Height, Weight, and Aerobic Fitness Level in Relation to the Risk of Atrial Fibrillation

Height, Weight, and Aerobic Fitness Level in Relation to the Risk of Atrial Fibrillation Abstract Tall stature and obesity have been associated with a higher risk of atrial fibrillation (AF), but there have been conflicting reports of the effects of aerobic fitness. We conducted a national cohort study to examine interactions between height or weight and level of aerobic fitness among 1,547,478 Swedish military conscripts during 1969–1997 (97%–98% of all 18-year-old men) in relation to AF identified from nationwide inpatient and outpatient diagnoses through 2012 (maximal age, 62 years). Increased height, weight, and aerobic fitness level (but not muscular strength) at age 18 years were all associated with a higher AF risk in adulthood. Positive additive and multiplicative interactions were found between height or weight and aerobic fitness level (for the highest tertiles of height and aerobic fitness level vs. the lowest, relative excess risk = 0.51, 95% confidence interval (CI): 0.40, 0.62; ratio of hazard ratios = 1.50, 95% CI: 1.34, 1.65). High aerobic fitness levels were associated with higher risk among men who were at least 186 cm (6 feet, 1 inch) tall but were protective among shorter men. Men with the combination of tall stature and high aerobic fitness level had the highest risk (for the highest tertiles vs. the lowest, adjusted hazard ratio = 1.70, 95% CI: 1.61, 1.80). These findings suggest important interactions between body size and aerobic fitness level in relation to AF and may help identify high-risk subgroups. atrial fibrillation, body height, body mass index, body size, body weight, muscle strength, physical fitness Atrial fibrillation (AF) is the most common cardiac arrhythmia; it affects more than 30 million people worldwide (1) and is an important cause of stroke and mortality (2, 3). AF is more common with increasing age or underlying heart disease (4, 5). However, many cases of AF also occur among younger adults who do not have known heart disease (6), which has led to investigations of other risk factors and potential targets for intervention. Researchers from prior studies have established that tall stature (7–11) and obesity (11–18) are associated with a higher risk of the development of AF. However, the reported effects of aerobic fitness have been conflicting. In a number of studies, investigators have reported an increased risk of AF among competitive athletes or individuals who report high levels of exercise (19–24). In contrast, results from several (12, 25–27) but not all (28) studies in which objectively measured levels of aerobic fitness were included have suggested that a high level of aerobic fitness is protective against AF. We hypothesized that these conflicting findings may be due to interactions between height or weight and aerobic fitness level that have not been examined. The pathogenesis of AF is a heterogeneous process that involves myocardial structural, hemodynamic, electrical, and neural factors (29). Physical exercise may potentially influence this process through its contribution to left ventricular hypertrophy, which leads to diastolic dysfunction and left atrial stretch (30). Because height and weight correlate positively with left atrial size (31), these variables may potentially modify the hemodynamic and autonomic effects of aerobic fitness on atrial remodeling and the development of AF (32). If so, elucidation of interactions among these factors may help identify high-risk subgroups and provide new insights into underlying mechanisms. We examined the interactions among height, weight, and aerobic fitness level in relation to the risk of AF in a large national cohort. Standardized measurements of these factors were obtained in approximately 1.5 million 18-year-old male military conscripts in Sweden who were subsequently followed up for AF in adulthood. In the present study, our aim was to determine the interactive effects of these common factors at age 18 years on the long-term risk of AF in a large, population-based cohort. METHODS Study population We identified 1,547,478 men (approximately 18 years of age) who underwent a military conscription examination in Sweden during 1969–1997. Nationally, all 18-year-old men were conscripted each year except for 2%–3% who either were incarcerated or had severe chronic medical conditions or disabilities that were documented by a physician. The Regional Ethics Committee of Lund University in Sweden approved this study. Ascertainment of height, weight, and aerobic fitness level We obtained measurements of height, weight, and aerobic fitness level from the Swedish Military Conscription Registry, which contains information from 2-day standardized physical and psychological examinations required of all conscripts starting in 1969 (33–37). We measured height and weight using standard protocols and examined them alternatively as continuous variables or categorical variables in tertiles (low, medium, and high tertiles for height: <175.0 cm, 175.0–181.4 cm, and ≥181.5 cm, respectively; for weight, <64.0 kg, 64.0–71.3 kg, and ≥71.4 kg, respectively). Categorizations for these and other predictors were based on dose-response analyses, which indicated that they well-characterized the underlying relationships with AF risk and had good interpretability. We examined body mass index (BMI) as an alternative to height and weight. We calculated BMI as weight in kilograms divided by the square of height in meters. We examined BMI alternatively as a continuous or categorical variable using Centers for Disease Control and Prevention definitions for children and adolescents 2–19 years of age to facilitate comparability with US studies. “Overweight” was defined as being in the 85th percentile or higher but less than 95th percentile and “obesity” as being in the 95th percentile or higher on the sex-specific BMI-for-age growth charts, issued in 2000; these values correspond to BMIs of 25.6–28.9 and 29.0 or higher, respectively, for 18-year-old men (38). The level of aerobic fitness was measured as the maximal aerobic workload in watts using a well-validated, electrically braked stationary bicycle ergometer test (39). After a warm-up period, each conscript performed 5–10 minutes of exercise on a stationary bicycle at a starting work rate of 75–175 W (determined on a sliding scale based on body weight) and increasing by 25 W per minute until volitional exhaustion. Maximal aerobic workload was calculated as the power output in watts before the last increase in intensity plus the prorated output for the last stage (39). Maximal aerobic workload correlates highly with maximal oxygen uptake (correlation ≈ 0.9) (40), and its measurement using this bicycle ergometer test is highly reproducible, with a test-retest correlation of 0.95 (41). We examined this measurement of aerobic fitness alternatively as a continuous variable or categorical variable in tertiles (<240 W, 240–288 W, and ≥289 W) (33–37). In addition to aerobic fitness level, we examined muscular strength as a secondary predictor of interest because it represents a different aspect of physical fitness. Muscular strength was measured in Newtons using well-validated isometric dynamometer tests and calculated as the weighted sum of maximal knee extension (weighted × 1.3), elbow flexion (weighted × 0.8), and hand grip (weighted × 1.7) (42). Each dynamometer test was performed 3 times, and the maximal value was recorded for analysis, except when the last value was highest, in which case testing was repeated until strength values stopped increasing. All testing equipment was calibrated daily (42). We examined muscular strength alternatively as a continuous variable or categorical variable in tertiles (<1,900 N, 1,900–2,170 N, and ≥2,171 N) (33–37). Ascertainment of AF We followed up the study cohort for the earliest diagnosis of AF, from the date of the military conscription examination through December 31, 2012. AF was identified based on any primary or secondary diagnosis according to codes from the International Classification of Diseases (ICD) listed in the Swedish Hospital and Outpatient Registries (ICD codes 427.4 (Eighth Revision), 427.3 (Ninth Revision), and I48 (Tenth Revision)). The Swedish Hospital Registry contains all primary and secondary hospital discharge diagnoses from 6 populous counties in southern Sweden starting in 1964, and with nationwide coverage starting in 1987. The Swedish Outpatient Registry also contains outpatient diagnoses from all specialty clinics nationwide starting in 2001 (33–37). Diagnosis records in the Hospital Registry are currently more than 99% complete and have a reported positive predictive value of 97% for either primary or secondary diagnoses of AF (43). Adjustment variables Data on other variables that may be associated with body size or physical fitness level and the risk of AF were obtained from the Swedish Military Conscription Registry and national census data, which were linked by use of an anonymous personal identification number (33–37). The following were used as adjustment variables: year that the military conscription examination was conducted (as a continuous variable); highest educational level attained during the study period (<12, 12–14, or ≥15 years); neighborhood socioeconomic status (SES) at baseline; and a family history of AF in a parent or sibling (yes or no, identified from the Swedish Hospital Registry during 1965–2012 and the Swedish Outpatient Registry during 2001–2012, using the same diagnosis codes noted above). SES was included because characteristics of neighborhood SES have been associated with obesity and physical fitness (44) and with the risk of cardiovascular disease (45). It was composed of an index that included low educational level, low income status, unemployment, and receipt of social welfare benefits, as previously described (46), and categorized as low (more than 1 SD below the mean), medium (1 SD below to 1 SD above the mean), or high (more than 1 SD above the mean). Because low levels of aerobic fitness and a high BMI are known to be associated with hypertension (33), diabetes mellitus (34), and ischemic heart disease (35), which are established risk factors for AF (47), we were interested in whether aerobic fitness level and BMI (or height and weight) are associated with AF independently of these conditions. To assess this, we further adjusted for hypertension (ICD codes 400–401 (Eighth Revision), 401 (Ninth Revision), and I10 (Tenth Revision)), diabetes mellitus (ICD code 250 (Eighth and Ninth Revisions) and E10–E14 (Tenth Revision)), and ischemic heart disease (ICD codes 410–414 (Eighth and Ninth Revisions) and I20–I25 (Tenth Revision)) as time-dependent variables. We imputed missing data for each variable using a standard multiple imputation procedure based on the relationship of the variable with all other covariates and AF (48). Missing data were relatively infrequent for height (7.2%), weight (7.3%), level of aerobic fitness (5.7%), muscular strength (5.0%), educational level (0.4%), and neighborhood SES (9.1%). Data were 100% complete for all other variables. Statistical analysis We used Cox proportional hazards regression models to compute hazard ratios and 95% confidence intervals for associations of height, weight, aerobic fitness level, or muscular strength with the risk of AF. The Cox model time scale was elapsed time since the military conscription examination, which also corresponds to attained age because the baseline age was the same (18 years) for all conscripts. Individuals were censored at emigration (n = 87,450; 5.7%) or death (n = 58,835; 3.8%). We used 3 models. The first was adjusted for attained age (as the time scale) and year of the military conscription examination; the second was also adjusted for height, weight, level of aerobic fitness, muscular strength, educational level, neighborhood SES, and family history of AF; and the third was further adjusted for hypertension, diabetes mellitus, and ischemic heart disease (as defined above). All time-varying factors were modeled as time-dependent covariates. The proportional hazards assumption was evaluated using graphical assessment of log-log plots, which showed a good fit in all models. Interactions among height, weight, aerobic fitness level, and muscular strength were examined on both the additive and multiplicative scale in relation to the risk of AF. We assessed additive interactions using the relative excess risk due to interaction, which is computed for binary variables as: RERIHR = HR11 − HR10 − HR01 + 1 (49, 50), where HR is the hazard ratio. Multiplicative interactions were assessed using the ratio of hazard ratios: HR11/(HR10 × HR01) (49, 50). These estimates and their respective 95% confidence intervals were determined using maximal likelihood estimation (49–51). A positive additive interaction is indicated if the relative excess risk due to interaction is greater than 0, and a positive multiplicative interaction is indicated if the ratio of hazard ratios is more than 1. See the Web Appendix (available at https://academic.oup.com/aje) for more details on these methods. We performed 2 sensitivity analyses. First, as an alternative to multiple imputation for missing data, we repeated all analyses after restricting to men for whom we had complete data for all variables (n = 1,361,083; 88.0%). Second, because of changes in AF ascertainment across time, we assessed different starting points for the follow-up period, alternatively starting in 1987 (at which time inpatient diagnoses were available nationwide instead of only for the most populous counties) or in 2001 (at which time both inpatient and outpatient diagnoses were available nationwide). All statistical tests were 2-sided, and an α-level of 0.05 was used. We conducted all analyses using Stata, version 14.1 (StataCorp LP, College Station, Texas). RESULTS Among the 1,547,478 men in this cohort, 23,600 (1.5%) were diagnosed with AF at 43.7 million person-years of follow-up (mean follow-up = 28.2 years). The median age at the end of follow-up was 47.2 years (mean = 47.4 (SD, 7.9) years; range, 19.0–62.0 years). The median age at diagnosis of AF was 48.2 years (mean = 47.1 (SD, 7.8) years; range, 18.2–62.0 years). Table 1 shows incidence rates of AF by height, weight, aerobic fitness level, and other factors. Table 1. Characteristics of Men Who Were or Were Not Subsequently Diagnosed With Atrial Fibrillation, Sweden, 1969–2012 Variable  No AF  AF  Ratea  No.  %  No.  %  Total  1,523,818  100.0  23,660  100.0  54.3  Tertile of height             Lowest (<175.0 cm)  527,967  34.6  6,888  29.1  44.8   Middle (175.0–181.4 cm)  504,217  33.1  6,801  28.7  47.4   Highest (≥181.5 cm)  491,634  32.3  9,971  42.1  71.9  Tertile of weight             Lowest (<64.0 kg)  500,063  32.8  6,784  28.7  45.1   Middle (64.0–71.3 kg)  506,366  33.2  7,061  29.8  49.2   Highest (≥71.4 kg)  517,389  34.0  9,815  41.5  69.1  BMIb             Normal  1,406,267  92.3  21,248  89.8  52.6   Overweight  83,124  5.4  1,544  6.5  68.8   Obesity  34,427  2.3  868  3.7  96.2  Tertile of aerobic fitness level             Lowest (<240 W)  501,335  32.9  10,015  42.3  60.3   Middle (240–288 W)  511,945  33.6  8,602  36.4  57.4   Highest (≥289 W)  510,538  33.5  5,043  21.3  42.0  Tertile of muscular strength             Lowest (<1,900 N)  504,507  33.1  6,418  27.1  46.4   Middle (1,900–2,170 N)  515,024  33.8  8,528  36.0  55.2   Highest (≥2,171 N)  504,287  33.1  8,714  36.8  60.9  Education, years             <12  232,151  15.2  4,693  19.8  65.9   12–14  674,001  44.2  9,626  40.7  51.0   ≥15  617,666  40.5  9,341  39.5  53.2  Neighborhood SES             Low  235,421  15.5  3,973  16.8  57.4   Medium  1,006,116  66.0  16,119  68.1  55.4   High  282,281  18.5  3,568  15.1  47.1  Hypertension             No  1,438,676  94.4  15,478  65.4  37.7   Yes  85,142  5.6  8,182  34.6  257.5  Diabetes mellitus             No  1,475,411  96.8  20,836  88.1  49.7   Yes  48,407  3.2  2,824  11.9  169.2  Ischemic heart disease             No  1,488,996  97.7  20,269  85.7  48.0   Yes  34,822  2.3  3,391  14.3  255.8  Family history of AF             No  1,223,751  80.3  14,986  63.3  44.0   Yes  300,067  19.7  8,674  36.7  91.2  Variable  No AF  AF  Ratea  No.  %  No.  %  Total  1,523,818  100.0  23,660  100.0  54.3  Tertile of height             Lowest (<175.0 cm)  527,967  34.6  6,888  29.1  44.8   Middle (175.0–181.4 cm)  504,217  33.1  6,801  28.7  47.4   Highest (≥181.5 cm)  491,634  32.3  9,971  42.1  71.9  Tertile of weight             Lowest (<64.0 kg)  500,063  32.8  6,784  28.7  45.1   Middle (64.0–71.3 kg)  506,366  33.2  7,061  29.8  49.2   Highest (≥71.4 kg)  517,389  34.0  9,815  41.5  69.1  BMIb             Normal  1,406,267  92.3  21,248  89.8  52.6   Overweight  83,124  5.4  1,544  6.5  68.8   Obesity  34,427  2.3  868  3.7  96.2  Tertile of aerobic fitness level             Lowest (<240 W)  501,335  32.9  10,015  42.3  60.3   Middle (240–288 W)  511,945  33.6  8,602  36.4  57.4   Highest (≥289 W)  510,538  33.5  5,043  21.3  42.0  Tertile of muscular strength             Lowest (<1,900 N)  504,507  33.1  6,418  27.1  46.4   Middle (1,900–2,170 N)  515,024  33.8  8,528  36.0  55.2   Highest (≥2,171 N)  504,287  33.1  8,714  36.8  60.9  Education, years             <12  232,151  15.2  4,693  19.8  65.9   12–14  674,001  44.2  9,626  40.7  51.0   ≥15  617,666  40.5  9,341  39.5  53.2  Neighborhood SES             Low  235,421  15.5  3,973  16.8  57.4   Medium  1,006,116  66.0  16,119  68.1  55.4   High  282,281  18.5  3,568  15.1  47.1  Hypertension             No  1,438,676  94.4  15,478  65.4  37.7   Yes  85,142  5.6  8,182  34.6  257.5  Diabetes mellitus             No  1,475,411  96.8  20,836  88.1  49.7   Yes  48,407  3.2  2,824  11.9  169.2  Ischemic heart disease             No  1,488,996  97.7  20,269  85.7  48.0   Yes  34,822  2.3  3,391  14.3  255.8  Family history of AF             No  1,223,751  80.3  14,986  63.3  44.0   Yes  300,067  19.7  8,674  36.7  91.2  Abbreviations: AF, atrial fibrillation; BMI, body mass index; SES, socioeconomic status. a AF incidence rate per 100,000 person-years. b Weight (kg)/height (m)2. Main effects of height, weight, and aerobic fitness level Tall stature was associated with a higher risk of AF after adjustment for weight and all other covariates (Table 2, model 3) (for the highest tertile vs. the lowest, HR = 1.53, 95% confidence interval (CI): 1.48, 1.59; P < 0.001). Weight in the highest (but not middle) tertile was associated with a modestly higher risk of AF compared with weight in the lowest tertile after adjustment for height and all other covariates (for the highest tertile vs. the lowest, HR = 1.18, 95% CI: 1.13, 1.23; P < 0.001). Both height and weight had strongly positive linear associations with the risk of AF across their full distribution (per each 5-cm increase in height, fully adjusted HR = 1.11, 95% CI: 1.10, 1.12; P for trend < 0.001; per each 5-kg increase in weight, fully adjusted HR = 1.05, 95% CI: 1.04, 1.06; P for trend < 0.001), but height was a stronger risk factor than was weight (P for heterogeneity < 0.001). Obesity (but not overweight) also was associated with an increased risk of AF in the fully adjusted model (Table 2, model 3). Table 2. Associations of Height, Weight, Aerobic Fitness Level, or Other Factors With Risk of Atrial Fibrillation, Sweden, 1969–2012 Variable  Model 1a  Model 2b  Model 3c  HR  95% CI  P Value  HR  95% CI  P Value  HR  95% CI  P Value  Height                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.09  1.05, 1.13  <0.001  1.00  0.96, 1.03  0.85  1.05  1.01, 1.09  0.006   High  1.70  1.65, 1.75  <0.001  1.47  1.42, 1.52  <0.001  1.53  1.48, 1.59  <0.001   Per 5 cm (trend test)  1.14  1.13, 1.15  <0.001  1.04  1.03, 1.05  <0.001  1.11  1.10, 1.12  <0.001  Weight                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.18  1.14, 1.22  <0.001  1.19  1.15, 1.24  <0.001  0.98  0.94, 1.02  0.28   High  1.79  1.74, 1.85  <0.001  1.75  1.68, 1.82  <0.001  1.18  1.13, 1.23  <0.001   Per 5 kg (trend test)  1.14  1.13, 1.15  <0.001  1.13  1.12, 1.13  <0.001  1.05  1.04, 1.06  <0.001  BMId                     Normal  1.00  Referent    1.00  Referent    1.00  Referent     Overweight  1.47  1.40, 1.55  <0.001  1.35  1.28, 1.42  <0.001  1.05  0.99, 1.10  0.09   Obesity  2.22  2.07, 2.38  <0.001  2.04  1.91, 2.19  <0.001  1.31  1.22, 1.41  <0.001   Per 1 BMI unit (trend test)  1.06  1.05, 1.06  <0.001  1.05  1.05, 1.06  <0.001  1.02  1.01, 1.03  <0.001  Aerobic fitness level                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.14  1.11, 1.18  <0.001  0.94  0.91, 0.97  <0.001  1.02  0.99, 1.05  0.31   High  1.36  1.31, 1.41  <0.001  1.00  0.96, 1.05  0.84  1.14  1.09, 1.19  <0.001   Per 100 W (trend test)  1.31  1.27, 1.35  <0.001  0.99  0.95, 1.02  0.41  1.12  1.08, 1.16  <0.001  Muscular strength                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.14  1.10, 1.17  <0.001  1.01  0.98, 1.05  0.42  1.04  1.00, 1.07  0.04   High  1.37  1.33, 1.42  <0.001  0.99  0.96, 1.03  0.66  1.03  0.99, 1.07  0.13   Per 1,000 N (trend test)  1.44  1.39, 1.50  <0.001  0.94  0.90, 0.98  0.006  1.00  0.95, 1.04  0.88  Education, years                     <12  1.00  Referent    1.00  Referent    1.00  Referent     12–14  0.92  0.89, 0.95  <0.001  0.89  0.86, 0.92  <0.001  0.90  0.87, 0.94  <0.001   ≥15  0.93  0.89, 0.96  <0.001  0.89  0.86, 0.92  <0.001  0.95  0.91, 0.98  0.02   Per higher category (trend test)  0.97  0.95, 0.99  <0.001  0.95  0.93, 0.97  <0.001  0.98  0.96, 1.00  0.06  Neighborhood SES                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.01  0.98, 1.05  0.48  1.04  1.00, 1.08  0.03  1.06  1.03, 1.10  0.001   High  0.96  0.92, 1.01  0.10  0.95  0.91, 1.00  0.04  0.98  0.93, 1.02  0.36   Per higher category (trend test)  0.98  0.96, 1.00  0.12  0.98  0.96, 1.00  0.06  0.99  0.97, 1.01  0.50  Hypertension                     No  1.00  Referent          1.00  Referent     Yes  5.67  5.52, 5.83  <0.001        4.59  4.45, 4.74  <0.001  Diabetes mellitus                     No  1.00  Referent          1.00  Referent     Yes  2.89  2.77, 3.00  <0.001        1.21  1.16, 2.26  <0.001  Ischemic heart disease                     No  1.00  Referent          1.00  Referent     Yes  4.25  4.10, 4.42  <0.001        2.09  2.01, 2.18  <0.001  Family history of AF                     No  1.00  Referent    1.00      1.00  Referent     Yes  1.80  1.76, 1.85  <0.001  1.73  1.68, 1.78  <0.001  1.72  1.67, 1.77  <0.001  Variable  Model 1a  Model 2b  Model 3c  HR  95% CI  P Value  HR  95% CI  P Value  HR  95% CI  P Value  Height                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.09  1.05, 1.13  <0.001  1.00  0.96, 1.03  0.85  1.05  1.01, 1.09  0.006   High  1.70  1.65, 1.75  <0.001  1.47  1.42, 1.52  <0.001  1.53  1.48, 1.59  <0.001   Per 5 cm (trend test)  1.14  1.13, 1.15  <0.001  1.04  1.03, 1.05  <0.001  1.11  1.10, 1.12  <0.001  Weight                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.18  1.14, 1.22  <0.001  1.19  1.15, 1.24  <0.001  0.98  0.94, 1.02  0.28   High  1.79  1.74, 1.85  <0.001  1.75  1.68, 1.82  <0.001  1.18  1.13, 1.23  <0.001   Per 5 kg (trend test)  1.14  1.13, 1.15  <0.001  1.13  1.12, 1.13  <0.001  1.05  1.04, 1.06  <0.001  BMId                     Normal  1.00  Referent    1.00  Referent    1.00  Referent     Overweight  1.47  1.40, 1.55  <0.001  1.35  1.28, 1.42  <0.001  1.05  0.99, 1.10  0.09   Obesity  2.22  2.07, 2.38  <0.001  2.04  1.91, 2.19  <0.001  1.31  1.22, 1.41  <0.001   Per 1 BMI unit (trend test)  1.06  1.05, 1.06  <0.001  1.05  1.05, 1.06  <0.001  1.02  1.01, 1.03  <0.001  Aerobic fitness level                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.14  1.11, 1.18  <0.001  0.94  0.91, 0.97  <0.001  1.02  0.99, 1.05  0.31   High  1.36  1.31, 1.41  <0.001  1.00  0.96, 1.05  0.84  1.14  1.09, 1.19  <0.001   Per 100 W (trend test)  1.31  1.27, 1.35  <0.001  0.99  0.95, 1.02  0.41  1.12  1.08, 1.16  <0.001  Muscular strength                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.14  1.10, 1.17  <0.001  1.01  0.98, 1.05  0.42  1.04  1.00, 1.07  0.04   High  1.37  1.33, 1.42  <0.001  0.99  0.96, 1.03  0.66  1.03  0.99, 1.07  0.13   Per 1,000 N (trend test)  1.44  1.39, 1.50  <0.001  0.94  0.90, 0.98  0.006  1.00  0.95, 1.04  0.88  Education, years                     <12  1.00  Referent    1.00  Referent    1.00  Referent     12–14  0.92  0.89, 0.95  <0.001  0.89  0.86, 0.92  <0.001  0.90  0.87, 0.94  <0.001   ≥15  0.93  0.89, 0.96  <0.001  0.89  0.86, 0.92  <0.001  0.95  0.91, 0.98  0.02   Per higher category (trend test)  0.97  0.95, 0.99  <0.001  0.95  0.93, 0.97  <0.001  0.98  0.96, 1.00  0.06  Neighborhood SES                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.01  0.98, 1.05  0.48  1.04  1.00, 1.08  0.03  1.06  1.03, 1.10  0.001   High  0.96  0.92, 1.01  0.10  0.95  0.91, 1.00  0.04  0.98  0.93, 1.02  0.36   Per higher category (trend test)  0.98  0.96, 1.00  0.12  0.98  0.96, 1.00  0.06  0.99  0.97, 1.01  0.50  Hypertension                     No  1.00  Referent          1.00  Referent     Yes  5.67  5.52, 5.83  <0.001        4.59  4.45, 4.74  <0.001  Diabetes mellitus                     No  1.00  Referent          1.00  Referent     Yes  2.89  2.77, 3.00  <0.001        1.21  1.16, 2.26  <0.001  Ischemic heart disease                     No  1.00  Referent          1.00  Referent     Yes  4.25  4.10, 4.42  <0.001        2.09  2.01, 2.18  <0.001  Family history of AF                     No  1.00  Referent    1.00      1.00  Referent     Yes  1.80  1.76, 1.85  <0.001  1.73  1.68, 1.78  <0.001  1.72  1.67, 1.77  <0.001  Abbreviations: AF, atrial fibrillation; BMI, body mass index; CI, confidence interval; HR, hazard ratio; SES, socioeconomic status. a Adjusted for age and year of military conscription examination. b Adjusted for age, year of military conscription examination, height, weight, aerobic fitness level, muscular strength, educational level, neighborhood SES, and family history of AF. c Adjusted for age, year of military conscription examination, height, weight, aerobic fitness level, muscular strength, educational level, neighborhood SES, hypertension, diabetes mellitus, ischemic heart disease, and family history of AF. d BMI was calculated as weight (kg)/height (m)2 and examined as an alternative to height and weight in a separate model. High levels of aerobic fitness were associated with a modestly higher risk of AF after adjustment for height, weight, and all other covariates (for the highest tertile vs. the lowest, HR = 1.14, 95% CI: 1.09, 1.19; P < 0.001). In contrast, muscular strength was not clearly associated with AF risk (Table 2). A first-degree family history of AF was associated with a 1.7-fold higher risk of AF. Sensitivity analyses that were restricted to men for whom we had complete data or in which we examined different follow-up periods (as noted above) yielded risk estimates similar to those from the main analyses, and the overall conclusions were unchanged. Interactions among height, weight, and aerobic fitness level The interaction between height and aerobic fitness level in relation to the risk for AF is shown in Table 3. High levels of aerobic fitness were protective against AF among shorter men (lowest tertile for height) but were associated with increased risk among taller men (medium or highest tertiles; see Web Table 1 for more complete reporting of stratum-specific HRs). Men with the combination of tall height and high aerobic fitness level had the highest AF risk, which was 70% higher relative to those with short height and low aerobic fitness level (for