Cardiovascular risk profile and frailty in Japanese outpatients: the Nambu Cohort Study

Sauver JL, Shah P, Vella A, Ryu E, Liu H, Marshall GD, Cerhan JR, Singh D, Summers RL. Predictors of Metformin Failure: Repurposing Electronic Health Record Data to Identify High-Risk Patients

CONTEXT

Metformin is the first-line drug for treating diabetes but has a high failure rate.

OBJECTIVE

To identify demographic and clinical factors available in the electronic health record (EHR) that predict metformin failure.

METHODS

A cohort of patients with at least 1 abnormal diabetes screening test that initiated metformin was identified at 3 sites (Arizona, Mississippi, and Minnesota). We identified 22 047 metformin initiators (48% female, mean age of 57 ± 14 years) including 2141 African Americans, 440 Asians, 962 Other/Multiracial, 1539 Hispanics, and 16 764 non-Hispanic White people. We defined metformin failure as either the lack of a target glycated hemoglobin (HbA1c) (<7%) within 18 months of index or the start of dual therapy. We used tree-based extreme gradient boosting (XGBoost) models to assess overall risk prediction performance and relative contribution of individual factors when using EHR data for risk of metformin failure.

RESULTS

In this large diverse population, we observed a high rate of metformin failure (43%). The XGBoost model that included baseline HbA1c, age, sex, and race/ethnicity corresponded to high discrimination performance (C-index of 0.731; 95% CI 0.722, 0.740) for risk of metformin failure. Baseline HbA1c corresponded to the largest feature performance with higher levels associated with metformin failure. The addition of other clinical factors improved model performance (0.745; 95% CI 0.737, 0.754, P < .0001).

CONCLUSION

Baseline HbA1c was the strongest predictor of metformin failure and additional factors substantially improved performance suggesting that routinely available clinical data could be used to identify patients at high risk of metformin failure who might benefit from closer monitoring and earlier treatment intensification.

Baseline Lipoprotein(a) Levels and Long-Term Cardiovascular Outcomes After Acute Myocardial Infarction

BACKGROUND

Lipoprotein(a) is a known independent risk factor for atherosclerotic cardiovascular disease. However, the prognostic impact of the baseline lipoprotein(a) levels on long-term clinical outcomes among patients with acute myocardial infarction remain unclear.

METHODS

We analyzed 1,908 patients with acute myocardial infarction from November 2011 to October 2015 from a single center in Korea. They were divided into 3 groups according to their baseline lipoprotein(a) levels: groups I (< 30 mg/dL, n = 1,388), II (30-49 mg/dL, n = 263), and III (≥50 mg/dL, n = 257). Three-point major adverse cardiovascular events (a composite of nonfatal myocardial infarction, nonfatal stroke, and cardiac death) at 3 years were compared among the 3 groups.

RESULTS

The patients were followed for 1094.0 (interquartile range, 1,033.8-1,095.0) days, during which a total of 326 (17.1%) three-point major adverse cardiovascular events occurred. Group III had higher rates of three-point major adverse cardiovascular events compared with Group I (23.0% vs. 15.7%; log-rank P = 0.009). In the subgroup analysis, group III had higher rates of three-point major adverse cardiovascular events compared with group I in patients with non-ST-segment elevation myocardial infarction (27.0% vs. 17.1%; log-rank P = 0.006), but not in patients with ST-segment elevation myocardial infarction (14.4% vs. 13.3%; log-rank P = 0.597). However, in multivariable Cox time-to-event models, baseline lipoprotein(a) levels were not associated with an increased incidence of three-point major adverse cardiovascular events, regardless of the type of acute myocardial infarction. Sensitivity analyses in diverse subgroups showed similar findings to those of the main analysis.

CONCLUSION

Baseline lipoprotein(a) levels in Korean patients with acute myocardial infarction were not independently associated with increased major adverse cardiovascular events at 3 years.

