Atrial fibrillation (AF) is the most common cardiac arrhythmia and is associated with significant morbidity, mortality and healthcare costs. In addition, AF patients have an impaired quality of life (QoL) and AF imposes a significant psychosocial burden including depression and anxiety on the individual (1). The efficacy of current treatment strategies, including antiarrhythmic drugs (AAD) and catheter ablation in AF rhythm control is quite variable and suboptimal (2). There is limited data on safety and efficacy of complementary or alternative form of therapy on AF reduction. Any such noninvasive interventions, if shown to reduce or control AF burden, will have a major public health impact.

Yoga is a combination of structured physical exercises, breathing techniques, and meditation, and is shown to positively influence cardiac autonomic function (3). It has been shown to reduce the symptoms of depression and anxiety and result in QoL improvement (4). There seems to be a complex relation among these factors, AF initiation, and maintenance. The impact of yoga in patients with AF has not been investigated. We sought to examine the effects of a structured 3-month yoga program on AF burden, QoL indicators, anxiety, and depression.

We performed a single center, prospective, self-controlled, pre-post cohort study. After screening 103 consecutive eligible paroxysmal atrial fibrillation (PAF) patients, 52 were enrolled and 49 completed the study. Patients who required changes in AAD regimen were excluded from the study. The study consisted of control (first 3 months) and a yoga intervention phase (next 3 months). Each patient acted as his or her own control. Three patients withdrew from the study during yoga training period. All patients were on stable medical therapy during the control and intervention phases. Clinical characteristics and quality of life/anxiety/depression scores were assessed at baseline (Day 0), end of the control phase (Day 90) and end of the intervention phase (Day 180). The details of patient screening and yoga intervention are outlined in the Online Appendix.

Patients with paroxysmal AF between 18 and 80 years of age and willing to participate in the study were enrolled. Patients with a history of AF ablation within 3 months, contraindications for yoga training, life expectancy <1 year, advanced heart failure, and patients who practiced any form of yoga in the preceding 6 months were excluded. The protocol was approved by the Human Subjects Committee at the University of Kansas Medical Center.

During the intervention period, all patients underwent the structured Iyengar yoga (details in the Online Appendix) training at least twice weekly. All training sessions were conducted in groups of 15 to 20 people in a yoga studio by a certified professional yoga instructor and lasted for 60 min. During each yoga session, 10 min of pranayamas, 10 min of warm-up exercises, 30 min of asanas, and 10 min of relaxation exercises were performed. An educational DVD was also provided to each participant, and depending on the comfort level, patients were encouraged to practice these postures on their own at home on a daily basis. Compliance was reinforced with biweekly phone calls.

The primary outcomes included change in symptomatic AF, symptomatic non-AF, and asymptomatic AF episodes. Secondary outcomes included change in Short Form 36 (SF-36) QoL score, Zung self-assessment anxiety score (SAS) and Zung self-assessment depression score (SDS).

These included any adverse effects related to the yoga intervention including injury or trauma during the interventional period.

AF (>30 s) during the study period was monitored using self-reporting (symptom diary) and cardiac nonlooping event monitors (Cardio Labs, Inc., Franklin, Tennessee) and logged as an episode. Patients were asked to log episodes that were consistent with symptoms of AF and at least 1 recording per day if they did not have symptoms. All recorded symptomatic episodes were correlated with the rhythm findings on the event monitor to distinguish between the symptomatic and asymptomatic AF episodes. Events associated with symptoms and AF on the event monitor was labeled as symptomatic AF episodes. Events associated with symptoms and no AF on the monitor was labeled as symptomatic non-AF episodes. Events associated with no symptoms but documented AF on the monitor were labeled as asymptomatic AF episodes.

SDS and SAS were used to assess depression and anxiety pre- and post-yoga intervention (8). SF-36 scoring system was used to assess the QoL (9).

The Kolmogorov-Smirnov test was used to test for data normality, and accordingly non-parametric methods were used where indicated. The Wilcoxon signed rank test was used to compare the primary outcome measures between the control and intervention phases; and the Friedman test was used to compare baseline, and pre- and post-yoga measurements of anxiety, depression, and QoL scores. Noninvasive hemodynamic parameters met normality assumptions so were compared using a repeated measures analysis of variance with post hoc testing utilizing the Bonferroni adjustment. The effect of yoga compliance was assessed using the Mann-Whitney U test and the Kruskal-Wallis H test. The nonparametric equivalent of the Pearson correlation coefficient, the Spearman's rho statistic, was used to evaluate the relationship of anxiety and depression with episodes of AF.

The mean age of study subjects (47% men) was 61 ± 11 years and mean body mass index was 28 ± 5.9 kg/m² (Table 1). The mean duration of AF since diagnosis was approximately 5 years with a mean left atrial size of 4.01 ± 0.5 cm and left ventricular ejection fraction of 59 ± 6%.

Yoga significantly reduced the number of symptomatic AF episodes (3.8 ± 3 vs. 2.1 ± 2.6; p < 0.001), symptomatic non-AF episodes (2.9 ± 3.4 vs. 1.4 ± 2.0; p < 0.001) and asymptomatic AF episodes (0.12 ± 0.44 vs. 0.04 ± 0.20; p < 0.001) from the end of control phase to the end of intervention phase (Figure 60_gr1). Eleven (22%) patients with documented AF during the control pre-yoga phase did not have any AF episode during the yoga phase.