the highest tertiles vs. the lowest for both variables, adjusted HR = 1.70, 95% CI: 1.61, 1.80; P < 0.001). There was a significant positive interaction between height and aerobic fitness level on both the additive and multiplicative scales (i.e., the association of both factors together with AF risk exceeded the sum or product of their associations considered separately; P < 0.001). Figure 1 shows the probability of AF among men in the 25th, 50th, and 75th percentiles of aerobic fitness across the full distribution of height after adjustment for all covariates. The nonparallel lines reflect a positive interaction. Specifically, high aerobic fitness levels were associated with a higher risk of AF among men who were at least 186 cm (6 feet, 1 inch) tall but were protective among men who were less than 186 cm tall. Table 3. Interaction of Height and Aerobic Fitness Level in Relation to Risk of Atrial Fibrillation,a Adjusted for Weight and Other Factors,b Sweden, 1969–2012 Tertile of Height  Tertile of Aerobic Fitness Level  Low  Medium  High  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  Low  4,569  54.1  1.00  Referent  1,756  39.3  0.89  0.84, 0.94  565  23.0  0.85  0.77, 0.93  Medium  2,633  55.3  0.95  0.90, 0.99  2,707  50.4  1.03  0.98, 1.09  1,460  34.6  1.15  1.07, 1.23  High  2,813  82.7  1.34  1.28, 1.41  4,139  80.5  1.49  1.42, 1.57  3,018  56.6  1.70  1.61, 1.80  Tertile of Height  Tertile of Aerobic Fitness Level  Low  Medium  High  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  Low  4,569  54.1  1.00  Referent  1,756  39.3  0.89  0.84, 0.94  565  23.0  0.85  0.77, 0.93  Medium  2,633  55.3  0.95  0.90, 0.99  2,707  50.4  1.03  0.98, 1.09  1,460  34.6  1.15  1.07, 1.23  High  2,813  82.7  1.34  1.28, 1.41  4,139  80.5  1.49  1.42, 1.57  3,018  56.6  1.70  1.61, 1.80  Abbreviations: CI, confidence interval; HR, hazard ratio; SES, socioeconomic status. a On an additive scale, for the highest tertiles versus the lowest, relative excess risk = 0.51, 95% CI: 0.40, 0.62; P < 0.001; on a multiplicative scale, for the highest tertiles versus the lowest, ratio of HRs = 1.50, 95% CI: 1.34, 1.65; P < 0.001. b HRs were adjusted for age, year of military conscription examination, weight, muscular strength, educational level, neighborhood SES, hypertension, diabetes mellitus, ischemic heart disease, and family history of AF. c AF incidence rate per 100,000 person-years. Figure 1. View largeDownload slide Probability of atrial fibrillation by height and aerobic fitness at age 18 years, adjusted for weight, muscular strength, and other covariates, Sweden, 1969–2012. The median attained age was 47 years, and the maximum was 62 years. Figure 1. View largeDownload slide Probability of atrial fibrillation by height and aerobic fitness at age 18 years, adjusted for weight, muscular strength, and other covariates, Sweden, 1969–2012. The median attained age was 47 years, and the maximum was 62 years. A similar overall pattern was seen for the interaction between weight and aerobic fitness level (Table 4). High levels of aerobic fitness were protective among men in the lowest tertile for weight but were associated with an increased risk of AF among those in the medium or highest tertiles (Table 4 and Web Table 2). The combination of high weight and high aerobic fitness level was associated with the highest AF risk (for the highest tertiles vs. the lowest for both variables, adjusted HR = 1.34, 95% CI: 1.27, 1.42; P < 0.001). These factors had significant positive interactions on both the additive and multiplicative scales (P < 0.001). Figure 2 shows the probability of AF among men in the 25th, 50th, and 75th percentiles of aerobic fitness across the full distribution of weight after adjustment for height and other covariates. In secondary analyses, we also found positive additive and multiplicative interactions of muscular strength with either height (Web Table 3) or weight (Web Table 4) in relation to the risk for AF, although these interactions were smaller in magnitude than the interactions between aerobic fitness level and height or weight. Table 4. Interaction of Weight and Aerobic Fitness Level in Relation to Risk of Atrial Fibrillaion,a Adjusted for Height and Other Factors,b Sweden, 1969–2012 Tertile of Weight  Tertile of Aerobic Fitness Level  Low  Medium  High  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  Low  5,213  53.0  1.00  Referent  1,309  33.9  0.86  0.80, 0.91  264  19.6  0.75  0.66, 0.86  Medium  2,643  62.0  0.90  0.86, 0.95  3,075  52.6  0.99  0.95, 1.04  1,342  31.7  1.01  0.94, 1.08  High  2,159  86.3  1.02  0.96, 1.07  4,218  80.1  1.15  1.09, 1.21  3,437  53.4  1.34  1.27, 1.42  Tertile of Weight  Tertile of Aerobic Fitness Level  Low  Medium  High  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  Low  5,213  53.0  1.00  Referent  1,309  33.9  0.86  0.80, 0.91  264  19.6  0.75  0.66, 0.86  Medium  2,643  62.0  0.90  0.86, 0.95  3,075  52.6  0.99  0.95, 1.04  1,342  31.7  1.01  0.94, 1.08  High  2,159  86.3  1.02  0.96, 1.07  4,218  80.1  1.15  1.09, 1.21  3,437  53.4  1.34  1.27, 1.42  Abbreviations: CI, confidence interval; HR, hazard ratio; SES, socioeconomic status. a On an additive scale, for the highest tertiles versus the lowest, relative excess risk = 0.57, 95% CI: 0.46, 0.68; P < 0.001; on a multiplicative scale, for the highest tertiles versus the lowest, ratio of HRs = 1.75, 95% CI: 1.51, 1.98; P < 0.001. b HRs were adjusted for age, year of military conscription examination, height, muscular strength, educational level, neighborhood SES, hypertension, diabetes mellitus, ischemic heart disease, and family history of AF. c AF incidence rate per 100,000 person-years. Figure 2. View largeDownload slide Probability of atrial fibrillation by weight and aerobic fitness level at age 18 years, adjusted for height, muscular strength, and other covariates, Sweden, 1969–2012. The median attained age was 47 years, and the maximum was 62 years. Figure 2. View largeDownload slide Probability of atrial fibrillation by weight and aerobic fitness level at age 18 years, adjusted for height, muscular strength, and other covariates, Sweden, 1969–2012. The median attained age was 47 years, and the maximum was 62 years. DISCUSSION In the present large national cohort study, we found that higher height or weight at age 18 years was associated with a higher risk of the development of AF in adulthood, and both had interactions with the level of aerobic fitness. High levels of aerobic fitness were associated with a higher risk of AF among men who were at least 186 cm (6 feet, 1 inch) tall but were protective among shorter men. In a similar manner, high levels of aerobic fitness were associated with a higher risk among men in the highest tertile of weight but not the lowest. The association between both exposures together (i.e., the combination of increased height or weight and aerobic fitness level) and the risk of AF exceeded the sum or product of their associations considered separately. These findings suggest that the underlying mechanisms for AF may include important interactions between height or weight and aerobic fitness level. The main associations that we observed between tall stature or obesity and AF are overall consistent with previous findings. In prior studies, researchers have reported that increased height (7–11) or high BMI and other measures of obesity (11–18) were associated with a higher risk of AF among men and women. Although most studies have focused on middle-aged or older adults, researchers of a study involving 12,850 young Danish men (median age, 19 years) reported that overweight or obesity was associated with a higher risk of AF in early adulthood (13). Results from other studies have suggested that high lean body mass (rather than body fat) may also be associated with a higher risk of AF (17, 52). In contrast, previous findings on aerobic fitness or exercise level in relation to AF risk have been conflicting. In some (19–24) but not all (53–55) studies, investigators have reported a higher risk of AF among athletes or other individuals who reported high levels of exercise. In a meta-analysis of 6 case-control studies that included 655 athletes and 895 control participants, researchers reported that athletes had more than a 5-fold higher odds of AF (odds ratio = 5.29, 95% CI: 3.57, 7.85) (56). However, results from most studies in which investigators examined objectively measured aerobic fitness (rather than self-reported exercise) have suggested that high levels of aerobic fitness are protective against AF (12, 25–27). The largest of these was a US cohort study of 64,561 middle-aged adults with a median follow-up of 5.4 years, in which investigators reported that each 1 higher metabolic equivalent achieved during treadmill testing was associated with a 7% lower risk of the development of AF (HR = 0.93, 95% CI: 0.92, 0.94; P < 0.001). In contrast, results from a Swedish study that included a subset of the present cohort (n = 1.1 million) suggested that aerobic fitness was positively associated with AF risk, but potential interactions with height were not reported (28). These discrepancies in reported associations between aerobic fitness level or exercise and the risk for AF may potentially be related to other modifying factors, such as body size. To our knowledge, the present study is the first to examine not only the independent associations between height, weight, or level of aerobic fitness and the risk of AF, but also the potential additive and multiplicative interactions. We found that the associations between aerobic fitness and the risk of AF varied substantially depending on height or weight. High levels of aerobic fitness were associated with a higher risk of AF among men who were at least 186 cm (6 feet, 1 inch) tall but were protective among men who were shorter than 186 cm. The interactions between height or weight and level of aerobic fitness were strongly positive on both the additive and multiplicative scales. These findings suggest that high levels of aerobic fitness accounted for significantly more cases of AF among tall versus short men or among those in the highest tertile of weight versus those in the lowest tertile. These results may provide additional insights into the underlying mechanisms of AF. AF is a heterogeneous disease process that is influenced by structural (e.g., left atrial size), hemodynamic (e.g., left atrial stretch), electrical (e.g., altered conduction patterns due to atrial myocardial fibrosis), and neural (e.g., autonomic dysregulation) factors (29). Exercise may influence this process through its contribution to left ventricular hypertrophy, which leads to some degree of diastolic dysfunction and left atrial stretch, which may increase the risk of AF among susceptible subgroups (30). Height and weight are known to correlate positively with left atrial and ventricular size (8, 31). Our findings suggest that these factors may potentially influence the hemodynamic or autonomic effects of aerobic fitness on atrial remodeling and the development of AF. However, left atrial and ventricular size have been reported to only partly explain the association between tall stature and the risk of AF (8), which also suggests the possibility of other mechanisms such as common genetic factors (e.g., pituitary homeobox 2 (PITX2) (57) and zinc finger homeobox 3 (ZFHX3) (58)) that are associated with growth pathways and with AF. Additional clinical and experimental studies are needed to elucidate these complex pathways, which may help further identify high-risk subgroups and ultimately new targets for intervention. Strengths of the present study include its large national cohort design with follow-up from age 18 years into adulthood. The national cohort design minimized the potential for selection bias, and the use of registry data with prospectively measured exposures prevented bias that may result from self-reporting. We examined well-validated, objective measurements of aerobic fitness level and muscular strength. We were able to adjust for other common risk factors that also were prospectively ascertained and not self-reported, including a family history of AF, individual and neighborhood-level SES, and chronic diseases that are associated with AF. Limitations included measurement of the study exposures at only a single age (18 years), which meant that we were unable to examine changes in these factors across time. Additional studies that contain data on longitudinal exposures are needed to delineate more specific age windows of susceptibility in relation to the risk of AF. The study cohort also was relatively young and exclusively male. The median age at the end of follow-up was 47 years (maximum, 62 years), and hence the findings may not necessarily apply to AF in older adults. The risk of AF is estimated to double with each decade of life (47), and 70% of prevalent cases are in adults 65 years of age or older (59). Further studies will be needed to assess our findings in older adults and in women. In summary, in the present study, we found that tall stature and high weight at age 18 years were independently associated with a higher risk of AF development in adulthood and interacted with the level of aerobic fitness. High aerobic fitness level was associated with a higher risk of AF among tall men but was protective among shorter men. These findings suggest important interactive effects of body size and aerobic fitness level on the development of AF. Tall, aerobically fit persons may be a relatively high-risk subgroup. ACKNOWLEDGMENTS Author affiliations: Alfred and Gail Engelberg Department of Family Medicine and Community Health, Icahn School of Medicine at Mount Sinai, New York, New York (Casey Crump, Jan Sundquist, Kristina Sundquist); Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, New York (Casey Crump); Center for Primary Health Care Research, Department of Clinical Sciences, Faculty of Medicine, Lund University, Malmö, Sweden (Jan Sundquist, Kristina Sundquist); and Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, California (Marilyn A. Winkleby). This work was supported by the National Heart, Lung, and Blood Institute at the National Institutes of Health (grant R01 HL116381); the Swedish Research Council; and the Avtal om Läkarutbildning och Forskning project grant from Region Skåne/Lund University, Malmö, Sweden. The funding agencies had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript. Conflicts of interest: none declared. Abbreviations AF atrial fibrillation BMI body mass index CI confidence interval HR hazard ratio ICD International Classification of Diseases SD standard deviation SES socioeconomic status REFERENCES 1 Chugh SS, Havmoeller R, Narayanan K, et al.  . Worldwide epidemiology of atrial fibrillation: a Global Burden of Disease 2010 Study. Circulation . 2014; 129( 8): 837– 847. Google Scholar CrossRef Search ADS PubMed  2 Benjamin EJ, Wolf PA, D’Agostino RB, et al.  . Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation . 1998; 98( 10): 946– 952. Google Scholar CrossRef Search ADS PubMed  3 Lip GY, Tse HF, Lane DA. Atrial fibrillation. Lancet . 2012; 379( 9816): 648– 661. Google Scholar CrossRef Search ADS PubMed  4 Kannel WB, Abbott RD, Savage DD, et al.  . Epidemiologic features of chronic atrial fibrillation: the Framingham study. N Engl J Med . 1982; 306( 17): 1018– 1022. Google Scholar CrossRef Search ADS PubMed  5 Krahn AD, Manfreda J, Tate RB, et al.  . The natural history of atrial fibrillation: incidence, risk factors, and prognosis in the Manitoba Follow-Up Study. Am J Med . 1995; 98( 5): 476– 484. Google Scholar CrossRef Search ADS PubMed  6 Wyse DG, Van Gelder IC, Ellinor PT, et al.  . Lone atrial fibrillation: does it exist? J Am Coll Cardiol . 2014; 63( 17): 1715– 1723. Google Scholar CrossRef Search ADS PubMed  7 Schmidt M, Botker HE, Pedersen L, et al.  . Adult height and risk of ischemic heart disease, atrial fibrillation, stroke, venous thromboembolism, and premature death: a population based 36-year follow-up study. Eur J Epidemiol . 2014; 29( 2): 111– 118. Google Scholar CrossRef Search ADS PubMed  8 Rosenberg MA, Patton KK, Sotoodehnia N, et al.  . The impact of height on the risk of atrial fibrillation: the Cardiovascular Health Study. Eur Heart J . 2012; 33( 21): 2709– 2717. Google Scholar CrossRef Search ADS PubMed  9 Schnabel RB, Sullivan LM, Levy D, et al.  . Development of a risk score for atrial fibrillation (Framingham Heart Study): a community-based cohort study. Lancet . 2009; 373( 9665): 739– 745. Google Scholar CrossRef Search ADS PubMed  10 Chamberlain AM, Agarwal SK, Folsom AR, et al.  . A clinical risk score for atrial fibrillation in a biracial prospective cohort (from the Atherosclerosis Risk in Communities [ARIC] study). Am J Cardiol . 2011; 107( 1): 85– 91. Google Scholar CrossRef Search ADS PubMed  11 Frost L, Hune LJ, Vestergaard P. Overweight and obesity as risk factors for atrial fibrillation or flutter: the Danish Diet, Cancer, and Health Study. Am J Med . 2005; 118( 5): 489– 495. Google Scholar CrossRef Search ADS PubMed  12 Grundvold I, Skretteberg PT, Liestol K, et al.  . Importance of physical fitness on predictive effect of body mass index and weight gain on incident atrial fibrillation in healthy middle-age men. Am J Cardiol . 2012; 110( 3): 425– 432. Google Scholar CrossRef Search ADS PubMed  13 Schmidt M, Botker HE, Pedersen L, et al.  . Comparison of the frequency of atrial fibrillation in young obese versus young nonobese men undergoing examination for fitness for military service. Am J Cardiol . 2014; 113( 5): 822– 826. Google Scholar CrossRef Search ADS PubMed  14 Berkovitch A, Kivity S, Klempfner R, et al.  . Body mass index and the risk of new-onset atrial fibrillation in middle-aged adults. Am Heart J . 2016; 173: 41– 48. Google Scholar CrossRef Search ADS PubMed  15 Aronis KN, Wang N, Phillips CL, et al.  . Associations of obesity and body fat distribution with incident atrial fibrillation in the biracial health aging and body composition cohort of older adults. Am Heart J . 2015; 170( 3): 498.e2– 505.e2. Google Scholar CrossRef Search ADS   16 Dublin S, French B, Glazer NL, et al.  . Risk of new-onset atrial fibrillation in relation to body mass index. Arch Intern Med . 2006; 166( 21): 2322– 2328. Google Scholar CrossRef Search ADS PubMed  17 Frost L, Benjamin EJ, Fenger-Gron M, et al.  . Body fat, body fat distribution, lean body mass and atrial fibrillation and flutter. A Danish cohort study. Obesity (Silver Spring) . 2014; 22( 6): 1546– 1552. Google Scholar CrossRef Search ADS PubMed  18 Karas MG, Yee LM, Biggs ML, et al.  . Measures of body size and composition and risk of incident atrial fibrillation in older people: the cardiovascular health study. Am J Epidemiol . 2016; 183( 11): 998– 1007. Google Scholar CrossRef Search ADS PubMed  19 Baldesberger S, Bauersfeld U, Candinas R, et al.  . Sinus node disease and arrhythmias in the long-term follow-up of former professional cyclists. Eur Heart J . 2008; 29( 1): 71– 78. Google Scholar CrossRef Search ADS PubMed  20 Karjalainen J, Kujala UM, Kaprio J, et al.  . Lone atrial fibrillation in vigorously exercising middle aged men: case-control study. BMJ . 1998; 316( 7147): 1784– 1785. Google Scholar CrossRef Search ADS PubMed  21 Mont L, Tamborero D, Elosua R, et al.  . Physical activity, height, and left atrial size are independent risk factors for lone atrial fibrillation in middle-aged healthy individuals. Europace . 2008; 10( 1): 15– 20. Google Scholar CrossRef Search ADS PubMed  22 Elosua R, Arquer A, Mont L, et al.  . Sport practice and the risk of lone atrial fibrillation: a case-control study. Int J Cardiol . 2006; 108( 3): 332– 337. Google Scholar CrossRef Search ADS PubMed  23 Molina L, Mont L, Marrugat J, et al.  . Long-term endurance sport practice increases the incidence of lone atrial fibrillation in men: a follow-up study. Europace . 2008; 10( 5): 618– 623. Google Scholar CrossRef Search ADS PubMed  24 Aizer A, Gaziano JM, Cook NR, et al.  . Relation of vigorous exercise to risk of atrial fibrillation. Am J Cardiol . 2009; 103( 11): 1572– 1577. Google Scholar CrossRef Search ADS PubMed  25 Faselis C, Kokkinos P, Tsimploulis A, et al.  . Exercise capacity and atrial fibrillation risk in veterans: a cohort study. Mayo Clin Proc . 2016; 91( 5): 558– 566. Google Scholar CrossRef Search ADS PubMed  26 Qureshi WT, Alirhayim Z, Blaha MJ, et al.  . Cardiorespiratory fitness and risk of incident atrial fibrillation: results from the Henry Ford exercise testing (FIT) project. Circulation . 2015; 131( 21): 1827– 1834. Google Scholar CrossRef Search ADS PubMed  27 Pathak RK, Elliott A, Middeldorp ME, et al.  . Impact of CARDIOrespiratory FITness on arrhythmia recurrence in obese individuals with atrial fibrillation: the CARDIO-FIT study. J Am Coll Cardiol . 2015; 66( 9): 985– 996. Google Scholar CrossRef Search ADS PubMed  28 Andersen K, Rasmussen F, Held C, et al.  . Exercise capacity and muscle strength and risk of vascular disease and arrhythmia in 1.1 million young Swedish men: cohort study. BMJ . 2015; 351: h4543. Google Scholar CrossRef Search ADS PubMed  29 Kapa S, Asirvatham SJ. A MET a day keeps arrhythmia at bay: the association between exercise or cardiorespiratory fitness and atrial fibrillation. Mayo Clin Proc . 2016; 91( 5): 545– 550. Google Scholar CrossRef Search ADS PubMed  30 McClean G, George K, Lord R, et al.  . Chronic adaptation of atrial structure and function in elite male athletes. Eur Heart J Cardiovasc Imaging . 2015; 16( 4): 417– 422. Google Scholar CrossRef Search ADS PubMed  31 Badano LP, Miglioranza MH, Mihaila S, et al.  . Left atrial volumes and function by three-dimensional echocardiography: reference values, accuracy, reproducibility, and comparison with two-dimensional echocardiographic measurements. Circ Cardiovasc Imaging . 2016; 9( 7): e004229. Google Scholar CrossRef Search ADS PubMed  32 Iwasaki YK, Nishida K, Kato T, et al.  . Atrial fibrillation pathophysiology: implications for management. Circulation . 2011; 124( 20): 2264– 2274. Google Scholar CrossRef Search ADS PubMed  33 Crump C, Sundquist J, Winkleby MA, et al.  . Interactive effects of physical fitness and body mass index on the risk of hypertension. JAMA Intern Med . 2016; 176( 2): 210– 216. Google Scholar CrossRef Search ADS PubMed  34 Crump C, Sundquist J, Winkleby MA, et al.  . Physical fitness among Swedish military conscripts and long-term risk for type 2 diabetes mellitus: a cohort study. Ann Intern Med . 2016; 164( 9): 577– 584. Google Scholar CrossRef Search ADS PubMed  35 Crump C, Sundquist J, Winkleby MA, et al.  . Interactive effects of obesity and physical fitness on risk of ischemic heart disease. Int J Obes (Lond) . 2017; 41( 2): 255– 261. Google Scholar CrossRef Search ADS PubMed  36 Crump C, Sundquist J, Winkleby MA, et al.  . Interactive effects of physical fitness and body mass index on risk of stroke: a national cohort study. Int J Stroke . 2016; 11( 6): 683– 694. Google Scholar CrossRef Search ADS PubMed  37 Crump C, Sundquist J, Winkleby MA, et al.  . Interactive effects of aerobic fitness, strength, and obesity on mortality in men. Am J Prev Med . 2017; 52( 3): 353– 361. Google Scholar CrossRef Search ADS PubMed  38 Ogden CL, Flegal KM. Changes in terminology for childhood overweight and obesity. Natl Health Stat Report . 2010;( 25): 1– 5. 39 Nordesjö LO, Schele R. Validity of an ergometer cycle test and measures of isometric muscle strength when predicting some aspects of military performance. Swedish J Defence Med . 1974; 10: 11– 23. 40 Patton JF, Vogel JA, Mello RP. Evaluation of a maximal predictive cycle ergometer test of aerobic power. Eur J Appl Physiol Occup Physiol . 1982; 49( 1): 131– 140. Google Scholar CrossRef Search ADS PubMed  41 Andersen LB. A maximal cycle exercise protocol to predict maximal oxygen uptake. Scand J Med Sci Sports . 1995; 5( 3): 143– 146. Google Scholar CrossRef Search ADS PubMed  42 Hook O, Tornvall G. Apparatus and method for determination of isometric muscle strength in man. Scand J Rehabil Med . 1969; 1( 3): 139– 142. Google Scholar PubMed  43 Smith JG, Platonov PG, Hedblad B, et al.  . Atrial fibrillation in the Malmö Diet and Cancer study: a study of occurrence, risk factors and diagnostic validity. Eur J Epidemiol . 2010; 25( 2): 95– 102. Google Scholar CrossRef Search ADS PubMed  44 Hoehner CM, Allen P, Barlow CE, et al.  . Understanding the independent and joint associations of the home and workplace built environments on cardiorespiratory fitness and body mass index. Am J Epidemiol . 2013; 178( 7): 1094– 1105. Google Scholar CrossRef Search ADS PubMed  45 Cubbin C, Sundquist K, Ahlen H, et al.  . Neighborhood deprivation and cardiovascular disease risk factors: protective and harmful effects. Scand J Public Health . 2006; 34( 3): 228– 237. Google Scholar PubMed  46 Crump C, Sundquist K, Sundquist J, et al.  . Neighborhood deprivation and psychiatric medication prescription: a Swedish national multilevel study. Ann Epidemiol . 2011; 21( 4): 231– 237. Google Scholar CrossRef Search ADS PubMed  47 Benjamin EJ, Levy D, Vaziri SM, et al.  . Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA . 1994; 271( 11): 840– 844. Google Scholar CrossRef Search ADS PubMed  48 Rubin DB. Multiple Imputation for Nonresponse in Surveys . New York, NY: Wiley; 1987. Google Scholar CrossRef Search ADS   49 Vanderweele TJ, Knol MJ. A tutorial on interaction. Epidemiol Methods . 2014; 3( 1): 33– 72. 50 Li R, Chambless L. Test for additive interaction in proportional hazards models. Ann Epidemiol . 2007; 17( 3): 227– 236. Google Scholar CrossRef Search ADS PubMed  51 Richardson DB, Kaufman JS. Estimation of the relative excess risk due to interaction and associated confidence bounds. Am J Epidemiol . 2009; 169( 6): 756– 760. Google Scholar CrossRef Search ADS PubMed  52 Azarbal F, Stefanick ML, Assimes TL, et al.  . Lean body mass and risk of incident atrial fibrillation in post-menopausal women. Eur Heart J . 2016; 37( 20): 1606– 1613. Google Scholar CrossRef Search ADS PubMed  53 Frost L, Frost P, Vestergaard P. Work related physical activity and risk of a hospital discharge diagnosis of atrial fibrillation or flutter: the Danish Diet, Cancer, and Health Study. Occup Environ Med . 2005; 62( 1): 49– 53. Google Scholar CrossRef Search ADS PubMed  54 Everett BM, Conen D, Buring JE, et al.  . Physical activity and the risk of incident atrial fibrillation in women. Circ Cardiovasc Qual Outcomes . 2011; 4( 3): 321– 327. Google Scholar CrossRef Search ADS PubMed  55 Mozaffarian D, Furberg CD, Psaty BM, et al.  . Physical activity and incidence of atrial fibrillation in older adults: the cardiovascular health study. Circulation . 2008; 118( 8): 800– 807. Google Scholar CrossRef Search ADS PubMed  56 Abdulla J, Nielsen JR. Is the risk of atrial fibrillation higher in athletes than in the general population? A systematic review and meta-analysis. Europace . 2009; 11( 9): 1156– 1159. Google Scholar CrossRef Search ADS PubMed  57 Gudbjartsson DF, Arnar DO, Helgadottir A, et al.  . Variants conferring risk of atrial fibrillation on chromosome 4q25. Nature . 2007; 448( 7151): 353– 357. Google Scholar CrossRef Search ADS PubMed  58 Benjamin EJ, Rice KM, Arking DE, et al.  . Variants in ZFHX3 are associated with atrial fibrillation in individuals of European ancestry. Nat Genet . 2009; 41( 8): 879– 881. Google Scholar CrossRef Search ADS PubMed  59 Go AS, Hylek EM, Phillips KA, et al.  . Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA . 2001; 285( 18): 2370– 2375. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Epidemiology Oxford University Press

Height, Weight, and Aerobic Fitness Level in Relation to the Risk of Atrial Fibrillation

Loading next page...
