- Email: [email protected]
Efficacy of gemfibrozil in the primary prevention of atrial fibrillation in a large randomized controlled trial
Electrophysiology
Efficacy of gemfibrozil in the primary prevention of atrial fibrillation in a large randomized controlled trial A. Selcuk Adabag, MD, MS, a,b,c Salima Mithani, MD, a,c Basel Al Aloul, MD, a,c Dorothea Collins, ScD, d Stefan Bertog, MD, a,c and Hanna E. Bloomfield, MD, MPH b,c,e for The Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group Minneapolis, MN; and West Haven, CT
Background
Peroxisome proliferator–activated receptor α (PPARα) activators reduce inflammation and oxidative stress. Inflammation plays an important role in the initiation and maintenance of atrial fibrillation (AF). It has been suggested that PPARα activators may have antiarrhythmic properties, but no clinical data exist. The objective of this study was to investigate whether the PPARα activator gemfibrozil prevents or delays the development of AF in patients with coronary heart disease.
Methods
We retrospectively analyzed the electrocardiograms (ECGs) performed in the Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial, a multicenter, randomized, double-blinded, secondary prevention trial of gemfibrozil and matching placebo. The ECGs were performed annually or biannually and when clinically indicated. Participants who were in AF on baseline ECG were excluded from the present analysis. Relative risk for AF was calculated from Cox regression with death as a competing risk factor.
Results
A total of 12,605 ECGs from 2,130 participants were interpreted (5.9 ± 2.1 ECGs per participant, range 2-20). At baseline, the gemfibrozil (n = 1,070) and placebo (n = 1,060) groups were well matched. Mean age was 64.1 ± 7.1 years. Over 4.4 ± 1.5 years of follow-up, 123 (5.8%) participants developed new AF. There was no difference in AF incidence between the gemfibrozil and placebo groups (64/1,070 vs 59/1,060, respectively; P = .33). In Cox regression, the risk of AF was similar between the 2 study groups (hazard ratio 1.04, 95% CI 0.73-1.49, P = .82).
Conclusions
In this post hoc analysis of a multicenter, double-blinded, randomized controlled trial, the PPARα activator gemfibrozil did not reduce the 4-year incidence of AF among men with coronary heart disease. (Am Heart J 2009;157:913-8.)
Treatment options for the primary prevention of atrial fibrillation/flutter (AF) are limited in the nonsurgical setting. Recent evidence suggests that lipid-lowering therapy (consisting mostly of statins but also fibrates) may reduce the incidence of ventricular tachycardia and AF in patients with cardiomyopathy, independent of their effects on serum lipid profile.1-4 Statins have
From the aDivision of Cardiology, bCenter for Chronic Disease Outcomes Research, Veterans Affairs Medical Center, Minneapolis, MN, cUniversity of Minnesota, Minneapolis, MN, dCooperative Studies Program, Veterans Affairs Medical Center, West Haven, CT, and eDepartment of General Internal Medicine, Veterans Affairs Medical Center, Minneapolis, MN. Submitted October 21, 2008; accepted February 19, 2009. Reprint requests: A. Selcuk Adabag, MD, Section of Cardiology (111 C), Veterans Affairs Medical Center, One Veterans Dr, Minneapolis, MN 55417. E-mail: [email protected] 0002-8703/$ - see front matter Published by Mosby, Inc. doi:10.1016/j.ahj.2009.02.015
also been suggested to prevent AF in the setting of coronary heart disease (CHD),5 heart failure,6 and after electrical cardioversion.7,8 The mechanism by which statins reduce AF is unknown; however, their antiinflammatory and antifibrotic effects have been considered to be responsible. Peroxisome proliferator–activated receptor α (PPARα) regulates fatty acid metabolism, inflammation, and endothelial function.9 Activation of PPARα by fibrates reduces the serum levels and cellular expression of inflammatory mediators in animal models10,11 and decreases adverse cardiac events in humans.12,13 Because these effects are similar to the pleiotropic properties of statins, the possibility that PPARα activators may also prevent cardiac arrhythmias has recently been raised.1,2 However, there are no clinical data to examine whether PPARα agonists have antiarrhythmic activity. The objective of the present study was to determine whether gemfibrozil, a PPARα agonist, prevents or
914 Adabag et al
American Heart Journal May 2009
Figure 1
Selection of the 2,130 participants for retrospective ECG analysis in the present study. Flowchart describing the selection of patients.
