The study was approved by the National Health Ethics Committee of the DRC (certificate number CNES 001/DPSK/229PP/2023, 29 January 2023). Verbal informed consent was obtained after clearly explaining the objectives and procedures in the local language. A brief educational session on bacterial resistance and carriage of MDR bacteria during pregnancy was conducted beforehand to facilitate understanding and acceptance. Consent was given by participants or mothers before sampling. Participants’ confidentiality and anonymity were strictly maintained throughout the study. Data confidentiality was ensured by anonymisation and secure storage. The study complied with the Declaration of Helsinki, International Council for Harmonisation – Good Clinical Practice (ICH-GCP) E6(R2), and national regulations.

This descriptive, multicentre laboratory-based study took place from April 2023 to October 2024 across 10 health facilities in South Kivu province, eastern DRC. Five urban hospitals in Bukavu city were included: Panzi General Reference Hospital, Centre Hospitalier de la 8e Communauté des Églises de Pentecôte en Afrique Centrale de CAHI, Bukavu, Baptist Community Hospital Nyamugo, Institut Supérieur des Techniques Médicales de Bukavu-Bukavu Hospital Center, and Ciriri Hôpital Général de Référence. Five rural hospitals were also selected: Ciriri Hôpital Général de Références of Miti-Murhesa, Uvira, Walungu, Kaziba, and Nyantende, representing both urban and rural settings in a region affected by prolonged instability.

The study included pregnant women with a gestational age of 37 weeks or more who voluntarily attended antenatal consultation and agreed to participate after receiving health education on AMR and pregnancy. Exclusion criteria were pregnant women with known risk factors for bacterial colonisation, such as gestational diabetes, prolonged hospitalisation (> 10 days) during pregnancy, malignancy, and kidney and/or liver disease.

Sample size was calculated using the Kish–Leslie formula¹⁹: n = [z² × p × (1–p)]/e², where n is the minimum required sample size, z is the confidence level value (1.96 for 95% confidence), p is the estimated prevalence of ESBL carriage in pregnant women (64.3%, based on Tanzanian data⁹), and e is the margin of error (5%). The minimum sample size was estimated at 354 pregnant women. Convenience sampling was used, with consecutive recruitment of eligible pregnant women during their clinic visits until the sample size was reached.

Women identified as carriers of ESCR and/or CRE were followed until delivery, when fresh rectal swabs were collected from the mother and newborn, then sent again to the Institut Supérieur des Techniques Médicales de Bukavu laboratory and strains forwarded to Rome for confirmation.

Rectal swabs were collected from pregnant women ≥ 37 weeks gestation after consent. Mothers identified as carriers at antenatal care had additional swabs taken at delivery from both them and their newborns. Samples were transported sterilely the same day to Institut Supérieur des Techniques Médicales de Bukavu.

Selective culture media used included CHROMagarTM ESBL (CHROMagar, Paris, France) for ESCR-E and CHROMagarTM mSuperCARBATM (CHROMagar, Paris, France) for CRE. Plates were incubated at 37°C for 18–24 h. Colonies with characteristic morphologies on CHROMagar were selected for further identification: on CHROMagar ESBL, ESBL-producing E. coli appeared as dark pink to reddish colonies; Klebsiella spp., Enterobacter spp., and Citrobacter spp. as metallic blue colonies, sometimes with a reddish halo; Proteus spp. as colonies with a brown halo; Pseudomonas spp. as cream translucent to blue colonies; and Acinetobacter spp. as opaque cream colonies. On CHROMagar mSuperCARBA, carbapenemase-producing Enterobacterales appeared as pink to mauve colonies. These isolates then underwent biochemical identification using oxidase tests, Kligler–Hajna agar, Simmons citrate, urea–indole medium, sulfide–indole–motility medium, lysine decarboxylase, and o-Nitrophenyl-β-D-galactopyranoside tests, as per established protocols.²⁰

Antimicrobial susceptibility testing was performed by disk diffusion on Muller Hinton agar following European Committee on Antimicrobial Susceptibility Testing 2022 (Version 12.0) guidelines.²¹ For ESCR-E isolates, antibiotics tested included amoxicillin-clavulanic acid, ticarcillin, cefixime, ciprofloxacin, cefotaxime, ceftriaxone, ceftazidime, cefepime, gentamicin, amikacin, chloramphenicol, and tetracycline.

