- Design
- Retrospective two-campus cohort
- Population
- 2,251 inpatients with acute pulmonary embolism
- Primary outcome
- 30-day all-cause mortality discrimination
- Effect
- AUC 0.885 (AHA/ACC) vs 0.763 (ESC); difference 0.122 (95% CI 0.084-0.160)
A retrospective study at a two-campus Chinese tertiary hospital compared the 2026 AHA/ACC acute pulmonary embolism clinical categories with the 2019 ESC risk classification in 2,251 inpatients, using 30-day mortality as the outcome; 5.6% died.
The AHA/ACC categories discriminated 30-day mortality better (AUC 0.885 vs 0.763; difference 0.122, 95% CI 0.084 to 0.160), and PE-related death even more so. At high-risk thresholds, the AHA/ACC top categories caught 84% of PE-related deaths against 59% for ESC high risk, with similar specificity. Thirty-seven patients placed in the AHA/ACC top categories but called intermediate risk by ESC had a 30-day mortality of 76%.
The difference narrowed when the extreme categories were excluded, so part of the gain comes from the broader scale itself. It is one retrospective cohort, and better prediction does not by itself show that treating differently helps. The practical lesson is that some normotensive patients the ESC scheme calls intermediate risk are very sick and need closer watching.
- Do not assume a normotensive patient with pulmonary embolism is safe because they are not ESC high risk.
- Look for signs of impending deterioration such as rising lactate, worsening right ventricular function or escalating oxygen need.
- Consider monitoring intermediate-risk patients with these features in a high-dependency setting.
- The newer AHA/ACC categories may help identify this group; local adoption should follow validation.
Why it matters
A small group labelled intermediate risk under the ESC scheme had a mortality closer to high risk.
Don't overread it
One retrospective cohort; better prognostic discrimination does not show that changing treatment improves outcomes.
The statistics, in plain English
The area under the curve measures how well a score ranks patients who died above those who survived: 0.5 is chance, 1.0 perfect. A rise from 0.76 to 0.89 is substantial, but it shrank when the extreme categories were removed, which means part of the advantage comes from how widely the new scale spreads patients.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for pulmonology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free