Pharmacogenetic Study of Antiarrhythmic Drugs for Atrial Fibrillation

1.0 Background

While atrial fibrillation (AF) is the most common sustained cardiac arrhythmia requiring
therapy, it is also associated with increased risk of stroke, heart failure, myocardial
infarction, dementia, and death. The number of Americans affected with AF is expected to
surge to nearly 16 million by the year 2050. The AF epidemic may in part be related to the
aging of the population and increasing prevalence of recently identified risk factors
including obesity, metabolic syndrome, obstructive sleep apnea, and inflammation.
Furthermore, there is increasing support for the idea that both common and rare genetic
variants also increase susceptibility to AF which can clinically manifest in the presence of
acquired risk factors. While clinical risk factors for AF are established, the genetic
components of this "multiple-hit" genetic model for the development of AF have only recently
been identified.

Despite recent advances in catheter-based and surgical therapies, anti-arrhythmic drugs
(AADs) remain the mainstay of treatment for patients with symptomatic AF. However, response
in an individual is highly variable with more than half of patients treated with AADs
suffering a recurrence of AF within 6 to 12 months. The limited success of therapy for AF is
related to poor understanding of the underlying pathophysiology, heterogeneity of the
electrical and structural substrate, and the lack of targeted mechanism-based therapies.
Thus, one major knowledge gap is predicting response to AADs in an individual patient.
Contemporary membrane-active drugs used to suppress AF are incompletely and unpredictably
effective and are associated with significant risks of proarrhythmia and non-cardiac
toxicities. Furthermore, the current 'one-size fits all' approach to selecting AAD therapy
for a patient with symptomatic AF is based largely on minimizing the risk of adverse events
rather than on the likelihood of efficacy. Recent advances in our understanding of genetic
mechanisms of AF support the overarching hypothesis we wish to test in future studies that
variability in response to AAD therapy is modulated by common genetic variants associated
with AF. Several AF susceptibility loci have been identified and validated in genome-wide
association studies. In addition, we have shown that common AF risk single nucleotide
polymorphisms (SNPs) at the chromosome (chr) 4q25 locus not only predict poor response to
AADs but also recurrence of AF after ablation therapy and cardioversion.

While genetic approaches to AF have revealed that susceptibility to AF and response to
therapy are modulated in part by the underlying genetic substrate, the translation of these
discoveries to the bedside management of AF patients has thus far been limited. This relates
to poor understanding of the underlying mechanisms associated with common AF risk alleles,
challenges associated with determining efficacy of AADs and lack of genotype-directed
prospective studies. Our preliminary data showed that a common chr4q25 SNP associated with AF
predicted successful symptom control in patients treated with AADs and individuals who
carried the risk variant responded better to Vaughan Williams class I vs. class III AADs.
Based on the information collected in this feasibility and pilot study, we propose to conduct
a prospective pharmacogenomic study where a cohort of patients with frequent symptomatic
paroxysmal AF will be randomized to a flecainide (class I AAD) or sotalol (class III AAD) in
order to determine if response to therapy is modified by chr4q25 SNPs using AF burden as a
metric of drug efficacy. This main study, like the pilot study proposed here, will utilize a
crossover design to minimize inter-individual variability and maximize statistical power to
detect an interaction between chr4q25 genotype and the reduction of AF burden with flecainide
vs. sotalol. After a run-in period during which the AAD will be up-titrated, subjects will be
monitored with the Medtronic Reveal LINQ Insertable Cardiac Monitor (ICM) system to assess AF
burden. Furthermore, subjects will be asked to complete a comprehensive, validated 20-item AF
specific questionnaire (AF Effect on QualiTy-of-life [AFEQT]) at baseline, and monthly
thereafter for the duration of the study. At the end of the 3-month trial period, the AAD
will be discontinued and participants will be switched to the other AAD and followed for
another 3-month period in a crossover trial design.