delays the development of AF in patients with CHD who participated in the Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial (VA-HIT).12
Methods Patient selection The VA-HIT was a multicenter, double-blinded, randomized controlled trial that tested whether gemfibrozil was effective in secondary prevention of CHD in men with low levels of high-density lipoprotein. Trial design and results have previously been published elsewhere.12,14 Briefly, men b74 years old with a history of CHD, high-density lipoprotein cholesterol ≤40 mg/dL (≤3.6 mmol/L), low-density lipoprotein cholesterol ≤140 mg/dL, and triglyceride level ≤300 mg/dL were randomized to gemfibrozil 1,200 mg/d or matching placebo. During the VA-HIT, participants had electrocardiograms (ECGs) recorded annually or biannually and more frequently if clinically indicated. For the present study, we retrospectively analyzed all ECGs. The participants who were in sinus rhythm
at baseline and who had at least 1 follow-up ECG were included in the present analysis. Of the 2,531 participants in the VA-HIT, 182 were excluded because no ECGs were available (Figure 1). An additional 218 participants were excluded because baseline rhythm was AF (n = 38) or paced/ uninterpretable (n = 15), or because there were no follow-up ECGs after the baseline (n = 155) or that they were uninterpretable (n = 10). The remaining 2,130 participants (1,070 assigned to gemfibrozil and 1,060 to placebo) who were in sinus rhythm at first ECG were included in the present analysis (Figure 1). The characteristics of the participants included in the present analysis did not statistically differ from the characteristics of the entire VA-HIT cohort (data not shown). This study was approved by the human studies subcommittee of the Research and Development committee at the Minneapolis Veterans Affairs Medical Center. Individual consent was waived. This work was supported by a grant from the Minneapolis Veterans Research Institute and the Cooperative Studies Program of the Department of Veterans Affairs Office of Research and Development. The authors are solely responsible
American Heart Journal Volume 157, Number 5
Adabag et al 915
Table I. Baseline characteristics of the 2130 participants of the present study Characteristics Age (y) Race (%) White Black Other Education N12 y (%) Current smoker (%) Alcohol consumption (%) b1 Drink/d 1-3 Drinks/d N3 Drinks/d Hypertension (%) Diabetes mellitus (%) Prior myocardial infarction Heart failure (%) CABG or PCI (%) Medication use (%) Aspirin Nitrates Calcium-channel blockers ACE inhibitors β-Blockers Body mass index (kg/m2) LDL cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL) Glucose (mg/dL) WBC (cells/mL)
Placebo (n = 1060) 64.1 ± 7
Gemfibrozil (n = 1070) 64.0 ± 7.2
P value .93 .20
92 6 2 68 17
91 6 3 67 20
94 6 0.1 56 30 62 8 60
92 8 0.3 55 29 62 8 56
.67 .75 .78 .95 .14
84 46 51
82 46 52
.22 .85 .53
20 44 29.1 ± 4.8 111 ± 22 31 ± 5 161 ± 67 113 ± 33 7.1 ± 1.9
20 44 28.9 ± 4.7 111 ± 22 31 ± 5 162 ± 69 115 ± 37 7.2 ± 2.5
.83 .88 .24 .90 .89 .70 .21 .42
.76 .09 .23
Plus-minus values are means ± SD. CABG, Coronary artery bypass surgery; PCI, percutaneous coronary intervention; ACE, angiotensin-converting enzyme; LDL, lowdensity lipoprotein; HDL, high-density lipoprotein; WBC, white blood cell.
for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.
Outcome ascertainment The primary outcome variable was time to the first ECG with AF. Two experienced physicians (S.M. and B.A.A.) independently reviewed all of the ECGs. There was almost perfect agreement (κ = 0.96) between the 2 interpreters. All ECGs with a rhythm other than sinus and those where there were disagreements in interpretation were reread by a cardiologist with extensive experience in ECG analysis (S.B.).
Sample size calculation Based on our previous experience from a large observational data set of veterans with CHD, we estimated a 10% dropout rate due to AF at baseline and expected a 14% incidence of new AF over 5 years of follow-up. Thus, the present study had 80% power to detect a 28% relative reduction (absolute from 14% to 10.1%) in AF incidence with gemfibrozil.
Statistical analysis All analyses followed an intention-to-treat approach. Categorical variables were presented as frequency and continuous
Table II. Baseline characteristics of the participants who developed AF versus those who maintained sinus rhythm Characteristics
AF (n = 123)
Age (y) Race (%) White Black Other Education N12 y (%) Current smoker (%) Alcohol consumption (%) b1 Drink/d 1-3 Drinks/d N3 Drinks/d Hypertension (%) Diabetes mellitus (%) Prior myocardial infarction Heart failure (%) CABG or PCI (%) Medication use (%) Aspirin Nitrates Calcium-channel blockers ACE inhibitors β-Blockers Body mass index (kg/m2) LDL cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL) Glucose (mg/dL) WBC (cells/mL)
67.1 ± 5
Sinus rhythm (n = 2007) P value 63.9 ± 7.2
b.0001 .74
93 5 2 67 17
92 6 2 68 19
93 7 0 67 28 60 12 50
93 7 0.2 55 29 62 7 59
.01 .76 .65 .05 .05
85 59 54
83 45 51
.45 .005 .63
29 44 28.9 ± 4.6 107 ± 21 31 ± 6 170 ± 76 115 ± 38 7.2 ± 2.1
20 44 29.0 ± 4.7 111 ± 22 31 ± 5 161 ± 68 114 ± 35 7.2 ± 2.2
.01 .92 .88 .03 .97 .14 .72 .67
.90 .59 .88
Plus-minus values are means ± SD.
variables as mean ± 1 SD or median and interquartile range. Gemfibrozil and placebo groups were compared using t tests and nonparametric Wilcoxon rank tests for continuous variables and χ2 test for categorical variables. Follow-up was calculated as the number of days from the baseline ECG to the last follow-up ECG. Analyses of time-to-event data were performed with the log-rank test.15 Patients were removed from all time-to-event analyses at the time of death. Relative risks were calculated from Cox models16 with treatment as the single covariate and death as a competing risk factor. Additional Cox models including age, hypertension, diabetes mellitus, smoking, heart failure, body mass index, medication compliance, and death were performed. All analyses were performed using SAS (Cary, NC) version 9.2. The proportional hazards assumption was tested and confirmed. A P value b .05 was interpreted as significant.
Results Baseline characteristics of the study participants are displayed in Table I. Mean age was 64.1 ± 7.1 years, and all participants were male. The participants assigned to gemfibrozil and those assigned to placebo were well matched with regard to baseline characteristics (Table I).
American Heart Journal May 2009
916 Adabag et al
Figure 2
Kaplan-Meier estimates of the incidence of AF in the gemfibrozil and placebo groups.
