Table of Contents

HK J Paediatr (New Series)
Vol 24. No. 3, 2019

HK J Paediatr (New Series) 2019;24:127-134

Original Article

Escherichia coli Meningitis in Neonatal Intensive Care Units: A Five Years Study

KK Chik, WK To, K Luk, B Lam, CC Shek, C Tse, D Ng


Introduction: Neonatal meningitis causes severe morbidity. The incidence of ampicillin-resistant Escherichia coli (E. coli) meningitis was reported to be increasing in overseas neonatal intensive care units (NICUs) recently. This is the first local study to investigate E. coli neonatal meningitis. Choice of empirical antibiotics for intrapartum period and neonates will be discussed. Methods: Retrospective study was performed in 2 NICUs between 1st January 2011 and 31st May 2016. Neonates with E. coli meningitis were included. Demographic data, laboratory results, antibiotics use and antimicrobial susceptibilities were reviewed. Results: Eleven neonates were identified and 55% had early-onset (EO) meningitis. Ninety-one percent of E. coli isolates were ampicillin-resistant. Twenty-seven percent were extended-spectrum-beta-lactamase (ESBL) producing and all were isolated in EO meningitis group. Use of intrapartum ampicillin and ESBL-E. coli infection in neonates showed no statistical significance. Conclusions: The risk of ampicillin-resistant E. coli meningitis in neonates continues despite of using intrapartum antibiotics in high risk group. Choice of intrapartum antibiotic and empirical treatment of neonatal sepsis has an urgent need to be reviewed.


Keyword : Antibiotic resistance; Escherichia coli; Neonatal meningitis



Neonatal meningitis is a serious infection causing high mortality and morbidity including neurological disability. The major pathogens causing neonatal meningitis are Streptococcus agalactiae (GBS) and Escherichia coli (E. coli). The United States started the national screening programme for GBS in pregnancy in 1996, since then the incidence of early onset group B streptococcus disease decreased from 1.7 per 1000 live births to 0.6 per 1000.1,2 However, the incidence of neonatal E. coli sepsis and meningitis showed an increasing trend.3,4 There were also case reports of ampicillin-resistant E. coli neonatal infection after implementing the universal intrapartum antibiotic prophylaxis to maternal Group B Streptococcus carriers.3,5 Not just ampicillin-resistance, infection caused by extended spectrum b-lactamases (ESBL) producers are also spreading rapidly. ESBL enzyme results in resistance to many commonly used beta-lactam antibiotics, including ampicillin, first, second and third generation cephalosporin. This may affect the choice of empirical antibiotic in neonatal sepsis. Hong Kong started the antenatal screening and treatment programme for maternal GBS in 2012. There was no local study on neonatal E. coli meningitis all along. To address this issue, we carried out the first local retrospective study to describe the demographic data of E. coli neonatal meningitis and the antibiotic susceptibility pattern of the isolates between 2011 and 2016. We investigate and analyse the association between demographic data, laboratory findings, use of intrapartum ampicillin and drug resistance E. coli neonatal infection, also review the choice of empirical intrapartum antibiotic and empirical neonatal antibiotics.

Figure 1 Cerebrospinal fluid results among neonates with E. coli meningitis. Normal CSF protein: <1 g/L in preterm and <1 g/L in term infants; Normal CSF glucose concentration: </=1.1 mmol/L in preterm and </= 2.5 mmol/L in term infants.


Study Design

A retrospective study performed in two Hong Kong neonatal intensive care units (NICU). The demographic data were reviewed including the gestational age, days of symptoms onset, gender, birth weight, maternal and neonates' laboratory results, intrapartum antibiotic and empirical antibiotic use in neonates. Their neurological outcomes were also reviewed. In addition, the antimicrobial susceptibilities of the isolates were analysed. Data were collected between 1st January 2011 and 31st May 2016 via Clinical Data Analysis and Reporting System of Hong Kong Hospital Authority.


Neonates </=28 days of age and microbiological confirmed E. coli meningitis were included. The diagnosis of meningitis was based on positive culture of E. coli from cerebrospinal fluid (CSF) and/or from blood culture associated with pleocytosis (>/=10 cells/mm3) in CSF. The included patients were categorised into early onset neonatal meningitis and late onset neonatal meningitis according to their gestational age and the time of symptoms onset. Early onset neonatal meningitis was defined as onset of bacterial meningitis at ≤72 hours in preterm infants and ≤ 7 days in term infants. Late onset meningitis was defined as the onset occurred after 72 hours of life in preterm babies and >7 days up to 90 days in term infants.