Carbapenem-resistant Enterobacteriaceae isolates were tested against the same antibiotics plus carbapenems: imipenem, meropenem, and ertapenem.

Zone diameters were interpreted using European Committee on Antimicrobial Susceptibility Testing 2022 breakpoints.²¹ Multidrug resistance was defined as resistance to at least one agent in three or more antimicrobial classes. The bacterial species were identified by biochemical characteristics using the conventional biochemical gallery.²²

Bacteriological data were entered into a standardised database. Descriptive statistics, including frequencies and percentages of resistance, were computed using Epi Info 7.2.5.0. Resistance profiles were stratified by bacterial species (e.g. E. coli, K. pneumoniae, Acinetobacter baumannii), and patient group (pregnant women, delivery mothers, newborns). No inferential statistics were performed as the study was descriptive.

Figure 1 shows the prevalence of faecal carriage of ESCR-E and CRE among pregnant women, women at delivery, and newborns. High rates of ESCR-E carriage were observed in all groups: 94.9% (336/354) in pregnant women, 100% (322/322) in women after delivery, and 83.90% (276/329) in newborns. Carbapenem-resistant Enterobacterales carriage was also high compared with the African average, with 25.1% (89/354) in pregnant women, 28.88% (93/322) in women at delivery, and 6.07% (20/329) in newborns. These results indicate a substantial faecal carriage of MDR Enterobacterales in both mothers and their newborns.

In this study, antibiotic susceptibility testing was performed on isolates recovered from pregnant women. The Enterobacterales tested included E. coli (n = 96), K. pneumoniae (n = 88), Klebsiella aerogenes (n = 78), Citrobacter freundii (n = 18), Proteus mirabilis (n = 9), and Salmonella enterica (n = 22). In addition, A. baumannii (n = 25), a non-Enterobacterales Gram-negative species, was analysed separately (Table 1).

High resistance rates were observed for several antibiotics commonly used in clinical practice. Among the Enterobacterales, resistance to amoxicillin was nearly universal, ranging from 98.9% in E. coli to 100% in the other species, while ticarcillin resistance ranged from 94.8% to 100%. Resistance to third-generation cephalosporins, including cefixime, cefotaxime, ceftriaxone, and ceftazidime, was consistently high, ranging from 89.6% to 100% among Enterobacterales. Ciprofloxacin resistance ranged from 77.8% to 100% among Enterobacterales. Aminoglycoside resistance was also substantial, with gentamicin resistance ranging from 66.7% (P. mirabilis) to 90.9% (K. aerogenes), and amikacin resistance between 55.6% and 88.6%. Resistance to chloramphenicol and tetracycline was generally above 70% (Table 1).

Acinetobacter baumannii, analysed separately, displayed similarly high resistance: 100% to amoxicillin and ticarcillin; 100% to cefotaxime, ceftriaxone, and cefixime; and high resistance to ciprofloxacin (100%) and aminoglycosides (gentamicin 96.0%, amikacin 92.0%) (Table 1).

Notably, none of the isolates, whether Enterobacterales or A. baumannii, showed resistance to the combination of ceftazidime and tazobactam (Table 1).

These findings reveal a multidrug-resistance pattern among Enterobacterales colonising pregnant women, with A. baumannii exhibiting a similar but separate MDR profile, which could complicate empirical treatment strategies (Table 1).

The antibiotic susceptibility of isolates recovered from women at delivery was evaluated across nine bacterial species. The Enterobacterales included C. freundii (n = 18), E. coli (n = 87), K. aerogenes (n = 67), K. pneumoniae (n = 71), Morganella morganii (n = 9), P. mirabilis (n = 14), S. enterica (n = 22), and Serratia marcescens (n = 1). In addition, A. baumannii (n = 34) was analysed separately (Table 2).

