Predictors of International Normalized Ratio Variability in Patients With Atrial Fibrillation Under Warfarin Therapy

Abstract

Background: Variability in the international normalized ratio (INR) of prothrombin time has been suggested to be related to outcome in patients with atrial fibrillation (AF) under warfarin therapy, but its determinants remain unclear.

Methods and Results: The study population consisted of 626 AF patients under warfarin therapy in the Shinken Database (n=22,230). INR variability was calculated by Fihn’s method. Determinants of high log INR variability (defined as over mean+standard deviation) were determined by logistic regression analyses. Symptomatic heart failure (odds ratio [OR] 3.974, 95% confidence interval [CI] 2.510–6.292), older age (≥75 years old; OR 2.984, 95% CI 1.844–4.826) and severe renal dysfunction (eGFR <30 mL/min/1.73 m²; OR 3.918, 95% CI 1.742–8.813) were identified as independent predictors of high INR variability on multivariate logistic regression analysis.

Conclusions: The determinants of INR variability in AF patients under warfarin therapy could assist Japanese clinicians in identifying patients likely to show unstable warfarin control irrespective of the definition of the target INR range.

Atrial fibrillation (AF) is associated with increased risks of ischemic stroke and death, both of which are reduced by anticoagulation therapy.¹ Despite increased usage of direct oral anticoagulants (DOACs) in real-world practice,^2,3 warfarin is still frequently used in patients with AF, presumably because of limited indications or the high medical cost of DOACs.

Editorial p 17

The anticoagulation intensity of warfarin therapy is monitored by the prothrombin time-international normalized ratio (INR). As poor INR control is closely associated with the incidence of adverse outcomes, including thromboembolism and bleeding,^4–9 it is important to predict patients expected to have poor INR control. The time in the therapeutic range (TTR), which reflects the proportion of days within a target INR range, has been widely used for evaluating the quality of warfarin control^10–12 and predictors of low TTR have been identified.^13–15

Apart from the TTR, the concept of INR variability, reflecting stability of the INR under warfarin therapy, has been developed.¹⁶ Although INR variability has not been as clinically relevant as TTR, several reports have identified that it can predict adverse events, including bleeding and thromboembolism, under warfarin therapy, independently of TTR.^8,9,17 The independency of INR variability and TTR is reasonable because the 2 measurements evaluate different aspects of INR control: INR variability measures the degree of instability of INR control,¹⁶ while TTR calculates the time period in which the INR is controlled within the therapeutic range.¹⁰ Accordingly, the predictors of poor control of INR based on INR variability may be different from those based on TTR. However, there have been few reports on the predictors of INR variability to date,¹⁸ and to our knowledge there have been no such studies in patients with AF.

In the present study, we identified independent predictors of INR variability in patients with AF under warfarin therapy in a single hospital-based cohort from the Shinken Database.^19–21

Methods

Study Population

The Shinken Database, which was established in June 2004,^19–21 contains data on all new patients attending the Cardiovascular Institute Hospital, Tokyo, Japan (abbreviated in Japanese as ‘Shinken’), excluding foreign travellers and patients with active cancer. A total of 22,230 new patients were registered in the database up to March 2014.

We identified 1,154 patients with AF on warfarin therapy in this database. Among these, we screened for patients who received their first warfarin prescription within 90 days after the first visit and continued treatment for more than 12 months in the hospital. Accordingly, 626 patients fulfilling these eligibility criteria were included in the study (Figure 1).

Figure 1.

Flowchart of patients selected for study of INR variability under warfarin treatment for AF. AF, atrial fibrillation; INR, international normalized ratio.

