SUDANESE JOURNAL OF PAEDIATRICS
2021; Vol 21, Issue No. 1
Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan
Osama Ahmed Braima (1), Manal Abdalaziz Ali (2), Enas Mohammed Abdulla (3)
(1) Department of Paediatrics, University Hospitals of Morecambe, NHS Foundation Trust, United Kingdom
(2) Microbiology Department, Omdurman Maternity Hospital, Sudan
(3) Omdurman Paediatrics Emergency Hospital, Omdurman, Sudan
Osama Ahmed Braima
Consultant Paediatrician & Neonatologist, Department of Paediatrics, University Hospitals of Morecambe-NHS Foundation Trust, United Kingdom
Email: osamabraima [at] gmail.com
Received: 14 October 2020 | Accepted: 1 January 2021
How to cite this article:
Braima OA, Ali MA, Abdulla EM. Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan. Sudan J Paediatr. 2021;21(1):13–22.
© 2021 SUDANESE JOURNAL OF PAEDIATRICS
Neonatal sepsis is a leading cause of morbidity and mortality, and knowledge of bacterial patterns and susceptibility to antibiotics is essential to design therapeutic guidelines. To determine the bacterial aetiology and antibiotic resistance patterns in neonatal sepsis, a prospective, cross-sectional, hospital-based study was conducted in a large paediatric emergency hospital in Khartoum State, Sudan, over a 6-month period. All newborn infants with a clinical diagnosis of sepsis were included in the study and had a blood sample collected for culture and sensitivity. The World Health Organization case definition of neonatal sepsis in conjunction with the Integrated Management of Childhood Illnesses criteria was used to select patients. A total of 170 newborn infants fulfilled the study inclusion criteria. The median age at presentation was 4-7 days and blood culture was positive in 31%. Early-onset neonatal sepsis was present in 30% of cases, while 70% were late-onset. Gram-positive organisms accounted for 58% of the total isolates, and 38% were Gram-negative organisms. Staphylococcus aureus was the most prevalent organism (55% of all isolates) and 72% of these were methicillin-resistant which showed 100% sensitivity to vancomycin and 90% to gentamycin. Pseudomonas aeruginosa was the commonest Gram-negative organism in both early and late-onset sepsis and the second commonest isolated organism, accounting for 19% of cases. All Gram-negative organisms were 100% sensitive to imipenem, meropenem and ciprofloxacin. Both Gram-positive and Gram-negative organisms were highly resistant to benzylpenicillin and cefotaxime, the commonly used empiric antibiotics in neonatal sepsis.
Neonatal sepsis; Community-acquired; Bacteria; Susceptibility to antibiotics; Sudan.
Despite the recent advances in neonatal care, neonatal sepsis remains a leading cause of morbidity and mortality across the world. Globally, bacterial infections are responsible for approximately one-third of neonatal deaths . Newborn infants are more prone to infections due to impaired innate immunity which is characterised by low response chemotactic factors, decrease in the expression of adhesion molecules in neutrophils, low cytokine productions  and impaired T-cell function .A case definition for neonatal sepsis is lacking and neonatal sepsis is variably defined [4,5]. However, specific criteria that classify sepsis into culture-proven, possible and probable sepsis have been developed using clinical and laboratory findings . Moreover, the term ‘clinical sepsis’ is commonly used by neonatologists and clinicians in the presence of negative blood culture, and clinical symptoms and signs of sepsis. However, this term also lacks consensus and no clear criteria exist for its use in research and clinical practice.Neonatal sepsis is classified by the time and onset of the disease into early and late-onset . Early-onset neonatal sepsis (EONS) is the one that occurs in the first 3 days of an infant’s life and is often due to vertical transmission of organisms from the maternal genitourinary tract, while late-onset neonatal sepsis (LONS) is defined as sepsis occurring in infants older than 72 hours and is commonly due to horizontal transmission . However, the time limits for defining early and late-onset sepsis lack consensus too and many authors extend the duration of EONS for up to 7 days following birth . Moreover, in many studies from developing countries, sepsis is classified as hospital-acquired or community-acquired, possibly due to the significant number of infants that are stillborn at home. Unfortunately, despite the extensive scientific research and improvement in diagnostic technology over the last few decades, a cheap, accurate and reliable single marker for the diagnosis of sepsis is lacking. Historically, blood culture is considered the “gold standard” test for the diagnosis of sepsis. However, the low yield of blood culture often leads to more reliance on clinical judgment in treating infants with suspected sepsis .The increasing antibiotic resistance in children and neonates in developing countries is alarming . A recent meta-analysis of a systematic review showed that approximately 40% of organisms causing community-acquired neonatal sepsis in lower and middle-income countries (LMIC) are resistant or have reduced susceptibility to the current World Health Organization (WHO) recommended antibiotics . Moreover, available data from hospital-based studies in LMICs suggest that Gram-negative organisms cause most of the neonatal infections in the first 7 days of life . Due to these challenges, robust bacterial and antibiotic surveillance is essential to provide the necessary data to inform local therapeutic and management guidelines.