Among Enterobacterales, resistance to amoxicillin and ticarcillin was nearly universal, ranging from 93% to 100%. Third-generation cephalosporins – including cefixime, cefotaxime, ceftriaxone, and ceftazidime – showed high resistance rates, generally exceeding 89%. Ciprofloxacin resistance varied from 58% in K. pneumoniae to 95% in C. freundii, with intermediate values observed in other species. Aminoglycoside resistance was substantial, with gentamicin resistance ranging from 68% to 88%, and amikacin resistance from 61% to 86%. Resistance to chloramphenicol and tetracycline was also high, generally between 53% and 79% (Table 2).

Acinetobacter baumannii, analysed separately from Enterobacterales, displayed similarly high resistance: 100% to amoxicillin and ticarcillin, 97% to cefixime, cefotaxime, ceftriaxone, and ceftazidime, and high resistance to ciprofloxacin (95%) and aminoglycosides (gentamicin 91%, amikacin 89%) (Table 2).

Notably, none of the isolates – whether Enterobacterales or A. baumannii – demonstrated resistance to the combination of ceftazidime and tazobactam (Table 2).

Overall, these results reveal a high prevalence of multidrug resistance among Enterobacterales colonising women at delivery, with A. baumannii exhibiting a separate but comparable MDR profile, representing a significant challenge for empirical antimicrobial therapy in this population (Table 2).

Among the 276 ESCR-E isolates recovered from newborns, three species predominated: E. coli (n = 120), K. aerogenes (n = 82), and K. pneumoniae (n = 74). Resistance to first-line antibiotics, including amoxicillin and ticarcillin, was universal across all isolates (100%), and all strains also showed complete resistance to cefixime (Table 3).

High levels of resistance were also observed to third-generation cephalosporins: 93.5% for ceftriaxone, 93.1% for ceftazidime, and 88.8% for cefotaxime. Ciprofloxacin resistance was moderate, affecting 62.7% of isolates overall, with slightly higher rates among K. pneumoniae (66.2%) and E. coli (63.3%) (Table 3).

Resistance to aminoglycosides was variable. While gentamicin resistance was alarmingly high at 94.6%, amikacin remained relatively effective, with only 26.1% of isolates resistant. Resistance to chloramphenicol (86.6%) and tetracycline (87.0%) was also consistently elevated across species (Table 3).

Notably, none of the ESCR-E isolates was resistant to the combination of ceftazidime–tazobactam (0%), indicating its potential retained efficacy (Table 3).

These findings highlight a worrying burden of MDR Enterobacterales among neonatal colonisers, with limited therapeutic options remaining, especially for aminoglycosides and third-generation cephalosporins (Table 3).

A total of 89 carbapenem-resistant isolates were obtained from pregnant women. Among them, the Enterobacterales included E. coli (n = 42), K. pneumoniae (n = 14), and K. aerogenes (n = 8). In addition, A. baumannii (n = 25) was also analysed (Table 4).

Across all Enterobacterales isolates, universal resistance (100%) was observed to β-lactam antibiotics, including amoxicillin, ticarcillin, cefixime, cefotaxime, ceftriaxone, and ceftazidime. Full resistance to ertapenem was also observed in Enterobacterales isolates, confirming their CRE phenotype. Ciprofloxacin resistance among Enterobacterales varied: 50% in K. pneumoniae, 81% in E. coli, and 100% in K. aerogenes. Gentamicin resistance reached 100% in all Enterobacterales, whereas amikacin resistance was lower in E. coli (62%) but 100% in K. pneumoniae and Klebsiella aerogenes. Resistance to chloramphenicol and tetracycline among Enterobacterales was 69% and 67% in E. coli, and 100% in K. pneumoniae and K. aerogenes. Resistance to imipenem and meropenem among Enterobacterales ranged from 84% to 100% and 43% to 100%, with the lowest meropenem resistance observed in K. pneumoniae (43%) (Table 4).

