Association between n−3 fatty acid consumption and ventricular ectopy after myocardial infarction

Abstract

Background: n−3 (omega-3) Fatty acids are associated with a reduced risk of cardiovascular disease; however, the relation between dietary intake of n−3 fatty acids and ventricular arrhythmias has not been investigated among acute post-myocardial infarction (AMI) patients—a group at elevated risk of malignant arrhythmias.

Objective: The objective was to examine the association between n−3 fatty acid consumption and ventricular ectopy among AMI patients.

Design: In 260 AMI patients, dietary intake of n−3 fatty acids was assessed by using the Harvard food-frequency questionnaire, and ventricular ectopy was estimated from 24-h electrocardiograph recordings.

Results: A greater intake of n−3 fatty acids (eicosapentaenoic acid + docosahexaenoic acid + docosapentaenoic acid + α-linolenic acid) was associated with lower ventricular ectopy (β = −0.35, P = 0.011), and this effect remained after cardiovascular comorbidities were controlled for (β = −0.47, P = 0.003). Higher concentrations of both marine-based (eicosapentaenoic acid + docosahexaenoic acid) (β = −0.21, P = 0.060) and plant-based (α-linolenic acid) (β = −0.33, P = 0.024) fatty acids remained associated with lower ventricular ectopy after cardiovascular comorbidities were controlled for.

Conclusion: These findings extend existing evidence linking n−3 fatty acid consumption to a reduced risk of ventricular arrhythmias by showing that a greater intake of n−3 fatty acids may be associated with low ventricular ectopy among AMI patients.

INTRODUCTION

Increasing evidence over the past 3 decades indicates that n−3 fatty acids may protect against the development of coronary heart disease (CHD) (1–4). In particular, the intake of the long-chain n−3 fatty acids found in fatty fish or fish-oil supplements, eicosapentaenoic acid (EPA; 20:5n−3), and docosahexaenoic acid (DHA; 22:6n−3) may provide CHD protection (2, 3, 5). Observational studies have shown that greater consumption of foods containing long-chain n−3 fatty acids is prospectively associated with improved CHD survival (2, 3) and a reduced risk of ventricular fibrillation (4) and sudden cardiac death (SCD) after myocardial infarction (MI) (6), although increased n−3 fatty acid consumption does not appear to consistently reduce the risk of nonfatal CHD (7), which suggests that the protective effects of long-chain n−3 fatty acids may be more strongly related to antiarrhythmic effects rather than to antiischemic effects.

Although previous studies have generally reported antiarrhythmic effects of long-chain n−3 fatty acids among cardiac patients (6, 8–10), not all studies have found protective effects (11, 12). Furthermore, few studies have provided direct evidence that plant-derived n−3 fatty acids, such as α-linolenic acid (ALA; 18:3n−3), are associated with reduced arrhythmic activity (13). Ventricular ectopy is common among patients with CHD (14) and is predictive of increased risk of cardiac mortality (15), particularly among post-MI patients (16). However, to our knowledge, no studies have examined the relation between n−3 fatty acid intake and ventricular ectopy in this population, despite the increased vulnerability to arrhythmia during the acute post-MI (AMI) period. We therefore examined the association of self-reported n−3 fatty acid consumption and ventricular ectopy among 260 patients evaluated within 72 h after acute MI.

SUBJECTS AND METHODS

Experimental design

A consecutive series of AMI patients were recruited during hospitalization at Duke University Medical Center (17). Patients in the current analysis were participating in an ongoing investigation of heart rate variability recovery after MI (the HARMONY Study). The study was approved by the Duke University Medical Center Institutional Review Board, and all patients gave written informed consent before participating in the research protocol.

Patient sample

MI was defined by serial changes in cardiac enzymes (eg, elevations in tropinin I or CKMB isoenzyme) >3 times the upper limit of normal and either electrocardiograph (ECG) changes or cardiac symptoms. AMI patients were identified within 72 h of MI and screened for enrollment from January 2004 to August 2007. Patients were excluded if they were too ill with cardiac complications, showed chronic arrhythmia or left ventricular ejection fraction (LVEF) <20%, or had a diagnosis of dementia, alcoholism, recent coronary artery bypass grafting, or significant communication barriers (Figure 1).