Current topics of frailty in association with hypertension and other medical conditions

Frailty is a state of increased vulnerability to stress resulting from physiological decline associated with aging. Topics of hypertension management and its association with frailty and cognitive function, recent studies of coronavirus disease 2019 infection (COVID-19) in elderly is discussed in this narrative review. While various guidelines for hypertension recommend that frailty is taken into account in treatment decisions, specific assessment tools and clinical decision criteria have not been explicitly established. Hypertension is prevalent in frail individuals, although a direct association has not been reported. Therefore, optimal blood pressure (BP) control is critical for managing cardiovascular risk reduction and preserving quality of life in frail hypertensive patients. BP typically decreases in later life or situations in which patients are dependent on nursing care. Mortality is reported to be high among frail patients with lower BP, raising questions about appropriate BP targets for frail patients. Cognitive decline is one of the domains of frailty, and is associated with a loss of autonomy, lack of self-management, and compromised quality of life. It remains to be clarified whether antihypertensive treatment is beneficial for cognitive function especially in older individuals. Increased severity and mortality of COVID-19 infection has been reported in older people. Clinical manifestations and biomarkers particular to older patients, and lifestyle changes including social isolation during the COVID-19 pandemic is reported. From the knowledge from recent literatures, future perspectives for holistic approach and management of frail older people is addressed.

Prevalence and Risk Factors of Frailty in Stroke Patients: A Meta-Analysis and Systematic Review

OBJECTIVES

Summarize the existing evidence regarding the prevalence and risk factors of frailty in stroke patients.

DESIGN

A meta-analysis and systematic review.

PARTICIPANTS

Stroke patients in hospitals or communities.

METHODS

We undertook a systematic review and meta-analysis using articles available in 8 databases, including PubMed, The Cochrane Library, Web of Science, Embase, Chinese Biomedical Database (CBM), China National Knowledge Infrastructure Database (CNKI), Wanfang Database, and Weipu Database (VIP) from January 1990 to April 2022. Studies were quality rated using the Newcastle-Ottawa Scale and Agency for Healthcare Research and Quality tool.

RESULTS

A total of 24 studies involving 30,423 participants were identified. The prevalence of frailty and pre-frailty in stroke patients was 27% (95%CI: 0.23-0.31) and 47.9% (95%CI: 0.43-0.53). Female gender (OR = 1.76, 95%CI: 1.63-1.91), advanced age (MD = 6.73, 95%CI: 3.55-9.91), diabetes (OR = 1.34, 95%CI: 1.06-1.69), hyperlipidemia (OR = 1.46, 95%CI: 1.04-2.04), atrial fibrillation (OR = 1.36, 95%CI: 1.01-1.82), National Institutes of Stroke Scale (NIHSS) admission scores (MD = 2.27, 95%CI: 1.72-2.81) were risk factors of frailty in stroke patients.

CONCLUSIONS

Frailty was more prevalent in stroke patients. Female gender, advanced age, diabetes, hyperlipidemia, atrial fibrillation, and National Institutes of Stroke Scale (NIHSS) admission scores were identified as risk factors for frailty in stroke patients. In the future, medical staff should pay attention to the early screening of frailty in high-risk groups and provide information on its prevention.

Impact of Low Baseline Low-Density Lipoprotein Cholesterol on Long-Term Postdischarge Cardiovascular Outcomes in Patients With Acute Myocardial Infarction

Background

Real‐world data on low baseline low‐density lipoprotein cholesterol (LDL‐C) levels and long‐term postdischarge cardiovascular outcomes in patients with acute coronary syndrome are limited.

Methods and Results

Of the 10 719 patients enrolled in the Korean registry of acute myocardial infarction between January 2004 and August 2014, we identified 5532 patients who were event free from death, recurrent myocardial infarction, or stroke during the in‐hospital period after successful percutaneous coronary intervention. The co–primary outcomes were 3‐point major adverse cardiovascular events (a composite of nonfatal stroke, nonfatal myocardial infarction, and cardiovascular death) and cardiovascular death at 5 years. Of 5532 patients with acute myocardial infarction (mean age, 62.1±12.8 years; 75.0% men), 446 cardiovascular deaths (8.1%) and 695 three‐point major adverse cardiovascular events (12.6%) occurred at 5 years. In the continuous analysis of LDL‐C, the risk of cardiovascular events increased steeply as LDL‐C levels decreased from 100 mg/dL. For categorical analysis of LDL‐C (<70, 70–99, and ≥100 mg/dL), as LDL‐C levels decreased, clinical outcomes worsened (237/3759 [6.3%] in LDL‐C ≥100 mg/dL versus 123/1291 [9.5%] in LDL‐C 70–99 mg/dL versus 86/482 [17.8%] in LDL‐C <70 mg/dL for cardiovascular death; P‐trend<0.001; and 417/3759 [11.1%] in LDL‐C ≥100 mg/dL versus 172/1291 [13.3%] in LDL‐C 70–99 mg/dL versus 106/482 [22.2%] in LDL‐C <70 mg/dL for 3‐point major adverse cardiovascular event; P‐trend<0.001). In a Cox time‐to‐event multivariable model with LDL‐C levels ≥100 mg/dL as the reference, the baseline LDL‐C level <70 mg/dL was independently associated with an increased incidence of cardiovascular death (adjusted hazard ratio, 1.68 [95% CI, 1.30–2.17]) and 3‐point major adverse cardiovascular event (adjusted hazard ratio, 1.37 [95% CI, 1.10–1.71]).