Predictors of AF A total of 12,605 ECGs from 2,130 participants were interpreted (5.9 ± 2.1 ECGs per participant, range 2-20). Over 4.4 ± 1.5 years of follow-up, 123 (5.8%) participants developed new AF. Participants who developed new AF were older and more likely to have a history of hypertension, heart failure, coronary revascularization, and nitrate use at baseline than those who did not develop AF (Table II). Effect of gemfibrozil on AF There was no difference in AF incidence between the gemfibrozil and placebo groups (64/1,070 vs 59/1,060, respectively; P = .33). In Cox regression, after considering death as a competing risk, the risk of AF was similar between the gemfibrozil and placebo groups (hazard ratio 1.04, 95% CI 0.73-1.49, P = .82) (Figure 2). Results did not change after adjusting for age, hypertension, diabetes mellitus, heart failure, obesity, and medication
compliance (hazard ratio 1.05, CI 0.73-1.51, P = .81). Results were similar in subgroups of patients with heart failure and those not taking β-blockers.
Discussion In this post hoc analysis of a multicenter, doubleblinded, randomized controlled trial of patients with CHD who were in sinus rhythm at baseline, we found that the PPARα activator gemfibrozil did not reduce the incidence of AF. To our knowledge, these are the first clinical data about the effects of a PPARα activator on AF incidence in humans. We also found, as expected, that incident AF occurred more frequently in patients who were older and had hypertension, heart failure, and previous coronary revascularization. Recently, there has been an increased interest in the antiarrhythmic functions of several classes of medications that are not traditionally considered antiarrhythmic
American Heart Journal Volume 157, Number 5
drugs. Post hoc evaluation of randomized clinical trials suggests that angiotensin-converting enzyme inhibitors and angiotensin receptor blockers reduce the incidence of AF and ventricular tachyarrhythmias in patients with heart failure, hypertension, and diabetes mellitus.17-19 Statins have also been suggested to reduce ventricular arrhythmias and AF in patients with CHD, with heart failure, and after electrical cardioversion based on data from observational studies and small randomized controlled trials.5-8 Furthermore, statin treatment has been shown to reduce postoperative AF after cardiac surgery.20 Whether the antiarrhythmic functions of these medications are due to the treatment of the underlying condition or some other pleiotropic activity is yet unclear. However, it is widely believed that the antiarrhythmic activity of statins is independent of their lipid-lowering effects. Furthermore, it is known that inflammation plays a prominent role in the initiation and maintenance of AF.21-23 Thus, through secondary reasoning, the antiinflammatory properties of the statins are thought to contribute to their antiarrhythmic effects. There is recent evidence that pioglitazone, an activator of PPARγ, reduces atrial fibrosis and susceptibility to AF in rabbits with heart failure.24,25 Activators of PPARα also have pleiotropic effects akin to pioglitazone and statins. These medications reduce inflammation and oxidative stress and improve endothelial function in animals.9-11 Furthermore, PPARα agonists reduce mortality and cardiac events in humans.12,13 In an observational study, Hanna et al1 and Kantharia2 observed that lipid-lowering therapy, consisting of statins and fibrates, was associated with lower AF prevalence among patients with heart failure. In addition, Mitchell et al3 reported that ventricular tachycardia and fibrillation were less common among participants of the Antiarrhythmics Versus Implantable Defibrillators study who were taking lipidlowering medications. However, both of these studies were observational in design; and in most (80%-90%) cases, the lipid-lowering therapy consisted of statins. Thus, teasing out the effects of statins from fibrates was not possible. In this regard, our study fills an important gap in knowledge, suggesting that statins (not fibrates) are likely responsible for the observed differences in AF incidence in these studies. In this study, patients who developed AF were older and more frequently had a history of hypertension, heart failure, coronary revascularization, and nitrate use at baseline. Age, hypertension, and heart failure have been well-established risk factors for AF in previous large studies, suggesting that our study findings are generalizable to other populations. Higher prevalence of coronary revascularization and nitrate use at baseline may suggest presence of more severe CHD and ischemia in those with AF. Our data showed that the cumulative incidences of AF started to separate in the gemfibrozil and placebo groups
Adabag et al 917
within 6 months after randomization and was higher in patients assigned to placebo until the 3-year follow-up (Figure 2), although the difference was not statistically significant. However, at the 3-year follow-up, a substantial increase in AF incidence was seen in patients assigned to gemfibrozil, merging the incidence curves. This observation might merely be due to chance, or confounding may have been introduced by starting new medications (eg, statins). Noncompliance of the study medication does not appear to be the issue, however, as our results did not change after including compliance in the statistical models. The incidence of AF was lower among the VA-HIT study participants in comparison to a large cohort of veterans (N13,000) we recently studied.6 The participants of both studies were similar in baseline characteristics and were contemporaneous. Thus, the lower-than-expected AF incidence in the present study cohort supports a previously known notion that the participants of randomized clinical trials have a lower risk of adverse outcomes than eligible patients in the general population who were not enrolled in these trials.26 The strengths of this investigation include its randomized controlled trial design. Randomized controlled trials are considered to be the finest clinical research design to answer a question because of their ability to control for known and unknown confounders by way of randomization. Furthermore, to our knowledge, these are the first data on the antiarrhythmic activity of a PPARα agonist. Finally, the utilization of competing risk analysis is another notable strength. However, post hoc analyses such as this one also have some inherent limitations. First, the diagnosis of AF was solely based on ECGs, raising the possibility that AF events that occurred between annual ECGs but resolved may have been missed. Although ECGs were also obtained when clinically indicated (eg, during symptomatic AF episodes), asymptomatic events might not have been captured; or AF might have resolved by the time an ECG was performed. Furthermore, the VA-HIT population consisted of men with CHD, which may differ from a typical study cohort exclusively designed to investigate AF. Finally, our lower-than-expected event rate has reduced our power to detect a statistically significant difference between the 2 treatment groups. In conclusion, in this post hoc analysis of a randomized controlled trial of patients with CHD who were in sinus rhythm at baseline, the PPARα activator gemfibrozil did not reduce the incidence of AF.
