Introduction
              Methicillin-resistant Staphylococcus aureus (MRSA) infections are exceedingly difficult to treat and sometimes have serious consequences, especially when they occur in a compromised host. Pregnant women are compromised hosts, and MRSA infection is a cause of severe diseases in pregnant women1,2. The most important compromised hosts in perinatal care are neonates, especially premature babies, who require admission to the neonatal intensive care unit (NICU). When these populations become infected with MRSA, they easily develop serious complications3-5. Therefore, MRSA infection in pregnant women and neonates has significant clinical impact in the obstetric field6.
              MRSA infection is difficult to treat after it develops. Therefore, great importance is placed on the prevention of MRSA outbreaks. Healthcare providers have the most important role in preventing MRSA, especially in the NICU7. Although the role  of healthcare providers remains important, there is no apparent association between the amount of effort that healthcare providers apply towards preventing MRSA transmission in neonates and the actual prevention of MRSA outbreaks in the NICU8. The prevention of MRSA transmission caused by neonates admitted to the NICU is also important. In Japan, NICU-admitted neonates are commonly screened for MRSA by the collection of cultures from various sites. However, it takes several days to obtain culture results. If neonates who are newly admitted to the NICU are MRSA carriers, MRSA may spread through the NICU during this waiting period. Therefore, it is extremely useful to obtain information regarding MRSA carriage before a neonate’s admission to the NICU. Unfortunately, such information is not available.
              MRSA colonization in mothers is sometimes found accidentally, and there is no clinical protocol for management. MRSA is an indigenous bacterium of the skin, and it may be vertically transmitted in a manner similar to Group B streptococcus ( GBS)9,10. Therefore, information on maternal MRSA colonization may be useful for estimating the MRSA status of neonates. However, there is no consensus regarding the frequency of vertical transmission of MRSA at full-term delivery based on  large-scale data.
              We performed a prospective cohort study of the vertical transmission of MRSA conducted in Tajima area (Prevalence-in-TAJIMR study) in Japan to investigate the prevalence of MRSA and its clinical impacts on pregnant women and their full-term neonates.
 
Material and Methods
Study design, period and participants
              The Prevalence-in-TAJIMR study was an observational, prospective cohort study conducted in the Tajima area, which is the northern part of the Hyogo Prefecture in Japan. Pregnant women who were >18 years of age and at 32 to 36 weeks of gestation who visited our center were recruited from August 2016 to December 2017. Women who were capable of providing written informed consent were enrolled in this study. MRSA screening was performed at the time of enrollment. The mothers and their neonates were followed until day 7 after birth and were assessed for adverse outcomes. Although MRSA may be detected in various sites, it is commonly detected in the nares11,12. We screened for maternal MRSA colonization using nasal and vaginal samples in our study. We used nasal and umbilical surface swabs for neonatal screening because these sites had the highest detection efficiency in previous reports11,12. All samples were collected by midwives and doctors at enrollment and immediately after delivery using a standardized sterile procedure. If MRSA was detected in the cultures of the fetal samples collected immediately after delivery, it should be considered almost entirely due to vertical transmission. The sample collection and MRSA detection methods are described in detail in the supporting information texts.
 
Study outcomes
              The primary outcome of the present study was the prevalence of neonatal MRSA colonization at birth. Therefore, we collected umbilical surface swabs and nasal swabs from infants at the time of delivery. We also investigated the cumulative incidence of maternal and neonatal events during the perinatal period as the secondary outcome. We investigated the following maternal events: preterm delivery before 37 completed weeks of gestation, cesarean section (C/S), prelabor rupture of the membranes (PROM), prolonged PROM duration (over 24 hours), and vacuum or forceps delivery. We investigated the following neonatal events: pH of the umbilical cord arterial blood, Apgar score (1 min and 5 min), birth weight, admission to the NICU, and development of SSTIs, which may be less serious but are potentially associated with MRSA. The development of SSTIs in neonates were also assessed. All neonates were observed daily by neonatologists. When the neonate had abnormal skin lesions, the neonatologist diagnosed and classified all symptoms as invasive (eczema, impetigo and abscess) or noninvasive (rash, inflammation). Swabs were collected from the lesions and cultured to determine the causative organism if such exist.
 
