To estimate the prevalence of maternal colonizers in South Asia and their potential to colonize the umbilicus, an important precondition causing neonatal sepsis.
We conducted a cross-sectional study at a maternity center in Dhaka with 1219 pregnant women and a subset of 152 newborns from 15 January to 31 October 2011. During labor, study paramedics collected vaginal swabs for bacterial culture and rectal swabs for Group B Streptococcus (GBS) testing. Community health workers collected neonatal umbilical swabs. Log-binomial regression models were used to estimate risk ratios.
In all, 454 women (37.2%, 95% confidence interval (CI) 34.5 to 40.0%) were colonized. The most common organisms isolated were Staphylococcus aureus, Non-GBS and GBS. A total of 94 women (7.7%, 95% CI 6.2 to 9.2%) were colonized with GBS. The risk of GBS umbilical colonization was higher (RR=12.98, 95% CI 3.97 to 42.64) among newborns of mothers with GBS colonization.
Newborns of mothers colonized with GBS are at higher risk of developing umbilical colonization.
Common colonizers of the maternal reproductive tract are associated with early-onset neonatal sepsis.1 In Bangladesh during the first week of life, the incidence of clinical sepsis defined by the WHO (World Health Organization) Young Infants criteria for very severe disease2 was 13.4% with a case fatality of 10.2%.3 Studies in South Asia have examined the prevalence of maternal colonization, although there have been no studies in Bangladesh.4, 5, 6, 7, 8, 9, 10, 11, 12 Maternal recto-vaginal colonization with Group B Streptococcus (GBS) is associated with early-onset GBS neonatal sepsis in the United States.13, 14 However, based on the limited data, it appears that Group B Streptococcal sepsis in newborns is uncommon in South Asia.15 Not much is known about the prevalence of this organism in maternal recto-vaginal colonization, hereafter referred to as maternal colonization, in this setting.
We know that newborns come in direct contact with bacterial flora in the vaginal canal and perineum during labor and delivery, in which case newborns may acquire infections through the mouth, the umbilicus or a crack in the skin. Furthermore, ascending infections from the mother to the fetus may occur during labor, when colonized organisms from the maternal perineum spread through the vaginal canal to the placenta, and into the once-sterile amniotic fluid.16 The amniotic fluid, which bathes the neonate, also circulates through the neonate’s lungs and intestinal tract, which are potential hot spots for bacterial translocation.13, 17
We conducted this study to determine the prevalences and etiologies of maternal colonization, to understand the factors associated with maternal colonization, to estimate the risk of neonatal umbilical colonization associated with maternal colonization and to describe neonatal umbilical colonization patterns during the first 7 days of life.
To determine the prevalence of maternal colonization during labor and the characteristics associated with colonization, we conducted a cross-sectional study at a maternity center in Dhaka, Bangladesh with 1219 pregnant women from 15 January 2011 to 31 October 2011.
We followed a subgroup of 152 newborns (newborns who delivered vaginally during the last 2 months of the study) to determine neonatal colonization patterns.
Pregnant women presenting to the maternity center were screened for eligibility. Our inclusion criteria were women of 30 gestational weeks or later who planned to deliver at the maternity center. Women who presented with obstructed labor, hemorrhage, severe pre-eclampsia or fetal distress were excluded to facilitate their need for urgent care. Women who used antibiotics or steroids within 2 weeks of labor were also excluded from the study.
Trained study paramedics collected vaginal and rectal swabs to measure maternal bacterial colonization during labor. Women with a positive bacterial vaginal culture (Staphylococcus aureus, Non-GBS species, GBS, Klebsiella pneumoniae, Escherichia coli, Staphylococcus, Pseudomonas or Acinetobacter) or positive GBS rectal culture were classified as colonized. Trained community health workers collected umbilical swabs from newborns within 12 h after birth and on the seventh day of life.
