Mid-Life Cardiorespiratory Fitness, Obesity, and Risk of Atrial Fibrillation

Background Lower cardiorespiratory fitness (CRF) and higher body mass index (BMI) are associated with a higher risk of myocardial infarction and heart failure. However, the independent contribution of these lifestyle factors to the risk of atrial fibrillation (AF) is less well established. Objectives The purpose of this study was to evaluation the association between midlife CRF, BMI, and risk of AF in older age. Methods This study included 18,493 participants without AF who underwent assessment of CRF (estimated using the maximal treadmill time) and BMI in middle age and had Medicare coverage after the age of 65 years. The association among midlife CRF, BMI, and risk of AF was assessed by fitting a proportional hazards intensity model to the failure time data with adjustment for potential confounders. The association between changes in CRF and BMI in middle age and the risk of AF was also assessed in the subset of participants with repeat CRF assessments. Results Among 18,493 participants (79% men), a higher midlife BMI was significantly associated with a higher risk of AF independent of CRF levels and other potential confounders (hazard ratio per 1-kg/m2: 1.05; 95% confidence interval: 1.03-1.06). Lower midlife CRF was also associated with higher risk of AF (hazard ratio per 1 MET higher CRF: 0.98; 95% confidence interval: 0.96-0.99). However, this association was attenuated and not significant after further adjustment for BMI. Change in CRF on follow-up was also not associated with the risk of AF after adjustment for other confounders. Conclusions The association between low fitness and AF was primarily driven by differences in BMI. In contrast, obesity was independently associated with excess AF risk.

A trial fibrillation (AF) is the most common car- diac arrhythmia encountered in clinical practice affecting up to 5.1 million adults in the United States.AF burden is expected to double over the next 20 years. 1 Furthermore, AF is predominantly a disease of the elderly, with the prevalence of AF increasing by 5% annually among individuals aged $65 years. 1 AF is associated with a high comorbidity burden, health care costs, and excess mortality risk. 2,3As the incidence of AF continues to rise, 1 effective preventive strategies to reduce the burden of AF in the aging population are needed.
8][9] However, the association between physical activity (PA) and incident AF has been inconsistent [10][11][12][13][14][15][16] possibly due to heterogeneity in selfreporting PA measurements, which may bias study findings toward no association.3][24][25][26][27][28][29] Moreover, few studies have assessed the independent contribution of BMI and CRF to the risk of AF.Furthermore, the association of changes in exercise capacity, BMI, and AF risk is not well established.Accordingly, we evaluated the association between CRF and BMI measures and changes in midlife and the risk of AF after the age of 65 years among participants in the Cooper Center Longitudinal Study (CCLS).

METHODS STUDY POPULATION.
The CCLS is an ongoing study that follows up patients of a preventive health clinic in Dallas, Texas (Cooper Clinic).The details of the study design and protocol have been previously well described. 30,31The majority of participants are male, Caucasian, and from the middle and upper socioeconomic levels.Patients are able to self-refer to the clinic or may be referred by their primary care physician or their employer.The initial cohort included 23,194  CCLS CLINICAL EXAMINATION.1][32] A comprehensive clinical exam was conducted by a physician, and complete medical history was taken, in which the participants selfreported personal and family medical histories.
Standard laboratory tests were performed on fasting venous blood, including serum glucose, total serum cholesterol, and serum triglycerides, via standardized and automated techniques.Additionally, resting blood pressures were collected using a mercury sphygmomanometer.BMI was calculated for each participant using height and weight values.

