Design: Retrospective analysis of prospective multicentre PE registry
Population: 1,731 ED patients with pulmonary embolism, six US centres
Primary outcome: In-hospital death or clinical deterioration
Effect: CPES ≥3 sensitivity 72%, specificity 57%, AUC 0.68; advanced intervention AUC 0.78

This study, published in May, used prospectively collected registry data from six US academic emergency departments (2016–2020) to compare the Composite Pulmonary Embolism Shock (CPES) score with the simplified PESI and the ESC risk classes in 1,731 patients with pulmonary embolism.

In-hospital death or deterioration occurred in 11.1%. A CPES score of 3 or more had 72% sensitivity and 57% specificity for that outcome (AUC 0.68), with likelihood ratios similar to sPESI and ESC. CPES separated patients better for advanced intervention, such as thrombolysis or catheter therapy (AUC 0.78; negative LR 0.20). Thirty-day mortality was predicted similarly by all three.

The practical message is that the familiar tools still work. CPES may help flag patients likely to need escalation, but it does not replace sPESI and ESC stratification for disposition.

Keep using sPESI and ESC risk classes to stratify PE in the emergency department.
Consider CPES as an additional flag for patients who may need advanced intervention.
Note central thrombus as a marker for likely escalation, not for death.
Combine scores with right ventricular assessment and troponin rather than relying on one number.

Why it matters

A new score is only useful if it beats the old ones, and for deterioration this one did not.

The statistics, in plain English

An AUC of 0.68 is modest — the score ranks a deteriorating patient above a stable one only about two thirds of the time. A negative likelihood ratio of 0.20 for advanced intervention means a low score substantially lowers that probability.