Methods: A retrospective review of infants admitted to NICU with a culture-positive BSI at > 72 hours of age was done using our electronic discharge database. AM therapy was assessed for compliance with CDC guidelines on the following criteria a) AM choice based on local data and practice (vancomycin and cefotaxime or cefepime if ventilated for > 28 days, previously colonized with pseudomonas species or tracheotomized) b) timing of AM initiation-within 12 hours of septic evaluation c) AM duration –continuation beyond planned duration or inappropriate duration for focal involvement d) timeliness of line removal and e) continuation of broad-spectrum antibiotics.
Results: We identified 159 episodes of BSIs in a cohort of 88 infants, with median (range) gestational age of 27 (23-35) weeks, birth weight of 770 (410-2130) grams and 41% males. Respiratory symptoms (75 episodes, 47%) were the most common indication for septic evaluation. Leucocyte (> 20, 000), platelets (<100, 000) and C reactive protein > 10 were noted in 26 (16%), 60 (38%) and 60 (75%) infants in whom it was done respectively. The organisms were gram positive (64%), gram negative (15%) bacteria and Candida (21%). Coagulase negative staphylococcus (29%) was the commonest organism. Concomittant cerebrospinal fluid and urinary culture was performed in 121 (76%) and 51 (32%) episodes. Meningitis was identified on 14 occasions. The median (range) age at initial positive blood culture was 24.5 (4-114) days and duration of positive cultures was 3 (1-16) days. An inappropriate choice of AM at initiation was identified in 17 (10.6%), delay in initiation in 13 (8%) and a delayed or premature discontinuation in 12 (7.5%). The central venous catheter was not removed or removed late in 24 (15%) cases and broad spectrum AMs were continued in 15 (10.6%) cases for either necrotizing enterocolitis (7) or the critical condition of the infants (6). Overall, non-compliance using predefined criteria was noted in 63 (40%) cases.
Conclusion: Non-compliance with CDC guidelines on appropriate AM therapy was noted in 40% of infants in the studied cohort. The commonest deviations were failure to remove central venous catheters and continuation of broad-spectrum AM in critically ill infants, rather than target the identified pathogen. We speculate that, while ongoing QA initiatives may increase awareness and improve compliance, some deviations are inevitable in this critical vulnerable population.