Hospitalization precision in acute pulmonary embolism: a comparison of contemporary risk stratification models

Abstract Background and aim Although current pulmonary embolism guidelines recommend outpatient management for carefully selected patients with acute pulmonary embolism, many patients at low risk continue to be hospitalized. Real data evaluating the appropriateness of hospitalization decisions and agreement among contemporary PE risk stratification models remain limited. The primary aim of this study was to evaluate the appropriateness of hospitalization decisions in patients with APE using​ contemporary risk stratification models. The secondary aim was to compare agreement among these models and their prognostic performance in relation to hospitalization precision. Methods We retrospectively analyzed a prospectively maintained cohort of 150 consecutive patients hospitalized with APE between January 2022 and January 2024. APE was confirmed by computed tomography pulmonary angiography. Hospitalization decisions were evaluated using the Hestia criteria, PESI, simplified PESI, the Bova score, the 2019 ESC risk classification, and the 2026 AHA/ACC PE categories. Cohen’s kappa and McNemar’s tests were used to assess agreement and differences in paired classifications between Hestia and the other models. Prognostic performance for all-cause mortality was evaluated using ROC curve analysis. Results The mean age was 52 ± 17.0 years, and 34 patients (22.7%) died during follow-up. According to the Hestia rule, 35 patients (23.3%) fulfilled the criteria for outpatient management despite having been hospitalized. Low-risk classification was assigned to 35.3% of patients by the ESC algorithm, 30.7% by sPESI, 38.0% by PESI, and 42.7% by the AHA/ACC categories. Combining Hestia with additional clinical or imaging criteria reduced outpatient eligibility to 16.7%–22.7%. Agreement between Hestia and the other risk models ranged from fair to moderate (κ = 0.35–0.58). The 2026 AHA/ACC classificati​on showed the highest discrimination for mortality (AUC, 0.889), followed by PESI (0.870), ESC (0.865), and Bova (0.810). Hospitalization precision ranged from 52.0% to 76.7% across the evaluated strategies. Conclusions A substantial proportion of hospitalized patients with APE fulfilled validated low risk criteria for outpatient management, suggesting that potentially avoidable hospitalization persists in routine practice. Contemporary risk stratification models showed fair to moderate agreement and should be considered complementary rather than interchangeable. Integrating outpatient eligibility criteria with multidimensional risk assessment may improve hospitalization decisions and optimize healthcare resource utilization. Clinical trial number and affiliation approval number Approval No. AU17101855 (Ethics Committee, Faculty of Medicine, Assiut University); ClinicalTrials.gov Identifier: NCT05469724.