2.0 Rationale and Specific Aims

We and others have shown common SNPs at the chr4q25 locus are associated with increased risk
of AF and modulate symptomatic response to AADs. Furthermore, our preliminary data suggests
that there is a differential response to class I vs. class III membrane-active drugs. Here,
we propose a pilot study to obtain preliminary data regarding AF burden in patients receiving
AADs and to demonstrate feasibility for a future study to test the hypothesis that chr4q25
risk SNPs modulate differential response to AADs in patients with frequent symptomatic
paroxysmal AF using reduction in mean AF burden as a metric of drug efficacy. Therefore, the
Specific Aim of this pilot/feasibility study is to obtain preliminary data regarding AF
burden in patients receiving AADs and to demonstrate feasibility for a future study to test
the hypothesis that chr4q25 SNPs modulate differential response to class I vs. class III AADs
in patients with frequent symptomatic AF.

3.0 Previous Human Studies

Chr4q25 SNPs modulate differential response to AADs in patients with AF In a preliminary
study we addressed whether symptomatic response to AAD therapy is modulated by the 3 common
AF susceptibility loci on chr4q25 (near PITX2), 16q22 (in ZFHX3), and 1q21 (in KCNN3). We
studied 478 (discovery cohort) and 198 (validation cohort) age and gender matched Caucasian
patients in the Vanderbilt AF Registry. Response to AAD therapy was defined as successful
rhythm control if the patient remained on the same AAD therapy for a minimum of 6 months with
≥ 75% reduction in AF symptoms. Multiple clinical variables (including age, hypertension,
lone AF) failed to predict response to AADs. However, a SNP at the 4q25 locus (rs10033464)
was significantly associated with successful symptom control (odds ratio [OR] 2.97, 95%
confidence interval [CI] 1.42-6.21, P=0.003). Furthermore, individuals who carried the 4q25
SNP responded better to class I vs. class III AADs in both the discovery and validation
cohorts. These preliminary findings provide the rationale for our future studies by
suggesting that common AF susceptibility variants differentially modify the response to class
I vs. class III AADs in patients with frequent symptomatic AF.

4.0 Inclusion/Exclusion Criteria

Inclusion criteria

- Caucasians ≥ 18 years of age

- History of typical or lone symptomatic AF

- AF symptoms present at least once per month

- ECG that was recorded within 12 months of randomization showing AF

- Starting a Class I or Class III AAD

- In sinus rhythm at enrollment

- Able to give informed consent

Exclusion criteria

- Permanent AF or isolated atrial flutter.

- Cardiac or thoracic surgery within the previous 6 months

- Previous use of amiodarone other than short-term use (e.g. for an acute arrhythmia in
hospital)

- Medical condition that is likely to be fatal in less than one year

- Received 2 or more AADs in past

- Creatinine clearance <40 ml/min

- Contra-indication to a class I AAD e.g., structural heart disease, or history of MI

- Contra-indication to a class III AAD, e.g., congenital or acquired long QT syndrome with
QTc>450 ms at baseline

- A reversible cause of AF (e.g., thyrotoxicosis)

- Previously treated with class I and class III antiarrhythmic drugs

- Females who are pregnant or nursing.

- History of severe AV node dysfunction unless an electronic pacemaker is present

- Any race other than Caucasian

- First- or second-degree relative has already participated in the study.

5.0 Enrollment/Randomization

Patients ≥18 years of age with symptomatic paroxysmal AF will be enrolled from the Arrhythmia
and Cardiology Clinics, the Adult Emergency Department, and inpatient Cardiology services at
Vanderbilt University Medical Center. For patients with whom researchers do not have a
relationship, introductions from a member of their care team will be sought prior to
discussion of the study. After obtaining the permission of the patient, a member of the study
team will explain the study, answer any questions, and allow sufficient time to be certain
that the subject understands the study and has given their consent to participate. Consent
will be documented by signing the informed consent document, a copy of which will be given to
the subject. For this pilot and feasibility study, we plan to recruit a total of 10 patients
with a target enrollment of 1-2 patients monthly.

6.0 Study Procedures

Randomization: Patients will be treated with either flecainide or sotalol for 3 months and
then cross over to the alternate drug for 3 months. The order of assignment will be randomly
assigned using a permuted block design with alternate block sizes of 2 and 4 subjects. The
randomization list will be generated by the PI and provided to the Investigational Drug
Service (IDS) pharmacy. The study will be open-label.

