Effect of Fish Oil on Heart Rate in Humans
http://www.100md.com
《循环学杂志》
the Channing Laboratory (D.M.), Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, and the Departments of Epidemiology and Nutrition (D.M.), Harvard School of Public Health, Boston, Mass, and Division of Human Nutrition (A.G., I.B., J.G., P.Z., M.K.), Wageningen University, and Wageningen Centre for Food Sciences (A.G., I.B., P.Z., M.K.), Wageningen, the Netherlands.
Abstract
Background— The effect of fish oil on heart rate (HR), a major risk factor for sudden death, is not well established. We calculated this effect in a meta-analysis of randomized, double-blind, placebo-controlled trials in humans.
Methods and Results— Randomized trials of fish oil that evaluated HR were identified through MEDLINE (1966 through January 2005), hand-searching of references, and contact with investigators for unpublished results. Two investigators independently extracted trial data. A pooled estimate was calculated from random-effects meta-analysis. Predefined stratified meta-analyses and meta-regression were used to explore potential heterogeneity. Of 197 identified articles, 30 met inclusion criteria. Evidence for publication bias was not present. In the overall pooled estimate, fish oil decreased HR by 1.6 bpm (95% CI, 0.6 to 2.5; P=0.002) compared with placebo. Between-trial heterogeneity was evident (Q test, P<0.001). Fish oil reduced HR by 2.5 bpm (P<0.001) in trials with baseline HR 69 bpm (median) but had little effect (0.04-bpm reduction; P=0.56) in trials with baseline HR <69 bpm (P for interaction=0.03). Fish oil reduced HR by 2.5 bpm (P<0.001) in trials with duration 12 weeks but had less effect (0.7-bpm reduction; P=0.27) in trials with duration <12 weeks (P for interaction=0.07). HR reduction with fish oil intake did not significantly vary by fish oil dose (range, 0.81 to 15 g/d), type of HR measure, population age, population health, parallel versus crossover design, type of control oil, or study quality by Delphi criteria (P for interaction >0.25 for each).
Conclusions— In randomized controlled trials in humans, fish oil reduces HR, particularly in those with higher baseline HR or longer treatment duration. These findings provide firm evidence that fish oil consumption directly or indirectly affects cardiac electrophysiology in humans. Potential mechanisms such as effects on the sinus node, ventricular efficiency, or autonomic function deserve further investigation.
Key Words: heart rate fatty acids, omega-3 fish oil meta-analysis randomized controlled trials
Introduction
Fatty fish and fish oil intake is associated with lower risk of cardiac arrhythmias, including sudden death, arrhythmic coronary heart disease death, and atrial fibrillation.1–8 Experimental studies in isolated rat myocytes, exercising dogs, and nonhuman primates suggest that fish oil has direct cardiac electrophysiological effects, including slowing of the heart rate (HR).9–11 However, such effects are not well established in humans. Because higher HR is a major independent risk factor for cardiovascular death, particularly sudden death,12–18 an effect of fish oil on HR would both confirm an influence on cardiac electrophysiology in humans and indicate a plausible potential mechanism for observed relations between fish intake and arrhythmic events. We therefore performed a meta-analysis of randomized placebo-controlled clinical trials to determine the effect of fish oil consumption on HR in humans.
Methods
Selection of Randomized Trials
We followed the Quality of Reporting of Meta-Analyses (QUOROM) standards19 during all phases of the design and implementation of this analysis. Randomized clinical trials of fish oil that included evaluation of HR were identified through MEDLINE (1966 through February 2005), including fish oil trials designed primarily to evaluate other outcomes such as blood pressure or coronary restenosis, hand-searching of reference lists of obtained articles, and contacting investigators for unreported HR data in published trials or for HR data from unpublished trials. To minimize publication bias, we attempted to identify all fish oil trials that may have measured and reported HR data, rather than limiting our search to trials designed primarily to evaluate HR. English-language trials in human subjects >18 years of age were included if oral fish oil supplementation was randomized and changes in HR or baseline and follow-up HR were measured; trials with organ transplant subjects, cointerventions that could not be separated from fish oil treatment, no placebo control, nonblinding of participants, or duration <2 weeks were excluded.
Trial Review
When potentially relevant trials were identified, abstracts and, if necessary, original articles were screened for obvious exclusions by an investigator. Of 197 identified trials, 161 were excluded for not being a randomized trial of fish oil (n=28), for having no available HR data (n=75), for occurring in organ transplant recipients (n=12), for having no placebo control (n=29), for having a cointervention that could not be separated from fish oil treatment (n=6), or for being a duplicate publication from the same study (n=11). The identified trials included 10 published and 2 unpublished trials for which we contacted the authors to determine whether unreported HR data might be available. A list of all reviewed trials and reasons for exclusion is available by request from the investigators. For the remaining 36 trials not excluded during initial screening, each original article was independently reviewed for inclusion by 2 investigators. Six of these trials were excluded for no placebo control (n=2), no follow-up HR data (n=2), duration <2 weeks (n=1), or being a duplicate publication from the same study (n=1). Thirty trials met inclusion and exclusion criteria, including 2 trials for which unpublished HR data were obtained from the authors (personal communications, William Harris, February 18, 2005, and Ingrid Toft, March 4, 2005).20–49 Concordance on inclusion and exclusion decisions was 100%.
Data Extraction
For each of the articles meeting inclusion and exclusion criteria, data were independently extracted by 2 investigators on study design; population; sample size and dropout; fish oil type, dose, and duration; method of HR assessment; change in HR or baseline and follow-up HR values; and HR variance measures. For studies reporting RR interval values (duration of 1 heartbeat in milliseconds), HR was calculated and its corresponding variance was estimated proportionally to the RR interval variance. Study quality was also independently assessed by 2 investigators according to the criteria for quality assessment of randomized clinical trials developed by Delphi consensus.50 The 9 criteria (1a, 1b, and 2 through 8) include, for example, whether a method of randomization was performed, whether the treatment groups were similar at baseline with regard to the most important prognostic indicators, and whether the analysis was of intention-to-treat design. For the last criterion, we considered analyses as having intention-to-treat design if all subjects not lost to follow-up were analyzed according to their original randomization group; exclusions were not made for noncompliance. We assessed the validity of data extraction by comparing the independently abstracted results for concordance, and any discrepancies were resolved by discussion and review of the original manuscript by the 2 investigators who extracted the data or, if necessary, a committee comprising all the investigators. When necessary, missing information (type of control oil, mean age of participants, etc) was obtained by direct contact with the original authors. We attempted to minimize clinical heterogeneity by excluding studies in children, in organ transplant recipients, or with duration <2 weeks. Remaining clinical heterogeneity was evaluated qualitatively by comparing the mean age, gender distribution, and general health of the study populations; the doses and durations of fish oil treatment; and the methods of HR assessment. Clinical heterogeneity was assessed quantitatively in prespecified stratified analyses (see Statistical Analysis).
Statistical Analysis
Our primary outcome was the change in HR resulting from fish oil treatment. For parallel-design trials, the HR change from baseline to study end in the control group was subtracted from the HR change from baseline to study end in the treatment group. For crossover design trials, the HR at the end of the control period was subtracted from the HR at the end of the treatment period. Within-individual changes were used when available; otherwise, group means were used. SEs were abstracted or, if not reported, derived from SDs, CIs, or probability values. The pooled variance for the net HR change resulting from fish oil treatment was calculated as (1) SE2T+SE2C–2(r)(SET)(SEC) for crossover design trials, where SET and SEC are the SE of the treatment and control period HR values, respectively, and r is the within-individual correlation between the treatment and control period HR values, and (2) SE2TG+SE2CG for parallel-design trials, where SETG and SECG are the SE of the HR change from baseline to study end in the treatment and control groups, respectively. For parallel-design trials that reported precision of baseline and final HR values (n=18) rather than HR changes, SETG and SECG were calculated according to the method of Follmann et al,51 which involves making an assumption for the unreported within-individual correlation between baseline and final HR values. On the basis of measured correlations in fish oil trials (Anouk Geelen, personal communication, January 27, 2005), the within-individual correlation between HR values was estimated to be 0.60 for trials using a single HR measure, 0.80 for trials using the average of multiple measures, and 0.85 for trials using a 24-hour measure, with the higher correlations consistent with less random error in the HR measurement. Sensitivity analyses were performed assuming a within-individual correlation of 0.60 for all trials. Data for the calculation of the change in HR and the variance of this change were not missing from any trial.
