Traumatic injury: another unjustified reason to stop oral anticoagulation for atrial fibrillation

This editorial refers to ‘Resumption of oral anticoagulation following traumatic injury and risk of stroke and bleeding in patients with atrial fibrillation: a nationwide cohort study’†, by L. Staerk et al., on page 1698. The most important public health problem with oral anticoagulants (OACs) for atrial fibrillation (AF) is their underuse, resulting in hundreds of thousands of preventable strokes yearly, worldwide. While the nearly 70%1 , 2 relative risk reduction in stroke has led to guideline recommendations for OACs for stroke prevention in AF,3 approximately half of AF patients at risk for stroke are not receiving OACs.4 , 5 In the USA, ∼84% of AF-related strokes occur in patients not on therapeutic OACs.6 The reason for OAC underuse is multifactorial. The most commonly reported reasons providers give for not prescribing OACs include previous bleeding episodes, high risk of bleeding, and frequent falls.7 However, providers commonly have misperceptions that aspirin is effective for stroke prevention in AF,8 and these beliefs contribute to OAC underuse. The good news is that targeted education interventions for patients and providers coupled with real-time reporting and feedback have been shown to increase the use of OACs, cutting the gap in undertreatment in half.9 Interventions improving treatment are especially effective for patients being inappropriately treated with aspirin, which has a strong recommendation to not be used for stroke prevention in the European AF Guidelines.3 Restarting anticoagulation after an event that required OAC interruption is another major clinical challenge, resulting in long-term underuse of OACs. Only half of patients with a gastrointestinal bleed have their OAC restarted, despite OAC use being associated with lower mortality rate, lower rate of thrombo-embolism, and no higher rates of recurrent gastrointestinal bleeding.10 Decision analysis modelling demonstrated a benefit for OACs after upper gastrointestinal bleeds for most patients.11 Similarly, a decision analysis focusing on falls found that although falls are common, especially among the elderly, traumatic falls are uncommon, and the benefits of OACs generally outweigh the risk of falling, with the estimate that a patient would need to fall 295 times a year for the risk of falls to outweigh the benefits of warfarin.12 Since non-vitamin K antagonists (NOACs) have much lower rates of intracranial haemorrhage (including traumatic intracranial haemorrhage) than vitamin K antagonists (VKAs), the threshold for not using NOACs is even higher. Edoxaban has been shown, compared with warfarin, to be associated with a greater reduction in all-cause mortality and major bleeding among patients at risk for falls relative to the reductions seen in the overall ENGAGE AF-TIMI 48 (Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation-Thrombolysis In Myocardial Infarction 48) trial population.13 In this issue of the journal, Staerk et al. describe results of an analysis from a Danish nationwide registry, evaluating resumption of OACs and associated clinical outcomes among 4541 patients on an OAC at the time of a traumatic injury.14 Traumatic injury was identified retrospectively through claims data for ICD-10 (International Classification of Disease 10th Revision) coding for traumatic brain injury, torso or abdominal injury, and hip fracture. At the time of the traumatic injury, 82.6% of patients were taking a VKA, while 17.4% were taking an NOAC. Within 90 days of the traumatic injury, 77.3% of patients filled prescriptions for OACs (60.6% VKAs and 16.7% NOACs). Patients that did not resume OACs had higher HAS-BLED scores (3 vs. 2), more frequently had traumatic brain injury as the initial event (40.8% vs. 27.9%), were more often on a VKA at the time of their initial event (90.2% vs. 80.4%), had higher rates of previous bleeding (45.2% vs. 29.5%), and were more likely to be treated with aspirin post-traumatic injury (27.1% vs. 14.9%). The rate of resumption of OACs increased over the study period from 68.7% in 2009 to 85.6% in 2016 (P < 0.0001). Resuming OACs after traumatic injury was associated with lower all-cause mortality and lower rates of ischaemic stroke. The observation that there was no association between resuming OACs and recurrence rates of traumatic injury suggests that risk of recurrent trauma was not what drove the decision to resume anticoagulation. Resuming VKAs was associated with higher major bleeding [hazard ratio (HR) 1.30, 95% confidence interval (CI) 1.03–1.64], while resuming NOACs was not associated with higher major bleeding (HR 1.15, 95% CI 0.81–1.63), relative to non-resumption. These findings were consistent across a variety of subgroup and sensitivity analyses: type of traumatic injury; resumption of VKAs vs. switching from a VKA pre-event to an NOAC post-event; only patients with HAS-BLED score ≥4; and medication filled within 30, 60, or 180 days instead of 90 days. The clinical implications of this study need to be viewed in the context of several limitations. First, claims data did not allow a direct linkage between anticoagulation status and the traumatic injury event. If a patient had a medication filled within 90 days prior to the traumatic injury, it was assumed that the patient was on an OAC at the time of the event; however, the timing of the last dose of an NOAC before the injury was not known, and