Comparison of Short vs. Extended Antibiotic Durations for Patients in the Neonatal Intensive Care Unit with Late Onset, Culture-negative Sepsis

Abstract Background Culture-proven late-onset sepsis in neonates and preterm infants, defined as symptom onset after 3-7 days of life, is a common cause of morbidity and mortality. Some patients are treated with antibiotics for sepsis despite lack of positive cultures after 48 hours of incubation, potentially due to the non-traditional signs of sepsis. Controversy remains over optimal treatment durations of culture-negative sepsis with durations reported as 2 to 7 days, and rarely longer than 7 days. The primary outcome of this study is to compare lack of symptom resolution at 7 days in patients with negative cultures treated with antibiotics for either ≤72 hours or >72 hours. Methods This was a single-center, retrospective, cohort study with Institutional Review Board approval at a 60 bed, level 4, surgical referral neonatal intensive care unit from January 1, 2020, to December 31, 2021. Patients were included for screening if they were ≥7 days of age, ≤ 44 weeks postmenstrual age, and received more than 1 day of antibiotics for possible late-onset, culture-negative sepsis. Patients were then excluded if they had positive culture results from any site, diagnosis of pneumonia, were being treated for suspicion of necrotizing enterocolitis, received antibiotics prior to obtaining cultures, or had continued treatment of prior infection. Patients were divided into two groups based on antibiotic duration: ≤72 hours or >72 hours. Presenting symptoms and symptom resolution at 72 hours and 7 days were assessed. Chi-squared or Fisher’s exact tests were used for categorical comparisons. Results Of the 909 patients screened, 68 were included: 42 in the ≤72 hours group and 26 in the >72 hours group. The most prevalent symptoms prompting antibiotic initiation were elevated CRP (64.7%) and bradycardia/oxygen desaturations (78%), with elevated CRP being the initiating symptom more frequently in the >72-hour group (92% vs 48%, p<0.001). If patients had 2 or more symptoms, they were more likely to receive >72 hours of antibiotics (85% vs 62%, p=0.034). At 7 days, 28 (41%) patients had complete symptom resolution. Antibiotic duration ≤72 hours was not associated with symptom resolution at 7 days (50% symptom resolution in the ≤72-hour group vs 27% in the > 72-hour group, p=0.06). Lack of symptom resolution at 72 hours of antibiotics was associated with lack of resolution at 7 days (p<0.001). Among patients without symptom resolution at 7 days, 65% of patients had a persistently elevated CRP and 68% of patients continued to have bradycardia/oxygen desaturations. Conclusion In neonates and preterm infants with negative cultures but possible signs of sepsis, antibiotic duration does not impact 7-day symptom resolution. Results indicate that if the initiating symptom is not resolved at 72 hours despite negative cultures, it is unlikely to be resolved with additional days of antibiotics. This is likely due to the multifactorial potentially non-infectious etiology of initiating symptoms in these patients. It is appropriate to discontinue antibiotics if cultures are negative after 48 hours of incubation in this population.