Basic research
Effect of dietary omega-3 fatty acid deficiency on heart rate variability in hooded rats

Introduction
Moment-to-moment variation in heart rate, which reflects the influence of autonomic tone on cardiac function, serves as a powerful prognostic indicator for survival following myocardical infarction [1, 2]. Some studies [3-7], but not all [8] suggest, in both healthy human adults, and survivors of myocardial infarction (MI), a relationship between omega-3 PUFA content in blood cells and altered heart rate dynamics. The above-mentioned studies by Christensen and colleagues are in agreement that increased omega-3 levels are correlated with an increase in heart rate variability (HRV), as measured using time-domain analyses, though the mechanisms that modulate these effects are yet to be elucidated. However, these studies did not find increases in HRV following short-term dietary supplementation with omega-3 PUFA. Heart rate dynamics are usually estimated using parameters obtained from time-domain or frequency-domain analyses [9, 10]. Results obtained with these methods are confounded by the changing statistical properties of heartbeat interval time-series over time, i.e. non-stationary signals. These signals are difficult to interpret with dispersional measures in the time-domain, such as the standard deviation, because the results are not stable with increasing data length. Additionally, these measures do not give any information regarding the internal dynamics of the time-series. Frequency-domain measures rely on assumptions of stationarity that are not met with interbeat interval (IBI) data, especially with longer recording times. We have used a technique that measures the internal dynamics of an IBI time-series by calculating the Hurst exponent (H). The Hurst exponent is a direct measure of the dynamics of a time-series, conceived by Hurst [11] and formalized by Mandelbrot [12]. The value of H varies, with 0ŁHŁ1. Visual inspection of a time-series with H=1.0 (as shown in Figure 1C), reveals an overall relative smoothness to the graph of IBI versus time. As H tends towards 0, trends are more rapidly reversed (Figure 1B, H=0.05), where there are large variations between adjacent values, which give it a very irregular look. When H=0.5, (Figure 1D), the magnitude of the sequential points of the time-series are independent (of one another) and therefore uncorrelated. The time-series can be considered a random walk. Thus, H values approaching 0.5 from either extreme are symptomatic of a breakdown in the long-range correlations of the heart interbeat interval signal. Inasmuch as these long-term correlations are associated with cardiac health [13], Hurst analysis of heart interbeat intervals can provide a measure of cardiac signal complexity in relationship to disease states, and, as in this study, to differing neural fatty acid profiles, subsequent to dietary manipulation. Furthermore, since the measure H is not significantly altered by changes in data length, statistical moments, or means of the signal [14], it allows comparisons of cardiac signal complexity measures between different studies. Cardiac signal complexity is related to the autocorrelation of the interbeat interval time-series. The increased autocorrelation is manifested by increased beat-to-beat variability for the time-series as H approaches 0. For these time-series, the parameter H corresponds to increased heart rate variability (HRV). Note however that increased autocorrelation is seen as H diverges from the value 0.5 (random walk) towards either 1 or 0. Therefore, the degree of autocorrelation is relatively equivalent to HRV measures only for time-series with 0ŁHŁ0.5. The term HRV will be used as a descriptor of beat-to-beat variability, in accordance with current usage. For interbeat interval time-series, H is a measure of both HRV as well as cardiac signal complexity. Therefore, increases or decreases in H correspond to decreases or increases in HRV, respectively. A time-domain measure such as standard deviation, a classic descriptor of variance in data, does not always correlate with HRV. This is because the relationship between each interbeat interval is lost. This point is illustrated in Figure 1. All time-series represented in Figure 1B-1D were generated, as outlined in the methods, to have the same mean and standard deviation as an IBI time-series obtained from one rat, (mean±SD 154.9 ms ±8.3), shown in Figure 1A. The time-series from the animal had an associated H value of 0.05±0.02. By visual inspection, the “roughness” of this time-series best matches the synthesized IBI series from Figure 1B, with H=0.05. It can also easily be seen that beat-to-beat variability is highest as H approaches 0. The current methodology allows a better representation of the internal dynamics of the system than can be obtained from simple dispersional measures such as mean and standard deviation. Inasmuch as a spectral transform of the IBI data is not required prior to analysis, short lengths of IBI data with differing statistical characteristics can be compared. We used this technique to determine H using IBI time-series data obtained from rats raised on diets that differed in the amount of omega-3 fatty acids supplied in the diet. Since decreases in heart rate variability (HRV) are associated with higher values of H [14], we surmised that a higher H value would characterize ECG data from omega-3 deficient rats when compared with those raised on an omega-3 adequate diet. We are not aware of any other study to assess the effects of severe omega-3 PUFA deficiency on HRV using Hurst analysis.
