Presented at the Neonatal Society 2007 Spring Meeting.
Samanta S1, Farrer K1, Breathnach A2, Heath PT3 (introduced by S Calvert)
1 Department of Neonatology, St George’s Hospital, London SW17 0QT, UK
2 Department of Microbiology, St George’s Hospital, London SW17 0QT, UK
3 Paediatric Infectious Diseases Unit, St George’s Hospital, London SW17 0QT, UK
Background: Late onset Gram negative sepsis (LOGNS) is a significant cause of mortality and morbidity in neonatal units. A previous audit in our unit had identified an increased incidence of LOGNS due to Klebsiella and this stimulated the study.
Aim: To describe consecutive late onset Gram negative infections over a 6 year period in our Neonatal Unit and determine the incidence and mortality from LOGNS. To assess risk factors and potential contributors to the development of LOGNS.
Methods: We performed a retrospective case control study from 1999 – 2005. Cases were defined as all infants with late onset (>48hrs of life) infection in which Gram negative bacteria were grown from a normally sterile site. Cases were compared with two controls, matched for gestation and time of admission to the neonatal unit. For cases with >1 episode of LOGNS, only the first episode was considered for analysis. Incidence was described per 1,000 live births and stratified for very low birth weight (VLBW) (<1500gm) infants. Data was analysed using SPSS v.10 statistical package.
Results: There were 73 cases of LOGNS identified over the study period of which 48 were inborn and included in the study. These cases experienced 53 episodes of LOGNS. The median gestation for cases and controls were 26 weeks and median birth weight 785gm and 889 gm respectively. Intra uter ine growth restriction (IUGR) was present in 23% of cases compared to 20% in the control group. 42% of case had chronic lung disease (CLD) defined as oxygen dependency at 36 weeks corrected gestational age compared to 45% in controls. A primary surgical pathology was twice more common in cases than controls (12.5% vs. 6.25%). The median age at presentation with LOGNS was 18 days. Enterobacter (28%), E. Coli (27%) and Klebsiella (21%) were the commonest pathogens identified. In the vast majority (90%) infection was identified in blood alone. Only 2 (4%) babies had confirmed meningitis with organism identified in cerebrospinal fluid (CSF) as well. 3 babies with blood culture positive for Gram negative infection had evidence of infection on their CSF with increased cellular ity or evidence of the pathogen on autopsy but negative CSF and was treated for clinical meningitis. In 6% of cases there was a positive urine and blood culture with Gram negative organism.
The overall incidence of LOGNS was 1.85/1000 live births. Thus 2.2% of total admissions and 4.4% of inborn VLBW admissions to the neonatal unit over the study period had LOGNS. 75% of cases were <30 weeks and 65% <1000gm. Mortality was 27% in cases compared to 13.5% in controls (p=0.07). There was no significant difference between cases and controls with respect to maternal risk factors for infection (clinical chorioamnionitis, maternal pyrexia, prolonged rupture of membrane, intrapartum antibiotic prophylaxis, antenatal steroids) or use of nasal CPAP. Mechanical ventilation, Total Parenteral Nutrition (TPN) and its duration, presence of central line and its duration, use of Tazocin and Vancomycin and occurrence of Necrotising Enterocolitis (NEC) at or before first positive culture were all significantly associated with case status in univariate analysis. In multivariate logistic regression analysis, however, only duration of Total Parenteral Nutrition at or before first positive blood culture remained independently associated with late onset Gram negative septicaemia (p<0.001).
Conclusions: Gram negative infection is a significant cause of late onset neonatal sepsis. It occurs predominantly in preterm and VLBW infants and is not associated with maternal or perinatal risk factors. Duration of TPN at or before first positive culture is significantly associated with LOGNS in our population although this relationship is likely to be complex.
Ethical Approval: Project was approved by Local Research and Ethics committee (LREC) mentioning that formal ethical approval was not required.