All potential participants will be invited to undertake 3 prerecorded online beginner flow yoga classes to be completed within 2 weeks. This screening step is to ensure a willingness and physical ability to partake in yoga.

Next, participants will attend an in-person appointment, during which the following baseline investigations are completed ( Figure 1 ):

• 1.

  Implantable loop recorder (ILR) inserted, or smartwatch provided

• 2.

  ECG

• 3.

  HR and BP measurement (the average of 3 BP measurements with the participant sitting quietly after resting for 5 minutes )

• 4.

  echocardiogram (TTE)

• 5.

  Cardiopulmonary exercise test (CPET)

• 6.

  Hba1c and cholesterol prolife

• 7.

  Quality of life (QoL) questionnaires (Atrial fibrillation Effect on Quality-of-life (AFEQT), hospital anxiety and depression scale (HADS), 36 item short form survey (SF-36))

• 8.

  Autonomic assessment: 10-minute recording of heart rate variability (HRV) and blood pressure variability (BPV)

• 9.

  Randomization

Flow chart showing trial investigations and follow up.

Participants will undergo assessment of HRV and BPV at baseline. HR and BP will be measured continuously for 10 minutes at rest. ECG will be recorded with standard chest electrodes and noninvasive BP will be recorded continuously from cuffs on the fingers (Nova, Finapres Medical Systems, The Netherlands). Respiration will be recorded via a strain gauge transducer around the lower chest. Time and frequency domain analysis of HRV and BPV will be performed.

HRV will also be assessed throughout the study period using data from the participants rhythm monitoring device. Only measurements during periods of sinus rhythm will be included in analysis. In those with loop recorders HRV is available for each 24-hour period, presented as the SDANN (standard deviation of the average normal normal-to-normal RR (NN) intervals for each 5-minute segment over a 24-hour period ). In those with smartwatches, pulse rate variability is measured via Photoplethysmography (PPG) and presented as SDNN (standard deviation of all normal RR (NN) intervals during a 24‐hour period ). HRV measured by wearable devices have been previously validated. ^,

HR Recovery, a measure of parasympathetic reactivation after exercise, will be assessed at baseline and 12 months on CPET.

A subgroup of 40 participants (20 participants from each group) will undergo microneurography to measure muscle nerve sympathetic activity (MSNA). At the time of enrolment all participants will be invited to undertake this additional testing until the required 40 participants have been recruited.

MNSA provides real time assessment of sympathetic outflow to the skeletal muscle. It involves insertion of a tungsten microelectrode into the peroneal nerve while delivering weak electrical stimulation through the microelectrode, until muscle twitches are evoked at 20 uA. The microelectrode tip is manipulated until spontaneous bursts of MSNA with clear cardiac rhythmicity are encountered. Neural activity will be amplified, filtered (bandpass 0.3-5.0 kHz) using a low noise headstage (NeuroAmpEX, ADInstruments, Sydney, Australia) and acquired (20 kHz sampling) using a computer-based data acquisition and analysis system (LabChart 7, PowerLab; ADInstruments). Spontaneous sympathetic baroreflex sensitivity will be estimated from the changes in MSNA to changes in diastolic pressure (slope of relationship between MSNA and diastolic BP). Spontaneous cardiovagal baroreflex sensitivity will be estimated by changes in HR to changes in systolic pressure.

Microneurography is time consuming, taking up to 3 hours per participant, and hence it is not feasible to perform this on all participants from a logistical perspective. Other autonomic assessments (HRV and BPV) will be collected in all.

Participants will be allocated 1:1 to yoga or control group by block randomization. Stratification by AF type (paroxysmal or persistent) and monitoring strategy (continuous monitoring from a cardiac device or monitoring from smart watch/wearable) has been included to ensure these variables are balanced between the groups. Computer generated randomization will be completed using the REDCap (Vanderbilt University, Tennessee) database hosted by Alfred Health. Due to the nature of the intervention, the participants cannot be blinded to group allocation. The study will be single blinded with investigators involved in interpretation of the primary endpoint blinded to study group allocation.

Participants will be followed up for a minimum of 12 months. Study visits will occur at 3,6,9, and 12 months.

At 3, 6, and 9 month visits the following information will be collected via phone:

• 1.

  Changes to AF management including medication, cardioversion or ablation

• 2.

  AF symptoms

• 3.

  Healthcare utilization (hospital admissions, including elective admission, emergency department attendance, general practitioner, cardiologist and other specialist appointments)

• 4.

