This analysis demonstrates improved prognostic accuracy for acute pulmonary embolism using a Bayesian model integrating plasma lactate levels, suggesting it enhances risk stratification.
Key Points
This research aims to evaluate the prognostic value of plasma lactate in risk assessment for acute pulmonary embolism when combined with a Bayesian probability model.
Defined pretest probabilities using simplified PE Severity Index (sPESI) and SI.
Analyzed sensitivity and specificity for plasma lactate >2 mmol/L from previous studies.
Utilized Bayesian nomogram to estimate posttest probabilities with likelihood ratios.
Assessed diagnostic efficiency employing Bayesian Diagnostic Gain (BDG) and Bayesian Number Needed to Diagnose (BNND).
Thirty-day mortality for SI >1 was 24.1%, compared to 10.7% for low-risk SI and sPESI.
Lactate >2 mmol/L showed sensitivity of 82.4% and specificity of 73.5%.
Integrating lactate into the SI model resulted in an Absolute Diagnostic Gain of 25.9%, exceeding the sPESI model's gain of 16.3%.