Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2682

DEPARTMENT OF HEALTH AND proposed general requirements for claims based on the standard that FDA HUMAN SERVICES health claims for food (November 27, is establishing for determining the 1991, 56 FR 60537). As amended in reliability of health claims under section Food and Drug Administration 1990, the Federal Food, Drug, and 403(r)(5)(D) of the act. In the November 21 CFR Part 101 Cosmetic Act (the act) provides that a 27,1991, proposal on general [Docket No. 91N-0103] food is misbranded if it bears a claim requirements for health claims, FDA RIN 0905-AB67 that characterizes the relationship of a proposed to adopt the standard that the nutrient to a disease or health-related 1990 amendments provide for Food Labeling: Health Claims and condition unless the claim is made in conventional foods, which is set forth in Label Statements: Omega-3 Fatty accordance with section 403(r)(3) or section 403(r)(3)(B)(i) of the act, as the Acids and Coronary Heart Disease (r)(5)(D) of the act (21 U.S.C. 343(r)(3) or standard for dietary supplements; Given (r)(5)(D)). this fact, and the fact that omega-3 fatty AGENCY: Food and Drug Administration, Congress enacted the health claims acids are found in numerous HHS. provisions of the 1990 amendments to conventional foods as well as in dietary help U.S. consumers maintain good supplements, FDA broadened its ACTION: Final rule. health through appropriate dietary inquiry to a determination as to whether it should grant a health claim on omega- patterns and to protect consumers from unfounded health claims. Section 3 fatty acids and CHD for any foods. SUMMARY: The Food and Drug 3(b)(1)(A) of the 1990 amendments Because the DS Act provides that FDA Administration (FDA) is announcing its specifically requires the agency to may grant claims using the significant decision not to authorize the use on the determine whether health claims for 10 scientific agreement standard specified label or labeling of foods of health nutrient-disease relationships meet the in section 403(r)(3)(B)(i) of the act, and claims relating to an association requirements of section 403(r)(3) or given the breadth of FDA’s November between omega-3 fatty acids and (r)(5)(D) of the act. The relationship of 1991, proposal on omega-3 fatty acids, coronary heart disease (CHD). The omega-3 fatty acids and heart disease is FDA has decided to move forward to agency has determined, based on: (1) one of the claims required to be determine whether it can authorize a The totality of the publicly available evaluated. In the Federal Register of claim under section 403(r)(3)(B)(i) for scientific evidence; and (2) the agency’s March 28, 1991 (56 FR 12932). FDA omega-3 fatty acids and CHD. review of comments received in published a notice requesting scientific However, this rule does not apply to response to its November 27, 1991, data and information on the 10 specific dietary supplements. While a proposed rule on omega-3 fatty acids topic areas identified in the 1990 manufacturer of a dietary supplement and CHD, including scientific amendments. Relevant scientific studies can make a claim on omega-3 fatty acids information included in those and data received in response to this and CHD without rendering its product comments, that there is not significant request were considered as part of the misbranded under section 403(r)(1)(B) scientific agreement among experts that agency’s review of the scientific of the act, the manufacturer should such evidence supports a health claim literature on omega-3 fatty acids and assure itself that the making of the claim for omega-3 fatty acids and CHD. CHD and were included in the proposed will not misbrand the product under Further, FDA has determined that the rule. section 403(a). new information does not change the In addition, on January 30 and 31, conclusions that the agency reached on 1992, FDA held public hearings on all II. Summary of Comments and the the basis of the information reviewed in aspects of the proposed rules (57 FR Agency’s Response its proposal. Therefore, FDA has 239). FDA received 80 letters, each concluded that such a claim is not FDA requested in the Federal Register containing one or more comments, from justified. This action is in response to of November 27,1991 (56 FR 60663), consumers, health care professionals, provisions of the Nutrition Labeling and written comments in response to its universities and research institutes, Education Act of 1990 (the 1990 proposed rule, FDA reviewed all of the health profession associations, amendments) that bear on health claims. comments it received, including new consumer advocacy organizations, State and is developed in accordance with the data submitted in the comments, and and local governments, foreign final rule on general requirements for scientific articles referred to in the governments, trade organizations, health claims, published elsewhere in comments. FDA also reviewed industry, and professional this issue of the Federal Register. additional scientific articles, reviews, organizations. In addition to these EFFECTIVE DATE: May 8, 1993. and recommendations published from comments, the agency also considered FOR FURTHER INFORMATION CONTACT: John August 1991 through February 1992. statements made on omega-3 fatty acids C. Wallingford, Center for Food Safety The Dietary Supplement Act of 1992 and CHD at the January 30 and 31,1992, and Applied Nutrition (HFS-465) Food (DS Act) established a moratorium on public hearings. Some of the comments and Drug Administration, 200 C St. SW., the implementation of the 1990 agreed with one or more of the aspects Washington, DC 20204, 202-205-5461. amendments with respect to dietary of the proposed rule, without providing SUPPLEMENTARY INFORMATION: supplements. The DS Act says that FDA further grounds for support other than can grant health claims for food, those provided by FDA in the preamble I. Background including dietary supplements, under to the proposal. Other comments section 403(r)(3)(B)(i) of the act. disagreed with one or more aspect of the In the Federal Register of November However, it may not act on such claims proposal without providing specific 27, 1991 (56 FR 60663), FDA proposed under section 403(r)(5)(D) of the act grounds for the disagreement. A few not to authorize a health claim relating until it establishes a standard to comments addressed issues outside of diets high in omega-3 fatty acids to the scope of this document and will not reduced risk of heart disease. The implement that section of the act, which the DS Act says may not occur until be discussed in this document. Most of proposed rule was issued in response to December 1993. Section 3(b)(1)(A)(x) of the comments provided specific provisions of the 1990 amendments grounds in support of their positions (Pub. L. 101-535) that bear on health the 1990 amendments directs the agency concerning aspects of this health claim claims and in accordance with the to evaluate the omega-3 fatty acids/CHD

