438 - Variations in antibiotic use for culture-negative sepsis in near preterm and term newborns: A nationwide population based cohort

Publication Number: 2426.438

Johan Gyllensvärd, Ryhov County Hospital, Jönköping, Jonkopings Lan, Sweden; Marie Kristina. Studahl, Gothenburg University, Institute of Biomedicine, Göteborg, Vastra Gotaland, Sweden; Lars Gustavsson, The Sahlgrenska Academy, University of Gothenburg, Gothenburg, Vastra Gotaland, Sweden; Elisabet Hentz, Saglgrenska university hospital Gothenburg Sweden, Goteborg, Vastra Gotaland, Sweden; Karin Åkesson, Linköping University, Jönköping, Jonkopings Lan, Sweden; Huiqi Li, University of Gothenburg Institute of Medicine: Goteborgs universitet Institutionen for medicin, Gothenburg, Vastra Gotaland, Sweden; Mikael Norman, Karolinska Institutet, Stockholm, Stockholms Lan, Sweden; Anders Elfvin, University of Gothenburg, Gothenburg, Vastra Gotaland, Sweden

Background: Antibiotic use for suspected culture-negative sepsis (CN-sepsis) varies across neonatal networks. Monitoring variations across regions and units is key to refine best practice.

Objective: To quantify regional and unit variations in antibiotic use for CN-sepsis in the first week of life.

Design/Methods: This nationwide population-based cohort included all hospital live births from 34 weeks’ gestation during 2012 to 2020. Data were extracted from the Swedish Neonatal Quality Register and the Swedish Medical Birth Register. CN-sepsis (ICD-10 code P36.9) was defined as newborns with obvious clinical signs and symptoms compatible with sepsis, negative blood cultures or when blood cultures were not taken, elevated biomarkers for infection, and intravenous antibiotic therapy for at least 5 days (or if the patient died before 5 days of treatment).

Results: Out of 1 025 515 live births, 19 286 (19 per 1000 live births) received antibiotics during the first week of life. The rate of CN-sepsis, culture-positive sepsis, and newborns treated with antibiotics but without having sepsis were 4.1 (4157 newborns), 0.7 (719 newborns), and 14.1 (14 410 newborns) per 1000 live births, respectively. The rate of CN-sepsis varied 7-fold to 79-fold across regions (1.6 to 10.8 per 1000 live births) and units (0.3 to 23.9 per 1000 live births), respectively. The median duration of antibiotic treatment in newborns with CN-sepsis was 7 days in all regions and varied between 5 to 10 days across units. The number of antibiotic days for CN-sepsis per 1000 live births were 29 with a 6-fold to 62-fold variation across regions (12 to 71 days per 1000 live births) and units (2 to 151 days per 1000 live births), respectively. All-cause mortality was 0.46 per 1000 live births of which CN-sepsis contributed to 5.6% (26 of 468 live births). In newborns with CN-sepsis the mortality rate per 1000 live births was 6.3 (26 of 4157 live births) with a 2-fold variation across regions (4.5 to 8.3 per 1000 live births).
CN-sepsis was 5.8 times more common in newborns as compared to newborns with culture-positive sepsis, with a 4-fold to 16-fold variation across regions (2.8 to 10.9) and units (1.0 to 15.5).

Conclusion(s): There were wide regional and unit variations in CN-sepsis, which greatly influenced early antibiotic use in newborns.
To improve care and long-term health, future antibiotic stewardship programs should implement more uniform and restrictive criteria for CN-sepsis and aim to discontinue treatment earlier to minimize unwarranted antibiotic use.