Acinetobacter baumannii, analysed separately, exhibited universal resistance to β-lactams (amoxicillin, ticarcillin, cefixime, cefotaxime, ceftriaxone, ceftazidime), ertapenem, gentamicin, amikacin, chloramphenicol, and tetracycline. Ciprofloxacin and carbapenem resistance were also high, with 100% of isolates resistant to ciprofloxacin, imipenem, and meropenem (Table 4).

Among the 93 carbapenem-resistant isolates collected from women at delivery, the Enterobacterales included E. coli (n = 43), K. pneumoniae (n = 19), and Klebsiella aerogenes (n = 8). Acinetobacter baumannii (n = 23), a non-Enterobacterales Gram-negative species, was analysed separately (Table 5).

Across all Enterobacterales isolates, resistance was universal (100%) to key β-lactam antibiotics, including amoxicillin, ticarcillin, cefixime, cefotaxime, ceftriaxone, and ceftazidime. Complete resistance to the carbapenems ertapenem and meropenem was also observed. Ciprofloxacin resistance among Enterobacterales ranged from 81.4% in E. coli to 87.5% in Klebsiella aerogenes, and gentamicin resistance ranged from 94.7% to 100%. Amikacin resistance was slightly lower in E. coli (81.4%) and K. pneumoniae (89.5%). Resistance to chloramphenicol and tetracycline ranged from 79.1% to 100%, with E. coli exhibiting the lowest rates. Resistance to imipenem varied among Enterobacterales, with 90.7% in E. coli and 75% in K. aerogenes, whereas K. pneumoniae showed complete resistance (100%) (Table 5).

Acinetobacter baumannii exhibited universal resistance to all β-lactams (amoxicillin, ticarcillin, cefixime, cefotaxime, ceftriaxone, ceftazidime), ertapenem, meropenem, gentamicin, amikacin, chloramphenicol, tetracycline, and ciprofloxacin. Resistance to imipenem and meropenem was complete (100%) in all A. baumannii isolates (Table 5).

Notably, all isolates – whether Enterobacterales or A. baumannii – remained fully susceptible to ceftazidime–tazobactam (0% resistance), indicating retained activity of this combination against CRE isolates from women at delivery (Table 5).

Among the 20 CRE isolates obtained from neonates, E. coli (n = 11), Klebsiella aerogenes (n = 5), and K. pneumoniae (n = 4) were identified. All isolates from the three species showed 100% resistance to first-line β-lactam antibiotics, including amoxicillin, ticarcillin, cefixime, cefotaxime, ceftriaxone, and ceftazidime (Table 6).

Resistance to ciprofloxacin was complete in K. pneumoniae, and high in E. coli (81.8%) and Klebsiella aerogenes (80.0%). Likewise, all isolates were resistant to gentamicin. Resistance to amikacin was slightly lower, with K. pneumoniae (75.0%) and K. aerogenes (80.0%) showing more susceptibility than E. coli (100%) (Table 6).

Regarding other antibiotic classes, resistance to chloramphenicol and tetracycline ranged from 75.0% to 100%, with K. pneumoniae isolates again showing relatively lower resistance. Notably, sensitivity to the combination ceftazidime–tazobactam was universal across all isolates (100%) (Table 6).

Resistance to carbapenems was also very high. All K. pneumoniae and K. aerogenes isolates were resistant to ertapenem, imipenem, and meropenem. Among E. coli isolates, resistance to ertapenem was 90.9%, while both imipenem and meropenem resistance were also 90.9% (Table 6).

In our study, we observed an alarmingly high faecal carriage of ESCR-E among pregnant women 94.9% during pregnancy and 100% at delivery, as well as a high rate (83.9%) among their newborns. The carriage of CRE was also substantial in mothers (25.1% in pregnant women; 28.9% at delivery) and notable in newborns (6.1%). These prevalences markedly exceed the rates generally reported in the literature, highlighting a potentially severe public health issue in our setting. The findings have been reported in a separate publication

This study investigated the AMR patterns of ESCR-E isolated from pregnant women, postpartum mothers, and their newborns in South Kivu, eastern DRC. Our findings highlight a critical public health threat, with widespread multidrug resistance compromising treatment options in maternal and neonatal care.

