A hospital-based cross-sectional study was conducted from February to May 2020 at MNH, the largest tertiary hospital in Tanzania. MNH serves as a university teaching hospital with 1,500 bed capacity and attends to approximately 2,000 outpatients daily. MNH have more than 100 bed capacities for neonatal intensive care unit (NICU) and neonatal wards. On average, 15 to 20 neonates are admitted daily, translating to approximately 450 to 500 admissions per month.

The study enrolled neonates aged ≤ 28 days whose parents/guardians consented to participate, excluding neonates with severe congenital malformations such as anal atresia. The sample size was calculated using the Kish Leslie formula for cross-sectional studies, considering a 25.4% prevalence of ESBL-PE reported in a study in Tanzania⁹ and 27.4% for CPE reported in Luanda, Angola.¹¹ The minimum sample size required was 291 for ESBL-PE prevalence, and 306 for CPE prevalence. A total of 340 neonates were conveniently recruited from the NICU and neonatal wards to meet the study objectives.

A structured data collection tool was used to collect socio-demographic information, hygienic behavioural characteristics, and clinical information of paired neonates and their mothers. Socio-demographic information included; age, sex, physical address, parents’ or guardians’ level of education, and neonates’ birth weight. The mother’s hygienic behavioural characteristics included; toilet use, food handling, hand washing, and new born bathing practices. Clinical information of neonates included; history of antibiotic use, mode of delivery, invasive procedures, duration of hospital exposure, previous hospitalisation, gestation age, and comorbidities. Maternal clinical information collected encompassed sexually transmitted infections (STI) such as; Human Immunodeficiency Virus (HIV) and Syphilis, other comorbidities (e.g. pre-eclampsia, eclampsia, anaemia, gestational diabetes), and history of antibiotic use during pregnancy.

Birth weight below 2.5 kg was classified as low birth weight. Pre-term birth was defined as a gestational age at birth of less than 37 weeks. Hospital stay referred to the duration neonates spent in hospital beds during the study period. Previous hospitalisation was defined as any hospital admission occurring within a specified time prior to neonate’s current hospitalisation. Sexually transmitted infections (STI) and other comorbidities like (preeclampsia, eclampsia, anaemia, and gestational diabetes) during pregnancy may contribute to the presence of antimicrobial resistance organisms. Pregnant women with these conditions often require close monitoring from healthcare workers, hence frequent hospital visits or hospitalisation, hence increasing their exposure to contaminated hospital environments or items, potentially transmitting resistance genes to their new-borns. Multi-drug resistance (MDR) bacteria is defined as antimicrobial resistance expressed by a bacteria to at least one antimicrobial drug in three or more antimicrobial classes. ESBL-PE are Gram-negative bacteria in the human gastrointestinal tract. They can produce ESBL enzymes that confer resistance to different beta-lactam antibiotics, including first, second, third, and fourth-generation cephalosporins, as well as aztreonam. CPE are Gram-negative bacteria that produce enzymes (carbapenemase) that inactivate carbapenems (meropenem, imipenem, doripenem, and ertapenem), and several other classes of beta-lactam antibiotics such as penicillin, cephalosporin, and monobactam.

The rectal swab was collected by gently inserting a moist sterile swab approximately 1 to 2 inches into the anal canal. The swab was then inserted into Cary-Blair transport media and transported to the microbiology laboratory at MUHAS to process and identify the bacteria.¹⁶ For screening ESBL-PE and CRE, the rectal swabs were inoculated onto two MacConkey agar (MCA) (Oxoid Ltd UK); one supplemented with ceftazidime at 2 μg/ml and another supplemented with meropenem at 1 μg/ml for ESBL-PE and CRE screening, respectively. Then, plates were incubated aerobically at 37 ºC for 18 to 24 hours.^17,18