FIGURE 1.

Patient flow chart. FFQ, food-frequency questionnaire; ECG, electrocardiogram; LVEF, left ventricular ejection fraction; CABG, coronary artery bypass grafting.

Harvard food-frequency questionnaire

Dietary intake was estimated from an interviewer-administered semiquantitative food-frequency questionnaire (FFQ) during hospitalization after acute MI (18). This FFQ includes 131 food items with specified serving sizes corresponding to natural portions (eg, 1 apple or 2 slices of pizza) or standard weight and volume measures of the servings commonly consumed in this study population. For each food item, participants indicated their average frequency of consumption over the past year in terms of the specified serving size by choosing 1 of 9 frequency categories ranging from “almost never” to “≥6 times/d.” The selected frequency category for each food item was converted to a daily intake value. For example, a response of “2–4 servings /wk” was converted to 0.43 servings/d. The FFQ provides estimated daily nutrient intakes for n−3 fatty acids. The FFQ has been validated in multiple cohorts with the use of diet diaries and plasma biomarker concentrations (19).

Ambulatory ECG monitoring

Patients underwent 24-h Holter monitoring with a Lifecard CF, 3-channel digital recorder (Del Mar Reynolds, Irvine, CA) during or immediately after hospitalization for MI. The ambulatory ECG recordings were analyzed for isolated ventricular premature beats (VPBs), couplets, triplets, bigeminy, trigeminy, salvo, and ventricular tachycardia. The total number of VPBs over this 24-h period was determined by summing the number of VPBs from each of the above categories of ventricular arrhythmias. All Holter monitoring assessments had ≥21.9 h of data (mean: 23.9 ± 0.3 h).

Data reduction and analysis

General linear modeling equations (version 9.1.3; SAS Inc, Cary, NC) were used to examine the association between self-reported n−3 fatty acid intakes and total VPBs over the 24-h assessment period and to determine whether this association was moderated by any clinically relevant cardiac variables. Separate models were examined for total n−3 fatty acid consumption, marine-based n−3 fatty acids (EPA + DHA), and plant-based n−3 fatty acids (ALA). Age, LVEF, sex, coronary artery disease severity, ethnicity (white or nonwhite), history of prior MI (yes or no), body mass index, diabetes (presence or absence), HDL, LDL, and cardiac medications (β-blockers and antiarrhythmics; 1 = currently taking, 0 = not currently taking) were selected a priori as covariates in the model with n−3 fatty acid intake as the independent predictor of 24-h VPBs. To account for the potentially confounding effects of caloric intake, we examined the association between n−3 fatty acid consumption and ventricular ectopy using the multivariate nutrient density method, as outlined by Willett et al (20, 21). With this method, the intake of n−3 fatty acids was expressed per unit increase in caloric intake (in this case, 1000 kcal). This method has intuitive understanding to nutritionists, has been used in national dietary guidelines, and adjusts for total energy intake. In addition to this energy-adjusted nutrient density term, we included total caloric intake as a separate covariate because it was directly associated with 24-h ventricular ectopy (r = −0.17, P = 0.005), which can lead to confounding in multivariate analyses (21). Coronary artery disease severity was indexed by the number of coronary vessels with ≥75% stenosis. LVEF was truncated such that patients with an LVEF ≥ 60% were analyzed as having an LVEF of 60%. On the basis of an analysis of model residuals, we used the natural logarithm of the count of VPBs over the 24-h period as the response variable in the regression model.

Continuous predictors were divided by their interquartile range, which rescales their corresponding regression coefficients such that it can be interpreted as comparing a “typical” person in the middle of the upper half of the predictor distribution (ie, 75th percentile) with a “typical” person in the middle of the lower half of the predictor distribution (25th percentile). This rescaling preserves the continuous nature of the predictor but places the regression coefficient on a clinically meaningful scale. Before analysis, model assumptions of additivity, linearity, and distribution of residuals were evaluated and found to be adequate.

