Association Between Intracranial Pressure Monitoring and Discharge Outcome in Pediatric Severe Traumatic Brain Injury

To determine whether intracranial pressure (ICP) monitoring is associated with decreased mortality among pediatric severe traumatic brain injury (TBI) patients.

Retrospective cohort study of the Virtual Pediatric Systems database.

Patients younger than 18 years with diagnosis codes for TBI and Glasgow Coma Scale (GCS) of less than or equal to 8 between 2010 and 2024 from centers submitting ICP monitor data.

ICP monitoring practices were described using summary statistics. To estimate time-dependent association between ICP monitor placement and 30-day mortality, time-varying Cox proportional hazards modeling was performed. In a second complementary analysis to minimize immortal time bias, landmarked unadjusted Kaplan-Meier analyses and adjusted Cox proportional hazards models of ICP monitor placement by 1, 12, and 24 hours were performed.

There were 6636 admissions included across 94 PICUs between January 1, 2010, and December 31, 2024; 2016 of 6636 (30.7%) had ICP monitors. Patients with ICP monitors were older (8 yr [3-14] vs. 6 yr [2-13], p < 0.001), had lower GCS (3 [3-6] vs. 6 [3-7], p < 0.001), and a higher Pediatric Risk of Mortality 3 score (12.0 [8.0-21.0] vs. 9.0 [5.0-21.0], p < 0.001). ICP monitor placement was associated with lower time-specific mortality hazard until 4.4 d (95% CI, 3.7-5.2) in unadjusted and 10.5 d (95% CI, 7.2-30.0) in adjusted time-varying Cox modeling, resulting in lower 30-day cumulative mortality in the adjusted model only (29.2% [95% CI, 28.0-30.6%]) in the unmonitored group vs. (25.1% [95% CI, 23.4-26.5%]) in the monitored group. In landmark analysis and adjustment for severity, ICP monitor placement was associated with a significantly lower 30-day-mortality hazard ratio (HR) at the 1-hour (adjusted HR 0.81 [0.69, 0.94]), 12-hr (0.83 [0.73, 0.95]), and 24-hr landmarks (0.80 [0.69, 0.92]).

ICP monitoring was associated with reduced mortality for 10 days after admission, followed by increased mortality between days 11 and 30. The stronger effect of reduced early mortality resulted in lower 30-day cumulative mortality.