Seven studies were included in the review (n=8,371): one RCT (n=1,693) and six prospective cohort management studies (n=6,678). Sample size ranged from 331 to 3,306 patients. Study quality was reported to be similar across included studies (detailed results not provided). Heterogeneity was reported to be low.
Diagnostic safety:
Three-month thromboembolic risk in patients untreated after a combined negative VIDAS D-dimer with a low/intermediate or unlikely pre-test probability of pulmonary embolism was three out of 2,166 patients (0.14%, 95% CI 0.05 to 0.41%). Sensitivity analysis increased this risk to 0.64% (95% CI 0.38 to 1.08%).
Diagnostic accuracy:
Overall results for combined non-high pre-test probability and negative D-dimer were: sensitivity 99.7% (95% CI 99.0% to 99.9%); specificity 47.4% (95% CI 46.0% to 48.9%); negative predictive value 99.9% (95%CI 99.6% to 100%); and negative likelihood ratio 0.01 (95% CI 0.00 to 0.02).
Sensitivity results were high for all combinations using any pre-test probability model (range 98.7% to 100%). Specificity ranged from 40.8% (low/intermediate pre-test probability) to 57.4% (unlikely pre-test probability).
In an analysis of six studies, the number of patients needed to be tested by VIDAS D-dimer to rule out one pulmonary embolism with no further investigations was 2.8 (95% CI 2.7 to 2.9).