Measurement error of pulse pressure variation.

Area of Science:

• Critical Care Medicine
• Anesthesiology
• Cardiovascular Physiology

Background:

• Dynamic preload parameters, such as pulse pressure variation (PPV), guide perioperative fluid management.
• Clinical decision-making is complicated by variable reported cut-off values and a 'gray zone' for PPV.
• Measurement error in PPV calculations has not been previously quantified but may contribute to uncertainty.

Purpose of the Study:

• To quantify and compare measurement errors associated with different algorithms for calculating PPV.
• To assess the impact of sampling period duration on PPV measurement uncertainty.
• To provide a framework for interpreting PPV values considering bias and precision.

Main Methods:

• Utilized hemodynamic data from liver transplant patients in the VitalDatabase.
• Applied three distinct algorithms to calculate PPV over 1-minute observation periods with varying sampling durations.
• Employed Best Linear Unbiased Prediction as the reference PPV and a Bayesian model to determine bias, precision, and simulate uncertainty.

Main Results:

• All tested PPV calculation methods exhibited measurement error, with differential bias ranging from -0.04% to 1.64% and proportional bias from 0.92% to 1.17%.
• Heteroscedasticity, influenced by sampling period, was detected across all methods.
• Simulations showed significant predicted ranges for reference PPV values and changes, even for a measured PPV of 12% or a 1% change.

Conclusions:

• All methods for calculating PPV are associated with inherent measurement uncertainty.
• The degree of bias and precision varies among different PPV calculation methods and is influenced by sampling parameters.
• Accounting for quantified bias and precision is crucial for accurate interpretation of PPV values and changes in clinical practice.