Group B streptococcal prevalence in internal organs and placentas of deceased neonates and stillbirths in South Asia

Group B streptococcus (GBS) has been associated with adverse pregnancy outcomes, but few prospective studies have assessed its prevalence in low‐ and middle‐income country settings. We sought to evaluate the prevalence of GBS by polymerase chain reaction (PCR) in internal organ tissues and placentas of deceased neonates and stillbirths.


| I N TRODUC T ION
Group B streptococcus (GBS) is a commensal microbe that may inhabit the gastrointestinal and vaginal tracts of women of reproductive age, with reported carriage rates ranging from 4% to 40%. 1 Colonisation with GBS during pregnancy has been implicated as a leading cause of neonatal infections, including sepsis, pneumonia and meningitis. 2,3Vertical transmission to the newborn may occur via fetal aspiration of infected amniotic fluid or during passage through the birth canal. 4Vertical transmission from the mother to the fetus accounts for up to 75% of cases of neonatal GBS colonisation; 1%-2% of these infants will develop early-onset GBS infection. 5,6][9] The prevalence of vaginal colonisation with GBS is thought to be highest in Africa (22.4%), followed by the USA (19.7%),Europe (19%), the Western Pacific (17.9%), the Eastern Mediterranean (16.7%) and Southeast Asia (11.1%). 109][20] The microbiological detection of GBS is dependent on sampling techniques and the method of laboratory processing.When both culture and polymerase chain reaction (PCR) methodologies have been compared for the detection of GBS, PCR yielded higher rates of detection. 21,22eonatal colonisation and neonatal disease are distinct entities with only a small fraction of colonised infants developing disease (e.g.sepsis).Although GBS is a leading cause of neonatal sepsis in some high-income countries (HICs), its burden in low-and middle-income countries (LMICs) is less clear.Several multicentre studies of infection in young infants conducted in South Asian countries did not commonly identify this pathogen. 23Similarly, a multicentre study from six LMICs that enrolled nearly 9000 infants did not identify a single case of GBS among the neonates. 24Several multicentre studies conducted in South Asia reported low levels of GBS in pregnant women and neonates, compared with cases reported from Africa. 25,26Studies evaluating neonatal sepsis in LMICs, including India and Pakistan, reported a higher proportion of Gram-negative bacteria relative to GBS. 24,27,28 The environmental abundance and colonisation of Gram-negative organisms in South Asia might be a source of the higher rate of infection caused by Gram-negative organisms. 29aternal colonisation is the primary risk factor for GBS transmission; therefore, intrapartum antibiotic prophylaxis (IAP) is recommended for colonised women and women with premature or prolonged rupture of membranes, fever and anticipated preterm birth in the USA. 3 In 2002, the US Centers for Disease Control and Prevention (CDC) recommended the universal culture-based screening of all pregnant women from 35 to 37 weeks of gestation, in order to treat those testing positive for GBS with intrapartum antibiotics prior to delivery, to prevent early-onset neonatal GBS sepsis. 30Although IAP has resulted in an 80% reduction in early-onset GBS disease in the USA, 31 not all countries have established prophylaxis protocols or can provide accurate GBS disease estimates.
There is a scarcity of PCR-based GBS data from South Asia, and the presence of GBS and its effects are broadly debated.In this study, using PCR, we present data to show the prevalence of GBS in the placenta and membranes, as well as in samples taken from heart blood, lungs, liver and brain, among stillbirths and preterm neonatal deaths, and microbial cultures from maternal rectovaginal swabs. 32