Isolates were subcultured on Nutrient Agar (NA) for further analysis, including conventional biochemical tests such as oxidase, urease, citrate utilisation, Kligler Iron Agar (KIA), and SIM tests for bacterial identification.¹⁹ In addition, Analytical Profile Index (API-20E, bioMérieux, Marcy-l’Étoile, France) test was used to identify inconclusive isolates. Finally, isolates were confirmed as ESBL-PE using the Double-Disc Diffusion Synergy Test (DDST) and as CPE using the Modified Carbapenem Inactivation Method (mCIM).^20,21 In addition, Antimicrobial Susceptibility Testing (AST) was performed using the Kirby Bauer disk diffusion method.²² Antimicrobial disks, meropenem (10µg), amikacin (30µg), gentamicin (10 μg), amoxicillin/clavulanic acid (20/10 μg) and ciprofloxacin (5 µg) were used for ESBL-PE isolates. Regarding antimicrobial susceptibility pattern for CPE aztreonam (30 µg), ciprofloxacin (5 µg), amikacin (30µg), gentamicin (10 μg), cefepime 30 µg and amoxicillin/clavulanic acid (20/10 μg) antimicrobial disks were used for testing CPE isolates.²³ The inhibition zones were measured and interpreted according to the Clinical Laboratory Standard Institute (CLSI) 2017 guideline.

The culture media were prepared following the manufacturer’s guidelines and subjected to quality-control to ensure performance and sterility. A 0.5 McFarland standard was used to standardise the turbidity of bacterial suspensions for ESBL-PE and CPE testing. As per CLSI guidelines, Klebsiella pneumoniae (ATCC-700603) and E.coli (ATCC-25922) were used as positive and negative control bacteria strains, respectively, for ESBL-PE. Also, for CPE, K.pneumoniae (ATCC-1705) and E.coli (ATCC-25922) were used as positive and negative control bacteria strains, respectively. ²¹

Descriptive analysis was performed using medians with interquartile range (IQR) for continuous variables and proportions for categorical variables. Univariable and multivariable logistic regression models were used to determine the factors independently associated with faecal carriage of ESBL-PE. Factors with a p-value of less than 0.1 were subjected to multivariable logistic regression analysis using a backward stepwise elimination method noting their Odds Ratios (OR) and 95% Confidence Intervals (CI). Fisher’s exact test was used to determine the factors associated with faecal carriage of CPE. A p-value of ≤.05 was considered statistically significant for all the statistical tests.

The study obtained ethical approval from the Institutional Review Board (IRB) of the Muhimbili University of Health and Allied Sciences (MUHAS) with reference number DA.287/298/01A. Written informed consent was obtained from the mothers or guardians of each study participant. For neonates identified with drug-resistant strains, the findings were promptly shared with the attending clinicians for proper action.

A total of 340 neonates were enrolled, with a median age of 3 days (interquartile range (IQR) 2–9) and median birth weight of 2.01 kg (IQR: 1.43–2.84). The majority, 185/340 (54.4%), were males, 243/340 (71.5%), were aged ≤ 7 days, and 220/340 (64.7%) had a birth weight of < 2.5 kg. About 98 % of the participants were delivered at the hospital, and 62.7% were admitted to the NICU (Table 1).

The majority of the neonates’ mothers, 220/340 (64.7%), shared toilets with family members, while 96/340 (28.2%) shared with the community, and 24/340 (7.1%) had self-contained toilets. Twenty-four neonates (7.1%) were bathed, and of these, 23/24 (95.8%) used stagnant water, and 13/24 (54.2%) used medicated soap. A total of 320/340 (94.1%) neonates did not have physical contact with other people besides their mothers and hospital caregivers. Moreover, 25/340 (7.3%) neonates were sucking fingers (Table 1).

The study observed that 235 (69.1%) neonates had used antibiotics such as ampiclox, gentamicin, and ceftriaxone either once or more than once during their hospital stay. Furthermore, 100/340 (29.4%) neonates had a history of previous hospitalisation before the current admission. The majority, 287/340 (84.4%) neonates, were hospitalised for ≤ 7 days. In addition, a total of 230/340 (67.7%) neonates had a premature birth, and 189/340 (55.6%) were delivered by Spontaneous Vaginal Delivery (SVD) (Table 1).

We observed that 232/340 (68.2%) mothers did not use antibiotics during pregnancy, and regarding HIV status, the majority, 321/340 (94.4%), were negative for HIV (Table 1). In addition, mothers of the study participants were diagnosed with several comorbidities during the current pregnancy, with UTI being the most frequent comorbidity, accounting for 26.5%, followed by preeclampsia at 20%, malaria at 4.4%, eclampsia at 4.1%, and chronic hypertension 2.9.

