Bacteriological profile and antibiotic susceptibility of early-onset neonatal bacterial infection in a neonatal unit in Cameroon

Abstract

Background

Neonatal infection is a major public health problem, especially in developing countries. According to the World Health Organisation, about 2.3 million newborns die every year, with infection being the main cause. This remain a major concern for clinicians due to the emergence of antibiotic resistance.

Objectives

The aim of this study was to assess the bacteriological profile and antibiotic susceptibility of Early-onset neonatal bacterial infections at the Douala Gynaecological-Obstetric and Paediatric Hospital (DGOPH).

Methods

This was an 18-month descriptive cross-sectional study. We included newborns aged 0 to 3 days who were admitted from 1st January 2020 to 31st January 2024 of the neonatology department of the DGOPH. Bacterial strains were identified using the Enterotube system and the antibiogram was performed using the disc diffusion method. Data analysis was performed using SPSS software and Pearson’s Chi-squared test was used to assess the level of difference between proportions.

Results

During the study period 1520 neonates were admitted in the neonatal unit, 761 were suspected of having an early neonatal infection. On the suspected cases, 179 had a positive bacteriological culture, giving a hospital prevalence of confirmed early neonatal infection of 23.5%. Gram-negative bacilli were the most common isolates (97.2%): Escherichia coli (22.7%), Burkholderia cepacia (18.8%) and Klebsiella pneumoniae (7.7%). Fluoroquinolones and penems were the classes of antibiotics that showed in vitro susceptibility: ciprofloxacin (85.4%), ofloxacin (88.7%), levofloxacin (91.1%) and meropenem (87.8%). The main bacterial agents showed 66.7% resistance to cefotaxime. Mortality was high, with 5.5% of deaths related to confirmed early neonatal bacterial infection.

Conclusion

This study shown a high prevalence of early neonatal infection with a variability of the local ecology. Also there was a high resistance of bacterial pathogens to the common antibiotics. All these findings were associated with a high mortality which is suggesting that a timely identification and initiation of effective antibiotherapy is necessary to reduce morbidity and mortality associated with neonatal infection.

Introduction

Early-onset neonatal bacterial infection (EONI) is defined clinically by signs and symptoms, affecting all children from 0 to 72 h of age, and biologically by a positive culture of bacterial pathogens in the blood or cerebrospinal fluid (CSF) [1]. It is a public health problem due to its frequency and high mortality rate. Worldwide, the number of newborn deaths has fallen from 5 million in 1990 to 2.4 million in 2019 and 2.3 million in 2022 [2]. In 2022, 47% of all deaths in children under the age of 5 involved newborns. Despite significant progress in child survival, Sub-Saharan Africa has the highest neonatal mortality rate, with 27 deaths per 1,000 live births, followed by Central and South Asia with 24 deaths per 1,000 live births in 2019 and 21 deaths per 1,000 live births in 2022 [2, 3]. In Cameroon, neonatal mortality, although still high, is decreasing: it felt from 32‰ in 2004 to 28‰ in 2020, with complications at the time of delivery, neonatal infection and prematurity as the main causes [4]. Various clinical manifestations including fever, neurological disorders, and respiratory distress are associated with neonatal infection [5, 6]. Gram-negative bacilli such as Klebsiella species and Enterobacteriaceae, and coagulase-negative Gram-positive staphylococci (CoNS) are the most common pathogenic bacteria in local and international studies [6–8]. In Cameroon, a study carried out by Chiabi et al. have identified gram-negative bacilli such as Klebsiella spp, Escherichia coli and Enterobacter spp [9]. The emerging antibiotic resistance in these pathogenic bacteria makes their management challenging in our context. Knowledge of the bacterial ecology and their antibiotic susceptibility is a key element in the early and effective management of neonatal infection [10]. Although the bacterial profile in some neonatology units in Cameroon has been assess in some studies, few of them focused specifically on early -onset neonatal bacterial infections and their susceptibility or resistance to commonly used antibiotics. This gap in knowledge of the local bacterial ecology and associated resistances reduces the effectiveness of treatment protocols and can delay the management.

