Obesity and Outcomes Among Patients With Established Atrial Fibrillation
Article Outline
Atrial fibrillation (AF) and obesity have reached epidemic proportions. The impact of obesity on clinical outcomes in patients with established AF is unknown. We analyzed 2,492 patients in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) study. Body mass index (BMI) was evaluated as a categorical variable (normal 18.5 to <25 kg/m2, overweight 25 to <30 kg/m2, obese ≥30 kg/m2). Rate of death from any cause was higher in the normal BMI group (5.8 per 100 patient-years) than in the overweight and obese groups (3.9 and 3.7, respectively). Cardiovascular death rate was highest in the normal BMI group (3.1 per 100 patient-years), lowest in the overweight group (1.5 per 100 patient-years), and intermediate in the obese group (2.1 per 100 patient-years). After adjustment for baseline factors, differences in risk of death from any cause were no longer significant. However, overweight remained associated with a lower risk of cardiovascular death (hazard ratio 0.47, p = 0.002). Obese patients were more likely to have an uncontrolled heart rate at rest, but rhythm-control strategy success was similar across BMI categories. In each BMI category, risk of death from any cause was similar for patients randomized to a rhythm- or rate-control strategy. In conclusion, in patients with established AF, overweight and obesity do not adversely affect overall survival. Obesity does not appear to affect the relative benefit of a rate- or rhythm-control strategy.
Obesity, which is epidemic in the United States, is well established as an independent risk factor for all-cause and cardiovascular mortality in the general population.1, 2, 3, 4 However, in patients with certain established cardiovascular conditions, including coronary heart disease5 and heart failure,6 overweight and mild obesity are associated with a lower risk of death—the “obesity paradox.”7, 8 Recently, obesity has been demonstrated to be an independent risk factor for development of atrial fibrillation (AF), a condition that is also rapidly increasing in prevalence.9, 10 Given this interconnection between obesity and AF, we sought to determine the relation between obesity and clinical outcomes in patients with established AF. The Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) study provided a large, well-characterized cohort of patients with high-risk AF in which to do so.
Methods
Details of the AFFIRM protocol have been reported previously.11, 12 Briefly, patients with AF requiring long-term treatment and with ≥1 risk factor for death or stroke were enrolled. Patients were randomized to a rate-control strategy or a rhythm-control strategy. Mean follow-up was 3.5 years (range 0 to 6).11 Heart rate control at rest was assessed at the 1-year visit and was considered adequate if ≤80 beats/min. Classification of all fatal events and strokes was adjudicated by an independent clinical events committee. The institutional review board of the University of Washington approved the protocol, and each clinical site obtained approval from its local institutional review board. Every patient gave signed, informed consent to participate in the study. The study was supported by the National Heart, Lung, and Blood Institute (http://www.ClinicalTrials.gov, identifier NCT00000556).
Height and weight were recorded for AFFIRM participants at the baseline visit. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of height in meters. BMI was categorized according to the World Health Organization/National Institutes of Health classification scheme, with normal defined as 18.5 to <25 kg/m2, overweight as 25 to <30 kg/m2, and obese as ≥30 kg/m2. The obese category was further subdivided into mildly obese (BMI 30 to <35 kg/m2), severely obese (BMI 35 to <40 kg/m2), and extremely obese (BMI ≥40 kg/m2).13 Underweight patients (BMI <18.5 kg/m2) were excluded from analysis.
Baseline characteristics were compared across the 3 BMI groups using analysis of variance. The primary end point was a comparison of death from any cause. Other clinical end points included cardiovascular death and stroke. Number of events per 100 person-years at risk was calculated for patients in each BMI group. Rate ratios were then calculated, comparing each BMI group with the other 2 groups. Corresponding p values were calculated using Fisher's 2-tailed exact chi-square test. Cox proportional hazards modeling, using a stepwise regression procedure, was used to determine the independent relation between BMI category and clinical events. Covariables that remained statistically significant and those that were deemed clinically relevant were included in the final model. These included age; gender; history of hypertension, diabetes mellitus, smoking, coronary artery disease, heart failure, or stroke; rhythm at randomization; left ventricular ejection fraction; and randomized treatment group.
