Correspondence
Fish Consumption and the Risk of Myocardial Infarction
N Engl J Med 1997; 337:497-499August 14, 1997
- Article
Our primary concern is with the reliance on information ascertained from death certificates to determine the timing of death and the specific underlying cause. The rate of misclassification is high when this information is not supplemented with data from hospital records, autopsies, or interviews with next of kin. Kuller et al.4 reported that even when the probability of sudden death as ascertained from the death certificate was high, half of deaths were classified as nonsudden after all available information was considered. Death certificates are also less accurate than autopsies with respect to the determination of the specific cause within the broader categories of cardiovascular disease.5 Therefore, we question the focus on deaths due to myocardial infarction by Daviglus et al. for two reasons. First, this categorization may not have represented a distinct pathologic entity. Second, bias could occur if physicians were more likely to categorize a “quick death” than a slower death as due to myocardial infarction. Such bias could, at least in theory, provide an alternative explanation for the authors' findings, especially if the data on the timing of the deaths were inaccurate.
Finally, the definition of “sudden” death as “death occurring no more than 12 hours after the onset of the terminal acute illness [or that] occurring out of the hospital” is at variance with the more standard definition of death within 1 hour from the onset of symptoms. The chief hypothesized mechanism by which fish might reduce the rate of sudden death from cardiac causes is through the possible antiarrhythmic effects of n-3 fatty acids. Thus, the “one-hour” definition of sudden death provides a higher degree of specificity for death from arrhythmia in a general population.6 As the time from the onset of symptoms lengthens, the percentage of deaths due to circulatory failure rather than to arrhythmia increases.6
If there are data that validate the death certificate as a source of information on the timing of sudden death, then the results of the analysis of Daviglus et al. may change if the entire category of deaths from coronary heart disease is classified as sudden or nonsudden on the basis of the one-hour definition. In any case, we agree with the authors that more data on these questions are needed before firm conclusions can be drawn.
7 ReferencesChristine M. Albert, M.D.
JoAnn E. Manson, M.D.
Charles H. Hennekens, M.D.
Brigham and Women's Hospital, Boston, MA 02215-1204Jeremy N. Ruskin, M.D.
Massachusetts General Hospital, Boston, MA 02114To the Editor: Daviglus et al. (April 10 issue) report an association between fish consumption in low amounts and a reduced risk of nonsudden death from myocardial infarction, but not of sudden death. The totality of the evidence, which includes cogent biologic mechanisms, supports a benefit of fish consumption in terms of sudden death from cardiac causes. Their findings are inconsistent with those of a case–control study as well as of our own cohort study. The documentation and definition of sudden death may account for the apparently discrepant findings.
1
Daviglus ML ,Stamler J ,Orencia AJ , et al. Fish consumption and the 30-year risk of fatal myocardial infarction. N Engl J Med 1997;336:1046-1053
Full Text | Web of Science | Medline2
Siscovick DS ,Raghunathan TE ,King I , et al. Dietary intake and cell membrane levels of long-chain n-3 polyunsaturated fatty acids and the risk of primary cardiac arrest. JAMA 1995;274:1363-1367
CrossRef | Web of Science | Medline3
Albert CM ,Manson JE ,O'Donnell CJ ,Ajani UA ,Hennekens CH . Fish consumption and the risk of sudden death in the Physicians' Health Study. Circulation 1996;94:Suppl I:I-578 abstract.4
Kuller L ,Lilienfeld A ,Fisher R . An epidemiological study of sudden and unexpected deaths in adults. Medicine (Baltimore) 1967;46:341-361
CrossRef | Web of Science | Medline5
Kircher T ,Nelson J ,Burdo H . The autopsy as a measure of accuracy of the death certificate. N Engl J Med 1985;313:1263-1269
Full Text | Web of Science | Medline6
Hinkle LE Jr ,Thaler HT . Clinical classification of cardiac deaths. Circulation 1982;65:457-464
CrossRef | Web of Science | Medline
To the Editor:
Daviglus et al. suggest that the inverse relation between fish consumption and the 30-year risk of fatal myocardial infarction may have been related to favorable effects of fish on levels of total and intermediate-density lipoprotein triglycerides, alterations in the fatty-acid content of cell membranes, or an independent effect of fish protein or a component thereof. An additional factor may be that fish displaces from the diet foods containing relatively high levels of saturated fat. Although this is not apparent from the data, the variability in assessing and quantitating dietary fat intake inherent in the methods available may have been great enough to preclude the detection of such a shift.
Alice H. Lichtenstein, D.Sc.
Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA 02111To the Editor:
Daviglus and colleagues have demonstrated a persuasive association between fish consumption and the risk of nonsudden myocardial infarction; their study adds to the evidence of causality and spurs the need for personal and public health action in changing eating habits.
