Previous studies have raised the question of whether an association exists between physical activity and atrial fibrillation (AF). We used the Multi-Ethnic Study of Atherosclerosis (MESA) database to examine the association between physical activity and AF in a diverse population without clinically recognized cardiovascular disease (CVD). MESA participants (n = 5,793) with complete baseline physical activity and covariate data were included. Cox proportional hazards models were used to calculate hazard ratios (HRs) for incident AF by levels of total intentional exercise and vigorous physical activity, independently and in combination. Multivariate models were adjusted for demographics and CVD risk factors. During a mean follow-up of 7.7 ± 1.9 years, 199 AF cases occurred. In the overall MESA population, neither vigorous physical activity nor total intentional exercise was independently associated with incident AF after adjusting for covariates. However, within the group that reported any vigorous physical activity, there was a statistically significant inverse association between total intentional exercise (modeled as a continuous variable) and incident AF. In those who reported any vigorous physical activity, the top tertile of total intentional exercise was associated with a significantly lower risk of incident AF compared with the group with no total intentional exercise in the fully adjusted model (HR 0.46, 95% confidence interval 0.22 to 0.98). In conclusion, neither total intentional exercise nor vigorous physical activity alone was associated with incident AF, but greater total intentional exercise was associated with a lower risk of incident AF in those who participated in any vigorous physical activity. As importantly, no subgroup of participants demonstrated an increased risk of incident AF with greater physical activity. The results re-emphasize the beneficial role of physical activity for cardiovascular health.
A number of studies have examined and presented data both for and against an association between physical activity and AF. There seems to be an increased propensity for AF in elite athletes such as marathon runners. There is also a suggestion that the intensity of exercise may play a role in risk of incident AF, and a number of studies have specifically implicated vigorous exercise as a risk factor for incident AF. The physiologic basis for an association exists, and physical activity is potentially implicated in providing the trigger, modulators, and substrate for AF onset. In contrast, other studies have suggested a decreased risk of AF in walkers and runners and in those who participate in low-to-moderate intensity activity. A recent meta-analysis of 4 prospective cohort studies demonstrated no association between regular physical activity and AF in nonathletes. The conflicting data suggest that there is still much to be uncovered regarding the association of physical activity and AF, especially in the general population. In this study, we sought to examine the association between physical activity and AF in the Multi-Ethnic Study of Atherosclerosis (MESA) cohort.
Methods
Inclusion criteria and methods of the MESA study have been described previously. In brief, from July 2000 to August 2002, a total of 6,814 men and women aged 45 to 84 year and free of clinically apparent cardiovascular disease were recruited from 6 US communities: Baltimore City and Baltimore County, Maryland; Chicago, Illinois; Forsyth County, North Carolina; Los Angeles County, California; Northern Manhattan and the Bronx, New York; and St. Paul, Minnesota. Participants filled out a survey documenting their baseline level of physical activity and the specific type and duration of physical activity. Patients with self-reported AF were excluded from enrollment in MESA. All participants were contacted every 9 months to inquire about interval hospitalizations, obtain medical records, and abstract International Classification of Diseases, Ninth Revision, codes from the records. Incident AF events were ascertained on the basis of hospital discharge International Classification of Diseases, Ninth Revision, codes (427.3x) and Medicare claims data for those who were enrolled in fee-for-service Medicare at any time during follow-up. The institutional review boards at all participating centers approved the study, and all participants gave informed consent. In this study, we included all participants with completed physical activity questionnaires and complete covariate data (n = 5,793).
During the baseline examination (2000 to 2002), standardized questionnaires and calibrated devices were used to obtain demographic data, tobacco use data, co-morbidities, current prescription medication usage, weight, and height. Resting, seated blood pressure was measured 3 times using a Dinamap automated oscillometric sphygmomanometer (model Pro 100; Critikon, Tampa, Florida); the last 2 measurements were averaged for analysis. Hypertension was defined on the basis of use of an antihypertensive medication, systolic blood pressure (SBP) ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg. Fasting blood samples were drawn and were sent to a central laboratory for measurement of glucose and lipids. Participants were considered to have diabetes if they used hypoglycemic drugs or if their fasting blood glucose was >7.0 mmol/l (126 mg/dl). Participants were considered to have impaired fasting glucose if they did not have diabetes according to the preceding criteria but their fasting blood glucose level was 5.6 to 7.0 mmol/l (100 to 126 mg/dl) in accordance with the 2004 American Diabetes Association definition. Incident coronary heart disease (CHD) during follow-up was defined as myocardial infarction, angina, resuscitated cardiac arrest, or death attributed to coronary heart disease.
