Potential of an age adjusted D-dimer cut-off value to improve the exclusion of pulmonary embolism in older patients: a retrospective analysis of three large cohorts

OBJECTIVES: In older patients, the the D-dimer test for pulmonary embolism has reduced specificity and is therefore less useful. In this study a new, age dependent cut-off value for the test was devised and its usefulness with older patients assessed. DESIGN: Retrospective multicentre cohort study. SETTING: General and teaching hospitals in Belgium, France, the Netherlands, and Switzerland. Patients 5132 consecutive patients with clinically suspected pulmonary embolism. INTERVENTION: Development of a new D-dimer cut-off point in patients aged >50 years in a derivation set (data from two multicentre cohort studies), based on receiver operating characteristics (ROC) curves. This cut-off value was then validated with two independent validation datasets. MAIN OUTCOME MEASURES: The proportion of patients in the validation cohorts with a negative D-dimer test, the proportion in whom pulmonary embolism could be excluded, and the false negative rates. RESULTS: The new D-dimer cut-off value was defined as (patient's agex10) microg/l in patients aged >50. In 1331 patients in the derivation set with an "unlikely" score from clinical probability assessment, pulmonary embolism could be excluded in 42% with the new cut-off value versus 36% with the old cut-off value (<500 microg/l). In the two validation sets, the increase in the proportion of patients with a D-dimer below the new cut-off value compared with the old value was 5% and 6%. This absolute increase was largest among patients aged >70 years, ranging from 13% to 16% in the three datasets. The failure rates (all ages) were 0.2% (95% CI 0% to 1.0%) in the derivation set and 0.6% (0.3% to 1.3%) and 0.3% (0.1% to 1.1%) in the two validation sets. CONCLUSIONS: The age adjusted D-dimer cut-off point, combined with clinical probability, greatly increased the proportion of older patients in whom pulmonary embolism could be safely excluded.