| Study design
This sub-study is a part of a prospective, observational study, the Project to Understand and Research Preterm Pregnancy Outcomes and Stillbirths in Southeast Asia (PURPOSe), performed in India and Pakistan from July 2018 to February 2020. 32he detailed protocol for the PURPOSe study has been published previously. 32The study was conducted in India and Pakistan: at three hospitals (Chigateri General Hospital, Bapuji Child Health Institute, and Women and Child Hospital) attached to Jagadguru Jayadeva Murugarajendra Medical college (JJMMC) in Davangere, India; and at the Jinnah Postgraduate Medical Centre (JPMC) and at the National Institute of Child Health (NICH), both large referral hospitals in Karachi, Pakistan.
The study planned to screen, consent and enrol consecutive women with stillbirths (350) and approximately 2500 preterm births at each site.To ensure that there were a sufficient number of preterm births to meet the target sample size of 350 neonatal deaths at each site, based on estimated neonatal mortality rates, we enrolled approximately 2500 consecutive, consenting pregnant women with preterm live births at each study site.We also enrolled 100 live term controls at each study site.
Our goal was to determine the prevalence of GBS using PCR in the placenta, membranes and cord blood, as well as the prevalence of GBS in samples taken from fetal tissues, including heart blood, lungs, liver and brain/cerebrospinal fluid (CSF), among stillbirths and preterm neonatal deaths, and to determine the prevalence of GBS from maternal rectovaginal swabs.
At the time of enrolment, study personnel conducted an interview with the mother and performed a clinical examination of all stillbirths and neonates.The interview addressed the mother's sociodemographic data, current and past pregnancy history, and medical and obstetric history.Before delivery, a rectovaginal swab of pregnant women was collected only at the site in Pakistan.The swabs were immediately transported to the microbiology department for processing using standard microbiological culturing methods.Following delivery, tissue samples from the placenta, membranes and cord blood were collected in a 2-mL cryovial containing 1.5 mL of RNAlater (ThermoFisher Scientific, Waltham, MA, USA) and stored at −20°C until processing for PCR using the TaqMan Array Cards (TAC; ThermoFisher Scientific).Following stillbirth and preterm neonatal death, consent was obtained for minimally invasive tissue sampling (MITS), a method of obtaining internal organ samples by needle biopsy using aseptic techniques.Using MITS, fetal tissues including heart blood, liver, lungs, brain and CSF were collected, each in 2-mL cryovials containing 1.5 mL of RNAlater, and stored at −20°C until further processing for PCR.

| Sample collection and processing for PCR
The cryopreserved tissue samples were initially processed using Proteinase K and homogenised using beads and the Precellys 24 homogeniser (Bertin Technologies, Montignyle-bretonneux, Rockville, MD, USA).Total nucleic acids were extracted from the homogenates using the QIAGEN EZ1 Virus mini kit on the QIAGEN EZ1 Advanced XL instrument (QIAGEN, Germantown, MD, USA).Purified nucleic acids were loaded onto the TAC containing customised assays developed by the CDC.Both the positive and negative template controls were added to the card to validate the proper amplification of each custom assay and to identify any exogenous contaminants, respectively.Each TAC was loaded onto the QuantStudio 7 Flex instrument (ThermoFisher Scientific) for PCR amplification.Samples were reported as GBS positive at any cycle threshold (C t ) if the amplification curve and fluorescence level were deemed appropriate.

| Sample processing for GBS by culture
At the site in Pakistan, women who presented with a likely preterm birth or stillbirth, or who were one of the controls, had a rectovaginal swab collected for microbiological culture analysis.The swab was first inserted approximately 5 cm into the vagina, gently rotated twice around the midvaginal wall and removed without touching the labia.Then, the same swab was inserted into the anal canal to a depth of about 1 cm to obtain a rectal sample.After collection, the swab was placed into a sterile collection tube with transport media (Copan Diagnostics, Murrieta, CA, USA) and immediately transported to the microbiology laboratory for further processing.
The rectovaginal swab was used to inoculate LIM broth (Todd Hewitt broth with nalidixic acid and colistin; Himedia, Thane, Maharashtra, India) and incubated overnight at 37°C.The overnight culture was subcultured on 5% blood agar plates and CHROMagar™ Strep B (CHROMagar, Paris, France), which was incubated at 37°C for 24 h.After incubation, the plates were inspected for beta-haemolytic colonies on blood agar and for mauve-coloured colonies on Strep B CHROMagar™.If no beta-haemolytic colonies were observed after 24 h, the plates were re-incubated for another 24 h and inspected again.Presumptive GBS colonies were verified by Gram staining and performing the catalase test, the Camp test and the Hippurate hydrolysis test, and confirmed by using the Streptex™ Latex B agglutination test (Remel, San Diego, CA, USA).

| Statistical analysis
Descriptive statistics and prevalence of GBS across various sample types are presented as frequencies and percentages.To assess the relationship of several potential risk factors for GBS PCR positivity in at least one placental or organ tissue, we performed Cochran-Mantel-Haenszel (CMH) tests, adjusting for site.Overall sample size requirements were met for these tests despite expected cell counts of <5 in stratified analyses for several variables.Analyses were conducted using SAS Enterprise Guide 7.13 (SAS Institute Inc.).