Statistical Analysis

Excel spreadsheets were used for tabulation of data. Statistical analysis was performed using the Statistical Package for the Social Sciences (Windows version 17; SPSS Inc, Chicago [IL], US) Fisher's exact test and Mann-Whitney U test were used for categorical variables and continuous variables respectively to analyse the association between study factors and ESBL E. coli infection among neonates. P values of <0.05 were considered statistically significant.

Ethics Review

This proposal was reviewed and approved by the Research Ethics Committee of the Kowloon West Cluster (Kowloon West; REC [KWC REC No.: 107-07]).


Demographics and Maternal Peripartum Risk Factors for Neonatal Sepsis

Between 2011 and 2016, the total number of live born in Hong Kong public hospitals was around 227, 300 and 58, 113 babies were born in the two study hospitals. It accounted for 25% of all living newborns in all Hong Kong public hospitals. With eleven cases of neonatal E. coli meningitis were identified in this study, the prevalence rate of E. coli meningitis infection in neonates of the study hospitals was estimated to be 0.02% (11/58113). This was based on the assumption that all cases delivered in these 2 hospitals would be admitted to the same hospital's NICU while babies with E. coli meningitis delivered in other hospitals were not admitted to these 2 NICUs. Demographic data were shown in Table 1a, 1b and Table 2. Six of them were male (male to female ratio 1.2:1). Mean gestational age was 34 weeks (range 29-40 weeks). Fifty-five percent were preterm babies. The mean age of onset was 10.7 days (range 0-28 days). Six had early onset infection and 5 had late onset. Mean birth weight was 2.31 kg (range 1.36-3.61 kg). Among those with early onset disease, 83% had maternal risk factors for neonatal sepsis, including prematurity, premature rupture of membranes, maternal intrapartum fever and acute chorioamnionitis confirmed by placental histology. Although intrapartum intravenous ampicillin was given as per protocol, their babies still had early onset meningitis with ampicillin-resistant E. coli.

Table 1a Demographic data of neonates with early onset and late onset sepsis
  Total cases
Early onset
Late onset
Mean age of symptoms onset in days (range) Day 10.7
Day 3.17
( 0-6)
Day 19.4
( 11-28)
 M:F 1.2:1 2:1 1:1
Mean birth weight
2.31 kg
2.29 kg
2.86 kg
Mean gestation
34 weeks
32.8 weeks
37.2 week
Full term 5 1 4
Preterm 6 5 1

Table 1b Demographic data of preterm and term neonates
  Preterm Full term
Early onset Late onset Early onset Late onset
Number of cases 5 1 1 4
Mean gestation 32 30 37 39
(range) (29-36)     (38-40)
M:F 3:2 1:0 1:0 1:1
Mean birth weight
1.49 3.61 3.2
Mean age of symptoms onset in days
Day 3.2
Day 11 Day 3 Day 24

Clinical Presentation and Laboratory Results

Eight neonates presented with fever, 2 had recurrent apnoea and 1 had malaise and feeding intolerance. Majority did not have congenital abnormalities except one had Hirschsprung's disease presented with abdominal distension and intestinal obstruction. For the laboratory results, all had elevated C-reactive protein level (range 48.3-173 mg/L). For the cerebrospinal fluid (CSF) study, 72.7% had polymorph predominant pleocytosis, CSF glucose level was ranged from 0.6 to 6.3 mmol/L and 54.5% had low CSF glucose level which was defined as </=1.1 mmol in preterm and </=2.5 mmol/L in term infant.6-8 CSF protein level was ranged from 1.02 to 6.47 g/L and seven of them had abnormal result which was defined as >1.50 g/L in preterm and >1 g/L in term infants.6-8 (Table 2) Among those six with abnormal CSF glucose level, they also had elevated CSF protein (Figure 1). Five neonates had Gram-negative bacilli detected on the Gram stain of the CSF and all of them had low CSF glucose level and elevated CSF protein level. Among these five babies, 80% had positive result on E. coli K1 latex agglutination test. None had positive agglutination result while the CSF Gram stain was negative. Three babies had E. coli isolated in the urine specimens with same susceptibility pattern as the CSF culture isolate. Ninety-one percent had coexisting E. coli bacteraemia.

