Evaluation of Plateau CO2 via Inspiratory Hold to Estimate PaCO2 in Swine Pediatric ARDS Model

Tuesday, March 24, 2026

10:00 AM - 11:00 AM Central Time

Location: Connections Central - RST 13

Kaisa Stolen, BS (she/her/hers)

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

Vicky Chan, BS

Research Assistant
University of Wisconsin School of Medicine and Public Health

Disclosure information not submitted.
PM

Paige Munns, CVT, RALAT

Research Specialist
University of Wisconsin-Madison/Center for Biomedical Swine Research (CBSRI)

Disclosure information not submitted.
ML

Michael Lasarev, MS

Biostatistician
Department of Biostatistics and Medical Informatics, University of Wisconsin-Madison

Disclosure information not submitted.
DK

Devon Klipsic, DVM, MS

Research Veterinarian & Anesthesia Core Veterinarian
University of Wisconsin-Madison/Center for Biomedical Swine Research (CBSRI)

Disclosure information not submitted.
ES

Eric Schmuck, PhD

Director of Research Operations
University of Wisconsin-Madison/Center for Biomedical Swine Research (CBSRI)

Disclosure information not submitted.
Awni Al-Subu, MD

University of Wisconsin Hospital and Clinics, Wisconsin

Disclosure information not submitted.

Introduction: Pediatric acute respiratory distress syndrome (PARDS) accounts for approximately 4% of PICU admissions, with severe cases often necessitating invasive mechanical ventilation and frequent arterial blood gas monitoring. End-tidal CO₂ (ETCO₂) is a non-invasive alternative but is limited by dead space in ARDS. We hypothesized that end-inspiratory pause (EIP) may reduce dead space and improve real-time estimation of PaCO₂.

Methods: Juvenile Yorkshire swine (n=11, 6 male, 5 female), mean age 68.6 (SD 14.1) days, weight of 20.7 (1.7) kg were sedated, paralyzed, and mechanically ventilated. ARDS was induced via continuous oleic acid infusion to achieve a PF ratio < 300 mmHg. Hemodynamic and respiratory parameters were continuously monitored using volumetric and mainstream capnography. Dead space ventilation to tidal volume (Vd/Vt) and end-tidal alveolar dead space fraction (AVDSf) were calculated. EIPs were applied at 3, 5, and 10 seconds before and after ARDS induction, with 5-minute stabilization periods between measurements. PaCO₂ and ETCO₂ values were aligned at corresponding time points, and Lin’s concordance correlation coefficient (ρc) with 95% confidence interval was used to evaluate agreement.

Results: Median [IQR] time to ARDS onset was 114 [88–158] minutes. Baseline oxygenation index (OI) was 1.5 [1.0–1.9], rising to 4.9 [4.0–6.2] in ARDS. AVDSf and Vd/Vt significantly increased following lung injury, with AVDSf correlating strongly with Vd/Vt (rp = 0.76). At baseline, ρc between PaCO₂ and ETCO₂ at 3, 5, and 10 seconds was 0.44, 0.52, and 0.66, respectively. Corresponding ρc values between PaCO₂ and ETCO₂ during EIP were 0.39, 0.33, and 0.12. In PARDS, ρc between PaCO₂ and ETCO₂ was 0.46, 0.50, and 0.52, while ρc for PaCO₂ and EIP-ETCO₂ was 0.37, 0.18, and 0.58, respectively.

Conclusions: This is the first study to examine the impact of varying EIP durations on dead space estimation and PaCO₂-ETCO₂ agreement in a swine model of PARDS. Contrary to our hypothesis, EIPs did not improve concordance between ETCO₂ and PaCO₂. These findings suggest limited utility of EIP for enhancing non-invasive CO₂ monitoring in PARDS.