Background: To simplify the diagnostic strategy of patients suspected for venous thromboembolism, the use of D-dimer tests has been advocated. Very important for the safety of such diagnostic strategies would be the capacity to recognise false-normal D-dimer results, in order to prevent inadequately withholding anticoagulant treatment in patients who actually have the disease. Insight in the causes of false-normal D-dimer results would therefore be necessary. We hypothesised that certain patient characteristics are associated with relatively low plasma D-dimer levels and, therefore, could increase the risk of false-normal results.
Methods: Consecutive patients with an objectively confirmed venous thromboembolic event and an independently obtained false-normal SimpliRED D-dimer test result were included in the study. For each patient, two controls with objectively confirmed venous thromboembolism and an adequate abnormal D-dimer result were selected. Baseline patient characteristics, obtained by standardised questionnaires, were compared between the two groups of patients.
Results: In total, 686 patients had a venous thromboembolic event and 47 of these patients had a false-normal SimpliRED result. Therefore, the overall sensitivity of the SimpliRED test for venous thromboembolism was 94% (95% CI: 92-95%). Although the prevalence of certain clinical characteristics was significantly higher in patients with a false-normal D-dimer result than in the controls [odds ratios for (LMW)heparin treatment and symptoms lasting more than 10 days: 5.1 (95% CI: 1.5-18.7) and 3.2 (95% CI:1.4-7.4), respectively], the prevalence of these characteristics was also high in the control group with an adequate abnormal D-dimer. Combining two or more of these characteristics had a low prevalence and did not further improve the ability to identify those patients with a false-normal D-dimer test at presentation.
Conclusions: Although these findings clearly indicate an association between certain baseline clinical characteristics and the occurrence of a false-normal SimpliRED test, the clinical utility for these characteristics is limited.