In the rate-control arm, we assessed mean heart rate at rest and the proportion of patients with uncontrolled heart rate at rest (>80 beats/min) at the 1-year visit in patients remaining in AF. In the rhythm-control arm, we assessed the proportion of patients never achieving sinus rhythm, the proportion of patients with ≥1 episode of AF during the study period, and the proportion of electrical cardioversions that were successful in restoring sinus rhythm. All analyses were conducted using SAS 8.2 (SAS Institute, Cary, North Carolina). A 2-sided p value <0.05 was considered to represent statistical significance.
Results
BMI was available for 2,518 of 4,060 patients enrolled in AFFIRM. After exclusion of underweight patients (n = 26), the cohort for this analysis consisted of 2,492 patients (61.4% of total AFFIRM population). Baseline characteristics of patients with and without missing BMI data were generally similar, although patients with missing BMI data were significantly more likely to have a history of coronary artery disease (40% vs 37%) and heart failure (26% vs 21%) and were less likely to be enrolled after a first episode of AF (31% vs 38%).
Mean BMI was 29.0 ± 5.9 kg/m2. The normal BMI, overweight, and obese groups included 637 patients (25.5%), 965 patients (38.7%), and 890 patients (35.7%), respectively. Of patients in the obese group, 556 were mildly obese (22.3% of total), 207 were severely obese (8.3% of total), and 127 were extremely obese (5.1% of total). Significant differences between BMI groups were noted with respect to numerous baseline co-morbidities and medications (Table 1).
Table 1. Baseline characteristics according to body mass index category
| Variable | BMI (kg/m2) | p Value | ||
|---|---|---|---|---|
| 18.5–<25 | 25–<30 | ≥30 | ||
(n =637) | (n=965) | (n=890) | ||
| Age (years), mean ± SD | 72.4±7.1 | 70.7±7.5 | 66.4±8.4 | <0.001 |
| Body mass index (kg/m2), mean ± SD | 22.7±1.6 | 27.5±1.5 | 35.2±5.1 | <0.001 |
| Women | 321(50%) | 305(32%) | 353(40%) | <0.001 |
| Rate control group | 310(49%) | 499(52%) | 440(49%) | 0.4 |
| Hypertension | 392(62%) | 658(68%) | 724(81%) | <0.001 |
| Diabetes mellitus | 73(12%) | 163(17%) | 264(30%) | <0.001 |
| Smoking | 86(14%) | 99(10%) | 104(12%) | 0.1 |
| History of coronary artery disease | 201(32%) | 404(42%) | 319(36%) | <0.001 |
| Previous myocardial infarction | 87(14%) | 185(19%) | 141(16%) | 0.01 |
| Previous heart failure | 124(20%) | 190(20%) | 217(24%) | 0.02 |
| Previous stroke or transient ischemic attack | 98(15%) | 123(13%) | 93(11%) | 0.02 |
| First episode of atrial fibrillation | 214(34%) | 358(37%) | 943(38%) | 0.005 |
| Atrial fibrillation at randomization | 288(46%) | 444(46%) | 396(45%) | 0.8 |
| Systolic blood pressure (mm Hg), mean ± SD | 136±18.5 | 134±18.8 | 134±18.6 | 0.07 |
| Ejection fraction (%), mean ± SD | 55.9±7.9 | 55.7±7.3 | 55.3±7.8 | 0.5 |
| Left atrial dimension >4 cm | 424(67%) | 702(73%) | 716(76%) | <0.001 |
| Warfarin | 555(87%) | 844(87%) | 771(87%) | 0.9 |
| Angiotensin-converting enzyme inhibitor | 208(33%) | 371(38%) | 411(46%) | <0.001 |
| β blocker | 266(42%) | 451(47%) | 404(45%) | 0.1 |
| Calcium channel blocker | 142(23%) | 195(21%) | 234(27%) | 0.06 |
| Digoxin | 286(47%) | 351(39%) | 375(44%) | 0.005 |
| Amiodarone | 122(19%) | 188(19%) | 178(20%) | 0.6 |
| Sotalol | 104(16%) | 147(15%) | 152(17%) | 0.5 |
| Class I antiarrhythmic drug | 93(15%) | 108(12%) | 90(11%) | 0.02 |
Clinical event rates are presented in Table 2. There was no significant difference in the rate of stroke in patients in any of the BMI categories. Death from any cause occurred in 304 patients (12.2%) during the study period. Observed death rate was significantly higher in patients in the normal BMI group (5.8 deaths per 100 patient-years) than in those in the overweight group (3.9 deaths per 100 patient-years, p = 0.005) and obese group (3.7 deaths per 100 patient-years, p = 0.002). Of patients in the obese group, death rates were 3.5 per 100 patient-years in mildly obese patients, 4.4 per 100 patient-years in severely obese patients, and 3.3 per 100 patient-years in extremely obese patients (p = NS for comparisons among these groups).
