(CVP005) RECURRENCE AFTER ANTICOAGULANT CESSATION IN UNPROVOKED VENOUS THROMBOEMBOLISM: IMPACT OF DEATH AS A COMPETING OUTCOME

Background: Despite guideline recommendations for long-term anticoagulation after the first diagnosis of unprovoked venous thromboembolism (VTE), benefits and harms associated with this approach remain unclear among older adults at risk of non VTE-related mortality. While death is a competing event to recurrent VTE, competing risk methods have not been used in determining VTE recurrence rates following anticoagulant discontinuation in unprovoked VTE. We sought to evaluate the risk of recurrent VTE by age strata among patients with first unprovoked VTE after anticoagulant cessation, accounting for all-cause mortality as a competing risk.

METHODS AND RESULTS: Using data from the REVERSE cohort, we analyzed VTE recurrence among patients who stopped anticoagulation after 6 months with stratification by age ( < 50 years, 50-64 years, 65-74 years, ≥75 years). We compared VTE recurrence estimates among patients within each age stratum using the Kaplan-Meier method and cumulative incidence function (CIF). We assessed the impact of death as competing outcome on the relationship between of age and recurrent VTE using the Fine and Gray sub-distribution and cause-specific hazards models.

Among 663 patients enrolled in the study, 38 (5.7%) died over a mean follow-up of 5.0 years. Incidence of all-cause mortality was concentrated among men aged ≥75 years at the time of anticoagulant cessation (8.3 vs. 1.1 per 100 patient-years in the overall cohort). We observed numerically lower cumulative incidence of recurrent VTE using competing risk methods among men aged ≥75 years (27.9% with CIF vs. 49.8% with Kaplan-Meier; Figure). However, risk of recurrent VTE at 1 year exceeded 5% with either method (9.8% with CIF vs. 8.2% with Kaplan-Meier). Using the Fine and Gray hazards model, we did not observe an effect of age on VTE recurrence among men (Table). On the other hand, women aged 65 - 74 at the time of anticoagulant cessation had a 2.8-fold increased risk of VTE recurrence during follow-up compared to women < 50 years (hazard ratio 2.80, 95% CI 1.42 - 5.52, Table).

Conclusion: Long-term risk of VTE recurrence among men ≥75 years differs by over 20% when death as competing outcome is accounted for. However, risk of recurrent VTE at 1 year exceeded the 5% threshold for anticoagulant continuation regardless of the methodology used. While the impact of all-cause mortality as a competing risk is limited in the overall unprovoked VTE population, its role among patients with >50% anticipated 1-year risk of death (e.g., in the palliative care setting) should be the focus of additional investigations.

Lay Abstract Content: Anticoagulants, or blood-thinners, are used in the treatment of blood clots in the veins (known as deep vein thrombosis) or those that travel to the lungs (known as pulmonary embolism). Too much blood-thinning, however, increases the risk of bleeding. Therefore, it is crucial that we limit blood-thinners only to individuals who are at high risk of clot recurrence. Most data that have looked at clot recurrence over time have not accounted for the risk of death, which is important as individuals who die will no longer be at risk of having recurrent clots (this phenomenon is called a competing risk). In this analysis, we used data from a study that followed 663 individuals over 7 years, to see whether accounting for the risk of death may change the estimated risk of venous clot recurrence. We found that at risk of clot recurrence among men over 75 years was 22% lower when we accounted for the risk of death (28% compared to 50%); however, the risks remained high using either approach. Our results are useful for understanding the impact of death on estimated risk of clot recurrence, especially in settings where anticipated risk of death is high (such as at palliative care or end-of-life settings) in order to inform shared decision-making between clinicians and patients on whether to continue or stop blood-thinners after a venous clot.