Study staff interviewed the enrolled women to collect demographic data on maternal age; maternal education; antenatal care provider type; receipt of tetanus toxoid during antenatal care visits; type of housing materials for the women’s roof, wall and floor; source of water supply; sanitation facility; household number; household number under 5 years; number of rooms where household members sleep; neonatal sex; gestational age and birth weight. Month of birth was used to determine seasonality: summer (March to May), rainy (June to September) and winter (October to February).
During labor, study paramedics observed and collected data on maternal risk factors (duration of labor, duration of rupture of membranes (ROM), premature rupture of membranes (PROM), number of vaginal exams performed by facility health workers) and clinical signs of maternal infection (axillary temperature). Information on maternal risk factors and clinical signs was collected at the time the vaginal and rectal swabs were collected.
This study received ethical approval from the Johns Hopkins Bloomberg School of Public Health Committee on Human Research and the International Centre for Diarrheal Disease Research, Bangladesh Ethical Review Committee. All study participants provided informed consent.
Swabs were transported daily from the maternity center to the microbiology laboratory at Dhaka Shishu Hospital using Amies transport medium and processed within 24 h of collection. Primary inoculation of the vaginal swabs was done on MacConkey and blood agar media with 5% sheep blood and incubated at 37 °C for 18 to 24 h. Identification of organisms was done by Gram staining and standard biochemical reactions. In addition, we tested specifically for GBS by placing vaginal and rectal swabs into a Todd-Hewitt broth (2 ml) supplemented with gentamicin (8 μg/ml) and nalidixic acid (15 μg/ml). The broth was incubated at 37 °C for 18 h and then inoculated onto 5% sheep blood agar and incubated overnight at 37 °C. Suspected colonies of Gram-positive β-hemolytic colonies in the blood agar plate were identified as GBS by CAMP-positive test and group B antigen latex agglutination test by a Streptococcal grouping kit (Oxoid, Basingstoke, Hants, UK). Antimicrobial susceptibility tests were performed only for GBS using disk diffusion methods according to the Clinical and Laboratory Standards Institute guidelines.
The total number of women with a positive bacterial vaginal culture or a positive GBS rectal culture was divided by the total number of women with samples to produce a crude estimate of the overall maternal colonization prevalence and confidence interval (CI). Log-binomial regression models were used to determine the risk ratios for maternal colonization based on demographics, neonatal characteristics, maternal risk factors and maternal clinical signs of infection. We repeated the analyses to examine pathogen-specific associations.
The total number of newborns with a positive bacterial umbilical culture was divided by the total number of newborns with samples to produce a crude estimate of the prevalence and CI of neonatal umbilical colonization. To determine the risk of neonatal umbilical colonization associated with maternal colonization, we used log-binomial regression models. We repeated the analyses to examine pathogen-specific associations. We used a McNemar’s test to test the difference between the paired proportions of newborns colonized with specific pathogens on day 1 compared with day 7. Two sided P-values were calculated using a cumulative binomial distribution given the relatively small sample size. Analyses were performed using the STATA v12 (StataCorp, College Station, TX, USA).
We enrolled 1235 women. During the screening period, 1441 women were eligible for enrollment of which 108 (7.5%) were not interested, 58 (4.0%) presented for an elective C-section and wished not to participate, 24 (1.7%) had a family member who did not wish to participate, 8 (0.6%) were excluded because of maternal complications, and 8 (0.6%) were missed due to the paramedics workload.
Maternal colonization prevalences and etiologies
Of the 1219 women with colonization data, 454 (37.2%, 95% CI 34.5 to 40.0%) were colonized during labor. Three hundred eighty-eight women were singly colonized and sixty-six women were colonized with two or more organisms. Among the 388 women who were singly colonized, the most common organisms were S. aureus (n=101, 26.0%), Non-GBS Streptococcus (n=85, 21.9%), GBS (n=71, 18.3%), E. coli (n=43, 11.1%), and K. pneumoniae (n=33, 8.5%) (Table 1). Among the 66 women who were co-colonized, the most common combinations were E. coli and K. pneumonia (n=15, 22.7%), E. coli and GBS (n=5, 7.6%), and S. aureus and GBS (n=5, 7.6%). A total of 94 (7.71%; 95% CI 6.2 to 9.2%) women (71 singly colonized and 23 co-colonized) were colonized with GBS. Among them 66 (70.2%) were identified by normal culture and an additional 28 (29.8%) by selective broth media.