Material and methods
Animals and diets
All procedures involving animals were conducted in accordance with local Animal Care guidelines and were approved by the NIAAA Animal Ethics Committee. Animals in this experiment were raised on one of two semi-purified diets (Table I). Dietary fats were supplied by the supplementary oils added to the diets, thus creating one diet deficient in omega-3 fatty acids (n-3 DEF), and the other omega-3 adequate with the addition of pre-formed DHA (n-3 ADQ). The assayed fatty acid compositions of both experimental diets are given in Table II. Other details regarding the composition of the salt and vitamin mixes have been described previously [15]. Animals were successively raised for three generations (F3), as previously reported by this and other laboratories [16-19]. Twelve (12) F3 male, hooded rats (Long-Evans) were used for the study (though we failed to obtain data from one of the rats). Animals were fed once per day with dietary pellets and water ad libitum. The ambient light was cycled in the animal room (12/12 light/dark cycle), with temperature maintained at 70°F. Animals were 33 weeks old, and matched for bodyweight at the time of testing (mean ±SD; n-3 DEF: 878±44 g; n-3 ADQ: 881±40 g).
Experimental procedure
ECG IBI data were collected as follows. Individual rats were placed in an anesthesia chamber containing isoflurane (Baxter, Deerfield, IL) for approximately 30 seconds. Upon sedation, animals were transferred to a downdraft table, where they were maintained under anesthesia via a nose-cone. The electrode positions corresponding to the maximal QRS potential was determined using sub-cutaneous needle electrodes, attached to a CD-200 oscilloscope monitor. Typically, this was 1 cm rostral to the sternal notch, with a lateral separation of approximately 3 cm either side of the midline. The chest wall was shaven, and surface ECG electrodes fixed with conductive gel and fast-setting glue. The contact was further protected by tape and a custom-made jacket and harness system that allowed unrestricted movement upon waking. Animals were allowed to awaken from anesthesia following the setup procedure, which took 12-15 minutes. Electrodes were wired to a Polar belt (Polar Electro, Woodbury, NY), in communication with a Mini-logger Series 2000 (Mini-mitter, Sun River, OR, USA), which was used to store the IBI data. The logger configuration was adjusted in consultation with the manufacturer to allow reliable capture of IBIs corresponding to heart rates in the order of 400 beats per minute. The sampling rate for the device is 500 Hz and results in a timing accuracy of <1 ms in the measurement of the IBIs [20]. IBIs were measured from the time of electrode contact until 2 hours after awakening, such that approximately 45,000 IBIs were recorded for each animal. Following the recording period, animals were briefly anesthetized again to enable removal of the electrodes and harness. Soon after completion of the HRV study, animals were decapitated. Brain lipid composition was determined using a lipid extraction method modified after Schwertner [21], and fatty acid analysis as described by Salem et al. [22]. As HRV is primarily a measure of central function, cardiac tissue was not obtained.
Data analysis
Cardiac IBI data, in milliseconds, were retrieved from the Mini-Logger using Version 3.5 of the Mini-Logger software on an IBM Pentium II personal computer running Windows NT. The IBI data were filtered using linear interpolation if any single IBI was more than twice the magnitude of the previous IBI. In most animals, the amount of filtering required was less than 5% (range 0-10%; mean±SEM, 3.8±1.4%). There were no differences in filtering between the two experimental groups. Data was analyzed as previously reported [14, 23] using software incorporating an algorithm which extracts the fractal dimension D of the time-series, and derives the Hurst value as H=2–D. The software application used for these analyses, running on Windows NT, 95 or 98 (Microsoft, USA), is available and can be obtained through a procedure outlined at ftp://helix.nih.gov/pbdp. The value of the embedding constant used for the analysis was estimated by increasing the embedding dimension in increments of 1 until D reached a stable value. Empirical testing showed that a dimension of 12 resulted in a stable measure of D for all time-series tested. H was determined for one continuous 4000 beat sequence (approximately 10 minutes), which commenced 90 minutes after the animal awoke. We chose this period to enable maximal recovery from anesthesia, while still maintaining a good surface electrode contact. Analysis of variance (ANOVA) and the Mann-Whitney sum of ranks were calculated (Sigmastat, SPSS) to determine whether there were differences in the values of H or IBI magnitude between animals raised on the n-3 ADQ versus n-3 DEF diets.
