Rivaroxaban versus warfarin in Chinese elder diabetic patients with non-valvular atrial fibrillation: A retrospective study on the effectiveness and associated risks (EAR) analysis

Purpose: To evaluate the effectiveness and associated risks of rivaroxaban against warfarin in Chinese elderly diabetic patients with non-valvular atrial fibrillation. Methods: Data regarding demographical characteristics, clinical conditions, ischemic stroke, intracranial bleeding, gastrointestinal tract bleeding, myocardial infractions, hip/pelvic fracture, asthma, breast/prostate cancer, and death during 3-years of treatment of 584 Chinese diabetic patients with confirmed non-valvular atrial fibrillation who were placed on rivaroxaban (RX Cohort, n = 201) or warfarin (WF Cohort, n = 383) were collected from hospital records and analyzed. Multivariate analysis was performed for the prediction of the incidence of the treatment-emergent event(s). Results: During the 3-year treatment period, higher numbers of patients were reported for intracranial bleeding (p = 0.042), ischemic stroke (p = 0.042), gastrointestinal bleeding (p = 0.0006), hip/pelvic fracture (p = 0.042), and asthma (p = 0.0007) in WF cohort than RX cohort. Also, higher mortality was reported in WF cohort than for RX cohort (24 vs. 4, p = 0.024). Female sex (p = 0.031), age (p = 0.035), and comorbidities (p = 0.021) were associated with incidence of treatment-emergent event(s). Conclusion: With rivaroxaban, a significant and safe risk-reduction of thromboembolic events are found in elderly diabetic patients with non-valvular atrial fibrillation.


INTRODUCTION
Diabetes is a common morbidity associated with non-valvular atrial fibrillation [1] and has risk for combined endpoint events and mortality [2]. Diabetes has a greater risk of embolic events [3]. Unlike Caucasians, the prevalence of stroke in Chinese patients is associated with non-valvular atrial fibrillation and diabetes [4]. Approximately 8 million Chinese people are affected with nonvalvular atrial fibrillation [2]. The current ACC/AHA/HRS (The American College of Cardiology/the American Heart Association/the Heart Rhythm Society) Guideline for the management of patients with non-valvular atrial fibrillation recommends the use of anticoagulants  [5] but the use of anticoagulants is less than 10 % in a Chinese population with a history of diabetes. Most of the patients received aspirin or nothing to prevent thromboembolism [6]. There is a big gap between the recommended guidelines and practice in China because antiplatelet is also frequently used instead of anticoagulant agents [4].
Warfarin is Vitamin K anticoagulant [7] and have been used for many years but warfarin treatment is contraindicated in elderly patients due to the risk of bleeding [4]. Also, United States Food and Drug Administration (FDA) in 2010 as an anticoagulant but has a high risk of intracranial and gastrointestinal tract bleeding [8]. Rivaroxaban is a factor Xa inhibitor and approved by FDA in 2011 as an anti-coagulant [9]. A once daily dose of 20 mg rivaroxaban is the recommended dose, 15 mg daily dose of rivaroxaban is recommended in patients with impaired kidney function, while a 10 mg daily dose of rivaroxaban is preferred in deep venous thrombosis. The primary objective of this retrospective analysis of prospectively collected data was to evaluate the effectiveness of rivaroxaban against warfarin in Chinese elderly diabetic patients with non-valvular atrial fibrillation. The secondary endpoint of the analysis was to evaluate the associated risks responsible for the treatment-emergent event(s).

Ethics approval and consent to participate
The protocol of the established study was approved by the Huilongguan Community Health Service Center Human Ethics Committee (no. HCH/CL/15/19 dated 18 July 2019). The reporting of the analysis adheres to the law of China, the strengthening the reporting of observational studies in epidemiology (STROBE) cohort study: statement, and the 2008 Helsinki Declaration. The informed consent form was signed by all the enrolled patients or their relatives (legally authorized person) for interventions, pathology, and publication of workup in all formats irrespective of time and language during hospitalization.

Inclusion criteria
Diabetic patients within the age of ≥ 60 years who had developed non-valvular atrial fibrillation (the arrhythmia duration > 7 days, confirmed by echocardiography [4]) and those the physician decided to put on any anticoagulant therapy were considered for the analysis.

Exclusion criteria
Non-diabetic patients, patients who had not developed non-valvular atrial fibrillation, and patients who had been on only antiplatelet therapies (e.g. aspirin, digoxin, and/or clopidogrel) were not considered for the analysis. Patients placed on any anticoagulants therapies but who had reduced kidney function were not considered for the analysis.

Interventions
Patients who received 20 mg rivaroxaban once in a day were considered for the RX Cohort and those who received 5 mg warfarin per day in two divided doses were included in the WF Cohort.