RESULTS

Patient characteristics

Participants ranged in age from 27 to 86 y (mean ± SD: 56.7 ± 11.3 y), were primarily male (62.9%) and white (72.6%), and had relatively preserved LV function (mean ± SD LVEF: 52.1 ± 10.8%) (Table 1). Most of the patients were taking β-blockers at the time of their assessment (88.5%), although few were taking antiarrhythmic medications (3.8%). Patients had expected cardiovascular comorbidities typical of AMI patients; 59% of patients reported a history of hypertension, 41% diabetes, 12% chronic obstructive pulmonary disorder, and 7% congestive heart failure. Seventy-three percent of patients reported a history of smoking, and 57% were current smokers. Sixteen percent had a history of MI, 17% had a previous coronary artery bypass grafting, and 15% had a history of percutaneous coronary intervention.

TABLE 1.

Demographic and clinical characteristics of the patient sample¹

		Tertile of total n−3 fatty acid consumption (EPA + DHA + DPA + ALA)
Variable	Overall	1	2	3
Age (y)	56.6 ± 11.3	57.4 ± 10.8	56.6 ± 9.9	56.1 ± 13.4
LVEF (%)	52.1 ± 10.8	51.1 ± 10.8	50.7 ± 11.0	53.0 ± 10.6
No. of vessels with significant stenosis	1.59 ± 0.87	1.62 ± 0.88	1.59 ± 0.91	1.56 ± 0.84
BMI (kg/m2)	29.9 ± 5.9	29.3 ± 5.3	29.6 ± 5.6	30.4 ± 6.4
Peak CKMB (ng/mL)	120.2 ± 133.1	122.4 ± 139.7	134.4 ± 147.3	114.1 ± 118.5
LDL cholesterol (mg/dL)	105.6 ± 39.8	106.4 ± 39.3	106.5 ± 41.0	104.6 ± 39.6
HDL cholesterol (mg/dL)	42.8 ± 11.9	43.2 ± 13.3	43.3 ± 12.6	42.0 ± 10.0
No. of daily VPBs	349.8 ± 930	446.1 ± 1090	385.0 ± 970	235.0 ± 716
Daily n−3 fatty acid consumption (g/1000 kcal)	0.76 ± 0.27	0.59 ± 0.19	0.72 ± 0.18	0.94 ± 0.31

¹All values are means ± SDs. CKMB, creatine kinase MB isoenzyme; LVEF, left ventricular ejection fraction; VPBs, ventricular premature beats; EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; ALA, α-linolenic acid.

Omega fatty acids and VPBs

Analysis of FFQ data showed that the patients’ diets generally contained modest amounts of n−3 fatty acids (mean ± SD daily n−3 consumption: 0.76 ± 0.27 g/1000 kcal) relative to prior studies (22). The mean (±SD) daily consumption of individual fatty acids was also modest: EPA + DHA (0.14 ± 0.14 g/1000 kcal) and ALA (0.60 ± 0.22 g/1000 kcal) (Table 1). Ventricular ectopy was relatively low, with substantial individual variation across the sample (median VPB: 22; interquartile range: 122).

A greater intake of n−3 fatty acids was associated with lower ventricular ectopy (β = −0.35, P = 0.011), and this effect remained after cardiovascular comorbidities were controlled for (β = −0.47, P = 0.003) (Table 2, Figure 2). Higher intakes of marine-based fatty acids (EPA + DHA) tended to be associated with reduced ventricular ectopy after adjustment for cardiovascular comorbidities (β = −0.21, P = 0.060). The association between plant-based fatty acids (ALA) remained significant after adjustment for cardiovascular comorbidities (β = −0.33, P = 0.024).

TABLE 2.

Association between total n−3 fatty acid intake (eicosapentaenoic acid + docosahexaenoic acid + docosapentaenoic acid + α-linolenic acid) and 24-h log-transformed ventricular premature beats¹

Predictor variable	Interquartile range of the predictor variable	Regression coefficient	SE	P value
Age	15	0.34	0.19	0.0984
LVEF	15	−0.69	0.18	<0.001
Male sex	—	0.68	0.30	0.025
Nonwhite ethnicity	—	−0.13	0.31	0.667
BMI	7.04	0.08	0.16	0.642
Occlusions	1	0.41	0.24	0.083
Previous MI	—	1.28	0.32	<0.001
Diabetes	—	−0.16	0.31	0.610
HDL cholesterol	15	0.41	0.18	0.021
LDL cholesterol	59	−0.37	0.20	0.060
Cardiac medication use	—	−0.52	0.40	0.202
Calorie intake	978	−0.29	0.13	0.029
Intake of n−3 fatty acids, g · d−1 · 1000 kcal−1	0.31	−0.47	0.16	0.003

¹The regression coefficients are from a multivariate model. LVEF, left ventricular ejection fraction; MI, myocardial infarction.

