Age-adjusted D-dimer for DVT with the Wells score

How the Wells score and D-dimer work together

The Wells DVT score and the D-dimer test work as two steps: the score sets the pretest probability, and the D-dimer is used to rule DVT out only when that probability is not high. In the 2003 New England Journal of Medicine trial by Wells and colleagues, outpatients with a score under 2, called DVT unlikely, and a negative D-dimer had no ultrasound, and only 0.4% of those in whom DVT had been ruled out had a confirmed venous thromboembolic event in 3 months. Patients scoring 2 or more, DVT likely, went to ultrasound whatever the D-dimer showed. The American Society of Hematology (ASH) 2018 guideline on diagnosing venous thromboembolism takes the same approach: for patients at low or unlikely pretest probability, D-dimer as the first test reduces the need for imaging, while patients at high or likely probability need imaging. The age-adjusted cut-off changes only the D-dimer step; the Wells step stays the same.

Why a fixed D-dimer cut-off works poorly with age

D-dimer levels rise with age even without a clot, so the conventional cut-off of 500 µg/L calls more and more older people positive, and the test rules out fewer of them. Schouten and colleagues measured this in a 2013 BMJ meta-analysis of 13 cohorts with 12,497 patients over 50 with a non-high clinical probability of venous thromboembolism. With the fixed 500 µg/L cut-off, specificity fell from 57.6% at ages 51 to 60, to 39.4% at 61 to 70, 24.5% at 71 to 80 and 14.7% over 80. In the oldest group, then, fewer than 1 in 6 people without a clot had a negative test, and almost everyone went on to imaging. With the age-adjusted cut-off, specificity was higher in every age band: 62.3%, 49.5%, 44.2% and 35.2%. Sensitivity of the age-adjusted cut-off stayed above 97% in all age groups, so the gain in specificity did not come at the cost of missing more venous thromboembolism.

The age-adjusted D-dimer formula

The age-adjusted D-dimer cut-off is the patient's age multiplied by 10, in µg/L, for patients over 50; at 50 or younger the conventional 500 µg/L cut-off is used, and at exactly 50 the two are the same. Douma and colleagues derived it in the BMJ in 2010 from 5,132 patients with suspected pulmonary embolism in four European countries, choosing it from receiver operating characteristic curves and testing it in two validation cohorts. The formula assumes an assay whose conventional cut-off is 500 µg/L in fibrinogen equivalent units (FEU), the setting in which it was studied. Units are a common source of error: a value of 500 µg/L is the same as 500 ng/mL and 0.5 mg/L, so a 75-year-old's cut-off is 750 µg/L, 750 ng/mL or 0.75 mg/L. Some laboratories report D-dimer units (DDU) instead, with a different cut-off, so check which unit and cut-off your laboratory uses before applying the formula.

What ADJUST-DVT found

ADJUST-DVT, published by Le Gal and colleagues in JAMA in 2026, was the first prospective management study of the age-adjusted cut-off in suspected leg DVT. It enrolled 3,205 emergency department outpatients at 27 centers in Belgium, Canada, France and Switzerland from 2015 to 2022, assessed with the Wells score, a highly sensitive D-dimer and compression ultrasound. DVT prevalence was 14%. Among the 2,169 patients with a non-high or unlikely clinical probability, 531 (24.5%) had a D-dimer under 500 µg/L, and 161 more (7.4%) had a D-dimer between 500 µg/L and their age-adjusted cut-off. In those 161 patients DVT was ruled out on the D-dimer result, and none had a venous thromboembolic event in 3 months of follow-up (0%, 95% CI 0% to 2.3%). In patients aged 75 or over, the share with a negative D-dimer rose from 8.7% with the fixed cut-off to 26.1% with the age-adjusted one, without any false-negative results.

The earlier ADJUST-PE study by Righini and colleagues found the same pattern in suspected pulmonary embolism: a 3-month failure rate of 0.3% in patients between the two cut-offs, and in those aged 75 or over, D-dimer excluded PE in 29.7% instead of 6.4%.

Worked example

Take a 78-year-old outpatient with a swollen calf. She has calf swelling of more than 3 cm compared with the other leg and pitting oedema, but another diagnosis, a ruptured Baker's cyst, is judged as likely as DVT. Her 2003 Wells score is 1 + 1 − 2 = 0, DVT unlikely. Her D-dimer is 650 µg/L FEU. With the conventional 500 µg/L cut-off it is positive, and she would go on to ultrasound. Her age-adjusted cut-off is 78 × 10 = 780 µg/L, so the same result is negative, and in the ADJUST-DVT pathway DVT would be ruled out without imaging. Had her score been 2 or more, DVT likely, she would have gone to ultrasound whatever the D-dimer, because neither cut-off is used to rule out DVT at a likely pretest probability. And a D-dimer of 820 µg/L would be positive on both cut-offs. The adjustment matters only for the band of results between 500 µg/L and age × 10.

Score the leg first in the Wells DVT calculator, then apply the D-dimer cut-off that matches the patient's age.