international normalized ratio (INR) values before or at the time of the traumatic injury were not known. Secondly, the exact timing of when patients resumed their OACs following the traumatic injury was not known. Given that 97.9% of patients that resumed an OAC after traumatic injury restarted the same OAC they had been taking prior to the event, these patients probably had a supply of their OAC at home, so the timings of subsequent medication fills after the event were probably not representative of the resumption date of OACs. Although the data are supportive of restarting OACs after a traumatic injury, the analysis did not provide insight into the optimal timing of OAC resumption. Finally, the 1-year absolute risk differences in all-cause mortality for VKA and NOAC resumption (–4.2% and –4.4%, respectively) were meaningfully higher than the absolute risk differences in stroke (–0.6% and –1.2%, respectively). It was unexpected that VKAs only had a 15% relative lower rate of stroke at 1 year, while NOACs had only a 30% relative lower rate of stroke at 1 year, compared with no anticoagulation. Regardless, since OACs would be expected to affect mortality based on decreased stroke, the higher mortality in the non-resumption cohort probably represented some degree of unmeasured confounding. Providers should note that risk of recurrent traumatic injury was low (1-year standardized absolute risks ranged from 4.0 to 4.4%), while the 1-year standardized absolute risk of stroke or thrombo-embolism in the non-resumption cohort was similarly 4.0%. This emphasized the importance of resuming an OAC after traumatic injury, as the risk of recurrent traumatic injury was similar to the risk of stroke in a non-anticoagulated population. A multivariable model in the analysis found an increased risk of bleeding with the use of aspirin (HR 1.27, 95% CI 1.03–1.55) and use of non-steroidal anti-inflammatory drugs (HR 1.25, 95% CI 1.01–1.55) combined with anticoagulation. As the European AF Guidelines recommend, providers should focus on avoiding these (and other) preventable contributors to bleeding.3 Although not the focus of the manuscript, underuse of OACs prior to traumatic injury was identified. There were 8365 patients excluded from the analysis due to not taking an OAC at the time of the traumatic injury, and this was 56.3% of the 14 871 patients otherwise not excluded from the study. The CHA2DS2-VASc scores for the untreated patients were not known, but this traumatic injury population was an older population with the patients on OACs having a median CHA2DS2-VASc score of 4 (interquartile range 3–5), so it was probably a small minority that did not meet guideline criteria for an OAC. This 56.3% of AF patients who were not treated with an OAC remains consistent with the previously described rates of underuse.4 , 5 The majority of patients did resume an OAC after their traumatic injury, but a substantial minority of patients (22.7%) did not resume an OAC. A portion of non-resumption patients probably had a true contraindication to OACs, but 27.1% of these non-resumption patients were treated with an aspirin and not an OAC. The AVERROES [Apixaban Versus Acetylsalicylic Acid (ASA) to Prevent Stroke in Atrial Fibrillation Patients Who Have Failed or Are Unsuitable for Vitamin K Antagonist Treatment] trial showed that apixaban resulted in substantially lower stroke than aspirin (HR 0.45, 95% CI 0.32–0.62), while having only a modest (non-significant) increase in major bleeding (HR 1.13, 95% CI 0.74–1.75) and numerically fewer intracranial haemorrhages (11 vs. 13) than aspirin (two-thirds of aspirin dosing was 81 mg).15 Especially with the improved safety profile of NOACs, many (or most) patients treated with aspirin after their traumatic injury should have been treated with an OAC. It is gratifying that rates of resuming an OAC after traumatic injury increased over the course of the study. Nonetheless, failure to resume an OAC after various types of bleeding and trauma (Figure 1), when evidence strongly suggests benefit, is an important factor resulting in preventable AF-related strokes. This study provides supportive evidence that most patients would benefit from resuming an OAC after traumatic injury. Implementation programmes addressing patient-level, provider-level, and system-level barriers have been effective at increasing rates of OAC use. Expansion of implementation programmes should include a focus on ensuring that most patients who have traumatic injury (or other events leading to discontinuation of OACs) resume anticoagulation to prevent stroke. Management of patients and oral anticoagulation after acute events. VKA, vitamin K antagonist; NOAC, non-vitamin K antagonist; OAC, oral anticoagulant; PPI, proton pump inhibitor; NSAID, non-steroidal anti-inflammatory drug. Conflicts of interest: S.D.P. reports modest research grant support from Bristol-Myers Squibb, Pfizer, Gilead, Boston Scientific, and the Food and Drug Administration; and modest Consultant/Advisory Board from Bristol-Myers Squibb, Pfizer, Portola, Boston Scientific, and Medtronic. C.B.G. reports research grants from Armetheon, Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, Bayer, Daiichi, Janssen, GlaxoSmithKline, Medtronic Foundation, Sanofi-Aventis, and The Medicines Company; and Consultancy fees from AstraZeneca, Boehringer-Ingelheim, Bristol-Myers Squibb/Pfizer, Daiichi, GlaxoSmithKline, Hoffman-La Roche, Janssen, Lilly, Sanofi-Aventis, and The Medicines Company.