Antibiotic Susceptibility

Antibiotic susceptibility patterns of the isolates were determined according to the Clinical & Laboratory Standards Institute guideline. Ninety-one percent of the isolates were resistant to ampicillin. One isolate (9.1%) was resistant to gentamicin. Three isolates (27.3%) were ESBL producing and resistant to third generation cephalosporins. Otherwise, all tested isolates were susceptible to amoxicillin-clavulanate; meropenem, and amikacin (Table3).

Table 3 Antibiotic susceptibility
  Ampicillin Amoxcillin-
Clavulanic acid
Cefotaxime Meropenem Gentamicin Amikacin ESBL+
E. coli (susceptible / total isolate) 1/11 11/11 8/11 11/11 10/11 11/11 3/11
EO (susceptible / total isolate) 0/6 6/6 3/6 6/6 6/6 6/6 3/6
LO (susceptible / total isolate) 1/5 5/5 5/5 5/5 4/5 5/5 0/5
ESBL: Extended spectrum beta lactamase; EO: Early onset meningitis; LO: Late onset meningitis

ESBL E. coli Isolate and Intrapartum Antibiotic Prophylaxis (IAP)

The demographic data, laboratory results and antibiotics use among neonates with and without ESBL - E. coli infection were analysed. It was identified that all cases infected with neonatal ESBL E. coli were presented with early onset meningitis and were given intrapartum ampicillin. There was no ESBL E. coli infection among the group without intrapartum ampicillin given. After analysis with Fisher's exact test for categorical variables and Mann-Whitney U test for continuous variables, there was no statistical significance identified between intrapartum ampicillin use and neonatal ESBL E. coli infection (p value: 0.061). There was also no significance association between ESBL E. coli infection and other demographic data and laboratory results (Table 4).

Table 4 Statistical analysis of demographic data, laboratory results, intrapartum antibiotic (IAP) use, neurological outcome in associate with neonatal ESBL E. coli infection
  ESBL negative
ESBL positive
P value*
Male 4 2 1.00
Ampicillin use (IAP) 2 3 0.061
Maternal HVS/blood grew ESBL-E. coli 0 0 ND
Placental swab grew ESBL-E. coli 0 2 ND
Acute chorioamnionitis 2 2 0.491
Gestation (week) 36±3.8 32.7±6.4 0.376
Birth weight (kg) 2.7±0 .8 2±1.4 0.497
Symptoms onset
(days of life)
13.4±9.7 3.0±3.0 0.133
Maximum CRP (mg/L) 94.1±53.9 83.3±29.7 1.00
Delay in neurological development 1 2 0.183
*Fisher exact test for categorical variables, Mann-Witney U test for continuous variables

Neurological Outcome

Three babies had developmental delay, one of them had hydrocephalus and seven had normal development on follow-up. One baby defaulted follow-up (Table 2). There was no statistical association between neonatal ESBL E. coli infection and delay in neurological development (Table 4).


Hong Kong does not have a registry on the incidence of neonatal E. coli infection. To our knowledge; this is the first local study of neonatal E. coli meningitis in NICUs. The study hospitals accounted for more than 25% of total live births in Hong Kong public hospitals; therefore the result is representative to Hong Kong situation. The estimated prevalence of neonatal E. coli meningitis was 0.02%, with the significant mortality and morbidity, it is a problem worthy of further investigations. There is no doubt that use of IAP, i.e. ampicillin in pregnant women with GBS colonisation can reduce early-onset neonatal GBS infection. The reported odd ratio was 0.17.1,2 However, some overseas studies reported that the incidence of neonatal E. coli sepsis and meningitis had increased after this practice, together with the emergence of ampicillin-resistant E. coli infection in neonates.3-5 A Taiwanese study also described an increase in the percentage of early onset neonatal E. coli sepsis from 40.9% to 70% since the implementation of the guideline on prevention of perinatal group B streptococcal disease by Center for Disease Control and Prevention.4 Hong Kong has started the GBS screening in 2012 but we lack of data on this.

