Bedside Diagnostic Tests for Pulmonary Embolism—Reply

In Reply: Mr Schuur hypothesizes that PE screening with the D-dimer assay or ADSM will not reduce the incidence of PVI because of increased use in low-risk patients. Our data support his assertion that the test will probably perform differently in this subgroup. Compared with the overall specificity of 51.6%, the specificity of the D-dimer assay or ADSM was 78% in the low-risk subgroup (assuming a 22% false-positive rate on screening), suggesting that the D-dimer assay or ADSM would be positive in 33.3% of low-risk patients (22.0% combined with the true positive rate of 11.3%). If clinicians increased their screening rate in low-risk patients, the overall screening rate would have to double to produce a net increase in PVI. Even if this unlikely hypothetical outcome occurred, the positive LR (2.03) applied to the pretest probability of PE in low-risk patients suggests that the rate of "positive" findings on PVI would increase to 20.5% compared with the 14.7%, which we observed in our study without screening.¹ The net effect of increased screening would be to correctly diagnose more low-risk patients with PE that might otherwise remain undetected.