The group eating 35 g of fish or more per day had an energy intake 416 kcal higher than the group eating no fish, yet the two groups were almost identical in age, weight, height, and body-mass index. How can this surprising finding be explained?
Trying to explain the inconsistencies among studies of this matter, Daviglus et al. focus on differences in study design and measurement. Another explanation that merits attention is that the nutritional composition of the fish eaten differs in different times, places, and populations. The composition of fish is also known to vary according to the season and place of catch. There is evidence from five of six studies with relevant data, reviewed by George and Bhopal,1 that farmed fish contain more total fat than wild fish. The amount and relative distribution of fatty acids also vary; for example, the ratio of n-3 to n-6 fatty acids is higher in wild fish than in farmed fish.
Since Daviglus et al. measured fish consumption some 30 years ago, the demand for seafood has seriously threatened stocks of wild fish and fish farming has become a huge industry. The health benefits conferred by farmed and wild fish are likely to differ. This fact should be considered a possible explanation for differences among past studies with similar designs. It needs to be incorporated into future research and is potentially relevant to public health messages advocating more eating of fish.
1 ReferencesRaj Bhopal, M.D.
University of North Carolina, Chapel Hill, NC 27599-74001
George R ,Bhopal R . Fat composition of free living and farmed seas species: implications for human diet and sea-farming techniques. Br Food J 1995;97:19-22
CrossRef
Author/Editor Response
The authors reply:
To the Editor: In response to Bhopal, a reasonable inference is that the higher intake of energy in consumers of 35 g or more of fish per day reflected greater physical activity. Differences in lipid composition between wild and farmed fish merit attention. When our study began in 1957–1959, only wild fish were available.
In response to Lichtenstein, dietary lipids — and, correspondingly, serum cholesterol concentrations — were similar across the four strata of fish consumption; we controlled for these variables in regression analyses. Given the in-depth assessment of usual eating patterns, it is unlikely that extensive misclassification of the intake of these dietary lipids occurred differentially among the strata.
In response to Albert et al., our paper briefly noted that when four definitions of sudden death coded to code 410 of the International Classification of Diseases, Eighth Revision (ICD-8) (acute myocardial infarction) were used, fish intake was not inversely related to that outcome.1 There were 94 instantaneous deaths (relative risk for consumers of 35 g or more of fish per day as compared with nonconsumers, 1.63; 95 percent confidence interval, 0.64 to 4.15; P for trend, 0.408). There were 137 deaths within one hour or less (relative risk, 1.08; 95 percent confidence interval, 0.52 to 2.25; P for trend, 0.882). There were 148 deaths within three hours or less (relative risk, 1.24; 95 percent confidence interval, 0.61 to 2.54; P for trend, 0.865). The data on deaths in 12 hours or less were as published.1 The results were similar for the relation between fish intake and sudden death from coronary heart disease (ICD-8 codes 410 through 414). These data are robust evidence to substantiate that there was no inverse relation between fish consumption and sudden death in our study. We disagree that the findings relating fish intake and sudden death could be biased by any hypothetical tendency among physicians to categorize “quick death” as due to myocardial infarction. Most “quick deaths” among middle-aged and older U.S. men are caused by major coronary events.2
The issue discussed here is secondary; the primary unresolved issue is whether fish intake protects against myocardial infarction, coronary heart disease, or both. Is the inverse relation that we and others have observed etiologically significant? We remain guarded on this point.1,3 No data are available on the relation between fish intake and severe coronary atherosclerosis, the condition underlying most cases of coronary heart disease, of all forms of clinical expression, including electrical and thrombotic sudden death. Nor are data in humans available to support the hypothesis that small amounts of n-3 fatty acids ingested with 50 g per day of fish (lean and fatty) have “antiarrhythmic effects.” Pending the resolution of the “cardioprotection” issue (trials seem necessary), it is reasonable to recommend moderate consumption of properly prepared fish as contributing to a comprehensive pattern of improved nutrition.4
4 ReferencesMartha L. Daviglus, M.D., Ph.D.
Jeremiah Stamler, M.D.
Philip Greenland, M.D.
Northwestern University Medical School, Chicago, IL 60611-44021
Daviglus ML ,Stamler J ,Orencia AJ , et al. Fish consumption and the 30-year risk of fatal myocardial infarction. N Engl J Med 1997;336:1046-1053
Full Text | Web of Science | Medline2
The Pooling Project Research Group
CrossRef | Medline3
Shekelle RB ,Stamler J . Fish and coronary heart disease: the epidemiologic evidence. Nutr Metab Cardiovasc Dis 1993;3:46-514
National Research Council. Diet and health: implications for reducing chronic disease risk. Washington, D.C.: National Academy Press, 1989.
- Citing Articles (2)
