Refining risk in normotensive acute pulmonary embolism

Introduction: Normotensive acute pulmonary embolism (aPE) has a wide spectrum of outcomes. The best method to identify patients at higher-risk remains unclear. Aims and Objectives: 1) develop a unique prognostic model for adverse outcomes in normotensive aPE, 2) validate the Bova score in a North American population. Methods: This was a multi-centre retrospective cohort of all aPE admitted from emergency departments in Calgary, Canada between 2012-2017. Logistic regression models with bootstrapping for internal validation were used to predict the composite primary outcome of in-hospital death or hemodynamic decompensation. Results: 2067 patients with normotensive aPE were assessed. A primary outcome occurred in 32 patients (1.5%). Stratified by simplified pulmonary embolism severity index (sPESI), 21.2% were low-risk (0% event rate) and 78.8% were high-risk (2.0% event rate). The multivariable model in sPESI high-risk patients (n=1179) retained high-sensitivity troponin ≥50 pg/ml, CT right-left ventricular diameter ratio ≥1.5, systolic blood pressure <100 mmHg, central pulmonary artery clot, & heart rate ≥100 beats per minute (c-index 0.88, 95% CI 0.82-0.93). Three risk groups were derived from the model using a weighted score (score, prevalence, event rate): group 1 (0-3, 73.8%, 0.34%), group 2 (4-6, 17.6%, 5.8%), group 3 (7-9, 8.65%, 12.8%) (c-index 0.85, CI 0.78-0.91). The prevalence (event rate) by Bova risk categories (n=1482) were: stage 1 55.7% (0.1%), stage 2 29.6% (2.3%) and stage 3 14.6% (7.8%) (c-index 0.80, CI 0.74-0.86). Conclusions: Our novel risk score discriminated normotensive aPE patients at high risk of in-hospital adverse events better than the Bova score. Further validation of our score is warranted.