To evaluate the utility of a modified (i.e., without the variable “Age >80 years”) simplified Pulmonary Embolism Severity Index (sPESI) in elderly patients with acute symptomatic pulmonary embolism (PE), and to derive and validate a refined version of the sPESI for identification of elderly patients at low risk of adverse events.
The study included normotensive patients aged >80 years with acute PE enrolled in the RIETE registry. We used multivariable logistic regression analysis to create a new risk score to predict 30-day all-cause mortality. We externally validated the new risk score in elderly patients from the COMMAND VTE registry.
Multivariable logistic regression identified four predictors for mortality: high-risk sPESI, immobilization, coexisting deep vein thrombosis (DVT), and plasma creatinine >2 mg/dL. In the RIETE derivation cohort, the new model classified fewer patients as low risk (4.0% [401/10,106]) compared to the modified sPESI (35% [3522/10,106]). Low-risk patients based on the new model had a lower 30-day mortality than those based on the modified sPESI (1.2% [95% CI, 0.4-2.9%] versus 4.7% [95% CI, 4.0-5.4%]). In the COMMAND VTE validation cohort, 1.5% (3/206) of patients were classified as having low risk of death according to the new model, and the overall 30-day mortality of this group was 0% (95% CI, 0-71%), compared to 5.9% (95% CI, 3.1-10.1%) in the high-risk group.
For predicting short-term mortality among elderly patients with acute PE, this study suggests that the new model has a substantially higher sensitivity than the modified sPESI. A minority of these patients might benefit from safe outpatient therapy of their disease.