Data Collection at the Initial Visit

After obtaining an ECG and chest X-ray, the cardiovascular status of each patient was evaluated by echocardiography, exercise test, 24-h Holter recordings and blood laboratory data when necessary. In addition to sex, age, height and weight, we collected data on cardiovascular diseases, including heart failure (HF: New York Heart Association class ≥2), valvular heart disease (moderate or severe stenosis or regurgitation on echocardiography), coronary artery disease, hypertrophic and dilated cardiomyopathy, and history of disabling cerebral infarction or transient ischemic attack. The presence of cardiovascular risk factors, including hypertension (use of antihypertensive agents, systolic blood pressure ≥140 mmHg, or diastolic blood pressure ≥90 mmHg on admission), diabetes mellitus (use of oral hypoglycemic agents or insulin, or glycosylated hemoglobin ≥6.5%), dyslipidemia (use of statins or drugs for lowering triglycerides, low-density lipoprotein cholesterol ≥140 mg/dL, high-density lipoprotein cholesterol <40 mg/dL or triglyceride ≥150 mg/dL), chronic kidney disease (estimated glomerular filtration rate [GFR] <60 mL/min/m²), chronic obstructive pulmonary disease and use of anticoagulant or antiplatelet medications. Body mass index was calculated as weight in kilograms divided by height in meters squared. The GFR was estimated using the new Japanese coefficient for the modified isotope dilution mass spectrometry (IDMS)-traceable 4-variable Modification of Diet in Renal Disease study equation (GFR=194×SCr^−1.094×Age^−0.287×0.739 [if female]).²²

Definition of AF

AF at the initial visit was diagnosed on 12-lead surface ECG and 24-h Holter recordings. It was also diagnosed based on any medical history of AF from referring physicians.

Measurement of Prothrombin Time

Prothrombin time was measured routinely (with intervals of 1–3 months) for dose adjustment of warfarin (reagent: HemosIL RecombiPlasTin 2G, Instrumentation Laboratory, Bedford, MA, USA; analyzer: Coagtron-180, Kyowa Hakko Kirin, Tokyo, Japan; normal range, 9.7–12.9 s). For calculation of INR variability, INR measured at the outpatient clinic between 3 and 12 months after initiation of warfarin therapy was used. The average frequency of INR measurement between 3 and 12 months was 7.4±2.7 times (Figure 2). We excluded the in-hospital data points so we could exclude measurements performed under temporal discontinuation of warfarin because of surgery or bleeding.

Figure 2.

Frequency of measurements in a study of INR variability under warfarin treatment for AF. AF, atrial fibrillation; INR, international normalized ratio; PT, prothrombin time.

INR Variability

INR variability was calculated by Fihn’s method with the following equation, which calculates the time-weighted INR variance.¹⁶

The difference in 2 INR measurements was squared and divided by the time interval in days, and the values at each time interval were averaged. The INR variability was defined as its square root. We log-transformed INR variability to be able to calculate the average and standard deviation (SD) based on the normal distribution (Figure 3). Previous study showed that a patient with log INR variability of −5.0 has a much more stable anticoagulation regimen compared with another patient with a log INR variability of −0.75.⁹

Figure 3.

Distribution of number of patients with log INR variability from −4.0 to 0. Dotted lines indicate the borderline of tertiles of INR variability (−1.245). INR, international normalized ratio.

Statistical Analysis

Statistical analysis was performed using SPSS version 19.0 (SPSS, Chicago, IL, USA). In all analyses, P<0.05 was taken to indicate statistical significance. Means and SD were calculated for continuous variables. Numbers and percentages were calculated for categorical variables.

We divided the study patients into the high-INR variability group and the non-high-INR variability group, where the high-INR variability indicated an upper outlier defined as over than the mean+1*SD. The differences in continuous and categorical variables were tested using the unpaired t-test and chi-square test, respectively. Logistic regression analysis was performed to identify the clinical factors related to the high-INR variability group. Covariates significantly associated with the high-INR variability group in the univariate models were incorporated into the multivariate model by the stepwise method.

Ethical Issues

The Ethical Committee at the Cardiovascular Institute granted approval for this study and all the patients gave written informed consent.

Results

Distribution of INR Variability

Figure 3 shows the distribution of log INR variability in our patients. The mean (median) log INR variability was −1.961 (−1.944) with SD of 0.716. Therefore, the cutoff of the high-INR variability (mean+SD) was set at log INR variability −1.245 (INR variability 0.287).

Patients’ Characteristics

The characteristics of the patients in the present study are shown in Table 1. Patients in the high-INR variability group were older, were more likely to be female and had lower body weight than those in the non-high-INR variability group. Moreover, the patients in the high-INR variability group had higher prevalence of ischemic heart disease, valvular heart disease, cardiomyopathy and symptomatic HF. Hypoalbuminemia, anemia and renal dysfunction were more prevalent, and renin-angiotensin inhibitors and class III/IV antiarrhythmic drugs were more often prescribed for patients in the high-INR variability group.