Disclosures Dr Adabag is supported, in part, by Veterans Affairs Clinical Science Research and Development Service (grant 04S-CRCOE 001), Washington, DC. Drs Mithani and Aloul have invested equal effort in this project.
American Heart Journal May 2009
918 Adabag et al
References 1. Hanna IR, Heeke B, Bush H, et al. Lipid-lowering drug use is associated with reduced prevalence of atrial fibrillation in patients with left ventricular systolic dysfunction. Heart Rhythm 2006;3:881-6. 2. Kantharia BK. Lipid lowering drugs: beyond lowering lipid levels to their influence on cardiac arrhythmias. Heart Rhythm 2006;3:887-8. 3. Mitchell LB, Powell JL, Gillis AM, et al. Are lipid-lowering drugs also antiarrhythmic drugs? An analysis of the Antiarrhythmic Versus Implantable Defibrillators (AVID) Trial. J Am Coll Cardiol 2003;42: 81-7. 4. De Sutter J, Tavernier R, De Buyzere M, et al. Lipid lowering drugs and recurrences of life-threatening ventricular arrhythmias in high-risk patients. J Am Coll Cardiol 2000;36:766-72. 5. Young-Xu Y, Jabbour S, Goldberg R, et al. Usefulness of statin drugs in protecting against atrial fibrillation in patients with coronary artery disease. Am J Cardiol 2003;92:1379-83. 6. Adabag AS, Nelson DB, Bloomfield HE, et al. Effects of statin therapy on preventing atrial fibrillation in coronary disease and heart failure. Am Heart J 2007;154:1140-5. 7. Siu CW, Lau CP, Tse HF. Prevention of atrial fibrillation recurrence by statin therapy in patients with lone atrial fibrillation after successful cardioversion. Am J Cardiol 2003;92:1343-5. 8. Ozaydin M, Varol E, Aslan SM, et al. Effect of atorvastatin on the recurrence rates of atrial fibrillation after electrical cardioversion. Am J Cardiol 2006;97:1490-3. 9. Staels B, Fruchart JC. Therapeutic roles of peroxisome proliferator–activated receptor agonists. Diabetes 2005;54: 2460-70. 10. Diep QN, Benkirane K Amiri F, et al. PPAR alpha activator fenofibrate inhibits myocardial inflammation and fibrosis in angiotensin II–infused rats. J Mol Cell Cardiol 2004;36:295-304. 11. Kleemann R, Verschuren L, de Rooji BJ, et al. Evidence for anti-inflammatory activity of statins and PPAR alpha activators in human C-reactive protein transgenic mice in vivo and in cultured human hepatocytes in vitro. Blood 2004;103:4188-94. 12. Rubins HB, Robins SJ, Collins D, et al. Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol. Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group. N Engl J Med 1999;341:410-8. 13. Bloomfield HE. The role of fibrates in a statin world. Arch Intern Med 2006;166:715-6. 14. Rubins HB, Robins SJ, Iwane MK, et al. Rationale and design of the Department of Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial (HIT) for secondary prevention of coronary artery
15.
16. 17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
disease in men with low high-density lipoprotein cholesterol and desirable low-density lipoprotein cholesterol. Am J Cardiol 1993;71: 45-52. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966: 3163-70. Cox DR. Regression models and life-tables. J R Stat Soc 1972;34: 187-202. Pedersen OD, Bagger H, Kober L, et al. Trandolapril reduces the incidence of atrial fibrillation after acute myocardial infarction in patients with left ventricular dysfunction. Circulation 1999;100: 376-80. Vermes E, Tardif JC, Bourassa MG, et al. Enalapril decreases the incidence of atrial fibrillation in patients with left ventricular dysfunction: insight from the Studies Of Left Ventricular Dysfunction (SOLVD) trials. Circulation 2003;107:2926-31. Maggioni AP, Latini R, Carson PE, et al. Valsartan reduces the incidence of atrial fibrillation in patients with heart failure: results from the Valsartan Heart Failure Trial (Val-HeFT). Am Heart J 2005;149: 548-57. Patti G, Chello M, Candura D, et al. Randomized trial of atorvastatin for reduction of postoperative atrial fibrillation in patients undergoing cardiac surgery: results of the ARMYDA-3 (Atorvastatin for Reduction of MYocardial Dysrhythmia After cardiac surgery) study. Circulation 2006;114:1455-61. Chung MK, Martin DO, Sprecher D, et al. C-reactive protein elevation in patients with atrial arrhythmias: inflammatory mechanisms and persistence of atrial fibrillation. Circulation 2001; 104:2886-91. Aviles RJ, Martin DO, Apperson-Hansen C, et al. Inflammation as a risk factor for atrial fibrillation. Circulation 2003;108: 3006-10. Engelmann MD, Svendsen JH. Inflammation in the genesis and perpetuation of atrial fibrillation. Eur Heart J 2005;26: 2083-92. Shimano M, Tsuji Y Inden Y, et al. Pioglitazone, a peroxisome proliferator–activated receptor–gamma activator, attenuates atrial fibrosis and atrial fibrillation promotion in rabbits with congestive heart failure. Heart Rhythm 2008;5:451-9. Wilson LD, Tsai CT. Heart failure–related atrial fibrillation: a new model for a new prevention strategy? Heart Rhythm 2008;5: 460-1. Steg PG, López-Sendón J, Lopez de Sa E, et al. External validity of clinical trials in acute myocardial infarction. Arch Intern Med 2007; 167:68-73.