Data on bacteriological profile and antibiotic sensitivity in neonatal sepsis is very limited in Sudan, especially for community-acquired neonatal sepsis. Due to limited diagnostic resources, a significant number of newborn infants receive inappropriate and lengthy courses of empiric antibiotics and that may be associated with adverse short and long-term outcomes and an increase in antimicrobial resistance rates. Therefore, the main aim of this study was to determine the microbiological patterns of possible community-acquired neonatal sepsis and the antibiotic susceptibility of the isolates to the commonly used empiric antibiotics in our healthcare institution. These data will help in informing our local therapeutic guidelines and choice of empiric antibiotics, especially when access to reliable microbiology services is very limited.
MATERIALS AND METHODS
Study design, setting and participants
This was a prospective cross-sectional, hospital-based study conducted at Omdurman Paediatrics Emergency Hospital in Khartoum State, Sudan, between February and July 2019. The hospital is a large teaching and training centre that provides secondary care services for a large number of children in the capital and accommodates referrals from different parts of the country. The hospital has 320 beds including the short stay and the latest statistics showed an annual admission of almost thirty-one thousand infants and children. All newborn infants admitted to this hospital are out born since the hospital is not affiliated to a maternity unit but the closet maternity hospital has an annual total birth of over 25,000.
All newborn infants (age ≤ 28 days) with a clinical diagnosis of sepsis were included in this study. The WHO case definition for neonatal sepsis in conjunction with the Integrated Management of Childhood Illnesses criteria  was used to select patients. The clinical diagnosis of neonatal sepsis was made in the presence of at least two of the symptoms or signs in the tool to minimise the chance of negative blood culture.
A standard pro forma was specifically constructed to collect patients’ demographic data and essential information such as delivery history, risk factors for infection, symptoms, laboratory results and clinical outcome. The attending clinician collected demographic data on admission, and the data collector completed any missing data the following day. Data regarding the results of investigations were collected from the laboratory records using a standard form and subsequently were entered into an excel spreadsheet.
Procedure and specimen collection
All infants who fulfilled the study inclusion criteria had a blood sample collected for culture and sensitivity. The blood samples were collected by experienced doctors following training on the appropriate method of obtaining a blood culture specimen. An amount of 1-2 ml of the blood sample was collected from a suitable peripheral vein following strict aseptic precautions using 10% iodine followed by 70% ethanol solution. The collected blood sample was directly inoculated aseptically into a 20 ml paediatric brain-heart infusion broth (HiMedia, India) to make a 1:10 dilution. The specimens were then immediately transported to the microbiology laboratory at the Maternity Hospital, Omdurman.
The blood sample specimens were incubated aerobically at 37°C for up to 7 days and observed daily for evidence of microbial growth by one of the following: air bubbles (gas production), turbidity, haemolysis and coagulation of broth. Regular subcultures were done after one day of incubation, at 48 hours, and then at 96 hours. At the same time, direct gram stain from the bottle and subculture was done for all specimens on enriched and selective media including blood agar plate (Oxoid LTD), MacConkey agar (Fulka medica), and chocolate agar plate and incubated aerobically at 37°C for 24 hours. The same protocol was repeated until day seven before the specimens were considered negative for growth.Bacterial identification was made based on colonial morphology, Gram stain and standard biochemical reactions. Oxidase test, Kligler Iron Agar test, citrate utilisation test, urease test, motility test and indole test were used for the identification of Gram-negative bacilli. Catalase test, coagulase test, growth on mannitol salt agar and haemolytic activity on blood agar plate were used for identification of Gram-positive organisms .
Antibiotic susceptibility testing
Antibiotic susceptibility testing was performed for all blood culture isolates by Kirby-Bauer disc diffusion susceptibility method on Muller Hinton agar according to the Clinical Laboratory Standards Institutes (CLSI) guidelines . Three to five pure colonies of isolates were emulsified in distilled water, the turbidity of the bacterial suspension was matched and checked with 0.5 McFarland standards and inoculated on Muller Hinton agar. Antimicrobial discs were placed on the inoculated agar and incubated for 24 hours at 37°C; then they were observed for zones of inhibition and determined as susceptible, intermediate, or resistant according to the CLSI guidelines . The applied antibiotic for Gram-positive cocci Staphylococcus aureus ATCC25923 and included oxacillin to check methicillin-resistant S. aureus MRSA and for Gram-negative rods, Escherichia coli ATTCC 25922 and Pseudomonas aeruginosa ATCC 27853 was used as a control strain.