Conclusions

In this Korean acute myocardial infarction registry, the baseline LDL‐C level <70 mg/dL was significantly associated with an increased incidence of long‐term cardiovascular events after discharge. (COREA [Cardiovascular Risk and Identification of Potential High‐Risk Population]‐Acute Myocardial Infarction Registry; NCT02806102).

Registration

URL: https://www.clinicaltrials.gov/; Unique identifier: NCT02806102.

Risk factors for frailty in elderly Japanese people who received Ningen Dock: a cross-sectional study

Background

Frailty is a clinical condition characterized by increased vulnerability to adverse health outcomes. Elderly people are screened for frailty as part of preventative care. However, the risk factors for frailty among older adults who undergo Ningen Dock, a comprehensive medical checkup, remain unclear. Thus, this cross-sectional study aims to identify the risk factors for frailty in older adults who received Ningen Dock. The study included 372 participants over 65 years of age who underwent Ningen Dock at the Health Care Center of Gunma Chuo Hospital between April 2019 and March 2020. Frailty was defined using the Kihon Checklist, a basic checklist. Clinical variables were obtained from Ningen Dock records, a vascular function test, and a questionnaire on medication, among others. Multivariate ordinal logistic regression models were used to assess risk factors.

Results

Prevalence for frailty and pre-frailty was 12.6% and 26.6%, respectively. The mean age of participants was 72.0 ± 5.1 years old, and 43.5% were female. Compared with systolic blood pressure (SBP) ≥ 130 mmHg, the odds ratios for the 100–129 mmHg and < 100 mmHg groups were 2.43 (P = 0.020) and 8.95 (P <0.001). The odds ratio for the ≥ 7 medications group medications was 3.64 (P = 0.003) compared to 0–2 medications. Compared with serum iron ≥ 126 μg/dL, the odds ratio for ≤ 85 μg/dL was 2.91 (P = 0.002). The odds ratio for total bilirubin ≤ 0.6 mg/dL was 2.49 (P = 0.011) compared with > 0.6 mg/dL. Compared with an exercise habit of ≥ 4 metabolic equivalents (METs), the odds ratio for < 2 METs/week was 2.45 (P < 0.001). The odds ratio for the cardio-ankle vascular index (CAVI) ≥ 9 group was 1.84 (P = 0.020) compared to < 9.

Conclusions

In older adults who received Ningen Dock, SBP < 100 mmHg, medications ≥ 7, serum iron ≤ 85 μg/dL, total bilirubin ≤ 0.6 mg/dL, exercise habits < 2 METs, and CAVI ≥ 9 were associated with frailty.

Association between cardiovascular risk factors and left ventricular strain distribution in patients without previous cardiovascular disease

BACKGROUND

Some cardiovascular (CV) risk factors, such as hypertension and diabetes mellitus, have been reported to reduce left ventricular (LV) longitudinal strain (LS) even in patients with preserved LV ejection fraction. We hypothesized that multiple CV risk factors might cause changes in myocardial strain. Our study aimed to assess the association between multiple CV risk factors and strain in patients without previous CV disease (CVD).

METHODS

We retrospectively evaluated 137 patients without CVD, who underwent echocardiography at our institution between May 2017 and February 2020. They were divided into four groups based on the number of risk factors (group 0: no risk factor, group 1: one risk factor, group 2: two risk factors, and groups 3: three or four risk factors). Risk factors were hypertension, dyslipidemia, diabetes mellitus, and chronic kidney disease. Absolute values of global LS (GLS) and relative apical LS ratio (RALSR) defined using the equation: average apical LS/(average basal LS + average mid LS) and was used as a marker of strain distribution.