Results
The r2-value for the linear regressions used for the calculation of H values were all 1.00, suggesting that a strong scale-dependent power law relationship [24] applies to the hooded rat IBI time-series. The relationship of the IBI time-series and the calculated H parameter is shown for one animal in Figure 2. Heart rate was observed to increase (i.e. a decrease in IBI) as the animal recovered from anesthesia, as shown in Figure 2. The observed effects of isoflurane on heart rate variability are not discussed in this report, and will be presented elsewhere. As shown in Table III, despite the profound influence of the diets on the level of DHA in the brain, the value for H was not affected. However, the power of the ANOVA performed was far lower (0.05) than that considered acceptable to avoid a Type II error (i.e. 0.8). The sample size required to adequately detect a significant difference was calculated assuming a Normal distribution, equal variances and a two-sided alternate hypothesis using Equation 1, in which n is the sample size required in each group to detect a difference in mean H-value of D (with a=0.05 and power (1–b)=0.8). sADQ and sDEF are the standard deviations for measures of H in the two diet groups. It was revealed that approximately 200 animals (100 from each dietary group) would have been required for this study, in which D represented a difference of approximately 20%.
Discussion
Changes in heart rate are primarily determined by autonomic innervation, whereby parasympathetic activation slows heart rate and sympathetic tone increases it [25]. Hence, the moment-to-moment variability of heart rate is an index of autonomic nervous system function. One likely mediator of HRV control is the type-3 serotonin receptor (5-HT3), since pharmacological manipulations of this receptor’s activity profoundly alter HRV [26]. Pronounced changes in HRV have been found in various disease states, most notably, alcoholism [23]. Moreover, assessment of HRV provides a means of stratifying risk for sudden cardiac death, or death following acute MI [27, 28]. Given the importance of HRV as both an index of neural integrity, and as a prognostic indicator in clinical medicine, it is not surprising that the study of HRV generators and the factors which may potentially modulate them has attracted considerable research interest [1, 2]. The present study assessed the effects of extremes of tissue omega-3 PUFA on HRV, with the neural omega-3 PUFA levels of those fed the omega-3 deficient diet less than one-third that of those fed the omega-3 adequate diet. Despite these large differences, we failed to find a significant effect of dietary omega-3 fatty acids on HRV (Table III). Furthermore, we did not find an effect on heart rate, though others have reported a decrease in HR with omega-3 supplementation [8, 29, 30]. Moreover, the study would have required a large number of animals to avoid a type II error. Our results differ from some reports on humans, following supplementation with dietary omega-3 fatty acids. The discrepancies may be explained by several differences between this study and the numerous human clinical trials, though aside from one clinical trial on survivors of myocardial infarctions, positive associations between omega-3 fatty acids and higher HRV were only found after post hoc analyses, which dichotomized the treatment groups by plasma omega-3 concentrations [3-6, 31, 32]. The conclusions drawn, though informative, are not definitive for several reasons. Firstly it is well known that blood PUFA profile does not necessarily reflect neural accretion levels [33, 34], unless the level of consumption has been maintained for some months. Rather, blood cell PUFA profile is a reflection of the current (or recent past) level of dietary fatty acid consumption. If HRV was correlated with blood PUFA profile, one would also expect a significant change with omega-3 supplementation in an interventional prospective study. Hence, the conclusions made following dichotomization are inconsistent with the failure to find a treatment effect. Secondly, the reported increases in the standard deviation of RR interval (SDNN), as an indicator of increased HRV following dietary omega-3 supplementation, did not consider a parallel increase in mean RR interval. Since the SDNN is intrinsically related to the mean RR interval, increases in the mean will necessarily increase the SDNN, thereby confounding the interpretation of the results. Additionally, our failure on finding a treatment effect may be due to species differences in comparing rodents and humans, and that our study did not directly use dietary eicosapentaenoic acid, whereas the human trials did. Furthermore, the experimental models differed in that these rats were omega-3 deficient throughout neurodevelopment while the human subjects were supplemented as adults. Though ketamine is typically used in small animal studies, it has been shown to exert marked effects on N-methyl-D-aspartate (NMDA) receptors, which, in turn, have been implicated in reflex control of cardiac rhythm [26, 35, 36]. In contrast, isoflurane is believed to act on g-amino-butyric acid (GABA) receptors, though its mechanism of action is unclear. Furthermore, since isoflurane has very fast transport clearance [37] and is not significantly metabolized [38], we believe that our measures of HRV were not significantly affected by anesthesia. Indeed, though several reports have demonstrated generalized power reductions for frequency-domain measures of heart rate [39, 40], these do not appear to persist following recovery [41].
