Abstract 498: Age-adjusted Cut-off Levels Improve Test Performance in the Diagnosis Strategy of Venous Thromboembolism in Elderly Patients With a Non-high Pre-test Probability

Abstract

Background D-dimer levels below a well defined cut-off level enable to safely rule out VTE in patients with non-high PTP. Ageing is associated with higher levels of D-dimer. To address the issue of their usefulness in elderly patients, we firstly managed data from 644 out-patients with non-high PTP included in a multicenter management study of a D-dimer assay (C.Legnani et al. 2010). We then validated our results in a cohort of 1,042 consecutive patients investigated using another assay. Results In the first part, D-dimer level, evaluated using the HemosIL D-dimer HS500 assay (Instrumentation Laboratory), was above the cut-off level (500 ng/ml) in all 88 patients with VTE and in 299 of those without (53.8%). Test negative predictive value (NPV) and sensitivity were 100%. Overall test specificity was 46.2%, but decreased in an age-dependent manner over 70 y. ROC-analysis of test results obtained in patients classified according to age allowed us to propose an age- adjustment of the cut-off level calculated by increasing its level by 100 ng/ml per 10 y-increments in patients aged over 59 y e.g. 600 ng/ml between 60 et 69 y, 700 ng/ml between 70 et 79 y. Using those age-adjusted cut-off levels yielded to significantly improve test specificity (56.1%), particularly in old patients. If an 84 y old patient (D-dimer=625 ng/ml) was missed, NPV remained high (99.6%). That strategy was then validated by evaluating data from 1,042 consecutive out-patients, investigated using the Vidas D-dimer assay (BioMérieux). When applied, that age- adjusted cut-off levels leaded to improve test specificity 62.2% vs. 55.1% using the fixed cut-off level. 3 additional patients, with D-dimer >500 ng/ml but below age- adjusted cut-off level, were missed but overall NPV remained high 99.0% vs. 99.4%. Conclusions Age-adjusted cut-off levels for D-dimer, calculated by increasing the traditional cut- off level (500 ng/ml) by 100 ng/ml per 10 y-increments in patients aged over 59 y, significantly increased the test specificity. Correlatively, NPV and sensitivity were slightly decreased, as some patients, with D-dimer above 500 ng/ml but below the age-adjusted cut-off level, could be misdiagnosed. However such a strategy was safe, as NPV remained above 99% in our studied populations.