Resistance to third-generation cephalosporins – cefotaxime, ceftriaxone, ceftazidime, and cefixime – was uniformly high, exceeding 90% in all major species including E. coli, K. pneumoniae, A. baumannii, and K. aerogenes. This resistance profile is consistent with an ESBL phenotype, likely dominated by CTX-M-type enzymes, which are well documented in both hospital and community settings across sub-Saharan Africa.²³

Beyond β-lactams, resistance to fluoroquinolones was also alarmingly high. Over 85% of K. pneumoniae, A. baumannii, and E. coli isolates were resistant to ciprofloxacin, suggesting co-selection by means of plasmids carrying multiple resistance determinants. Such plasmids often harbor genes such as qnr, aac(6’)-Ib-cr, and qepA, which confer fluoroquinolone resistance and are frequently co-located with ESBL genes.^24,25 This co-resistance pattern reflects both poor antibiotic stewardship and widespread empirical use of broad-spectrum agents without microbiological confirmation, a scenario common in resource-limited settings.

Similarly, high resistance levels were observed for aminoglycosides such as gentamicin, as well as older agents such as tetracycline and chloramphenicol. The diminished efficacy of these traditionally first-line antibiotics severely restricts the therapeutic options available for treating neonatal and postpartum infections, potentially increasing morbidity and mortality in these vulnerable populations.

Interestingly, ceftazidime–tazobactam remained fully active against all ESCR-E isolates in this study. This suggests either limited prior use or absence of resistance mechanisms targeting β-lactam/β-lactamase inhibitor combinations in the region. As such, this drug may serve as a critical therapeutic option in severe infections, particularly where carbapenems are inaccessible. However, caution is warranted: resistance to this combination has emerged rapidly elsewhere because of misuse, and over-reliance could threaten its effectiveness.^26,27

Our findings are consistent with other studies from sub-Saharan Africa. In Tanzania, over 95% of neonatal E. coli isolates were resistant to third-generation cephalosporins, and over 80% to ciprofloxacin.^28,29 In Nigeria, maternal ESBL-E colonisation was strongly associated with neonatal ESBL-E colonisation,³⁰ and in Ethiopia, high ESBL production rates with fluoroquinolone co-resistance were also reported.³¹ This convergence in resistance profiles across East and West Africa underscores shared systemic drivers – poor antimicrobial regulation, limited diagnostics, and inadequate infection prevention in maternity wards.

Such conditions foster the propagation of ESCR-E and other MDR organisms, exacerbating already fragile healthcare systems. The presence of these resistant pathogens not only limits treatment efficacy but also heightens the risk of nosocomial transmission and therapeutic failure. This emphasises the need for coordinated interventions, including improved hygiene protocols, enhanced antimicrobial stewardship, and community-level awareness campaigns. These efforts must be reinforced by robust laboratory-based surveillance systems capable of tracking resistance trends and guiding empirical therapy in real time.

In addition to ESCR-E, this study identified an equally concerning resistance profile among CRE isolates. All CRE isolates exhibited universal resistance to penicillins and third-generation cephalosporins, strongly suggesting the presence of multiple β-lactamases, including both ESBLs and carbapenemases.³²

The most likely mechanisms include carbapenemase genes such as bla_KPC, bla_NDM, bla_VIM, bla_IMP, and bla_OXA-48-like, which encode KPC, NDM, VIM, IMP and OXA-48 carbapenemase enzymes. In Africa, NDM and OXA-48-like carbapenemases are particularly prevalent and have been reported in neonatal outbreaks.^33,34

These enzymes confer broad-spectrum resistance and are typically plasmid-borne, facilitating horizontal gene transfer within healthcare environments.

