Background

Antimicrobial resistance (AMR) is a critical public health threat that disproportionately affects low- and middle-income countries (LMICs) [1]. The threat is particularly profound for vulnerable populations such as newborns, whose immature immune systems and microbiomes are less able to effectively protect them from invasive infections. Neonatal sepsis is increasingly driven by antibiotic-resistant bacteria, which has hindered progress in reaching the United Nations sustainable development goal 3 of reducing newborn mortality [2]. Moreover, poverty and infections caused by antibiotic-resistant bacteria may interact to further increase the risk for adverse infectious outcomes, as neonatal sepsis has also been linked with lower household wealth metrics [3].

Colonization with a bacterial pathogen typically precedes infection and puts individuals at higher risk of developing an infection caused by the same bacteria [4–6]. This relationship between bacterial colonization and infection has been well documented in the hospital setting but has also been demonstrated in the community [6,7]. In the case of Group B Streptococcus (GBS), maternal colonization has been associated with a 9-fold increased risk for neonatal infections [8]. This has resulted in implementation of screening and antibiotic prophylaxis for pregnant women who are colonized with GBS in many high-resource settings [9–11]. However, in low-resource settings, Gram negative bacteria, particularly antibiotic-resistant Gram negative bacteria, are increasingly a dominant cause of neonatal infections, whereas GBS is relatively rare [3,12–15]. This contrasts with high-resource settings where neonatal infections remain predominantly caused by GBS followed by Escherichia coli [16–18]. Although less is known about the relationship between maternal colonization with Gram negative bacteria and neonatal outcomes, recent studies identified an association between maternal colonization with antibiotic-resistant Gram negative bacteria and a variety of adverse outcomes, including premature birth, preterm premature rupture of membranes, perinatal asphyxia, and breech birth [8,19].

Compared to adults, newborns lack well-developed, diverse microbiomes, which can provide resistance to colonization by pathogens and infection. As such, colonization at birth largely reflects the environment they are born into. Traditionally, the newborn microbiome is thought to be heavily shaped by the maternal gut and vaginal microbiome, particularly among vaginally-delivered infants [20–22]. This pattern can be disrupted among newborns undergoing Cesarean deliveries (C-section), exposed to antibiotics (either directly or indirectly through maternal administration pre-delivery), or experiencing prolonged hospitalization [23–25]. These nuances in care and the birth environment can play a crucial role in shaping the newborn’s gut microbiome.

A key strategy for reducing maternal and infant mortality has been shifting from home-based deliveries to facility-based deliveries, ensuring access to lifesaving intrapartum services as well as antenatal and postnatal care [26]. As a result, substantial improvements have been made: since 1990, maternal and neonatal mortality have been cut in half [27,28]. Although these improvements may not be solely attributable to facility-based deliveries, current strategies continue to emphasize the important role that healthcare plays in improving birth outcomes [29].

However, it is unknown to what extent this shift toward facility-based deliveries may be influencing the exposure of newborns to antibiotic-resistant bacteria. In a previous report, we found a high prevalence of colonization with extended-spectrum beta-lactamase-producing bacteria (ESBL-PB) and carbapenem-resistant bacteria (CRB) among newborns and pregnant mothers undergoing delivery in a hospital setting in Bangladesh [30]. Maternal colonization with ESBL-PB increased from 77% on admission to 98% following delivery, while colonization with CRB increased from 15% to 89%. Among the newborns, 89% were found to be colonized with ESBL-PB and 72% with CRB. Paired with high rates of antibiotic usage and C-sections as well as frequent contamination of hospital environments with these pathogens, these findings implicated the hospital setting in the transmission of antibiotic-resistant bacteria to newborns. Similarly, a large multicenter study of facility-based deliveries found very few genetically related antibiotic-resistant bacteria between mother and newborn pairs, indicating horizontal transmission as a key exposure pathway [19].

Although hospital-based deliveries have been increasing in Bangladesh, one-third of deliveries are still taking place in the home [31]. Roughly half of these deliveries are attended by a medically trained provider, which includes nurses, midwives, paramedics, family welfare visitors, community skilled birth assistants, and sub-assistant community medical officers. In contrast with facility-based deliveries, little is known about the infection prevention and control (IPC) practices of home-based deliveries. Additionally, the exposure to antibiotic-resistant bacteria may vary by birth setting. This is supported by findings of increased newborn colonization with CRB among those born in a study facility compared with those born outside a study facility, which included home births [19]. This study was conducted to evaluate ESBL-PB and CRB colonization patterns among newborns delivered at home and assess perinatal exposures and IPC practices that may contribute to colonization with antibiotic-resistant bacteria in home births.