CARDIORESPIRATORY FITNESS ASSESSMENT:
MODIFIED BALKE PROTOCOL.CRF was measured with a symptom-limited maximal treadmill exercise test using a modified-Balke protocol, which has been described previously. 30,31,33The CCLS used the  Pandey et al Fitness, Obesity, and Risk of Atrial Fibrillation J U N E 2 0 2 2 : 1 0 0 0 4 0 correlated using this protocol in both women and men (r for women ¼ 0.94, r for men ¼ 0.92). 34,35dividual participant's treadmill times were categorized into age-and sex-specific categories based on normative treadmill performance data according to standard approaches. 30,31The treadmill time thresholds for CRF quintiles across the age and sex strata developed in the CCLS have been published previously and reported in the Supplemental Table 1. 36,37The categories were then consolidated into mutually exclusive groups: "low fitness," comprised of participants in the first category, "moderate fitness," comprised of participants in the second and third categories, and "high fitness," comprised of participants in the fourth and fifth categories.The use of the Balke protocol allows for the estimation of the participants' fitness levels in units of metabolic equivalents (METs) using proven and established regression equations.groups were more commonly obese and had a higher burden of traditional cardiovascular (CV) risk factors, including higher blood pressure, higher prevalence of smoking, and diabetes in the overall cohort (Table 1).
A similar pattern of clinical characteristics was also observed across the midlife CRF strata in a sexstratified comparison among men and women separately (Supplemental Tables 2 and 3).1).In adjusted analyses, higher midlife CRF was associated with a lower risk of AF after adjustment for age, sex, and exam year (Table 2, model 1).However, this association was attenuated and no longer significant after further adjustment for BMI and other risk factors (Table 2, Pandey et al Fitness, Obesity, and Risk of Atrial Fibrillation similar pattern of association was observed using continuous measures of CRF in the adjusted models (Table 2).In contrast, higher BMI was significantly associated with a higher risk of AF independent of baseline CRF and other baseline risk factors (hazard ratio [HR] per 1 kg/m 2 : 1.05; 95% CI:  3).

DISCUSSION
The present study demonstrated several important findings related to later-life AF and midlife CRF and BMI.First, lower CRF in midlife was associated with a higher risk of AF in later life.However, this association was primarily driven by differences in BMI and attenuated after adjustment for the same.Second, higher BMI in midlife was significantly associated with a higher risk of AF after the age of 65 years, independent of other midlife risk factors and CRF.Previous studies have evaluated the association between CRF and the risk of AF with inconsistent findings.Some studies from European cohorts have demonstrated a U-shaped association between CRF and the risk of AF. 22,27,28 In a Finnish cohort study, Khan et al 27 observed the lowest risk of AF among individuals with CRF levels between 6 and 9 METs and an increased risk at the lower and higher ends of CRF distribution.Similarly, in a cohort of Swedish men, a U-shaped association was noted between CRF in young adulthood and risk of AF in middle age. 22In contrast, studies from the United States in cohorts of older individuals who underwent clinically indicated stress tests have demonstrated a linear, inverse association between CRF levels and risk of AF. 23,24,26,44 In the present study, we also observed that a higher midlife CRF was associated with a lower risk of AF.
However, this association was largely driven by differences in BMI and was attenuated after adjustment for the same.
The discordant findings between our study and others may be related to study population and design differences.For example, prior studies from US cohorts were limited by the referral nature of the study population, 24,26,44 the use of nonexercise test-based estimated measures of CRF, 23 or inadequate adjustment for confounders such as BMI. 23,44Furthermore, most U.S. cohort studies evaluated the association between CRF assessed in older age and downstream risk of AF over the next few years.Thus, the observed associations between higher CRF and lower risk of AF in these studies are also susceptible to reverse causation.In contrast, our study was conducted among community-based, relatively healthy individuals who underwent symptom-limited maximal treadmill test in midlife and had AF outcomes assessed decades later, minimizing the potential for referral bias and reverse causation bias.Our study   Our study also evaluated the association between short-term longitudinal changes in CRF and the risk of AF in later life.We observed that these changes in CRF and BMI were not associated with the risk of AF.Consistent with our observation, Khan et al 27 also did not observe a significant association between changes in CRF and risk of AF in a cohort of The association between CRF and the risk of AF noted in this study is different from that observed for CV outcomes.Previous studies from the CCLS and other cohorts have demonstrated that a higher CRF is associated with a significantly lower risk of CVD, including myocardial infarction, stroke, and HF. 19,21,38,45,46Furthermore, improvements in CRF have been associated with a lower risk of CVD, such as HF. 20,21In the present study, we observed that the higher CRF related to lower risk of AF was largely driven by BMI.The differences in how CRF may influence the risk of AF vs other CV conditions may be related to the effect of exercise and higher CRF on the cardiac structure.Specifically, left atrial (LA) dilation is considered 1 of the key cardiac substrates that underlie the development of AF.8][49][50][51][52][53] Thus, the favorable downstream effects of lowering the CV risk factor burden in preventing AF may be counterbalanced by dilating LA in high-fit individuals.
In contrast with CRF, we observed a significant association between BMI in midlife and risk of AF in older age, independent of CRF and other risk factors.
Several prior studies have demonstrated that obesity The association of low midlife fitness with higher risk of incident atrial fibrillation (AF) is driven by higher burden of obesity and cardiovascular (CV) risk factors in less fit individuals.In contrast, the association between obesity in midlife and risk of AF is independent of fitness levels and CV risk factor burden.BMI ¼ body mass index.
Pandey et al Fitness, Obesity, and Risk of Atrial Fibrillation J U N E 2 0 2 2 : 1 0 0 0 4 0 is an independent risk factor for AF. 7,9,14,54However, unlike the present study, most previous studies did not account for the contribution of CRF to the association between BMI and AF.The findings from the present study highlight the primacy of obesity over low CRF for the development of AF in older individuals.
Obesity may promote the development of AF via systemic and local pathways.Excess adipose tissue alters the metabolic, neurohormonal, and inflammatory milieu of the CV system, resulting in volume expansion, myocardial fibrosis, and autonomic dysfunction. 55Furthermore, epicardial and pericardial fat accumulation may induce local changes to the myocardium leading to structural abnormalities such as LA enlargement. 55Consistently, LA size has been shown to attenuate the obesity-associated risk of AF in epidemiological studies. 9r observations may have implications for the primary prevention of AF.The present study findings suggest that a higher BMI in middle age may be a more important and independent risk factor for AF than CRF.Prior studies with intensive lifestyle interventions targeting CRF improvement and modest weight loss have failed to demonstrate a significant reduction in the risk of AF. 56 In contrast, aggressive weight-loss strategies, such as bariatric surgery, have been shown to lower the AF risk significantly. 57Such aggressive weight-loss strategies need to be tested in prospective, randomized controlled trials to determine the optimal approaches for the prevention of AF.
STUDY LIMITATIONS.Our study is not without limitations.First, the participants of the CCLS were predominantly men of the White race and had high income and education levels.[40][41] Fourth, CRF was not assessed using the gold standard of directly measured peak oxygen uptake in the CCLS.However, prior reports have demonstrated a strong correlation between directly measured peak oxygen uptake and treadmill speed-and grade-based CRF levels (r for men ¼ 0.92, r for women ¼ 0.94).
Fifth, echocardiographic data are not available at the baseline visit for most study participants, limiting our ability to evaluate how LA size and function may modify the association of BMI and CRF with the risk of AF.Finally, given the observational design of our study, results may be susceptible to residual confounding.

