Original Contributions
Criteria for the safe use of D -dimer testing in emergency department patients with suspected pulmonary embolism: A multicenter us study*,**,*,**

https://doi.org/10.1067/mem.2002.121398Get rights and content

Abstract

Study Objective: We derive a decision rule to partition emergency department patients with suspected pulmonary embolism (PE) into a small, high-risk group (>40% pretest probability) that is unsafe for D -dimer testing and a larger group that is safe to have PE ruled out with either a whole-blood D -dimer plus alveolar deadspace measurement or a quantitative D -dimer assay. Methods: Nine hundred thirty-four patients with suspected PE were studied at 7 urban EDs in the United States. Patients were prospectively interviewed and examined for recognized symptoms, signs, and risk factors associated with PE. These data were collected before standard objective imaging for PE. Selected variables were analyzed by multivariate logistic analysis to determine factors associated with PE (P <.05). A decision rule was then constructed to categorize approximately 80% of ED patients as safe for D -dimer testing. Results: Pretest prevalence of PE was 19.4% (181/934; 95% confidence interval [CI] 16.3% to 21.7%). Six variables found to be significant on multivariate analysis were used to construct the decision rule. Unsafe patients had either a shock index (heart rate/systolic blood pressure) more than 1.0 or age older than 50 years, together with any one of the following conditions: unexplained hypoxemia (SaO2 <95%; no prior lung disease), unilateral leg swelling, recent major surgery, or hemoptysis. These criteria were met by 197 (21.0%) of 934 patients, and 83 of 197 (42.1%; 95% CI 35.3% to 49.6%) patients had PE. Exclusion of these 197 unsafe patients significantly decreased the probability of PE in the remaining 737 (79.0%) safe patients to 13.3% (95% CI 10.9% to 15.9%). Conclusion: Simple clinical criteria can permit safe D -dimer testing in the majority of ED patients with suspected PE. These criteria warrant prospective validation. [Kline JA, Nelson RD, Jackson RE, Courtney DM. Criteria for the safe use of D -dimer testing in emergency department patients with suspected pulmonary embolism: a multicenter US study. Ann Emerg Med. February 2002;39:144-152.]

Introduction

A growing body of evidence indicates that a normal D -dimer result can help rule out the diagnosis of pulmonary embolism (PE). Among the commercially available rapid D -dimer assays, the lowest likelihood ratio negative (LR–) can be generated by either a normal quantitative D -dimer assay (<500 ng/L)1, 2, 3, 4 or the combination of a normal whole-blood D -dimer assay together with a normal index of pulmonary gas exchange.5, 6, 7 Aggregated data from published reports suggest that the best-case LR– for the rapid quantitative D -dimer assays (rapid enzyme-linked immunoassays or turbidimetric assays) is approximately 0.03 to 0.07.4 Likewise, in a multicenter study, the combination of a normal whole-blood D -dimer plus a normal alveolar deadspace measurement generated an LR– of 0.03.8 However, authors have cautioned that no D -dimer–based strategy can safely rule out PE in the entire spectrum of patients with suspected PE and that clinical criteria should be used to exclude patients with excessive risk.3, 9, 10 To define excessive risk, we have previously argued that the posttest probability for PE should not exceed 2.0% (and preferably should not exceed 1.0% overall).11 The 2.0% value is based in part on data published by Gosselin et al,12 indicating that, among 785 patients without symptoms of PE who underwent contrast-enhanced chest computed tomographic (CT) scanning for other reasons, 15/785 (1.9%) had CT scans read by board-certified radiologists as positive for PE. A screening test with an LR– of 0.03 will not reduce the probability of PE to below 2.0% if the pretest prevalence of PE exceeds 40%.13

Accordingly, we sought to derive a set of clinical criteria that would define an ED subgroup with a pretest probability of PE greater than 40% and therefore would have too high of a risk to have the diagnosis of PE excluded on the basis of any contemporary D -dimer–based screening strategy. We analyzed data that were prospectively collected on patients with suspected PE at 7 large urban emergency departments in the United States. The primary hypothesis we tested was that a set of clinical criteria, which can be obtained immediately at the bedside in the ED, could be used in a decision rule to identify a subgroup of patients with a probability of PE above 40% (high risk) while significantly reducing the probability of PE in the remaining patients who are not at high risk.

