ureaplasma would have a longer latency period. Rather, the longer period reflected Ureaplasma that was not reduced by antibiotic therapy grew intravaginally. The longer latency from PPROM may increase the likelihood of ascending infection of Ureaplasma to the amnion despite the antibiotic therapy. The possibility of avoiding invasive amniocentesis was considered if Ureaplasma could be detected in maternal blood in the presence of intrauterine Ureaplasma infection. However, it was impractical because Ureaplasma in maternal blood was undetectable in most cases.

Antibiotics were recommended for women with PPROM to prolong the delivery latency interval and reduce chorioamnionitis and neonatal sepsis.[18,19,21,30] A regimen of intravenous ampicillin and erythromycin followed by oral amoxicillin and erythromycin has been commonly used. [18] Recently, azithromycin has been used instead of erythromycin because of its low cost, better side effect profile, ease of administration, and antibiotic resistance. Azithromycin was associated with a similar latency interval and lower rate of clinical chorioamnionitis for women with PPROM.[25,31] Tanaka reported the eradication of Ureaplasma in the amniotic fluid of patients with PPROM treated with intravenous ampicillin–sulbactam and oral azithromycin.[32] Kacerovsky showed that intravenous clarithromycin was related to the reduced Ureaplasma DNA load in the amniotic fluid of patients with PPROM.[20] Conversely, Gomez reported that antibiotic administration (ceftriaxone, clindamycin, and erythromycin for 10–14 days) rarely eradicates intra-amniotic infection in patients with preterm PROM.[33] A report showed that azithromycin effectively eradicated respiratory tract Ureaplasma colonization in preterm infants. [34] In this study, most vaginal Ureaplasma was increased after the antibiotic regimen. The differences in results from similar papers may be related to the different types of antibiotics and the duration of administration in this study. Although the ampicillin and erythromycin or azithromycin protocols are relatively widely used, it has been suggested that they may not be effective when targeting Ureaplasma or Mycoplasma. Several studies have been reported investigating the antimicrobial susceptibility patterns of genital Mycoplasmas and Ureaplasma in pregnant women. The resistance of isolated Ureaplasma spp. to erythromycin was 80%, and it was increased to 97% in mixed Ureaplasma and Mycoplasma infections. [35] In recent years, increasing rates of resistance of Mycoplasma and Ureaplasma to macrolide antimicrobials have been reported. [36]According to global analysis on the mutations associated with multidrug-resistant urogenital Mycoplasmas and Ureaplasma infection, macrolide resistance showed a marked surge, increasing from 25.6% in the period 2003–2015 to 56.6% in 2020–2023.[37] Su Zhang et al. had reported that the drug resistance rates in vitro of U. urealyticum and M. hominis showed resistance to azithromycin and erythromycin were 19.50% and 47.00%,respectively. [38] In a 2021 survey conducted in central China, the resistance rates of U. urealyticum and M. hominis to azithromycin were 65.79% and 88.89%, respectively.[39] In previous antimicrobial resistance studies of Mycoplasma pneumoniae, the A2063G mutation in 23S rRNA allows high-level macrolide resistance. On the other hand, regarding antimicrobial resistance in Ureaplasma, 23S rRNA C2243T mutant strains have been reported to be highly resistant to the 15-membered-ring macrolide, azithromycin. On the other hand, 23S rRNA A2149C and A2181T double mutants showed intermediate susceptibility to azithromycin. [40] In addition, it has been reported that S21A in the L4 protein, G2654T and T2245C in 23S rRNA, and the ermB gene were identified in erythromycin-resistant Ureaplasma spp. [41] Based on these findings, it is presumed that 23S rRNA mutations play a major role in macrolide resistance in Ureaplasma. Since this study did not search for resistance genes, it cannot be determined that azithromycin resistance in Ureaplasma or Mycoplasma is responsible for the present results.