The MESA Typical Week Physical Activity Survey (TWPAS), completed during the baseline examination, identifies the amount of time spent in and the frequency of various physical activities during a typical week in the previous month. The survey has 28 items in 9 categories of activities (1, household chores; 2, lawn/yard/garden/farm; 3, care of children/adults; 4, transportation; 5, nonoccupational walking; 6, dancing and sport activities; 7, conditioning activities; 8, leisure activities; 9, work). Respondents were asked whether they participated in these activities, and if applicable, they answered questions regarding the average number of days per week and time per day engaged in each activity. Where appropriate, the physical activity survey accounted for exercise intensity at 3 levels (heavy, moderate, or light), which was determined by the type of activity in any given category (i.e., sitting or standing vs pushing or lifting). Minutes of activity were summed for each discrete activity type and multiplied by metabolic equivalent (MET) level.
In our study, we used 2 summary variables derived from the MESA TWPAS data: total intentional exercise and vigorous physical activity. Total intentional exercise focused on the areas of physical activity recommended by US Department of Health and Human Services guidelines and was a sum of walking for exercise, sports, dancing, and conditioning. We were interested in the role of exercise intensity because the conflicting results of previous studies suggest that it may mediate the association between physical activity and AF. Vigorous physical activity was a sum of MET-min/wk spent as “heavy effort” in only the following modes of activity: household chores, lawn/yard/garden/farm work, conditioning activities, and occupational/volunteering work. Conditioning was the only form of physical activity included in both the total intentional exercise and vigorous physical activity summary measurements.
All participants who completed the TWPAS and had completed covariate data were included in this study, with the vigorous physical activity and total intentional exercise summary measurements used as independent variables. Baseline characteristics for the entire group and each subgroup were summarized as mean (standard deviation) and numbers (percentages) and compared using the Student t test or the chi-square tests as appropriate. Cox proportional hazards models were used to examine the association of total intentional exercise or vigorous physical activity, modeled as continuous variables, with incident AF for the entire cohort. Because of the wide range of total intentional exercise, the variable was scaled down by dividing by its standard deviation. Model 1 included covariates for demographics including age, gender, race or ethnicity, insurance status, and education level. Model 2 included all variables in model 1 in addition to risk factors including cigarette smoking, resting heart rate, SBP, use of antihypertensives, low-density lipoprotein, high-density lipoprotein, body mass index, diabetes, and left ventricular hypertrophy (by electrographic criteria). Model 3 included all model 2 variables, with the addition of incident CHD. We subsequently modeled total intentional exercise and vigorous physical activity as categorical variables by stratifying them into tertiles and performed similar analyses.
To further elucidate the effect of vigorous physical activity, the participants were then stratified into 2 subgroups: those that performed any vigorous physical activity and those that did not perform any. For each subgroup, Cox proportional hazards models, as mentioned previously, were used to examine the association of total intentional exercise with incident AF. We additionally stratified total intentional exercise into tertiles and applied the same Cox models for each subgroup. To provide detailed analyses of the dose–response relation of total intentional exercise with incident AF in those that participated in vigorous physical activity, we modeled total intentional exercise with restricted quadratic splines with knots at the twenty-fifth, fiftieth, and seventy-fifth percentiles of the total intentional exercise distribution. In spline analyses, we used the twenty-fifth percentile of the total intentional exercise distribution as the reference value. We tested for potential interactions with age, gender, and race or ethnicity, using multiplicative models as a part of model 2.
Results
Of the total 6,814 MESA participants, 5,793 had complete TWPAS and covariate data and were included in our study. Overall, 199 participants (3.4%) developed AF over a mean follow-up of 7.7 ± 1.9 years. There were 1,300 participants (22.4%) who reported no total intentional exercise. The average ± standard deviation of weekly total intentional exercise was 1,567 ± 2,345 MET-min/wk overall and was 2,007 ± 2,471 MET-min/wk in those who reported some total intentional exercise. In the entire population, the average age was 62 ± 10 years, and 47% were men. Of the entire study population, 1,866 (32.2%) reported participation in vigorous physical activity, and of those who reported some total intentional exercise, 1,612 (35.9%) participated in any vigorous physical activity.