All CRE isolates in this study also showed high resistance to ciprofloxacin and gentamicin, particularly among A. baumannii, K. aerogenes, and K. pneumoniae. This co-resistance likely reflects plasmid-mediated gene clusters, which include fluoroquinolone resistance determinants such as qnr and aminoglycoside-modifying enzymes such as aac(6’)-Ib, often accompanied by efflux pump overexpression.^35,36 Resistance to additional agents – tetracycline, amikacin, and chloramphenicol – further classifies these strains as MDR and highlights the limited therapeutic landscape in South Kivu.

Remarkably, no phenotypic resistance to ceftazidime–tazobactam was detected among the CRE isolates under the testing conditions used in this study. However, previous in vitro studies have reported variable activity of this combination against Gram-negative bacteria, depending on the bacterial species and underlying β-lactamase mechanisms.^37,38 Because molecular characterisation of the isolates was not performed, the mechanisms underlying the observed activity could not be determined. Therefore, these findings should be interpreted cautiously, and phenotypic confirmation remains essential before empirical use.

Colistin, often considered the last resort for CRE infections, was not tested in this study. However, its use remains extremely rare in the region, and the absence of routine testing for plasmid-mediated resistance (e.g. mcr genes) represents a critical gap. The global spread of mcr-positive strains is deeply concerning because of the risk of horizontal gene transfer and community-level dissemination.^39,40

The emergence of CRE in high-risk populations such as mothers and neonates demands urgent attention. These infections are associated with high mortality, prolonged hospital stays, and significant economic burden.^32,41 Yet in much of sub-Saharan Africa, surveillance of carbapenem resistance is either weak or nonexistent. Laboratory testing for carbapenemases is not routine, leading to underestimation of prevalence and missed opportunities for containment.⁴²

To address this challenge, investments in diagnostic infrastructure are essential, alongside Infection Prevention and Control training and strict regulation of antibiotic use. The limited but preserved susceptibility to ceftazidime–tazobactam must be safeguarded through targeted, rational use supported by local susceptibility data. The implementation of national AMR control programmes, coupled with international collaboration, is crucial for early detection and containment of CRE and other high-risk pathogens.

This study has several important limitations. Firstly, its cross-sectional design limits the ability to draw causal inferences or assess temporal resistance trends. Secondly, the absence of molecular analysis precluded confirmation of the genetic basis of resistance, such as the presence of bla_CTX-M, bla_NDM, or mcr genes. Thirdly, susceptibility testing did not include colistin, which is a key last-resort antibiotic in the context of carbapenem resistance. Fourth, the study was limited to selected healthcare facilities in South Kivu and may not reflect resistance patterns across other regions of the DRC. Fifth, because of resource constraints, not all antibiotics could be tested, and only phenotypic methods were employed, which may fail to detect certain resistance mechanisms. Finally, minimum inhibitory concentrations were not determined, limiting the precision in assessing the degree of bacterial susceptibility or resistance.

Our findings reveal a high prevalence of ESCR-E and CRE carriage among pregnant women and neonates in South Kivu, indicating a concerning burden of MDR Enterobacterales in the maternal–neonatal ecosystem. The universal resistance to third-generation cephalosporins, alongside elevated resistance to aminoglycosides, fluoroquinolones, and other agents, severely restricts therapeutic options in this resource-limited context. The preserved susceptibility to ceftazidime–tazobactam offers a narrow but critical treatment opportunity that must be safeguarded.

Strengthening antimicrobial stewardship, improving access to microbiology diagnostics, and reinforcing infection prevention and control are urgently required to mitigate potential mother-to-child transmission and broader community dissemination. These interventions should be guided by local surveillance data and adapted to the operational realities of low-resource healthcare systems.

Given the impact of antimicrobial exposure on intestinal microbiota, complementary strategies such as targeted probiotics may support decolonisation and bolster current infection prevention and control efforts. Future studies assessing the dynamics of acquisition and transmission within mother–newborn pairs are warranted.

Sustained integration of AMR surveillance into maternal and neonatal care, aligned with national and global action plans, is essential to protect vulnerable populations and preserve the effectiveness of life-saving antibiotics.