FIGURE 2.

n–3 Fatty acid consumption and log-transformed 24-h ventricular premature beats (VPBs), adjusted for age, left ventricular ejection fraction, sex, coronary artery disease severity, ethnicity, history of prior myocardial infarction, BMI, diabetes, HDL cholesterol, LDL cholesterol, cardiac medication use, and total caloric intake (n = 260).

DISCUSSION

We found that greater self-reported intakes of omega-3 (n−3) fatty acids was associated with reduced ventricular ectopy after MI. Greater intake of n−3 fatty acids was associated with fewer VPBs during 24-h Holter-monitoring after acute MI, with every 1.0-g/d increase in total n−3 fatty acid intake associated with a reduction in VPBs of ≈2850/d. Similarly, a 1.0-g/d increase in EPA + DHA was associated with a reduction in VPBs of 800/d. Although a 1.0-g/d increase may represent a substantial increase in the present sample, supplementation with 1.0 g EPA + DHA/d has been shown to improve survival in previous trials (6). These findings extend previous research showing a protective effect of n−3 fatty acids after acute MI by showing that n−3 fatty acids may have a protective, antiarrhythmic effect in the high-risk period after acute MI.

n−3 Fatty acids, derived primarily from fish, have been shown in many studies to have a protective effect against CHD (1, 3, 7). In a meta-analytic study of prospective cohort studies incorporating data from >222,000 individuals, He et al (2) found that higher fish intakes were associated with a reduced risk of CHD mortality and that every 20-g/d increase was associated with a 7% lower risk of CHD mortality. Burr et al (23) showed a similar effect among AMI patients, ie, that diets higher in fatty fish were associated with a 29% reduction in 2-y all-cause mortality in the Diet and Reinfarction Trial (DART). Several meta-analytic studies have reported similar findings (3, 24, 25). In a meta-analysis of randomized controlled trials among cardiac patients, Bucher et al (24) found that both dietary and nondietary intakes of n−3 fatty acids were associated with reduced cardiac and all-cause mortality as well as sudden death. Leon et al (25), in a meta-analysis of randomized controlled trials using EPA and DHA as dietary supplements, found that n−3 fatty acids were associated with reductions in cardiac mortality, but did not have an effect on arrhythmias or all-cause mortality. Similarly, a recent meta-analysis of controlled trials showed that n−3 consumption was not associated with reduced fatal or nonfatal CHD events (26).

Our findings are consistent with multiple prospective studies and randomized controlled trials that have investigated the antiarrhythmic effects of n−3 fatty acids among cardiac patients (13, 27). Albert et al (28), in a prospective case-control study, found that higher blood concentrations of n−3 fatty acids were associated with a reduced risk of SCD during a 17-y follow-up among men in the Physicians’ Health Study. Aarsetoy et al (4) examined this association among 460 acute coronary syndrome patients, finding that those patients with lower serum n−3 concentrations were more likely to exhibit ventricular fibrillation than were patients who did not experience arrhythmic complications. Marchioli et al (6), in a randomized controlled trial of 11,323 patients, found that 3 mo of n−3 supplementation was associated with substantial reductions in SCD, and this effect was observable as early as 4 mo after initiation of treatment. Leaf et al (29) also found modest support for a protective effect of n−3 fatty acids among 402 high-risk patients with implanted cardioverter/defibrillator (ICD), showing that fish-oil supplementation prolonged the time to the first arrhythmic event. Similarly, Sellmayer et al (10) found that n−3 supplementation reduced VPBs among 79 patients with suspected ischemic heart disease and moderate ventricular ectopy at baseline.

