Early-Onset Neonatal Sepsis: Group B Streptococcal Compared to E. coli Disease

Background: Early onset sepsis (EOS) of the newborn is a severe disease and associated with high morbidity and mortality. Aim of the study was to compare perinatal, short-term outcome and laboratory data of neonates with early-onset sepsis (EOS) either due to Group B Streptococci (GBS) or Escherichia coli (E. coli) infection. Methods: Retrospective cohort analysis of all neonates with culture proven GBS and E. coli EOS born between 1993 and 2011 and hospitalized at the NICU of the Medical University of Graz, Austria. Data were analyzed regarding perinatal, laboratory and short-term outcome data. Results: During the study period 100 neonates with EOS due to GBS and 11 neonates with E. coli infection were hospitalized at our NICU. Perinatal and short-term outcome data differed between GBS and E. coli infection regarding gestational age (median 38 vs. 32 weeks, p=.005), birth weight (median 3095 vs. 1836 grams, p=.031), presence of hypothermia (0 vs. 18%, p=.009), duration of mechanical ventilation (4 vs. 8 days, p=.019), duration of therapy with supplemental oxygen (9 vs. 2 days, p=.031), length of hospitalization (15 vs. 22 days, p=.039), presence of chorioamnionitis (17 vs. 46%, p=.041) and maternal fever (2 vs. 18%, p=.049). Mortality rates did not differ significantly (6 vs. 18%, p=.180). Laboratory data regarding white blood cell count, IT-ratio, and CRP value were not different between groups within the first 72 hours of life. There was a significant decrease of GBS sepsis during the study period (p=0,014). Conclusion: Main differences between GBS and E. coli infections were due to higher rates of preterm birth in the E. coli group, clinical and laboratory characteristics only differed marginally.


Introduction
Early onset sepsis (EOS) of the newborn is a severe disease and associated with high morbidity and mortality. At present Group B Streptococci (GBS) and Escherichia coli (E. coli) are the most common pathogens in developing countries, most infants with GBS are born at term, infants with E. coli are more common born preterm [1,2]. After the implementation of preventing strategies based on administration of intrapartum antibiotic prophylaxis the incidence of GBS decreased over the last 15 years from 1.7 cases per 1,000 live births in the early 1990s to 0.34-0.37 cases per 1,000 live births in the last years [3]. In 1996 CDC recommended either a risk-based approach for intrapartum administration of antibiotics or an universal screening for vaginal or rectal GBS colonization for all pregnant woman between 35 and 37 weeks of gestation. In the revised guidelines of 2002 CDC only recommended the universal screening. In Austria there is still a riskbased approach and no universal GBS screening [4,5]. There are concerns that the widespread use of antibiotics could increase the frequencies of non-GBS or antimicrobial-resistant pathogens [6][7][8][9].
Aim of the study was to compare perinatal data, morbidities, mortality, therapies and laboratory data of neonates with EOS either due to GBS or E. coli infection.

Study design
This was a retrospective study at the NICU of the Pediatric Department of the Medical University of Graz. Data were collected between 1993 and 2011. Inclusion criteria were a culture proven sepsis caused by GBS or E. coli diagnosed within the first 72 hours of life. Exclusion criteria were missing or incomplete data, a culture-negative clinical sepsis and an unknown state of infection.

Definition of EOS
EOS was defined as culture proven sepsis (a positive blood or cerebrospinal fluid culture) within the first 72 hours of life.
In addition to a positive blood or cerebrospinal fluid culture they had to meet the following criteria: clinical signs of sepsis in ≥ 1 with either ≥ 1 maternal risk factor or ≥ 1 abnormal laboratory marker.

Statistical analysis
Statistical analyses were performed with SPSS version 20 (SPSS, Chicago, IL, USA). Descriptive statistics were obtained for all categorical variables. Statistical significance was determined for unadjusted comparisons by Mann-Whitney-U-test for continuous variables and by Fisher´s exact test for categorical variables. A univariate analysis was performed and the significance level was set at p<.05. The Pearson correlation coefficient was used to analyze the association between the frequency of occurrence of pathogens and the time.

