Comparative validation of AHA/ACC 2026 and ESC 2019 risk stratification in hospitalized patients with acute pulmonary...

The 2026 AHA/ACC guideline introduced an A-E classification with a respiratory (R) modifier for pulmonary embolism (PE), but real-world validation is limited.

We classified 1050 inpatients with CTPA-confirmed acute PE (2014-2025) at a single tertiary center by ESC 2019 and AHA/ACC 2026 criteria. The primary endpoint was in-hospital all-cause mortality; the key secondary endpoint was a composite of death, hemodynamic decompensation, or rescue therapy; the secondary endpoint added unplanned ICU transfer or invasive ventilation. Discrimination was compared by AUC.

Median age was 63 years; 46/1050 (4.4%) died in hospital. ESC 2019 classified 792/1050 (75.4%) as intermediate-low risk; AHA/ACC 2026 reclassified 611/792 (77.1%) of these to C2 or higher, including 272/792 (34.3%) to D2/D2R (mortality 22/272 [8.1%], two-thirds of intermediate-low deaths). AHA/ACC 2026 showed higher discrimination for in-hospital death (AUC, 0.716 vs 0.610; P < .001), the key secondary endpoint (0.611 vs 0.545; P = .007), and the secondary endpoint (0.676 vs 0.535; P < .001). The advantage persisted after collapsing AHA/ACC into a 4-group ordinal score. In patients with active cancer, AHA/ACC 2026 discriminated in-hospital death better than ESC 2019 (AUC, 0.745 vs 0.535; P < .001), whereas no significant difference was observed in those without cancer (0.698 vs 0.683; P = .734).

AHA/ACC 2026 improved discrimination over ESC 2019 mainly by identifying normotensive shock (D2/D2R) within the ESC intermediate-low stratum. The advantage was concentrated in patients with active cancer and may partly reflect comorbidity rather than PE-specific risk. Prospective multicenter validation is needed.