Methods

Study site

We conducted a descriptive cross-sectional study from January 1st 2020 to January 31st 2024 with retrospective and prospective data collection in the paediatric department of DGOPH. This is a Category I hospital located in Douala (Yassa), Cameroon. In this hospital, the paediatrics department is made up of three units: the general paediatrics unit, the ambulatory care unit and the neonatology unit. The neonatology department has been operational since 1st August 2016 with two sub-units: one internal for newborns born at the DGOPH and the other external for children born out of the DGOPH or admitted after their discharge from maternity with an average number of 525 admissions per year according to DGOPH records. The medical staff includes a neonatologist, a neuropaediatrician, three general practitioners and nineteen nurses. The DGOPH also has an equipped laboratory to perform bacteriological diagnosis of neonatal bacterial infections (haemoculture, CSF analysis).

Study population and inclusion criteria

We included in our study all neonates aged 0–72 h admitted in the neonatology department of the DGOPH for suspected neonatal infection on the basis of the presence of at least one maternal and neonatal infectious risk criterion and/or clinical signs of infection. The following criteria selected were as infectious risk: rupture of the membranes for more than 12 h, Spontaneous and unexplained prematurity, maternal fever > 38.0 °C intrapartum, meconium stained amniotic fluid, genitourinary infections of the third trimester, home delivery [11].

Procedure and data collection

Data collection was carried out at the hospital after the participant information sheet and informed consent form had been signed. The retrospective phase involved collecting data from medical records between 1st January 2020 and 31st December 2022. The prospective phase stage included, in consecutive, newborns suspected of having early neonatal bacterial infection between 1st January 2023 and the 31st January 2024. The patient’s informations in the register were entered into a main list which only the researchers had access and served as the database for data analysis and results. The questionnaire was divided into four main sections: sociodemographic data, bacteriological profile, antibiotic susceptibility, and outcomes for admitted patients. This data were collected from information contained in patients’ medical records as well as information provided by parents.

In the case of any suspected neonatal infection admitted to neonatology, approximately 1 to 2 ml of peripheral or central blood were collected with rigorous aseptic measures, before initiating a broad spectrum antibiotic therapy. Then immediately, the blood was inoculated into a Bact/Alert FA Plus culture broth. Lumbar puncture was performed after rigorous asepsis if the patient’s neurological condition worsened and/or if blood cultures were positive, and a volume of 1–2 ml of fluid was obtained by inserting the puncture needle into the L3-L4 space and inserting it into Bact/Alert FA culture broth. These vials were sent directly to the laboratory. Bacterial strains were identified using the Enterotube system and the antibiogram was performed using the disc diffusion method.

Operational definitions

Date of onset of infection: date of collection of the first sample yielding a pathogenic strain.

Early neonatal infection: infection occurring within the first 72 h after birth.

Neonatal infection suspected: presence of at least one maternal or neonatal risk factor, with or without clinical signs.

Confirmed case of neonatal infection: any newborn presenting one or more infectious risk factors and/or clinical signs and a positive culture.

Positive culture: culture isolating a bacterial agent.

Ethical considerations

This study was approved by the Faculty of Medicine and Pharmaceutical Sciences of the University of Dschang (N° B13/4171/23/UDs/FMSP/D/VD-RC) and received ethical clearance (N°2023/00935/HGOPED/DG/CEI), from the DGOPH institutional committee. In addition, an authorization was issued by DGOPH (N° 2023/0475/L/HGOPED/DG/DFRI). Informed consent forms were signed by parents who agreed to participate in the study. Each patient was assigned a unique code to ensure anonymity.

Statistical analysis

The data previously collected on a questionnaire sheet were transferred to the Kobotoolbox collection software version 2023.2.4. Once the database had been cleared, the data were exported to IBM SPSS version 26.0 for statistical analysis. Frequencies and percentages were calculated for categorical variables. A mean and standard deviation were used to estimate continuous quantitative variables according to the distribution of the data. Results were presented in tables, bar charts and pie charts. The Pearson chi-square test was used to assess the level of difference between the proportions, and a p-value of less than 0.05 within a 95% confidence interval was considered statistically significant.