Not all studies have reported a protective effect of n−3 fatty acids, however. Several controlled trials with smaller samples have shown that n−3 supplementation may not reduce VPBs in cardiac patients (30, 31). More recently, Brouwer et al (32) found no protective effect of n−3 fatty acids in a double-blind multicenter trial of 546 ICD patients. Although it is unclear why findings have varied to such a degree between studies, drug-nutrient interactions and/or variations in the comorbidities of the participants may partially explain the observed pattern of results (33). The discrepancy between our positive findings and the negative results reported among ICD patients may be partially explained by the superior relative health of our sample because of specific exclusion criteria requiring greater patient mobility, small sample sizes, and the low statistical power of these previous trials.

ALA has been prospectively associated with a reduced risk of CHD development in many studies (34–38) and has been shown to have antiarrhythmic properties in animal studies (39). In a systematic review of clinical trials, Harper and Jacobson (7) found that interventions using ALA supplements or ALA-enriched diets showed possible benefits on SCD and nonfatal infarction, although many of the existing studies at this time were limited by design flaws and reduced statistical power. Brouwer et al (34) reported similar findings, ie, that ALA consumption may be associated with a reduced risk of fatal heart disease but an increased risk of prostate cancer. More recent systematic reviews have failed to support these findings, however, showing that ALA supplementation does not reduce cardiac and sudden death (40). At present, the potentially antiarrhythmic effects of ALA are largely supported by epidemiologic observational studies and have yet to be supported in a large-scale randomized controlled trial (13).

n−3 Fatty acids may be associated with reduced ventricular ectopy by modulation of calcium ion fluxes (41, 42), stabilization of cardiac myocytes (43), reduced platelet adhesion and reactivity (44, 45), improved endothelial function (46), and altered metabolism of adhesion molecules (47). Intravenous administration of fish-oil extract, DHA, EPA, and ALA have all been shown to effectively reduce the ventricular fibrillation threshold in canine models of exercise-induced adrenergic stress and myocardial ischemia (48, 49). However, the short-term benefits of n−3 administration remain unclear in humans.

The present study had several limitations. First, because our study was cross-sectional, the causal relation between n−3 fatty acid consumption and ventricular ectopy could not be determined. It is possible that individuals who experienced more symptomatic ectopic beats were advised to alter their diets to include greater amounts of n−3 fatty acid–rich foods or decided to do so of their own accord. However, ancillary analyses in which we controlled for any variables associated with previous cardiac medical attention (eg, previous coronary artery bypass grafting) did not alter the n−3 and ventricular ectopy association, which indicated that decreased ectopy was more likely a result of improved diet and not the cause thereof. Second, our measure of n−3 fatty acid consumption was based on self-reported dietary intake. Although the Harvard FFQ has been widely validated using other more objective measures of dietary intake (18, 50), it is unclear whether self-report was influenced by the hospital setting or other sources of reporting bias. Furthermore, self-reported dietary intake is an inherent source of error, and it is unclear how this error may have influenced the observed association between n−3 fatty acid intake and VPBs. Third, because of our study criteria, which excluded individuals with compromised LVEF, our results may only be generalized to AMI patients with relatively preserved LV function. Although frequent VPBs are independently associated with SCD among patients with a history of MI (51–53), VPBs in persons with preserved LV function may be less predictive of clinical events. Fourth, the present analysis was conducted during the initial days after acute MI and did not incorporate data from other time points. It is unclear from the present analysis whether n−3 fatty acid levels would be associated with reduced ventricular ectopy at later time points after MI. Similarly, given the modest levels of both n−3 fatty acids and ventricular ectopy in our sample, it is unclear how the magnitude of this relation may differ among other cardiac samples. A final limitation inherent to our observational methodology is the potential for unmeasured variables to have influenced the observed pattern of results.

Results of the present analysis indicate that a greater intake of n−3 fatty acids is associated with reduced ventricular ectopy after MI. The association between n−3 intake and VPBs was approximately half the strength of that observed with coronary occlusions, which indicated a clinically meaningful magnitude. These results could have important implications for the management of AMI patients, who are at increased risk of malignant ventricular arrhythmias and SCD during the early post-MI period. Future randomized controlled trials should investigate whether the administration of fish-oil supplements during hospitalization for AMI is associated with reduced SCD during this time period. Furthermore, future studies would benefit from more extensive ECG follow-up assessments that examine patients prospectively after AMI and incorporate longer follow-up time-points.