2026 Critical Care Congress Skip to main content
2026 Critical Care Congress Main banner
Final Program


Advance Program
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.
Session Icons
Thought Leader Session
Thought Leader Session
Guidelines Session
Guidelines Session
Aging Adult Track
Aging Adult Track
Additional Registration Fee
Additional Registration Fee

Research Snapshot Theater: Cardiovascular, Pediatric IV

Perfusion Index Patterns Preceding Cardiac Arrest in the Pediatric Cardiac Intensive Care Unit

Tuesday, March 24, 2026
12:30 PM - 1:30 PM Central Time
Location: Connections Central - RST 07
  • DC

    Daniel Cater, MD

    Disclosure(s): No relevant financial relationship(s) to disclose.

  • LM

    Lee Murphy, DO

    Associate Professor of Clinical Pediatrics
    Riley Hospital for Children at Indiana University Health

    Disclosure information not submitted.

  • MF

    Matthew Friedman, MD

    Associate Professor of Clinical Pediatrics
    Riley Hospital for Children at Indiana University Health

    Disclosure information not submitted.

  • RL

    Riad Lutfi, MD

    Professor of Clinical Pediatrics
    Riley Hospital For Children at Indiana University Health

    Disclosure information not submitted.

Introduction: Cardiac arrest remains a major cause of morbidity and mortality in pediatric cardiac intensive care units (CICU). Children may show subtle warning signs before arrest, making early detection of poor circulation critical. The peripheral perfusion index (PI), a continuous, non-invasive measure from pulse oximetry, has shown promise as a marker of hemodynamic instability. We hypothesized that both declining PI and changes in PI variability occur prior to arrest and could serve as early biomarkers of deterioration.


Methods: We performed a retrospective matched cohort study using high-fidelity vital sign data. Patients who experienced CICU arrest were matched 2:1 to controls by cardiac complexity, age, weight, and operative features. For controls, pseudo-events were defined using matched arrest times or 75% of CICU stay. PI values were extracted for the 24-hours pre-event from 5-second data and aggregated at minute and hourly levels. Analyses included PI trends and variability metrics (standard deviation, coefficient of variation, and sequential variability) using Wilcoxon rank-sum tests. Variability was assessed in 10-minute rolling windows, with Bonferroni correction and effect size thresholds (Cohen’s d≥0.5) to account for multiple testing.


Results: A total of 155 patients were included (58 arrests, 97 controls). Groups were well matched (Composite Standardized Mean Difference=0.85). Males comprised 61% of patients, though females had higher arrest rates (47% vs 30%, p=0.036). Thirty-day mortality was higher in cases (22% vs 6%, p=0.005). One hour pre-event, PI was lower in arrests (0.63 vs 1.1, p< 0.001). High resolution review showed a significant decline beginning five minutes pre-arrest (p=0.017). Arrest patients demonstrated a 1.36-fold variability increase vs 0.77-fold decrease in controls (p < 0.001). Robust group separation was noted in the final two hours with 67% of time points meeting stringent significance criteria.


Conclusions: Declines in PI and increases in PI variability preceded cardiac arrest in our cohort. While absolute PI separated only minutes before arrest, variability changes appeared earlier and culminated with a distinct terminal spike. This variability signature may represent a novel early marker of impending arrest and warrants further investigation.