Pooled estimates of the effect of fish oil on HR were calculated through the use of random-effects meta-analysis, which accounts for heterogeneity in treatment effects among trials, using the method of DerSimonian and Laird52 with inverse-variance (SE) weighting. Because some trials compared multiple intervention groups with a single control group (n=7), we performed sensitivity analyses in which separate pooled estimates and variances for the effect of fish oil on HR were calculated using separate meta-analyses for each of these trials; these trial-specific estimates then were used in a second meta-analysis evaluating all trials. Heterogeneity between studies was tested with the DerSimonian and Laird Q statistic.52,53 To assess publication bias, a funnel plot of the treatment effect versus SE was visually inspected.54 Potential publication bias was also evaluated with the Begg adjusted-rank correlation test,55 a statistical analog of the visual funnel graph, and the regression asymmetry test according to the method of Egger et al.54
We performed predefined stratified meta-analyses to explore potential heterogeneity by dose of eicosapentaenoic acid and docasohexaenoic acid (EPA+DHA) (at the median), duration of treatment (12 weeks versus less), type of HR measure (single measure, average of multiple resting measures, or 24-hour measure), baseline HR (at the median), type of control oil (olive oil versus other), population age (at the median), general health (healthy versus otherwise), study design (parallel versus crossover), and study quality (meeting at least 8 Delphi criteria versus fewer). We used meta-regression to test for heterogeneity of the pooled treatment effect by these factors,56 testing for significance of the stratifying variable by using the Wald test in a mixed-effects meta-regression model. We also performed sensitivity analyses excluding trials with 20% dropout of randomized participants at baseline. All analyses were performed with Stata version 8.2 (Stata Corp). Statistical significance was defined as 2-tailed <0.05.
Results
Overview of Trials
Of the 30 trials meeting inclusion and exclusion criteria, 6 had 2 separate intervention groups, and 1 had 3 separate intervention groups, for a total of 38 intervention groups in the 30 trials (Table 1). Although single-blind trials were acceptable, all were double-blind trials. Eight were crossover design trials, and 22 were parallel-design trials. Median study size was 30 participants; in total, this meta-analysis included 1678 individuals treated with fish oil or placebo for 27 615 person-weeks. The mean ages of the study populations ranged from 23 to 68 years (median, 54 years). Sixteen intervention groups were made up of generally healthy populations; 22 comprised individuals with 1 underlying chronic condition. The median EPA+DHA dose was 3.5 g/d (range, 0.81 to 15 g/d), and the median treatment duration was 8 weeks (range, 4 to 52 weeks). Thirteen intervention groups assessed HR with a single resting measure; 14 used the average of 2 or 3 resting measures; and 11 used the average of ambulatory or continuous monitoring. Twenty-five trials (30 intervention groups) met at least 8 Delphi criteria for study quality; 5 trials (8 intervention groups) met <8.
Our broad search methods appeared to be successful in minimizing the effect of publication bias. Among the 30 included trials, 12 reported HR findings in the abstract (7 reporting an effect, 5 reporting the absence of an effect); 10 reported HR findings in the results text but not the abstract (5 reporting an effect, 5 reporting the absence of an effect); 6 presented HR findings in a table only (all 6 showing no significant effect); and 2 constituted unpublished results. Little evidence for publication bias was present by visual inspection of a funnel plot (Figure 1), Begg’s test (P=0.87), or Egger’s test (P=0.69).
Effect of Fish Oil on HR
The individual trial results and the pooled estimate are presented in Figure 2. In the overall pooled estimate, fish oil decreased HR by 1.6 bpm (95% CI, 0.6 to 2.5; P=0.002) compared with placebo. Exclusion of trials with 20% dropout (n=5) had little effect on the pooled estimate, with fish oil decreasing HR by 1.3 bpm (95% CI, 0.3 to 2.4; P=0.009). Assuming a within-individual HR correlation of 0.60 for all trials also had little effect, with fish oil decreasing HR by 1.5 bpm (95% CI, 0.5 to 2.5; P=0.003). The pooled estimate was also similar in sensitivity analyses accounting for multiple intervention groups in some trials, with fish oil decreasing HR by 1.4 bpm (95% CI, 0.4 to 2.5; P=0.007).
Between-trial heterogeneity was evident (Q test, P<0.001). We evaluated prespecified study characteristics to explore reasons for potential heterogeneity (Table 2). The HR reduction with fish oil consumption was greater in study populations with a mean baseline HR 69 bpm (P for interaction=0.03), among whom fish oil reduced HR by 2.5 bpm (95% CI, 1.4 to 3.5; P<0.001), and in study populations receiving 12 weeks of fish oil treatment (P for interaction=0.07), among whom fish oil reduced HR by 2.5 bpm (95% CI, 1.1 to 4.0; P=0.001). Although other differences related to study characteristics were not statistically significant (Table 2), several findings were consistent with intuition; eg, the effect of fish oil on HR appeared possibly greater with increasing precision of the measurement method used (single versus average of 2 or 3 measures versus ambulatory/continuous), consistent with reduced measurement error reducing bias toward the null.
Little evidence was present for a dose-response effect. Stratified at the median dose of fish oil (3.5 g/d), the reduction in HR was not significantly different at higher versus lower doses (each compared with placebo) (P for interaction=0.72) (Table 2). Similarly, stratified into quartiles of fish oil dose, HR was reduced by 1.1 (95% CI, –0.9 to 3.1), 1.8 (95% CI, –0.1 to 3.6), 1.9 (95% CI, 0.1 to 3.8), and 1.5 (95% CI, –0.6 to 3.6) bpm in quartiles 1 through 4, respectively, compared with placebo (P for ordinal interaction=0.72). Evaluated continuously, the dose of fish oil was not a predictor of treatment effect (P=0.63), above and beyond being on fish oil treatment (yes/no). In the 2 trials with EPA+DHA doses 1 g/d, HR was reduced by 5.0 bpm (95% CI, 2.3 to 7.7; P<0.001) compared with 1.4 bpm in the trials with EPA+DHA doses >1 g/d (95% CI, 0.4 to 2.3; P<0.001).
When we evaluated different factors simultaneously in the meta-regression model, there appeared to be potential independent heterogeneity related to both baseline HR (P for interaction=0.04) and treatment duration (P for interaction=0.09). Among the 9 trials with mean baseline HR 69 bpm and treatment duration 12 weeks, fish oil reduced HR by 2.9 bpm (95% CI, 1.5 to 4.4; P<0.001) compared with placebo, without significant between-trial heterogeneity (Q test, P>0.05).
Discussion
In this meta-analysis of randomized, double-blind, placebo-controlled clinical trials, fish oil consumption reduced HR in humans. Although the overall effect was modest (1.6-bpm reduction), on a population level, even modest differences in risk factors can have a significant impact on health. These findings provide firm evidence for an effect of fish oil consumption on cardiac electrophysiology in humans.
The regulation of HR is a complex physiological process, with components related to vagal tone, sympathetic input, responsiveness of the sinus node, and systolic and diastolic left ventricular function. The decrease in HR with fish oil consumption indicates that marine n-3 fatty acids influence at least 1 of these parameters. The n-3 fatty acids are incorporated into myocyte membranes and may influence ion channel function9,10; this could directly alter the automaticity or responsiveness of the sinus node. Fish oil also lowers blood pressure in humans,57 possibly by reducing systemic vascular resistance.58 In one observational study, such an effect was apparent at dietary levels of fish intake.58 Such a decrease in systemic vascular resistance would reduce left ventricular afterload and improve diastolic function, which could indirectly reduce HR as a result of better ventricular efficiency. Experimental studies in nonhuman primates support the hypothesis that fish oil consumption improves left ventricular efficiency.59,60 Intake of n-3 fatty acid may also improve measures of HR variability,21–23,27 suggesting a potential effect on autonomic tone. Our findings substantiate an electrophysiological effect of fish oil in humans and support the need for further investigation of these potential mechanisms.
Higher HR is associated with increased cardiovascular risk, including greater risk of sudden death,12–15,17 coronary heart disease death,13,14 and cardiovascular death.16 A higher HR could directly increase cardiovascular risk, eg, by increasing myocardial vulnerability to ischemia or arrhythmia. On the basis of work by Jouven et al,17 our finding of a 1.6-bpm HR reduction with fish oil consumption would correspond to an 5% lower risk of sudden death. Thus, in addition to effects on HR, other mechanisms are likely to contribute to the reductions in sudden death risk with fish or fish oil consumption seen in observational studies and randomized trials. A higher HR may indicate less optimal underlying cardiovascular health as manifested by increased sympathetic tone, decreased vagal tone, or decreased ventricular efficiency. The HR reduction with fish oil consumption could therefore indicate beneficial effects of fish oil on these other physiological parameters that might reduce cardiovascular risk to a greater extent than that resulting from the change in HR alone.
