Incidence Rates of Bleeding and Emergency Surgery Due to Trauma or Fracture Among Japanese Patients with Non-valvular Atrial Fibrillation Receiving Oral Anticoagulation Therapy

Nonvalvular atrial fibrillation (NVAF) is a major risk factor for thromboembolism and increased incidence of stroke in affected individuals [1, 2]. While the oral anticoagulant (OAC) warfarin, a vitamin K antagonist (VKA), prevents stroke in patients with NVAF [1, 3], it increases the risk of hemorrhage and adverse events due to its drug–drug interactions, food–drug interactions, and variable patient response via genetic polymorphisms [4‐6]. Direct OACs (DOACs) have been developed to overcome these limitations [7], and include dabigatran, rivaroxaban, apixaban, and edoxaban [8]. Patients taking DOACs who undergo major emergency surgeries or experience medical trauma may require rapid reversal of the anticoagulant effect to reduce their risk of bleeding due to anticoagulation [9]. For all reversal agents, the benefits of stopping or reducing bleeding must be balanced against the risk of thrombosis while anticoagulation is reversed [10, 11].

Given its aging population—about 27% of the population in Japan are aged ≥ 65 years [12]—clinical scenarios requiring anticoagulant reversal are expected to be encountered more frequently in Japan, and this is for two main reasons. Firstly, around one-half of the patients experiencing severe trauma in Japan are aged > 60 years (52.9% of 127,303 trauma patients), with the most frequent cause being ground-level falls (55.2%) [13]. Emergency surgery for bone fractures was frequent among elderly trauma patients in a study in Japan (n = 631); it was reported in 25% of those aged 65–84 and in 18% of those aged ≥ 85 years [14]. Secondly, anticoagulant usage (warfarin or DOACs) among 8855 adult patients with NVAF in Japan increased between 2006 and 2013 from 20.0 to 49.1%, with a more dramatic increase in the subgroup aged 65–74 years (7.9–59.9%) [15]. Specific prescription data for DOACs in 2014 suggests that they comprise ~ 40% of all anticoagulant prescriptions in patients with NVAF (aged 18–74 years, n = 4380) [16].

What is not yet known is the incidence rate of emergency surgery or bleeding related to trauma or fracture specifically among patients with NVAF taking DOACs in Japan. The aim of the present study was to assess the need for reversal agents in Japanese patients with NVAF on OAC therapy by determining the incidence of emergency surgeries and major bleeding episodes associated with fractures and trauma in this patient population.

Within the database, 62,888 patients with NVAF who initiated OAC monotherapy were identified, of whom 53,969 met the criteria for study inclusion (Fig. 1). Overall, 40% of the study population were female and 60% of the overall cohort were aged ≥ 75 years. The mean age ± SD of the study cohort was 76 ± 10 years (Table 1).

We have determined that the annual incidence rate of emergency surgery or major hemorrhage associated with fracture and injury was ~ 0.5% among 53,969 adult NVAF patients on OAC therapy. In the subgroup of very elderly patients (aged ≥ 75 years) this annual rate was ~ 0.6%, which was almost double that of those aged ≤ 64 years (~ 0.3%).

Since our study did not include assessment of OAC-related bleeding, only bleeding related to fractures or trauma/injury and emergency surgery, this may explain why our incidence rates were lower than reported in previous retrospective or observational analyses conducted in adult patients with NVAF receiving OAC therapy (warfarin or DOAC) that reported major bleeding incidence rates (2.4–7.5 per 100 person years [19‐21]) or cumulative incidence (1.2–4.7% [22, 23]).

