Abstract Background Recent studies demonstrate an improved prognostic performance of the 2014 European Society of Cardiology (ESC) algorithm for risk stratification of patients with pulmonary embolism (PE) compared to the 2008 ESC algorithm. The modified FAST and Bova scores appear especially helpful to identify PE patients at intermediate-high risk. Methods We validated the prognostic performance of the modified FAST score compared to other scores for risk stratification in a post-hoc analysis of 868 normotensive PE patients included in the prospective Italian Pulmonary Embolism Registry. In-hospital adverse outcome was defined as PE-related death, mechanical ventilation, cardiopulmonary resuscitation or administration of catecholamines. Results Overall, 27 patients (3.1%) had an adverse outcome and 32 patients (3.7%) died. The rate of an adverse outcome was highest in the intermediate-high risk classes of the 2019 ESC algorithm (7.5%) and the modified FAST score (5.3%) while the Bova score failed to discriminate between intermediate-low and intermediate-high-risk patients. Patients classified as intermediate-high risk by the 2019 ESC algorithm (Odds Ratio [OR], 4.2 [95% CI, 1.9–9.0]) and modified FAST score (OR, 2.8 [1.3–6.2]) had a higher risk of an adverse outcome compared to patients classified by the Bova score (OR, 1.6 [0.7–3.7]). The c-index was higher for the 2019 ESC algorithm and the modified FAST score (AUC, 0.69 [0.58–0.79] and 0.67 [0.59–0.76]) compared to the Bova score (AUC, 0.64 [0.55–0.73]). Conclusions The 2019 ESC algorithm provided the best prognostic performance, but also the modified FAST score accurately stratified normotensive PE patients in different risk classes while the Bova score failed to identify patients at highest risk. Graphic abstract