Our exploration of heterogeneity revealed several interesting findings. First, the reduction in HR appeared larger in trials with longer duration of intake (12 weeks). This may relate in part to the time needed for EPA and DHA to be incorporated into the tissues where they exert their effects and suggests that regular consumption over time may have greater effects than short-term intake. Second, HR was reduced to a greater extent in populations with higher baseline HR. Because fish oil was compared with placebo in each trial, this result would not be due to regression toward the mean. This finding suggests that fish oil may have greater effects on HR in populations with higher intrinsic sinus node automaticity, greater sympathetic tone, lower vagal tone, or lower ventricular efficiency. Third, although power was insufficient to prove equivalence of different doses, very high consumption of fish oil did not appear to have substantially greater effects than modest consumption. This is consistent with observational studies and randomized trials indicating clinical benefits of fatty fish or fish oil consumption at relatively modest intake, 1 to 2 servings per week or 500 to 1000 mg/d EPA+DHA, respectively.1–8 In the present meta-analysis, the lowest EPA+DHA doses were 1 g/d, and it is possible that a dose-response effect may exist at lower (eg, dietary) levels of intake, as suggested by one observational analysis.58 Finally, although the differences were not statistically significant, the HR reduction was smaller in trials using a single resting measure of HR, intermediate in trials using the average of 2 or 3 resting measures, and greatest in trials using ambulatory or continuous measures. This is consistent with a greater degree of misclassification (random measurement error) when only a single or a few resting measures were used, suggesting that such trials may underestimate the true effect of fish oil on HR. Alternatively, the results of trials using ambulatory and continuous monitoring represent the averaged effect of fish oil consumption on both resting and activity-related HR responses, which may be somewhat greater than effects on resting HR alone.
Publication bias is a major potential limitation of meta-analyses. Our broad, prespecified search methods and contacting of investigators for unpublished results appeared to be successful in minimizing the effect of publication bias; in only a minority of included trials was a significant HR effect prominently reported, and little evidence was present for publication bias in the final included studies. Additionally, given the large number of included trials, it is unlikely that the results of even several additional studies would greatly alter the pooled estimate. Between-trial heterogeneity may limit the generalizability of the overall pooled estimate; we attempted to account for potential heterogeneity by using a random-effects model and by assessing factors that may explain between-trial differences.
In this meta-analysis of randomized, double-blind, placebo-controlled clinical trials, fish oil reduced HR, particularly with higher baseline HR or longer durations of treatment. These results provide strong evidence that fish oil consumption directly or indirectly influences cardiac electrophysiology in humans. This effect may directly account for part of the observed benefits of fish intake on cardiovascular risk, particularly risk of arrhythmic events, and may indicate favorable effects on physiological systems such as on autonomic tone, vascular resistance, or ventricular efficiency that improve cardiovascular health.
Acknowledgments
The Wageningen Centre for Food Sciences is an alliance of major Dutch food industries, Maastricht University, TNO Nutrition and Food Research in Zeist, and Wageningen University and Research Centre, with financial support by the Dutch Government. Dr Mozaffarian was supported by a Mentored Clinical Scientist Award from the National Heart, Lung, and Blood Institute, National Institutes of Health (K08-HL-075628) and thanks Drs Eric Rimm, David Siscovick, and David Herrington for their invaluable guidance and support. The authors thank Drs William Harris and Ingrid Toft for sharing unpublished results for this analysis.
Footnotes
Guest Editor for this article was Robert H. Eckel, MD.
Medline search criteria included (heart rate or blood pressure or restenosis) and (fish oil or n-3 fatty acids or omega-3 or eicosapentaenoic or docosahexaenoic). Limits were adults 19 years of age, English language, clinical trial, and humans.
References
Burr ML, Fehily AM, Gilbert JF, Rogers S, Holliday RM, Sweetnam PM, Elwood PC, Deadman NM. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: Diet and Reinfarction Trial. Lancet. 1989; 2: 757–761.
Siscovick DS, Raghunathan TE, King I, Weinmann S, Wicklund KG, Albright J, Bovbjerg V, Arbogast P, Smith H, Kushi LH, et al. Dietary intake and cell membrane levels of long-chain n-3 polyunsaturated fatty acids and risk of primary cardiac arrest. JAMA. 1995; 274: 1363–1367.
Albert CM, Hennekens CH, O’Donnell CJ, Ajani UA, Carey VJ, Willett WC, Ruskin JN, Manson JE. Fish consumption and risk of sudden cardiac death. JAMA. 1998; 279: 23–28.
Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto miocardico. Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: GISSI-Prevenzione trial. Lancet. 1999; 354: 447–455.
Hu FB, Bronner L, Willett WC, Stampfer MJ, Rexrode KM, Albert CM, Hunter D, Manson JE. Fish and omega-3 fatty acid intake and risk of coronary heart disease in women. JAMA. 2002; 287: 1815–1821.
Mozaffarian D, Lemaitre RN, Kuller LH, Burke GL, Tracy RP, Siscovick DS. Cardiac benefits of fish consumption may depend on the type of fish meal consumed: the Cardiovascular Health Study. Circulation. 2003; 107: 1372–1377.
Mozaffarian D, Ascherio A, Hu FB, Stampfer MJ, Willett WC, Siscovick DS, Rimm EB. Interplay between different polyunsaturated fatty acids and risk of coronary heart disease in men. Circulation. 2005; 111: 157–164.
Mozaffarian D, Psaty BM, Rimm EB, Lemaitre RN, Burke GL, Lyles MF, Lefkowitz D, Siscovick DS. Fish intake and risk of incident atrial fibrillation. Circulation. 2004; 110: 368–373.
McLennan PL Myocardial membrane fatty acids and the antiarrhythmic actions of dietary fish oil in animal models. Lipids. 2001; 36: 111S–114S.
Leaf A, Kang JX, Xiao YF, Billman GE. Clinical prevention of sudden cardiac death by n-3 polyunsaturated fatty acids and mechanism of prevention of arrhythmias by n-3 fish oils. Circulation. 2003; 107: 2646–2652.
Billman GE, Kang JX, Leaf A. Prevention of ischemia-induced cardiac sudden death by n-3 polyunsaturated fatty acids in dogs. Lipids. 1997; 32: 1161–1168.
Dyer AR, Persky V, Stamler J, Paul O, Shekelle RB, Berkson DM, Lepper M, Schoenberger JA, Lindberg HA. Heart rate as a prognostic factor for coronary heart disease and mortality: findings in three Chicago epidemiologic studies. Am J Epidemiol. 1980; 112: 736–749.
Kannel WB, Kannel C, Paffenbarger RS Jr, Cupples LA. Heart rate and cardiovascular mortality: the Framingham study. Am Heart J. 1987; 113: 1489–1494.
Shaper AG, Wannamethee G, Macfarlane PW, Walker M. Heart rate, ischaemic heart disease, and sudden cardiac death in middle-aged British men. Br Heart J. 1993; 70: 49–55.
Wannamethee G, Shaper AG, Macfarlane PW, Walker M. Risk factors for sudden cardiac death in middle-aged British men. Circulation. 1995; 91: 1749–1756.
Palatini P, Casiglia E, Julius S, Pessina AC. High heart rate: a risk factor for cardiovascular death in elderly men. Arch Intern Med. 1999; 159: 585–592.
Jouven X, Zureik M, Desnos M, Guerot C, Ducimetiere P. Resting heart rate as a predictive risk factor for sudden death in middle-aged men. Cardiovasc Res. 2001; 50: 373–378.
Palatini P, Thijs L, Staessen JA, Fagard RH, Bulpitt CJ, Clement DL, de Leeuw PW, Jaaskivi M, Leonetti G, Nachev C, O’Brien ET, Parati G, Rodicio JL, Roman E, Sarti C, Tuomilehto J, for the Systolic Hypertension in Europe (Syst-Eur) Trial Investigators. Predictive value of clinic and ambulatory heart rate for mortality in elderly subjects with systolic hypertension. Arch Intern Med. 2002; 162: 2313–2321.
Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement: Quality of Reporting of Meta-Analyses. Lancet. 1999; 354: 1896–1900.
Bairati I, Roy L, Meyer F. Effects of a fish oil supplement on blood pressure and serum lipids in patients treated for coronary artery disease. Can J Cardiol. 1992; 8: 41–46.
Christensen JH, Christensen MS, Dyerberg J, Schmidt EB. Heart rate variability and fatty acid content of blood cell membranes: a dose-response study with n-3 fatty acids. Am J Clin Nutr. 1999; 70: 331–337.