More than 97% of GBS isolates were sensitive to penicillin, cephalexin, ceftriaxone and imipenem. There were high rates of intermediate or complete resistance to ampicillin (10.1%), chloramphenicol (20.2%), cotrimoxazole (83.2%), ciprofloxacin (19.1%), gentamicin (97.8%), netilmicin (75.3%), erythromycin (23.6%) and azithromycin (24.7%).
Maternal colonization and demographics, neonatal characteristics, maternal risk factors and clinical signs
Colonization status was not associated with: maternal age; maternal education; antenatal care provider type; receipt of tetanus toxoid during antenatal care visits; wealth; neonatal sex; gestational age; birth weight; duration of labor; prolonged ROM⩾8 h; PROM; number of vaginal exams or maternal temperature. Women from households where everyone slept in one room demonstrated a decreased risk of maternal colonization (relative risk (RR)=0.79, 95% CI 0.68 to 0.92). Women who had vaginal and rectal swabs collected after ROM were more likely colonized (RR=1.17, 95% CI 1.01 to 1.36) than women who had swabs collected before ROM. The rainy season was associated with a 14% higher risk (RR=1.14, 95% CI 0.95 to 1.36) of maternal colonization than the summer season, although this was not statistically significant, and a 27% higher risk (RR=1.27, 95% CI 1.04 to 1.55) of maternal colonization than the winter season, which was statistically significant (Table 2).
Specific pathogens were associated with certain maternal risk factors. GBS maternal colonization was associated with ROM for ⩾8 h or more before swab collection (RR=2.60, 95% CI 1.37 to 4.94) and PROM (RR=2.29, 95% CI 1.23 to 4.25). No pattern of demographic or baseline characteristics emerged for specific pathogens with the exception of seasonality. The rainy season was associated with increased S. aureus colonization (RR=1.96, 95% CI 1.29 to 2.95). Women had an increased risk of being co-colonized during the rainy season (RR=1.68, 95% CI 1.03 to 2.74) than in the summer and winter seasons combined. Otherwise, women who were co-colonized did not differ from women who were not co-colonized in their observed demographic and neonatal characteristics and in their maternal risk factors and clinical signs of infection.
Neonatal umbilical colonization prevalences and etiologies
We collected initial umbilical swabs from 152 newborns on the first day of life. Eighty-three newborns (54.6%, 95% CI 46.6 to 62.6%) were colonized, of which 46 (55.4%) were singly colonized. Of the 46 singly colonized newborns, the most common organisms were E. coli (n=14, 30.4%), S. aureus (n=13, 28.3%), K. pneumoniae (n=6, 13.0%), Acinetobacter species (n=6, 13.0) and GBS (n=3, 6.5%). A total of 9 newborns (5.9%, 95% CI 2.1 to 9.7%, 3 singly colonized and 6 co-colonized) were colonized with GBS.
Of the 148 mother–newborn pairs with complete colonization data on day 1, the risk of umbilical colonization with any organism was not associated with maternal colonization (RR=1.04, 95% CI 0.77 to 4.40). The risk of GBS umbilical colonization was 12.98 times higher (RR=12.98, 95% CI 3.97 to 42.64) among newborns of mothers with GBS colonization compared with newborns of mothers without GBS colonization. The risk of Acinetobacter umbilical colonization was 7.74 times higher (RR=7.74 95% CI 5.32 to 12.32) among newborns of mothers with Acinetobacter species colonization compared with newborns of mothers without Acinetobacter species colonization. The risk of E. coli umbilical colonization was 3.22 times higher (RR=3.22, 95% CI 1.92 to 5.39) among newborns of mothers with E. coli colonization compared with newborns of mothers without E. coli colonization. Staphylococcus aureus umbilical colonization was not associated with maternal S. aureus colonization (RR=1.41, 95% CI 0.57 to 3.48). None of the newborns with K. pneumoniae umbilical colonization were born to women with K. pneumoniae colonization.