Baseline visit procedures: The baseline visit will take place at the Vanderbilt Clinical
Research Center (CRC). Patients will present to the CRC and be met by a study physician, who
will perform a history and physical exam. The CRC nurse will draw 30 ml of blood for DNA
extraction and subsequent genotyping. Plasma will be stored at -70 degrees C for future
studies. Blood will be sent to the Vanderbilt inpatient laboratory for analysis of the basic
metabolic panel. A baseline 12-lead ECG will be obtained for all patients and interpreted at
the bedside by the study physician. For patients randomized to start sotalol first, a
peripheral intravenous catheter will be placed. Once inclusion and exclusion criteria are
confirmed, the patient will receive his/her first dose of study medication (sotalol 120 mg by
mouth or flecainide 50 mg by mouth), which will be dispensed by the IDS and given to the
patient. Per routine clinical practice, patients initiating flecainide will also be started
on atenolol (25 mg by mouth daily) or diltiazem (120 mg long-acting by mouth daily) if the
patient has contraindications to β-blockers to prevent the development of 1:1 conducted
atrial flutter.

Two hours after the first dose of study drug is given a 12-lead ECG will be recorded and
interpreted by the study physician. For patients initiating flecainide the PR interval and
QRS duration will be carefully scrutinized. If these remain unchanged, the patient will be
discharged from the CRC. For patients initiating sotalol, the QTc will be examined. If the
duration is >500 ms, the sotalol dose will be reduced to 80 mg by mouth twice daily. Patients
initiating sotalol will be admitted for continuous telemetric monitoring and will remain in
the CRC until after the 5th dose is administered. Additional 12-lead ECGs will be obtained
two hours after the 3rd and 5th doses of sotalol to assess QTc. If the duration remains >500
ms on the reduced (80 mg) dose of sotalol, the drug will be stopped and the patient will be
removed from the study. If the QTc remains ≤500 ms after the 5th dose of sotalol, the patient
will be discharged from the CRC.

A study physician who has undergone specific training will insert the Medtronic Reveal LINQ
ICM according to manufacturer instructions in the CRC. The device will be inserted on day 1.
Briefly, using sterile technique and local anesthesia, a very small (6 to 8 mm) skin incision
is made over the left chest. The ICM, which is approximately cylindrical and measures roughly
2 cm long and 5 mm in diameter, is inserted using a proprietary insertion tool and a sterile
dressing is applied at the end of the procedure. The device will be programmed to optimize
the memory for storing AF episodes. "AF-only mode" will be activated and programmed to "All
episodes" to ensure that the shortest possible AF episodes are detected and stored in the
device memory. The patients will be taught how to perform a CareLink transmission using the
small handheld "Patient Assistant" device. Medtronic personnel will be on hand to aid in the
insertion and initial programing of the ICM as well as patient education with the Patient
Assistant device.

For patients initiating sotalol, no further dose adjustments will be planned after discharge
from the CRC. Patients initiating flecainide will be contacted by telephone 3 to 4 days after
initiation and, if they are tolerating the medication, will be instructed to increase their
dose to 100 mg twice daily.

One-week follow-up visit: All patients will return for a brief visit in the CRC 1 week after
ICM insertion for a wound check and to obtain a 12-lead ECG. A study physician will perform a
brief focused history to assess medication tolerance and compliance. A pill count will be
performed to further assess compliance. Barring any ECG abnormalities or clinical side
effects, patients initiating flecainide will be instructed to increase the dose to the final
dose of 150 mg twice daily.

Three-month follow-up visit and crossover procedures: Patients will be scheduled to return to
the CRC 3 months after study entry. Once a visit is confirmed, each patient will be asked to
stop his or her current AAD 5 half-lives (approximately 60 hours) prior to arrival to the
CRC. Patients initially taking flecainide will also stop their AV nodal blocker (atenolol or
diltiazem) at this time. Patients will then present to the CRC and undergo initiation of the
alternative drug using the same procedures that were used at the baseline study visit, with
the exception of the ICM insertion. A full interrogation of the ICM will be performed. A pill
count will be performed to further assess compliance. Patients crossing over to flecainide
will be discharged from the CRC after a 2 hour 12-lead ECG is obtained and interpreted as
normal. Patients initiating sotalol will remain in the CRC for loading and continuous
telemetric monitoring. All patients will be scheduled for a brief follow-up visit in the CRC
one week after crossover to assess for medication compliance and tolerance and for final
titration of flecainide.