Efficacy of gemfibrozil in the primary prevention of atrial fibrillation in a large randomized controlled trial A. Selcuk Adabag, MD, MS, a,b,c Salima Mithani, MD, a,c Basel Al Aloul, MD, a,c Dorothea Collins, ScD, d Stefan Bertog, MD, a,c and Hanna E. Bloomfield, MD, MPH b,c,e for The Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group Minneapolis, MN; and West Haven, CT
Background
Peroxisome proliferator–activated receptor α (PPARα) activators reduce inflammation and oxidative stress. Inflammation plays an important role in the initiation and maintenance of atrial fibrillation (AF). It has been suggested that PPARα activators may have antiarrhythmic properties, but no clinical data exist. The objective of this study was to investigate whether the PPARα activator gemfibrozil prevents or delays the development of AF in patients with coronary heart disease.
Methods
We retrospectively analyzed the electrocardiograms (ECGs) performed in the Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial, a multicenter, randomized, double-blinded, secondary prevention trial of gemfibrozil and matching placebo. The ECGs were performed annually or biannually and when clinically indicated. Participants who were in AF on baseline ECG were excluded from the present analysis. Relative risk for AF was calculated from Cox regression with death as a competing risk factor.
Results
A total of 12,605 ECGs from 2,130 participants were interpreted (5.9 ± 2.1 ECGs per participant, range 2-20). At baseline, the gemfibrozil (n = 1,070) and placebo (n = 1,060) groups were well matched. Mean age was 64.1 ± 7.1 years. Over 4.4 ± 1.5 years of follow-up, 123 (5.8%) participants developed new AF. There was no difference in AF incidence between the gemfibrozil and placebo groups (64/1,070 vs 59/1,060, respectively; P = .33). In Cox regression, the risk of AF was similar between the 2 study groups (hazard ratio 1.04, 95% CI 0.73-1.49, P = .82).
Conclusions
In this post hoc analysis of a multicenter, double-blinded, randomized controlled trial, the PPARα activator gemfibrozil did not reduce the 4-year incidence of AF among men with coronary heart disease. (Am Heart J 2009;157:913-8.)
Treatment options for the primary prevention of atrial fibrillation/flutter (AF) are limited in the nonsurgical setting. Recent evidence suggests that lipid-lowering therapy (consisting mostly of statins but also fibrates) may reduce the incidence of ventricular tachycardia and AF in patients with cardiomyopathy, independent of their effects on serum lipid profile.1-4 Statins have
From the aDivision of Cardiology, bCenter for Chronic Disease Outcomes Research, Veterans Affairs Medical Center, Minneapolis, MN, cUniversity of Minnesota, Minneapolis, MN, dCooperative Studies Program, Veterans Affairs Medical Center, West Haven, CT, and eDepartment of General Internal Medicine, Veterans Affairs Medical Center, Minneapolis, MN. Submitted October 21, 2008; accepted February 19, 2009. Reprint requests: A. Selcuk Adabag, MD, Section of Cardiology (111 C), Veterans Affairs Medical Center, One Veterans Dr, Minneapolis, MN 55417. E-mail: [email protected] 0002-8703/$ - see front matter Published by Mosby, Inc. doi:10.1016/j.ahj.2009.02.015
also been suggested to prevent AF in the setting of coronary heart disease (CHD),5 heart failure,6 and after electrical cardioversion.7,8 The mechanism by which statins reduce AF is unknown; however, their antiinflammatory and antifibrotic effects have been considered to be responsible. Peroxisome proliferator–activated receptor α (PPARα) regulates fatty acid metabolism, inflammation, and endothelial function.9 Activation of PPARα by fibrates reduces the serum levels and cellular expression of inflammatory mediators in animal models10,11 and decreases adverse cardiac events in humans.12,13 Because these effects are similar to the pleiotropic properties of statins, the possibility that PPARα activators may also prevent cardiac arrhythmias has recently been raised.1,2 However, there are no clinical data to examine whether PPARα agonists have antiarrhythmic activity. The objective of the present study was to determine whether gemfibrozil, a PPARα agonist, prevents or
914 Adabag et al
American Heart Journal May 2009
Figure 1
Selection of the 2,130 participants for retrospective ECG analysis in the present study. Flowchart describing the selection of patients.