Faecal carriage of ESBL-PE was detected in 134/340 (39.4%) neonates, whereby 166 Enterobacterales were isolated. The most commonly isolated ESBL-PE was K. pneumoniae 111/166 (66.9%), followed by E. coli, 36/166 (21.7%). Faecal carriage of CPE was detected in 6/340 (1.8%) neonates, whereby three different species were isolated. The most commonly isolated CPE was E. coli 4/6 (66.7%), followed by K. pneumoniae 1/6 (16.7%) and Kluyvera spp 1/6 (16.7%) (Figure 1).

On univariable logistic regression analysis, it was observed that when age increased by one day, the odds of carrying ESBL-PE increased by 1.17 times (cOR 1.17, 95% CI: 1.12–1.22, p<.001). The neonates admitted to the NICU had 1.54 odds of carrying ESBL-PE than those admitted to other neonatal wards (Table 2). Giving a bath to a newborn had five times the odds of carrying ESBL-PE (cOR 5.17, 95% CI: 1.99–13.40, p=.001) compared to neonates who did not receive a bath. A neonate-sucking finger was approximately three times more likely to carry ESBLPE than neonates who did not suck their fingers (cOR 2.97, 95% CI: 1.27–6.93, p=.012) (Table 2). The neonates who used antibiotics were five times more likely to have ESBL-PE (cOR 5.13, 95% CI: 2.88–9.16, p<.001) than those who did not. Furthermore, neonates with a history of hospitalisation were approximately three times more likely to carry ESBL-PE (cOR 2.81, 95% CI 1.74–4.53, p<.001). Neonates hospitalised for > 7 days were eight times more likely to carry ESBL-PE (cOR 8.09, 95% CI: 3.98–16.44, p<.001) compared to being hospitalised for ≤ 7 days. Premature neonates had 1.72 times more likely to carry ESBL-PE than full-term neonates (cOR 1.72, 95% CI: 1.06–2.78, p=.027). In addition, neonates whose mothers were HIV positive were four times more likely to carry ESBL-PE than mothers who were HIV negative (cOR 4.69, 95% CI: 1.65–13.35, p=.004) (Table 2).

On multivariable logistic regression analysis, it was observed that age (aOR 1.09, 95% CI: 1.02–1.15, p=.006) was independently associated with faecal carriage of ESBL-PE. The use of antibiotics (aOR 2.73, 95% CI: 1.38–5.39, p=.04), hospitalisation for >7 days (aOR 2.92, 95% CI: 1.17–7.29, p=.02), and neonate sucking their fingers (aOR 2.98, 95% CI: 1.04–8.58, p=.04) were also independently associated with faecal carriage of ESBLPE (Table 2). In this study, only prolonged hospitalisation showed a trend toward association with CPE carriage (p=.05)

ESBL-PE showed variable antibiotic resistance to gentamicin (19.4%-100 %) and amoxicillin/clavulanic acid (11.1%–100%). The least antibiotic resistance was observed in meropenem, whereby only 1(0.9%) K. pneumoniae isolate was resistant. Resistance to amikacin was low (2.7%-6.7%). Resistance to ciprofloxacin was observed among E. aerogenes 1(100%), R. ornithinolytica (66.7%), K. pneumoniae 28(25.2%), E. coli 14(38.9%) and E. cloacae 4(26.7%). Among ESBL-PE isolates, K. pneumoniae was found to resist most antibiotics tested, while R. ornithinolytica had the least resistance compared to all other isolates (Table 3).

All CPE isolates showed (100%) resistance to tested drugs such as ciprofloxacin, cefepime, amoxicillin/clavulanic acid, and aztreonam. Furthermore, E. coli showed 100% resistance to gentamicin. However, all isolates were sensitive to amikacin (Table 3).

It was not possible to perform the genotypic tests for molecular characterisation of ESBL-PE and CPE genes due to budget constraints. However, the isolates have been stored at −80° centigrade for further analysis once funds become available.