 
/lp/ou_press/height-weight-and-aerobic-fitness-level-in-relation-to-the-risk-of-b0QqKMgTJo
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0002-9262
eISSN
1476-6256
D.O.I.
10.1093/aje/kwx255
Publisher site
See Article on Publisher Site

Abstract

Abstract Tall stature and obesity have been associated with a higher risk of atrial fibrillation (AF), but there have been conflicting reports of the effects of aerobic fitness. We conducted a national cohort study to examine interactions between height or weight and level of aerobic fitness among 1,547,478 Swedish military conscripts during 1969–1997 (97%–98% of all 18-year-old men) in relation to AF identified from nationwide inpatient and outpatient diagnoses through 2012 (maximal age, 62 years). Increased height, weight, and aerobic fitness level (but not muscular strength) at age 18 years were all associated with a higher AF risk in adulthood. Positive additive and multiplicative interactions were found between height or weight and aerobic fitness level (for the highest tertiles of height and aerobic fitness level vs. the lowest, relative excess risk = 0.51, 95% confidence interval (CI): 0.40, 0.62; ratio of hazard ratios = 1.50, 95% CI: 1.34, 1.65). High aerobic fitness levels were associated with higher risk among men who were at least 186 cm (6 feet, 1 inch) tall but were protective among shorter men. Men with the combination of tall stature and high aerobic fitness level had the highest risk (for the highest tertiles vs. the lowest, adjusted hazard ratio = 1.70, 95% CI: 1.61, 1.80). These findings suggest important interactions between body size and aerobic fitness level in relation to AF and may help identify high-risk subgroups. atrial fibrillation, body height, body mass index, body size, body weight, muscle strength, physical fitness Atrial fibrillation (AF) is the most common cardiac arrhythmia; it affects more than 30 million people worldwide (1) and is an important cause of stroke and mortality (2, 3). AF is more common with increasing age or underlying heart disease (4, 5). However, many cases of AF also occur among younger adults who do not have known heart disease (6), which has led to investigations of other risk factors and potential targets for intervention. Researchers from prior studies have established that tall stature (7–11) and obesity (11–18) are associated with a higher risk of the development of AF. However, the reported effects of aerobic fitness have been conflicting. In a number of studies, investigators have reported an increased risk of AF among competitive athletes or individuals who report high levels of exercise (19–24). In contrast, results from several (12, 25–27) but not all (28) studies in which objectively measured levels of aerobic fitness were included have suggested that a high level of aerobic fitness is protective against AF. We hypothesized that these conflicting findings may be due to interactions between height or weight and aerobic fitness level that have not been examined. The pathogenesis of AF is a heterogeneous process that involves myocardial structural, hemodynamic, electrical, and neural factors (29). Physical exercise may potentially influence this process through its contribution to left ventricular hypertrophy, which leads to diastolic dysfunction and left atrial stretch (30). Because height and weight correlate positively with left atrial size (31), these variables may potentially modify the hemodynamic and autonomic effects of aerobic fitness on atrial remodeling and the development of AF (32). If so, elucidation of interactions among these factors may help identify high-risk subgroups and provide new insights into underlying mechanisms. We examined the interactions among height, weight, and aerobic fitness level in relation to the risk of AF in a large national cohort. Standardized measurements of these factors were obtained in approximately 1.5 million 18-year-old male military conscripts in Sweden who were subsequently followed up for AF in adulthood. In the present study, our aim was to determine the interactive effects of these common factors at age 18 years on the long-term risk of AF in a large, population-based cohort. METHODS Study population We identified 1,547,478 men (approximately 18 years of age) who underwent a military conscription examination in Sweden during 1969–1997. Nationally, all 18-year-old men were conscripted each year except for 2%–3% who either were incarcerated or had severe chronic medical conditions or disabilities that were documented by a physician. The Regional Ethics Committee of Lund University in Sweden approved this study. Ascertainment of height, weight, and aerobic fitness level We obtained measurements of height, weight, and aerobic fitness level from the Swedish Military Conscription Registry, which contains information from 2-day standardized physical and psychological examinations required of all conscripts starting in 1969 (33–37). We measured height and weight using standard protocols and examined them alternatively as continuous variables or categorical variables in tertiles (low, medium, and high tertiles for height: <175.0 cm, 175.0–181.4 cm, and ≥181.5 cm, respectively; for weight, <64.0 kg, 64.0–71.3 kg, and ≥71.4 kg, respectively). Categorizations for these and other predictors were based on dose-response analyses, which indicated that they well-characterized the underlying relationships with AF risk and had good interpretability. We examined body mass index (BMI) as an alternative to height and weight. We calculated BMI as weight in kilograms divided by the square of height in meters. We examined BMI alternatively as a continuous or categorical variable using Centers for Disease Control and Prevention definitions for children and adolescents 2–19 years of age to facilitate comparability with US studies. “Overweight” was defined as being in the 85th percentile or higher but less than 95th percentile and “obesity” as being in the 95th percentile or higher on the sex-specific BMI-for-age growth charts, issued in 2000; these values correspond to BMIs of 25.6–28.9 and 29.0 or higher, respectively, for 18-year-old men (38). The level of aerobic fitness was measured as the maximal aerobic workload in watts using a well-validated, electrically braked stationary bicycle ergometer test (39). After a warm-up period, each conscript performed 5–10 minutes of exercise on a stationary bicycle at a starting work rate of 75–175 W (determined on a sliding scale based on body weight) and increasing by 25 W per minute until volitional exhaustion. Maximal aerobic workload was calculated as the power output in watts before the last increase in intensity plus the prorated output for the last stage (39). Maximal aerobic workload correlates highly with maximal oxygen uptake (correlation ≈ 0.9) (40), and its measurement using this bicycle ergometer test is highly reproducible, with a test-retest correlation of 0.95 (41). We examined this measurement of aerobic fitness alternatively as a continuous variable or categorical variable in tertiles (<240 W, 240–288 W, and ≥289 W) (33–37). In addition to aerobic fitness level, we examined muscular strength as a secondary predictor of interest because it represents a different aspect of physical fitness. Muscular strength was measured in Newtons using well-validated isometric dynamometer tests and calculated as the weighted sum of maximal knee extension (weighted × 1.3), elbow flexion (weighted × 0.8), and hand grip (weighted × 1.7) (42). Each dynamometer test was performed 3 times, and the maximal value was recorded for analysis, except when the last value was highest, in which case testing was repeated until strength values stopped increasing. All testing equipment was calibrated daily (42). We examined muscular strength alternatively as a continuous variable or categorical variable in tertiles (<1,900 N, 1,900–2,170 N, and ≥2,171 N) (33–37). Ascertainment of AF We followed up the study cohort for the earliest diagnosis of AF, from the date of the military conscription examination through December 31, 2012. AF was identified based on any primary or secondary diagnosis according to codes from the International Classification of Diseases (ICD) listed in the Swedish Hospital and Outpatient Registries (ICD codes 427.4 (Eighth Revision), 427.3 (Ninth Revision), and I48 (Tenth Revision)). The Swedish Hospital Registry contains all primary and secondary hospital discharge diagnoses from 6 populous counties in southern Sweden starting in 1964, and with nationwide coverage starting in 1987. The Swedish Outpatient Registry also contains outpatient diagnoses from all specialty clinics nationwide starting in 2001 (33–37). Diagnosis records in the Hospital Registry are currently more than 99% complete and have a reported positive predictive value of 97% for either primary or secondary diagnoses of AF (43). Adjustment variables Data on other variables that may be associated with body size or physical fitness level and the risk of AF were obtained from the Swedish Military Conscription Registry and national census data, which were linked by use of an anonymous personal identification number (33–37). The following were used as adjustment variables: year that the military conscription examination was conducted (as a continuous variable); highest educational level attained during the study period (<12, 12–14, or ≥15 years); neighborhood socioeconomic status (SES) at baseline; and a family history of AF in a parent or sibling (yes or no, identified from the Swedish Hospital Registry during 1965–2012 and the Swedish Outpatient Registry during 2001–2012, using the same diagnosis codes noted above). SES was included because characteristics of neighborhood SES have been associated with obesity and physical fitness (44) and with the risk of cardiovascular disease (45). It was composed of an index that included low educational level, low income status, unemployment, and receipt of social welfare benefits, as previously described (46), and categorized as low (more than 1 SD below the mean), medium (1 SD below to 1 SD above the mean), or high (more than 1 SD above the mean). Because low levels of aerobic fitness and a high BMI are known to be associated with hypertension (33), diabetes mellitus (34), and ischemic heart disease (35), which are established risk factors for AF (47), we were interested in whether aerobic fitness level and BMI (or height and weight) are associated with AF independently of these conditions. To assess this, we further adjusted for hypertension (ICD codes 400–401 (Eighth Revision), 401 (Ninth Revision), and I10 (Tenth Revision)), diabetes mellitus (ICD code 250 (Eighth and Ninth Revisions) and E10–E14 (Tenth Revision)), and ischemic heart disease (ICD codes 410–414 (Eighth and Ninth Revisions) and I20–I25 (Tenth Revision)) as time-dependent variables. We imputed missing data for each variable using a standard multiple imputation procedure based on the relationship of the variable with all other covariates and AF (48). Missing data were relatively infrequent for height (7.2%), weight (7.3%), level of aerobic fitness (5.7%), muscular strength (5.0%), educational level (0.4%), and neighborhood SES (9.1%). Data were 100% complete for all other variables. Statistical analysis We used Cox proportional hazards regression models to compute hazard ratios and 95% confidence intervals for associations of height, weight, aerobic fitness level, or muscular strength with the risk of AF. The Cox model time scale was elapsed time since the military conscription examination, which also corresponds to attained age because the baseline age was the same (18 years) for all conscripts. Individuals were censored at emigration (n = 87,450; 5.7%) or death (n = 58,835; 3.8%). We used 3 models. The first was adjusted for attained age (as the time scale) and year of the military conscription examination; the second was also adjusted for height, weight, level of aerobic fitness, muscular strength, educational level, neighborhood SES, and family history of AF; and the third was further adjusted for hypertension, diabetes mellitus, and ischemic heart disease (as defined above). All time-varying factors were modeled as time-dependent covariates. The proportional hazards assumption was evaluated using graphical assessment of log-log plots, which showed a good fit in all models. Interactions among height, weight, aerobic fitness level, and muscular strength were examined on both the additive and multiplicative scale in relation to the risk of AF. We assessed additive interactions using the relative excess risk due to interaction, which is computed for binary variables as: RERIHR = HR11 − HR10 − HR01 + 1 (49, 50), where HR is the hazard ratio. Multiplicative interactions were assessed using the ratio of hazard ratios: HR11/(HR10 × HR01) (49, 50). These estimates and their respective 95% confidence intervals were determined using maximal likelihood estimation (49–51). A positive additive interaction is indicated if the relative excess risk due to interaction is greater than 0, and a positive multiplicative interaction is indicated if the ratio of hazard ratios is more than 1. See the Web Appendix (available at https://academic.oup.com/aje) for more details on these methods. We performed 2 sensitivity analyses. First, as an alternative to multiple imputation for missing data, we repeated all analyses after restricting to men for whom we had complete data for all variables (n = 1,361,083; 88.0%). Second, because of changes in AF ascertainment across time, we assessed different starting points for the follow-up period, alternatively starting in 1987 (at which time inpatient diagnoses were available nationwide instead of only for the most populous counties) or in 2001 (at which time both inpatient and outpatient diagnoses were available nationwide). All statistical tests were 2-sided, and an α-level of 0.05 was used. We conducted all analyses using Stata, version 14.1 (StataCorp LP, College Station, Texas). RESULTS Among the 1,547,478 men in this cohort, 23,600 (1.5%) were diagnosed with AF at 43.7 million person-years of follow-up (mean follow-up = 28.2 years). The median age at the end of follow-up was 47.2 years (mean = 47.4 (SD, 7.9) years; range, 19.0–62.0 years). The median age at diagnosis of AF was 48.2 years (mean = 47.1 (SD, 7.8) years; range, 18.2–62.0 years). Table 1 shows incidence rates of AF by height, weight, aerobic fitness level, and other factors. Table 1. Characteristics of Men Who Were or Were Not Subsequently Diagnosed With Atrial Fibrillation, Sweden, 1969–2012 Variable  No AF  AF  Ratea  No.  %  No.  %  Total  1,523,818  100.0  23,660  100.0  54.3  Tertile of height             Lowest (<175.0 cm)  527,967  34.6  6,888  29.1  44.8   Middle (175.0–181.4 cm)  504,217  33.1  6,801  28.7  47.4   Highest (≥181.5 cm)  491,634  32.3  9,971  42.1  71.9  Tertile of weight             Lowest (<64.0 kg)  500,063  32.8  6,784  28.7  45.1   Middle (64.0–71.3 kg)  506,366  33.2  7,061  29.8  49.2   Highest (≥71.4 kg)  517,389  34.0  9,815  41.5  69.1  BMIb             Normal  1,406,267  92.3  21,248  89.8  52.6   Overweight  83,124  5.4  1,544  6.5  68.8   Obesity  34,427  2.3  868  3.7  96.2  Tertile of aerobic fitness level             Lowest (<240 W)  501,335  32.9  10,015  42.3  60.3   Middle (240–288 W)  511,945  33.6  8,602  36.4  57.4   Highest (≥289 W)  510,538  33.5  5,043  21.3  42.0  Tertile of muscular strength             Lowest (<1,900 N)  504,507  33.1  6,418  27.1  46.4   Middle (1,900–2,170 N)  515,024  33.8  8,528  36.0  55.2   Highest (≥2,171 N)  504,287  33.1  8,714  36.8  60.9  Education, years             <12  232,151  15.2  4,693  19.8  65.9   12–14  674,001  44.2  9,626  40.7  51.0   ≥15  617,666  40.5  9,341  39.5  53.2  Neighborhood SES             Low  235,421  15.5  3,973  16.8  57.4   Medium  1,006,116  66.0  16,119  68.1  55.4   High  282,281  18.5  3,568  15.1  47.1  Hypertension             No  1,438,676  94.4  15,478  65.4  37.7   Yes  85,142  5.6  8,182  34.6  257.5  Diabetes mellitus             No  1,475,411  96.8  20,836  88.1  49.7   Yes  48,407  3.2  2,824  11.9  169.2  Ischemic heart disease             No  1,488,996  97.7  20,269  85.7  48.0   Yes  34,822  2.3  3,391  14.3  255.8  Family history of AF             No  1,223,751  80.3  14,986  63.3  44.0   Yes  300,067  19.7  8,674  36.7  91.2  Variable  No AF  AF  Ratea  No.  %  No.  %  Total  1,523,818  100.0  23,660  100.0  54.3  Tertile of height             Lowest (<175.0 cm)  527,967  34.6  6,888  29.1  44.8   Middle (175.0–181.4 cm)  504,217  33.1  6,801  28.7  47.4   Highest (≥181.5 cm)  491,634  32.3  9,971  42.1  71.9  Tertile of weight             Lowest (<64.0 kg)  500,063  32.8  6,784  28.7  45.1   Middle (64.0–71.3 kg)  506,366  33.2  7,061  29.8  49.2   Highest (≥71.4 kg)  517,389  34.0  9,815  41.5  69.1  BMIb             Normal  1,406,267  92.3  21,248  89.8  52.6   Overweight  83,124  5.4  1,544  6.5  68.8   Obesity  34,427  2.3  868  3.7  96.2  Tertile of aerobic fitness level             Lowest (<240 W)  501,335  32.9  10,015  42.3  60.3   Middle (240–288 W)  511,945  33.6  8,602  36.4  57.4   Highest (≥289 W)  510,538  33.5  5,043  21.3  42.0  Tertile of muscular strength             Lowest (<1,900 N)  504,507  33.1  6,418  27.1  46.4   Middle (1,900–2,170 N)  515,024  33.8  8,528  36.0  55.2   Highest (≥2,171 N)  504,287  33.1  8,714  36.8  60.9  Education, years             <12  232,151  15.2  4,693  19.8  65.9   12–14  674,001  44.2  9,626  40.7  51.0   ≥15  617,666  40.5  9,341  39.5  53.2  Neighborhood SES             Low  235,421  15.5  3,973  16.8  57.4   Medium  1,006,116  66.0  16,119  68.1  55.4   High  282,281  18.5  3,568  15.1  47.1  Hypertension             No  1,438,676  94.4  15,478  65.4  37.7   Yes  85,142  5.6  8,182  34.6  257.5  Diabetes mellitus             No  1,475,411  96.8  20,836  88.1  49.7   Yes  48,407  3.2  2,824  11.9  169.2  Ischemic heart disease             No  1,488,996  97.7  20,269  85.7  48.0   Yes  34,822  2.3  3,391  14.3  255.8  Family history of AF             No  1,223,751  80.3  14,986  63.3  44.0   Yes  300,067  19.7  8,674  36.7  91.2  Abbreviations: AF, atrial fibrillation; BMI, body mass index; SES, socioeconomic status. a AF incidence rate per 100,000 person-years. b Weight (kg)/height (m)2. Main effects of height, weight, and aerobic fitness level Tall stature was associated with a higher risk of AF after adjustment for weight and all other covariates (Table 2, model 3) (for the highest tertile vs. the lowest, HR = 1.53, 95% confidence interval (CI): 1.48, 1.59; P < 0.001). Weight in the highest (but not middle) tertile was associated with a modestly higher risk of AF compared with weight in the lowest tertile after adjustment for height and all other covariates (for the highest tertile vs. the lowest, HR = 1.18, 95% CI: 1.13, 1.23; P < 0.001). Both height and weight had strongly positive linear associations with the risk of AF across their full distribution (per each 5-cm increase in height, fully adjusted HR = 1.11, 95% CI: 1.10, 1.12; P for trend < 0.001; per each 5-kg increase in weight, fully adjusted HR = 1.05, 95% CI: 1.04, 1.06; P for trend < 0.001), but height was a stronger risk factor than was weight (P for heterogeneity < 0.001). Obesity (but not overweight) also was associated with an increased risk of AF in the fully adjusted model (Table 2, model 3). Table 2. Associations of Height, Weight, Aerobic Fitness Level, or Other Factors With Risk of Atrial Fibrillation, Sweden, 1969–2012 Variable  Model 1a  Model 2b  Model 3c  HR  95% CI  P Value  HR  95% CI  P Value  HR  95% CI  P Value  Height                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.09  1.05, 1.13  <0.001  1.00  0.96, 1.03  0.85  1.05  1.01, 1.09  0.006   High  1.70  1.65, 1.75  <0.001  1.47  1.42, 1.52  <0.001  1.53  1.48, 1.59  <0.001   Per 5 cm (trend test)  1.14  1.13, 1.15  <0.001  1.04  1.03, 1.05  <0.001  1.11  1.10, 1.12  <0.001  Weight                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.18  1.14, 1.22  <0.001  1.19  1.15, 1.24  <0.001  0.98  0.94, 1.02  0.28   High  1.79  1.74, 1.85  <0.001  1.75  1.68, 1.82  <0.001  1.18  1.13, 1.23  <0.001   Per 5 kg (trend test)  1.14  1.13, 1.15  <0.001  1.13  1.12, 1.13  <0.001  1.05  1.04, 1.06  <0.001  BMId                     Normal  1.00  Referent    1.00  Referent    1.00  Referent     Overweight  1.47  1.40, 1.55  <0.001  1.35  1.28, 1.42  <0.001  1.05  0.99, 1.10  0.09   Obesity  2.22  2.07, 2.38  <0.001  2.04  1.91, 2.19  <0.001  1.31  1.22, 1.41  <0.001   Per 1 BMI unit (trend test)  1.06  1.05, 1.06  <0.001  1.05  1.05, 1.06  <0.001  1.02  1.01, 1.03  <0.001  Aerobic fitness level                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.14  1.11, 1.18  <0.001  0.94  0.91, 0.97  <0.001  1.02  0.99, 1.05  0.31   High  1.36  1.31, 1.41  <0.001  1.00  0.96, 1.05  0.84  1.14  1.09, 1.19  <0.001   Per 100 W (trend test)  1.31  1.27, 1.35  <0.001  0.99  0.95, 1.02  0.41  1.12  1.08, 1.16  <0.001  Muscular strength                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.14  1.10, 1.17  <0.001  1.01  0.98, 1.05  0.42  1.04  1.00, 1.07  0.04   High  1.37  1.33, 1.42  <0.001  0.99  0.96, 1.03  0.66  1.03  0.99, 1.07  0.13   Per 1,000 N (trend test)  1.44  1.39, 1.50  <0.001  0.94  0.90, 0.98  0.006  1.00  0.95, 1.04  0.88  Education, years                     <12  1.00  Referent    1.00  Referent    1.00  Referent     12–14  0.92  0.89, 0.95  <0.001  0.89  0.86, 0.92  <0.001  0.90  0.87, 0.94  <0.001   ≥15  0.93  0.89, 0.96  <0.001  0.89  0.86, 0.92  <0.001  0.95  0.91, 0.98  0.02   Per higher category (trend test)  0.97  0.95, 0.99  <0.001  0.95  0.93, 0.97  <0.001  0.98  0.96, 1.00  0.06  Neighborhood SES                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.01  0.98, 1.05  0.48  1.04  1.00, 1.08  0.03  1.06  1.03, 1.10  0.001   High  0.96  0.92, 1.01  0.10  0.95  0.91, 1.00  0.04  0.98  0.93, 1.02  0.36   Per higher category (trend test)  0.98  0.96, 1.00  0.12  0.98  0.96, 1.00  0.06  0.99  0.97, 1.01  0.50  Hypertension                     No  1.00  Referent          1.00  Referent     Yes  5.67  5.52, 5.83  <0.001        4.59  4.45, 4.74  <0.001  Diabetes mellitus                     No  1.00  Referent          1.00  Referent     Yes  2.89  2.77, 3.00  <0.001        1.21  1.16, 2.26  <0.001  Ischemic heart disease                     No  1.00  Referent          1.00  Referent     Yes  4.25  4.10, 4.42  <0.001        2.09  2.01, 2.18  <0.001  Family history of AF                     No  1.00  Referent    1.00      1.00  Referent     Yes  1.80  1.76, 1.85  <0.001  1.73  1.68, 1.78  <0.001  1.72  1.67, 1.77  <0.001  Variable  Model 1a  Model 2b  Model 3c  HR  95% CI  P Value  HR  95% CI  P Value  HR  95% CI  P Value  Height                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.09  1.05, 1.13  <0.001  1.00  0.96, 1.03  0.85  1.05  1.01, 1.09  0.006   High  1.70  1.65, 1.75  <0.001  1.47  1.42, 1.52  <0.001  1.53  1.48, 1.59  <0.001   Per 5 cm (trend test)  1.14  1.13, 1.15  <0.001  1.04  1.03, 1.05  <0.001  1.11  1.10, 1.12  <0.001  Weight                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.18  1.14, 1.22  <0.001  1.19  1.15, 1.24  <0.001  0.98  0.94, 1.02  0.28   High  1.79  1.74, 1.85  <0.001  1.75  1.68, 1.82  <0.001  1.18  1.13, 1.23  <0.001   Per 5 kg (trend test)  1.14  1.13, 1.15  <0.001  1.13  1.12, 1.13  <0.001  1.05  1.04, 1.06  <0.001  BMId                     Normal  1.00  Referent    1.00  Referent    1.00  Referent     Overweight  1.47  1.40, 1.55  <0.001  1.35  1.28, 1.42  <0.001  1.05  0.99, 1.10  0.09   Obesity  2.22  2.07, 2.38  <0.001  2.04  1.91, 2.19  <0.001  1.31  1.22, 1.41  <0.001   Per 1 BMI unit (trend test)  1.06  1.05, 1.06  <0.001  1.05  1.05, 1.06  <0.001  1.02  1.01, 1.03  <0.001  Aerobic fitness level                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.14  1.11, 1.18  <0.001  0.94  0.91, 0.97  <0.001  1.02  0.99, 1.05  0.31   High  1.36  1.31, 1.41  <0.001  1.00  0.96, 1.05  0.84  1.14  1.09, 1.19  <0.001   Per 100 W (trend test)  1.31  1.27, 1.35  <0.001  0.99  0.95, 1.02  0.41  1.12  1.08, 1.16  <0.001  Muscular strength                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.14  1.10, 1.17  <0.001  1.01  0.98, 1.05  0.42  1.04  1.00, 1.07  0.04   High  1.37  1.33, 1.42  <0.001  0.99  0.96, 1.03  0.66  1.03  0.99, 1.07  0.13   Per 1,000 N (trend test)  1.44  1.39, 1.50  <0.001  0.94  0.90, 0.98  0.006  1.00  0.95, 1.04  0.88  Education, years                     <12  1.00  Referent    1.00  Referent    1.00  Referent     12–14  0.92  0.89, 0.95  <0.001  0.89  0.86, 0.92  <0.001  0.90  0.87, 0.94  <0.001   ≥15  0.93  0.89, 0.96  <0.001  0.89  0.86, 0.92  <0.001  0.95  0.91, 0.98  0.02   Per higher category (trend test)  0.97  0.95, 0.99  <0.001  0.95  0.93, 0.97  <0.001  0.98  0.96, 1.00  0.06  Neighborhood SES                     Low  1.00  Referent    1.00  Referent    1.00  Referent     Medium  1.01  0.98, 1.05  0.48  1.04  1.00, 1.08  0.03  1.06  1.03, 1.10  0.001   High  0.96  0.92, 1.01  0.10  0.95  0.91, 1.00  0.04  0.98  0.93, 1.02  0.36   Per higher category (trend test)  0.98  0.96, 1.00  0.12  0.98  0.96, 1.00  0.06  0.99  0.97, 1.01  0.50  Hypertension                     No  1.00  Referent          1.00  Referent     Yes  5.67  5.52, 5.83  <0.001        4.59  4.45, 4.74  <0.001  Diabetes mellitus                     No  1.00  Referent          1.00  Referent     Yes  2.89  2.77, 3.00  <0.001        1.21  1.16, 2.26  <0.001  Ischemic heart disease                     No  1.00  Referent          1.00  Referent     Yes  4.25  4.10, 4.42  <0.001        2.09  2.01, 2.18  <0.001  Family history of AF                     No  1.00  Referent    1.00      1.00  Referent     Yes  1.80  1.76, 1.85  <0.001  1.73  1.68, 1.78  <0.001  1.72  1.67, 1.77  <0.001  Abbreviations: AF, atrial fibrillation; BMI, body mass index; CI, confidence interval; HR, hazard ratio; SES, socioeconomic status. a Adjusted for age and year of military conscription examination. b Adjusted for age, year of military conscription examination, height, weight, aerobic fitness level, muscular strength, educational level, neighborhood SES, and family history of AF. c Adjusted for age, year of military conscription examination, height, weight, aerobic fitness level, muscular strength, educational level, neighborhood SES, hypertension, diabetes mellitus, ischemic heart disease, and family history of AF. d BMI was calculated as weight (kg)/height (m)2 and examined as an alternative to height and weight in a separate model. High levels of aerobic fitness were associated with a modestly higher risk of AF after adjustment for height, weight, and all other covariates (for the highest tertile vs. the lowest, HR = 1.14, 95% CI: 1.09, 1.19; P < 0.001). In contrast, muscular strength was not clearly associated with AF risk (Table 2). A first-degree family history of AF was associated with a 1.7-fold higher risk of AF. Sensitivity analyses that were restricted to men for whom we had complete data or in which we examined different follow-up periods (as noted above) yielded risk estimates similar to those from the main analyses, and the overall conclusions were unchanged. Interactions among height, weight, and aerobic fitness level The interaction between height and aerobic fitness level in relation to the risk for AF is shown in Table 3. High levels of aerobic fitness were protective against AF among shorter men (lowest tertile for height) but were associated with increased risk among taller men (medium or highest tertiles; see Web Table 1 for more complete reporting of stratum-specific HRs). Men with the combination of tall height and high aerobic fitness level had the highest AF risk, which was 70% higher relative to