Methods

Results

Demographics of enrolled participants

We enrolled 50 mothers and newborns. All deliveries were non-assisted (i.e., without use of instruments such as forceps or vacuum cups), live vaginal births that took place in the household. The median age of the mothers was 23 years (range 18–26) (Table 1). The median monthly household income was $118 USD, and the median household size was 6 people. This amounts to $0.66 USD per person per day, which meant that all households fell below the World Health Organization poverty level ($1.90 USD per person per day at the time of the study). Approximately half (46%, n = 23) of mothers reported their highest education level as primary school or lower, and 48 (96%) identified family caregiver as their occupation. All participants reported using a pit latrine, either shared or private, as their toilet type, and all reported handwashing facilities with soap and water at their residence. Additionally, all mothers reported physical contact with domestic animals during their pregnancy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographics of mothers undergoing home-based deliveries, Baliakandi, Bangladesh, 2022 (N = 50).

https://doi.org/10.1371/journal.pone.0325404.t001

One-third of the women had no prior pregnancies. Thirty-eight (76%) mothers reported at least some antenatal care, but 42 (84%) mothers had less than 4 antenatal care visits prior to delivery. Only 4 (8%) mothers reported using antibiotics in the 30 days prior to delivery. The most commonly self-reported pregnancy complications were anemia (n = 12, 24%) and urinary tract infection (n = 6, 12%). Most deliveries occurred at term (n = 46, 92%). One participant was hospitalized during her pregnancy, but none of the enrolled mothers or babies were hospitalized after delivery.

Colonization prevalence and associated risk factors

ESBL-PB colonization was more frequent among mothers compared with newborns (Fig 1). Forty-seven (94%) mothers were colonized with ESBL-PB, while 37 (74%) were colonized with CRB. Thirty-seven (74%) mothers were colonized with both. Of the newborns, 36 (72%) were colonized with ESBL-PB, and 27 (54%) were colonized with CRB. Twenty-six (48%) newborns were colonized with both.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Colonization prevalence among mothers and newborns following home-based delivery, Baliakandi, Bangladesh, 2022 (N = 50).

Colonization was determined based on bacterial growth on selective media from rectal and/or vaginal swabs. ESBL-PB = extended-spectrum beta-lactamase-producing bacteria; CRB = carbapenem-resistant bacteria.

https://doi.org/10.1371/journal.pone.0325404.g001

Given the ubiquity of ESBL-PB colonization, no differences were found between mothers with ESBL-PB colonization compared to those without ESBL-PB colonization. However, the 3 mothers without ESBL-PB colonization all had incomes above the median and reported having a private pit latrine (rather than a shared pit latrine). Common exposures among the 37 mothers with CRB colonization included having at least a secondary education (59.5%), an income above the median (64.9%), a private pit latrine (56.8%), contact with domestic animals inside the home (73.0%), any antenatal care (81.1%), and no pregnancy complications (70.3%), though these differences were not significantly different from mothers without CRB colonization (Table 2). The most common animals contacted inside the home were chickens and cats (S2 File Supplementary Fig and Tables).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Comparison of maternal colonization patterns following home based delivery across community variables, Baliakandi, Bangladesh (N = 50).

https://doi.org/10.1371/journal.pone.0325404.t002

Discussion

Mothers and newborns who underwent home deliveries in a rural setting in Bangladesh were found to have an alarmingly high incidence of ESBL-PB and CRB colonization following delivery. This occurred despite limited interaction with healthcare settings 一 only one mother reported hospitalization during pregnancy 一 and rare antibiotic use during pregnancy. This contrasts with studies from high-resource settings that frequently link antibiotic resistance with healthcare settings and antibiotic use [32,33]. The high prevalence of colonization with antibiotic-resistant bacteria in a community setting in the absence of recognized risk factors highlights that the epidemiology of antibiotic-resistant bacteria is context-dependent and varies between high- and low-resource settings.

