P5019Derivation and validation of a new probability score in pulmonary embolism suspicion allowing safely reduction of imaging testing: PEPS (Pulmonary Embolism Probability Score)

Abstract Background In pulmonary embolism (PE) suspicion, several strategies based on clinical criteria and D-dimer (Dd) measurement have been developed in order to reduce resource utilization. However, they used different clinical probability (CP) assessment methods limiting their combination. Purpose To develop and validate a unique probability score integrating most of previous proposals to allow safely reduction of imaging testing. Methods 4 CP levels were previously defined in order to obtain a false negative rate <1.9%: 1) without Dd test: very low CP (PE prevalence <1.9%), 2) with Dd <1000 μg/L: low CP (<15%), 3) with Dd <500 or age x10μg/L: moderate CP (<60%) and 4) precluding PE exclusion on Dd: high CP. We used individual data from 4 prospective cohorts of suspected PE patients in Europe and America (n=11 066) for derivation and internal validation. The variables significantly associated with PE in univariate analysis were included in a multivariate logistic regression model. Points were assigned according to the regression coefficients. The score was validated in two external independent cohorts (n=1554, n=1669). Results PEPS comprised 13 variables: age <50 years (−2), age 50–64 years (−1), heart rate <80 beats/min (−1), chronic lung disease (−1), chest pain and recent dyspnea (+1), syncope (+1), male sex (+1), previous venous thromboembolism (+2), medical or orthopaedic immobilization (+2), estrogenic treatment (+2), oxygen saturation <95% (+3), unilateral lower limb pain (+3) and PE is the most likely diagnosis (+ 5). The rates of false negative and avoidable imaging tests if the PEPS strategy would have been applied were 0.6% [95% CI: 0.3–1.1] and 22.7% [20.2–25.3] in the first external validation cohort, and 0.85 [0.5–1.45] and 26.6% [23.5–29.9] in the second one. Applied retrospectively, PEPS strategy compared favourably with other strategies and combinations. Derivation Int. validation Ext. validation 1 Ext. validation 2 nPE/N % [95% CI] nPE/N % [95% CI] nPE/N % [95% CI] nPE/N % [95% CI] TOTAL 615/5588: 11.0% [10.2–11.9] 432/3726: 11.6% [10.6–12.7] 327/1546: 21.2% [19.2–23.2] 196/1669: 11.7% [10.3–13.4] Very low CP PEPS<0 16/1445: 1.1% [0.7–1.8] 16/946: 1.7% [1.0–2.7] 3/118: 2.5% [0.7–6.8] 5/347: 1.4% [0.6–3.3] Low CP 0≤PEPS<5 127/2620: 4.9% [4.1–5.7] 106/1805: 5.9% [4.9–7.1] 49/611: 8.0% [6.1–10.4] 61/647: 7.2% [5.7–9.1] Moderate CP 5≤PEPS<12 347/1334: 26.0% [23.7–28.4] 243/867: 28.0% [25.1–31.1] 206/715: 28.8% [25.6–32.2] 107/430: 24.9% [21.0–29.2] High CP 12≤PEPS 125/179:69.8% [62.8–76.1] 67/108: 62.0% [52.6–70.6] 69/102: 67.7% [58.1–76.2] 23/45: 51.1% [37.0–65.0] AUC 0.84 [0.83–0.86] 0.82 [0.80–0.84] 0.79 [0.76–0.82] 0.77 [0.74–0.80] CP: Clinical probability; PEPS: Pulmonary Embolism Probability Score. Conclusions A strategy based on the proposed score may lead to a safely substantial reduction of imaging testing. It should now be tested in an outcome interventional study.