A comparison of baseline characteristics for the entire population and the subgroups stratified by participation in vigorous physical activity is presented in Table 1 . The distribution of total intentional exercise by tertile for the 2 subgroups is shown in Figure 1 and demonstrates generally higher total intentional exercise ranges for the subgroup with any vigorous physical activity. Besides exercising more, the participants that took part in vigorous physical activity were younger, were more likely to be male, and were ethnically different compared to those that did not. Participants with any vigorous physical activity also had a lower heart rate, SBP, higher low-density lipoprotein and lower high-density lipoprotein, lower body mass index; were less likely to be on antihypertensive agents; tended to be better educated; and had a lower prevalence of diabetes.
| Variable | Overall (N=5793) | Vigorous Physical Activity | ||
|---|---|---|---|---|
| = 0 (N=3927) | > 0 (N=1866) | p value | ||
| Total Intentional Exercise (MET-min/wk) | 1566 ± 2345 | 1168 ± 1674 | 2406 ± 3184 | < 0.010 |
| Age (years) | 62 ± 10 | 63 ± 10 | 59 ± 10 | < 0.010 |
| Resting Heart Rate (bpm) | 63 ± 9 | 64 ± 10 | 61 ± 9 | < 0.010 |
| Systolic Blood Pressure (mm Hg) | 126 ± 21 | 128 ± 22 | 123 ± 20 | < 0.010 |
| Low density lipoprotein (mg/dl) | 117 ± 31 | 117 ± 32 | 119 ± 31 | = 0.012 |
| High density lipoprotein (mg/dl) | 51 ± 15 | 52 ± 15 | 50 ± 14 | < 0.010 |
| Body mass index (kg/m 2 ) | 28 ± 5 | 29 ± 6 | 28 ± 5 | < 0.010 |
| Female | 53% | 62% | 36% | < 0.010 |
| White | 38% | 33% | 48% | <0.010 |
| Chinese-American | 12% | 13% | 9% | |
| Black | 27% | 29% | 25% | |
| Hispanic | 22% | 24% | 17% | |
| Education | < 0.010 | |||
| Less than HS grade | 19% | 22% | 11% | |
| HS Grade | 18% | 20% | 14% | |
| Post-HS degree | 29% | 27% | 34% | |
| Grad degree | 18% | 15% | 23% | |
| Smoker | < 0.010 | |||
| Never | 51% | 52% | 47% | |
| Former | 36% | 35% | 40% | |
| Current | 13% | 13% | 13% | |
| HTN Meds | 36% | 39% | 29% | < 0.010 |
| Diabetes mellitus | 12% | 14% | 9% | < 0.010 |
There was no association between vigorous physical activity (continuous variable) and incident AF, and this finding persisted after adjustment for confounders and inclusion of incident CHD. Additionally, when vigorous physical activity was classified into tertiles and compared with the group with no reported vigorous physical activity, it was not associated with incident AF ( Table 2 ). There were no interactions between vigorous physical activity and age, gender, or race on AF risk.
| Total Intentional Exercise | Vigorous Physical Activity | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| N | HR | CI | p | N | HR | CI | p | ||
| No total intentional exercise | 1300 | 1.00 | — | — | No vigorous physical activity | 3927 | 1.00 | — | — |
| Lower tertile | 1497 | 1.16 | 0.78-1.72 | 0.467 | Lower tertile | 534 | 0.93 | 0.54-1.60 | 0.804 |
| Middle tertile | 1497 | 0.86 | 0.56-1.31 | 0.482 | Middle tertile | 690 | 1.14 | 0.73-1.77 | 0.559 |
| Upper tertile | 1499 | 0.85 | 0.56-1.30 | 0.465 | Upper tertile | 642 | 1.29 | 0.81-2.04 | 0.296 |
In the overall cohort, we observed no association between total intentional exercise and incident AF after adjusting for covariates and including incident CHD. Similarly, no significant association was observed between total intentional exercise and incident AF after categorization of the total intentional exercise variable into 4 levels: no reported total intentional exercise and tertiles of those with any total intentional exercise ( Table 2 ). Additionally, we found no interactions between total intentional exercise and age, gender, or race on AF risk.
In the participants without vigorous physical activity, total intentional exercise as a continuous variable was not associated with incident AF in model 2 (hazard ratio [HR] 0.984, p = 0.89) or after inclusion of incident CHD in model 3 (HR 1.00, p = 0.99). Similarly, categorization of total intentional exercise into 4 levels did not produce any significant findings in terms of association with incident AF ( Table 3 ). Although there were no interactions with gender or ethnicity, we did find a significant interaction with age (P for interaction 0.04). Specifically, in the second age quartile (ages 53 to 60 years) compared with the remainder of the population, greater total intentional exercise (modeled as a continuous variable) was negatively associated with incident AF (HR 0.84, p = 0.03).