In addition, OAC-related major bleeding rates may generally be lower among Japanese patients than in patients from other countries, as exemplified by the Global Anticoagulant Registry in the FIELD-Atrial Fibrillation (GARFIELD-AF) ongoing multi-national observational study of stroke prevention in patients with newly diagnosed NVAF [24]. In this study, the major bleeding event rate in Japan was 0.32 (95% CI 0.19–0.53) per 100 patient-years versus 0.91 (0.82–1.00) per 100 patient-years in other countries [24]. Major bleeding incidence rates in one Japanese study were 2.2 per 100 patient-years (patients aged ≥ 75 years) and 1.4 per 100 patient-years (patients aged < 75 years) among 9578 NVAF patients receiving rivaroxaban [25]. Major bleeding rates associated with DOAC therapy (n = 1676) that varied by age were also reported by Nishida and colleagues [26]: 0.973 per 100 patient-years (age < 75 years), 1.720 per 100 patient-years (age 75–84 years), and 0.985 per 100 patient-years (≥ 85 years). Annualized major bleeding incidence rates in Japanese NVAF patients on warfarin were 1.2% per year and on a DOAC were 0.5% per year in an analysis of 3-year clinical outcomes of a Japanese registry study [27]. Another Japanese claims database study reported the incidence rates per 1000 patient-years of major bleeding (6.39), GI bleeding (1.59), and intracranial bleeding (0) among 4606 dabigatran recipients [28].

We did not determine the incidence rate of bleeding associated with fractures or trauma separately from emergency surgery. Falls are the primary cause of severe trauma in the elderly [13]. Falls may lead to traumatic brain injury (TBI [29]), traumatic spinal cord injury (TSCI [30]), or fractures. The incidence of falls is higher with increasing age, as is the incidence of fractures. Several Japanese studies have reported an age-related increase in the incidence of ground-level falls resulting in severe trauma [13], annual hip fracture incidence [31], bone fractures requiring emergency surgery in trauma patients [14], annual TSCI incidence [30], and TBI from non-motor vehicle accident causes [32]. While much of these data on falls, fractures, and severe traumatic injury are not specific for patients with NVAF receiving OAC therapy, the age-related increased incidence of such events makes it likely that trauma clinicians and surgeons will encounter patients with NVAF receiving OAC therapy.

Our study has defined the incidence rate of patients undergoing emergency surgery or who have bleeding associated with fractures or trauma in Japan. Together with data indicating that falls are a large contributory factor to the incidence of trauma in Japan [13], that OAC-related bleeding may occur in up to 2.2% of patients with NVAF, and Japan’s aging population, there is a clear reason to anticipate an increased need for anti-coagulation reversal agents.

The main limitation of the present study was the fact that, as a retrospective database analysis, it relied on the integrity of the source information. However, the definitions of outcomes used as endpoints would likely prevent any incorrectly coded events from being included in the analysis, so if anything, is likely to underestimate the incidence of events. Also, given its retrospective nature, it was not possible to capture information about the magnitude of change in hemoglobin levels or the severity of bleeding. As previously mentioned, the current analysis of outcomes was limited to hospitalized or transfused bleeding associated with fracture or trauma, and did not include life-threatening intracranial or GI bleeding, which were both indications for idarucizumab use in the RE-VERSE AD study [33], or other OAC-related major bleeding. Therefore, this analysis probably underestimates the incidence of bleeding events that may require rapid reversal. Furthermore, the definition of emergency surgery used in the RE-VERSE AD study could not be replicated in this study since the database lacked hourly time stamp information. As a consequence, we used the definition of same day or 1 day after hospitalization, which could result in over-estimation of the incidence rates. Another limitation was that only Japanese patients were included, which could limit the applicability of the findings in other populations. Lastly, we did not conduct any analyses comparing outcomes in patients receiving different types of OACs (warfarin or individual DOACs) or in patients who received reversal agents such as FFP or vitamin K. Since prescribing decisions in real-world clinical practice are individualized for each patient, based on patient demographics and comorbidities, such analyses would be subject to selection and channeling bias that would be difficult to control for statistically, undermining the possibility of obtaining meaningful comparative data.

In conclusion, although relatively few patients with NVAF receiving OACs experience emergency surgeries and major bleeding episodes associated with fractures and trauma, the incidence rate of these events is markedly higher in patients of advanced age. This study highlights the importance of having an effective reversal agent for patients on OACs, particularly those in older age groups.