delays the development of AF in patients with CHD who participated in the Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial (VA-HIT).12
Methods Patient selection The VA-HIT was a multicenter, double-blinded, randomized controlled trial that tested whether gemfibrozil was effective in secondary prevention of CHD in men with low levels of high-density lipoprotein. Trial design and results have previously been published elsewhere.12,14 Briefly, men b74 years old with a history of CHD, high-density lipoprotein cholesterol ≤40 mg/dL (≤3.6 mmol/L), low-density lipoprotein cholesterol ≤140 mg/dL, and triglyceride level ≤300 mg/dL were randomized to gemfibrozil 1,200 mg/d or matching placebo. During the VA-HIT, participants had electrocardiograms (ECGs) recorded annually or biannually and more frequently if clinically indicated. For the present study, we retrospectively analyzed all ECGs. The participants who were in sinus rhythm
at baseline and who had at least 1 follow-up ECG were included in the present analysis. Of the 2,531 participants in the VA-HIT, 182 were excluded because no ECGs were available (Figure 1). An additional 218 participants were excluded because baseline rhythm was AF (n = 38) or paced/ uninterpretable (n = 15), or because there were no follow-up ECGs after the baseline (n = 155) or that they were uninterpretable (n = 10). The remaining 2,130 participants (1,070 assigned to gemfibrozil and 1,060 to placebo) who were in sinus rhythm at first ECG were included in the present analysis (Figure 1). The characteristics of the participants included in the present analysis did not statistically differ from the characteristics of the entire VA-HIT cohort (data not shown). This study was approved by the human studies subcommittee of the Research and Development committee at the Minneapolis Veterans Affairs Medical Center. Individual consent was waived. This work was supported by a grant from the Minneapolis Veterans Research Institute and the Cooperative Studies Program of the Department of Veterans Affairs Office of Research and Development. The authors are solely responsible
American Heart Journal Volume 157, Number 5
Adabag et al 915
Table I. Baseline characteristics of the 2130 participants of the present study Characteristics Age (y) Race (%) White Black Other Education N12 y (%) Current smoker (%) Alcohol consumption (%) b1 Drink/d 1-3 Drinks/d N3 Drinks/d Hypertension (%) Diabetes mellitus (%) Prior myocardial infarction Heart failure (%) CABG or PCI (%) Medication use (%) Aspirin Nitrates Calcium-channel blockers ACE inhibitors β-Blockers Body mass index (kg/m2) LDL cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL) Glucose (mg/dL) WBC (cells/mL)
Placebo (n = 1060) 64.1 ± 7
Gemfibrozil (n = 1070) 64.0 ± 7.2
P value .93 .20
92 6 2 68 17
91 6 3 67 20
94 6 0.1 56 30 62 8 60
92 8 0.3 55 29 62 8 56
.67 .75 .78 .95 .14
84 46 51
82 46 52
.22 .85 .53
20 44 29.1 ± 4.8 111 ± 22 31 ± 5 161 ± 67 113 ± 33 7.1 ± 1.9
20 44 28.9 ± 4.7 111 ± 22 31 ± 5 162 ± 69 115 ± 37 7.2 ± 2.5
.83 .88 .24 .90 .89 .70 .21 .42
.76 .09 .23
Plus-minus values are means ± SD. CABG, Coronary artery bypass surgery; PCI, percutaneous coronary intervention; ACE, angiotensin-converting enzyme; LDL, lowdensity lipoprotein; HDL, high-density lipoprotein; WBC, white blood cell.
for the design and conduct of this study, all study analyses, the drafting and editing of the paper, and its final contents.
Outcome ascertainment The primary outcome variable was time to the first ECG with AF. Two experienced physicians (S.M. and B.A.A.) independently reviewed all of the ECGs. There was almost perfect agreement (κ = 0.96) between the 2 interpreters. All ECGs with a rhythm other than sinus and those where there were disagreements in interpretation were reread by a cardiologist with extensive experience in ECG analysis (S.B.).
Sample size calculation Based on our previous experience from a large observational data set of veterans with CHD, we estimated a 10% dropout rate due to AF at baseline and expected a 14% incidence of new AF over 5 years of follow-up. Thus, the present study had 80% power to detect a 28% relative reduction (absolute from 14% to 10.1%) in AF incidence with gemfibrozil.
Statistical analysis All analyses followed an intention-to-treat approach. Categorical variables were presented as frequency and continuous
Table II. Baseline characteristics of the participants who developed AF versus those who maintained sinus rhythm Characteristics
AF (n = 123)
Age (y) Race (%) White Black Other Education N12 y (%) Current smoker (%) Alcohol consumption (%) b1 Drink/d 1-3 Drinks/d N3 Drinks/d Hypertension (%) Diabetes mellitus (%) Prior myocardial infarction Heart failure (%) CABG or PCI (%) Medication use (%) Aspirin Nitrates Calcium-channel blockers ACE inhibitors β-Blockers Body mass index (kg/m2) LDL cholesterol (mg/dL) HDL cholesterol (mg/dL) Triglycerides (mg/dL) Glucose (mg/dL) WBC (cells/mL)
67.1 ± 5
Sinus rhythm (n = 2007) P value 63.9 ± 7.2
b.0001 .74
93 5 2 67 17
92 6 2 68 19
93 7 0 67 28 60 12 50
93 7 0.2 55 29 62 7 59
.01 .76 .65 .05 .05
85 59 54
83 45 51
.45 .005 .63
29 44 28.9 ± 4.6 107 ± 21 31 ± 6 170 ± 76 115 ± 38 7.2 ± 2.1
20 44 29.0 ± 4.7 111 ± 22 31 ± 5 161 ± 68 114 ± 35 7.2 ± 2.2
.01 .92 .88 .03 .97 .14 .72 .67
.90 .59 .88
Plus-minus values are means ± SD.
variables as mean ± 1 SD or median and interquartile range. Gemfibrozil and placebo groups were compared using t tests and nonparametric Wilcoxon rank tests for continuous variables and χ2 test for categorical variables. Follow-up was calculated as the number of days from the baseline ECG to the last follow-up ECG. Analyses of time-to-event data were performed with the log-rank test.15 Patients were removed from all time-to-event analyses at the time of death. Relative risks were calculated from Cox models16 with treatment as the single covariate and death as a competing risk factor. Additional Cox models including age, hypertension, diabetes mellitus, smoking, heart failure, body mass index, medication compliance, and death were performed. All analyses were performed using SAS (Cary, NC) version 9.2. The proportional hazards assumption was tested and confirmed. A P value b .05 was interpreted as significant.
Results Baseline characteristics of the study participants are displayed in Table I. Mean age was 64.1 ± 7.1 years, and all participants were male. The participants assigned to gemfibrozil and those assigned to placebo were well matched with regard to baseline characteristics (Table I).
American Heart Journal May 2009
916 Adabag et al
Figure 2
Kaplan-Meier estimates of the incidence of AF in the gemfibrozil and placebo groups.
