Table 1

Cohort studies

STUDYSTUDY CHARACTERISTICSMEAN AGE OF PARTICIPANTS, YDEFINITION OF HIGH-INTENSITY EXERCISE AND END POINTSOUTCOMESLIMITATIONSRISK OF BIAS
Karjalainen et al,31 1998• Top-level veteran orienteers (ie, runners)
• 228 in orienteer group vs 212 in control group; 100% men
• 10-y follow-up
47.5• High position in veteran ranking is an indicator of y of intense training
• Lone AF
• AF developed in 5.3% of orienteers vs 8.9% of control group (RR = 5.5; P = .012)• Men only
• Cohort information insuffcient
• No correction for confounding
• Unreliable outcome assessment
• Potential recall bias
High
Frost et al,32 2005• Danish Diet, Cancer, and Health Study data
• 19–593 (51%) men and 18 807 women
• Mean 5.7-y follow-up
56• Heavy physical workload (self-report)
• AF or atrial futter
• No differences between the sexes*
• No differences overall*
• Few subjects exposed to heavy work or vigorous PA
• Relied on registry outcomes
• Could not separate AF from atrial futter
• Potential recall bias
Moderate
Heidbuchel et al,33 2006• 137 patients after ablation for atrial futter; 83% men
• 31 of the 137 patients regularly engaged in sports before ablation
• Mean 2.5-y follow-up
58• Intense competitive activity ≥ 3 h per wk
• Development of AF after ablation for atrial futter
• AF development in high PA (HR = 1.81; P = .02)
• No differences between the sexes
• Status after ablation not comparable
• Higher rate of cardiovascular disease
• Potential for recall bias
• No correction for confounding
Moderate
Molina et al,34 2008• 252 Barcelona marathon runners vs 305 sedentary men; 100% men
• Retrospective cohort study
• Mean 11.6-y follow-up
39• Marathon running
• Lone AF
• Endurance sport practice associated with higher risk of lone AF (HR = 8.80, 95% CI 1.26 to 61.29; P = .028)*• Men only
• Control group recruited from different population
• 5-y difference in follow-up duration
• Risk of recall bias
• All AF events not considered
Moderate
Mozaffarian et al,35 2008• Cardiovascular Health Study Medicare eligibility lists in US communities
• 5446 participants; 42% men
• Post hoc analysis of RCT
• 12-y follow-up
73• ≥6 MET of intensity
• ≥ 1840 kcal/wk (self-report)
• AF on annual examination or electrocardiogram
• Graded reduction in AF with light to moderate PA intensity*
• No reduction in AF with high PA intensity
(RR = 0.87, 95% CI 0.64 to 1.19)*• With an older cohort, there was less high-intensity PA
• No sex differences examined
Low
Aizer et al,36 2009• 16–921 men in Physicians’ Health Study
• Post hoc analysis of RCT
• 12-y follow-up
51• Suffcient to “work up a sweat” 5–7 d/ wk (self-report)
• AF
• RR for 0 vs 5–7 d/wk (joggers) of vigorous exercise was 1.20 (95% CI 1.02 to 1.41; P = .04)*
• Elevated risk observed if aged < 50 y (RR = 1.74, 95% CI 1.23 to 2.47;
P < .01)
• Men only
• Some retrospective subgroup analysis
• Association seen at 3-y evaluation but not at 9 y
Low
Pelliccia et al,37 2010• 114 Olympic athletes vs 97 people in control group; 78% n
• 8.6-y follow-up
22• Participation in Olympic-endurance disciplines and having multiple games experience
• Cardiac symptoms or events
• No cardiac events developed for more than 8 y
• Reduced incidence compared with general population
• Very young study cohort
• Small group and therefore few potential outcomes
• Screening of multiple cardiovascular systems before enrolment
• Control group was very ft; had participated in 1 Olympic game
High
Everett et al,38 2011• 34 759 women who had been part of the Women’s Health Study
• 20-y follow-up
57.5• ≥ 6 MET of intensity
• ≥ 15 MET h/wk
• AF