Outcome measures
Data regarding the event(s) of ischemic stroke, myocardial infractions, intracranial bleeding, gastrointestinal tract bleeding, breast/prostate cancer, hip/pelvic fracture, and asthma were collected from the institutional records. If patients had occlusion of cerebral arteries in imaging diagnosis (angiography and/or the Computed Tomography), it was considered as ischemic stroke. If patients had a subarachnoid hemorrhage in imaging diagnosis (angiography and/or the Computed Tomography), it was considered as intracranial bleeding. Clinical confirmation of gastrointestinal tract bleeding was carried out as per the published algorithm [10]. Myocardial infractions and breast/prostate cancer were evaluated by imaging diagnosis (angiography and/or the Computed Tomography) or biopsies wherever applicable. Hip/pelvic fracture was validated by the x-ray image. Asthma was defined by spirometry test of lung functions [11].

Statistical analysis
InStat window version 3.0, GraphPad, San Diego, CA, USA was used for statistical analysis. The Fisher's exact test was used for categorical variable and Mann-Whitney U-test was used for continuous variables. Multivariate analysis was performed for the prediction of the incidence of the treatment-emergent event(s). Data were considered significant at 95 % of confidence level.

Enrollment
From the 1 st of June 2015 to 15 th of June 2016, there were 920 diabetic patients recorded or consulted in the outpatient center of the Huilongguan Community Health Service Center, Beijing, China and the referring hospitals. Among them, 85 patients had not developed nonvalvular atrial fibrillation, therefore, they were not placed on anticoagulant or antiplatelet therapies. Also, 110 patients were below the age of 60 years. So, they were not considered for the analysis.
Among the remaining patients, 115 patients were placed on only antiplatelet therapies, so these were also excluded from the study and 26 patients who received anticoagulant therapy had reduced kidney functions. Therefore, they were not considered for the analysis. Data regarding outcome measures of 584 patients who were placed on rivaroxaban (RX Cohort, n = 201) or warfarin (WF Cohort, n = 383) were included in the analysis (Figure 1).

Demographic characteristics
All the enrolled patients were new users of rivaroxaban or warfarin and no history of ischemic stroke, myocardial infractions, intracranial bleeding, gastrointestinal tract bleeding, breast/prostate cancer, and/ or hip/pelvic fracture. There was no significant difference between clinical and demographical characteristics for both groups at the start of anticoagulant therapy (p < 0.05 for all, Table 1).

Risk factors for the incidence of treatmentemergent event(s)
Among the enrolled 584 patients of diabetes, 135 had reported one or more adverse events (stroke, bleeding, myocardial infarction, fracture, and/ or asthma) during 3-years of treatment. Among the risk factors, female sex (p = 0.031), age (p = 0.035) and comorbidities (p = 0.021) were associated with incidence of treatmentemergent event(s) ( Table 3).

DISCUSSION
The study reported that diabetic patients with non-valvular atrial fibrillation who received rivaroxaban had the fewer risk of intracranial bleeding, ischemic stroke, and mortality than those who received warfarin during 3-years of treatment. The results of the study were consistent with the results of retrospective analysis [9] and a ROCKET trial on Caucasians patients [12]. With rivaroxaban, a significant risk reduction of thromboembolic events can be found in Chinese diabetic patients with nonvalvular atrial fibrillation.
Fewer numbers of patients suffered from gastrointestinal bleeding in the cohort who received rivaroxaban than those received warfarin. The results of the study were parallel with the results of retrospective analyses [8][9][10] but were not consistent with a ROCKET trial on Caucasians patients [12]. Rivaroxaban is safe anticoagulant than warfarin.
Fewer numbers of patients suffered from hip/ pelvic fracture and asthma in a cohort who received rivaroxaban than those who received warfarin. The results of the study were parallel with the results of the retrospective analysis [9]. Warfarin inhibits the activation of bone proteins and impairs bone material hardness [13]. Hip/ pelvic fracture during 3-years of treatment is not an acceptable adverse effect hence rivaroxaban can be a good substitute for warfarin.  Fisher's exact test was performed for statistical analysis; p < 0.05 was considered significant. * Significantly higher than RX cohort Female sex, age, and the presence of comorbidities were responsible for stroke, bleeding, fracture, and the other treatmentemergent event(s). The results of the study were parallel with the results of the retrospective analysis [9]. Older women have comparatively fewer bone mass density than older men and younger women [14]. Diabetes with the other comorbidities increases the risk of stroke event(s) [3]. Anticoagulants always have the risk of bleeding [10]. The other independent risk factors responsible for stroke, bleeding, fracture, and the other event(s) are required to evaluate for treatment-emergent event(s).

Limitations of the study
There are several limitations of the analysis, for example, the limited sample size has chances of type-I errors in the statistical analysis. The large randomized phase III trial is required to state the hypothesis strongly. The study is not completely generalized because the risk factors for the incidence of embolic events and treatmentemergent adverse effects are different in Caucasians and Chinese patients [1]. The quality of anticoagulation control was not possible because the treatment was for 3-years and the study lacks initial stage pathological results.

CONCLUSION
Warfarin is an effective treatment for embolic events in elderly Chinese patients with nonvalvular atrial fibrillation but its management is difficult and has high adverse effects. Rivaroxaban appear to show a significant risk reduction of thromboembolic events with manageable treatment-emergent adverse effects.