Quality of life assessment: Patients will be asked to fill out the AFEQT quality of life
questionnaire at baseline and at the end of each month of the study. The questionnaire at the
end of the 3rd month will be administered just prior to stopping the first study drug. The
validated questionnaire is comprised of 20 questions about AF symptoms with answers given on
a Likert scale. The questionnaire will be provided to the patients in paper format either in
person or by mail, as applicable. A follow-up phone call will be made to each patient to
ensure that the questionnaire is completed and returned.

End of study procedures: Patients will be scheduled for a clinic visit with their primary
cardiac electrophysiologist to coincide with the termination of the study (6 months). A study
physician will also be available to conduct a brief patient interview, pill count, and ICM
interrogation. All pertinent clinical data including AF burden on each AAD, maximal tolerated
dose, and any reported side effects will be given to the primary cardiac electrophysiologist
to aide in clinical decision making about future AAD treatment.

Early study termination, early crossover, and procedures to treat persistent AF: Based on our
prior clinical experience, a proportion of patients (20 to 30%) are expected to experience
residual symptomatic episodes of AF while taking either or both study AADs. Usually these
episodes will be less bothersome or lengthy. Patients will be encouraged to continue the
study protocol if possible. However, a smaller proportion of patients might have intolerable
symptoms necessitating an unscheduled change in therapy. The study team will work with the
primary cardiac electrophysiologist to facilitate any necessary change in therapy. An early
crossover to the alternative study AAD will be encouraged over a change to a non-study AAD
when possible. A change in therapy to a non-study AAD or AF ablation procedure will be
allowed when necessary. In addition, if patients acquire persistent AF during the course of
the study, direct current cardioversion will be allowed and will be coordinated by the study
team and the primary cardiac electrophysiologist. AF burden data for patients deviating from
the study protocol will be analyzed using an "on treatment" analysis. In other words, the AF
burden will be calculated per day that patients are on each study AAD. Patients will also be
allowed to withdraw from the study for personal or other reasons at any time.

Insertable cardiac monitor management after study termination: At the conclusion of the
six-month study, patients will be given the option to remove or retain the ICM. We anticipate
that most patients and their primary electrophysiologists will opt to retain the device to
allow for continued cardiac monitoring, which is not only useful to gauge response to AAD
therapy but also frequently used to monitor AF recurrence for patients undergoing catheter
ablation. If the ICM is retained, its management, including removal, will be transferred to
the primary electrophysiologist. Patients wishing to have the ICM removed at the end of the
study period will be scheduled for an outpatient visit in the CRC for removal by a study
physician.

Primary end point and its determination: The primary endpoint will be mean AF burden over 3
months as measured by the Medtronic Reveal LINQ ICM system. All patients are scheduled to
remain in the study for 6 months. The FDA-approved Medtronic Reveal LINQ ICM system is able
to continuously monitor the heart through a single-lead ECG and has a specific algorithm that
detects AF by evaluating the irregularity of R-R intervals. The device is able to correctly
classify AF in 96.1% of patients and correctly exclude AF in 97.4% of subjects. The device is
able to monitor the heart for up to 3 years and is compatible with the Medtronic CareLink
Network remote monitoring system. This system allows the treating physician to download
device and diagnostic data through a secure network. Patients transmit data manually after
the device indicates that one of the alert criteria have been met. The occurrence of symptoms
with the AF episode will be recorded, but both asymptomatic and symptomatic AF will be
included in the primary endpoint measurement of AF burden.

