Douma et al compared the Wells score, Geneva score, and the simplified versions of each, combined with D-dimer, on a prospective cohort of 807 patients with suspected PE (~23% of whom were found to have PE). With a negative D-dimer and a low-probability score using any rule, risk of pulmonary embolism was ~0.5%. However, only ~23% of patients fit this category. CT-angiogram was only done in moderate/high risk patients. Three-month follow-up was the gold standard for patients at low-risk by rules/D-dimer.

Notably, 7 "low-risk" patients were scanned in violation of protocol, and one (14%) had a PE -- this lone patient generated the "0.5%" risk. Which begs the question, how many more patients in the low-risk group would have had PEs discovered on CT-angiogram, had they gotten one? Ann Intern Med 2011;154:709-718. FREE FULL TEXT