Table 1. Characteristics of AF Patients in a Study of INR Variability Under Warfarin Treatment

	Total (n=626)	Non-high-INR var (n=519)	High-INR var (n=107)	P value
Demographics
Age, years	65.8±11.5	65.2±10.9	68.7±13.6	0.004
≥75 years	147 (23.5)	102 (19.7)	45 (42.1)	<0.001
Male sex	457 (73.0)	383 (73.8)	74 (69.2)	0.340
Body weight, kg	64.4±13.8	65.0±13.9	61.5±13.0	0.020
Comorbidities and risk factors
Current smoker	67 (10.7)	56 (10.8)	11 (10.3)	1.000
Hypertension	329 (52.6)	258 (49.7)	71 (66.4)	0.002
Dyslipidemia (treated)	190 (30.4)	137 (26.4)	53 (49.5)	<0.001
Diabetes	123 (19.6)	95 (18.3)	28 (26.2)	0.081
Chronic kidney disease	451 (72.0)	366 (70.5)	85 (79.4)	0.075
Ischemic heart disease	74 (11.8)	54 (10.4)	20 (18.7)	0.021
Valvular heart disease	105 (16.8)	71 (13.7)	34 (31.8)	<0.001
Cardiomyopathy	54 (8.6)	39 (7.5)	15 (14.0)	0.037
Congenital heart disease	5 (0.8)	3 (0.6)	2 (1.9)	0.204
Symptomatic HF (NYHA ≥2)	181 (28.9)	123 (23.7)	58 (54.2)	<0.001
Previous stroke or TIA	41 (6.5)	29 (5.6)	12 (11.2)	0.050
Previous hemorrhage	9 (1.4)	9 (1.7)	0 (0)	0.370
COPD	5 (0.8)	3 (0.6)	2 (1.9)	0.204
Paroxysmal AF	290 (46.3)	240 (46.2)	50 (46.7)	1.000
Persistent AF	336 (53.7)	279 (53.8)	57 (53.3)	1.000
Medications
β-blockers	74 (11.8)	64 (12.3)	10 (9.3)	0.510
RAS inhibitors	288 (46.0)	224 (43.2)	64 (59.8)	0.002
Class I antiarrhythmic agents	167 (26.7)	146 (28.1)	21 (19.6)	0.073
Class III antiarrhythmic agents	48 (7.7)	38 (7.3)	10 (9.3)	0.432
Class IV antiarrhythmic agents	127 (20.3)	101 (19.5)	26 (24.3)	0.291
Digoxin	144 (23.0)	120 (23.1)	24 (22.4)	1.000
Laboratory data
Albumin, g/dL	4.1±0.5	4.1±0.4	3.9±0.6	0.001
<3.5 g/dL	46 (7.3)	29 (5.6)	17 (15.9)	<0.001
eGFR, mL/min/1.73 m2	50.5±15.4	51.6±14.5	45.4±18.1	<0.001
≥60 mL/min/1.73 m2	141 (22.5)	120 (23.1)	21 (19.6)	<0.001
30≤eGFR<60 mL/min/1.73 m2	409 (65.3)	344 (66.3)	65 (60.7)
<30 mL/min/1.73 m2	42 (6.7)	22 (4.2)	20 (18.9)
Unknown	34 (5.4)	33 (6.4)	1 (0.9)
Hemoglobin, mg/dL	13.9±1.7	14.03±1.6	13.5±2.2	0.005
<11 mg/dL	31 (5.0)	19 (3.7)	12 (11.2)	0.003
BNP, pg/mL	342.8±44.8	302.9±385.2	482.0±602.4	<0.001
Time in therapeutic range, %	62.5±27.2	63.7±27.6	56.6±24.6	0.015

Categorical and consecutive data are presented as numbers (%) and mean±standard deviation, respectively. AF, atrial fibrillation; BNP, B-type natriuretic peptide; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; HF, heart failure; INR var, international normalized ratio variability; NYHA, New York Heart Association; RAS, renin-angiotensin system; TIA, transient ischemic attack.

Predictors for High INR Variability

The results of univariate models of logistic regression analysis for high INR variability are shown in Table 2A. Using the factors significantly associated with high INR variability, the multivariate logistic regression analysis was performed (Table 2B). In the multivariate model, symptomatic HF (odds ratio [OR]; 3.974, 95% confidence interval [CI]; 2.510–6.292), older age (≥75 years old) (OR; 2.984, 95% CI; 1.844–4.826) and severe renal dysfunction (eGFR <30 mL/min/1.73 m²; OR; 3.918, 95% CI; 1.742–8.813) were identified as independent predictors of high INR variability.

