PESI score calculator (Pulmonary Embolism Severity Index and sPESI)

How this PESI score calculator works

This PESI score calculator adds the patient's age in years to fixed points for 10 other characteristics, as Aujesky and colleagues published the Pulmonary Embolism Severity Index in the American Journal of Respiratory and Critical Care Medicine in 2005. They built it from 15,531 inpatient discharges with pulmonary embolism at 186 Pennsylvania hospitals, using 10,354 to derive the rule and 5,177 to validate it, then tested it again in 221 inpatients from Switzerland and France. The outcome was death from any cause within 30 days. The 11 predictors are age, male sex, cancer, heart failure, chronic lung disease, pulse of 110 beats/min or more, systolic blood pressure below 100 mmHg, respiratory rate of 30 or more, temperature below 36 °C, altered mental status and arterial oxygen saturation below 90%. The simplified PESI of Jiménez and colleagues (Archives of Internal Medicine, 2010) keeps six of them and gives each 1 point. The calculator scores both from the same answers.

Points for each item in the original PESI (Aujesky 2005, Table 2) and the simplified PESI (Jiménez 2010, Table 1):

Aujesky defines altered mental status as disorientation, lethargy, stupor or coma, and counts an oxygen saturation below 90% with or without supplemental oxygen. Both papers treated a missing value as normal, so an unticked box here means the finding is absent or was not recorded.

PESI classes, sPESI results and 30-day mortality

The PESI score places a patient in one of five classes, and each class carries the 30-day mortality that Aujesky and colleagues observed. A total of 65 or less is class I (very low risk), 66 to 85 is class II (low risk), 86 to 105 is class III (intermediate risk), 106 to 125 is class IV (high risk) and more than 125 is class V (very high risk). Across the derivation and both validation samples, 30-day mortality ranged from 0 to 1.6% in class I, 1.7 to 3.5% in class II, 3.2 to 7.1% in class III, 4.0 to 11.4% in class IV and 10.0 to 24.5% in class V. The external sample's rates are lower because it excluded severely ill patients. The simplified PESI has two results: 0 points is low risk and 1 point or more is high risk. For comparison with it, Jiménez's paper groups PESI classes I and II together as low risk and classes III to V as high risk.

30-day mortality by sPESI result (Jiménez 2010, Table 3):

PESI or sPESI: how the two scores compare

PESI and sPESI predicted 30-day mortality equally well when Jiménez and colleagues compared them in 995 patients with acute symptomatic pulmonary embolism at a Madrid hospital: both had an area under the ROC curve of 0.75. The simplified version dropped male sex, respiratory rate, temperature and altered mental status because they were not significantly linked with death in that cohort, merged heart failure and chronic lung disease into one item, and turned age into a single cut-off at 80 years. Its advantage is speed: six yes-or-no items instead of eleven weighted ones and arithmetic that needs no calculator. Its trade-off is detail. The sPESI labels fewer patients as low risk (30.7% against 36.3% for PESI classes I and II in the same cohort), and the authors note it was not built to grade patients into rising levels of risk. In that cohort, sPESI low-risk patients had 1.0% mortality, against 2.5% for PESI classes I and II.

For the step before this one, when PE is suspected but not yet confirmed, see the Wells score for PE, and the guide to PESI vs the Wells score.

When the PESI score is less reliable

The PESI score is less reliable outside the setting it was built for, or when used alone. Aujesky's derivation patients were identified from hospital discharge codes for pulmonary embolism rather than imaging criteria, and the external validation sample was small and left out severely ill patients, so its confidence intervals for classes IV and V are wide. The authors wrote that the rule is intended to supplement, not replace, clinical judgment, and gave the example of a previously healthy 40-year-old woman with severe hypoxemia who scores class I. The sPESI was derived retrospectively in a single Spanish hospital and validated retrospectively in the RIETE registry. Neither score includes imaging of the right ventricle or cardiac biomarkers. The 2019 ESC guideline states that a PESI of class I–II or an sPESI of 0 is a reliable predictor of low-risk PE, and that patients with right ventricular dysfunction or raised biomarkers despite such a score are classed as intermediate risk. Both scores apply only after PE is confirmed.