Christensen JH, Aaroe J, Knudsen N, Dideriksen K, Kornerup HJ, Dyerberg J, Schmidt EB. Heart rate variability and n-3 fatty acids in patients with chronic renal failure: a pilot study. Clin Nephrol. 1998; 49: 102–106.
Christensen JH, Gustenhoff P, Korup E, Aaroe J, Toft E, Moller J, Rasmussen K, Dyerberg J, Schmidt EB. Effect of fish oil on heart rate variability in survivors of myocardial infarction: a double blind randomised controlled trial. BMJ. 1996; 312: 677–678.
Conquer JA, Holub BJ. Effect of supplementation with different doses of DHA on the levels of circulating DHA as non-esterified fatty acid in subjects of Asian Indian background. J Lipid Res. 1998; 39: 286–292.
Deslypere JP. Influence of supplementation with N-3 fatty acids on different coronary risk factors in men: a placebo controlled study. Verh K Acad Geneeskd Belg. 1992; 54: 189–216.
Dyerberg J, Eskesen DC, Andersen PW, Astrup A, Buemann B, Christensen JH, Clausen P, Rasmussen BF, Schmidt EB, Tholstrup T, Toft E, Toubro S, Stender S. Effects of trans- and n-3 unsaturated fatty acids on cardiovascular risk markers in healthy males: an 8 weeks dietary intervention study. Eur J Clin Nutr. 2004; 58: 1062–1070.
Geelen A, Zock PL, Swenne CA, Brouwer IA, Schouten EG, Katan MB. Effect of n-3 fatty acids on heart rate variability and baroreflex sensitivity in middle-aged subjects. Am Heart J. 2003; 146: E4.
Geelen A, Brouwer IA, Schouten EG, Maan AC, Katan MB, Zock PL. Effects of n-3 fatty acids from fish on premature ventricular complexes and heart rate in humans. Am J Clin Nutr. 2005; 81: 416–420.
Gray DR, Gozzip CG, Eastham JH, Kashyap ML. Fish oil as an adjuvant in the treatment of hypertension. Pharmacotherapy. 1996; 16: 295–300.
Grimsgaard S, Bonaa KH, Hansen JB, Myhre ES. Effects of highly purified eicosapentaenoic acid and docosahexaenoic acid on hemodynamics in humans. Am J Clin Nutr. 1998; 68: 52–59.
Landmark K, Thaulow E, Hysing J, Mundal HH, Eritsland J, Hjermann I. Effects of fish oil, nifedipine and their combination on blood pressure and lipids in primary hypertension. J Hum Hypertens. 1993; 7: 25–32.
Leaf A, Jorgensen MB, Jacobs AK, Cote G, Schoenfeld DA, Scheer J, Weiner BH, Slack JD, Kellett MA, Raizner AE, et al. Do fish oils prevent restenosis after coronary angioplasty Circulation. 1994; 90: 2248–2257.
Levinson PD, Iosiphidis AH, Saritelli AL, Herbert PN, Steiner M. Effects of n-3 fatty acids in essential hypertension. Am J Hypertens. 1990; 3: 754–760.
McVeigh GE, Brennan GM, Cohn JN, Finkelstein SM, Hayes RJ, Johnston GD. Fish oil improves arterial compliance in non-insulin-dependent diabetes mellitus. Arterioscler Thromb. 1994; 14: 1425–1429.
Mehta JL, Lopez LM, Lawson D, Wargovich TJ, Williams LL. Dietary supplementation with omega-3 polyunsaturated fatty acids in patients with stable coronary heart disease: effects on indices of platelet and neutrophil function and exercise performance. Am J Med. 1988; 84: 45–52.
Mills DE, Mah M, Ward RP, Morris BL, Floras JS. Alteration of baroreflex control of forearm vascular resistance by dietary fatty acids. Am J Physiol. 1990; 259: R1164–R1171.
Mills DE, Prkachin KM, Harvey KA, Ward RP. Dietary fatty acid supplementation alters stress reactivity and performance in man. J Hum Hypertens. 1989; 3: 111–116.
Miyajima T, Tsujino T, Saito K, Yokoyama M. Effects of eicosapentaenoic acid on blood pressure, cell membrane fatty acids, and intracellular sodium concentration in essential hypertension. Hypertens Res. 2001; 24: 537–542.
Monahan KD, Wilson TE, Ray CA. Omega-3 fatty acid supplementation augments sympathetic nerve activity responses to physiological stressors in humans. Hypertension. 2004; 44: 732–738.
Mori TA, Bao DQ, Burke V, Puddey IB, Beilin LJ. Docosahexaenoic acid but not eicosapentaenoic acid lowers ambulatory blood pressure and heart rate in humans. Hypertension. 1999; 34: 253–260.
Nestel P, Shige H, Pomeroy S, Cehun M, Abbey M, Raederstorff D. The n-3 fatty acids eicosapentaenoic acid and docosahexaenoic acid increase systemic arterial compliance in humans. Am J Clin Nutr. 2002; 76: 326–330.
O’Keefe JH, Abulssa H, Sastre A, Steinhaus D, Harris W AHA-recommended intakes of omega-3 fatty acids improve indicators of cardiac autonomic tone but not lipids or inflammatory markers. Presented at: American College of Cardiology Scientific Session; March 6–9, 2005; Orlando, Fla. Abstract.
Solomon SA, Cartwright I, Pockley G, Greaves M, Preston FE, Ramsay LE, Waller PC. A placebo-controlled, double-blind study of eicosapentaenoic acid-rich fish oil in patients with stable angina pectoris. Curr Med Res Opin. 1990; 12: 1–11.
Stark KD, Holub BJ. Differential eicosapentaenoic acid elevations and altered cardiovascular disease risk factor responses after supplementation with docosahexaenoic acid in postmenopausal women receiving and not receiving hormone replacement therapy. Am J Clin Nutr. 2004; 79: 765–773.
Toft I, Bonaa KH, Ingebretsen OC, Nordoy A, Jenssen T. Effects of n-3 polyunsaturated fatty acids on glucose homeostasis and blood pressure in essential hypertension: a randomized, controlled trial. Ann Intern Med. 1995; 123: 911–918.
Vacek JL, Harris WS, Haffey K. Short-term effects of omega-3 fatty acids on exercise stress test parameters, angina and lipoproteins. Biomed Pharmacother. 1989; 43: 375–379.
Vandongen R, Mori TA, Burke V, Beilin LJ, Morris J, Ritchie J. Effects on blood pressure of omega 3 fats in subjects at increased risk of cardiovascular disease. Hypertension. 1993; 22: 371–379.
Wing LM, Nestel PJ, Chalmers JP, Rouse I, West MJ, Bune AJ, Tonkin AL, Russell AE. Lack of effect of fish oil supplementation on blood pressure in treated hypertensives. J Hypertens. 1990; 8: 339–343.
Woodman RJ, Mori TA, Burke V, Puddey IB, Watts GF, Beilin LJ. Effects of purified eicosapentaenoic and docosahexaenoic acids on glycemic control, blood pressure, and serum lipids in type 2 diabetic patients with treated hypertension. Am J Clin Nutr. 2002; 76: 1007–1015.
Verhagen AP, de Vet HC, de Bie RA, Kessels AG, Boers M, Bouter LM, Knipschild PG. The Delphi List: a criteria list for quality assessment of randomized clinical trials for conducting systematic reviews developed by Delphi consensus. J Clin Epidemiol. 1998; 51: 1235–1241.
Follmann D, Elliott P, Suh I, Cutler J. Variance imputation for overviews of clinical trials with continuous response. J Clin Epidemiol. 1992; 45: 769–773.
DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986; 7: 177–188.
Takkouche B, Cadarso-Suarez C, Spiegelman D. Evaluation of old and new tests of heterogeneity in epidemiologic meta-analysis. Am J Epidemiol. 1999; 150: 206–215.
Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. Br Med J. 1997; 315: 629–634.
Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994; 50: 1088–1101.
Stram DO. Meta-analysis of published data using a linear mixed-effects model. Biometrics. 1996; 52: 536–544.
Geleijnse JM, Giltay EJ, Grobbee DE, Donders AR, Kok FJ. Blood pressure response to fish oil supplementation: metaregression analysis of randomized trials. J Hypertens. 2002; 20: 1493–1499.
Mozaffarian D, Gottdiener JS, Siscovick DS. Fish intake and cardiac structure, function, and hemodynamics: the Cardiovascular Health Study. Am Heart J. In press.
Charnock JS, McLennan PL, Abeywardena MY. Dietary modulation of lipid metabolism and mechanical performance of the heart. Mol Cell Biochem. 1992; 116: 19–25.
