To compare the prognostic accuracy of the 2014 risk model of the European Society of Cardiology (ESC) and of Bova and TELOS scores for identification of normotensive patients with pulmonary embolism (PE) at high risk for short-term adverse events (i.e., intermediate-high risk patients), we retrospectively applied these tests to a prospective cohort of 994 normotensive patients with objectively confirmed PE. Sixty-three (6.3 %) patients reached the primary outcome, a composite of hemodynamic collapse and death within 7 days from diagnosis. The Bova and TELOS scores classified the same proportion of patients in intermediate-high risk category (5.9 and 5.7 %, respectively), with a similar primary outcome rate (18.6 and 21.1 %, respectively). The 2014 ESC model classified in the intermediate-high risk category the largest proportion of patients (12.5 %, p < 0.001 vs Bova and TELOS), with the lowest primary outcome rate (13 %, p = ns vs Bova and TELOS). When lactate determination was added to the Bova score, 112 patients (11.2 %) were classified in the intermediate-high risk category (p < 0.05 vs Bova and TELOS), with a slight increase in the primary outcome rate (25.9 %, p = 0.014 vs 2014 ESC model), allowing the recognition of a twofold higher number of patients reaching the primary outcome (29 vs 15, 11 and 12 patients in the 2014 ESC model, Bova and TELOS scores, respectively, p < 0.01 for all). The 2014 ESC model, Bova and TELOS scores identify a small number of intermediate-high risk patients with PE, without differences among tests. Adding plasma lactate to the Bova score significantly improves the identification of intermediate-high risk patients.