On day 7, among the 147 mother–newborn pairs with complete colonization data, the risk of GBS umbilical colonization is decreased although it is still persisted (RR=6.25, 95% CI 2.49 to 15.68) compared with the risk of GBS umbilical colonization on day 1 among newborns of GBS-colonized mothers. However, the risk of Acinetobacter species or E. coli umbilical colonization was lower and no longer significant on day 7 among newborns of mothers colonized by these organisms during labor (Table 3).
Neonatal umbilical colonization during the first 7 days of life
One hundred twenty-eight newborns had umbilical swabs collected on day 1 and day 7. Seven newborns (5.5%) were colonized with GBS on day 1, four (3.1%) of which continued to be colonized throughout day 7. There were nine (7.0%) new cases of GBS colonization on day 7 suggesting an environmental source. There was no statistically significant difference between neonatal GBS colonization status on day 1 and day 7 (5.5 vs 10.2%, respectively, P=0.146). Similarly, there was no statistically significant difference between neonatal K. pneumoniae colonization status on day 1 and day 7 (14.06 vs 9.38%, respectively, P=0.308). Acinetobacter and E. coli appeared to be transient colonizers that presented early in life (day 1) and then decreased in prevalence, whereas S. aureus colonization occurred later in life (day 7) (Table 4).
We identified common skin and enteric flora in the maternal reproductive tract: S. aureus, non-GBS, E. coli and K. pneumoniae. These organisms are similar to those identified in South Asia by other studies.4, 5, 9
Maternal carriage of GBS in Dhaka (8%) is less common than in the United States where the prevalence is estimated to be 28%.18 Our findings are similar to prior studies in India that found the prevalence of GBS colonization to be 6 to 9%.7, 11 A study in Pakistan among middle-class women found a higher prevalence of 23%.6 The women sampled in the Pakistan study may be different than our study population, although we do not have information to compare our baseline characteristics. GBS in this setting is susceptible to penicillin and cephalosporins, suggesting possible therapies for intervention and treatment if necessary. Antibiotic resistance to aminoglycosides and macrolides is as expected.19
While most demographic, maternal and neonatal characteristics observed were not associated with colonization, the number of rooms available for household members to sleep in, ROM and seasonality affected colonization status. Colonization status was negatively associated with all household members sleeping in one room. Further work could investigate possible differences in personal hygiene by the number of household rooms available in which to sleep and the role of socioeconomic status and GBS colonization. Colonization status was positively associated with ROM. As expected with most bacterial growth, colonization was positively associated with the rainy season. In particular, GBS was associated with prolonged ROM ⩾8 h and PROM, which are also factors associated with early-onset neonatal sepsis. In vitro studies have found increased growth of GBS after incubation with amniotic fluid, whereas amniotic fluid inhibited E. coli growth.20 Preventive strategies could be developed targeting high risk women.
Neonatal umbilical colonization and mother to newborn transmission
Newborns were colonized with organisms similar to those found colonizing the maternal reproductive tract: GBS, Acinetobacter, E. coli, S. aureus and K. pneumoniae. The risk of GBS neonatal umbilical colonization was 13 times higher among newborns of mothers with GBS colonization compared with newborns of mothers without GBS colonization. GBS colonization remains throughout the first 7 days of life and may be more persistent than other organisms. Although the prevalence of maternal and neonatal GBS colonization was lower in Bangladesh than in the United States, we may see an increase in the prevalence of colonization along with GBS neonatal sepsis as Bangladesh develops. Therapies preventing maternal GBS colonization may reduce early-onset GBS sepsis similar to GBS intrapartum antibiotic prophylaxis strategies in the United States.
