Current issues of ACP Journal Club are published in Annals of Internal Medicine


D-dimer assay was useful for ruling out suspected pulmonary embolism

ACP J Club. 1999 May-June;130:75. doi:10.7326/ACPJC-1999-130-3-075

Source Citation

Ginsberg JS, Wells PS, Kearon C, et al. Sensitivity and specificity of a rapid whole-blood assay for D-dimer in the diagnosis of pulmonary embolism. Ann Intern Med. 1998 Dec 15; 129:1006-11.



How accurate is a D-dimer assay in patients with suspected pulmonary embolism (PE)?


A blinded comparison of D-dimer assay results, ventilation-perfusion (V/Q) lung scanning, and bilateral compression ultrasonography (CU) with watchful waiting (3 mo).


3 Canadian tertiary care hospitals.


1177 consecutive adults (mean age 53.4 y, 59% women) who were referred for suspected PE. Exclusion criteria were suspected upper-extremity deep venous thrombosis, no symptoms within the previous 48 hours, receipt of anticoagulants for 72 hours, limited life expectancy, or contraindication to contrast media.

Description of tests and diagnostic standard

History and physical data were used to classify patients as having a high, moderate, or low pretest probability for PE. V/Q lung scans and bilateral CU from the common femoral vein to the calf trifurcation were done within 24 hours. Patients with nondiagnostic V/Q scans or high-probability V/Q scans and a low pretest probability had further testing done using a complex algorithm. A D-dimer assay was done concurrently with initial testing. All patients were followed for 3 months.

Main outcome measures

Diagnostic test properties for D-dimer assay in patients with various combinations of pretest probabilities, V/Q lung scans, and bilateral CU results.

Main results

17% of patients had PE. For all patients, D-dimer assay results had a sensitivity of 84.8% and a specificity of 68.4%. The sensitivity, specificity, and likelihood ratios for patients with low, moderate, and high pretest probabilities and normal, nondiagnostic, and high-probability V/Q scans are shown in the Table.


A normal D-dimer assay result was useful for ruling out PE in patients with a low pretest probability of PE or a nondiagnostic V/Q lung scan.

Source of funding: Medical Research Council of Canada.

For correspondence: Dr. J.S. Ginsberg, McMaster University Medical Centre, 1200 Main Street West, Room 3W15, Hamilton, Ontario L8N 3Z5, Canada. FAX 905-521-6068.

Table. Test properties of D-dimer assay for pulmonary embolism (PE)*

Variable PE rate Sensitivity (95% CI) Specificity (CI) +LR (CI) -LR (CI)
Pretest probability
Low 3.4% 79% (58 to 93) 76% (73 to 79) 3.3 (2.6 to 4.2) 0.27 (0.13 to 0.60)
Moderate 26.4% 80% (71 to 87) 52% (46 to 57) 1.7 (1.4 to 1.9) 0.38 (0.26 to 0.58)
High 78.3% 93% (85 to 98) 45% (23 to 68) 1.7 (1.1 to 2.5) 0.15 (0.06 to 0.41)
Lung scan
Normal 1.3% 75% (19 to 99) 77% (72 to 81) 3.2 (1.8 to 4.2) 0.33 (0.06 to 0.91)
Nondiagnostic 7.4% 77% (63 to 87) 65% (61 to 69) 2.2 (1.8 to 2.6) 0.36 (0.22 to 0.59)
High probability 88.1% 88% (81 to 93) 47% (24 to 71) 1.7 (0.1 to 2.5) 0.26 (0.13 to 0.49)

*Abbreviations defined in Glossary; sensitivity, specificity, and CI calculated from data in article.


The diagnosis of acute PE remains troublesome for physicians because the signs and symptoms of this condition are not specific (i.e., many signs and symptoms that suggest PE occur in the absence of PE). Tests that allow confident exclusion of acute PE are very useful. When a physician suspects acute PE, a normal V/Q scan allows anticoagulants to be withheld with a very low likelihood of symptomatic thromboembolism during 3 months of follow-up (1). Unfortunately, more than half of patients with suspected PE have abnormal, but nondiagnostic, lung scans. Thus, a test that allows a confident decision to be made about withholding anticoagulation therapy when lung scans are not diagnostic would be enormously helpful. Negative serial examinations of the legs for deep venous thrombosis permit anticoagulants to be withheld safely in this situation (2), but these tests are expensive and burdensome. It is in this context that Ginsberg and colleagues studied a rapid whole-blood assay for D-dimer.

Their study provides new evidence that a negative test result for D-dimer allows the physician to avoid further tests and anti-coagulation when the clinical pretest probability is low and the lung scan indicates a low or intermediate probability of acute PE. Additional tests are necessary when the clinical probability is high, even if the D-dimer result is negative and lung scans suggest a low probability of PE.

Practitioners must recognize that the bedside D-dimer test is not a stand-alone screening test for acute PE. They must also recognize that many D-dimer tests exist and that most are insensitive screening tests for PE. Finally, practitioners should look for the results of additional clinical management studies, particularly in their own or similar patient populations, before incorporating the D-dimer assay into their algorithm for the diagnosis of acute PE.

C. Gregory Elliott, MD
LDS Hospital and the University of UtahSalt Lake City, Utah, USA


1. Hull RD, Raskob GE, Coates G, Panju AA. Chest. 1990;97:23-6.

2. Hull RD, Raskob GE, Ginsberg JS, et al. Arch Intern Med. 1994;154:289-97.