Third-generation cephalosporin such as ceftazidime (30 μg), cefotaxime (30 μg) and ceftriaxone (30 μg) were used to detect extended-spectrum B-lactamase (ESBL). If the organism was sensitive, it was considered as β-lactamase sensitive but not an ESBL producer. On the other hand, if the organism was resistant, it was considered an ESBL producer, and all ESBL isolates were confirmed by the double-disk synergy test . The following antimicrobial disks were used: penicillin (10 μg), amoxicillin-clavulanate (20/10 μg), vancomycin (30 μg), gentamicin (CN10:10 μg), amikacin (30 μg), ciprofloxacin (10 μg), cefuroxime (30 μg), ceftazidime (30 μg), ceftriaxone (30 μg), meropenem (10 μg).
Ethical considerations and data analysis
An ethical clearance certificate was obtained from the hospital ethics and research committee. The blood sample taken was considered part of standard care, and as such, it did not require explicit consent. However, informed consent to participate in the study was obtained from at least one of the infant’s parents/guardians. The data were collected and entered into Microsoft Excel and exported into a statistical package for social sciences (SPSS software, Version 24) for editing and statistical analysis and simple descriptive statistics and percentages were used.
Characteristics of all study participants
During the study period, a total of 170 newborn infants fulfilled the study inclusion criteria and were screened for sepsis (Table 1). They were 104 (61%) male and 66 (39%) female infants, two-thirds of them had a birth weight of ≥2.5 kg and 84% were full term. At least one risk factor for infection was detected in 32% of the study participants and approximately 60% of babies had at least one dose of intravenous antibiotic prior to the collection of blood culture specimens. In the present study, early-onset sepsis was defined as that occurring from 0-72 hours of life while late-onset sepsis that occurred from 72 hours to 28 days. Using this definition, early-onset sepsis was detected in 51 (30%) cases and late-onset sepsis in 119 (70%) cases.
Characteristics of infants with culture-proven sepsis
Among the 53 infants with positive blood cultures, 38 (72%) were males while 15 (28%) were females, with a male: female ratio of 2.5:1. The majority (45, 84%) of infants were born at term while 16% were late preterm, with only 2% of babies below 30 weeks of gestation. Early-onset sepsis was present in 20 (38%) cases while late-onset was detected in 33 (62 %) cases. In this category, approximately 52% of infants were born at a healthcare facility, while 48% were born at home. Regarding risk factors for infections, there were 9 infants (17%) with at least one risk factor for infection while almost 77% of infants with positive blood culture received at least one dose of intravenous antibiotic prior to the collection of blood specimens.
Of the 170 collected blood samples, blood culture was positive in 53 (31%) cases. Gram-positive organisms were more prevalent and accounted for 58% of the total isolates, while Gram-negative organisms were 42%. Staphylococcus aureus was the commonest isolated organism in both early and LONS and accounted for 55% of all isolates (Figure 1, Table 2). This organism was responsible for 55% of cases of EONS and also 55% of cases in LONS. More than two-thirds of the isolated cases of S. aureus were MRSA. On the other hand, Pseudomonas aeruginosa was the second commonest isolated organism and the commonest isolated Gram-negative organism in both early and LONS. Combined with S. aureus, the two organisms were responsible for almost 74% of cases of culture-proven sepsis in this study. However, Klebsiella pneumoniae was the third commonest isolated organism and accounted for 13% of cases, followed by Escherichia coli (6%) while Group B streptococcus (GBS) was only isolated in 2% of cases.
Antibiotic susceptibility patterns of bacterial isolates
The in-vitro sensitivity patterns showed that the isolated Gram-positive organisms were highly sensitive to vancomycin and gentamycin (Table 3). MRSA showed 100% sensitivity to vancomycin and 90% to gentamycin while methicillin-susceptible S. aureus (MSSA) revealed a modest sensitivity to vancomycin and ciprofloxacin. On the other hand, all Gram-negative organisms were highly susceptible to ciprofloxacin, imipenem and meropenem while both Gram-positive and negative organisms have shown significant resistance was to benzylpenicillin and cefotaxime (Table 4).
Table 1. Demographic characteristics of the study participants.