Conclusions
We conclude that, according to this model of dietary PUFA manipulation, omega-3 fatty acids are unlikely to exert a large effect on the autonomic functions that control heart rate variability. Prospective studies into the effect of omega-3 fatty acids on HRV should consider the need for large sample size as estimated by the results contained in this report.
Acknowledgments
The authors thank Dr. Lee Chedester, Dr. Raouf Kechrid (logistics and animal care), Ms. Toni Calzone, Mr. James Loewke (technical support), and the support of the Australian National Health and Medical Research (NHMRC) and Australian Research (ARC) Councils (HSW).
References
1. Kiviniemi AM, Tulppo MP, Wichterle D, et al. Novel spectral indexes of heart rate variability as predictors of sudden and non-sudden cardiac death after an acute myocardial infarction. Ann Med 2007; 39: 54-62. 2. Chattipakorn N, Incharoen T, Kanlop N, Chattipakorn S. Heart rate variability in myocardial infarction and heart failure. Int J Cardiol 2007; 120: 289-96. 3. 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: 331-7. 4. Christensen JH, Aarøe 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: 102-6. 5. Christensen JH, Gustenhoff P, Korup E, et al. Effect of fish oil on heart rate variability in survivors of myocardial infarction: a double blind randomised controlled trial. BMJ 1996; 312: 677-8. 6. Christensen JH, Gustenhoff P, Korup E, et al. n-3 polyunsaturated fatty acids, heart rate variability and ventricular arrhythmias in post-AMI-patients. A clinical controlled trial [Danish]. Ugeskr Laeger 1997; 159: 5525-9. 7. Villa B, Calabresi L, Chiesa G, Risè P, Galli C, Sirtori CR. Omega-3 fatty acid ethyl esters increase heart rate variability in patients with coronary disease. Pharmacol Res 2002; 45: 475. 8. O’Keefe JH, Abuissa H, Sastre A, Steinhaus DM, Harris WS. Effects of omega-3 fatty acids on resting heart rate, heart rate recovery after exercise, and heart rate variability in men with healed myocardial infarctions and depressed ejection fractions. Am J Cardiol 2006; 97: 1127-30. 9. Kleiger RE, Stein PK, Bosner MS, Rottman JN. Time domain measurements of heart rate variability. Cardiol Clin 1992; 10: 487-98. 10. Stein PK, Bosner MS, Kleiger RE, Conger BM. Heart rate variability: a measure of cardiac autonomic tone. Am. Heart J 1994; 127: 1376-81. 11. Hurst HE. Long-term storage capacity of reservoirs. Trans Am Soc Civ Engrs 1951; 116: 770-808. 12. Mandelbrot BB, Van Ness JW. Fractional Brownian motions, fractional noises, and applications. SIAM Rev 1968; 10: 422-37. 13. Peng CK, Havlin S, Hausdorff JM, Mietus JE, Stanley HE, Goldberger AL. Fractal mechanisms and heart rate dynamics. Long-range correlations and their breakdown with disease. J. Electrocardiol 1995; 28 (Suppl.): 59-65. 14. DePetrillo PB, Speers D, Ruttimann UE. Determining the Hurst exponent of fractal time series and its application to electrocardiographic analysis. Comput Biol Med 1999; 29: 393-406. 15. Reeves PG. Nielsen FH, Fahey GC. AIN-93 purified diets for laboratory rodents: final report of the American Institute of Nutrition ad hoc writing committee on the reformulation of the AIN-76A rodent diet. J Nutr 1993; 123: 1939-51. 16. Leat WM. Retinal function in rats and guinea-pigs reared on diets low in essential fatty acids and supplemented with linoleic or linolenic acids. Ann Nutr Metab 1986; 30: 166-74. 17. Moriguchi T, Greiner RS, Salem N. Behavioral deficits associated with dietary induction of decreased brain docosahexaenoic acid concentration. J Neurochem 2000; 75: 2563-73. 18. Wainwright PE, Huang YS, Coscina DV, Lévesque S, McCutcheon D. Brain and behavioral effects of dietary n-3 deficiency in mice: a three generational study. Dev Psychobiol 1994; 27: 467-87. 19. Weisinger HS, Vingrys AJ, Sinclair AJ. Dietary manipulation of long-chain polyunsaturated fatty acids in the retina and brain of guinea pigs. Lipids 1995; 30: 471-3. 