Design: Systematic review and meta-analysis of 6 prospective management studies
Population: 12,194 patients with suspected PE managed with adaptive D-dimer thresholds
Primary outcome: 3-month diagnostic failure (VTE or PE-related/unexplained death)
Effect: Overall 0.12% (95% CI 0.06–0.26%); adaptive window 0.50% (0.20–1.11%)

Age-adjusted and clinical probability-adjusted D-dimer thresholds let more patients avoid CT pulmonary angiography, but doubts remained because much of the safety data came from cohorts managed with the fixed 500 ng/mL cut-off. This meta-analysis included only prospective management studies in which an adaptive threshold actually guided imaging: six studies, 12,194 patients, PE prevalence 7–19%.

Among all patients in whom PE was excluded without imaging, the 3-month failure rate (venous thromboembolism or PE-related or unexplained death) was 0.12% (95% CI 0.06–0.26%; I² 0%). In the key group — D-dimer above 500 but below the adaptive threshold — failure was 0.50% (0.20–1.11%). About 61% of patients avoided imaging across algorithms.

A failure rate below about 2% is the conventional safety benchmark for ruling out PE. Both figures are well below it, with no heterogeneity. This supports using adaptive thresholds routinely in patients without high clinical probability — which, where CT capacity and contrast safety are concerns, spares many patients unnecessary scans.

Use an age-adjusted or probability-adjusted D-dimer threshold for patients with non-high PE probability.
Patients with a D-dimer between 500 and their adjusted threshold can safely forgo CTPA — failure was about 0.5%.
Always assess clinical probability first; the strategy does not apply to high-probability patients or pregnancy-specific pathways.
Provide safety-netting advice to return if symptoms worsen.
Expect to spare imaging in more than half of suspected PE.

Why it matters

It removes the main safety doubt about adaptive thresholds, supporting fewer CT scans.

The statistics, in plain English

A pooled failure rate of 0.50% with an upper limit of 1.11% means at worst about 1 in 90 patients in the adaptive window were missed — still under the usual 2% safety benchmark. I² of 0% means the studies agreed.