https://orcid.org/0000-0002-7690-663X
This editorial refers to ‘Physical activity, cardiorespiratory fitness, and cardiovascular outcomes in individuals with atrial fibrillation: the HUNT Study’†, by L.E. Garvnik et al., on page .
Atrial fibrillation (AF) is the most common sustained arrhythmia (lifetime risk of 37% at age 55 years), with increasing prevalence globally and having significant adverse impacts on morbidity and mortality.1 , 2 Substantial evidence supports benefits of physical activity (PA) and cardiorespiratory fitness (CRF) among patients who have had a heart attack for preventing future cardiovascular disease (CVD) events.3 , 4 However, while PA is associated with lower incidence of AF, whether PA and CRF are associated with better health outcomes among patients with prevalent AF has been unclear.
In this issue of the European Heart Journal, Garvnik and colleagues5 sought to assess the associations of self-reported PA and estimated CRF with subsequent CVD and death among 1117 patients with prevalent AF at baseline within the Norwegian HUNT3 cohort. PA was estimated using three questions on the frequency, intensity, and duration of habitual exercise. CRF (VO2 peak) was estimated using sex-specific prediction models based on age, waist circumference, resting heart rate, and estimated PA. Events were assessed over ∼8 years of follow-up by linkage to national disease registries. Adjusting for age, sex, smoking, alcohol use, occupational status, and prevalent CVD, AF patients who met PA guidelines (either 75 min of vigorous activity or 150 min of moderate-to-vigorous activity per week6) experienced ∼45% lower risk of all-cause mortality and CVD mortality, compared with inactive patients. In addition, each one MET (metabolic equivalent) increase in estimated CRF was associated with 12% lower all-cause mortality and 15% lower CVD mortality. Both inactive AF patients and inactive subjects without AF (in the full HUNT3 cohort) had ∼50% higher risk of all-cause mortality and a more than two-fold higher risk of CVD mortality compared with their counterparts who met PA guidelines. AF patients meeting PA targets had similar mortality risk to inactive subjects without AF, which is important to note considering that AF patients have a higher absolute risk of mortality than the general population.
Given the lack of studies addressing these relationships in AF, these findings are novel and informative. However, several methodological issues are relevant to consider. First, PA was assessed by a simple survey that did not consider activities beyond leisure-time exercise and also broadly categorized participants into ‘at/above the PA guidelines’, ‘below guidelines’, or ‘inactive’. The instrument had also been previously shown to be best suited for detecting vigorous activity in a healthy population.7 While such a short survey has practical utility in a large cohort, it provides less precision for measuring more moderate activities that may be more clinically relevant to subjects with a diagnosed chronic condition such as AF. Similarly, the equation for estimating CRF was derived from another healthy sample from the HUNT3 cohort and included the aforementioned PA measure.8 Thus, the likelihood of misclassification of both PA and CRF is increased in this study.
The challenges in accurately assessing PA and CRF in these patients would generally lead to underestimation of associations with outcomes such as mortality and CVD. For example, 33% of the participants were calculated to be meeting PA recommendations, which is substantially greater than what is typically observed in a healthy older population when measured objectively.9 , 10 Some research suggests higher PA in certain regions of Norway, but generally fitness levels appear similar to those of the USA.11 Interestingly, lower risk was also observed among patients in the middle category of PA—those not meeting targets but not completely inactive—compared with inactive patients. This may have clinical and public health relevance, suggesting that even modest increases in PA can be beneficial among AF patients who are sedentary. More precise measures of light and moderate activity may be of high utility to better understand the importance of PA for long-term health benefits in AF, informing future patient-specific recommendations.
Beyond the issues of measurement error, from a temporal perspective, one cannot determine whether the PA levels and CRF of these AF patients preceded or followed their AF diagnosis. If these patients were simply continuing, on average, their preceding activities, then the implications are that being more active (or fit) prior to an AF diagnosis predicts better outcomes following development of AF. In this regard, the measure of CRF in particular is likely to be a predictor of longer term fitness, i.e. incorporating fitness prior to AF. This would support the need for the general population to be as active and fit as possible, but not necessarily the benefits of increasing activity or fitness following AF per se.
It is also unclear why associations were stronger for mortality outcomes vs. morbidity (e.g. total CVD or stroke). While some trends toward lower CVD morbidity were seen, these were weaker than for mortality and generally not statistically significant, even though the numbers of CVD morbidity events were similar to the mortality outcomes. In the general population, PA is similarly associated with lower risk of fatal vs. non-fatal CVD events.12–14 In sum, the lack of association with CVD morbidity is puzzling and indicates a need for further investigation of the range of potential benefits of activity and fitness in patients with AF.
Finally, the greatest potential limitation is the possibility of residual confounding: that PA and CRF are markers for other risk factors or for severity of illness rather than the cause of associated lower mortality. For example, adjustment for AF subtype (paroxysmal, persistent, or permanent) attenuated all associations, including for mortality. More patients with paroxysmal AF, compared with persistent or permanent AF, met PA targets (and presumably also had higher CRF), suggesting that type of AF could be an important confounder. However, data on AF subtype were available in less than half of patients, making this an important question for further study. Not surprisingly, greater amounts of PA and CRF also clustered with younger age, lower body mass index, less prevalent hypertension, heart failure, heart attack, stroke, and diabetes, and lower CHADS-VASc score (clinical prediction score for stroke risk). Other key lifestyle behaviours that were not accounted for that could correlate with PA and CRF include dietary habits, which are strongly linked to mortality and CVD risk, and other measures of socioeconomic status beyond occupation, such as education and family income. On the other hand, the associations of PA and CRF with mortality were generally consistent in a range of sensitivity and subgroup analyses, PA has a range of physiological benefits which provide biological plausibility (Figure 1), and the magnitude of the associations suggests that residual confounding may partly but not fully explain the observed lower risk.
Mitigating mortality risk with physical activity and fitness.
The HUNT3 study population included mostly white males of Norwegian descent with similar sociodemographics, and thus findings may be limited in generalizability to more diverse populations, although fitness levels, and PA perhaps to a lesser degree, in HUNT3 are reflective of levels across Europe and the USA.11 However, this study also had key strengths, including the prospective cohort design, the large numbers of patients with AF, the substantial follow-up period, and good measures of mortality and CVD outcomes. In sum, this investigation provides new evidence on the potential benefits of PA and fitness among patients with prevalent AF. However, these associations cannot confirm causality in order to directly inform recommendations for exercise prescriptions in AF patients.
Currently, there are no established PA guidelines for patients with AF. A better understanding of the relationship between PA and health outcomes in AF has great clinical and public health relevance, based on the increasing prevalence of AF on a global scale and also the widespread prevalence of inactivity and insufficient PA and fitness.11 , 15 These new findings by Garvnik and colleagues5 support the need for appropriately powered randomized controlled trials to test the long-term effects of different types and intensities of PA in patients with AF. In the meantime, practical steps should be taken to continue to encourage PA and actions to achieve fitness in the general population, especially among inactive individuals. Given absence of evidence for any harm in the present study, it also seems reasonable to extend general PA guidelines to patients with prevalent AF, while awaiting further data.
Conflict of interest: none declared.
The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.
Footnotes
doi:10.1093/eurheartj/ehaa032.
References
Department of Health and Human Services.
