Researchers evaluating risk stratification tools for acute pulmonary embolism have found that a recalibrated multimarker calculator and a phenotype-based framework offer improved mortality predictions for haemodynamically stable patients. The findings, published following a multicentre cohort investigation across China, compare several predictive models to determine how best to identify patients at higher risk of 30-day mortality.
Evaluating Acute Pulmonary Embolism Risk Models in Chinese Hospitals
Accurate risk stratification for haemodynamically stable acute pulmonary embolism remains clinically challenging. To address this, investigators assessed data from the prospective, multicentre China pUlmonary thromboembolism REgistry Study covering the years 2012 to 2021. The primary outcome measured across the cohort was all-cause mortality within 30 days of presentation.
The study included 11,568 adult patients. Among them, 331 individuals died within the 30-day window, representing an overall mortality rate of 2.9%. Researchers evaluated the multimarker calculator, the 2019 European Society of Cardiology model, and the American Heart Association and American College of Cardiology clinical categories using discrimination, decision-curve analysis, risk classification, and subgroup evaluations.
Comparing Model Performance and Mortality Rates
Performance metrics demonstrated that both the multimarker calculator and the 7-category AHA/ACC classification achieved better discrimination than the 2019 European Society of Cardiology model, yielding area under the curve values of 0.704 and 0.680, respectively, compared to 0.606 for the European model. However, the original multimarker calculator overestimated mortality within the intermediate-high-risk group, predicting a mean risk of 16.2% against an observed risk of 9.4%.
To correct this discrepancy, researchers applied logistic recalibration to update the linear predictor of the original multimarker calculator. This adjustment improved calibration and isolated a smaller intermediate-high-risk subgroup comprising 2.2% of patients, where the observed mortality reached 13.2%.
Phenotype Classifications and Cardiopulmonary Involvement
When assessing the AHA/ACC classification using a simplified 3-category scheme, researchers observed a stepwise increase in mortality linked to the extent of phenotype involvement. Patients classified in category B experienced a 0.7% mortality rate, while those with partial phenotype involvement faced a 2.7% mortality rate. Patients exhibiting dual cardiopulmonary phenotype involvement recorded the highest mortality at 6.7%.
These distinct risk tiers highlight the utility of incorporating clinical severity, right ventricular dysfunction, cardiac biomarker elevation, and respiratory compromise into a single framework. The recalibrated multimarker calculator provides individualized numerical estimates, whereas the AHA/ACC categories deliver a streamlined, phenotype-based grouping mechanism for clinicians managing acute pulmonary embolism.
Worth a look