Data analysis
              The clinical data were collected from our medical records and in-person interview questionnaires. In this study, we defined “multiparous woman” as those who had given birth at least once13. Other factors we analyzed are described in detail in the supporting information texts.
              The chi-squared test or Fisher’s exact test was used to examine associations between categorical variables, and the unpaired t-test was used to examine  associations between continuous variables. A multivariable logistic regression model was used to determine the independent predictors of the primary outcome (prevalence of neonatal MRSA colonization at birth) and secondary outcomes. Statistical significance was considered when p<0.05. Odds ratios (ORs) and 95% confidence intervals (95% CI) were calculated, and  variables were considered significantly associated with the outcomes when the CI did not include 1.0. The statistical methods are described in detail in the supporting information.
 
Ethical approval
              The Institutional Review Board of Toyooka Public Hospital approved this study (reference number: 109, approval: June 22, 2016). Written consent was obtained from each pregnant woman prior to her participation in the study.
 
Results
Characteristics of the study population
              Figure 1 shows the flow chart of the study. In brief, 898 pregnant women out of 1223 pregnant women who visited our hospital agreed to participate and were enrolled in the study. The clinical characteristics are shown in supporting information Table S1. 55 of the pregnant women (6.1%) were colonized with MRSA according to nasal or vaginal swabs. Vaginal MRSA was detected in nine cases (1.0%). We categorized these 55 MRSA-positive pregnant women as the MRSA-group, and the remaining 843 pregnant women as the no-MRSA group. These women were followed prospectively, and there was no loss to follow-up.
              All mothers had live births, and there were no perinatal deaths during the study period. There were seven twin pregnancies, one in the MRSA-group and six in the no-MRSA group, resulting in a total of 905 neonates (56 neonates in the MRSA-positive group and 849 neonates in the non-MRSA group, Figure1). We monitored the clinical course of all neonates throughout the study period.
 
Risk factors for maternal MRSA colonization
              We first examined the differences in maternal backgrounds between the MRSA-group and the no-MRSA group (Table 1). Among the maternal background information, the percentage of multiparous mothers and mothers who worked as healthcare providers was significantly higher in the MRSA-group than in the no-MRSA group (MRSA-group:37/55, 67.3% vs. no-MRSA group: 408/843, 48.4%, OR: 2.09, 95% CI: 1.23 to 3.01 and MRSA-group:16/55, 29.1% vs. no-MRSA group: 116/843, 13.8%,OR: 2.57, 95% CI: 1.39 to 4.75, respectively). The frequency of a parity of two or more was also significantly higher in the MRSA-group than in  the no-MRSA group (15/55: 27.3% vs. 119/843: 14.1%, OR: 2.28, 95% CI: 1.22 to 4.26). Although it was not significant, the proportion of mothers who used antibiotics during pregnancy tended to be higher in the MRSA-group.
              Multivariable logistic regression analysis revealed that multiparity and occupation (healthcare provider) were independent risk factors for maternal MRSA colonization (OR: 2.35, 95% CI; 1.25-4.42 and OR: 2.58, 95% CI; 1.39-4.79, respectively).
              To analyze the association between maternal background and MRSA colonization in the vagina, we compared the maternal background of the vaginal MRSA-positive patients (vaginal MRSA group, n=9) and the vaginal MRSA-negative patients (no-vaginal MRSA group, n=889, Table 1). The multiparity rate was significantly higher in the vaginal MRSA group than in the no-vaginal MRSA group (9/9: 100% vs. 436/889: 49.0%, OR: 19.74, 95% CI: 1.15 to 340.4). The proportion of women with a parity of two or more was also significantly higher in the vaginal MRSA group than in the no-vaginal MRSA group (6/9: 66.6% vs. 128/889: 14.4%, OR: 11.89, 95% CI: 2.94 to 48.16). There was no significant difference in the proportion of patients who worked as healthcare providers (Table 1).
              These analyses showed that being multiparous and working as a healthcare provider were independent risk factors for maternal MRSA colonization. Multiparity, particularly a parity of two or more, was an obvious significant risk factor for vaginal and maternal MRSA colonization.
 