Table 2. Rate of clinical events according to body mass index category
| Normal | Overweight | Obese | |
|---|---|---|---|
| (n=637) | (n=965) | (n=890) | |
| Death from any cause per 100 person-years (95% CI) | 5.8(4.8–7.1) | 3.9(3.3–4.7)⁎ | 3.7(3.0–4.5)⁎ |
| Cardiovascular death per 100 person-years (95% CI) | 3.1(2.4–4.0) | 1.5(1.1–2.0)⁎ | 2.1(1.6–2.8) |
| Stroke per 100 person-years (95% CI) | 1.1(0.7–1.8) | 1.2(0.8–1.7) | 1.2(0.8–1.7) |
⁎p <0.05 for comparison with normal body mass index group. |
Cardiovascular death occurred in 148 patients (5.9%). Observed cardiovascular death rate was highest in patients in the normal BMI group (3.1 per 100 patient-years), lowest in those in the overweight group (1.5 per 100 patient-years, p = 0.001 vs normal BMI group), and intermediate in those in the obese group (2.1 per 100 patient-years, p = 0.07 vs normal BMI group). Of patients in the obese group, cardiovascular death rates were 1.9 per 100 patient-years in mildly obese patients, 2.7 per 100 patient-years in severely obese patients, and 2.2 per 100 patient-years ing extremely obese patients (p = NS for comparisons among these groups).
Results of multivariable Cox proportional hazards models are presented in Table 3. After adjustment for relevant covariables, neither overweight nor obesity was significantly associated with risk of death from any cause. Overweight was significantly associated with a lower risk of cardiovascular death (hazard ratio [HR] 0.47, p = 0.002).
Table 3. Multivariable analysis of death from any cause and cardiovascular death
| Variable | HR (95% CI) | p Value |
|---|---|---|
| Death from any cause | ||
| Overweight | 0.74(0.53–1.03) | 0.07 |
| Obese | 0.82(0.57–1.18) | 0.3 |
| Age (per year) | 1.06(1.04–1.08) | <0.001 |
| Female gender | 0.85(0.63–1.15) | 0.3 |
| Hypertension | 1.29(0.94–1.75) | 0.12 |
| Diabetes | 1.99(1.48–2.66) | <0.001 |
| Smoking | 1.84(1.27–2.65) | 0.001 |
| Coronary artery disease | 1.82(1.37–2.42) | <0.001 |
| Congestive heart failure | 2.05(1.53–2.76) | <0.001 |
| Stroke | 1.69(1.20–2.39) | 0.003 |
| Ejection fraction <50% | 1.17(1.01–1.36) | 0.04 |
| Atrial fibrillation at randomization | 0.82(0.63–1.07) | 0.14 |
| Rhythm-control arm | 1.23(0.94–1.6) | 0.14 |
| Cardiovascular death | ||
| Overweight | 0.47(0.29–0.76) | 0.002 |
| Obese | 0.69(0.42–1.14) | 0.15 |
| Age (per year) | 1.04(1.01–1.07) | 0.005 |
| Female gender | 1.12(0.74–1.71) | 0.6 |
| Hypertension | 1.36(0.86–2.16) | 0.2 |
| Diabetes | 2.36(1.57–3.57) | <0.001 |
| Smoking | 1.55(0.92–2.63) | 0.10 |
| Coronary artery disease | 1.80(1.20–2.72) | 0.005 |
| Congestive heart failure | 2.56(1.69–3.89) | <0.001 |
| Stroke | 1.97(1.19–3.24) | 0.008 |
| Ejection fraction <50% | 1.41(1.16–1.72) | <0.001 |
| Atrial fibrillation at randomization | 0.80(0.55–1.17) | 0.3 |
| Rhythm-control arm | 1.14(0.78–1.66) | 0.5 |
