Early Onset Escherichia coli Sepsis and Severe Intraventricular Hemorrhage in Extremely Premature Infants: Cases Series and Literature Review

Ipsita Goswami1*, Abhay Lodha1, Stacey Dalgleish3 and Khorshid Mohammad1 1Department of Pediatrics, Cumming School of Medicine, University of Calgary, Canada 3Alberta Health Services, Foothills Medical Centre, Canada Submission: February 1, 2017; Published: February 23, 2017 *Corresponding author: Dr. Ipsita Goswami, Neonatal Neurocritical Care Fellow, Department of Pediatrics, Cumming School of Medicine, University of Calgary, Foothills Medical Centre, 1403 29th Street NW Calgary, Alberta, Canada Email:

parenteral ampicillin and erythromycin for 48 hours followed by amoxicillin and erythromycin orally for 5 days [4]. Although effective in reducing overall incidence of EOS, administration of intrapartum broad-spectrum antibiotics has raised concerns about selective colonization by antibiotic resistant organisms [5]. In premature infants, prevention of E coli EOS continues to be a challenging. E coli is associated with more fulminant disease than non-E coli gram-negative organisms [6]. Early onset E coli sepsis in premature infants may result in systemic involvement as evidenced by thrombocytopenia and hypotension requiring aggressive interventions such as fluid boluses and inotropes, increasing the risk of germinal matrix hemorrhage in these infants. We report a possible association between early onset E coli sepsis in 5 extremely preterm infants and severe intraventricular hemorrhage (IVH).

Case Presentation
We report a series of 5 very low birth weight infants (VLBWI) with early onset E coli sepsis in our regional NICU between January and August 2015. In all cases there was history of pPROM more than 24 hours prior to delivery and mothers had received intrapartum antibiotics as per unit policy and complete courses of antenatal corticosteroids. All infants had positive blood cultures growing E. coli within 12-16 hours after birth. Two cases had culture proven E Coli meningitis and one had positive eye swab culture for E coli as well as positive sputum cultures for the same organism. Maternal and perinatal characteristics of the infants are enumerated in Table 1. Isolated strains in cases 1-3 were resistant to Ampicillin and Ciprofloxacin but sensitive to Gentamicin and 3 rd generation cephalosporin. Twins (Case 4 & 5) were infected with multidrug resistant extended spectrum beta lactamase (ESBL) producing E.coli sensitive only to Carbapenem. E coli strains isolated from CSF were Ampicillin resistant but sensitive to cephalosporin and Carbapenem. Of three surviving infants, two were successfully treated with monotherapy of Cefotaxime and Meropenem for 10 days respectively, one was initially treated with Meropenem and later switched to Cefotaxime for 3 weeks to treat meningitis. All the infants developed IVH in the first week of life. Case 3 & 4 had severe septic shock syndrome along with Grade III/IV IVH and cerebellar hemorrhages, thus care was withdrawn with parental consent. Case 1 had bilateral Grade II IVH complicated with post-hemorrhagic hydrocephalus. Of note, this infant had a relatively uneventful stay with invasive ventilation for 7 days, was on full enteral feeds by 10 days, neither had meningitis, thrombocytopenia nor medical management of patent ductus arteriosus (PDA). The post-hemorrhagic hydrocephalus was initially managed with a reservoir and serial CSF tapping and later required choroid plexus coagulation and 3 rd ventriculostomy followed by ventriculo-peritoneal shunt placement. Case 5 had progressive ventricular dilatation during NICU stay that stabilised at corrected term gestational age, thus did not require invasive management. Case 1&5 had white matter injury manifested as porencephalic cysts while Case 3 had no noted complications of IVH ( Figure 1).