Colonization with ESBL-PB reported in this study is similar to other reports from Bangladesh, though CRB colonization was found to be higher than in other studies. One study evaluating colonization among children less than 1-year of age in rural Bangladesh found 82% were colonized with E. coli resistant to third-generation cephalosporins, and 74% of infants were colonized with ESBL-producing E. coli [34]. This study did not report prevalence of CRB. Another study evaluating colonization prevalence in urban Bangladesh reported 78% of adults were colonized with Enterobacterales resistant to third-generation cephalosporins but only 9% were colonized with CRB [35]. These estimates align with other reports from South Asia but differ greatly from reports in high-resource settings, where the prevalences of ESBL-PB and CRB colonization are much lower among community populations [36–40]. The difference in CRB colonization is particularly striking given that carbapenem antibiotics are infrequently used in Bangladesh. This could indicate that non-specific mechanisms of antibiotic resistance (e.g., efflux pumps, ampC overproduction) are mediating carbapenem resistance. Alternatively, antibiotic-resistance genes against carbapenems (i.e., carbapenemases) may be co-located on plasmids with other antibiotic-resistance genes where they are being selected for following exposure to non-carbapenem antibiotics or other biocides [41]. Additional genomic analyses could provide further insight into the molecular epidemiology of CRB. For example, the presence of non-specific resistance mechanisms may point away from antibiotics as a key driver of resistance, potentially implicating more generic exposures such as heavy metals or other chemicals in the upregulation of these genes. Plasmid analyses could reveal whether antibiotic-resistance genes located on the same plasmid might be upregulated through exposure to any one of the antibiotics to which a resistance gene is present. Thus, these findings could help identify potential drivers for CRB dissemination.

Although colonization of mothers was not evaluated pre-delivery, colonization post-delivery might be expected to be similar to pre-delivery in the absence of any healthcare exposures or interventions. Yet, a previous study conducted among mothers presenting to a hospital in an adjacent community for delivery were found to have substantially lower rates of colonization with antibiotic-resistant bacteria pre-delivery [30]. While the participants in this study were largely similar to the mothers in the hospital cohort, the mothers who underwent home births were of lower income. In addition, all participants in the current study listed pit latrines as their toilet type compared with 79% of mothers in the hospital-based study; the remainder utilized pour-flush toilets, which are generally regarded as more sanitary and hygienic. Moreover, 98% of mothers who underwent home births indicated not treating their water, while only 72% of mothers did in the hospital cohort. These statistics outline a population that is lower income and with less access to water, sanitation, and hygiene (WASH) infrastructure, which may be mediating exposure to antibiotic-resistant bacteria [42].

Although risk factor analyses did not reveal any clear perinatal exposures linked with colonization, all participants met criteria for living in poverty, which has been directly associated with increased risk for colonization or infection with antibiotic-resistant bacteria [43]. A prior study evaluating risk factors for colonization among pregnant women in India identified that low socioeconomic status was associated with increased risk for urinary colonization with ESBL-PB [44]. Similarly, more global analyses have linked increased antibiotic resistance with lower development indices, such as governance, health expenditure, and infrastructure [45,46]. Additionally, WASH is a poverty-related factor that has been linked with increased antibiotic resistance [42,45,47]. In this study, participants reported primarily relying on pit latrines and tubewells. While these are frequently classified as improved WASH infrastructure, improved infrastructure does not always translate to a lack of exposure to fecal pathogens [48]. Additionally, because humans are a major reservoir for antibiotic-resistant bacteria, increased fecal exposure likely parallels increased exposure to antibiotic-resistant bacteria [49]. Several studies have found fecal bacterial contamination in tubewells, regardless of depth, which has been attributed to potential breaks in the structures [50,51]. A meta-analysis of studies from LMICs found fecal contamination to be higher from boreholes compared with piped water (43% vs. 25%) [52]. Similarly, another systematic review found 3.45-times higher odds of fecal contamination among self-supplied water (primarily boreholes) compared with piped public water in LMICs [53]. Additionally, pit latrines with unsealed septic tanks can lead to bacterial contamination of ground water through seepage, which may, in turn, contaminate drinking water sources, particularly when located in close proximity [54]. This is especially problematic in areas with a high-water table such as Bangladesh, where nearly a third of the land area of the country is underwater during rainy seasons. As these water sources are used for a range of activities 一 drinking, washing, cooking, bathing 一 and treatment of water is uncommon, they may be a source of fecal bacteria, including antibiotic-resistant bacteria, permeating household environments and being ingested, leading to colonization of mothers that could also lead to contact transmission to newborns.