Concerning the microbiological aspect and the pathogenesis, E. coli is a normal flora of the human gastrointestinal tract. It has diverse and complex antigenic structures. Some strains carry a specific virulence factor K1 capsular antigen, a polysialic acid, which impairs host's opsonophagocytic killing. The amount and persistence of K1 antigen in the CSF is closely linked to neonatal meningitis and the disease severity. Infants infected with K1 antigenic strains have increased morbidity and mortality compared to those infected with other strains. Unfortunately, up to 50% of E. coli colonised in maternal vagina can express K1 antigen which poses risk for vertical transmission of this virulence strain.9 In the past, researchers suggested the E. coli strains with K1 capsular antigen, i.e. virulent strain, were able to cause invasive infection but they were more susceptible to antimicrobial drugs, as they could not possess resistant genes. While the more antimicrobials resistant organisms, the more difficult for them to reconcile bacterial virulence and thus could only cause noninvasive infection.9 Later on, this postulation was being challenged by the case reports of ESBL E. coli neonatal meningitis.10 Our findings also opposed the postulation, as among four E. coli strain with positive K1 antigen detected in CSF, one of them was ESBL producing.

The global burden of neonatal ESBL producing Enterobacteriaceae infection is high. An overseas study reported 17% of organisms causing neonatal sepsis were ESBL producing Enterobacteriaceae among all culture positive cases.11 While our study showed 27.3% of E. coli meningitis were ESBL producing. This was much higher than the overseas study which probably due to small sample size and different in the denominator. The known risk factors of neonatal ESBL producing pathogens infection were prematurity, and extremely low birth weight.4 There is a controversy about the association between the use of intrapartum ampicillin and neonatal ESBL- Gram negative organism infection. Cordero et al12 reported there was no increase in neonatal sepsis caused by drug-resistant Gram-negative pathogen after national maternal GBS screening and prophylaxis, while other study suggested IAP might play a role in neonatal ESBL Gram negative organisms' infection.5 Though our study did not show a statistical significant relationship between IAP and neonatal ESBL E. coli meningitis, further study and monitoring deemed necessary on this issue. Meanwhile, conclusion is difficult to draw.

Besides the above identifiable risk factors for neonatal infection by ESBL producing organism, the global increase in the prevalence of ESBL producers infection among general population also played a part. In United States, 36% of all community-acquired E. coli infections were ESBL producing,13 while our local data in 2015 was 26%.14 The high rate of ESBL E. coli infection in our community has been postulated to the overuse of antibiotics in livestock farming. One study identified 53.6% of the local live pigs had faecal carriage of ESBL-producing E. coli.15 This emergence of ESBL-producing organisms in the community poses a major challenge to the choice of empirical antibiotic in newborns. According to the international guideline, intrapartum intravenous benzylpenicillin to mother or combination of ampicillin and gentamicin to the neonates should be used to prevent early-onset neonatal infection in those mothers had risk factors of neonatal sepsis. If there is microbiological evidence of neonatal Gram-negative bacterial sepsis or meningitis; the guideline suggests benzylpenicillin/ampicillin, gentamicin and cefotaxime until further results are available. This is a widely adapted regime with good coverage for GBS and E. coli, and even L. monocytogenes infection.16-18 However, this empirical antibiotic regime cannot treat meningitis caused by ESBL producing Enterobacteriaceae as it is intrinsically resistant to third generation cephalosporin and benzylpenicillin/ ampicillin. Even adding gentamicin is inadequate as E. coli can also be resistant to gentamicin as shown in one of our patient. Furthermore aminoglycoside (gentamicin and amikacin) is not a good agent to penetrate the blood brain barrier. Nevertheless, empirical use of broad-spectrum antibiotics such as carbapenem is not a way of solving this problem. Broad-spectrum antibiotic would alter the neonatal intestinal flora colonisation patterns and thus increase the risk of necrotising enterocolitis and candidaemia. The dilemma thus comes up. However, at the time of writing, there are no well-powered randomised trials on empirical antibiotic therapy for neonatal sepsis. Further studies are deemed necessary to determine the most appropriate choice of empirical intrapartum antibiotic and empirical antibiotic for neonatal sepsis in the era of increasing drug resistance organisms especially when Gram negative organism meningitis is suspected.

The limitation of our study is the small sample size despite of taking five-year data in 2 large neonatal intensive care units. We suggest continuous monitoring of the incidence of Gram negative bacterial infection in neonates and a cross centre study on the issue of intrapartum antibiotic use and empirical antibiotic use in neonatal sepsis is urgently needed.