RESULTS

Out of 137 patients, group 0 had 35 patients, group 1 had 35 patients, group 2 had 32 patients, and group 3 had 35 patients. GLS was 22.4 ± 2.0%, 21.7 ± 2.1%, 21.3 ± 1.8%, 20.7 ± 2.2%, and RALSR was 0.64 ± 0.06, 0.66 ± 0.06, 0.68 ± 0.08, 0.69 ± 0.07 in groups 0-3, respectively. The one-way ANOVA detected significant differences between groups in GLS (p = 0.005) and RALSR (p = 0.037), respectively. Group 3 had a significantly lower GLS and higher RALSR than group 0 (p < 0.05).

CONCLUSION

In patients without previous CVD, LS decreased especially from the basal segment as the number of cardiovascular risks increased. The segmental LS may be markers of occult LV dysfunction in patients with CV risk factors.

Comparative accuracies of automated and manual office blood pressure measurements in a Chinese population

We aimed to assess the difference in the accuracy of readings from automated office blood pressure machines with each other or with manual office blood pressure measurements in Chinese individuals. We collected awake 48-h ambulatory blood pressure monitoring, two automated office blood pressure device (BpTRU and WatchBP) readings, and manual office blood pressure measurements in Chinese patients (n = 135) with hypertension in a randomized sequence. Differences were compared using paired t-tests and Bland-Altman plots. The sensitivity and specificity of the techniques for detecting elevated blood pressure were calculated using awake ambulatory blood pressure monitoring as the reference standard. The WatchBP device's and awake ambulatory blood pressure readings were similar. The BpTRU device provided significantly lower mean systolic (P < 0.001) and diastolic (P < 0.001) blood pressure readings, while manual office BP provided significantly higher mean systolic (P = 0.008) and diastolic (P < 0.001) blood pressure readings than the awake automated office blood pressure readings. Automated and manual office blood pressure measurements showed similar sensitivity, specificity, and 95% limits of agreement as based on Bland-Altman plots. The mean systolic (P < 0.001) and diastolic (P < 0.02) blood pressure readings of WatchBP and BpTRU differed, and their diagnostic performances were not superior than those of manual office blood pressure measurements in Chinese patients. Therefore, automated office blood pressure measurements cannot be routinely recommended for Chinese individuals in clinical practice. More studies are needed to confirm these results.

The Role of Traditional Obesity Parameters in Predicting Frailty among Coronary Artery Disease Patients Undergoing Cardiac Catheterization

BACKGROUND

Obesity has been reported to be associated with frailty and coronary artery disease (CAD).

OBJECTIVE

The present study aimed to investigate the role of the seven traditional obesity parameters body mass index (BMI), waist-height ratio (WHtR), waist-hip ratio (WHR), body adiposity index (BAI), body shape index (BSI), waist circumference (WC), and hip circumference (HC) in the prediction of frailty among CAD patients undergoing cardiac catheterization.

DESIGN

A secondary data analysis was conducted. Setting. Three main hospitals were located at the northern and middle regions of Jordan. Participants. 220 hospitalized patients undergoing cardiac catheterization were recruited. Measurements. The traditional obesity parameters were measured using an anthropometric tape and weight scale and frailty was measured using the Tilburg Frailty Indicator (TFI). Data were analyzed using bivariate Pearson's correlation and forward linear regression analysis.

RESULTS

Total cholesterol, HC, triglycerides, age, random blood sugar, and WC had significant positive associations with and were predictors of frailty (p < 0.05). The model of the seven predictors explained 32.4% of the variance in frailty (p = 0.02).

CONCLUSION

The incidence of frailty can be predicted by the increase in total cholesterol, HC, triglycerides, age, random blood sugar, and WC. The results of this study may help healthcare providers, including nurses, to identify the factors that could lead to frailty among CAD patients undergoing cardiac catheterization.