Table 2. Univariate and Multivariate Logistic Regression Analysis Showing the Factors Significantly Associated With High INR Variability

Variables	OR (95% CI)	P value
A. Univariate models
Female	1.256 (0.797–1.979)	0.326
Age ≥75 years	2.967 (1.910–4.610)	<0.001
Body weight <50 kg	1.687 (0.969–2.936)	0.065
Hypertension	1.995 (1.290–3.087)	0.002
Dyslipidemia	2.737 (1.787–4.191)	<0.001
Ischemic heart disease	1.980 (1.129–3.472)	0.017
Valvular heart disease	2.939 (1.823–4.739)	<0.001
Cardiomyopathy	2.007 (1.062–3.790)	0.032
Symptomatic HF (NYHA ≥2)	3.811 (2.477–5.863)	<0.001
2 or more comorbidities*	3.454 (2.250–5.301)	<0.001
Albumin <3.5 mg/dL	3.192 (1.684–6.050)	<0.001
eGFR ≥60 mL/min/1.73 m2	Ref.
30≤eGFR<60 mL/min/1.73 m2	1.080 (0.623–1.842)	0.778
eGFR <30 mL/min/1.73 m2	5.195 (2.423–11.137)	<0.001
eGFR unknown	0.173 (0.022–1.335)	0.092
Hemoglobin <13 mg/dL	2.158 (1.388–3.355)	0.001
B. Multivariate model
Symptomatic HF (NYHA ≥2)	3.974 (2.510–6.292)	<0.001
Age ≥75 years	2.984 (1.844–4.826)	<0.001
eGFR ≥60 mL/min/1.73 m2	Ref.
30≤eGFR<60 mL/min/1.73 m2	1.012 (0.578–1.773)	0.966
eGFR <30 mL/min/1.73 m2	3.918 (1.742–8.813)	0.001
eGFR unknown	0.297 (0.038–2.344)	0.249
C. Multivariate model (INR measurement ≥6)
Symptomatic HF (NYHA ≥2)	3.635 (2.151–6.144)	<0.001
Age ≥75 years	2.270 (1.296–3.974)	0.004
eGFR ≥60 mL/min/1.73 m2	Ref.
30≤eGFR<60 mL/min/1.73 m2	1.455 (0.738–2.867)	0.279
eGFR <30 mL/min/1.73 m2	3.398 (1.267–9.116)	0.015
eGFR unknown**	–	–

*Two or more of following comorbidities: hypertension, diabetes, dyslipidemia, ischemic heart disease, valvular heart disease, cardiomyopathy, previous stroke or TIA, previous hemorrhage, COPD, and chronic kidney disease. **OR for eGFR unknown was not determined for model C because of a lack of numbers. CI, confidence interval; OR, odds ratio. Other abbreviations as in Table 1.

Given that INR variability, by definition, is affected by the intervals of INR measurement, we additively performed a multivariate logistic regression analysis in subjects with relatively frequent INR measurement (INR measurement ≥6 times; excluding the lowest tertile of frequency of INR measurement). Accordingly, we obtained mostly consistent results as compared with those in the total patient cohort: the ORs of symptomatic HF, older age (≥75 years old) and severe renal dysfunction (eGFR <30 mL/min/1.73 m²) were 3.635 (95% CI; 2.151–6.144), 2.270 (95% CI; 1.296–3.974) and 3.398 (95% CI; 1.267–9.116), respectively.

Discussion

In the present study, we determined the predictors of high INR variability in AF patients under warfarin therapy using a single-center cohort in a cardiovascular hospital. The independent predictors of high INR variability in the present study were symptomatic HF, older age (≥75 years old) and severe renal function (estimated GFR <30 mL/min/1.73 m²). The results were mostly consistent when we excluded the patients with infrequent INR measurements.

