Comparison of the safety and efficacy of YEARS, PEGeD, 4PEPS or the sole item "PE is the most likely diagnosis" strategies for the diagnosis of pulmonary embolism in the emergency department: post-hoc analysis of two European cohort studies.

Background: The optimal strategy for the diagnosis of pulmonary embolism (PE) in the emergency department (ED) remains debated. To reduce the need of imaging testing, several rules have been recently validated using an elevated D-dimer threshold.

Objective: To validate the safety of different diagnostic strategies and compare the efficacy in terms of chest imaging testing.

Design And Patients: Post-hoc analysis of individual data of 3330 adult patients without a high clinical probability of PE in the ED followed-up at 3 months in France and Spain (1916 from the PROPER cohort, 1414 from the MODIGLIANI cohort).

Exposure: Four diagnostic strategies with an elevated D-dimer threshold if PE is unlikely. The YEARS combined with Pulmonary Embolism Rule-out Criteria (PERC) the pulmonary embolism graduated D-dimer (PEGeD) combined with PERC and the 4-level pulmonary embolism probability score (4PEPS) rules were assessed. A modified simplified (MODS) rule with a simplified YEARS reduced to the sole item of "Is PE the most likely diagnosis" combined with PERC was also tested.

Outcome Measure And Analysis: The primary outcome was the proportion of diagnosed PE or deep venous thrombosis at 3 months in patients in whom PE could have been excluded without chest imaging according to the tested strategy. The safety of a strategy was confirmed if the failure rate was less than 1.85%. The secondary outcome was the use of imaging testing according to each rule.

Results: Among 3330 analyzed patients, 150 (4.5%) had a PE. The number of missed PEs were 25, 29, 30 and 26 for the PERC+YEARS, PERC+PEGeD, 4PEPS and MODS rules respectively, with a failure rate of 0.75% (95% CI 0.51% to 1.10%), 0.87% (0.61% to 1.25%), 0.90% (0.63% to 1.28%) and 0.78% (0.53% to 1.14%) respectively. There was no significant difference in the failure rate between rules. Except for a significant lower use of chest imaging for 4PEPS compared to YEARS (14.9% vs 16.3%, difference -1.4% [95%CI -2.1% to -0.8%]), there was no difference in the proportion of imaging testing.

Conclusion: In this post-hoc analysis of patients with suspicion of PE, YEARS and PEGeD combined with PERC, and 4PEPS were safe to exclude PE. The safety of the modified simplified MODS strategy was also confirmed. There was no significant difference of the failure rate between strategies.