McLennan PL, Barnden LR, Bridle TM, Abeywardena MY, Charnock JS. Dietary fat modulation of left ventricular ejection fraction in the marmoset due to enhanced filling. Cardiovasc Res. 1992; 26: 871–877.(Dariush Mozaffarian, MD, )
Abstract
Background— The effect of fish oil on heart rate (HR), a major risk factor for sudden death, is not well established. We calculated this effect in a meta-analysis of randomized, double-blind, placebo-controlled trials in humans.
Methods and Results— Randomized trials of fish oil that evaluated HR were identified through MEDLINE (1966 through January 2005), hand-searching of references, and contact with investigators for unpublished results. Two investigators independently extracted trial data. A pooled estimate was calculated from random-effects meta-analysis. Predefined stratified meta-analyses and meta-regression were used to explore potential heterogeneity. Of 197 identified articles, 30 met inclusion criteria. Evidence for publication bias was not present. In the overall pooled estimate, fish oil decreased HR by 1.6 bpm (95% CI, 0.6 to 2.5; P=0.002) compared with placebo. Between-trial heterogeneity was evident (Q test, P<0.001). Fish oil reduced HR by 2.5 bpm (P<0.001) in trials with baseline HR 69 bpm (median) but had little effect (0.04-bpm reduction; P=0.56) in trials with baseline HR <69 bpm (P for interaction=0.03). Fish oil reduced HR by 2.5 bpm (P<0.001) in trials with duration 12 weeks but had less effect (0.7-bpm reduction; P=0.27) in trials with duration <12 weeks (P for interaction=0.07). HR reduction with fish oil intake did not significantly vary by fish oil dose (range, 0.81 to 15 g/d), type of HR measure, population age, population health, parallel versus crossover design, type of control oil, or study quality by Delphi criteria (P for interaction >0.25 for each).
Conclusions— In randomized controlled trials in humans, fish oil reduces HR, particularly in those with higher baseline HR or longer treatment duration. These findings provide firm evidence that fish oil consumption directly or indirectly affects cardiac electrophysiology in humans. Potential mechanisms such as effects on the sinus node, ventricular efficiency, or autonomic function deserve further investigation.
Key Words: heart rate fatty acids, omega-3 fish oil meta-analysis randomized controlled trials
Introduction
Fatty fish and fish oil intake is associated with lower risk of cardiac arrhythmias, including sudden death, arrhythmic coronary heart disease death, and atrial fibrillation.1–8 Experimental studies in isolated rat myocytes, exercising dogs, and nonhuman primates suggest that fish oil has direct cardiac electrophysiological effects, including slowing of the heart rate (HR).9–11 However, such effects are not well established in humans. Because higher HR is a major independent risk factor for cardiovascular death, particularly sudden death,12–18 an effect of fish oil on HR would both confirm an influence on cardiac electrophysiology in humans and indicate a plausible potential mechanism for observed relations between fish intake and arrhythmic events. We therefore performed a meta-analysis of randomized placebo-controlled clinical trials to determine the effect of fish oil consumption on HR in humans.
Methods
Selection of Randomized Trials
We followed the Quality of Reporting of Meta-Analyses (QUOROM) standards19 during all phases of the design and implementation of this analysis. Randomized clinical trials of fish oil that included evaluation of HR were identified through MEDLINE (1966 through February 2005), including fish oil trials designed primarily to evaluate other outcomes such as blood pressure or coronary restenosis, hand-searching of reference lists of obtained articles, and contacting investigators for unreported HR data in published trials or for HR data from unpublished trials. To minimize publication bias, we attempted to identify all fish oil trials that may have measured and reported HR data, rather than limiting our search to trials designed primarily to evaluate HR. English-language trials in human subjects >18 years of age were included if oral fish oil supplementation was randomized and changes in HR or baseline and follow-up HR were measured; trials with organ transplant subjects, cointerventions that could not be separated from fish oil treatment, no placebo control, nonblinding of participants, or duration <2 weeks were excluded.
Trial Review
When potentially relevant trials were identified, abstracts and, if necessary, original articles were screened for obvious exclusions by an investigator. Of 197 identified trials, 161 were excluded for not being a randomized trial of fish oil (n=28), for having no available HR data (n=75), for occurring in organ transplant recipients (n=12), for having no placebo control (n=29), for having a cointervention that could not be separated from fish oil treatment (n=6), or for being a duplicate publication from the same study (n=11). The identified trials included 10 published and 2 unpublished trials for which we contacted the authors to determine whether unreported HR data might be available. A list of all reviewed trials and reasons for exclusion is available by request from the investigators. For the remaining 36 trials not excluded during initial screening, each original article was independently reviewed for inclusion by 2 investigators. Six of these trials were excluded for no placebo control (n=2), no follow-up HR data (n=2), duration <2 weeks (n=1), or being a duplicate publication from the same study (n=1). Thirty trials met inclusion and exclusion criteria, including 2 trials for which unpublished HR data were obtained from the authors (personal communications, William Harris, February 18, 2005, and Ingrid Toft, March 4, 2005).20–49 Concordance on inclusion and exclusion decisions was 100%.
Data Extraction
For each of the articles meeting inclusion and exclusion criteria, data were independently extracted by 2 investigators on study design; population; sample size and dropout; fish oil type, dose, and duration; method of HR assessment; change in HR or baseline and follow-up HR values; and HR variance measures. For studies reporting RR interval values (duration of 1 heartbeat in milliseconds), HR was calculated and its corresponding variance was estimated proportionally to the RR interval variance. Study quality was also independently assessed by 2 investigators according to the criteria for quality assessment of randomized clinical trials developed by Delphi consensus.50 The 9 criteria (1a, 1b, and 2 through 8) include, for example, whether a method of randomization was performed, whether the treatment groups were similar at baseline with regard to the most important prognostic indicators, and whether the analysis was of intention-to-treat design. For the last criterion, we considered analyses as having intention-to-treat design if all subjects not lost to follow-up were analyzed according to their original randomization group; exclusions were not made for noncompliance. We assessed the validity of data extraction by comparing the independently abstracted results for concordance, and any discrepancies were resolved by discussion and review of the original manuscript by the 2 investigators who extracted the data or, if necessary, a committee comprising all the investigators. When necessary, missing information (type of control oil, mean age of participants, etc) was obtained by direct contact with the original authors. We attempted to minimize clinical heterogeneity by excluding studies in children, in organ transplant recipients, or with duration <2 weeks. Remaining clinical heterogeneity was evaluated qualitatively by comparing the mean age, gender distribution, and general health of the study populations; the doses and durations of fish oil treatment; and the methods of HR assessment. Clinical heterogeneity was assessed quantitatively in prespecified stratified analyses (see Statistical Analysis).
Statistical Analysis
Our primary outcome was the change in HR resulting from fish oil treatment. For parallel-design trials, the HR change from baseline to study end in the control group was subtracted from the HR change from baseline to study end in the treatment group. For crossover design trials, the HR at the end of the control period was subtracted from the HR at the end of the treatment period. Within-individual changes were used when available; otherwise, group means were used. SEs were abstracted or, if not reported, derived from SDs, CIs, or probability values. The pooled variance for the net HR change resulting from fish oil treatment was calculated as (1) SE2T+SE2C–2(r)(SET)(SEC) for crossover design trials, where SET and SEC are the SE of the treatment and control period HR values, respectively, and r is the within-individual correlation between the treatment and control period HR values, and (2) SE2TG+SE2CG for parallel-design trials, where SETG and SECG are the SE of the HR change from baseline to study end in the treatment and control groups, respectively. For parallel-design trials that reported precision of baseline and final HR values (n=18) rather than HR changes, SETG and SECG were calculated according to the method of Follmann et al,51 which involves making an assumption for the unreported within-individual correlation between baseline and final HR values. On the basis of measured correlations in fish oil trials (Anouk Geelen, personal communication, January 27, 2005), the within-individual correlation between HR values was estimated to be 0.60 for trials using a single HR measure, 0.80 for trials using the average of multiple measures, and 0.85 for trials using a 24-hour measure, with the higher correlations consistent with less random error in the HR measurement. Sensitivity analyses were performed assuming a within-individual correlation of 0.60 for all trials. Data for the calculation of the change in HR and the variance of this change were not missing from any trial.