Predictors of AF A total of 12,605 ECGs from 2,130 participants were interpreted (5.9 ± 2.1 ECGs per participant, range 2-20). Over 4.4 ± 1.5 years of follow-up, 123 (5.8%) participants developed new AF. Participants who developed new AF were older and more likely to have a history of hypertension, heart failure, coronary revascularization, and nitrate use at baseline than those who did not develop AF (Table II). Effect of gemfibrozil on AF There was no difference in AF incidence between the gemfibrozil and placebo groups (64/1,070 vs 59/1,060, respectively; P = .33). In Cox regression, after considering death as a competing risk, the risk of AF was similar between the gemfibrozil and placebo groups (hazard ratio 1.04, 95% CI 0.73-1.49, P = .82) (Figure 2). Results did not change after adjusting for age, hypertension, diabetes mellitus, heart failure, obesity, and medication
compliance (hazard ratio 1.05, CI 0.73-1.51, P = .81). Results were similar in subgroups of patients with heart failure and those not taking β-blockers.
Discussion In this post hoc analysis of a multicenter, doubleblinded, randomized controlled trial of patients with CHD who were in sinus rhythm at baseline, we found that the PPARα activator gemfibrozil did not reduce the incidence of AF. To our knowledge, these are the first clinical data about the effects of a PPARα activator on AF incidence in humans. We also found, as expected, that incident AF occurred more frequently in patients who were older and had hypertension, heart failure, and previous coronary revascularization. Recently, there has been an increased interest in the antiarrhythmic functions of several classes of medications that are not traditionally considered antiarrhythmic
American Heart Journal Volume 157, Number 5
drugs. Post hoc evaluation of randomized clinical trials suggests that angiotensin-converting enzyme inhibitors and angiotensin receptor blockers reduce the incidence of AF and ventricular tachyarrhythmias in patients with heart failure, hypertension, and diabetes mellitus.17-19 Statins have also been suggested to reduce ventricular arrhythmias and AF in patients with CHD, with heart failure, and after electrical cardioversion based on data from observational studies and small randomized controlled trials.5-8 Furthermore, statin treatment has been shown to reduce postoperative AF after cardiac surgery.20 Whether the antiarrhythmic functions of these medications are due to the treatment of the underlying condition or some other pleiotropic activity is yet unclear. However, it is widely believed that the antiarrhythmic activity of statins is independent of their lipid-lowering effects. Furthermore, it is known that inflammation plays a prominent role in the initiation and maintenance of AF.21-23 Thus, through secondary reasoning, the antiinflammatory properties of the statins are thought to contribute to their antiarrhythmic effects. There is recent evidence that pioglitazone, an activator of PPARγ, reduces atrial fibrosis and susceptibility to AF in rabbits with heart failure.24,25 Activators of PPARα also have pleiotropic effects akin to pioglitazone and statins. These medications reduce inflammation and oxidative stress and improve endothelial function in animals.9-11 Furthermore, PPARα agonists reduce mortality and cardiac events in humans.12,13 In an observational study, Hanna et al1 and Kantharia2 observed that lipid-lowering therapy, consisting of statins and fibrates, was associated with lower AF prevalence among patients with heart failure. In addition, Mitchell et al3 reported that ventricular tachycardia and fibrillation were less common among participants of the Antiarrhythmics Versus Implantable Defibrillators study who were taking lipidlowering medications. However, both of these studies were observational in design; and in most (80%-90%) cases, the lipid-lowering therapy consisted of statins. Thus, teasing out the effects of statins from fibrates was not possible. In this regard, our study fills an important gap in knowledge, suggesting that statins (not fibrates) are likely responsible for the observed differences in AF incidence in these studies. In this study, patients who developed AF were older and more frequently had a history of hypertension, heart failure, coronary revascularization, and nitrate use at baseline. Age, hypertension, and heart failure have been well-established risk factors for AF in previous large studies, suggesting that our study findings are generalizable to other populations. Higher prevalence of coronary revascularization and nitrate use at baseline may suggest presence of more severe CHD and ischemia in those with AF. Our data showed that the cumulative incidences of AF started to separate in the gemfibrozil and placebo groups
Adabag et al 917
within 6 months after randomization and was higher in patients assigned to placebo until the 3-year follow-up (Figure 2), although the difference was not statistically significant. However, at the 3-year follow-up, a substantial increase in AF incidence was seen in patients assigned to gemfibrozil, merging the incidence curves. This observation might merely be due to chance, or confounding may have been introduced by starting new medications (eg, statins). Noncompliance of the study medication does not appear to be the issue, however, as our results did not change after including compliance in the statistical models. The incidence of AF was lower among the VA-HIT study participants in comparison to a large cohort of veterans (N13,000) we recently studied.6 The participants of both studies were similar in baseline characteristics and were contemporaneous. Thus, the lower-than-expected AF incidence in the present study cohort supports a previously known notion that the participants of randomized clinical trials have a lower risk of adverse outcomes than eligible patients in the general population who were not enrolled in these trials.26 The strengths of this investigation include its randomized controlled trial design. Randomized controlled trials are considered to be the finest clinical research design to answer a question because of their ability to control for known and unknown confounders by way of randomization. Furthermore, to our knowledge, these are the first data on the antiarrhythmic activity of a PPARα agonist. Finally, the utilization of competing risk analysis is another notable strength. However, post hoc analyses such as this one also have some inherent limitations. First, the diagnosis of AF was solely based on ECGs, raising the possibility that AF events that occurred between annual ECGs but resolved may have been missed. Although ECGs were also obtained when clinically indicated (eg, during symptomatic AF episodes), asymptomatic events might not have been captured; or AF might have resolved by the time an ECG was performed. Furthermore, the VA-HIT population consisted of men with CHD, which may differ from a typical study cohort exclusively designed to investigate AF. Finally, our lower-than-expected event rate has reduced our power to detect a statistically significant difference between the 2 treatment groups. In conclusion, in this post hoc analysis of a randomized controlled trial of patients with CHD who were in sinus rhythm at baseline, the PPARα activator gemfibrozil did not reduce the incidence of AF.
