Whether post-anticoagulation D-dimer levels are useful in predicting recurrence in elderly patients with unprovoked venous thromboembolism is unknown.
We followed up 157 patients aged ≥65 years with acute symptomatic, unprovoked venous thromboembolism in a prospective, multicenter cohort study. All patients completed 3-12 months of anticoagulation and then underwent quantitative D-dimer testing (enzyme-linked immunosorbent assay) 12 months after the index venous thromboembolism. The outcome was recurrent symptomatic venous thromboembolism after D-dimer measurement. We examined associations between log-transformed and dichotomized D-dimer values and the time to venous thromboembolism recurrence using competing risk regression, adjusting for age, sex, and overt pulmonary embolism.
There was no statistically significant association between quantitative or dichotomized D-dimer levels and venous thromboembolism recurrence. The area under the receiver operating characteristic curve for predicting recurrent venous thromboembolism was moderate (0.66; 95% confidence interval [CI], 0.51-0.82). The negative likelihood ratios were 0.34 (95% CI, 0.05-2.38) at the usual and 0.34 (95% CI, 0.09-1.29) at the age-adjusted cutoff values. Among patients with normal D-dimer results, venous thromboembolism recurrence rates were 6.8 (95% CI, 2.2-21.2) per 100 patient-years using the usual and 7.1 (95% CI, 3.2-15.8) per 100 patient-years using the age-adjusted cutoff values.
D-dimer testing alone may not be useful in identifying elderly patients with unprovoked venous thromboembolism who are at low risk of recurrent venous thromboembolism and in whom anticoagulants may be safely stopped.