Results

Prevalence of confirmed early bacterial neonatal infection at DGOPH

During the 49 months of the study, we recorded 1520 admissions at the neonatology unit. Among them, 761 newborns were admitted for suspected early neonatal bacterial infection, representing 50% (761/1520) of hospital admissions. On the 761 suspected cases, 179 were confirmed as early neonatal bacterial infection, representing a prevalence of 23.5% (179/761) excluding fungal and viral infections.

Socio-demographic characteristics of newborns admitted for confirmed EONI to DGOPH

Among the newborns admitted with confirmed EONI, 120/179 (67.0%) were born at DGOPH. Female were the most affected by neonatal infection, with 90/179 (50.3%) cases, resulting in a sex ratio of 0.9:1, male to female. (Table 1)

Table 1.

Socio-demographic characteristics of mothers and infectious risk criteria for confirmed EONI

Variables	n	%
Gender
Feminine	90	50.3
Masculine	89	49.7
Place of birth
DGOPH	120	67.0
Out of DGOPH	59	33.0
Mother’s age in years
15–25	25	14.0
25–35	126	60.4
35–45	28	15.6
Marital status
Bride	120	67.0
Single	56	31.3
Concubinage	3	1.7
Profession
Employee	128	71.5
Non employee	51	28.5
Infectious risk
Spontaneous prematurity < 37 weeks	58	26.0
Premature rupture of membranes	45	20.2
Membrane rupture > 12 h	33	14.8
Neonatal resuscitation	20	9.0
Meconium amniotic fluid	30	13.5
Fetal heart rate abnormalities	19	8.5
Peripartum maternal fever	14	6.3
Genitourinary infection in 3rd semester	1	0.4
Foul amniotic fluid	2	0.9
Home delivery	1	0.4

DGOPH  Douala gynaeco-obstetric and paediatric hospital

Socio-demographic characteristics of mothers and infectious risk criteria for confirmed EONI

Two-thirds (126, 60.4%) of the mothers were aged between 25 and 35, 120 (67.0%) were married and 128 (71.5%) were employed. Spontaneous prematurity (26.0%), premature rupture of membranes (20.2%) and prolonged rupture of membranes > 12 h (14.8%) were the most frequent infectious risk criteria. There was only one case of a home delivery (Table 1).

Clinical caracteristics of newborns with confirmed EONI

Among the newborns presenting confirmed EONI, 131/179 (73.2%) were admitted within the first 24 h of life. More than two-thirds of the population (61.4%), were born vaginally. Premature babies (62.0%) were the most affected group, and 44.7% of the population had a low birth weight (< 2500 g) (Table 2).

Table 2.

Clinical parameters and manifestations of newborns with confirmed EONI

Variable	n	%
Age on admission (in days)
H0-H23	131	73.2
H24-H47	24	13.4
H48-H72	24	13.4
Gestational age at birth
< 28 SA	8	4.5
28-31SA + 6 J	23	12.8
32-36SA + 6j	80	44.7
≥ 37SA	68	38.0
Delivery mode
Cesarean section - Emergency	54	30.2
Cesarean section - Elective	15	8.4
Low track	110	61.4
Birth weight
< 1000 g	3	1.7
1000–1500 g	19	10.6
1500–2000 g	31	17.3
2000–2500 g	27	15.1
2500–3500 g	72	40.2
≥ 3500 g	27	15.1
Clinical signs
Respiratory distress	121	31.9
Fever	20	5.3
Hypothermia	37	9.8
Tachypnea/Bradypnea	70	18.5
Neurological disorders	44	11.6
Cyanosis/acrocyanosis	51	13.5
Jaundice	10	2.6
Digestive disorders	10	2.6
Poor adaptation to extra-uterine life (asphyxia)	8	2.1
ORL Signs	3	0.8
Other clinical signs (Omphalocele)	2	0.5
Hypoglycemia	0	0.0
Signs of shock	3	0.8
Splenomegaly	0	0.0

Clinical signs observed for confirmed EONI

The most frequent clinical signs were respiratory distress, tachypnoea/bradypnoea and thermoregulation abnormalities in 121 (31.9%), 70 (18.5%) and 57 (15.1%) neonates respectively admitted for confirmed early onset neonatal infection. (Table 2)

Bacteriological profile

Among the 761 cultures taken, 179 were positive, mainly from blood cultures (100%). No CSF analysis yielded any bacterial pathogens. Gram-negative bacilli were the most common isolates (97.2%): Escherichia coli (22.7%), Burkholderia cepacia (18.8%) and Klebsiella pneumoniae (7.7%) (Table 3). Gram-positive bacteria were rarely found, represented by coagulase-negative Staphylococcus (CONS) (2.2%) and Streptococcus mitis (0.6%).