Pooled estimates of the effect of fish oil on HR were calculated through the use of random-effects meta-analysis, which accounts for heterogeneity in treatment effects among trials, using the method of DerSimonian and Laird52 with inverse-variance (SE) weighting. Because some trials compared multiple intervention groups with a single control group (n=7), we performed sensitivity analyses in which separate pooled estimates and variances for the effect of fish oil on HR were calculated using separate meta-analyses for each of these trials; these trial-specific estimates then were used in a second meta-analysis evaluating all trials. Heterogeneity between studies was tested with the DerSimonian and Laird Q statistic.52,53 To assess publication bias, a funnel plot of the treatment effect versus SE was visually inspected.54 Potential publication bias was also evaluated with the Begg adjusted-rank correlation test,55 a statistical analog of the visual funnel graph, and the regression asymmetry test according to the method of Egger et al.54
We performed predefined stratified meta-analyses to explore potential heterogeneity by dose of eicosapentaenoic acid and docasohexaenoic acid (EPA+DHA) (at the median), duration of treatment (12 weeks versus less), type of HR measure (single measure, average of multiple resting measures, or 24-hour measure), baseline HR (at the median), type of control oil (olive oil versus other), population age (at the median), general health (healthy versus otherwise), study design (parallel versus crossover), and study quality (meeting at least 8 Delphi criteria versus fewer). We used meta-regression to test for heterogeneity of the pooled treatment effect by these factors,56 testing for significance of the stratifying variable by using the Wald test in a mixed-effects meta-regression model. We also performed sensitivity analyses excluding trials with 20% dropout of randomized participants at baseline. All analyses were performed with Stata version 8.2 (Stata Corp). Statistical significance was defined as 2-tailed <0.05.
Results
Overview of Trials
Of the 30 trials meeting inclusion and exclusion criteria, 6 had 2 separate intervention groups, and 1 had 3 separate intervention groups, for a total of 38 intervention groups in the 30 trials (Table 1). Although single-blind trials were acceptable, all were double-blind trials. Eight were crossover design trials, and 22 were parallel-design trials. Median study size was 30 participants; in total, this meta-analysis included 1678 individuals treated with fish oil or placebo for 27 615 person-weeks. The mean ages of the study populations ranged from 23 to 68 years (median, 54 years). Sixteen intervention groups were made up of generally healthy populations; 22 comprised individuals with 1 underlying chronic condition. The median EPA+DHA dose was 3.5 g/d (range, 0.81 to 15 g/d), and the median treatment duration was 8 weeks (range, 4 to 52 weeks). Thirteen intervention groups assessed HR with a single resting measure; 14 used the average of 2 or 3 resting measures; and 11 used the average of ambulatory or continuous monitoring. Twenty-five trials (30 intervention groups) met at least 8 Delphi criteria for study quality; 5 trials (8 intervention groups) met <8.
Our broad search methods appeared to be successful in minimizing the effect of publication bias. Among the 30 included trials, 12 reported HR findings in the abstract (7 reporting an effect, 5 reporting the absence of an effect); 10 reported HR findings in the results text but not the abstract (5 reporting an effect, 5 reporting the absence of an effect); 6 presented HR findings in a table only (all 6 showing no significant effect); and 2 constituted unpublished results. Little evidence for publication bias was present by visual inspection of a funnel plot (Figure 1), Begg’s test (P=0.87), or Egger’s test (P=0.69).
Effect of Fish Oil on HR
The individual trial results and the pooled estimate are presented in Figure 2. In the overall pooled estimate, fish oil decreased HR by 1.6 bpm (95% CI, 0.6 to 2.5; P=0.002) compared with placebo. Exclusion of trials with 20% dropout (n=5) had little effect on the pooled estimate, with fish oil decreasing HR by 1.3 bpm (95% CI, 0.3 to 2.4; P=0.009). Assuming a within-individual HR correlation of 0.60 for all trials also had little effect, with fish oil decreasing HR by 1.5 bpm (95% CI, 0.5 to 2.5; P=0.003). The pooled estimate was also similar in sensitivity analyses accounting for multiple intervention groups in some trials, with fish oil decreasing HR by 1.4 bpm (95% CI, 0.4 to 2.5; P=0.007).
Between-trial heterogeneity was evident (Q test, P<0.001). We evaluated prespecified study characteristics to explore reasons for potential heterogeneity (Table 2). The HR reduction with fish oil consumption was greater in study populations with a mean baseline HR 69 bpm (P for interaction=0.03), among whom fish oil reduced HR by 2.5 bpm (95% CI, 1.4 to 3.5; P<0.001), and in study populations receiving 12 weeks of fish oil treatment (P for interaction=0.07), among whom fish oil reduced HR by 2.5 bpm (95% CI, 1.1 to 4.0; P=0.001). Although other differences related to study characteristics were not statistically significant (Table 2), several findings were consistent with intuition; eg, the effect of fish oil on HR appeared possibly greater with increasing precision of the measurement method used (single versus average of 2 or 3 measures versus ambulatory/continuous), consistent with reduced measurement error reducing bias toward the null.
Little evidence was present for a dose-response effect. Stratified at the median dose of fish oil (3.5 g/d), the reduction in HR was not significantly different at higher versus lower doses (each compared with placebo) (P for interaction=0.72) (Table 2). Similarly, stratified into quartiles of fish oil dose, HR was reduced by 1.1 (95% CI, –0.9 to 3.1), 1.8 (95% CI, –0.1 to 3.6), 1.9 (95% CI, 0.1 to 3.8), and 1.5 (95% CI, –0.6 to 3.6) bpm in quartiles 1 through 4, respectively, compared with placebo (P for ordinal interaction=0.72). Evaluated continuously, the dose of fish oil was not a predictor of treatment effect (P=0.63), above and beyond being on fish oil treatment (yes/no). In the 2 trials with EPA+DHA doses 1 g/d, HR was reduced by 5.0 bpm (95% CI, 2.3 to 7.7; P<0.001) compared with 1.4 bpm in the trials with EPA+DHA doses >1 g/d (95% CI, 0.4 to 2.3; P<0.001).
When we evaluated different factors simultaneously in the meta-regression model, there appeared to be potential independent heterogeneity related to both baseline HR (P for interaction=0.04) and treatment duration (P for interaction=0.09). Among the 9 trials with mean baseline HR 69 bpm and treatment duration 12 weeks, fish oil reduced HR by 2.9 bpm (95% CI, 1.5 to 4.4; P<0.001) compared with placebo, without significant between-trial heterogeneity (Q test, P>0.05).
Discussion
In this meta-analysis of randomized, double-blind, placebo-controlled clinical trials, fish oil consumption reduced HR in humans. Although the overall effect was modest (1.6-bpm reduction), on a population level, even modest differences in risk factors can have a significant impact on health. These findings provide firm evidence for an effect of fish oil consumption on cardiac electrophysiology in humans.
The regulation of HR is a complex physiological process, with components related to vagal tone, sympathetic input, responsiveness of the sinus node, and systolic and diastolic left ventricular function. The decrease in HR with fish oil consumption indicates that marine n-3 fatty acids influence at least 1 of these parameters. The n-3 fatty acids are incorporated into myocyte membranes and may influence ion channel function9,10; this could directly alter the automaticity or responsiveness of the sinus node. Fish oil also lowers blood pressure in humans,57 possibly by reducing systemic vascular resistance.58 In one observational study, such an effect was apparent at dietary levels of fish intake.58 Such a decrease in systemic vascular resistance would reduce left ventricular afterload and improve diastolic function, which could indirectly reduce HR as a result of better ventricular efficiency. Experimental studies in nonhuman primates support the hypothesis that fish oil consumption improves left ventricular efficiency.59,60 Intake of n-3 fatty acid may also improve measures of HR variability,21–23,27 suggesting a potential effect on autonomic tone. Our findings substantiate an electrophysiological effect of fish oil in humans and support the need for further investigation of these potential mechanisms.
Higher HR is associated with increased cardiovascular risk, including greater risk of sudden death,12–15,17 coronary heart disease death,13,14 and cardiovascular death.16 A higher HR could directly increase cardiovascular risk, eg, by increasing myocardial vulnerability to ischemia or arrhythmia. On the basis of work by Jouven et al,17 our finding of a 1.6-bpm HR reduction with fish oil consumption would correspond to an 5% lower risk of sudden death. Thus, in addition to effects on HR, other mechanisms are likely to contribute to the reductions in sudden death risk with fish or fish oil consumption seen in observational studies and randomized trials. A higher HR may indicate less optimal underlying cardiovascular health as manifested by increased sympathetic tone, decreased vagal tone, or decreased ventricular efficiency. The HR reduction with fish oil consumption could therefore indicate beneficial effects of fish oil on these other physiological parameters that might reduce cardiovascular risk to a greater extent than that resulting from the change in HR alone.