GBS, Group B Streptococcus; MRSA, methicillin-resistant S. aureus; MSSA, methicillin-susceptible S. aureus; Staph epiderm, Staphylococcus epidermidis.
E. coli, Escherichia coli; K. pneumoniae, Klebsiella pneumoniae.
The overall mortality rate amongst study participants was 10%. However, only 50% of infants who died had culture-proven sepsis; MRSA was isolated in more than one-third of these cases.
Figure 1. The frequency of isolated organisms causing possible community-acquired neonatal sepsis. E. coli, Escherichia coli; GBS, Group B Streptococcus.
Table 2. Distribution of isolated organisms among early and late-onset neonatal sepsis.
In the present study, the prevalence of culture-proven sepsis was 31%. This percentage is comparable to some studies from other developing countries including Egypt and Nigeria [17,18]. However, a lower prevalence of only 10% and a higher prevalence of 46% have been reported in studies from Botswana and Ethiopia, respectively [19,20]. This variation may be due to the difference in study settings, methodological variations between studies, and other factors such as the efficiency of infection prevention and control policies.
The study analysed the data of 170 infants with clinical symptoms and signs of sepsis. Male infants accounted for 61% while 39% were females. Similar sex distribution was reported by Ejiofor et al. . Muenchhoff and Goulder suggested that the increase in the prevalence of sepsis in male infants may be due to their natural predisposition to infection . However, other studies reported no significant sex difference between infants with sepsis . In this study, approximately 50% of infants were born at home. This is comparable to the results of many studies from sub-Saharan Africa, where facility-based delivery is disappointingly low . Moreover, in contrast to most infants in many studies from high- and low-income countries, infants in this study were born at term and were previously healthy with no other co-morbidities.
EONS accounted for 70% of cases while the remaining percentage was late-onset. In infants with positive blood culture, LONS accounted for approximately 60% of cases. However, in a recent laboratory-based study from the largest maternity hospital in Sudan, the prevalence result was reversed; with LONS accounting for only 1/3 of the cases . Among other factors, the high prevalence of LONS in the community has been attributed to poor cord care, poor hygiene and bottle feeding in this setting .
Table 3. The average percentage of antibiotic susceptibility of isolated Gram-positive organisms.
Table 4. The average percentage of antibiotic susceptibility of isolated Gram-negative organisms.
Staphylococcus aureus accounted for most of the isolated organisms and more than two-thirds were MRSA. A recent systematic review in developing countries reported S. aureus as the commonest isolated organism responsible for causing neonatal sepsis . However, other studies from Africa reported a high prevalence of Gram-negative organisms; with Klebsiella species accounting for most of the isolates . The horizontal transmission from mothers to their new-born infants was shown to be a major source of neonatal infection with Staphylococcus aureus .
The study showed that MRSA was the commonest cause of both early and late-onset neonatal sepsis. MRSA is known to be associated with LONS and outbreaks; although it was isolated in newborns right from the time of their admission to the neonatal intensive care units (NICUs) . Factors known to increase the MRSA infection include LBW, prematurity, mechanical ventilation and umbilical catheterisation . However, infants with MRSA in the present study were mostly born at term and none has been exposed to any of the reported above risk factors. Therefore, the reasons and sources of MRSA in our study population are yet to be established. Pseudomonas aeruginosa was the second most prevalent organism causing both early and late-onset neonatal sepsis. However, Pseudomonas aeruginosa is a rare cause of neonatal sepsis in developed countries and is mostly known to cause outbreaks and colonisation in NICUs .
Klebsiella pneumoniae was the third commonest isolated organisms and accounted for 13% of cases in this study. However, a NICU-based study from Soba University Hospital in Sudan showed that Klebsiella pneumoniae was the most prevalent organism causing neonatal sepsis, accounting for more than 70% of cases , and their findings were in accordance with many hospital-based studies from developing countries where Klebsiella was a common cause of both early and late-onset neonatal sepsis . On the other hand, GBS was isolated in only one infant which is comparable to many studies in Sub-Saharan Africa where GBS is relatively isolated in small numbers, in contrast to high-income countries where GBS is a major cause of EONS. In the present study, E. coli was responsible for 5% of cases of early-onset sepsis and 8% of LONS. This result is in contrast to a study from Nigeria where E.coli was reported as the only Gram-negative organism in EONS . However, E. coli was reported as a leading cause of EONS in both developing and developed countries and it is frequently associated with severe infections .