Table 3.

Profile of bacterial pathogens responsible for EONI at DGOPH

Isolated bacterial pathogens	n	%
Gram negative bacilli Enterobacteriaceae
Escherichia coli	41	22.7
Klebsiella pneumoniae	14	7.7
Enterobacter spp	10	5.5
Serratia Liquefaciens	12	6.6
Hafnia Alvei	6	3.3
Klebsiella spp	6	3.3
Salmonella spp	8	4.4
Serratia Odorifera	2	1.1
Citrobacter Freundii	4	2.2
Proteus Mirabilis	7	3.9
Raoultella Ornithinolytica	3	1.7
Serratia Marcescens	1	0.6
Other isolated bacterial pathogens	13	7.2
Non fermentative gram negative bacilli
Burkholderia Cepacia	34	18.8
Pseudomonas spp	7	3.9
Acinetobacter Calcoaceticus	4	2.2
Acinetibacter metnnii	4	2.2
Gram positive cocci
Staphylococcus aureus	0	0.0
CoNS*	4	2.2
Streptococcus mitis	1	0.6
*CoNS (Staphylococcus epidermidis, Staphylococcus haemolyticus, Staphylococcus xylosus)

Burkholderia cepacia was present in high frequency in preterm neonates, neonates born at DGOPH, born by cesarean section and those admitted to the internal neonatology department with a statistically significant difference in proportion (p-value < 0.05) (Table 4).

Table 4.

Frequency and distribution of Burkholderia Cepacia

Variables	Effective (n)	Effective (n)	Test
Gestationnal age at birth	< 37 weeks	≥ 37 weeks	Chi-2 pearson	p-value
Burkholderia Cepacia	24	10	1.31	0.25
Others	87	58
Place of birth	DGOPH	OUT DGOPH	Chi-2 pearson	p-value
Burkholderia Cepacia	22	12	0.12	0.71
Others	89	56
Delivery mode	Vaginal delivery	Cesarean section	Chi-2 pearson	p-value
Burkholderia Cepacia	15	19	5.70	0.01
Others	96	49
Admission mode	Internal unit	External unit	Chi-2 pearson	p-value
Burkholderia Cepacia	22	12	0.12	0.71
Others	89	56

Figure 1 shows the annual trend in the predominant bacteria from 2020 to 2024.

Fig. 1.

Annual trend of the main pathogenic strains of early neonatal bacterial infection in DGOPH

Antibiotic susceptibility

Overall, low bacterial pathogens sensitivity was obtained with ampicillin (0%), amoxicillin (5.6%) and cefotaxime (33.3%). The most effective antibiotics were fluoroquinolones: ciprofloxacin (85.4%), ofloxacin (88.7%), levofloxacin (91.1%), pefloxacin (91.3%) and penems including meropenem (87.8%) (Table 5).

Table 5.

Comparative table of bacterial sensitivity and resistance to antibiotics

Antibiotics	E.coli (S) %	E.coli (R) %	B. cepacia (S) %	B. cepacia (R) %	K. pneumoniae (S) %	K. pneumoniae (R) %
Ciprofloxacine	95	5	60	40	100	0
Levofloxacine	100	0	80	20	100	0
Ofloxacine	100	0	80	20	80	20
Imipénème	100	0	90	10	100	0
Méropénème	75	25	85	15	100	0
Céfoxitine	100	0	100	0	100	0
Ceftriaxone	100	0	40	60	45	55
Ceftazidime	80	20	60	40	35	65
Cefuroxime	50	50	75	25	25	75
Amoxicilline	0	100	0	100	0	100
Ampicilline	0	100	0	100	0	100
Amoxi + Ac. Clav	0	100	0	100	0	100
Amikacine	100	0	30	70	100	0
Gentamycine	55	45	70	30	75	25
Pipéracilline	100	0	100	0	40	60
Pipéracilline + Tazobactam	100	0	100	0	100	0
Ticarcilline	100	0	100	0	30	70

