Real-world validation of the 2026 ACC/AHA acute pulmonary embolism clinical categories and comparison with the 2019 ESC model

Background Risk assessment in acute pulmonary embolism (PE) is crucial to guide monitoring intensity and therapeutic decisions, yet remains challenging due to heterogeneous clinical presentations. The 2019 ESC guidelines proposed a four-tiered classification integrating hemodynamic status, biomarkers and right ventricular (RV) dysfunction. Recently, the 2026 ACC/AHA guidelines introduced a novel five-category (A–E) framework with additional subcategories and a respiratory modifier (R) to capture respiratory compromise. Real-world distribution, prognostic implications and comparative performance of this new model remain incompletely explored. Methods We retrospectively analyzed 397 consecutive patients with imaging-confirmed acute PE hospitalized within a regional hub-and-spoke network between January 2022 and December 2024. Only conservatively managed patients were included. Risk was adjudicated at admission according to ACC/AHA and ESC classifications, including ACC/AHA subcategories and R modifier. The primary outcome was a composite of PE-related in-hospital death and emergent systemic thrombolysis. Discrimination was assessed using area under the curve (AUC) analysis. Results The ACC/AHA framework yielded a broad distribution: Class A 29.7%, B 14.1%, C 35.8% (predominantly C2 19.1% and C1 11.8%), D 17.6% and E 2.8%. The composite outcome increased progressively across categories, reaching 23.8% in Class D and 63.6% in Class E. For prediction of the composite outcome, the ACC/AHA macro-class model achieved an AUC of 0.839 (95% CI 0.769–0.909), increasing to 0.857 (95% CI 0.792–0.922) with full subclass stratification. Both outperformed the four-tier ESC model (p=0.006 and p=0.001, respectively).Among patients in Classes C–E, R+ was associated with a numerically higher risk after adjustment for macro-class severity (adjusted OR 1.68, 95% CI 0.68–4.18; p=0.258). Addition of R modestly increased discrimination (ΔAUC +0.033, 95% CI −0.007 to +0.072; p=0.11). Within-class analyses suggested a heterogeneous effect, with a marked signal in Class C (OR 7.61, 95% CI 1.18–148.89) and negligible impact in Classes D–E. Conclusions In a real-world cohort of conservatively managed PE patients, the 2026 ACC/AHA classification provided superior risk discrimination compared with the ESC model, largely by refining the intermediate-risk spectrum and identifying a sizeable D phenotype. The respiratory modifier appears to offer incremental prognostic value primarily in Class C, whereas its impact diminishes in more advanced cardiocirculatory compromise.