Predictors of High INR Variability

Although the association between symptomatic HF and poor INR control has been reported previously, proper explanations have not been proposed.^5,23 One possible mechanism may be unstable drug metabolism via liver dysfunction caused by congestive liver and/or renal dysfunction from the use of diuretics by patients with symptomatic HF, although it is not assumed there are so many outpatients in such an unstable state. Similarly, polypharmacy is often seen in patients with HF, especially those with a greater number of comorbidities, which would substantially modify the pharmacokinetics. Indeed, polypharmacy increases bleeding events in NVAF patients under anticoagulant therapy,²⁴ which would be, in part, the result of INR instability.

Older age was associated with high INR variability in our study, suggesting that these patients were at high risk of poor INR control, consistent with previous reports that elderly patients were also associated with low TTR.⁵ These patients would have unstable pharmacokinetics because of low body weight,¹⁸ polypharmacy and/or liver and renal dysfunction.

Severe renal dysfunction was also associated with high INR variability in our study. It is rather difficult to explain the direct association between renal function and high INR variability because warfarin is almost entirely dependent on hepatic elimination. The possible explanation may be the linkage of higher age and estimated GFR, based on the equation itself,²² and accordingly concomitant medications and comorbidities,^25,26 which are both characteristics of elderly patients and/or renal dysfunction.²⁷

It remains a matter of discussion whether INR variability and TTR measure different characteristics of anticoagulation control.⁸ TTR reflects the achievement of appropriate anticoagulation intensity, and not necessarily stability, of the treatment regimen.²⁸ In contrast, INR variability measures the stability, but not intensity of anticoagulation. Therefore, the predictive capabilities of INR variability and TTR for thromboembolism and/or bleeding can be independent,^8,9,17 and in some cases INR variability may be superior to TTR, as reported in a cohort of hospitalized AF patients.^8,9 In patients with non-adherence to the therapeutic regimen, TTR and INR variability would be closely associated with each other, which may partially explain why some of the factors associated with high INR variability were the same as those reported for low TTR.¹⁵

Clinical Implications

The situation for estimating INR control is complex in Japan,^13,14 because the Japanese guidelines recommend different target ranges of INR control according to age (i.e., 2.0–3.0 and 1.6–2.6 for patients <70 and ≥70 years, respectively). Moreover, several reports have suggested that a target INR of 1.6–2.6 for younger NVAF patients may be acceptable based on the results of risk-benefit analysis.²⁹ In this regard, using TTR to evaluate the quality of INR control in NVAF patients is potentially confusing in current Japanese clinical practice. In contrast, it could be advantageous to discuss INR variability, which does not depend on the definition of target INR.

Our results will help physicians to identify patients likely to show unstable warfarin control irrespective of the definition of the TTR. If patients have risk factors for high INR variability, they may require frequent hospital visits to help achieve acceptable INR control.

Study Limitations

First, at present there is no agreed-upon method for calculating INR variability. We selected our method based on the conclusion of its superiority in an earlier study.¹⁶ Similarly, there is no agreement regarding a cutoff value for INR variability that can be used to define stable and unstable anticoagulation. In this study, we determined the cutoff as over than mean+1*SD, which should indicate the borderline of outliers (log INR variability−1.245 in our data was close to the unstable level in a previous study⁹ (log INR variability−1.37)), although the reliability of this cutoff value of INR variability should be assessed in different datasets. Second, as our database consisted of a cohort at a single cardiovascular hospital, selection bias should be taken into consideration. Third, in this analysis, INR variability was calculated by INR measurements between 3 and 12 months after commencement of warfarin administration, which would reflect the quality of INR control only in the early phase. Finally, our dataset did not include information on alcohol abuse or patients with severe liver/renal impairment, which potentially could affect the INR control.

Conclusions

The present study identified clinical factors related to high INR variability in AF patients under warfarin therapy. The determinants of INR variability in AF patients under warfarin therapy could assist Japanese clinicians in identifying patients likely to show unstable warfarin control irrespective of the definition of the target INR range.

Acknowledgments

We thank Shiro Ueda and Nobuko Ueda at Medical Edge Company, Ltd., for assembling the database by the Clinical Study Supporting System and Ineko Hayakawa, Hiroaki Arai, and Hirokazu Aoki for data management and system administration.

Conflict of Interest

S.S. received research funding and remuneration from Nippon Boehringer Ingelheim. T.Y. received research funding from Nippon Boehringer Ingelheim and Daiichi-Sankyo, and remuneration from Nippon Boehringer Ingelheim, Daiichi-Sankyo, Bayer Healthcare, Pfizer, Bristol-Myers Squibb, Eisai and Ono Pharmaceutical.