Outcome of newborns hospitalised for confirmed EONI

Among the 761 patients admitted for suspected EONI in our study, 42 neonates who had a positive culture died, giving a mortality rate of 5.5% for confirmed early bacterial neonatal infection. The neonatal mortality rate linked to confirmed EONI at the DGOPH was 14 deaths per 1000 live births (42/2947*1000), with 2947 representing the number of births recorded in the maternity department and hospitalisations in the outpatient neonatology department of the DGOPH.

Discussion

The aim of our study was to assess the bacteriological profile and antibiotic susceptibility of EONI at the DGOPH.

The hospital prevalence of confirmed early bacterial neonatal infection in our study was 23.5% (179/761) with a high prevalence of premature (62.0%) and low birth weight (44.6%). This result is similar to that found by Kemeze et al. in Cameroon in 2016, which was 24%, to the overall prevalence of bacterial neonatal infection in East Africa (29.7%) in 2020 and to the 22.7% obtained by Thomas et al. in Niger in 2022 [5, 12, 13]. Aku et al. in 2018 in Ghana obtained a lower prevalence of 17.3%, Konate et al. in 2019 in Mali,11.04%, and Yadav et al. in 2018 in Nepal, 16.9% [11, 14, 15]. This difference could be explained by a difference in sample size, which was larger in our study due to the fact that DGOPH is a referral hospital for high-risk pregnancies and newborns, with a high request for bacteriological sample. Neonatal infection is a real problem in hospitals with limited resources for the adequate management of newborns [16], the high prevalence of neonatal bacterial infection in our study could be due to the fact that the majority of our population was premature and low-birth-weight newborns admitted in their first week of life, who are at high risk of developing bacterial infection due to an humoral immune deficiency [5, 14, 17–19].

The bacterial ecology was dominated by gram-negative bacilli (97.2%). A similar result was found by Kemeze et al. in 2016 in Cameroon, Chemsi et al. in Morocco, and Jatsho et al. in a study published in 2020 in Bhutan in referral hospitals similar to ours [5, 17, 20]. Different results were described by Folquet et al. in Ivory Coast in 2016, by Aku et al. in Ghana in 2018, by Konate et al. in Mali in 2019, Berhane et al. in Ethiopia in 2021, and Das et al. in India in 2022, who found a predominance of gram-positive bacteria, most of which are part of the normal skin flora, suggesting that in these studies infections could be transmitted by parents and medical staff [6, 14, 15, 19, 21]. Since gram-negative bacteria are most often implicated in the pathogenesis of early bacterial neonatal infection, suggesting vertical transmission of the disease, the high frequency of neonates presenting for consultation within the first seventy-two hours of life in our study could explain the predominance of gram-negative bacterial strains [14, 17, 19]. Among the isolates obtained, E. coli, Burkholderia cepacia and Klebsiella pneumoniae were the most frequent. Kemeze et al. obtained E. coli and S. aureus [5]. In north-western Nigeria, E. coli was the most common, and in Morocco, E. coli, Klebsiella pneumoniae and Enterobacter cloacae were the predominant bacterial pathogens [20, 22]. In China, Gao et al. found group B Streptococcus and E. coli, but CONS and K. pneumoniae were the bacterial pathogens found by Berhane et al. in Ethiopia and Das et al. in India [6, 19, 23]. This suggests a variation in bacterial ecology depending on the year and the geographical environment.

One of the observations made in our study was the isolation of Burkholderia cepacia, the second most frequent bacterium, found mainly in patients admitted to the in-patient neonatology department, in preterm babies, and those born by caesarean section, with a statistically significant difference in proportion. Early neonatal infection with Burkholderia cepacia was only considered if clinical criteria were present, the germ was isolated in a sterile site (CSF, blood), and there was no obvious alternative diagnosis. It was also found frequently in studies conducted by Mali et al. in India and by Salah et al. in Yemen, where it was identified as a source of healthcare-associated infections [24, 25]. B. cepacia, a non-fermentative, aerobic, gram-negative bacillus, mainly colonises hospital fluids and can contaminate hosts with weakened immune systems.