CONCLUSIONS
Among participants in the CCLS, baseline CRF and improvement in CRF in midlife were not associated with incident AF after accounting for differences in BMI.In contrast, a higher BMI in midlife was associated with a higher risk of AF, independent of AF risk factors and CRF.These findings suggest that obesity, but not low CRF, in midlife may be an important driver of the higher burden of AF in older individuals.
ACKNOWLEDGMENTS The authors thank the participants, staff, and investigators of the Cooper Center for Longitudinal study.

FUNDING SUPPORT AND AUTHOR DISCLOSURES
CCLS participants with a qualifying midlife exam who later entered a surveillance period of Medicare administrative claims data from 1999 to 2009.Qualifying exams occurred before the surveillance period and had complete data on required covariates, including BMI, systolic blood pressure, cholesterol, fasting blood glucose, and a modified-Balke treadmill exercise test.Exclusions consisted of 338 participants with a history of myocardial infarction or stroke, 500 participants who entered Medicare before the age of 65 years due to disability or renal dialysis, 3,106 participants without traditional fee-for-service Medicare, and 757 participants diagnosed with AF before the surveillance period, leaving a final cohort of 18,493 CCLS participants (Supplemental Figure 1).In a subset of study participants (N ¼ 7,435), a repeat treadmill exercise test was performed an average of 2.3 years after the qualifying examination.No CCLS participant was excluded from the present analysis based on their maximal treadmill test results.Participants in the CCLS provide informed consent for their data to be used for research, and the CCLS is approved annually by the Cooper Institute Institutional Review Board (IRB# IORG0000452-00000767).

