Journal of Tropical Pediatrics Advance Access first published online on December 6, 2006
This version published online on March 25, 2007
Journal of Tropical Pediatrics, doi:10.1093/tropej/fml073
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Study of Nosocomial Primary Bloodstream Infections in a Pediatric Intensive Care Unit
Departments of aPediatrics and bMicrobiology, Postgraduate Institute of Medical Education and Research, Chandigarh, India
Correspondence: Dr Jayashree Muralidharan, Additional Professor of Pediatrics, Advanced Pediatric Centre, PGIMER, Chandigarh–160 012, 160 012, India. E-mail: < mjshree{at}hotmail.com>.
 |
Abstract |
|---|
Bloodstream infections (BSI) are the commonest cause of nosocomial infections (NI) in PICU. Knowledge about their magnitude, risk factors and outcome are important in devising appropriate prevention and control measures. Our objective was to study the incidence, etiology, risk factors and outcome of primary BSI in PICU. A prospective cohort of 285 patients consecutively admitted to PICU from July 2003–04, having a stay of >48 h, were enrolled and monitored for BSI till discharge from ICU or death. Primary BSI was defined as per CDC criteria 1988. Data of patients with BSI was compared with those without BSI with respect to demographic details, PRISM III, primary diagnosis, nutritional status, device utilization and invasive procedures to identify risk factors for BSI. Variables significant on univariate analysis were subjected to multiple logistic regression analysis. Outcome was measured as length of PICU stay (LOS) and survival or death. There were 116 episodes of primary BSI in 86 (30%) patients; the incidence being 31.2 episodes/1000 patient days. The mean age of the patients with BSI was 3.7 ± 3.5 years. Predominant isolates were Gram-negative (53.5%); Klebsiella pneumoniae (n = 21) being the commonest. Staphylococcus aureus (n = 18) was the most common Gram-positive organism. Seven of the 9 (77.8%) yeast isolates were Candida tropicalis. Younger age, higher PRISM III, lower hemoglobin, pre-existing infection, higher frequency and duration of device utilization (CVC, urinary catheter, endotracheal tube, mechanical ventilation) were significant risk factors on univariate analysis. On multiple logistic regressions, hemoglobin (OR 1.24, 95% CI 1.1–1.4, p = 0.002) duration of urinary catheter (OR 0.91, 95% CI 0.84–0.98, p = 0.015) and pre-existing infection (OR 0.46, 95% CI 0.23–0.93, p = 0.03) were independent risk factors for primary BSI. The median LOS was significantly longer in patients with BSI compared to those without (16 vs. 7 days, p = 0.0001) 47% of patients with BSI died as compared to 26% deaths in the whole cohort (p = 0.002). Just over half the cases of BSI in our PICU were caused by Gram-negative bacteria. Lower hemoglobin, pre-existing infection and prolonged duration of urinary catheter were independent risk factors identified on multivariate analysis. BSI was associated with significantly higher mortality and longer stay in our PICU.