Funding

This study is partially supported by the Practical Research Project for Life-style Related Diseases including Cardiovascular Diseases and Diabetes Mellitus from Japan Agency for Medical Research and Development, AMED (15656344).

References

• 1.   Singer DE, Chang Y, Fang MC, Borowsky LH, Pomernacki NK, Udaltsova N, et al. The net clinical benefit of warfarin anticoagulation in atrial fibrillation. Ann Intern Med 2009; 151: 297–305.
• 2.   Kakkar AK, Mueller I, Bassand JP, Fitzmaurice DA, Goldhaber SZ, Goto S, et al. Risk profiles and antithrombotic treatment of patients newly diagnosed with atrial fibrillation at risk of stroke: Perspectives from the international, observational, prospective GARFIELD registry. PLoS One 2013; 8: e63479.
• 3.   Huisman MV, Rothman KJ, Paquette M, Teutsch C, Diener HC, Dubner SJ, et al. Antithrombotic treatment patterns in patients with newly diagnosed nonvalvular atrial fibrillation: The GLORIA-AF Registry, phase II. Am J Med 2015; 128: 1306–1313.e1301.
• 4.   Inoue H, Okumura K, Atarashi H, Yamashita T, Origasa H, Kumagai N, et al. Target international normalized ratio values for preventing thromboembolic and hemorrhagic events in Japanese patients with non-valvular atrial fibrillation: Results of the J-RHYTHM Registry. Circ J 2013; 77: 2264–2270.
• 5.   Van Spall HG, Wallentin L, Yusuf S, Eikelboom JW, Nieuwlaat R, Yang S, et al. Variation in warfarin dose adjustment practice is responsible for differences in the quality of anticoagulation control between centers and countries: An analysis of patients receiving warfarin in the randomized evaluation of long-term anticoagulation therapy (RE-LY) trial. Circulation 2012; 126: 2309–2316.
• 6.   Piccini JP, Hellkamp AS, Lokhnygina Y, Patel MR, Harrell FE, Singer DE, et al. Relationship between time in therapeutic range and comparative treatment effect of rivaroxaban and warfarin: Results from the ROCKET AF trial. J Am Heart Assoc 2014; 3: e000521.
• 7.   Amin A, Deitelzweig S, Jing Y, Makenbaeva D, Wiederkehr D, Lin J, et al. Estimation of the impact of warfarin’s time-in-therapeutic range on stroke and major bleeding rates and its influence on the medical cost avoidance associated with novel oral anticoagulant use: Learnings from ARISTOTLE, ROCKET-AF, and RE-LY trials. J Thromb Thrombolysis 2014; 38: 150–159.
• 8.   Lind M, Fahlen M, Kosiborod M, Eliasson B, Oden A. Variability of INR and its relationship with mortality, stroke, bleeding and hospitalisations in patients with atrial fibrillation. Thromb Res 2012; 129: 32–35.
• 9.   Razouki Z, Ozonoff A, Zhao S, Jasuja GK, Rose AJ. Improving quality measurement for anticoagulation: Adding international normalized ratio variability to percent time in therapeutic range. Circ Cardiovasc Qual Outcomes 2014; 7: 664–669.
• 10.   Rosendaal FR, Cannegieter SC, van der Meer FJ, Briet E. A method to determine the optimal intensity of oral anticoagulant therapy. Thromb Haemost 1993; 69: 236–239.
• 11.   Schmitt L, Speckman J, Ansell J. Quality assessment of anticoagulation dose management: Comparative evaluation of measures of time-in-therapeutic range. J Thromb Thrombolysis 2003; 15: 213–216.
• 12.   Rose AJ, Hylek EM, Ozonoff A, Ash AS, Reisman JI, Berlowitz DR. Risk-adjusted percent time in therapeutic range as a quality indicator for outpatient oral anticoagulation: Results of the Veterans Affairs Study to Improve Anticoagulation (VARIA). Circ Cardiovasc Qual Outcomes 2011; 4: 22–29.
• 13.   J-Rhythm Registry Investigators. Determinants of warfarin use and international normalized ratio levels in atrial fibrillation patients in Japan: Subanalysis of the J-RHYTHM Registry. Circ J 2011; 75: 2357–2362.
