(1679) Arterial Line Placement on Admission Does Not Independently Increase Risk of Longer PICU Stay

Introduction: The effects of arterial line (A-line) placement on outcomes in the pediatric ICU (PICU) are not well described. We previously demonstrated that A-line placement associated with higher vasoactive infusion score (VIS) and with longer PICU lengths of stay (LOS), even when controlling for illness severity. The aim of the study was to assess if early A-line placement, excluding later placement, was still associated with longer LOS.

Methods: A retrospective review of patients admitted to a 37-bed tertiary-care PICU from Mar 2020 - Aug 2021 exposed to >8h of vasoactive therapy was performed. Subjects who received A-lines for severe pARDS or outside of the PICU were excluded, as were patients with LOS < 2 days. For this analysis, patients who had A-lines placed beyond 2 days of PICU admission were also excluded. Pediatric Sequential Organ Failure Assessment (pSOFA) and VIS were recorded. Fine and Grey subdistribution hazard modeling was used to evaluate the impact of A-line placement within 2 days of ICU admission on LOS.

Results: 121 patients exposed to vasoactive medications for >8h were included. Of 70 subjects who received A-lines, 57 subjects (81%) received A-lines within 2 days of admission. Patients who did not undergo A-line placement within the first 2 days had a 2 times greater increase in the rate of being discharged alive sooner compared to those with A-line placement (p=0.0004, hazard ratio (HR): 2.0 (1.3-2.8). Max VIS within the first 2 days was not associated with LOS when A-line placement was considered. Higher max pSOFA within the first 2 days decreased the incidence of earlier discharge after adjusting for A-line status. When both max VIS and pSOFA were considered, A-line placement was no longer significantly associated with LOS (p=0.07). Our final model showed that max pSOFA within the first 2 days, primary diagnosis, having comorbidities and neurologic/metabolic disease status were associated with LOS (p < 0.008 for all); A-line placement was no longer significant.

Conclusions: Early A-line placement associated with longer LOS univariately. When adjusted for illness severity, diagnosis and comorbidities (chronic illness), early A-line placement does not seem to impact LOS. Further study of A-line placement practice and utility, and its effects on patient outcomes is warranted.