For the primary analysis, AF burden will begin to be tabulated after an initial run in period
for each study drug. For patients starting sotalol, the run-in period will coincide with the
first 5 doses (60 hours). For patients starting flecainide, the run-in period will last until
a dose of at least 200 mg per day (in divided doses) is achieved, unless a smaller dose is
the maximally tolerated dose. Ideally, this will be achieved within 3-4 days after initiation
of flecainide. If, during the course of the study, a patient develops persistent AF, a
cardioversion will be scheduled in coordination with the patient's primary cardiologist or
electrophysiologist. The patient will then be switched to the other study drug. In case of
such an occurrence, AF burden will be tabulated using an "on treatment" analysis. Conditions
for which a patient will be withdrawn from the study and data censored will be
discontinuation of AAD therapy, initiation of amiodarone, AF ablation or AV node
ablation/permanent pacing, loss to follow-up, elective withdrawal from the study, or death.

Blood processing, DNA extraction, and genotyping: Blood samples will be drawn into
ethylenediamenetetraacetic acid (EDTA) tubes and immediately refrigerated at 4° C. Plasma
will be separated by centrifugation and stored at -80° C. DNA will be extracted from the
buffy coat using a commercially available kit (Qiagen Puregene, Valencia, California) and
stored at -20° C. Study participants will be genotyped for three common chr4q25 SNPs
(rs2200733, rs17570669, and rs3853445) using Sanger sequencing.

7.0 Risks

Risks will be minimized by performing study procedures in the Vanderbilt CRC. A complete
history and physical examination will be performed to ensure that subjects fulfill the entry
criteria as defined. All identifying documents, data, and specimens collected as a result of
this study will be retained by the investigator. Data will be entered into the secure
Research Electronic Data Capture (REDCap) database. Access to this material will be available
only to the research investigator and his staff. If results of this study are to be
published, only code numbers will be used for identification purposes. Participants will not
be identified by name.

Risks associated with the Medtronic Reveal LINQ ICM: Risks associated with insertion of the
ICM include pain and discomfort, bleeding, bruising, hematoma, infection, and an allergic
reaction to the skin prep or lidocaine used for local anesthesia. There is a very small risk
of long-term discomfort associated with having the device. According to the manufacturer, the
ICM is fully MRI-compatible immediately after implantation.

Risks associated with sotalol: Therapeutic doses of sotalol range from 160 mg to 240 mg per
day. Common side effects of sotalol therapy include bradycardia (13% to 16%), chest pain (3%
to 16%), palpitations (14%), fatigue (20%), dizziness (20%), lightheadedness (12%), weakness
(13%), dyspnea (21%), edema (8%), hypotension (6%), proarrhythmia (5%), syncope (5%), heart
failure (5%), torsade de pointes (dose related; 1% to 4%), peripheral vascular disorders
(3%), ventricular tachycardia worsened (1%), QTc interval prolongation (dose related),
headache (8%), sleep problems (8%), mental confusion (6%), anxiety (4%), depression (4%),
rash (5%), sexual side effects (3%), nausea/vomiting (10%), diarrhea (7%), stomach discomfort
(3% to 6%), flatulence (2%), impotence (2%), bleeding (2%), extremity pain (7%), paresthesia
(4%), back pain (3%), visual problems (5%), upper respiratory problems (5% to 8%), and asthma
(2%). Rare (< 1%) side effects of therapeutic doses of sotalol include alopecia,
bronchiolitis obliterans with organized pneumonia (BOOP), cold extremities, diaphoresis,
eosinophilia, leukocytoclastic vasculitis, leukopenia, paralysis, phlebitis, photosensitivity
reaction, pruritus, pulmonary edema, Raynaud's phenomenon, red crusted skin, retroperitoneal
fibrosis, elevated liver transaminases, thrombocytopenia, and vertigo.