The in vitro sensitivity patterns showed that Gram-positive organisms were highly susceptible to vancomycin and gentamycin. Similar findings were reported in a study from Egypt, with slightly lower susceptibility of organisms to gentamycin compared to the present study . On the other hand, Gram-negative organisms were highly susceptible to imipenem, meropenem and ciprofloxacin. Both Gram-positive and Gram-negative organisms were highly resistant to benzylpenicillin and cefotaxime. The high resistance shown to benzylpenicillin and cefotaxime in the treatment of neonatal sepsis is in accordance with many studies from developing countries . However, in the United Kingdom, for instance, the combination of penicillin and gentamicin remains the first-line empiric antibiotic and was reported to cover more than 95% of pathogens responsible for early-onset sepsis .
MRSA was highly sensitive to vancomycin in this study. The efficacy of vancomycin in the treatment of Gram-positive bacteria including MRSA was reported in a previous study by Chua and Howden . Moreover, the American Academy of Paediatrics (AAP) recommended vancomycin plus gentamycin and/or rifampicin for the treatment of life-threatening community-acquired (CA)-MRSA infection with unknown susceptibility .The overall mortality rate amongst infants with suspected sepsis was 10%. However, only 50% of infants who died had culture-proven sepsis and MRSA was responsible for more than one-third of those deaths. Factors that may have contributed to the high prevalence of negative blood culture among infants who died may include the excessive use of intravenous antibiotics prior to the collection of blood culture, low bacteraemia which is a recognised cause for false-negative blood culture if an insufficient blood volume is inoculated  or infants may simply have had an alternative diagnosis other than sepsis.
CONCLUSIONS AND RECOMMENDATIONS
This study provides crucial information regarding organisms causing possible community-acquired neonatal sepsis and their susceptibility to commonly used empiric antibiotics, and highlighted the indispensable contribution of a clinical microbiologist. MRSA was the most prevalent organism causing community-acquired neonatal sepsis and the organism was highly sensitive to vancomycin and gentamycin which are both nephrotoxic. On the other hand, all isolated Gram-negative organisms were highly sensitive to meropenem, imipenem and ciprofloxacin. However, significant resistance to benzylpenicillin and cefotaxime was shown by most of the isolated organisms in the study. To combat the escalating problem of antibiotic resistance, the authors suggest regular surveillance, rigorous infection control measures and development of local therapeutic guidelines. To inform the development of evidence-based therapeutics guidelines in Sudan, more studies are needed in this field.
CONFLICTS OF INTEREST
The authors declare that there is no conflict of interest regarding the publication of this article.
The study was self-funded.
|How to Cite this Article|
Braima OA, Ali MA, Abdulla EM. Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan. Sudan J Paed. 2021; 21(1): 13-22. doi:10.24911/SJP.106-1601909519
Braima OA, Ali MA, Abdulla EM. Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan. http://www.sudanjp.com/?mno=7349 [Access: May 13, 2021]. doi:10.24911/SJP.106-1601909519
AMA (American Medical Association) Style
Braima OA, Ali MA, Abdulla EM. Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan. Sudan J Paed. 2021; 21(1): 13-22. doi:10.24911/SJP.106-1601909519
Braima OA, Ali MA, Abdulla EM. Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan. Sudan J Paed. (2021), [cited May 13, 2021]; 21(1): 13-22. doi:10.24911/SJP.106-1601909519
Braima, O. A., Ali, . M. A. & Abdulla, . E. M. (2021) Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan. Sudan J Paed, 21 (1), 13-22. doi:10.24911/SJP.106-1601909519
Braima, Osama Ahmed, Manal Abdalaziz Ali, and Enas Mohammed Abdulla. 2021. Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan. Sudanese Journal of Paediatrics, 21 (1), 13-22. doi:10.24911/SJP.106-1601909519
Braima, Osama Ahmed, Manal Abdalaziz Ali, and Enas Mohammed Abdulla. "Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan." Sudanese Journal of Paediatrics 21 (2021), 13-22. doi:10.24911/SJP.106-1601909519
MLA (The Modern Language Association) Style
Braima, Osama Ahmed, Manal Abdalaziz Ali, and Enas Mohammed Abdulla. "Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan." Sudanese Journal of Paediatrics 21.1 (2021), 13-22. Print. doi:10.24911/SJP.106-1601909519
APA (American Psychological Association) Style
Braima, O. A., Ali, . M. A. & Abdulla, . E. M. (2021) Bacteriological profile and antibiotic resistance in newborn infants with possible community-acquired neonatal sepsis in Khartoum State, Sudan. Sudanese Journal of Paediatrics, 21 (1), 13-22. doi:10.24911/SJP.106-1601909519