Statistics from Altmetric.com
Clinical variables were derived and validated in two independent cohorts of patients suspected of having a pulmonary embolism (PE). There were 965 patients in the derivation group and 756 in the validation group.
Univariate analysis identified 10 variables statistically associated with PE, eight of which remained significant in a multivariate analysis. Each variable was assigned points based on the regression coefficient: age ⩾65 years (1 point), previous deep venous thrombosis or PE (3 points), surgery or fracture within 1 month (2 points), active malignancy (2 points), unilateral lower limb pain (3 points), haemoptysis (2 points), heart rate 75–94/min (3 points) or ⩾95/min (5 points), and pain on lower limb venous palpation with unilateral oedema (4 points). Using this revised score in the validation dataset, 8% of low probability patients (0–3 points), 28% of independent probability patients (4–10 points), and 74% of high probability patients (⩾11 points) had PE. This system may therefore assist in identifying patients who can be safely left untreated and those who require invasive investigations.
This study improves on the original Geneva score as it is standardised, entirely clinical (not requiring assessment of arterial blood gases), and has internal and external validity. However, the clinical utility now needs to be addressed in a prospective study.