In this study, there was an increased frequency of hCAM, funisitis, and BPD in cases that completed the antibiotic regimen. An increased vaginal Ureaplasma increased the frequency of funisitis but not the incidence of neonatal infection and BPD. These suggest that even if the antibiotic regimen contributes to a prolonged gestational age, it may increase the frequency of intrauterine infection or neonatal infection from proliferated Ureaplasma and decreased Lactobacillus spp. Many studies presented that Ureaplasma was associated with adverse pregnancy outcomes and neonatal morbidities, such as preterm delivery, PPROM, chorioamnionitis, BPD,[9,42] intraventricular hemorrhage,[43] and necrotizing enterocolitis.[44] Macrolides have anti-inflammatory properties that might reduce lung damage in preterm infants and treat bacterial infections[45]; they reduce BPD in preterm infants. [46,47] In this study, respiratory samples from infants were not obtained for Ureaplasma culture, so it was unknown whether the infants had Ureaplasma infection. The sample size was insufficient to validate the neonatal morbidity, so further research is needed.

Lactobacillus, the most frequently isolated microorganism from a healthy human vagina, is touted to prevent the invasion of pathogens and ensure vaginal epithelial homeostasis. [48] The loss of Lactobacillus and increase in Gram-negative rods, such as Gardnerella vaginalis, lead to bacterial vaginosis. [49] In a recent study*, L. crispatus*, L. gasseri, L. jensenii, and L. iners, were the most frequently isolated in the vaginas of healthy childbearing age or pregnant women. [50] A vaginal microbiota that is rich in L. crispatus, L. gasseri, and L. jensenii is often associated with a lower risk of bacterial vaginosis and preterm birth. Conversely, L. iners may play a role in the pathogenesis of bacterial vaginosis associated with preterm birth.[51,52] L. iners alone detected in vaginal smears of healthy women in early pregnancy might be related to preterm delivery.[51] L. iners secretes the cholesterol-dependent cytolysins (CDC), inerolysin, as well as one of the CDC family, vaginolysin, produced by G. vaginalis. [53] In this study, about half of the cases had inerolysin, which may be related to the fact that the subjects in this study were cases of PPROM. There was no difference in the increase in vaginal Ureaplasma with or without inerolysin. Lactobacillus spp. was decreased after the antibiotics therapy regardless of the presence or absence of inerolysin. It is unclear whether Lactobacillus spp., reduced by antibiotic therapy, will recover spontaneously. It may be challenging to reduce Ureaplasma with this antibiotic therapy; we may rather focus on Lactobacillus formulations.

The strength of this study is that it is the first to investigate both vaginal Ureaplasma and Lactobacillus changes caused by ampicillin and azithromycin regimens. This was also an examination at a single institution with a uniform management policy. However, it has several limitations. In this study, vaginal cultures were collected rather than from amniotic fluid through transabdominal amniocentesis, which does not directly indicate amniotic fluid infection. There were three reasons why the specimens were vaginal secretions. First, the purpose of this study was to observe changes in both Ureaplasma and Lactobacillus due to antibiotic administration. Second, in PPROM cases, amniotic fluid testing after antibiotic administration might not be available due to the almost complete loss of amniotic fluid in the uterine cavity or the rapid onset of labor. Third, amniocentesis was invasive. In almost all vaginal *Ureaplasma-*positive cases, Ureaplasma was found on the surface of the placenta, regardless of the delivery mode and vaginal Ureaplasma changes due to the antibiotic regimen. Although there may be contaminations during delivery in PPROM cases, most vaginal Ureaplasma may be ascending into amniotic fluid or placenta. Another limitation was that the vaginal mucosa was scraped and vaginal secretions were collected, but there was a possibility of contamination with amniotic fluid and a decrease in vaginal mucus, especially when collecting specimens at the time of PPROM. In addition, the sample size was insufficient to validate the neonatal morbidity.

Conclusions

Almost all Lactobacillus spp. decreased while most vaginal Ureaplasma and M. hominis increased after the antibiotic regimen. This suggests that Ureaplasma and M. hominis became resistant to azithromycin. Future studies are needed to revalidate current antibiotic therapy for PPROM because azithromycin-resistant Ureaplasma and M. hominis are widespread in the vaginal tract.