• No difference in incident of AF among quintiles after adjusting for hypertension and obesity*• Women only
• Very few women underwent strenuous activity
• Self-assessment of PA intensity
• Self-assessment of outcomes
Low
Andersen et al,39 2013• 52 755 Swedish participants in a 90-km cross-country skiing event; 87% men
• Mean 9.7-y follow-up
38.5• Fast finishing time or high number of races completed
• AF or atrial flutter
• Higher risk of AF among those who completed > 5 races (HR = 1.29, 95% CI 1.04 to 1.61) and among those who had the fastest relative fnishing times (HR = 1.20, 95% CI 0.93 to 1.55)*• Outcomes include AF or atrial futter
• Cohort was still extremely active and not representative of general population
Low
Thelle et al,40 2013• 309 540 Norwegians in a public health screening program; 48% men
• 4-y follow-up
41.4• Frequent hard training or competitive PA
• Flecainide prescription
• Increase in AF in men with high intensity PA (HR = 3.14, 95% CI 2.17 to 4.54)*• High attrition rate over time
• Surrogate outcome (fecainide) for lone AF
• Self-assessment of PA intensity
• Risk of recall bias
Moderate
Williams and Franklin,41 2013• 46 807 participants in the National Runners’ and Walkers’ Health studies; 41% men
• Age range was 33–72 y
• 6.2-y follow-up
NA• ≥ 6 MET of intensity
• ≥ 5.4 MET h/d
• Any cardiac arrhythmia
• Arrhythmia risk declined by 4.8% each MET h/d over baseline for runners and walkers*
• Beneft was higher for those aged < 50 y
• Self-report of PA intensity
• Self-report of physician diagnosis
• No differentiation of arrhythmia type
• Risk of selection bias
• Incomplete statistical reporting
Moderate
Bapat et al,42 2014• Participants from MESA databaseNA• Highest of 3 PA intensity categories was > 2383 MET min/wk
• AF
• At highest intensity PA in 1 model, HR = 0.79, 95% CI 0.61 to 1.02; P < .05*• There are insuffcient data to properly evaluate this studyHigh
Drca et al,43 2014• 44 410 Swedish men
• Mean follow-up of 12 y
60• Any leisure-time PA of > 5 h/wk
• AF or atrial futter
• Walking or cycling at age 30 y showed beneft in AF reduction
• Risk of AF increased at age 30 y for > 5 h/wk of PA (RR = 1.19, 95% CI 1.05 to 1.36; P = .008)*
• Men only
• Risk of recall bias
• Self-assessment of PA intensity
• Outcomes included AF and atrial futter
Moderate
Ghorbani et al,44 2014• 28 169 US men
• 8-y follow-up
68• PA > 6 MET with high MET h/wk• No correlation between AF and PA*• Men only
• Self-assessment of PA intensity
• Self-report of physician diagnosis
• Participation and dropout rates were unclear
Moderate
Knuiman et al,45 2014• 4267 adults from Busselton, Western Australia; 44% men
• 15-y follow-up
52• Vigorous exercise in a usual week
• AF
• Higher level of PA associated with non-signifcant trend to reduced AF*• 57% survey response
• Self-assessment of PA intensity
• Risk of recall bias
Moderate
Myrstad et al,46 2014• 2366 Norwegian men participating in 56-km cross-country ski race vs 1179 men from general population
• 9-y follow-up
66• Endurance PA for > 30 min > 3 times per wk
• AF or atrial futter
• Increased risk of AF for 10 y of vigorous PA (HR = 1.16, 95% CI 0.06 to 1.28) in skiers*• Men only
• Controls from different population
• Self report of recent PA
• Risk of selection bias
Moderate
  • AF—atrial fibrillation, HR—hazard ratio, MESA—Multi-Ethnic Study of Atherosclerosis, MET—metabolic equivalent task, NA—not available, PA—physical activity, RCT—randomized controlled trial, RR—relative risk.

  • * After adjustment for confounding variables.

  • A physiologic measure of energy cost of a physical activity: 1 MET = 1 kcal/kg/h.