Complete data were available for 1,249 patients randomized to a rate-control strategy. There were 522 patients (42%) in AF at the time of the baseline visit, and mean heart rate at rest was similar in patients in each BMI group (normal 80.7 ± 14.8 beats/min, overweight 80.0 ± 13.4 beats/min, obese 80.7 ± 14.3 beats/min, p = 0.9). At the 1 year visit, the number of patients in AF increased to 642, comprising 56% of patients with available 1-year follow-up. The proportion of patients receiving β blockers and calcium channel blockers at 1 year was similar in each BMI group, whereas digoxin use was significantly lower in overweight patients (normal BMI 59.2%, overweight 44.1%, obese 53.7%, p <0.01). For patients in AF, mean heart rate at rest at 1 year was significantly higher in obese patients (78.7 ± 13.2 beats/min) than in patients in the normal BMI group (75.1 ± 12.8 beats/min, p = 0.02) and tended to be higher than in overweight patients (76.1 ± 12.5 beats/min, p = 0.06). The proportion of patients in AF with uncontrolled heart rate at rest (>80 beats/min) at 1 year was significantly higher in obese patients than in patients in each of the other BMI groups (Figure 1). After adjustment for baseline differences and rate-control medication use at 1 year, obese patients remained more likely to have an uncontrolled heart rate at rest compared to patients in the normal BMI group (HR 1.9, 95% confidence interval [CI] 1.1 to 3.1, p = 0.01). Few patients in AF had a heart rate at rest >100 beats/min at the 1-year visit (5 in normal BMI group, 3.4%; 9 in overweight group, 3.4%; 11 in obese group, 4.6%; p = NS for all comparisons).
Figure 1.
Proportion of patients in AF with uncontrolled heart rate at rest at 1 year according to BMI category.
Complete data were available for 1,243 patients randomized to a rhythm-control strategy. The proportion of patients never achieving sinus rhythm was similar in patients with normal BMI (4.0%), overweight patients (4.8%), and obese patients (5.8%, p = NS). In addition, there was no significant difference in the proportion of patients with ≥1 episode of AF during the study period (66.1%, 67.3%, and 67.1%, respectively, p = NS).
Electrical cardioversion was attempted on 602 occasions in 400 patients. The proportion of patients who had ≥1 successful cardioversion was similar in patients with normal BMI (90.7%), overweight patients (85.2%), and obese patients (83.0%, p = NS), as was the proportion of first attempted cardioversions that were successful (87.2%, 83.2%, and 77.6%, respectively, p = NS).
In each BMI category, risk of death from any cause was similar for patients randomized to a rhythm-control strategy or rate-control strategy. Adjusted HRs were 0.92 (95% CI 0.56 to 1.51, p = 0.7) in patients with normal BMI, 1.45 (95% CI 0.93 to 2.23, p = 0.10) in overweight patients, and 1.35 (95% CI 0.84 to 2.17, p = 0.2) in obese patients. A test for interaction between BMI category and randomized treatment strategy with respect to risk of death was not significant (p >0.1).