Additionally, the risk of E. coli umbilical colonization among newborns of E. coli colonized mothers is three times higher. E. coli neonatal colonization is more prevalent during the first day of life and decreases by day 7, suggesting that these early days are most vulnerable. Strategies preventing E. coli maternal colonization may reduce E. coli early-onset neonatal sepsis, one of the most common organisms of neonatal sepsis in South Asia.21
There does not appear to be an association between maternal S. aureus colonization and S. aureus umbilical colonization. Furthermore, the prevalence of S. aureus umbilical colonization increases over time and is higher on day 7, at which time S. aureus was the most common colonizer of the umbilical cord. Although S. aureus is another common organism of neonatal sepsis in South Asia, our findings suggest that it may be community acquired from the environment rather than from mother.21, 22
Our study has several limitations. The results may be generalizable only to women who deliver at first-level facilities. Patterns of maternal reproductive tract flora may differ among women in communities who deliver at home or at higher level tertiary care centers. Excluding high-risk mothers may have underestimated our effect measure, however <1% of the women presented with complications and were excluded before enrollment. Thirteen percent of eligible women were not interested in participating in the study and these women may be different than those who enrolled in the study. We used laboratory data on neonatal umbilical colonization rather than diagnosing laboratory-confirmed early-onset neonatal sepsis because of the logistical challenges of obtaining blood, urine and cerebrospinal fluid cultures in this setting and the low incidence of early-onset neonatal sepsis.
Maternal colonization is associated with early-onset neonatal sepsis in Dhaka. In a subsample of this population, we found that newborns of colonized mothers had a 71% higher risk of sepsis in the first 7 days of life than their non-exposed counterparts.23 Future steps include focusing on women with risk factors associated with GBS colonization or E. coli colonization, examining maternal/neonatal antibody responses to develop targeted vaccines, understanding the relationship between umbilical colonization and the development of early-onset sepsis, serotyping mother–newborn paired isolates for further evidence of vertical transmission and testing interventions that will prevent early-onset neonatal sepsis globally.
We identified similar organisms in the maternal reproductive tract and around the neonatal umbilical cord. Newborns of mothers colonized with GBS or E. coli are at significantly higher risk of developing umbilical colonization with the same organism. Umbilical colonization with GBS persists over the first 7 days of life and appears to be more persistent than other organisms. Maternal risk factors, such as prolonged ROM and PROM, which expose the vaginal canal to amniotic fluid over an extended period of time, were associated with increased maternal GBS colonization. Future studies are needed to understand the relationship between umbilical colonization and early-onset sepsis, serotype mother–newborn isolates, and test interventions targeting maternal colonization by addressing risk factors, specific bacterial strains, prophylaxis with antibiotics or vaccines to decrease early-onset neonatal sepsis globally.
We thank our colleagues at the International Centre for Diarrheal Disease Research, Bangladesh Shams El Arifeen, Marzia Zaman, Taufiqur Rahman, Qazi Sadequr Rahman and Abu Salaheen for their field and data management support. We are grateful to Kazi Moksedur Rahman and the paramedics and community health workers from Shimantik who worked tirelessly for collecting data. We acknowledge Maksuda Islam, Anik Sarkar and Iftekhar Quasem from the Child Health Research Foundation, Bangladesh for laboratory work contributions. Thank you to Aletta Nonyane at Johns Hopkins Bloomberg School of Public Health and Daniel Roth from the University of Toronto for their technical expertise. Finally, we thank the mothers and newborns who generously gave their time to participate in this study. This research was supported by the Thrasher Research Fund Early Career Award and grant# 5KL2RR025006 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH), and the NIH Roadmap for Medical Research.