those with short height and low aerobic fitness level (for the highest tertiles vs. the lowest for both variables, adjusted HR = 1.70, 95% CI: 1.61, 1.80; P < 0.001). There was a significant positive interaction between height and aerobic fitness level on both the additive and multiplicative scales (i.e., the association of both factors together with AF risk exceeded the sum or product of their associations considered separately; P < 0.001). Figure 1 shows the probability of AF among men in the 25th, 50th, and 75th percentiles of aerobic fitness across the full distribution of height after adjustment for all covariates. The nonparallel lines reflect a positive interaction. Specifically, high aerobic fitness levels were associated with a higher risk of AF among men who were at least 186 cm (6 feet, 1 inch) tall but were protective among men who were less than 186 cm tall. Table 3. Interaction of Height and Aerobic Fitness Level in Relation to Risk of Atrial Fibrillation,a Adjusted for Weight and Other Factors,b Sweden, 1969–2012 Tertile of Height  Tertile of Aerobic Fitness Level  Low  Medium  High  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  Low  4,569  54.1  1.00  Referent  1,756  39.3  0.89  0.84, 0.94  565  23.0  0.85  0.77, 0.93  Medium  2,633  55.3  0.95  0.90, 0.99  2,707  50.4  1.03  0.98, 1.09  1,460  34.6  1.15  1.07, 1.23  High  2,813  82.7  1.34  1.28, 1.41  4,139  80.5  1.49  1.42, 1.57  3,018  56.6  1.70  1.61, 1.80  Tertile of Height  Tertile of Aerobic Fitness Level  Low  Medium  High  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  Low  4,569  54.1  1.00  Referent  1,756  39.3  0.89  0.84, 0.94  565  23.0  0.85  0.77, 0.93  Medium  2,633  55.3  0.95  0.90, 0.99  2,707  50.4  1.03  0.98, 1.09  1,460  34.6  1.15  1.07, 1.23  High  2,813  82.7  1.34  1.28, 1.41  4,139  80.5  1.49  1.42, 1.57  3,018  56.6  1.70  1.61, 1.80  Abbreviations: CI, confidence interval; HR, hazard ratio; SES, socioeconomic status. a On an additive scale, for the highest tertiles versus the lowest, relative excess risk = 0.51, 95% CI: 0.40, 0.62; P < 0.001; on a multiplicative scale, for the highest tertiles versus the lowest, ratio of HRs = 1.50, 95% CI: 1.34, 1.65; P < 0.001. b HRs were adjusted for age, year of military conscription examination, weight, muscular strength, educational level, neighborhood SES, hypertension, diabetes mellitus, ischemic heart disease, and family history of AF. c AF incidence rate per 100,000 person-years. Figure 1. View largeDownload slide Probability of atrial fibrillation by height and aerobic fitness at age 18 years, adjusted for weight, muscular strength, and other covariates, Sweden, 1969–2012. The median attained age was 47 years, and the maximum was 62 years. Figure 1. View largeDownload slide Probability of atrial fibrillation by height and aerobic fitness at age 18 years, adjusted for weight, muscular strength, and other covariates, Sweden, 1969–2012. The median attained age was 47 years, and the maximum was 62 years. A similar overall pattern was seen for the interaction between weight and aerobic fitness level (Table 4). High levels of aerobic fitness were protective among men in the lowest tertile for weight but were associated with an increased risk of AF among those in the medium or highest tertiles (Table 4 and Web Table 2). The combination of high weight and high aerobic fitness level was associated with the highest AF risk (for the highest tertiles vs. the lowest for both variables, adjusted HR = 1.34, 95% CI: 1.27, 1.42; P < 0.001). These factors had significant positive interactions on both the additive and multiplicative scales (P < 0.001). Figure 2 shows the probability of AF among men in the 25th, 50th, and 75th percentiles of aerobic fitness across the full distribution of weight after adjustment for height and other covariates. In secondary analyses, we also found positive additive and multiplicative interactions of muscular strength with either height (Web Table 3) or weight (Web Table 4) in relation to the risk for AF, although these interactions were smaller in magnitude than the interactions between aerobic fitness level and height or weight. Table 4. Interaction of Weight and Aerobic Fitness Level in Relation to Risk of Atrial Fibrillaion,a Adjusted for Height and Other Factors,b Sweden, 1969–2012 Tertile of Weight  Tertile of Aerobic Fitness Level  Low  Medium  High  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  Low  5,213  53.0  1.00  Referent  1,309  33.9  0.86  0.80, 0.91  264  19.6  0.75  0.66, 0.86  Medium  2,643  62.0  0.90  0.86, 0.95  3,075  52.6  0.99  0.95, 1.04  1,342  31.7  1.01  0.94, 1.08  High  2,159  86.3  1.02  0.96, 1.07  4,218  80.1  1.15  1.09, 1.21  3,437  53.4  1.34  1.27, 1.42  Tertile of Weight  Tertile of Aerobic Fitness Level  Low  Medium  High  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  No. of Cases  Ratec  HR  95% CI  Low  5,213  53.0  1.00  Referent  1,309  33.9  0.86  0.80, 0.91  264  19.6  0.75  0.66, 0.86  Medium  2,643  62.0  0.90  0.86, 0.95  3,075  52.6  0.99  0.95, 1.04  1,342  31.7  1.01  0.94, 1.08  High  2,159  86.3  1.02  0.96, 1.07  4,218  80.1  1.15  1.09, 1.21  3,437  53.4  1.34  1.27, 1.42  Abbreviations: CI, confidence interval; HR, hazard ratio; SES, socioeconomic status. a On an additive scale, for the highest tertiles versus the lowest, relative excess risk = 0.57, 95% CI: 0.46, 0.68; P < 0.001; on a multiplicative scale, for the highest tertiles versus the lowest, ratio of HRs = 1.75, 95% CI: 1.51, 1.98; P < 0.001. b HRs were adjusted for age, year of military conscription examination, height, muscular strength, educational level, neighborhood SES, hypertension, diabetes mellitus, ischemic heart disease, and family history of AF. c AF incidence rate per 100,000 person-years. Figure 2. View largeDownload slide Probability of atrial fibrillation by weight and aerobic fitness level at age 18 years, adjusted for height, muscular strength, and other covariates, Sweden, 1969–2012. The median attained age was 47 years, and the maximum was 62 years. Figure 2. View largeDownload slide Probability of atrial fibrillation by weight and aerobic fitness level at age 18 years, adjusted for height, muscular strength, and other covariates, Sweden, 1969–2012. The median attained age was 47 years, and the maximum was 62 years. DISCUSSION In the present large national cohort study, we found that higher height or weight at age 18 years was associated with a higher risk of the development of AF in adulthood, and both had interactions with the level of aerobic fitness. High levels of aerobic fitness were associated with a higher risk of AF among men who were at least 186 cm (6 feet, 1 inch) tall but were protective among shorter men. In a similar manner, high levels of aerobic fitness were associated with a higher risk among men in the highest tertile of weight but not the lowest. The association between both exposures together (i.e., the combination of increased height or weight and aerobic fitness level) and the risk of AF exceeded the sum or product of their associations considered separately. These findings suggest that the underlying mechanisms for AF may include important interactions between height or weight and aerobic fitness level. The main associations that we observed between tall stature or obesity and AF are overall consistent with previous findings. In prior studies, researchers have reported that increased height (7–11) or high BMI and other measures of obesity (11–18) were associated with a higher risk of AF among men and women. Although most studies have focused on middle-aged or older adults, researchers of a study involving 12,850 young Danish men (median age, 19 years) reported that overweight or obesity was associated with a higher risk of AF in early adulthood (13). Results from other studies have suggested that high lean body mass (rather than body fat) may also be associated with a higher risk of AF (17, 52). In contrast, previous findings on aerobic fitness or exercise level in relation to AF risk have been conflicting. In some (19–24) but not all (53–55) studies, investigators have reported a higher risk of AF among athletes or other individuals who reported high levels of exercise. In a meta-analysis of 6 case-control studies that included 655 athletes and 895 control participants, researchers reported that athletes had more than a 5-fold higher odds of AF (odds ratio = 5.29, 95% CI: 3.57, 7.85) (56). However, results from most studies in which investigators examined objectively measured aerobic fitness (rather than self-reported exercise) have suggested that high levels of aerobic fitness are protective against AF (12, 25–27). The largest of these was a US cohort study of 64,561 middle-aged adults with a median follow-up of 5.4 years, in which investigators reported that each 1 higher metabolic equivalent achieved during treadmill testing was associated with a 7% lower risk of the development of AF (HR = 0.93, 95% CI: 0.92, 0.94; P < 0.001). In contrast, results from a Swedish study that included a subset of the present cohort (n = 1.1 million) suggested that aerobic fitness was positively associated with AF risk, but potential interactions with height were not reported (28). These discrepancies in reported associations between aerobic fitness level or exercise and the risk for AF may potentially be related to other modifying factors, such as body size. To our knowledge, the present study is the first to examine not only the independent associations between height, weight, or level of aerobic fitness and the risk of AF, but also the potential additive and multiplicative interactions. We found that the associations between aerobic fitness and the risk of AF varied substantially depending on height or weight. High levels of aerobic fitness were associated with a higher risk of AF among men who were at least 186 cm (6 feet, 1 inch) tall but were protective among men who were shorter than 186 cm. The interactions between height or weight and level of aerobic fitness were strongly positive on both the additive and multiplicative scales. These findings suggest that high levels of aerobic fitness accounted for significantly more cases of AF among tall versus short men or among those in the highest tertile of weight versus those in the lowest tertile. These results may provide additional insights into the underlying mechanisms of AF. AF is a heterogeneous disease process that is influenced by structural (e.g., left atrial size), hemodynamic (e.g., left atrial stretch), electrical (e.g., altered conduction patterns due to atrial myocardial fibrosis), and neural (e.g., autonomic dysregulation) factors (29). Exercise may influence this process through its contribution to left ventricular hypertrophy, which leads to some degree of diastolic dysfunction and left atrial stretch, which may increase the risk of AF among susceptible subgroups (30). Height and weight are known to correlate positively with left atrial and ventricular size (8, 31). Our findings suggest that these factors may potentially influence the hemodynamic or autonomic effects of aerobic fitness on atrial remodeling and the development of AF. However, left atrial and ventricular size have been reported to only partly explain the association between tall stature and the risk of AF (8), which also suggests the possibility of other mechanisms such as common genetic factors (e.g., pituitary homeobox 2 (PITX2) (57) and zinc finger homeobox 3 (ZFHX3) (58)) that are associated with growth pathways and with AF. Additional clinical and experimental studies are needed to elucidate these complex pathways, which may help further identify high-risk subgroups and ultimately new targets for intervention. Strengths of the present study include its large national cohort design with follow-up from age 18 years into adulthood. The national cohort design minimized the potential for selection bias, and the use of registry data with prospectively measured exposures prevented bias that may result from self-reporting. We examined well-validated, objective measurements of aerobic fitness level and muscular strength. We were able to adjust for other common risk factors that also were prospectively ascertained and not self-reported, including a family history of AF, individual and neighborhood-level SES, and chronic diseases that are associated with AF. Limitations included measurement of the study exposures at only a single age (18 years), which meant that we were unable to examine changes in these factors across time. Additional studies that contain data on longitudinal exposures are needed to delineate more specific age windows of susceptibility in relation to the risk of AF. The study cohort also was relatively young and exclusively male. The median age at the end of follow-up was 47 years (maximum, 62 years), and hence the findings may not necessarily apply to AF in older adults. The risk of AF is estimated to double with each decade of life (47), and 70% of prevalent cases are in adults 65 years of age or older (59). Further studies will be needed to assess our findings in older adults and in women. In summary, in the present study, we found that tall stature and high weight at age 18 years were independently associated with a higher risk of AF development in adulthood and interacted with the level of aerobic fitness. High aerobic fitness level was associated with a higher risk of AF among tall men but was protective among shorter men. These findings suggest important interactive effects of body size and aerobic fitness level on the development of AF. Tall, aerobically fit persons may be a relatively high-risk subgroup. ACKNOWLEDGMENTS Author affiliations: Alfred and Gail Engelberg Department of Family Medicine and Community Health, Icahn School of Medicine at Mount Sinai, New York, New York (Casey Crump, Jan Sundquist, Kristina Sundquist); Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, New York (Casey Crump); Center for Primary Health Care Research, Department of Clinical Sciences, Faculty of Medicine, Lund University, Malmö, Sweden (Jan Sundquist, Kristina Sundquist); and Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, California (Marilyn A. Winkleby). This work was supported by the National Heart, Lung, and Blood Institute at the National Institutes of Health (grant R01 HL116381); the Swedish Research Council; and the Avtal om Läkarutbildning och Forskning project grant from Region Skåne/Lund University, Malmö, Sweden. The funding agencies had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript. Conflicts of interest: none declared. Abbreviations AF atrial fibrillation BMI body mass index CI confidence interval HR hazard ratio ICD International Classification of Diseases SD standard deviation SES socioeconomic status REFERENCES 1 Chugh SS, Havmoeller R, Narayanan K, et al.  . Worldwide epidemiology of atrial fibrillation: a Global Burden of Disease 2010 Study. Circulation . 2014; 129( 8): 837– 847. Google Scholar CrossRef Search ADS PubMed  2 Benjamin EJ, Wolf PA, D’Agostino RB, et al.  . Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation . 1998; 98( 10): 946– 952. Google Scholar CrossRef Search ADS PubMed  3 Lip GY, Tse HF, Lane DA. Atrial fibrillation. Lancet . 2012; 379( 9816): 648– 661. Google Scholar CrossRef Search ADS PubMed  4 Kannel WB, Abbott RD, Savage DD, et al.  . Epidemiologic features of chronic atrial fibrillation: the Framingham study. N Engl J Med . 1982; 306( 17): 1018– 1022. Google Scholar CrossRef Search ADS PubMed  5 Krahn AD, Manfreda J, Tate RB, et al.  . The natural history of atrial fibrillation: incidence, risk factors, and prognosis in the Manitoba Follow-Up Study. Am J Med . 1995; 98( 5): 476– 484. Google Scholar CrossRef Search ADS PubMed  6 Wyse DG, Van Gelder IC, Ellinor PT, et al.  . Lone atrial fibrillation: does it exist? J Am Coll Cardiol . 2014; 63( 17): 1715– 1723. Google Scholar CrossRef Search ADS PubMed  7 Schmidt M, Botker HE, Pedersen L, et al.  . Adult height and risk of ischemic heart disease, atrial fibrillation, stroke, venous thromboembolism, and premature death: a population based 36-year follow-up study. Eur J Epidemiol . 2014; 29( 2): 111– 118. Google Scholar CrossRef Search ADS PubMed  8 Rosenberg MA, Patton KK, Sotoodehnia N, et al.  . The impact of height on the risk of atrial fibrillation: the Cardiovascular Health Study. Eur Heart J . 2012; 33( 21): 2709– 2717. Google Scholar CrossRef Search ADS PubMed  9 Schnabel RB, Sullivan LM, Levy D, et al.  . Development of a risk score for atrial fibrillation (Framingham Heart Study): a community-based cohort study. Lancet . 2009; 373( 9665): 739– 745. Google Scholar CrossRef Search ADS PubMed  10 Chamberlain AM, Agarwal SK, Folsom AR, et al.  . A clinical risk score for atrial fibrillation in a biracial prospective cohort (from the Atherosclerosis Risk in Communities [ARIC] study). Am J Cardiol . 2011; 107( 1): 85– 91. Google Scholar CrossRef Search ADS PubMed  11 Frost L, Hune LJ, Vestergaard P. Overweight and obesity as risk factors for atrial fibrillation or flutter: the Danish Diet, Cancer, and Health Study. Am J Med . 2005; 118( 5): 489– 495. Google Scholar CrossRef Search ADS PubMed  12 Grundvold I, Skretteberg PT, Liestol K, et al.  . Importance of physical fitness on predictive effect of body mass index and weight gain on incident atrial fibrillation in healthy middle-age men. Am J Cardiol . 2012; 110( 3): 425– 432. Google Scholar CrossRef Search ADS PubMed  13 Schmidt M, Botker HE, Pedersen L, et al.  . Comparison of the frequency of atrial fibrillation in young obese versus young nonobese men undergoing examination for fitness for military service. Am J Cardiol . 2014; 113( 5): 822– 826. Google Scholar CrossRef Search ADS PubMed  14 Berkovitch A, Kivity S, Klempfner R, et al.  . Body mass index and the risk of new-onset atrial fibrillation in middle-aged adults. Am Heart J . 2016; 173: 41– 48. Google Scholar CrossRef Search ADS PubMed  15 Aronis KN, Wang N, Phillips CL, et al.  . Associations of obesity and body fat distribution with incident atrial fibrillation in the biracial health aging and body composition cohort of older adults. Am Heart J . 2015; 170( 3): 498.e2– 505.e2. Google Scholar CrossRef Search ADS   16 Dublin S, French B, Glazer NL, et al.  . Risk of new-onset atrial fibrillation in relation to body mass index. Arch Intern Med . 2006; 166( 21): 2322– 2328. Google Scholar CrossRef Search ADS PubMed  17 Frost L, Benjamin EJ, Fenger-Gron M, et al.  . Body fat, body fat distribution, lean body mass and atrial fibrillation and flutter. A Danish cohort study. Obesity (Silver Spring) . 2014; 22( 6): 1546– 1552. Google Scholar CrossRef Search ADS PubMed  18 Karas MG, Yee LM, Biggs ML, et al.  . Measures of body size and composition and risk of incident atrial fibrillation in older people: the cardiovascular health study. Am J Epidemiol . 2016; 183( 11): 998– 1007. Google Scholar CrossRef Search ADS PubMed  19 Baldesberger S, Bauersfeld U, Candinas R, et al.  . Sinus node disease and arrhythmias in the long-term follow-up of former professional cyclists. Eur Heart J . 2008; 29( 1): 71– 78. Google Scholar CrossRef Search ADS PubMed  20 Karjalainen J, Kujala UM, Kaprio J, et al.  . Lone atrial fibrillation in vigorously exercising middle aged men: case-control study. BMJ . 1998; 316( 7147): 1784– 1785. Google Scholar CrossRef Search ADS PubMed  21 Mont L, Tamborero D, Elosua R, et al.  . Physical activity, height, and left atrial size are independent risk factors for lone atrial fibrillation in middle-aged healthy individuals. Europace . 2008; 10( 1): 15– 20. Google Scholar CrossRef Search ADS PubMed  22 Elosua R, Arquer A, Mont L, et al.  . Sport practice and the risk of lone atrial fibrillation: a case-control study. Int J Cardiol . 2006; 108( 3): 332– 337. Google Scholar CrossRef Search ADS PubMed  23 Molina L, Mont L, Marrugat J, et al.  . Long-term endurance sport practice increases the incidence of lone atrial fibrillation in men: a follow-up study. Europace . 2008; 10( 5): 618– 623. Google Scholar CrossRef Search ADS PubMed  24 Aizer A, Gaziano JM, Cook NR, et al.  . Relation of vigorous exercise to risk of atrial fibrillation. Am J Cardiol . 2009; 103( 11): 1572– 1577. Google Scholar CrossRef Search ADS PubMed  25 Faselis C, Kokkinos P, Tsimploulis A, et al.  . Exercise capacity and atrial fibrillation risk in veterans: a cohort study. Mayo Clin Proc . 2016; 91( 5): 558– 566. Google Scholar CrossRef Search ADS PubMed  26 Qureshi WT, Alirhayim Z, Blaha MJ, et al.  . Cardiorespiratory fitness and risk of incident atrial fibrillation: results from the Henry Ford exercise testing (FIT) project. Circulation . 2015; 131( 21): 1827– 1834. Google Scholar CrossRef Search ADS PubMed  27 Pathak RK, Elliott A, Middeldorp ME, et al.  . Impact of CARDIOrespiratory FITness on arrhythmia recurrence in obese individuals with atrial fibrillation: the CARDIO-FIT study. J Am Coll Cardiol . 2015; 66( 9): 985– 996. Google Scholar CrossRef Search ADS PubMed  28 Andersen K, Rasmussen F, Held C, et al.  . Exercise capacity and muscle strength and risk of vascular disease and arrhythmia in 1.1 million young Swedish men: cohort study. BMJ . 2015; 351: h4543. Google Scholar CrossRef Search ADS PubMed  29 Kapa S, Asirvatham SJ. A MET a day keeps arrhythmia at bay: the association between exercise or cardiorespiratory fitness and atrial fibrillation. Mayo Clin Proc . 2016; 91( 5): 545– 550. Google Scholar CrossRef Search ADS PubMed  30 McClean G, George K, Lord R, et al.  . Chronic adaptation of atrial structure and function in elite male athletes. Eur Heart J Cardiovasc Imaging . 2015; 16( 4): 417– 422. Google Scholar CrossRef Search ADS PubMed  31 Badano LP, Miglioranza MH, Mihaila S, et al.  . Left atrial volumes and function by three-dimensional echocardiography: reference values, accuracy, reproducibility, and comparison with two-dimensional echocardiographic measurements. Circ Cardiovasc Imaging . 2016; 9( 7): e004229. Google Scholar CrossRef Search ADS PubMed  32 Iwasaki YK, Nishida K, Kato T, et al.  . Atrial fibrillation pathophysiology: implications for management. Circulation . 2011; 124( 20): 2264– 2274. Google Scholar CrossRef Search ADS PubMed  33 Crump C, Sundquist J, Winkleby MA, et al.  . Interactive effects of physical fitness and body mass index on the risk of hypertension. JAMA Intern Med . 2016; 176( 2): 210– 216. Google Scholar CrossRef Search ADS PubMed  34 Crump C, Sundquist J, Winkleby MA, et al.  . Physical fitness among Swedish military conscripts and long-term risk for type 2 diabetes mellitus: a cohort study. Ann Intern Med . 2016; 164( 9): 577– 584. Google Scholar CrossRef Search ADS PubMed  35 Crump C, Sundquist J, Winkleby MA, et al.  . Interactive effects of obesity and physical fitness on risk of ischemic heart disease. Int J Obes (Lond) . 2017; 41( 2): 255– 261. Google Scholar CrossRef Search ADS PubMed  36 Crump C, Sundquist J, Winkleby MA, et al.  . Interactive effects of physical fitness and body mass index on risk of stroke: a national cohort study. Int J Stroke . 2016; 11( 6): 683– 694. Google Scholar CrossRef Search ADS PubMed  37 Crump C, Sundquist J, Winkleby MA, et al.  . Interactive effects of aerobic fitness, strength, and obesity on mortality in men. Am J Prev Med . 2017; 52( 3): 353– 361. Google Scholar CrossRef Search ADS PubMed  38 Ogden CL, Flegal KM. Changes in terminology for childhood overweight and obesity. Natl Health Stat Report . 2010;( 25): 1– 5. 39 Nordesjö LO, Schele R. Validity of an ergometer cycle test and measures of isometric muscle strength when predicting some aspects of military performance. Swedish J Defence Med . 1974; 10: 11– 23. 40 Patton JF, Vogel JA, Mello RP. Evaluation of a maximal predictive cycle ergometer test of aerobic power. Eur J Appl Physiol Occup Physiol . 1982; 49( 1): 131– 140. Google Scholar CrossRef Search ADS PubMed  41 Andersen LB. A maximal cycle exercise protocol to predict maximal oxygen uptake. Scand J Med Sci Sports . 1995; 5( 3): 143– 146. Google Scholar CrossRef Search ADS PubMed  42 Hook O, Tornvall G. Apparatus and method for determination of isometric muscle strength in man. Scand J Rehabil Med . 1969; 1( 3): 139– 142. Google Scholar PubMed  43 Smith JG, Platonov PG, Hedblad B, et al.  . Atrial fibrillation in the Malmö Diet and Cancer study: a study of occurrence, risk factors and diagnostic validity. Eur J Epidemiol . 2010; 25( 2): 95– 102. Google Scholar CrossRef Search ADS PubMed  44 Hoehner CM, Allen P, Barlow CE, et al.  . Understanding the independent and joint associations of the home and workplace built environments on cardiorespiratory fitness and body mass index. Am J Epidemiol . 2013; 178( 7): 1094– 1105. Google Scholar CrossRef Search ADS PubMed  45 Cubbin C, Sundquist K, Ahlen H, et al.  . Neighborhood deprivation and cardiovascular disease risk factors: protective and harmful effects. Scand J Public Health . 2006; 34( 3): 228– 237. Google Scholar PubMed  46 Crump C, Sundquist K, Sundquist J, et al.  . Neighborhood deprivation and psychiatric medication prescription: a Swedish national multilevel study. Ann Epidemiol . 2011; 21( 4): 231– 237. Google Scholar CrossRef Search ADS PubMed  47 Benjamin EJ, Levy D, Vaziri SM, et al.  . Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA . 1994; 271( 11): 840– 844. Google Scholar CrossRef Search ADS PubMed  48 Rubin DB. Multiple Imputation for Nonresponse in Surveys . New York, NY: Wiley; 1987. Google Scholar CrossRef Search ADS   49 Vanderweele TJ, Knol MJ. A tutorial on interaction. Epidemiol Methods . 2014; 3( 1): 33– 72. 50 Li R, Chambless L. Test for additive interaction in proportional hazards models. Ann Epidemiol . 2007; 17( 3): 227– 236. Google Scholar CrossRef Search ADS PubMed  51 Richardson DB, Kaufman JS. Estimation of the relative excess risk due to interaction and associated confidence bounds. Am J Epidemiol . 2009; 169( 6): 756– 760. Google Scholar CrossRef Search ADS PubMed  52 Azarbal F, Stefanick ML, Assimes TL, et al.  . Lean body mass and risk of incident atrial fibrillation in post-menopausal women. Eur Heart J . 2016; 37( 20): 1606– 1613. Google Scholar CrossRef Search ADS PubMed  53 Frost L, Frost P, Vestergaard P. Work related physical activity and risk of a hospital discharge diagnosis of atrial fibrillation or flutter: the Danish Diet, Cancer, and Health Study. Occup Environ Med . 2005; 62( 1): 49– 53. Google Scholar CrossRef Search ADS PubMed  54 Everett BM, Conen D, Buring JE, et al.  . Physical activity and the risk of incident atrial fibrillation in women. Circ Cardiovasc Qual Outcomes . 2011; 4( 3): 321– 327. Google Scholar CrossRef Search ADS PubMed  55 Mozaffarian D, Furberg CD, Psaty BM, et al.  . Physical activity and incidence of atrial fibrillation in older adults: the cardiovascular health study. Circulation . 2008; 118( 8): 800– 807. Google Scholar CrossRef Search ADS PubMed  56 Abdulla J, Nielsen JR. Is the risk of atrial fibrillation higher in athletes than in the general population? A systematic review and meta-analysis. Europace . 2009; 11( 9): 1156– 1159. Google Scholar CrossRef Search ADS PubMed  57 Gudbjartsson DF, Arnar DO, Helgadottir A, et al.  . Variants conferring risk of atrial fibrillation on chromosome 4q25. Nature . 2007; 448( 7151): 353– 357. Google Scholar CrossRef Search ADS PubMed  58 Benjamin EJ, Rice KM, Arking DE, et al.  . Variants in ZFHX3 are associated with atrial fibrillation in individuals of European ancestry. Nat Genet . 2009; 41( 8): 879– 881. Google Scholar CrossRef Search ADS PubMed  59 Go AS, Hylek EM, Phillips KA, et al.  . Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA . 2001; 285( 18): 2370– 2375. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Journal

American Journal of EpidemiologyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off