Except for the General Health domain on SF-36, the SDS, SAS, or the SF-36 scores did not change from baseline to the end of the control (pre-yoga) phase (Table 2). However, at the end of yoga intervention phase, the SDS and SAS scores improved significantly (p < 0.001 for both). Similarly, SF-36 scores improved after yoga therapy on the following domains: physical functioning (p = 0.017), general health (p < 0.001), vitality (p < 0.001), social functioning (p = 0.019), and mental health (p < 0.01). At the end of the control phase, the heart rate and systolic and diastolic blood pressures decreased from baseline (p < 0.001). Greater reductions in all 3 parameters were noted from the end of control phase to the end of intervention phase (p < 0.001).

The changes in anxiety and depression scores at 3 time points (at baseline and at the end of control and yoga phases) did not significantly correlate with changes in symptomatic non-AF, asymptomatic AF, or symptomatic AF episodes (Table 3).

Changes in SAS scores correlated with changes in heart rate (p = 0.024), indicating a decrease in heart rate after yoga may be related to improvement in anxiety (Table 4). Absolute changes in SBP following yoga not only correlated with improvement in symptomatic AF episodes (p = 0.03), asymptomatic AF episodes (p = 0.015), and symptomatic non-AF episodes (p = 0.04), but also with SAS scores (p = 0.04). These data indicate that the reduction in SBP with the practice of yoga influences arrhythmia recurrence as well as QoL.

All patients had at least 2 sessions/week of intervention with a mean of 3 sessions/week and a range of 2 to 7 sessions/week. There was no correlation between the number of practice sessions and the primary or secondary outcomes (Table 5). All of the participants tolerated the practice of yoga well. No major adverse effects or complications were noted or reported due to yoga therapy during the entire study period.

This is the first study to evaluate the role of yoga, a noninvasive complementary and alternative medicine intervention, in the management of AF. Our results show that yoga therapy significantly reduces symptomatic and asymptomatic AF burden, and improves anxiety, depression, resting heart rate and blood pressure and the QoL in patients with PAF. These findings underscore the therapeutic value of a low-cost noninvasive therapy such as yoga to effectively complement the conventional treatment strategies in improving AF patient care. Given the high prevalence of AF and costs of conventional therapy, the public health relevance of these findings is very pertinent.

Although the precise mechanisms underlying the benefits in AF remain unclear, existing literature supports the multisystem benefits of yoga that may explain our findings. Extreme and nonphysiologic fluctuations in the autonomic tone, especially surges in sympathetic tone, often precede the onset of PAF episodes (5). Continued systemic and regional inflammation, oxidative stress and endothelial dysfunction can promote substrate changes that help sustain AF. Several studies suggest that practicing yoga can decrease systemic stress, down regulate the hypothalamic-pituitary-adrenal axis, which in turn can reduce sympathetic nervous system activity and increase parasympathetic activity. Yoga practice was shown to decrease blood pressure, improve endothelial function and reduce inflammation ((6),(7),8). Yoga may prevent the AF initiation and perpetuation through its pleiotropic effects such as: 1) increasing the baseline parasympathetic tone; 2) suppressing extreme fluctuations in the 2 autonomic nervous system components; and 3) decreasing the progression of the arrhythmia by preventing or minimizing atrial remodeling.

Our results indicate that the practice of yoga can result in significant improvement in QoL, anxiety, and depression scores in patients with AF. This is likely explained by yoga related attenuation of neurohormonal response to triggers of stress (9). Finally, the benefit from the emotionally supportive atmosphere at yoga training centers, and the positive impact by the caring relationships, change in diet and life style modification associated with yoga practice on physiological parameters cannot be underestimated.

Very little information currently exists on the efficacy of alternative therapy for AF. Recently, in a small cohort of patients Lombardi et al. have shown that acupuncture is effective in decreasing AF recurrences after electrical cardioversion in persistent AF (10). In a smaller follow-up study on paroxysmal AF they have shown similar effects of acupuncture on AF burden reduction (11). Use of therapeutic hypnosis was found to be associated with a statistically significant lower incidence of PAF (6% vs. 24%) and AAD use (14% vs. 28%) (12). However, the impact of yoga therapy on AF has not been evaluated heretofore, and our current results constitute the first evidence that yoga is effective as a complementary therapy for the alleviation of AF burden and consequences.

We did not study the variations in autonomic tone, systemic inflammatory markers, and endothelial function that could have provided a better foundation in explaining our findings. Additionally, our study was not designed to assess if the initiation of the AF episodes in the study population was vagally mediated or not. The overall asymptomatic AF episodes could have been grossly underestimated due to the low sampling rates inherent to the non-looping event recorders.

This is a small, proof-of-concept study and future large focused randomized controlled studies will be necessary to examine the previous postulates and potential additional mechanisms through which yoga may exert beneficial effects in patients with AF.

The practice of yoga improves symptoms and arrhythmia burden, reduces anxiety and depression, and improves QoL in patients with AF. Yoga is an effective complementary and alternative therapy in the management of AF and can be incorporated in comprehensive AF management strategies.