  BP recorded from a home monitor, or at doctor’s office/pharmacy. Participants will be asked to perform 3 measurements while sitting quietly after resting for 5 minutes and the average will be used.

• 5.

  Review of exercise and yoga diaries

The final 12-month study visit will be conducted in person and includes the 3 monthly assessments as well as repeating all baseline investigations.

The preferred rhythm monitoring strategy is an ILR (Medtronic LINQ II), which will be inserted at the baseline visit. Where a pre-existing dual chamber pacemaker, defibrillator or ILR is present this will be utilized. Participants unwilling to undergo implantation of an ILR will be provided with an ECG capable smartwatch, this strategy has been chosen to allow otherwise eligible and motivated patients to still participate in the trial . A smartwatch compatible with the participants mobile device will be provided. Participants who already own a smart watch with ECG capabilities may use this. Participants will be required to always wear their smartwatch, except when charging and provide twice daily ECGs, plus additional ECGs if experiencing symptoms. Finally, participants not willing/able to have an ILR implanted, or use a smartwatch, will be provided with a Kardia Mobile (Alivecor, USA) ECG monitoring device.

ILRs will be set up on a remote monitoring system (Medtronic MyCareLink Heart Mobile) with transmissions sent to a dedicated study account. For participants with pre-existing devices, remote monitoring history and any device clinic reviews during the study period will be requested from the treating physician. Additionally, the device will be interrogated at baseline and at the final 12-month study visits.

Participants with a smartwatch will record a 30 second, single lead ECG twice daily and take additional recordings at the onset/offset of any symptoms suggestive of arrhythmia. The rhythm strips will be sent to a dedicated study email at the Alfred Hospital (Melbourne, Australia) and reviewed regularly by research clinicians. Rhythm strips adjudicated as sinus rhythm by the device will not undergo additional review. All ECGs labelled as AF or inconclusive by the device will be reviewed and adjudicated by 2 cardiologists blinded to study group allocation.

AF burden quantification will be performed using the following methods:

• 1.

  Implantable cardiac device—the AF burden data will be derived from transmitted data on remote monitoring networks and/or data from device interrogation during follow up visits.

• 2.

  Smart watch or Kardia Mobile device—the time of the recorded AF ECG will be taken as the onset of the episode, with the next recorded sinus rhythm taken as the offset. The duration of all episodes combined will then be expressed as a percentage of a predefined monitoring period (ie, 0-3 months, 3-6 months, 6-9 months, 9-12 months) to derive the burden for that period.

• 3.

  Any other sources of rhythm documentation such as ambulatory or inpatient care will be included.

The dual primary endpoints are;

• 1.

  AF recurrence: time to recurrence, as defined by any sustained atrial arrhythmia lasting >60 minutes after an initial 12-week blanking period, and

• 2.

  AF burden: commencing after an initial 12-week blanking period and calculated at 12 months follow up.

A 12- week “blanking” or introductory period has been arbitrarily defined to allow time for participants to learn basic yoga principles and gain a level of competence in performing yoga postures, breathing and meditation, before the impact on the endpoints is assessed.

• 1.

  A prespecified analysis of the endpoint events (time to recurrence and AF burden) will be performed in the subgroup with continuous monitoring from an implanted cardiac device.

• 2.

  Clinical success defined as freedom from documented symptomatic atrial arrhythmia recurrence (episodes >1 hour)

• 3.

  QoL and mental health questionnaires : AFEQT, HADS and SF-36

• 4.

  AF related and total healthcare utilization

• 5.

  Escalation in AF management defined as change in antiarrhythmic medication or dose increase, cardioversion, or catheter ablation

• 6.

  Changes in HRV and BPV

  • •

    Time and frequency domain analysis of HRV and BP from 10-minute recording from baseline and 12 months

  • •

    24-hour HRV from ILR or smartwatch collected throughout the study period (HRV from periods of AF will not be utilized).

    • ○

      ILR: Daily HRV available via cardiac compass, weekly average will then be calculated

    • ○

      Smartwatch: weekly average HRV recorded

• 7.

  Change in cardiorespiratory fitness as assessed by CEPT

• 8.

  Change in resting HR and BP

• 9.

  Average daytime and nighttime HR in sinus rhythm as recorded by ILR

• 10.

  Change in BMI

• 11.

  Change in left atrial size/volume as assessed by echocardiogram

• 12.

  Reduction in MSNA in a subgroup of 40 participants