Role of an age-adjusted D-dimer cutoff level in the diagnosis algorithm of lower limb deep venous thrombosis in...

Our goal was to analyze the utility of the age-adjusted D-dimer cutoff value in patients with clinically suspected deep venous thrombosis (DVT) in an ambulatory care setting, including distal DVTs.

This was an observational cohort study of 606 outpatients older than 18 years presenting with low or moderate clinical suspicion of lower limb DVT (measured by Wells scale). D-dimer levels were obtained, and duplex ultrasound was performed (including femoropopliteal and below-knee veins). We calculated sensitivity, specificity, and positive and negative predictive D-dimer values and when to apply the age-adjusted D-dimer cutoff value (D-dimer threshold = age × 10 μg/L). We split patients older than 50 years into 10-year age groups. We constructed receiver operating characteristic curves of the D-dimer test for each group to find the best threshold (defined as the value of D-dimer that gives more specificity, maintaining the maximum possible sensitivity).

There were 249 men and 357 women with a mean age of 69.3 years; 41 patients were diagnosed with DVT. At a D-dimer threshold of 250 μg/L, sensitivity was 93%, specificity was 8%, positive predictive value was 7%, and negative predictive value was 94%. When the age-adjusted cutoff level was applied, global sensitivity was 76% and specificity 61%; positive predictive value was 12%, and negative predictive value was 97%. False-negative rate was 24%. We split patients older than 50 years into 10-year age groups: 50 to 60 years, 60 to 70 years, 70 to 80 years, and >80 years. The optimum thresholds were, respectively, 526 μg/L, 442.5 μg/L, 475 μg/L, and 549. μg/L.

In our series, the age-adjusted D-dimer cutoff level is not useful in the diagnostic algorithm of DVT.