Discussion
Despite the large number of patients with obesity and AF, the literature regarding the relation between these 2 epidemics is sparse. Obesity is an important risk factor for development of AF.9, 10, 14 In patients with established AF, obesity is associated with progression from paroxysmal to permanent AF,15 and obesity does not appear to affect outcomes after radiofrequency catheter ablation of AF.16 To our knowledge, however, ours is the first study to assess the impact of obesity on morbidity and mortality in patients with established AF.
In our study, patients in the overweight and obese BMI categories comprised a remarkable 74% of the study population. Although obesity is associated with an increased risk of overall and cardiovascular mortality in the general population,1, 2, 3, 4 we found that in established AF overweight and obesity—even severe and extreme obesity—were not associated with a higher risk of death, and overweight was independently associated with a substantially lower risk of death from cardiovascular causes. We also demonstrated that strict rate control was more difficult to achieve in obese patients, but that there was no relation between BMI category and ability to maintain rhythm control. Perhaps most importantly from a clinical standpoint, the relative benefit of a rate-control versus rhythm-control strategy was similar in each BMI group.
The obesity paradox—an association between overweight and obesity and a more favorable prognosis—is poorly understood but has been observed consistently in patients with established cardiovascular disease, including chronic coronary heart disease,5 acute myocardial infarction,17, 18 acute19 and chronic6, 8, 17, 20 heart failure, and peripheral arterial disease.21 In this first study to address this issue in patients with established AF, our findings were generally in line with previous studies: overweight and obesity were not associated with an increased risk of death from any cause and overweight was strongly associated with a lower risk of cardiovascular death.
Perhaps the relation between increasing BMI and improved outcomes is not physiologic but rather completely explained by confounding.6 In our study, however, after comprehensive adjustment for baseline factors, overweight remained significantly associated with a markedly lower risk of cardiovascular death. Hence, differences in baseline risk do not appear to completely explain this phenomenon.
Some authorities have suggested that the obesity paradox implies the presence of poor nutritional status and/or cachexia in patients with a lower BMI, and that this leads to the association between a higher BMI and a lower mortality risk.5, 8, 22 Indeed, it is a consistent finding that underweight patients have a substantially increased mortality risk.8, 23 However, in our study we excluded underweight patients yet still observed a lower risk of cardiovascular death in patients in the overweight group.
Perhaps the presence of additional body mass, although predisposing to development of cardiovascular disease and therefore premature death, is truly protective once cardiovascular disease has become established. Overweight and obese patients may have greater metabolic reserve than lean patients, allowing a greater tolerance for metabolic stress.6 Adipose tissue produces soluble tumor necrosis factor-α receptors, which may play an anti-inflammatory, cardioprotective role.24 The higher blood pressure levels seen in overweight and obese patients may allow for greater uptitration of therapies such as angiotensin-converting enzyme inhibitors and β blockers, drugs that are likely to be life-extending in many patients with AF.6, 25 Each of these potential mechanisms, however, remains speculative.
The current classification of “overweight” may be suboptimal. In the present study and in general population studies, there is little evidence that a BMI in the overweight range is associated with excess mortality.1, 2 With this in mind, and in the absence of definitive data regarding the impact of purposeful weight loss in patients with AF (or in other disease states exhibiting an obesity paradox), the approach to increased BMI remains uncertain. Therefore, aggressive modification of conventional risk factors and an increase in moderate physical activity should remain the focus of lifestyle interventions.
With respect to treatment of AF, we found that the ability to successfully perform electrical cardioversion and the long-term success of a rhythm-control strategy were similar irrespective of BMI category. In contrast, the likelihood of obtaining adequate heart rate control was substantially lower in obese patients. Importantly, however, we found similar outcomes for rate- and rhythm-control strategies in each BMI category. Therefore, our results suggest that the decision to pursue a rate- or rhythm-control strategy in an individual patient should be based on factors other than BMI.