Section snippets

Materials and methods

All methods were approved by the institutional review boards of all participating hospitals. Written informed consent was obtained from all study participants. Data were collected from ED patients at 7 urban teaching hospitals: Carolinas Medical Center, Charlotte, NC (1996-2000); William Beaumont Hospital, Royal Oak, MI (1998-1999); Detroit Medical Center and Grace Hospital, Detroit, MI (1998-1999); Barnes-Jewish Hospital, St. Louis, MO (1998-1999); St. Vincent Mercy Medical Center, Toledo, OH

Results

Data were collected from 934 patients at 7 EDs in the United States. These 934 patients constituted the entire sample of patients enrolled during the study period. Once the clinical data were collected, patients were excluded only if the criterion standard was not met, as in the cases in which physicians in charge of patient care abandoned the effort to diagnose or exclude PE altogether. PE was diagnosed in 181 patients (19.4%; 95% CI 16.9% to 22.0%). The criterion standard for the diagnosis

Discussion

This study tested the hypothesis that a set of clinical criteria could be developed to define the ED patient with a greater than 40% probability of PE based on data collected prospectively from EDs at 7 hospitals in the United States. We determined that a patient who is either older than 50 years or who has a shock index (HR/SBP) of more than 1.0, together with any 1 of 4 risk factors (unexplained hypoxemia, unilateral leg swelling, recent history of surgery, or hemoptysis) would have a 42.1%

Acknowledgements

Author contributions: JAK, REJ, and DMC conceived the study. JAK and REJ obtained study funding and supervised data collection including quality control. RDN and DMC contributed substantially to data collection, data entry, and preliminary data analysis. JAK performed statistical analysis in consultation with REJ. JAK drafted the manuscript, and all authors contributed substantially to its revision. JAK takes responsibility for the paper as a whole.

We thank the following individuals for their

References (32)

  • G Reber et al.

    Performances of a new, rapid, and automated microlatex D -dimer assay for the exclusion of pulmonary embolism in symptomatic outpatients

    Thromb Haemost

    (1998)
  • JA Kline et al.

    Use of the alveolar dead space fraction (Vd/Vt) and plasma D -dimers to exclude acute pulmonary embolism in ambulatory patients

    Acad Emerg Med

    (1997)
  • P Egermayer et al.

    Usefulness of D -dimer, blood gas, and respiratory rate measurements for excluding pulmonary embolism

    Thorax

    (1998)
  • JA Kline et al.

    Diagnostic accuracy of a bedside D -dimer assay and alveolar dead-space measurement for rapid exclusion of pulmonary embolism: a multicenter study

    JAMA

    (2001)
  • MR De Groot et al.

    The use of a rapid D -dimer blood test in the diagnostic work-up for pulmonary embolism: a management study

    Thromb Haemost

    (1999)
  • JA Kline et al.

    Bedside diagnostic tests for pulmonary embolism

    JAMA

    (2001)
  • Cited by (129)

    View all citing articles on Scopus
    *

    Dr. Courtney is now affiliated with the Division of Emergency Medicine, Northwestern Memorial Hospital, Chicago, IL.

    **

    Author contributions are provided at the end of this article.

    *

    Supported in part by the Established Investigator Award from the Emergency Medicine Foundation.

    **

    Address for reprints: Jeffrey A. Kline, MD, Emergency Medicine Clinical Research, PO Box 32861, Charlotte, NC 28323-2861;,704-355-7092, fax 704-355-7047.

    View full text