20. Ruha A, Sallinen S, Nissila S. A real-time microprocessor QRS detector system with a 1-ms timing accuracy for the measurement of ambulatory HRV. IEEE Trans Biomed Eng 1997; 44: 159-67. 21. Schwertner HA, Mosser EL. Comparison of lipid fatty acids on a concentration basis vs weight percentage basis in patients with and without coronary artery disease or diabetes. Clin Chem 1993; 39: 659-63. 22. Salem N, Reyzer M, Karanian J. Losses of arachidonic acid in rat liver after alcohol inhalation. Lipids 1996; 31 (Suppl.): S153-6. 23. DePetrillo PB, White KV, Liu M, Hommer D, Goldman D. Effects of alcohol use and gender on the dynamics of EKG time-series data. Alcohol Clin Exp Res 1999; 23: 745-50. 24. Peitgen HO, Jurgens H, Saupe D. Chaos and Fractals: New Frontiers of Science. Springer-Verlag, New York 1992. 25. Hainsworth R. The control and physiological importance of heart rate. In: Heart rate variability. Malik M, Camm AJ (eds). Futura Publishing Co, New York 1995: 3-20. 26. DePetrillo PB, Bennett AJ, Speers D, et al. Ondansetron modulates pharmacodynamic effects of ketamine on electrocardiographic signals in rhesus monkeys. Eur J Pharmacol 2000; 391: 113-9. 27. Turner A, Malik M, Camm AJ. Autonomic function following myocardial infarction. Br J Hosp Med 1994; 51: 89-96. 28. Voss A, Hnatkova K, Wessel N, et al. Multiparametric analysis of heart rate variability used for risk stratification among survivors of acute myocardial infarction. Pacing Clin Electrophysiol 1998; 21: 186-92. 29. Harris WS, Gonzales M, Laney N, Sastre A, Borkon AM. Effects of omega-3 fatty acids on heart rate in cardiac transplant recipients. Am J Cardiol 2006; 98: 1393-5. 30. Mozaffarian D, Prineas RJ, Stein PK, Siscovick DS. Dietary fish and n-3 fatty acid intake and cardiac electrocardiographic parameters in humans. J Am Coll Cardiol 2006; 48: 478-84. 31. Christensen JH, Dyerberg J, Schmidt EB. n-3 fatty acids and the risk of sudden cardiac death assessed by 24-hour heart rate variability. Lipids 1999; 34 (Suppl): S197. 32. Christensen JH, Korup E, Aarøe J, et al. Fish consumption, n-3 fatty acids in cell membranes, and heart rate variability in survivors of myocardial infarction with left ventricular dysfunction. Am J Cardiol 1997; 79: 1670-3. 33. Rioux FM, Innis SM, Dyer R, MacKinnon M. Diet-induced changes in liver and bile but not brain fatty acids can be predicted from differences in plasma phospholipid fatty acids in formula- and milk-fed piglets. J Nutr 1997; 127: 370-7. 34. Salem N. Omega-3 fatty acids: molecular and biochemical aspects. In: New Protective Roles for Selected Nutrients. Spiller GA, Scala J (eds), Alan R. Liss, New York 1989: 127. 35. Chianca DA, Machado BH. Microinjection of NMDA antagonist into the NTS of conscious rats blocks the Bezold-Jarisch reflex. Brain Res 1996; 718: 185-8. 36. Lo WC, Lin HC, Ger LP, Tung CS, Tseng CJ. Cardiovascular effects of nitric oxide and N-methyl-D-aspartate receptors in the nucleus tractus solitarii of rats. Hypertension 1997; 30: 1499-503. 37. Wissing H, Kuhn I, Rietbrock S, Fuhr U. Pharmacokinetics of inhaled anaesthetics in a clinical setting: comparison of desflurane, isoflurane and sevoflurane. Br J Anaesth 2000; 84: 443-9. 38. Wenker O. Review of Currently Used Inhalation Anesthetics. Part I. http://www.ispub.com/journals/IJA/Vol3N2/inhal1.htm. 39. Galletly DC, Westenberg AM, Robinson BJ, Corfiatis T. Effect of halothane, isoflurane and fentanyl on spectral components of heart rate variability. Br J Anaesth 1994; 72: 177-80. 40. Kato M, Komatsu T, Kimura T, Sugiyama F, Nakashima K, Shimada Y. Spectral analysis of heart rate variability during isoflurane anesthesia. Anesthesiology 1992; 77: 669-74; erratum in Anesthesiology 1993; 78: 224. 41. Widmark C, Olaison J, Reftel B, Jonsson LE, Lindecrantz K. Spectral analysis of heart rate variability during desflurane and isoflurane anaesthesia in patients undergoing arthroscopy. Acta Anaesthesiol Scand 1998; 42: 204-10.