| R E SU LTS
This study describes maternal rectovaginal GBS culture results from Pakistan and PCR results from India and Pakistan for various tissues including the placenta, membranes, umbilical cord and MITS-collected tissues (brain, lung, liver, CSF and heart blood) from stillbirths and from deaths following a preterm live birth.PCR results were also obtained from cord blood on a subset of participants.Altogether, there were 1793 mothers, 896 stillbirths, 770 preterm neonatal deaths and 200 live-birth term controls in the analytic sample who were screened, eligible, consented and had at least one PCR result reported (Figure 1).
The majority of women enrolled were 20-30 years of age (77.8%) (Table 1).Overall, 33.4% were illiterate or had no formal education, and 56.3% had between 5 and 12 years of education.There were substantial differences in educational levels between women at the sites in India and Pakistan.Most women were multigravidas (65.1%) and nearly all had received at least some antenatal care (ANC) (93.9%).As an indicator of the quality of ANC, almost all women in the Indian sites (98.7%) but less than half of the women at the Pakistani site (37.8%) had received a tetanus toxoid vaccine.In this population, 5.7% of the study fetuses/infants were between 20 and <25 weeks of gestation, 25.7% were between 25 and <30 weeks of gestation, 44.8% were 30-37 weeks of gestation and 23.8% were more than 37 weeks of gestation.Among the fetuses/infants, just over half were male (53.9%) and weighed 1000-2499 g at birth (58.2%).
We also evaluated GBS PCR results from the tissues of internal organs obtained by MITS in both stillbirths and neonatal deaths (Table 2).Overall, 1.2% of the cases (1.1% of stillbirths and 1.3% of preterm neonatal deaths) had at least one internal organ that tested positive for GBS.Table 3 presents the frequency of GBS found in various placental tissues of stillbirths, neonatal deaths and the term controls.The frequencies ranged from zero and 6.4%.
We also assessed the relationship between potential risk factors for GBS infection and GBS positivity, while adjusting for site (Table 4).Among the characteristics evaluated, only increased GBS positivity in primigravidae was statistically significant.
Finally, in Pakistan we obtained rectovaginal cultures on hospital admission from 577 women with a stillbirth, 428 women who eventually had a preterm neonatal death, 817 women with a preterm infant who lived to 28 days and 101 women with a term live birth (controls) (Table 5).For each of these categories, the GBS positivity rates were between 4.8% and 6.1%.There was no difference in the prevalence of GBS among women with these outcomes.

| DISCUS SION
][3][4][5][6][7][8][9][10] In this analysis of data from India and Pakistan, we were interested in assessing the prevalence of GBS identified by PCR in a variety of internal organ tissues of deceased fetuses and neonates, and in their placentas.We also evaluated the placentas of term live-born infants, as well as the rectovaginal cultures of pregnant women in Pakistan as they entered the hospital for delivery.The most obvious finding from this series of analyses was that no matter the country, the condition of the subject, the tissue studied or the methodology used, the prevalence of GBS was low, ranging between 0% and 6%.We also evaluated several potential risk factors for GBS positivity in placental and organ tissues.Among the characteristics evaluated, only primigravidae had increased GBS positivity.
][3][4][5][6][7][8][9][10] This pathogen is of particular interest because intrapartum antibiotics given to colonised mothers can reduce early-onset infection in the newborn.This intervention requires the screening of pregnant women late in pregnancy, or in labour, and the administration of antibiotics to women who have been colonised.
Findings similar to this study are found in other studies.For example, a meta-analysis of the prevalence of maternal GBS colonisation in India, evaluating 9778 pregnant women, mostly using culture, estimated that the prevalence of GBS colonisation is 7.8%. 8They found that the risk of preterm delivery was higher among women with GBS colonisation compared with women  without GBS colonisation.They also found the prevalence of GBS was higher in primigravidae than in multigravidae.A retrospective study conducted in Jiangsu, China, between 2017 and 2019, where 9022 pregnant women were screened for GBS by PCR and culture, found higher rates of detection by PCR (8.7%) than by culture (3.5%). 33A study conducted in 750 pregnant women in central Vietnam between 2018 and 2019 found that the prevalence of vaginal GBS colonisation was 9.2%. 34In contrast to our findings, a retrospective cross-sectional study conducted in a large tertiary-care university teaching hospital in Virginia, USA, between 2011 and 2019, found that the prevalence of maternal GBS colonisation was 35%.Smoking and young age were identified as independent risk factors for GBS colonisation during pregnancy. 7A longitudinal study in Nigeria in 500 women and their newborns found that 34% of the mothers and 19% of their neonates were colonised with GBS. 35They also found that prolonged membrane rupture and young maternal age were associated with maternal colonisation, whereas maternal rectal colonisation and prolonged rupture of membranes were associated with neonatal colonisation.
To ensure our GBS PCR results were consistent with rectovaginal GBS culture results, we performed rectovaginal cultures at the Pakistan site and found that the prevalence of GBS varied between 4.8% and 6.1%.This is in agreement with another study conducted in Rawalpindi, Pakistan, in 2009 that found the prevalence of GBS in mothers to be 8%. 36In contrast, a study conducted in Lahore, Pakistan, showed a higher prevalence (16%) of GBS colonisation. 18,19These differences show that the prevalence of GBS may vary even within a country.As compared with HICs, there appears to be a low prevalence of GBS colonisation in mothers and neonates in South Asia.
One possible reason for the low prevalence of GBS found in this study is the very high use of antibiotics during pregnancy in South Asian sites. 37It is likely that these antibiotics reduce the prevalence of susceptible Gram-positive organisms, such as GBS, but have little impact on the majority of Gram-negative organisms that cause disease in mothers and neonates in South Asia.As an example, in this same population, we demonstrated very high rates of neonatal infection with Acinetobacter baumannii. 38A B L E 2 Group B streptococcus (GBS) identified by PCR in internal organ samples obtained by MITS among stillbirths and preterm neonatal deaths in India and Pakistan.a n/N (%).

