Effective shunt

Effective shunt estimates the effective shunt fraction non-invasively from arterial blood gas data to improve pulmonary oxygenation assessment in critically ill patients.

Key Features:

• Non-Invasive Estimation: Calculates shunt fraction by rearranging the indirect Fick equation, incorporating standard constants and inverting the relationship between content and tension to allow estimation from a single arterial blood gas (ABG) sample.
• Predictive Validity: Validated in a retrospective analysis of 78,159 ABG results and demonstrated the lowest median absolute error (MAE) in predicting PaO₂ after FiO₂ changes, with an MAE of 1.88 kPa.
• Comparison with Other Methods: Compared against the alveolar-arterial difference, the P/F ratio, and inferred values from a DB derived from an integrated mathematical model of gas exchange, with ES outperforming these measures (p < 10⁻¹⁰ in all comparisons).
• Computational Efficiency: Computation is efficient and performed using routinely collected ABG data.
• Enhanced Precision with Dual Measurements: Analysis indicates that two blood gas measurements at different FiO₂ levels provide a more precise description of pulmonary oxygenation when using the DB method, suggesting potential enhancements to ES methodology.

Scientific Applications:

• Assessing disease severity: Provides non-invasive quantification of shunt fraction for evaluating pulmonary impairment in critically ill patients.
• Quantifying pulmonary outcomes: Supports research studies that require accurate measurement of oxygenation and changes in PaO₂ in response to FiO₂ adjustments.

Methodology:

Rearrangement of the indirect Fick equation with standard constants and inversion of the content–tension relationship to estimate shunt from a single ABG sample; validated by retrospective analysis of 78,159 ABG results and compared to alveolar-arterial difference, P/F ratio, and DB-inferred values from an integrated mathematical model of gas exchange; two-measurement FiO₂ comparisons evaluated for DB precision.