Vertical transmission of MRSA to neonates at birth
              Of the 905 neonates delivered in the present study, MRSA was detected in the umbilical surface swabs or nasal cavity swabs taken at the time of delivery in eight neonates (0.8%) (Table 2A). We assessed the primary outcome of this study, the vertical transmission of MRSA to neonates during delivery and analyzed this outcome in the mother-neonate pairs.
              Maternal MRSA carriers had a significantly higher risk of vertical transmission  at the time of delivery (MRSA-group; 7/55: 12.7% vs no-MRSA group; 1/843: 0.12%, OR 124; 95% CI 20.3-1393, Table 2A).
              We found that the vertical transmission of MRSA occurred significantly more often in mothers with MRSA in only the nasal cavity than in the no-MRSA group (3/46, 6.38%, OR: 59.2, 95% CI: 6.0-581). Vertical transmission occurred even more frequently in mothers with MRSA in the vagina than in the no-MRSA group (4/9, 44.4%, OR: 678.4, 95% CI: 64.0-7197). Considering MRSA colonization in the vagina, the vertical transmission in the vaginal MRSA-group was also significantly higher than that in the no-vaginal MRSA group (vaginal MRSA group: 4/9, 44.4%, no-vaginal MRSA group: 4/896, 0.45%; OR: 178, 95% CI 39.9-782, Table 2A). The multivariable regression model analysis revealed that maternal MRSA colonization remained significantly associated with vertical transmission (Table 2B).
               A stratified analysis showed that no factors increased the vertical transmission rate at birth. None of the neonates born via cesarean section had MRSA colonization, which suggests that MRSA is transmitted directly by vaginal delivery (supporting information Table S2).
              Maternal MRSA colonization was an independent risk factor for vertical transmission in the present analysis. Even when MRSA was detected only in the maternal nasal cavity, 6.38% (3/46) of the neonates born to these mothers had MRSA colonization. Among the pregnant women who had MRSA colonization in the vagina, the vertical transmission rate of MRSA increased to 44.4% (4/9) at the time of delivery (Table 2A).
 
Association between maternal and neonatal adverse outcomes and MRSA colonization
              We assessed the relationship between maternal MRSA colonization and the clinical courses of mothers and their neonates. We did not find any significant difference in maternal adverse outcomes in maternal MRSA colonization (supporting information Table S3). However, the frequency of SSTIs during the neonatal period was significantly higher in the neonates in the MRSA-group than in the neonates in the no-MRSA group (MRSA-group: 3/56, 5.4%; no-MRSA group: 7/849, 0.82%; OR: 7.47, 95% CI: 2.50-22.3).
              We performed the same analysis based on MRSA colonization in the vagina (Table 3). We found that maternal vaginal MRSA colonization was also significantly associated with the occurrence of SSTIs during the neonatal period (vaginal MRSA group: 2/9, 22.2%; no-vaginal MRSA group: 8/896, 0.89%; OR:31.7 95% CI: 5.68-177). The vaginal MRSA group had a significantly higher frequency of SSTIs during the neonatal period than the no-MRSA group did(vaginal MRSA group: 2/9, 22.2%; no-MRSA group: 7/849, 0.82%; OR: 34.4 95% CI: 6.04-195, Table 3). No factors other than maternal MRSA colonization were significantly associated with SSTIs (data not shown).
              Table 4 shows the details for all neonatal SSTIs. MRSA was the causative bacterium in all three cases in the MRSA-group but only one of the seven cases in the no-MRSA group (p=0.033). Two of the three cases in the MRSA-group developed symptoms on the day after birth, and only one of the seven cases in the no-MRSA group showed symptoms on the day after birth. We classified all symptoms as invasive (eczema, impetigo and abscess) or noninvasive (rash, inflammation). All three cases in the MRSA-group and one case in the no-MRSA group showed invasive symptoms, and the MRSA-group had a significantly higher frequency of invasive symptoms (p=0.033).
              Most SSTIs resolved spontaneously within a few days or resolved with simple local disinfection procedures. However, one case of MRSA SSTI in the MRSA-group did not resolve with local treatment and required incision and drainage procedures.
 
