As part of a quality improvement project to reduce the incidence of nosocomial sepsis in VLBW babies, we did a retrospective cohort study to identify attributable risks for nosocomial sepsis in our Intensive Care Nursery. Data were abstracted for all <1500 gram birth weight admissions between 1/1/94 and 3/31/96. Of the 174 admissions during the study period, 51(29%) had nosocomial sepsis (+ blood culture at >48 hr age and antibiotic therapy for ≥5 days) with 70% due to coag negative staph.

Infants with nosocomial sepsis were more likely to be male (73% vs 52% p=0.013); lower birth weight (mean 821 gm vs 1074 gm p<0.001); and lower gestation (26 vs 29 wks p<0.001). There was no significant difference between groups for: multiple gestation, cesarian delivery, prolonged rupture of membranes, antenatal steroids or location of birth. Septic infants had higher SNAP scores at day 1 (19 vs 13 p<0.001) and at day 14 (9 vs 4 p<0.001); more ventilator days (24 vs 4 p<0.001); and more days in oxygen (40 vs 10 p<0.001). They took longer to reach full enteral feeds (33 vs 12 days p<0.001); to regain birth weight (14 vs 11 days p=0.009); and were exposed to parenteral nutrition longer (24 vs 11 days p<0.001). More septic infants had central venous lines (76% vs 30% p<0.001), umbilical artery lines (94% vs 68% p<0.001) and umbilical venous lines (94% vs 58% p<0.001).

A baseline logistic regression prediction model using birth weight, gestational age and sex accounted for 79% of the variability between septic and non-septic infants (0.79 area under the ROC curve). Variables present prior to the onset of nosocomial sepsis that were significant when added to the model included: SNAP score on Day 14 (adj odds ratio 10.0 for scores of 10-13, p=0.004; and 14.0 for scores >14, p=0.01); placement of a central venous line (adj OR 2.2, p<0.001); days of mechanical ventilation (adj OR 6.0 for 5-10 ventilator days, p<0.001; 21.0 for 11-15 ventilator days, p<0.001 and 3.8 for >16 days of ventilation, p= 0.018); and days of TPN(adj OR 9.6 for 4-10 TPN days, p<0.001; 12.0 for 11-15 TPN days, p<0.001; and 3.7 for >16 TPN days, p=0.044).

Directing quality improvement efforts toward decreasing exposure to central venous lines, and reducing TPN and ventilator days may decrease the incidence of nosocomial sepsis in VLBW infants.