Scott M. Stevens, MD
In patients with suspected acute pulmonary embolism (PE), what is the accuracy of gestalt and clinical decision rules (CDRs) for detecting PE?
Included studies evaluated gestalt or CDRs for estimating probability of PE in patients ≥ 16 years of age with suspected acute PE, enrolled patients consecutively, blinded assessors to results of D-dimer testing or pulmonary vascular imaging, used an appropriate reference standard to confirm diagnosis of PE (or deep venous thrombosis as a surrogate for PE), had ≥ 45 days of follow-up for patients with negative test results and no pulmonary imaging, included > 50 patients with confirmed PE if a decision rule was being derived, and provided sufficient data to create 2 × 2 tables. Gestalt estimates were unstructured and based on patient history, physical examination with or without basic laboratory tests, chest radiographs, or electrocardiographs. Decision rule estimates were structured and based on multivariate logistic regression models. Outcomes were pooled sensitivity, specificity, and likelihood ratios (calculated from data in article).
MEDLINE and EMBASE/Excerpta Medica (both to Jun 2011), and reference lists were searched for original, prospective studies published in English, French, German, Italian, Spanish, or Dutch. Investigators were contacted. 52 studies (n = 55 268, mean or median age 45 y to 72 y, 47% to 74% women) met the selection criteria: 19 evaluated gestalt; 26 the Wells or simplified Wells rule; 12 the Geneva, revised Geneva, or simplified revised Geneva rule; 4 the Pisa or revised Pisa rule; 3 the Charlotte rule; and 3 the Pulmonary Embolism Rule-out Criteria. Data for rules evaluated in ≥ 4 studies were pooled using random-effects bivariate analysis that simultaneously modeled sensitivity and specificity pairs. 20 studies that combined gestalt or CDRs with D-dimer testing are not reported here.
Prevalence of PE ranged from 4% to 44%. Pooled sensitivities and specificities are shown in the Table. Studies were heterogeneous for sensitivity and specificity (data not reported). Wells rule with a cutpoint ≤ 4 differed from other diagnostic tools for sensitivity and specificity (P < 0.01), and Wells rule with a cutpoint < 2 had higher specificity than the revised Geneva rule (P = 0.026).
In patients with suspected acute pulmonary embolism, gestalt and clinical decision rules have limited sensitivity and specificity for detecting pulmonary embolism.
*Diagnostic terms defined in Glossary. Only rules evaluated in ≥ 4 studies are reported here.
†LRs calculated on the basis of pooled sensitivity and specificity.
Stevens SM. Review: Gestalt or clinical decision rules have limited sensitivity and specificity for detecting acute PE. Ann Intern Med. ;156:JC1–11. doi: 10.7326/0003-4819-156-2-201201170-02011
Download citation file:
Published: Ann Intern Med. 2012;156(2):JC1-11.
Emergency Medicine, Pulmonary Embolism, Pulmonary/Critical Care, Venous Thromboembolism.
Results provided by:
Copyright © 2018 American College of Physicians. All Rights Reserved.
Print ISSN: 0003-4819 | Online ISSN: 1539-3704
Conditions of Use