Results
During the study period 100 neonates with EOS due to GBS and 11 neonates with E. coli infection were hospitalized at our NICU.
Forty-two percent of the infants with GBS and 73 percent with E. coli EOS were born preterm. Thirteen percent of the infants with GBS and 36 percent with E. coli infection had a very-low-birth-weight (VLBW, < 1500 grams)). GBS was the most common pathogen among term (95%), preterm (84%) and VLBW (76%) infants.
There was a significant reduction of GBS during the study period (p=.0.014). The Pearson correlation coefficient was -0.601.
In comparison to neonates with GBS sepsis, neonates with E. coli sepsis had a lower GA and a lower BW. The risk factors chorioamnionitis and maternal fever occurred more frequently in the E. coli group. Perinatal data are shown in Table 1.
Clinical signs, therapy, morbidities and the length of hospitalization are shown in Table 2. There were significant differences regarding hypothermia, duration of mechanical ventilation and length of hospitalization.
Laboratory data not differed between groups (Table 3).

Discussion
In spite of the implementation of preventing strategies, GBS remained the leading cause of EOS among term, preterm and VLBW infants in our study. There was a significant decrease of GBS sepsis over the study period, but no increase of EOS caused by E. coli neither among term nor preterm or VLBW infants like in other studies [13,14]. Intrapartum antibiotic treatment is suggested to cause negative blood cultures and thus, maybe some infants with culture proven EOS were not identified.
Stoll et al. [1] reported that neonates with GBS sepsis were mostly born at term, and those with E. coli sepsis preterm. In their study the risk factors chorioamnionitis and premature rupture of membranes occurred more frequently among infants with E. coli sepsis compared to infants with GBS sepsis. Also mortality rates were higher among infants with E. coli EOS, but differences were not significant anymore after adjustment for GA. Mayor-Lynn et al. [15] reported that E. coli sepsis cases had a lower birth weight, a higher percentage with 5minute Apgar score <7, and a longer stay in the hospital neonatal intensive care unit and required mechanical ventilation more frequently. Death after early-onset neonatal sepsis with E. coli was also more frequent. Additionally the risk factor chorioamnionitis was more frequently found among neonates with GBS sepsis. The authors concluded that EOS with E. coli was associated with more morbidity and a higher mortality rate compared with early-onset GBS disease.  Our results were comparable to those reported by Stoll et al. [1] and Mayor-Lynn et al. [15]. Most infants with GBS sepsis were born at term and those with E. coli sepsis preterm. In comparison to neonates with GBS sepsis, neonates with E. coli sepsis had a lower BW, a lower GA and a longer stay at the neonatal ward. The Apgar scores at 5 and 10 minutes were lower. Mortality rate did not differed significantly between the two groups. The risk factor chorioamnionitis and maternal fever occurred more frequently in the E. coli group. We found no difference regarding the frequency of mechanical ventilation between the two groups, but neonates with E. coli sepsis had a longer duration of ventilation and longer therapy with supplemental oxygen. The higher rate of hypothermia in the E. coli group might be the result of the higher rate of preterm birth, as far as preterm infants are known Citation: Renoldner B, Hofer N, Resch B (2015) Early-Onset Neonatal Sepsis: Group B Streptococcal Compared to E. coli Disease. J Neonatal to more likely react with hypothermia to a bacterial infection than term infants [16].
Limitations of this study include the relatively small study population and the long study time period, the retrospective design of the study even when data were collected carefully and the missing control group of uninfected neonates -the latter was not the aim of our study. Nevertheless our results are proven by larger studies in the field [1,15]. The long study time period enabled to demonstrate a global reduction in GBS disease. This might be the positive result of the risk-factor based GBS screening in Austria with subsequent intrapartum antibiotics therapy.  Preventive strategies for E. coli sepsis might be based on risk factor analysis that identified chorioamnionitis and maternal fever occurring significantly more frequent in cases of E. coli infection. Despite a reduction of cases with GBS EOS over the study period vaccines for the prevention of GBS disease are urgently warranted and might further contribute to a reduction on the burden of GBS disease [17,18].
In conclusion our study confirmed that the main differences between GBS and E. coli sepsis were attributable to higher rates of preterm birth in the E. coli group. Laboratory data were not helpful in the differentiation of both pathogens.