Tutorial 2a

Article: Sensitivity and Specificity of a Rapid Whole-Blood Assay for D-Dimer in the Diagnosis of Pulmonary Embolism

Authors: Ginsberg JS, Wells PS, Kearon C, et al

Publication: Ann Intern Med. 1998;129:1006-1011

Study: Prospective cohort of 1177 consecutive patients with suspected PE underwent whole-blood D-dimer assay compared with diagnostic strategy for diagnosis of PE

Inclusion Criteria: consecutive patients 18 years or older with clinically suspected PE

Exclusion Criteria: suspected upper extremity DVT; no symptoms within 48 hours of presentation; treatment with anticoagulants for greater than 72 hours; life expectancy < 3 months; contraindication to contrast dye; geographic inaccessibility

Step 1: Are the results valid?

Step 1a: Was there an independent, blind comparison to a reference standard?

Yes, but did not use pulmonary angiography as reference standard and instead used a diagnostic strategy in which all patients had a VQ scan and US. Patients at highest risk underwent venography and pulmonary angiography if venography was negative. All patients with negative test results were followed for 3 months.

Step 1b: Did patient sample include appropriate spectrum of patients?

Yes, all patients suspected of having PE were included regardless of low, moderate, or high pre-test probability.

Step 1c: Did results influence the decision to perform reference standard?

No

Step 1d: Were methods of performing the test described in sufficient detail?

Yes

Step 2: What are the results?

 D-dimer result  PE present  PE absent  LR (95% CI)
 positive  167  310  2.7 ( 2.3-3.0)
 negative  30  670  0.22 (0.16-0.31)
 total  197  980  

 Patient Subgroup  PE present  PE present  PE absent  PE absent

 LR

(95% CI)

 LR

(95% CI)

   + D-dimer  - D-dimer  + D-dimer  - D-dimer  + D-dimer  - D-dimer
 low pretest probability  19  5  163  516

 3.3

(2.6-4.2)

 0.27

(0.13-0.60)

 moderate pretest probability  81  20  136  145

 1.7

(1.4-1.9)

 0.38

(0.26-0.58)

 high pretest probability  67  5  11  9

 1.7

(1.1-2.5)

 0.15

(0.06-0.41)

Step 3: Will the results help me in caring for my patients?

Limitations

1. used complex diagnostic strategy instead of pulmonary angiography as "gold standard"

2. valid for whole-blood D-dimer by simpliRED assay with results reported as normal or abnormal

Strength

1. included data on pre-test probability

Conclusion

In patients with suspected PE, a normal D-dimer is useful for excluding the diagnosis in those with a low pre-test probability. A normal D-dimer does not exclude PE in pateints with moderate to high pretest probability