Comparative accuracy of the STIMATE and ESC risk models for predicting adverse outcomes in acute pulmonary embolism

Reliable early risk assessment is essential in acute pulmonary embolism (PE) to guide the therapeutical approach in these patients, that is why several scores are used in daily clinical practice.

The STIMATE (eStiMaTe®) score is a validated quantitative model that estimates 30-day mortality by integrating clinical and biomarker variables. Nevertheless, the ESC risk classification remains the reference standard, but due to being categorical may underestimate risk as a result of the heterogeneity within intermediate-risk groups.

We intended to compare the discriminative and calibration performance of both models in a contemporary real-world cohort.

Methods: We analyzed 208 consecutive patients from the TEP-HUC Registry (2025). Each patient had an ESC score (0–3) and a continuous STIMATE score (% estimated 30-day mortality).

The assessed endpoints included: In-hospital death, PE readmission and a combined of death or PE readmission.

Discrimination was assessed by ROC-AUC (95% CI) and compared using the DeLong test.

Calibration was evaluated by the Hosmer–Lemeshow test and Brier score.

Both models showed adequate calibration (Hosmer–Lemeshow p > 0.6; Brier ≈ 0.09).

In the intermediate-high risk ESC stratum, STIMATE retained good discrimination (AUC = 0.82 for mortality).

Conclusions: In this unselected PE cohort, the STIMATE and ESC risk models demonstrated comparable predictive performance for short-term adverse outcomes.

Although differences were not statistically significant, STIMATE showed a consistent trend toward better discrimination for mortality and combined events.

These results suggest that STIMATE may complement the ESC classification, providing a more precise stratification within the heterogeneous intermediate-risk population, which is needed.

Comparative performance