Blood pressure, frailty status, and all-cause mortality in elderly hypertensives; The Nambu Cohort Study

Antihypertensive therapy is pivotal for reducing cardiovascular events. The 2019 Guidelines for the Management of Hypertension set a target blood pressure (BP) of <140/90 mmHg for persons older than 75 years of age. Optimal BP levels for older persons with frailty, however, are controversial because evidence for the relationship between BP level and prognosis by frailty status is limited. Here, we evaluated the relationship between systolic BP and frailty status with all-cause mortality in ambulatory older hypertensive patients using data from the Nambu Cohort study. A total of 535 patients (age 78 [70-84] years, 51% men, 37% with frailty) were prospectively followed for a mean duration of 41 (34-43) months. During the follow-up period, 49 patients died. Mortality rates stratified by systolic BP and frailty status were lowest in patients with systolic BP < 140 mmHg and non-frailty, followed by those with systolic BP ≥ 140 mmHg and non-frailty. Patients with frailty had the highest mortality regardless of the BP level. The adjusted hazard ratios (95% confidence intervals) of each category for all-cause mortality were as follows: ≥140 mmHg/Non-frailty 3.19 (1.12-11.40), <140 mmHg/Frailty 4.72 (1.67-16.90), and ≥140 mmHg/Frailty 3.56 (1.16-13.40) compared with <140 mmHg/Non-frailty as a reference. These results indicated that frail patients have a poor prognosis regardless of their BP levels. Non-frail patients, however, with systolic BP levels <140 mmHg had a better prognosis. Frailty may be a marker to differentiate patients who are likely to gain benefit from antihypertensive medication among older hypertensives.

A Simple Risk Score Based on Routine Clinical Parameters Can Predict Frailty in Hospitalized Heart Failure Patients

Frailty syndrome (FS) has recently attracted attention as one of the major predictors of heart failure (HF) course severity. We aimed to develop a simple tool for predicting frailty in hospitalized HF patients using routine clinical parameters. A total of 153 hospitalized patients diagnosed with heart failure with reduced ejection fraction (HFrEF) were included in the study. Presence of FS was assessed with the SHARE-FI questionnaire. Clinical and biochemical parameters were collected. Using ROC curves and logistic regression analysis, a model predicting FS presence was developed and tested. Proposed model includes five variables with following cut-off values (1 point for each variable): age > 50 years, systolic pressure on admission < 110 mmHg, total cholesterol < 4.85 mmol/L, bilirubin ≥ 15.5 mmol/L, and alanine aminotransferase ≤ 34 U/L. Receiving 5 points was considered a high risk of FS with positive and negative predictive values (NPV), 83% and 72%, respectively, and specificity of 97%. Awarding 2 points or less ruled out FS in the studied group with negative predictive value 94%. The presented novel, simple score predicts FS in HFrEF patients with routine clinical parameters and has good positive and negative predictive values.

The association between frailty and chronic kidney disease; cross-sectional analysis of the Nambu Cohort Study

BACKGROUND

Both frailty and chronic kidney disease (CKD) increase with age and share many similarities. Many studies have demonstrated an association between frailty and chronic kidney disease (CKD), but an association with dipstick proteinuria is limited.

METHODS

This is the cross-sectional analysis of the Nambu Cohort Study at the beginning of observation. Frailty was diagnosed using Kihon Checklist. Logistic analysis was used to evaluate the association between frailty and CKD or dipstick proteinuria.

RESULTS

Among a total of 630 outpatients [age, 78 (70-84) years, men, 50%], the prevalence of patients with pre-frailty and frailty was 32% and 40%, respectively. The proportion of patients with pre-frailty and frailty increased with decreasing estimated glomerular filtration rate (eGFR) and increasing dipstick proteinuria levels. The odds ratios (95% confidence intervals) for CKD stage of 60 < eGFR ≤ 45 ml/min/1.73 m², and 45 ml/min/1.73 m² < eGFR for frailty was 0.87 (0.56-1.35) and 2.54 (1.46-4.53), respectively, compared with non-CKD as a reference. Furthermore, the odds ratios for the frailty of dipstick proteinuria with ± and + or over were 1.36 (0.88-2.09) and 1.78 (1.00-3.17), respectively, when dipstick proteinuria-was used as a reference. Moreover, the combination of eGFR and dipstick proteinuria levels increased the odds ratio for pre-frailty and frailty.

CONCLUSION

Elderly patients with CKD had a higher prevalence of pre-frailty and frailty. By adding urinary protein information to eGFR, the link between CKD and frailty becomes even more robust.

Silent Myocardial Infarction and Long-Term Risk of Frailty: The Atherosclerosis Risk in Communities Study

Background

Silent myocardial infarction (SMI) accounts for more than half of all MIs, and common risk factors and pathophysiological pathways coexist between SMI and frailty. The risk of frailty among patients with SMI is not well established. This study aimed to examine the association between SMI and frailty.