Our exploration of heterogeneity revealed several interesting findings. First, the reduction in HR appeared larger in trials with longer duration of intake (12 weeks). This may relate in part to the time needed for EPA and DHA to be incorporated into the tissues where they exert their effects and suggests that regular consumption over time may have greater effects than short-term intake. Second, HR was reduced to a greater extent in populations with higher baseline HR. Because fish oil was compared with placebo in each trial, this result would not be due to regression toward the mean. This finding suggests that fish oil may have greater effects on HR in populations with higher intrinsic sinus node automaticity, greater sympathetic tone, lower vagal tone, or lower ventricular efficiency. Third, although power was insufficient to prove equivalence of different doses, very high consumption of fish oil did not appear to have substantially greater effects than modest consumption. This is consistent with observational studies and randomized trials indicating clinical benefits of fatty fish or fish oil consumption at relatively modest intake, 1 to 2 servings per week or 500 to 1000 mg/d EPA+DHA, respectively.1–8 In the present meta-analysis, the lowest EPA+DHA doses were 1 g/d, and it is possible that a dose-response effect may exist at lower (eg, dietary) levels of intake, as suggested by one observational analysis.58 Finally, although the differences were not statistically significant, the HR reduction was smaller in trials using a single resting measure of HR, intermediate in trials using the average of 2 or 3 resting measures, and greatest in trials using ambulatory or continuous measures. This is consistent with a greater degree of misclassification (random measurement error) when only a single or a few resting measures were used, suggesting that such trials may underestimate the true effect of fish oil on HR. Alternatively, the results of trials using ambulatory and continuous monitoring represent the averaged effect of fish oil consumption on both resting and activity-related HR responses, which may be somewhat greater than effects on resting HR alone.
Publication bias is a major potential limitation of meta-analyses. Our broad, prespecified search methods and contacting of investigators for unpublished results appeared to be successful in minimizing the effect of publication bias; in only a minority of included trials was a significant HR effect prominently reported, and little evidence was present for publication bias in the final included studies. Additionally, given the large number of included trials, it is unlikely that the results of even several additional studies would greatly alter the pooled estimate. Between-trial heterogeneity may limit the generalizability of the overall pooled estimate; we attempted to account for potential heterogeneity by using a random-effects model and by assessing factors that may explain between-trial differences.
In this meta-analysis of randomized, double-blind, placebo-controlled clinical trials, fish oil reduced HR, particularly with higher baseline HR or longer durations of treatment. These results provide strong evidence that fish oil consumption directly or indirectly influences cardiac electrophysiology in humans. This effect may directly account for part of the observed benefits of fish intake on cardiovascular risk, particularly risk of arrhythmic events, and may indicate favorable effects on physiological systems such as on autonomic tone, vascular resistance, or ventricular efficiency that improve cardiovascular health.
Acknowledgments
The Wageningen Centre for Food Sciences is an alliance of major Dutch food industries, Maastricht University, TNO Nutrition and Food Research in Zeist, and Wageningen University and Research Centre, with financial support by the Dutch Government. Dr Mozaffarian was supported by a Mentored Clinical Scientist Award from the National Heart, Lung, and Blood Institute, National Institutes of Health (K08-HL-075628) and thanks Drs Eric Rimm, David Siscovick, and David Herrington for their invaluable guidance and support. The authors thank Drs William Harris and Ingrid Toft for sharing unpublished results for this analysis.
Footnotes
Guest Editor for this article was Robert H. Eckel, MD.
Medline search criteria included (heart rate or blood pressure or restenosis) and (fish oil or n-3 fatty acids or omega-3 or eicosapentaenoic or docosahexaenoic). Limits were adults 19 years of age, English language, clinical trial, and humans.
References
Burr ML, Fehily AM, Gilbert JF, Rogers S, Holliday RM, Sweetnam PM, Elwood PC, Deadman NM. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: Diet and Reinfarction Trial. Lancet. 1989; 2: 757–761.
Siscovick DS, Raghunathan TE, King I, Weinmann S, Wicklund KG, Albright J, Bovbjerg V, Arbogast P, Smith H, Kushi LH, et al. Dietary intake and cell membrane levels of long-chain n-3 polyunsaturated fatty acids and risk of primary cardiac arrest. JAMA. 1995; 274: 1363–1367.
Albert CM, Hennekens CH, O’Donnell CJ, Ajani UA, Carey VJ, Willett WC, Ruskin JN, Manson JE. Fish consumption and risk of sudden cardiac death. JAMA. 1998; 279: 23–28.
Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto miocardico. Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: GISSI-Prevenzione trial. Lancet. 1999; 354: 447–455.
Hu FB, Bronner L, Willett WC, Stampfer MJ, Rexrode KM, Albert CM, Hunter D, Manson JE. Fish and omega-3 fatty acid intake and risk of coronary heart disease in women. JAMA. 2002; 287: 1815–1821.
Mozaffarian D, Lemaitre RN, Kuller LH, Burke GL, Tracy RP, Siscovick DS. Cardiac benefits of fish consumption may depend on the type of fish meal consumed: the Cardiovascular Health Study. Circulation. 2003; 107: 1372–1377.
Mozaffarian D, Ascherio A, Hu FB, Stampfer MJ, Willett WC, Siscovick DS, Rimm EB. Interplay between different polyunsaturated fatty acids and risk of coronary heart disease in men. Circulation. 2005; 111: 157–164.
Mozaffarian D, Psaty BM, Rimm EB, Lemaitre RN, Burke GL, Lyles MF, Lefkowitz D, Siscovick DS. Fish intake and risk of incident atrial fibrillation. Circulation. 2004; 110: 368–373.
McLennan PL Myocardial membrane fatty acids and the antiarrhythmic actions of dietary fish oil in animal models. Lipids. 2001; 36: 111S–114S.
Leaf A, Kang JX, Xiao YF, Billman GE. Clinical prevention of sudden cardiac death by n-3 polyunsaturated fatty acids and mechanism of prevention of arrhythmias by n-3 fish oils. Circulation. 2003; 107: 2646–2652.
Billman GE, Kang JX, Leaf A. Prevention of ischemia-induced cardiac sudden death by n-3 polyunsaturated fatty acids in dogs. Lipids. 1997; 32: 1161–1168.
Dyer AR, Persky V, Stamler J, Paul O, Shekelle RB, Berkson DM, Lepper M, Schoenberger JA, Lindberg HA. Heart rate as a prognostic factor for coronary heart disease and mortality: findings in three Chicago epidemiologic studies. Am J Epidemiol. 1980; 112: 736–749.
Kannel WB, Kannel C, Paffenbarger RS Jr, Cupples LA. Heart rate and cardiovascular mortality: the Framingham study. Am Heart J. 1987; 113: 1489–1494.
Shaper AG, Wannamethee G, Macfarlane PW, Walker M. Heart rate, ischaemic heart disease, and sudden cardiac death in middle-aged British men. Br Heart J. 1993; 70: 49–55.
Wannamethee G, Shaper AG, Macfarlane PW, Walker M. Risk factors for sudden cardiac death in middle-aged British men. Circulation. 1995; 91: 1749–1756.
Palatini P, Casiglia E, Julius S, Pessina AC. High heart rate: a risk factor for cardiovascular death in elderly men. Arch Intern Med. 1999; 159: 585–592.
Jouven X, Zureik M, Desnos M, Guerot C, Ducimetiere P. Resting heart rate as a predictive risk factor for sudden death in middle-aged men. Cardiovasc Res. 2001; 50: 373–378.
Palatini P, Thijs L, Staessen JA, Fagard RH, Bulpitt CJ, Clement DL, de Leeuw PW, Jaaskivi M, Leonetti G, Nachev C, O’Brien ET, Parati G, Rodicio JL, Roman E, Sarti C, Tuomilehto J, for the Systolic Hypertension in Europe (Syst-Eur) Trial Investigators. Predictive value of clinic and ambulatory heart rate for mortality in elderly subjects with systolic hypertension. Arch Intern Med. 2002; 162: 2313–2321.
Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement: Quality of Reporting of Meta-Analyses. Lancet. 1999; 354: 1896–1900.
Bairati I, Roy L, Meyer F. Effects of a fish oil supplement on blood pressure and serum lipids in patients treated for coronary artery disease. Can J Cardiol. 1992; 8: 41–46.
Christensen JH, Christensen MS, Dyerberg J, Schmidt EB. Heart rate variability and fatty acid content of blood cell membranes: a dose-response study with n-3 fatty acids. Am J Clin Nutr. 1999; 70: 331–337.
Christensen JH, Aaroe J, Knudsen N, Dideriksen K, Kornerup HJ, Dyerberg J, Schmidt EB. Heart rate variability and n-3 fatty acids in patients with chronic renal failure: a pilot study. Clin Nephrol. 1998; 49: 102–106.