FIGURE 1 2 Fitness
FIGURE 1 Incidence of Atrial Fibrillation After the Age of 65 Years Across Sex-Specific Deciles of Midlife CRF Among the Participants of Cooper Center Longitudinal Study Fitness, Obesity, and Risk of Atrial Fibrillation findings add to the existing literature by evaluating the independent associations of objectively measured CRF and BMI in midlife and the risk of AF in older age.

Finnish
participants.In contrast to our observations, Garnvik et al 25 demonstrated that a 1-MET increase in estimated CRF levels was associated with a 10% lower risk of AF.However, Garnvik et al 25 used non-exercise-based estimates of CRF and did not adjust for baseline BMI in the models evaluating the association of CRF change with the risk of AF.

Dr
Pandey has received grant funding outside the present study from Applied Therapeutics and Gilead Sciences; has received honoraria outside of the present study as an advisor/consultant for Tricog Health Inc, Lilly USA, Rivus, and Roche Diagnostics; has received nonfinancial support from Pfizer and Merck; and has received research support for this study from the Texas Health Resources Clinical Scholarship, the Gilead Sciences Research Scholar Program, and the National Institute on Aging GEMSSTAR Grant (1R03AG067960-01).All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.ADDRESS FOR CORRESPONDENCE: Dr Ambarish Pandey, Division of Cardiology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, Texas 75390-9047, USA.E-mail: ambarish.pandey@utsouthwestern.edu.

2 Fitness
Fitness, Obesity, and Risk of Atrial Fibrillation PERSPECTIVES COMPETENCY IN MEDICAL KNOWLEDGE: Obesity but not low fitness in midlife is independently associated with the risk of AF in older age.TRANSLATIONAL OUTLOOK: Future studies are needed to determine if lifestyle interventions aimed at intentional weight loss may lower the risk of AF in older age.Pandey et al J A C C : A D V A N C E S , V O L . 1 , N O . 2 , 2 0 2

Table 3 )
after adjustment for midlife measures of BMI, CRF, CV risk factors, and change in BMI over the same period.Change in BMI was also not significantly associated with risk of AF in the most adjusted model 1.03-1.06).In stratified analysis by BMI, the association between (HR [95% CI] per 1 kg/m 2 increase in BMI: 1.03 [0.97-1.08];P ¼ 0.33) (Table

TABLE 1
Midlife Characteristics of Participants in the Cooper Center Longitudinal Study by Fitness Categories Values are mean AE SD or n (%).BMI ¼ body mass index; CRF ¼ cardiorespiratory fitness; MET ¼ metabolic equivalent of task; SBP ¼ systolic blood pressure.

TABLE 2
Multivariable-Adjusted Association Between Midlife CRF Levels, Body Mass Index, and Risk of Atrial Fibrillation in Older Age Separate models were constructed using categorical and continuous measures of CRF.Model 1: adjusted for age, sex, and exam year.Model 2: Model 1 þ BMI.Model 3: Model 2 þ resting systolic blood pressure, cholesterol, glucose, and smoking status.Low fit ¼ category 1, moderate fit ¼ categories 2 to 3, high fit ¼ categories 4 to 5.

TABLE 3
Multivariable-Adjusted Association in 7,435 Participants With Multiple Visits Between Baseline and Change Values of Midlife CRF Levels and Body Mass Index on Risk of Atrial Fibrillation in Older Age Model 1: adjusted for age, sex, exam year, baseline CRF, change CRF, and difference in exam years.Model 2: Model 2 þ baseline and change in BMI.Model 3: Model 3 þ resting systolic blood pressure, cholesterol, glucose, and smoking status.BMI ¼ body mass index; CI ¼ confidence interval; CRF ¼ cardiorespiratory fitness; HR ¼ hazard ratio; MET ¼ metabolic equivalent of task; NA ¼ not available.

Table 1 .
Third, incident AF was captured nority race/ethnicity, socioeconomically disadvantaged individuals, and those with a high burden of CV risk factors.Second, due to the limited availability of these data, income and education level are not reported in