The Bleeding Risk of Vitamin K Antagonists Hardly Increases after the Age of 80 Years: Data from a Large Real-Life Cohort Study

Introduction: Many physicians are reluctant to prescribe anticoagulants in the elderly, because of the increased bleeding risk. On the other hand, the potential benefit of anticoagulants also increases as the thrombotic risk rises as well. The BAFTA-trial demonstrated that patients over 75 years still benefit from vitamin K antagonists (VKA) use for atrial fibrillation, and the RIETE-study showed that patients over 80 years have a higher risk to die of recurrent thrombosis than of bleeding while on VKA. However, even in these studies the number of nonagenarians was low. Given the aging population, data on the real-life safety and efficacy of VKA in these very old patients are urgently needed.

Aim: To determine the safety and efficacy of VKA in octogenarians and nonagenarians compared to septuagenarians.

Methods: We selected all nonagenarians from a consecutive cohort of 26,089 patients treated with VKA at Certe Thrombosis Service Groningen (2009-2012). Subsequently, every nonagenarian was randomly matched with one septuagenarian and one octogenarian. The primary endpoint was bleeding events (clinically relevant non-major and major). Secondary endpoints were thrombotic events, VKA related deaths, and quality of VKA control. All events were adjudicated by two experienced physicians who were unaware of age. The relation with age was analyzed in Cox regression models, septuagenarians being the reference group.

Results: In total, 3313 patients were included. The 713 patients (22%) with at least one bleeding event had in total 986 clinically relevant and 64 major bleeding events (Table 1). The bleeding risk of the octogenarians (HR 1.07; CI 0.89-1.27) was comparable with the septuagenarians. Nonagenarians had a mildly increased risk (HR 1.26; CI 1.05-1.50), and a non-significantly higher risk of major bleeding (HR 1.20; CI 0.65-2.22) and fatal bleeding (HR 1.32; 0.58-3.01). The outcomes of the analyses in the inception cohort did not differ essentially from the main analyses.

The risk to develop a thrombosis was higher in the nonagenarians (HR 2.14; CI 1.22-3.75) and octogenarians (HR 1.75; CI 1.002-3.05), and this was even more pronounced in the inception cohort (HR 5.57; CI 1.22-25.4 and HR 5.00;CI 1.10-22.8, respectively). The risk of thrombosis related death was also increased in the nonagenarians (HR 4.53; CI 1.94-10.5).

The control of individual time in the therapeutic range and variability became significantly poorer with rising age (Table 1). Interestingly, the time under and above the range remained well balanced. The increased bleeding risk of nonagenarians reduced from 1.26 to 1.16 (CI 0.96-1.39) after correction for VKA control. In contrast, the increased thrombotic risk was independent of VKA control. Thus, VKA users over 80 years probably had a relatively high baseline thrombotic risk compared to their bleeding risk.

The number of thrombotic events (20) was comparable with the number of major bleeds (22) in the septuagenarians. However, the thrombotic events (32 and 33, respectively) outweighed the major bleeding rate (20 and 22, respectively) in the nonagenarians and octogenarians.

Conclusion: In this real-life cohort study, the bleeding risk was not increased in octogenarians and only mildly increased in nonagenarians, compared to septuagenarians. Moreover, the thrombotic risk outweighed the bleeding risk in both octogenarians and nonagenarians. Therefore, any age, even above 90 years, should in itself not be a reason to withhold anticoagulants.