Discussion
Our findings suggest that antibiotic resistant E coli vertically transmitted from mothers can particularly affect the immature brain and may associated with an increased risk of severe IVH. A cluster of 5 cases within a span of 6 months raises concern about changing trends in microbiological pattern of EOS and antibiotic sensitivity pattern. There is an emerging trend towards organisms resistant to empirical antibiotics used in NICU. Severe systemic infection and resulting cytokine response may lead to germinal matrix bleeds especially in the first 48hrs of life when the culture sensitivity of the infecting organism is pending. With the initiation of antibiotics specific to the particular organism, the systemic infection can be promptly treated and controlled but severe intraventricular hemorrhages that have occurred within first 48hrs may have major effect on future neurological outcome among survivors.
Overall incidence of EOS in infants <1500g is 10 per 1000 live births. E coli is the pathogen in 5 per 1000, significantly more than GBS (2 per 1000) [3]. Use of intrapartum antibiotics for prevention of GBS sepsis has led to reduction in EOS but E coli still accounts for 30% of the disease burden [3]. Broad-spectrum antibiotics given to mothers with pPROM reduces neonatal morbidities [7] but there is a potential problem of significant increase in Ampicillin resistant E coli causing more fulminant disease especially in mothers with an intrapartum fever [8]. We report a series of five infants born before 28wks to mothers who had pPROM and developed EOS with E coli resistant to the empirical antibiotics used in NICU. Certain clinical signs like intrapartum fever and pPROM > 2 days may help identify the mothers who need to be screened for E coli so that culture sensitivity is available for the suspected organism even before the infant is born.
All E coli isolates in our case series were resistant to Ampicillin and Ciprofloxacin and two strains were ESBL producing organism resistant to all antibiotics except Carbapenem. Mothers of infants did not visit known countries where there is an increased risk of ESBL producing organisms. Whether use of broad-spectrum antibiotic prophylaxis has increased the risk of resistant E coli infection is still controversial [5,6,[8][9][10]. Alarmingly, all isolates in our case series were Ampicillin resistant while the empirical antibiotic used in our NICU was Ampicillin with an aminoglycoside. Fortunately, most of the isolates were Gentamicin sensitive but due to the severity of illness clinicians opted to treat with Carbepenem in 3/5 cases. Choice of empirical antibiotics that provides adequate E coli coverage may be vital in this scenario as the culture sensitivity takes at-least 48 hours after birth, a time period very critical for the immature brain.

Academic Journal of Pediatrics & Neonatology
The majority of IVH occurs within first 72 hours of life and all infants in our series developed moderate to severe IVH. Association between E coli sepsis and IVH is unclear. Tsai et al. found no difference in the incidence of IVH in neonates with E. Coli sepsis compared to uninfected controls of same gestational age.
[6] A retrospective study identified all E coli EOS cases and found 27% had IVH while 33% of the survivors assessed at infancy had impaired motor development, abnormal cognitive outcome and seizure disorders. [11] Clinical outcome of infants with E coli EOS from previously published studies are summarized in Table 2. Among the studies that have reported the clinical characteristics of E coli sepsis, only two studies mention the incidence of IVH ranging from 15-30% and only one reports longterm outcome. Systemic inflammation either intrauterine or within the first week of life increases the likelihood of IVH in extreme premature infants [12,13]. The other organism, associated with higher risk IVH is Ureaplasma Urealyticum colonizing maternal genital tract [14]. Given the fact that E coli is a more virulent organism, we speculate that the virulence of the pathogen is an important trigger for severe inflammatory response and possibly causing severe intraventricular haemorrhage [15].
E coli EOS may lead to increased risk of IVH directly as well as secondary to associated thrombocytopenia and coagulopathy, which increases the propensity to bleed. Lower platelet count during first 3 days of life increases the risk of IVH in VLBWI with gram negative EOS [16]. The effect of neonatal sepsis on platelet count is believed to be organism specific, the platelet nadir being significantly of lower in gram-negative sepsis [15]. Four cases in our case series had significant thrombocytopenia in the first week of life requiring transfusions. Affected infants are more likely to receive volume expanders due to hypotension and have a hemodynamically significant PDA contributing to the fluctuations in cerebral blood flow. Premature infants with their limited capacity of cerebral autoregulation, further impaired by systemic illness, are unable to buffer the effect of fluctuations leading to germinal matrix hemorrhage [17].

Conclusion
We report a possible association between early onset Ampicillin resistant E coli infection and IVH in VLBWI. There is need of continued surveillance to monitor changes in sensitivity pattern of pathogens causing EOS in NICU and emergence of strains resistant to the antibiotics used for antepartum prophylaxis.
Clinicians need to be aware of the risk factors that precede E coli sepsis and its implications on VLBWI. Timely interventions like aggressive antimicrobial therapy, platelet transfusions and early PDA closure may reduce the risk of severe IVH. Further prospective studies would be needed to establish temporal relationship.

Author's Contributions
IG performed literature search and started writing the manuscript. KM conceived the idea, obtained results, manuscript revision, and pictures. AL contributed to manuscript, finalized the manuscript. SD did critical analysis of the manuscript and obtaining consent for publication. All the authors read and approved the final manuscript.

Consent
Written consent was obtained from parents in institutional forms for publication of data.