Age‐adjusted D‐dimer cut‐off levels to rule out venous thromboembolism in patients with non‐high pre‐test probability: Clinical performance and cost‐effectiveness analysis

Background D-dimer levels below a well-defined cut-off level enable to safely rule out VTE in patients with a low or intermediate pre-test probability (PTP), as the test negative predictive value (NPV) was high. As ageing is associated with increased D-dimer levels, the question was raised of the usefulness of their measurement to rule out VTE in elderly patients. Various attempts were made in the recent years to address that issue, particularly the use of age-adjusted cut-off values calculated by multiplying the patient's age by 10, in patients aged over 50. Aim of that study The aim of the study was to validate such a strategy, and to evaluate its economic impact in unselected outpatients referred to the emergency departments of 5 French centres (2 university hospitals, and 3 general hospitals). Patients, Materials and Methods After exclusion of those on anticoagulant treatment at the time of the diagnosis (n=23), a total of 1255 consecutive patients with a non-high PTP of VTE were included in the study. The same standardized procedure was used in the 5 centres, i.e. D-dimer measurement in patients with a non-high PTP, and imaging techniques (usually computed pulmonary angiography in case of suspected PE and Doppler ultrasonography in case of suspected DVT) only when D-dimer was above the cut-off level. D-dimer was evaluated using the same latex-based fully automated assay (HemosIL D-dimer HS500 assay, Instrumentation Laboratory). Test results were expressed in ng/mL fibrinogen equivalent units (FEU), and the traditional cut-off level for VTE exclusion was 500 ng/mL. Results VTE was diagnosed in 105 patients of the 1255 included patients (8.4%): 88 patients were diagnosed out of the 1082 patients referred for suspected PE and 27 out the 173 referred for suspected DVT. D-dimer levels were above 500 ng/mL in all patients with VTE and in 521 of the 1150 patients without VTE (45.3%). The test NPV was 100% as well as its sensitivity. If the overall test specificity was 54.7%, it significantly decreased in an age-dependent manner over 70 years, related to a high percentage of increased D-dimer levels in elderly patients, particularly in those above 80 years. Using age-adjusted cut-off levels, calculated by multiplying the patient's age by 10, induced a significantly improvement in test specificity, particularly in very old patients with an overall NPV=60.2% vs. 54.7% using the fixed cut-off value. The overall NPV remained high (99.9%), even if a 78 y-old female with a low PTP of PE would have been misdiagnosed as her D-dimer level (540 ng/mL) was above 500 ng/mL but below the age-adjusted cut-off value. Such an improvement in test performance was found both in patients with suspected PE and DVT (Table). As this increase in test specificity would have led to exclude VTE in a higher percentage of patients in the studied population, we evaluated the cost-effectiveness of that strategy, taking into account the local cost of D-dimer testing, angiography and Doppler US (16.20, 58.72, and 75.60 Euros respectively). In the case of suspected PE, the economic impact of the proposed diagnosis strategy was a decreased of 6.9% of total costs (45,023.4 vs. 48,356.4 Euros). In the case of DVT, the overall saving was 5.1% (9,909 vs. 10,438.2 Euros). If such an analysis was used in the US, where angiography and Doppler US were more expensive (648 and 226 US$ respectively), and D-dimer less costly (14 US$), the cost savings would have been even higher (-11.0% for PE, and -6.3% for DVT). Conclusions The use age-adjusted cut-off levels for D-dimer, in patients aged over 50 years old, leaded to a significant increase in the test specificity, but correlatively to slightly decreased NPV and sensitivity, as some patients with D-dimer levels above 500 ng/mL but below the age-adjusted cut-off value could be misdiagnosed. However such a strategy was found to be safe in our studied populations, as the NPV remained high (99.9%), and cost-effective. Table Table. Disclosures No relevant conflicts of interest to declare.