A comparative analysis of the diagnostic performances of four clinical probability models to rule out pulmonary embolism

Abstract Funding Acknowledgements Type of funding sources: None. Background Ruling out pulmonary embolism (PE) through a combination of clinical assessment and Ddimer is crucial to avoid excessive computed tomography pulmonary angiography (CTPA), and different algorithms should be considered as an alternative to the fixed cutoff to achieve that goal. Purpose To compare diagnostic performance of 4 algorithms to rule out PE: fixed Ddimer cutoff, age-adjusted, YEARS and PEGeD. Methods Retrospective study of consecutive outpatients who presented to the emergency department and underwent CTPA for PE suspicion from April 2019 to May 2020. In fixed and age-adjusted cut-off, high probability patients are directly selected for CTPA. Low to moderate probability patients perform CTPA if Ddimer ≥500µg/L in fixed cutoff, and in age-adjusted cutoff if Ddimer ≥500µg/L in patients who are ≤ 50 years of age, and if Ddimer level was more than 10 times the patient’s age in patients older than 50 years. YEARS includes 3 items (signs of deep vein thrombosis, haemoptysis and whether PE is the most likely diagnosis): patients without any YEARS items and Ddimer ≥1000ng/mL or with ≥1 items and Ddimer 500ng/mL perform CTPA. In the PEGeD, patients with high clinical probability or with intermediate and Ddimers >500µg/L or low probability and Ddimer >1000 µg/L are selected for CTPA. Results We enrolled 571 patients and PE was confirmed by CTPA in 172. Compared with a fixed Ddimer cutoff, age-adjusted was associated with a significant increase of specificity (p < 0.001), correctly avoiding 38 CTPAs, without losing sensitivity. YEARS and PEGED resulted in a increase in specificity, compared to the fixed cutoff, but with an impairment of sensitivity (p < 0.001). PEGeD had the worst sensitivity (13 more false negatives (FN) than the fixed cutoff). Despite the lack of difference between PEGed and YEARS strategies regarding sensitivity, PEGED had a significantly higher specificity (p < 0.001) and allowed to correctly avoid a higher number of CTPA (95 vs 85), compared to the fixed cutoff. Conclusion Compared to fixed d-dimer cutoff, all algorithms were associated with an increased specificity. Age-adjusted cutoff was the only that is not associated with a significant decrease in sensitivity when compared to fixed cutoff, allowing to safely reduce the need to perform CTPA. Sen(%)Spec(%)Correctly avoid CTPAs(n)FN(n)Fixed cutoff251017Age-adjusted933513912YEARS894718619PEGED884919620