Overall
b Cord blood collected from enrolled stillbirths and term controls only.
Many demographic factors associated with a higher prevalence of GBS, but data are conflicting.In our study, we found primagravidity to be a significant risk factor for GBS colonisation.This is consistent with a study conducted by Khatoon et al. 12 Another meta-analysis in India analysed the influence of gravidity on GBS prevalence and found that the prevalence of GBS was higher in primigravidae than in multigravidae, although the difference was not statistically significant. 8However, in a study involving Turkish pregnant women, parity was unrelated to GBS carriage. 39In the present study, younger women showed a higher prevalence of GBS, as compared with older women; however, this difference was not statistically significant.This finding is consistent with a study from Korea, where younger age was a significant risk factor for GBS colonisation. 9We did not find any significant correlation between GBS colonisation and demographic risk factors such as educational status and socio-economic factors.
This study had several and weaknesses.First, the study was performed at sites in two countries of South Asia with different levels of antenatal care and neonatal mortality rates.They had similar oversight and data forms, but different investigators.To the best of our knowledge, this study is the first of its kind with a large sample size, using PCR for the detection of GBS from various tissues of deceased fetuses and neonates, and their placentas, as well as placentas of preterm and term live-born infants.The similar results from the two countries indicate the robustness of the findings.The study has some limitations.First, we did not evaluate the clinical presentation of mothers or neonates for various GBS infections and correlate them with microbiological findings.Second, our study recruited women for the PCR analyses only at delivery.Third, our study recruited patients in urban teaching hospitals.Hence, the data cannot be assumed to reflect the overall levels of GBS colonisation for these countries.

| CONCLUSION
Neonatal disease caused by GBS is traditionally classified as early-onset disease that occurs in neonates of <7 days of age and late-onset disease that occurs in infants aged 7-90 days.Understanding the burden of GBS disease in young infants (0-89 days), including neonates (0-27 days), is important to guide public health decision-making on interventions.Many HICs have implemented IAP, aiming to reduce earlyonset GBS disease for women with rectovaginal GBS colonisation detected through microbiological screening or with clinical risk factors. 31,32However, this strategy will not reduce late-onset infant GBS disease, and in LMICs where there are more home deliveries and where women present later for delivery, IAP may be less feasible and effective than other potential strategies for prevention, such as a maternal GBS vaccine.With the availability of more extensive data, we can now better understand that the prevalence of GBS is low in and in South Asia, when compared with many HICs.However, data from the literature and the PURPOSe study suggest a high incidence of neonatal sepsis caused by Gram-negative organisms, often nosocomially acquired, such as Acinetobacter. 38,40Hence, we believe that in South Asian countries, less emphasis might be placed on reducing GBS infection and more emphasis should be placed on neonatal sepsis caused by Gram-negative organisms.

T A B L E 1
Maternal and infant characteristics.

T A B L E 4
Maternal characteristics and association with Group B streptococcus (GBS) positivity.: p-values calculated from a Cochran-Mantel-Haenszel test adjusted by site; b row mean scores differ statistic; or c general statistic.(6.1) 39/816 (4.8) 5/101 (5.0)

India Pakistan Stillbirths Preterm, neonatal death Stillbirths Preterm, neonatal death
Number and percent of infants with at least one positive GBS result.Group B streptococcus in placental tissues of stillbirths, preterm neonatal deaths and term controls.
bT A B L E 3 MGK, EMM and RLG conceived of the analyses and wrote the initial draft of the manuscript with EF; MGK, NKG, SUH, MSS, IA and JK performed and oversaw the microbiology analyses; EF, and KH performed statistical analyses with support from EMM. SS, SSG, SST, SMD, GG, HY, SY oversaw study implementation; EMM, SS, SSG and RLG worked on the protocol.All authors read and reviewed the final manuscript.
38AU T HOR C ON T R I BU T ION S