Discussion
              The Prevalence-in-TAJIMR study is the largest study to date in Japan to examine the actual proportion of MRSA carriers among pregnant women. It found a proportion of 6.1%; in particular, the proportion of vaginal MRSA carriers was 1%.
              The prevalence of MRSA carriers can vary across time and regions6, 14, 15. Our hospital is the only perinatal medical center within 2,000 km2, and approximately 90% of deliveries in this area are performed in our facility. Therefore, we anticipate that this study nearly accurately reflects the proportion of MRSA carriers among pregnant women living in this area. As of 2018, the prevalence rate of community-acquired MRSA was estimated to be approximately 6% in Japan.
              Among the clinical background characteristics related to MRSA carriers in pregnant women, being multiparous and being a healthcare provider were found to be independent risk factors. Several other studies have also shown that multiparous women and healthcare providers are at greater risk of being MRSA carriers7, 16, 17. It is still unclear why the proportion of MRSA carriers is higher among multiparous women. We assume that these women visit the hospital more frequently for obstetrical and neonatal issues, which might increase their probability of acquiring MRSA.
              The high rate of vertical transmission found in our study differs from the results of previous studies that reported the vertical transmission of MRSA from mothers to their neonates16, 18, 22, 24-26. In these reports, the vertical transmission rate of MRSA was as high as 10%, and no study reported a rate as high as the 40% observed in this study. This inconsistent observation may in part be due to the small sample size in our study, in which only nine pregnant women had MRSA detected in the vagina. Another possible reason for the high rate of vertical MRSA transmission in our study may be that since the cultures were taken immediately after delivery, a large amount of maternal vaginal secretions could remain, leading to a high detection rate. However, such cases should be included within vertical transmission rates and not be considered  a false-positive result. Additionally, the vertical transmission rate of GBS by delivery is reported to be approximately 40% 23, 24. Taking those factors into account, a vertical transmission rate of 44.4% (4/9 pairs) for vaginal MRSA may be reasonable.
Consistent with several previous reports, our study found that maternal MRSA carrier status had little influence on maternal and neonatal adverse outcomes18, 19. Interestingly, however, maternal MRSA carrier status was found to be associated with SSTIs during the neonatal period. Although MRSA infections in preterm neonates are known to be associated with poor prognostic outcomes3, the clinical significance of MRSA infection in full-term neonates has not been well studied thus far. We revealed for the first time that the vertical transmission of MRSA was associated with an increased occurrence of SSTIs in full-term neonates.
              Given that MRSA is generally found in the epidermis, we think this is a plausible observation and consistent with other reports7, 20. MRSA was the causative bacterium in the neonates delivered by the three mothers who were MRSA carriers (Table 4). There was a tendency for MRSA infections to occur shortly after birth, sometimes on the day of birth. These three patients showed moderate to severe symptoms such as abscess. Moreover, one patient required incision and drainage of skin infections. On the other hand, all patients with SSTIs who were delivered by non-MRSA carriers showed relatively mild symptoms. When an SSTI with moderate to severe symptoms, such as abscess, occurs immediately after birth, we may consider MRSA the causative bacterium and assume that the mother is a MRSA carrier.
              There are some limitations in this study. As described above, the number of MRSA-positive cases was small. While the variables were selected based on our clinical hypotheses, our small sample size limited the power to conduct multivariable analyses. Thus, there is a possibility of type 2 error in several of the multivariable regression model analyses conducted in this study. Our small sample size  also limited the power to conduct multivariable analyses. Nevertheless, the data obtained from this prospective study, which involved nearly 1,000 pregnant women, are considered valuable even if the number of MRSA-positive participants was small. Another limitation of the study is that it was not blinded. However, all of the neonates were seen daily by neonatologists, regardless of the mother’s MRSA status. We think that the fact that this was not a blinded study does not pose a major problem.
                           
Conclusion
              In conclusion, the current Prevalence-in-TAJIMAR study was the first to report that approximately 6% of healthy pregnant women had community-acquired MRSA colonization, and 1% had vaginal MRSA colonization. MRSA was highly transmitted to neonates by vertical transmission during delivery. In particular, when the mother had vaginal MRSA, vertical transmission occurred  in four out of nine cases (approximately 44%), which is a fairly high rate. MRSA transmission can cause SSTIs in full-term neonates. For mothers whose neonates are expected to enter the NICU, such as those with preterm birth, checking their MRSA colonization status should be useful. Additionally, if a neonate is to be admitted to the NICU suddenly and the mother is a MRSA carrier, the possibility of vertical MRSA transmission should be taken into account. At present, there is no concrete consensus regarding strategies for when MRSA is detected in the vaginal secretions of pregnant women. Our study shed light on the clinical significance of maternal MRSA colonization.