The risk of E. coli sepsis among neonates continues despite of improvement in perinatal care and empirical use of intrapartum antibiotic. Though this study did not identify any statistical significance between intrapartum ampicillin use and ESBL infection, with the widespread drug resistance, choice of intrapartum antibiotic and empirical treatment of neonatal sepsis has an urgent need to be reviewed.

Conflict of Interest

The authors have no conflicts of interest to disclose.


1. Center for Disease Control and Prevention(CDC). Prevention of perinatal group B streptococcal disease: a public health perspective. MMWR 1996;45(RR-7):1-24.

2. Ohlsson A, Shah VS. Intrapartum antibiotics for known maternal group B streptococcal colonization. Cochrane Database Syst Rev 2014;6:CD007467.

3. Stoll BJ, Hansen NI, Sanchez PJ, et al. Early onset neonatal sepsis: the burden of group B streptococcal and E. coli disease continues. Pediatrics 127:817-26.

4. Lin CY, Hsu CH, Huang FY, et al. The changing face of early-onset neonatal sepsis after the implementation of a maternal group B Streptococcus screening and intrapartum prophylaxis policy-a study in one medical center. Pediatr Neonatol 2011;52:78-84.

5. Bizzarro MJ, Dembry LM, Baltimore RS, Gallagher PG. Changing patterns in neonatal Escherichia coli sepsis and ampicillin resistance in the era of intrapartum antibiotic prophylaxis. Pediatrics 2008;121:689-96.

6. Byington CL, Kendrick J, Sheng X. Normative cerebrospinal fluid profiles in febrile infants. J Pediatr 2011;158:130-4.

7. Chadwick SL, Wilson JW, Levin JE, Martin JM. Cerebrospinal fluid characteristics of infants who present to the emergency department with fever: establishing normal values by week of age. Pediatr Infect Dis J 2011;30:e63-7.

8. Rodriguez AF, Kaplan SL, Mason EO Jr. Cerebrospinal fluid values in the very low birth weight infant. J Pediatr 1990;116:971-4.

9. Lavigne JP, Blanc-Potard AB, Bourg G et al Virulence genotype and nematode-killing properties of extra-intestinal Escherichia coli producing CTX-M Beta-lactamases. Clin Microbiol Infect 2006;12:1199-206.

10. Dubois D, Prasadarao NV, Mittal R, Bret L, Roujou-Gris M, Bonnet R. CTX-M beta-lactamase production and virulence of Escherichia coli K1. Emerg Infect Dis 2009;15:1988-90.

11. Vijayakanthi N, Bahl D, Kaur N, Maria A, Dubey NK. Frequency and characteristics of infections caused by extended-spectrum betalactamase-producing organisms in neonates: a prospective cohort study. Biomed Res Int 2013;2013:756209.

12. Cordero L, Sananes M, Ayers LW. Bloodstream infections in a neonatal intensive-care unit: 12 years' experience with an antibiotic control program. Infect Control Hosp Epidemiol 1999;20:242-6.

13. Doi Y, Park YS, Rivera JI, et al. Community-associated extended-spectrum beta-lactamase-producing Escherichia coli infection in the United States. Clin Infect Dis 56:641-8.

14. Hong Kong Princess Margaret Hospital Antibiotic susceptibility profile 2015. http://pmh.home (accessible via Hong Kong Hospital Authority intranet).

15. Ho PL, Chow KH, Lai EL, et al. Extensive dissemination of CTX-M-producing Escherichia coli with multidrug resistance to 'critically important' antibiotics among food animals in Hong Kong, 2008-10. J Antimicrob Chemother 2011:66:765-8.

16. National Institute for Health and Care Excellence guideline Neonatal infection (early onset): antibiotics for prevention and treatment. Clinical guideline 22 August 2012. Available at:

17. Simonsen KA, Anderson-Berry AL, Delair SF, Davies HD. Early-onset neonatal sepsis. Clin Microbiol Rev 2014;27:21-47.

18. MacGowan A, Wootton M, Bowker K, Holt HA, Reeves D. Ampicillin-aminoglycoside interaction studies using Listeria monocytogenes. J Antimicrob Chemother 1998;41:417-8.


This web site is sponsored by Johnson & Johnson (HK) Ltd.
©2022 Hong Kong Journal of Paediatrics. All rights reserved. Developed and maintained by Medcom Ltd.