This analysis has several potential limitations that require consideration. First, baseline data on BMI were missing for a considerable number of AFFIRM patients, and there was a modest imbalance in several baseline characteristics and in clinical event rates in this group. Because the distribution of BMI in those patients with missing data is unknown, the potential effect on our results is uncertain. Second, data on BMI were collected only at the baseline visit. We are therefore unable to explore the relation between changes in BMI over time and clinical outcomes. Third, despite making comprehensive adjustments for baseline differences between patients in each BMI category, residual confounding might remain. Fourth, although BMI is the most commonly used measurement of obesity, it does not directly distinguish between adipose and lean tissue or central and peripheral adiposity.6
References
- . General and abdominal adiposity and risk of death in Europe. N Engl J Med. 2008;359:2105–2120
- . Body mass index and mortality: a meta-analysis based on person-level data from twenty-six observational studies. Ann Epidemiol. 2005;15:87–97
- . Body weight and mortality among women. N Engl J Med. 1995;333:677–685
- . Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med. 1999;341:1097–1105
- . Association of bodyweight with total mortality and with cardiovascular events in coronary artery disease: a systematic review of cohort studies. Lancet. 2006;368:666–678
- . Body mass index and mortality in heart failure: a meta-analysis. Am Heart J. 2008;156:13–22
- . The obesity paradox: impact of obesity on the prevalence and prognosis of cardiovascular diseases. Postgrad Med. 2008;120:34–41
- . Reverse epidemiology of conventional cardiovascular risk factors in patients with chronic heart failure. J Am Coll Cardiol. 2004;43:1439–1444
- . Obesity and the risk of new-onset atrial fibrillation. JAMA. 2004;292:2471–2477
- . Atrial fibrillation and obesity—results of a meta-analysis. Am Heart J. 2008;155:310–315
- . A comparison of rate control and rhythm control in patients with atrial fibrillation. N Engl J Med. 2002;347:1825–1833
- . Atrial fibrillation follow-up investigation of rhythm management—the AFFIRM study design. Am J Cardiol. 1997;79:1198–1202
- Physical status: the use and interpretation of anthropometry (Report of a WHO Expert Committee). World Health Organ Tech Rep Ser. 1995;854:1–452
- . Metabolic syndrome and risk of development of atrial fibrillation: the Niigata preventive medicine study. Circulation. 2008;117:1255–1260
- . Obesity as a risk factor for the progression of paroxysmal to permanent atrial fibrillation: a longitudinal cohort study of 21 years. Eur Heart J. 2008;29:2227–2233
- . Body mass index, obstructive sleep apnea, and outcomes of catheter ablation of atrial fibrillation. J Cardiovasc Electrophysiol. 2008;19:668–672
- . Impact of obesity as a mortality predictor in high-risk patients with myocardial infarction or chronic heart failure: a pooled analysis of five registries. Eur Heart J. 2008;29:594–601
- . The impact of anthropomorphic indices on clinical outcomes in patients with acute ST-elevation myocardial infarction. Eur Heart J. 2007;28:415–424
- . An obesity paradox in acute heart failure: analysis of body mass index and in-hospital mortality for 108,927 patients in the Acute Decompensated Heart Failure National Registry. Am Heart J. 2007;153:74–81
- . The obesity paradox: body mass index and outcomes in patients with heart failure. Arch Intern Med. 2005;165:55–61
- . The obesity paradox in patients with peripheral arterial disease. Chest. 2008;134:925–930
- . Wasting as independent risk factor for mortality in chronic heart failure. Lancet. 1997;349:1050–1053
- . Obesity, cardiology, and beyond. J Am Coll Cardiol. 2008;52:986–987
- . Production of soluble tumor necrosis factor receptors by human subcutaneous adipose tissue in vivo. Am J Physiol Endocrinol Metab. 1999;277:E971–E975
- . Prognostic importance of weight loss in chronic heart failure and the effect of treatment with angiotensin-converting-enzyme inhibitors: an observational study. Lancet. 2003;361:1077–1083
The AFFIRM study was supported by the National Heart, Lung, and Blood Institute, Bethesda, Maryland. The present analysis was supported by the Cardiovascular Research Institute, Washington Hospital Center, Washington, DC.
PII: S0002-9149(10)00785-X
doi:10.1016/j.amjcard.2010.03.036
© 2010 Elsevier Inc. All rights reserved.