• 14.   Okumura K, Komatsu T, Yamashita T, Okuyama Y, Harada M, Konta Y, et al. Time in the therapeutic range during warfarin therapy in Japanese patients with non-valvular atrial fibrillation: A multicenter study of its status and infuential factors. Circ J 2011; 75: 2087–2094.
• 15.   Apostolakis S, Sullivan RM, Olshansky B, Lip GY. Factors affecting quality of anticoagulation control among patients with atrial fibrillation on warfarin: The SAMe-TT(2)R(2) score. Chest 2013; 144: 1555–1563.
• 16.   Fihn SD, Gadisseur AA, Pasterkamp E, van der Meer FJ, Breukink-Engbers WG, Geven-Boere LM, et al. Comparison of control and stability of oral anticoagulant therapy using acenocoumarol versus phenprocoumon. Thromb Haemost 2003; 90: 260–266.
• 17.   van Leeuwen Y, Rosendaal FR, Cannegieter SC. Prediction of hemorrhagic and thrombotic events in patients with mechanical heart valve prostheses treated with oral anticoagulants. J Thromb Haemost 2008; 6: 451–456.
• 18.   Kooistra HA, Gebel M, Sahin K, Lensing AW, Meijer K. Independent predictors of poor vitamin K antagonist control in venous thromboembolism patients: Data from the EINSTEIN-DVT and PE studies. Thromb Haemost 2015; 114: 1136–1143.
• 19.   Suzuki S, Yamashita T, Ohtsuka T, Sagara K, Uejima T, Oikawa Y, et al. Prevalence and prognosis of patients with atrial fibrillation in Japan: A prospective cohort of Shinken Database 2004. Circ J 2008; 72: 914–920.
• 20.   Suzuki S, Yamashita T, Otsuka T, Sagara K, Uejima T, Oikawa Y, et al. Recent mortality of Japanese patients with atrial fibrillation in an urban city of Tokyo. J Cardiol 2011; 58: 116–123.
• 21.   Suzuki S, Otsuka T, Sagara K, Semba H, Kano H, Matsuno S, et al. Nine-year trend of anticoagulation use, thromboembolic events, and major bleeding in patients with non-valvular atrial fibrillation: Shinken database analysis. Circ J 2016; 80: 639–649.
• 22.   Matsuo S, Imai E, Horio M, Yasuda Y, Tomita K, Nitta K, et al. Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis 2009; 53: 982–992.
• 23.   Tomita H, Kadokami T, Momii H, Kawamura N, Yoshida M, Inou T, et al. Patient factors against stable control of warfarin therapy for Japanese non-valvular atrial fibrillation patients. Thromb Res 2013; 132: 537–542.
• 24.   Piccini JP, Hellkamp AS, Washam JB, Becker RC, Breithardt G, Berkowitz SD, et al. Polypharmacy and the efficacy and safety of rivaroxaban versus warfarin in the prevention of stroke in patients with nonvalvular atrial fibrillation. Circulation 2016; 133: 352–360.
• 25.   Reinecke H, Brand E, Mesters R, Schabitz WR, Fisher M, Pavenstadt H, et al. Dilemmas in the management of atrial fibrillation in chronic kidney disease. J Am Soc Nephrol 2009; 20: 705–711.
• 26.   Marinigh R, Lane DA, Lip GY. Severe renal impairment and stroke prevention in atrial fibrillation: Implications for thromboprophylaxis and bleeding risk. J Am Coll Cardiol 2011; 57: 1339–1348.
• 27.   Shlipak MG, Stehman-Breen C, Fried LF, Song X, Siscovick D, Fried LP, et al. The presence of frailty in elderly persons with chronic renal insufficiency. Am J Kidney Dis 2004; 43: 861–867.
• 28.   Singer DE, Hellkamp AS, Piccini JP, Mahaffey KW, Lokhnygina Y, Pan G, et al. Impact of global geographic region on time in therapeutic range on warfarin anticoagulant therapy: Data from the ROCKET AF clinical trial. J Am Heart Assoc 2013; 2: e000067.
• 29.   Kodani E, Atarashi H, Inoue H, Okumura K, Yamashita T, Investigators JRR. Target intensity of anticoagulation with warfarin in Japanese patients with valvular atrial fibrillation: Subanalysis of the J-RHYTHM Registry. Circ J 2015; 79: 325–330.