Disclosures Dr Adabag is supported, in part, by Veterans Affairs Clinical Science Research and Development Service (grant 04S-CRCOE 001), Washington, DC. Drs Mithani and Aloul have invested equal effort in this project.
American Heart Journal May 2009
918 Adabag et al
References 1. Hanna IR, Heeke B, Bush H, et al. Lipid-lowering drug use is associated with reduced prevalence of atrial fibrillation in patients with left ventricular systolic dysfunction. Heart Rhythm 2006;3:881-6. 2. Kantharia BK. Lipid lowering drugs: beyond lowering lipid levels to their influence on cardiac arrhythmias. Heart Rhythm 2006;3:887-8. 3. Mitchell LB, Powell JL, Gillis AM, et al. Are lipid-lowering drugs also antiarrhythmic drugs? An analysis of the Antiarrhythmic Versus Implantable Defibrillators (AVID) Trial. J Am Coll Cardiol 2003;42: 81-7. 4. De Sutter J, Tavernier R, De Buyzere M, et al. Lipid lowering drugs and recurrences of life-threatening ventricular arrhythmias in high-risk patients. J Am Coll Cardiol 2000;36:766-72. 5. Young-Xu Y, Jabbour S, Goldberg R, et al. Usefulness of statin drugs in protecting against atrial fibrillation in patients with coronary artery disease. Am J Cardiol 2003;92:1379-83. 6. Adabag AS, Nelson DB, Bloomfield HE, et al. Effects of statin therapy on preventing atrial fibrillation in coronary disease and heart failure. Am Heart J 2007;154:1140-5. 7. Siu CW, Lau CP, Tse HF. Prevention of atrial fibrillation recurrence by statin therapy in patients with lone atrial fibrillation after successful cardioversion. Am J Cardiol 2003;92:1343-5. 8. Ozaydin M, Varol E, Aslan SM, et al. Effect of atorvastatin on the recurrence rates of atrial fibrillation after electrical cardioversion. Am J Cardiol 2006;97:1490-3. 9. Staels B, Fruchart JC. Therapeutic roles of peroxisome proliferator–activated receptor agonists. Diabetes 2005;54: 2460-70. 10. Diep QN, Benkirane K Amiri F, et al. PPAR alpha activator fenofibrate inhibits myocardial inflammation and fibrosis in angiotensin II–infused rats. J Mol Cell Cardiol 2004;36:295-304. 11. Kleemann R, Verschuren L, de Rooji BJ, et al. Evidence for anti-inflammatory activity of statins and PPAR alpha activators in human C-reactive protein transgenic mice in vivo and in cultured human hepatocytes in vitro. Blood 2004;103:4188-94. 12. Rubins HB, Robins SJ, Collins D, et al. Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol. Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group. N Engl J Med 1999;341:410-8. 13. Bloomfield HE. The role of fibrates in a statin world. Arch Intern Med 2006;166:715-6. 14. Rubins HB, Robins SJ, Iwane MK, et al. Rationale and design of the Department of Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial (HIT) for secondary prevention of coronary artery
15.
16. 17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
disease in men with low high-density lipoprotein cholesterol and desirable low-density lipoprotein cholesterol. Am J Cardiol 1993;71: 45-52. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966: 3163-70. Cox DR. Regression models and life-tables. J R Stat Soc 1972;34: 187-202. Pedersen OD, Bagger H, Kober L, et al. Trandolapril reduces the incidence of atrial fibrillation after acute myocardial infarction in patients with left ventricular dysfunction. Circulation 1999;100: 376-80. Vermes E, Tardif JC, Bourassa MG, et al. Enalapril decreases the incidence of atrial fibrillation in patients with left ventricular dysfunction: insight from the Studies Of Left Ventricular Dysfunction (SOLVD) trials. Circulation 2003;107:2926-31. Maggioni AP, Latini R, Carson PE, et al. Valsartan reduces the incidence of atrial fibrillation in patients with heart failure: results from the Valsartan Heart Failure Trial (Val-HeFT). Am Heart J 2005;149: 548-57. Patti G, Chello M, Candura D, et al. Randomized trial of atorvastatin for reduction of postoperative atrial fibrillation in patients undergoing cardiac surgery: results of the ARMYDA-3 (Atorvastatin for Reduction of MYocardial Dysrhythmia After cardiac surgery) study. Circulation 2006;114:1455-61. Chung MK, Martin DO, Sprecher D, et al. C-reactive protein elevation in patients with atrial arrhythmias: inflammatory mechanisms and persistence of atrial fibrillation. Circulation 2001; 104:2886-91. Aviles RJ, Martin DO, Apperson-Hansen C, et al. Inflammation as a risk factor for atrial fibrillation. Circulation 2003;108: 3006-10. Engelmann MD, Svendsen JH. Inflammation in the genesis and perpetuation of atrial fibrillation. Eur Heart J 2005;26: 2083-92. Shimano M, Tsuji Y Inden Y, et al. Pioglitazone, a peroxisome proliferator–activated receptor–gamma activator, attenuates atrial fibrosis and atrial fibrillation promotion in rabbits with congestive heart failure. Heart Rhythm 2008;5:451-9. Wilson LD, Tsai CT. Heart failure–related atrial fibrillation: a new model for a new prevention strategy? Heart Rhythm 2008;5: 460-1. Steg PG, López-Sendón J, Lopez de Sa E, et al. External validity of clinical trials in acute myocardial infarction. Arch Intern Med 2007; 167:68-73.