Risks associated with flecainide: Therapeutic doses of flecainide range from 200 mg to 300 mg
per day. Common side effects of flecainide therapy include dizziness (19% to 30%), visual
disturbances (16%), dyspnea (10%), palpitations (6%), chest pain (5%), edema (3.5%),
tachycardia (1% to 3%), proarrhythmia (4% to 12%), sinus node dysfunction (1.2%), syncope,
headache (4% to 10%), fatigue (8%), nervousness (5%), fever, malaise, hypoesthesia, paresis,
ataxia, vertigo, somnolence, tinnitus, anxiety, insomnia, depression, rash (1% to 3%), nausea
(9%), constipation (1%), abdominal pain (3%), anorexia (1% to 3%), diarrhea (0.7% to 3%),
tremor (5%), weakness (5%), paresthesia (1%), diplopia (1% to 3%), and blurred vision. Rare
(< 1%) side effects include alopecia, altered pacing threshold, amnesia, angina, AV block,
bradycardia, bronchospasm, heart failure, corneal deposits, depersonalization, euphoria,
exfoliative dermatitis, granulocytopenia, heart block, increased PR interval, leukopenia,
metallic taste, neuropathy, paradoxical increase in ventricular rate in atrial
fibrillation/flutter, paresthesia, photophobia, pneumonitis, pruritus, increased QRS
duration, swollen lips/tongue/mouth, tardive dyskinesia, thrombocytopenia, urinary retention,
urticaria, and ventricular arrhythmias.

Risks associated with oral atenolol use: Therapeutic dose of atenolol ranges from 50 mg to
200 mg per day. Common side effects of chronic atenolol therapy include tiredness (26%),
dizziness (13%), depression (12%), cold, tingling, or numbness in the hands or feet (12%),
and shortness of breath (6%). Other common adverse effects of atenolol include slow heart
rate (bradycardia), a decrease in blood pressure when going from a lying-down or sitting
position to standing, a spinning sensation (vertigo), lightheadedness, diarrhea, and nausea.
We will use a low dose of atenolol (25 mg) in this study to minimize the risk of occurrence
of these adverse effects. Rare (< 1%) side effects of therapeutic doses of atenolol include
an increase in liver enzymes, allergic reactions, headache, impotence (also known as erectile
dysfunction), Peyronie's disease, worsening of psoriasis, reversible hair loss, vision
problems, dry mouth, Raynaud's phenomenon, unexplained rash, and dry eyes. These adverse
effects are not anticipated with a single low dose of atenolol that will be administered to
our volunteers.

Atenolol can rarely cause an allergic reaction (<1%) that manifests as rash, itching,
swelling, severe dizziness, or trouble breathing. Any subject with a previous history of
allergy or intolerance to beta-blockers will be excluded from the study.

Risks associated with oral diltiazem use: The therapeutic dose of diltiazem ranges from 120
mg to 480 mg per day. Common side effects of chronic diltiazem therapy include edema (2% to
15%), headache (5% to 12%), AV block (first degree 2% to 8%), bradycardia (2% to 6%),
hypotension (2% to 4%), vasodilation (2% to 3%), extrasystoles (2%), flushing (1% to 2%),
palpitation (1% to 2%), dizziness (3% to 10%), nervousness (2%), rash (1% to 4%), gout (1% to
2%), dyspepsia (1% to 6%), constipation (<2% to 4%), vomiting (2%), diarrhea (1% to 2%),
weakness (1% to 4%), myalgia (2%), rhinitis (<2% to 10%), pharyngitis (2% to 6%), dyspnea (1%
to 6%), bronchitis (1% to 4%), cough (≤3), and sinus congestion (1% to 2%). Rare (< 1%) side
effects of diltiazem include an increase in alkaline phosphatase, allergic reaction, increase
in liver transaminases, amblyopia, amnesia, arrhythmia, AV block (second or third degree),
bundle branch block, heart failure, depression, dysgeusia, extrapyramidal symptoms, gingival
hyperplasia, hemolytic anemia, petechiae, photosensitivity, Stevens-Johnson syndrome,
syncope, tachycardia, thrombocytopenia, tremor, and toxic epidermal necrolysis.

Risks associated with venipuncture: Potential problems related to establishing an intravenous
access include pain, bleeding, hematoma, fainting, and rarely infection at the venipuncture
site.