Christensen JH, Gustenhoff P, Korup E, Aaroe J, Toft E, Moller J, Rasmussen K, Dyerberg J, Schmidt EB. Effect of fish oil on heart rate variability in survivors of myocardial infarction: a double blind randomised controlled trial. BMJ. 1996; 312: 677–678.
Conquer JA, Holub BJ. Effect of supplementation with different doses of DHA on the levels of circulating DHA as non-esterified fatty acid in subjects of Asian Indian background. J Lipid Res. 1998; 39: 286–292.
Deslypere JP. Influence of supplementation with N-3 fatty acids on different coronary risk factors in men: a placebo controlled study. Verh K Acad Geneeskd Belg. 1992; 54: 189–216.
Dyerberg J, Eskesen DC, Andersen PW, Astrup A, Buemann B, Christensen JH, Clausen P, Rasmussen BF, Schmidt EB, Tholstrup T, Toft E, Toubro S, Stender S. Effects of trans- and n-3 unsaturated fatty acids on cardiovascular risk markers in healthy males: an 8 weeks dietary intervention study. Eur J Clin Nutr. 2004; 58: 1062–1070.
Geelen A, Zock PL, Swenne CA, Brouwer IA, Schouten EG, Katan MB. Effect of n-3 fatty acids on heart rate variability and baroreflex sensitivity in middle-aged subjects. Am Heart J. 2003; 146: E4.
Geelen A, Brouwer IA, Schouten EG, Maan AC, Katan MB, Zock PL. Effects of n-3 fatty acids from fish on premature ventricular complexes and heart rate in humans. Am J Clin Nutr. 2005; 81: 416–420.
Gray DR, Gozzip CG, Eastham JH, Kashyap ML. Fish oil as an adjuvant in the treatment of hypertension. Pharmacotherapy. 1996; 16: 295–300.
Grimsgaard S, Bonaa KH, Hansen JB, Myhre ES. Effects of highly purified eicosapentaenoic acid and docosahexaenoic acid on hemodynamics in humans. Am J Clin Nutr. 1998; 68: 52–59.
Landmark K, Thaulow E, Hysing J, Mundal HH, Eritsland J, Hjermann I. Effects of fish oil, nifedipine and their combination on blood pressure and lipids in primary hypertension. J Hum Hypertens. 1993; 7: 25–32.
Leaf A, Jorgensen MB, Jacobs AK, Cote G, Schoenfeld DA, Scheer J, Weiner BH, Slack JD, Kellett MA, Raizner AE, et al. Do fish oils prevent restenosis after coronary angioplasty Circulation. 1994; 90: 2248–2257.
Levinson PD, Iosiphidis AH, Saritelli AL, Herbert PN, Steiner M. Effects of n-3 fatty acids in essential hypertension. Am J Hypertens. 1990; 3: 754–760.
McVeigh GE, Brennan GM, Cohn JN, Finkelstein SM, Hayes RJ, Johnston GD. Fish oil improves arterial compliance in non-insulin-dependent diabetes mellitus. Arterioscler Thromb. 1994; 14: 1425–1429.
Mehta JL, Lopez LM, Lawson D, Wargovich TJ, Williams LL. Dietary supplementation with omega-3 polyunsaturated fatty acids in patients with stable coronary heart disease: effects on indices of platelet and neutrophil function and exercise performance. Am J Med. 1988; 84: 45–52.
Mills DE, Mah M, Ward RP, Morris BL, Floras JS. Alteration of baroreflex control of forearm vascular resistance by dietary fatty acids. Am J Physiol. 1990; 259: R1164–R1171.
Mills DE, Prkachin KM, Harvey KA, Ward RP. Dietary fatty acid supplementation alters stress reactivity and performance in man. J Hum Hypertens. 1989; 3: 111–116.
Miyajima T, Tsujino T, Saito K, Yokoyama M. Effects of eicosapentaenoic acid on blood pressure, cell membrane fatty acids, and intracellular sodium concentration in essential hypertension. Hypertens Res. 2001; 24: 537–542.
Monahan KD, Wilson TE, Ray CA. Omega-3 fatty acid supplementation augments sympathetic nerve activity responses to physiological stressors in humans. Hypertension. 2004; 44: 732–738.
Mori TA, Bao DQ, Burke V, Puddey IB, Beilin LJ. Docosahexaenoic acid but not eicosapentaenoic acid lowers ambulatory blood pressure and heart rate in humans. Hypertension. 1999; 34: 253–260.
Nestel P, Shige H, Pomeroy S, Cehun M, Abbey M, Raederstorff D. The n-3 fatty acids eicosapentaenoic acid and docosahexaenoic acid increase systemic arterial compliance in humans. Am J Clin Nutr. 2002; 76: 326–330.
O’Keefe JH, Abulssa H, Sastre A, Steinhaus D, Harris W AHA-recommended intakes of omega-3 fatty acids improve indicators of cardiac autonomic tone but not lipids or inflammatory markers. Presented at: American College of Cardiology Scientific Session; March 6–9, 2005; Orlando, Fla. Abstract.
Solomon SA, Cartwright I, Pockley G, Greaves M, Preston FE, Ramsay LE, Waller PC. A placebo-controlled, double-blind study of eicosapentaenoic acid-rich fish oil in patients with stable angina pectoris. Curr Med Res Opin. 1990; 12: 1–11.
Stark KD, Holub BJ. Differential eicosapentaenoic acid elevations and altered cardiovascular disease risk factor responses after supplementation with docosahexaenoic acid in postmenopausal women receiving and not receiving hormone replacement therapy. Am J Clin Nutr. 2004; 79: 765–773.
Toft I, Bonaa KH, Ingebretsen OC, Nordoy A, Jenssen T. Effects of n-3 polyunsaturated fatty acids on glucose homeostasis and blood pressure in essential hypertension: a randomized, controlled trial. Ann Intern Med. 1995; 123: 911–918.
Vacek JL, Harris WS, Haffey K. Short-term effects of omega-3 fatty acids on exercise stress test parameters, angina and lipoproteins. Biomed Pharmacother. 1989; 43: 375–379.
Vandongen R, Mori TA, Burke V, Beilin LJ, Morris J, Ritchie J. Effects on blood pressure of omega 3 fats in subjects at increased risk of cardiovascular disease. Hypertension. 1993; 22: 371–379.
Wing LM, Nestel PJ, Chalmers JP, Rouse I, West MJ, Bune AJ, Tonkin AL, Russell AE. Lack of effect of fish oil supplementation on blood pressure in treated hypertensives. J Hypertens. 1990; 8: 339–343.
Woodman RJ, Mori TA, Burke V, Puddey IB, Watts GF, Beilin LJ. Effects of purified eicosapentaenoic and docosahexaenoic acids on glycemic control, blood pressure, and serum lipids in type 2 diabetic patients with treated hypertension. Am J Clin Nutr. 2002; 76: 1007–1015.
Verhagen AP, de Vet HC, de Bie RA, Kessels AG, Boers M, Bouter LM, Knipschild PG. The Delphi List: a criteria list for quality assessment of randomized clinical trials for conducting systematic reviews developed by Delphi consensus. J Clin Epidemiol. 1998; 51: 1235–1241.
Follmann D, Elliott P, Suh I, Cutler J. Variance imputation for overviews of clinical trials with continuous response. J Clin Epidemiol. 1992; 45: 769–773.
DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986; 7: 177–188.
Takkouche B, Cadarso-Suarez C, Spiegelman D. Evaluation of old and new tests of heterogeneity in epidemiologic meta-analysis. Am J Epidemiol. 1999; 150: 206–215.
Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. Br Med J. 1997; 315: 629–634.
Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994; 50: 1088–1101.
Stram DO. Meta-analysis of published data using a linear mixed-effects model. Biometrics. 1996; 52: 536–544.
Geleijnse JM, Giltay EJ, Grobbee DE, Donders AR, Kok FJ. Blood pressure response to fish oil supplementation: metaregression analysis of randomized trials. J Hypertens. 2002; 20: 1493–1499.
Mozaffarian D, Gottdiener JS, Siscovick DS. Fish intake and cardiac structure, function, and hemodynamics: the Cardiovascular Health Study. Am Heart J. In press.
Charnock JS, McLennan PL, Abeywardena MY. Dietary modulation of lipid metabolism and mechanical performance of the heart. Mol Cell Biochem. 1992; 116: 19–25.
McLennan PL, Barnden LR, Bridle TM, Abeywardena MY, Charnock JS. Dietary fat modulation of left ventricular ejection fraction in the marmoset due to enhanced filling. Cardiovasc Res. 1992; 26: 871–877.(Dariush Mozaffarian, MD, )