Methods and Results

This analysis included data from the Atherosclerosis Risk in Communities study. Patients without MI at baseline were eligible for inclusion. SMI was defined as electrocardiographic evidence of MI without clinical MI (CMI) after the baseline and until the fourth visit. Frailty was assessed during the fifth visit. A total of 4953 participants were included with an average age of 52.2±5.1 years. Among these participants, 2.7% (n=135) developed SMI, and 2.9% (n=146) developed CMI. After a median follow-up time of 14.7 (14.0–15.3) years, 6.7% (n=336) of the participants developed frailty. Patients with SMI and CMI were significantly more likely to become frail than those without MI (15.6% vs 6.2%, P<0.001 and 16.4% vs 6.2%, P<0.001, respectively). After adjusting for confounders, SMI and CMI were found to be independent predictors of frailty (odds ratio [OR]=2.243, 95% confidence interval [CI]=1.307–3.850, P=0.003 and OR=2.164, 95% CI=1.259–3.721, P=0.005, respectively). The association was consistent among the subgroups of age, sex, race, diabetes, and hypertension.

Conclusion

In conclusion, both SMI and CMI were found to be associated with a higher risk of frailty. Future studies are needed to confirm the beneficial effects of screening for SMI as well as to implement standardized preventive treatment to reduce the risk of frailty.

Clinical Trial Registration

URL: https://www.clinicaltrials.gov; Unique identifier: NCT00005131.

Cardiometabolic diseases, frailty, and healthcare utilization and expenditure in community-dwelling Chinese older adults

This study investigated associations between cardiometabolic diseases, frailty, and healthcare utilization and expenditure among Chinese older adults. The participants were 5204 community-dwelling adults aged at least 60 years from the China Health and Retirement Longitudinal Study. Five cardiometabolic diseases were assessed including hypertension, dyslipidemia, diabetes, cardiac diseases and stroke. Frailty status was based on five criteria: slowness, weakness, exhaustion, inactivity, and shrinking. Participants were deemed frailty if they met at least three criteria. As the number of cardiometabolic diseases increased, so did the prevalence of frailty, and the proportion of healthcare utilization, including outpatient visit and inpatient visit. Moreover, the total healthcare expenditure and the odds of catastrophic health expenditure were increased with the number of cardiometabolic disorders. After adjusting for covariates, cardiometabolic diseases were positively associated with higher odds of frailty, incurring outpatient and inpatient visit. And individuals with 2 or more cardiometabolic diseases had a higher odds of catastrophic health expenditure than persons with non-cardiometabolic disease. Participants who were frailty were more likely to report higher odds of healthcare utilization. These findings suggest that both cardiometabolic diseases and frailty assessment may improve identification of older adults likely to require costly, extensive healthcare.

Cause of Death and Associated Factors in Elderly Patients With Atrial Fibrillation - Long-Term Retrospective Study

Background: Oral anticoagulant (OAC) therapy reduces the risk of stroke in patients with atrial fibrillation (AF). This study elucidated the causes of death and related factors in elderly Japanese AF patients.

Methods and Results: Over a median (interquartile range [IQR]) follow-up period of 46 (20–76) months, there were 171 all-cause deaths (28% cardiovascular, 46% non-cardiovascular, and 26% unknown causes) among 389 AF patients (median [IQR] age 80 [74–85] years; CHAD₂DS₂-VASc score 5 [4–6]). Cox regression analysis indicated that diabetes was associated with an increase in all-cause death (hazard ratio [HR] 1.48; 95% confidence interval [CI] 1.02–2.13), whereas hypercholesterolemia (HR 0.53; 95% CI 0.35–0.79), pre-existing heart failure (HR 0.67; 95% CI 0.48–0.95), and OAC use (HR 0.62; 95% CI 0.44–0.88) were associated with reductions in all-cause death. Pre-existing heart failure was associated with both cardiovascular (HR 3.03; 95% CI 1.33–8.20) and non-cardiovascular (HR 0.44; 95% CI 0.30–0.65) deaths, in opposite directions. OAC use was associated with a reduction in cardiovascular death (HR 0.34, 95% CI 0.17–0.69). The predominance of non-cardiovascular death and death-related factors were equivalent regardless of when observations started (before 2009 or in 2009 and later).

Conclusions: The predominant cause of death in elderly Japanese AF patients was non-cardiovascular. Distinct clinical factors were associated with cardiovascular and non-cardiovascular death.