8.0 Reporting of Adverse Events or Unanticipated Problems Involving Risk to Participants or
Others

All adverse events and incidents of noncompliance with the protocol will be reported to the
institutional review board (IRB). A serious adverse event is defined as an untoward medical
occurrence that results in death, is life-threatening, requires hospitalization, results in
persistent or significant disability, or requires intervention to prevent permanent
disability or death. Other untoward medical occurrences that do not meet the above criteria
will be classified as adverse events. Study personnel who are administering the research
protocol in the CRC will monitor for adverse events, with the assistance of the CRC nursing
staff. All suspected or confirmed adverse events will promptly be reported to the principal
investigator (PI), who will collect data on these occurrences. The principal investigator
will report serious adverse events to the IRB, NIH, and any other applicable authority,
within 10 days of the PI's knowledge of the occurrence. Non-serious adverse events and
incidents of noncompliance with the protocol will be reported at the time of annual
continuing review.

9.0 Study Withdrawal/Discontinuation

Participation in the study is strictly voluntary, and patients will be able to withdraw at
any time. Following the subject's withdrawal, any samples collected will be destroyed and not
analyzed further, although any analyzed data will be maintained.

10.0 Statistical Considerations

Sample size: This proposed pilot/feasibility study is not expected to be adequately powered
to detect a variable response to AADs based on chr4q25 genotype. Rather, this study will
provide vitally important and currently unavailable data on mean and standard deviation of AF
burden during therapy with the study drugs. These data will then be used to conduct a formal
sample size calculation for the main trial. This study will also examine the feasibility of
the crossover design which, if feasible, will be utilized for the main trial.

Statistical analysis: The primary study endpoint will be AF burden (the proportion of time a
patient is in AF) and will be calculated for each patient while undergoing therapy with each
study drug. The primary endpoint will be expressed for each study drug as median, 25th
percentile, and 75th percentile. The difference in AF burden while on class I and class III
AADs will be compared using the Wilcoxon signed-rank test for paired data. We will conduct an
exploratory analysis using a linear regression model to detect an association between
genotype and AF burden after adjustment for age and sex. First, we will analyze a model
including separate terms for all three chr4q25 SNPs simultaneously with adjustment for age
and sex. Then we will create a second model that replaces the separate chr4q25 SNP terms for
a combined chr4q25 risk score calculated by adding the beta-coefficients for the genotypes at
rs2200733, rs17570669, and rs3853445 found in the first model. If multivariable analysis
demonstrates the combined chr4q25 risk score and study drug assignment are both significantly
associated with AF burden, they will be tested for a statistical interaction to determine
whether the chr4q25 AF susceptibility locus determines a differential response to class I and
class III AADs. It should be emphasized that this analysis will be underpowered and
over-fitted in this pilot/feasibility study.

11.0 Privacy/Confidentiality Issues

All research hard copy records will be stored in a locked cabinet within a locked office with
access only to the research personnel. Digitalized ECG recordings will be de-identified and
stored in a password-protected electronic database. Blood samples will be de-identified and
given a specific code number before storage and analysis. Only the investigators and members
of the study team will have access to the code key.

12.0 Follow-up and Record Retention

We expect that the study will span a period of 12 months. All data will be archived and
stored indefinitely.

Inclusion Criteria:

- All races/ethnicities ≥ 18 years of age

- History of typical or lone symptomatic AF

- AF symptoms present at least once per month

- ECG that was recorded within 12 months of randomization showing AF

- Starting a Class I or Class III AAD

- In sinus rhythm at enrollment

- Able to give informed consent

Exclusion Criteria:

- Permanent AF or isolated atrial flutter.

- Cardiac or thoracic surgery within the previous 6 months

- Previous use of amiodarone other than short-term use (e.g. for an acute arrhythmia in
hospital)

- Medical condition that is likely to be fatal in less than one year

- Received 2 or more AADs in past

- Creatinine clearance <40 ml/min

- Contra-indication to a class I AAD e.g., structural heart disease, or history of MI

- Contra-indication to a class III AAD, e.g., congenital or acquired long QT syndrome
with QTc>450 ms at baseline

- A reversible cause of AF (e.g., thyrotoxicosis)

- Previously treated with class I and class III antiarrhythmic drugs

- Females who are pregnant or nursing.

- History of severe AV node dysfunction unless an electronic pacemaker is present

- Any race other than Caucasian

- First- or second-degree relative has already participated in the study.