Additionally, inadvertent exposures to antibiotic residues may be mediating high levels of antibiotic-resistant bacteria in this population. Bangladesh has a flourishing pharmaceutical manufacturing industry that supplies 97% of the domestic market and exports to 131 other countries [55]. Pharmaceutical production is known to promote AMR through release of effluents from the manufacturing process into the environment, which frequently contain biocides with antimicrobial properties [56]. Prior sampling from a river in Bangladesh that was downstream of a pharmaceutical manufacturing plant observed concentrations of metronidazole that were more than 300 times higher than the safe target [57]. These biocides may be entering into food or water supplies and, in turn, exerting their own selective pressures for AMR. Further exploration is warranted to better understand how this could be contributing to community AMR in Bangladesh.

Little information is available on IPC practices during home deliveries. Data from this study revealed gaps in IPC practices, particularly related to hand hygiene. Lapses in IPC have been linked with newborn exposure to pathogens through both direct and indirect means, though these studies have largely been conducted in hospitals where there is also shared medical equipment, patient crowding, simultaneous provision of care to multiple patients, and use of invasive devices [58]. In the case of home births, many of these pathways are eliminated, with the possible exception of shared medical equipment, reducing the likelihood of horizontal transmission between households. The more direct route of transmission may be between mothers and their home environment. Although not systematically evaluated in the current study, the majority of the households in this region utilize earthen floors, which allow limited cleaning strategies. These floors permit persistent contamination with human fecal bacteria (such as through tracking by foot traffic or defecation by young children) as well as animal fecal bacteria (who often have a fluid range between indoor and outdoor dwellings and whose feces permeate the household premises). Home births are occurring on these same floors, covered with some bedding materials, providing opportunities for transmission of antibiotic-resistant bacteria. During the birthing process, bacteria are likely transmitted between the floors and the bedding, potentially providing a route for newborn exposure. While we did not do any environmental sampling of these houses, earthen floors could be an important reservoir for antibiotic-resistant bacteria and should be evaluated in future studies [59,60].

This study has several limitations, with the key limitation being the homogeneity of the population. The similarities in exposures across mothers and newborns limited the ability to detect significant associations, particularly in the setting of ubiquitous colonization. This high colonization prevalence overall demonstrates intense colonization pressure at the community level. Therefore, colonization outcomes among individuals may not be independent, potentially obscuring the identification of meaningful associations using individual-level analyses [61]. For example, community-level sanitation may matter as much or more than household-level sanitation [62]. Moreover, the small sample size also limits the capacity to identify risk factors.

Another major limitation is that colonization was assigned based only on CHROMagar culture results, which may be an imperfect assessment of colonization. Basing colonization on CHROMagar growth may overcall resistance, which could have contributed to the lack of significant associations identified. However, other studies have demonstrated that CHROMagar findings from rectal swabs give comparable phenotypic findings to standard microbiologic workflows with automated susceptibility and AST, justifying its use for routine screening in healthcare settings [63–65]. Additionally, compared with other studies, we did not perform culture enrichment, and we relied on rectal swabs compared with stool, which may have led to undercalled estimates, potentially also compromising our analysis [66–70]. Therefore, true colonization estimates could be higher or lower than reported. However, even with a substantial margin of error, the burden of colonization in this population remains concerningly high. Another limitation is that the data from mothers and birth attendants was self-reported and may be subject to recall or social desirability bias.

This study did not address the role of Gram positive bacteria in contributing to neonatal infections. While GBS is relatively rare in LMICs, other Gram positive bacteria such as Staphylococcus aureus and Enterococcus spp. are known to contribute to neonatal infections in LMICs [3,15]. Future studies should explore potential risk factors and transmission pathways through which Gram positive bacteria may cause neonatal infections to provide a more holistic understanding of neonatal sepsis and opportunities for intervention.

Conclusions

Overall, this study identified alarmingly high rates of ESBL-PB and CRB colonization among mother-newborn pairs undergoing home deliveries in rural Bangladesh, highlighting the critical need for further understanding of community based exposure and transmission. This is especially important in low-resource settings, where socioeconomic factors may play a vital role in the spread of AMR.

Supporting information

Acknowledgments

We are grateful to all the participants of this study for providing consent to collect samples and invaluable information. We would also like to thank the CHAMPS Bangladesh team for their collaboration in the use of their surveillance network for identifying births and for all their preceding work to build trust within the community that facilitated participation in the current study. icddr,b acknowledges with gratitude the Governments of Bangladesh and Canada for providing core/unrestricted support.