Overall, the most effective antibiotics were the fluoroquinolones: ciprofloxacin (85.4%), ofloxacin (88.7%), levofloxacin (91.1%), pefloxacin (91.3%) and the penems: meropenem (87.8%). In Yaounde, the best sensitivity were obtained with imipenem (91.7%), ofloxacin (90%), ciprofloxacin (85.3%) and ceftazidime (69.4%) in a study done by Chiabi et al [9].

Among the gram-negative bacteria isolated, Escherichia coli was 100% sensitive to piperacillin, piperacillin + tazobactam and levofloxacin. High sensitivities were also obtained with ceftazidime (92.9%), ciprofloxacin (91.7%), ofloxacin (90.5%) and meropenem (85.7%). Burkholderia cepacia was 100% sensitive to piperacillin and piperacillin + tazobactam. This bacterial pathogen was 87.5% sensitive to imipenem, 84.2% to meropenem and 75% to ceftazidime. Klebsiella pneumoniae strains were 100% sensitive to levofloxacin, ciprofloxacin, imipenem, meropenem and piperacillin. They were 80% for ofloxacin, 75% for amikacin and 66.7% for ceftazidime. In Douala, 100% of gram-negative patients were sensitive to imipenem and 60% to amikacin, according to Kemeze et al [5]. In Palestine, Ibrahim et al. found that gram-negative patients were highly sensitive to amikacin (63%), meropenem (70%) and piperacillin-tazobactam (65.6%) [26].

The main bacterial isolates, E. coli, Burkholderia cepacia and Klebsiella species, showed high levels of resistance to β-lactam antibiotics, ranging from 66.7% to cefotaxime to 100% to ampicillin. These data are similar to those found by Jatsho et al. in which Klebsiella species were 100% resistant to ampicillin, 92.3% resistant to ceftriaxone and E. coli was resistant to 83.3% ampicillin. Pokhrel et al. found 66.7% to ampicillin/sulbactam and 90.5% to cefotaxime. Aku et al. found 100% resistance to ampicillin [8, 14, 17].

Out of the 761 patients admitted for suspected early neonatal infection in our study, 42 neonates having a positive culture neonates died, giving a mortality rate of 5.5% linked to confirmed neonatal bacterial infection. This result is similar to those obtained by Berhane et al (12.2%), Folquet et al (11.6%) and Gao et al (9.0%) [6, 21, 23]. Prematurity, low birth weight, early acquisition of infection and the level of resistance to antibiotic therapy could explain the high mortality rate in our study [15].

Conclusion

This study highlights the burden of early-onset neonatal infections, with a confirmed incidence of one in seven cases, particularly among female, premature, and low-birth-weight infants. The high prevalence of bacterial ecology, and resistance to common antibiotics complicate timely and effective management. Escherichia coli emerged as the most prevalent gram-negative pathogen, demonstrating sensitivity to fluoroquinolones and penems. The 5.5% mortality rate emphasises on the importance of early detection and targeted antibiotherapy to reduce neonatal morbidity and mortality.

Abbreviations

DGOPH

Douala Gynaecological Obstetric and Paediatric Hospital

EONI

Early onset neonatal bacterial Infection

CSF

Cerebrospinal fluid

CoNS

Gram-positive staphylococci

Author contributions

Authors’ contributionConception and design: DNN, SG; Data collection: DNN, AREE; Data analysis and interpretation: DNN, AREE, ED; Write-up and revision of the manuscript: DNN, DE, AREE, DFAKM, YDPF, AAST, MCA, CE, SG; Supervision: DNN, SG; Approval of final version of the manuscript: All authors.

Funding

This research did not receive a specific grant.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on request.

Declarations

Ethics approval and consent to participate

This research adhered to the Declaration of Helsinki and received approval from the Ethics Committee of Douala Gynaecological-Obstetric and Paediatric Hospital. Informed consent to participate was obtained from the parents of any participant.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.