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2683 as proposed. The agency has converted in the body to EPA and DHA that interested persons may petition summarized and addressed the issues (Ref.100). Therefore, FDA believes it FDA under § 101.70 (21 CFR 101.70) to raised in the sections of this document has represented the potential nutrient- issue a regulation regarding a health that follow. disease relationship appropriately by claim that relates these ratios to the risk limiting its attention to EPA and DHA. of CHD. A. General Comments 3. Two comments stated that FDA’s 5. Another comment pointed out that 1. Definition of omega-3 fatty adds and position on fish as opposed to the supplements used currently have composition of omega-3 fatty acid omega-3 fatty acids in fish was a contained various amounts of short- and supplements tautology, because: “if polyunsaturated long-chain omega-3 fatty acids and that 1. One comment criticized the fatty acids have beneficial effects on many supplements also contain definition ofomega-3 fatty acids used in CHD, and if fish oils are a member of saturated fat. The comment stated that the proposed rule, on the basis that this class of fatty acids, it should not be some of the discrepancies in reported omega-3 fatty acids were not counted against their beneficial effects findings may be due to the type of distinguished from other on CHD.” supplement used. polyunsaturated fatty acids (PUFA’s). FDA disagrees with this comment. FDA agrees that numerous In the proposed rule, FDA limited the FDA considers the claim for omega-3 supplements varying in fatty acid term omega-3 fatty acids to fatty acids to reflect the unique composition have been used, and that eicosapentaenoic acid ((EPA), 20 biochemistry of these fatty acids. In the variation in the fatty acid carbons, 5 double bonds) and particular, the prevailing theory about composition of supplements may have docosahexaenoic acid ((DHA), 22 the mode of action of omega-3 fatty influenced the outcome. FDA carbons, 6 double bonds) (56 FR 60663 acids is that they compete with omega- reexamined the studies cited in its at 60664). FDA noted that most of the 6 fatty acids (fatty acids with their first proposal and the new data submissions relevant research has used fish or fish double bond at the sixth carbon from for evidence that the nature of the oils rich in these two fatty acids. the methyl end, and which comprise the supplement used was related to the FDA acknowledges that its statement largest amount of dietary PUFA’s). outcome. However, the agency found Defining omega-3 fatty acids did not Thus, a clear separation of effects of that the same results are observed explicitly refer to omega-3 -fatty acids. omega-3 fatty acids from effects of other regardless of the source of omega-3 fatty The sentence: “Their unique (primarily omega-6) PUFA’s is needed acids. For example, in eight well- characteristic is the location of the first to support a claim. designed studies cited in the proposal double bond, which occurs at the third 4. One comment stated that FDA did on the total serum cholesterol response carbon from the methyl (or omega) end not consider the importance of the ratio among normal subjects (Refs., 6, 9,14, of the fatty acid.” (56 FR 60663 at of omega-3 fatty acids to arachidonic 49, 54, 73, 156, and 166,)six different 60664) was intended to refer to omega- acid (AA), a 20 carbon omega-6 PUFA sources of omega-3 fatty acids were 3 fatty acids. This definition with four double bonds, and stated that used: Salmon oil, SuperEPA, MaxEPA, distinguishes omega-3 fatty acids from evidence exists for a relationship a fish oil triglyceride, Promega, and other PUFA’s, which have their first, between the saturated fat:unsaturated fat mackerel paste. None of these unsaturation at the sixth or ninth carbon ratio in the diet and the omeg-6:omega- supplements produced a change in total from the omega end of the fatly acid. 3 fatty acid ratio in the diet and the risk serum cholesterol. Similarly, four 2. One comment argued that the of CHD. different sources of omega-3 fatty acids definition of omega-3 fatty acids should FDA considers concerns about the (fresh water fish, salmon oil, purified include land-based.(primarily plant) ratio of AA to omega-3 fatty acids and EPA, MaxEPA) were shouwn in seven omega-3 fatty acids (i.e., linolenic acid). the ratio of omega-6 fatty acids to Well-designed studies to reduce platelet (For the purposes of this document, the omega-3 fatly acids to be reasonable in aggregation in normal subjects (Refs. 2, term linolenic acid is used to indicate view of the competition between these 6, 24, 54, 96, 143, and 166). the omega-3 fatty acid, alpha linolenic classes of fatty acids in human FDA did note that some differences in acid. In contrast, gamma linolenic acid biochemistry. FDA considered all types response have been produced by has its first double bond at the sixth of foods and supplements used to supplements that vary in ratio of EPA to carbon from the omega end of the fatty provide omaga-3 fatty acids in its DHA. For example, one fish oil (pollock acid, and is not an ornega-3 fatty acid.) evaluation of the claim. Although the oil) with a high EPA :DHA ratio FDA disagrees with this comment. AA content of the supplement was often increased low-density lipoprotein (LDL) reported, studies did not report data for FDA defined omega-3 fatty acids as EPA cholesterol, LDL triglyceride and total dietary AA. FDA is aware of only and DHA, primarily as a functional apoprotein B (apoB) (a protein very limited data regarding the ratios of definition derived from the scientific component of LDL) in comparison to a AA to omega-3 fatty acids, and of the butter-rich diet, but two fish oils with a literature. The hypothesis for a omega-6 fatty acid:omega-3 fatty acid relationship between omega-3 fatty low EPA:DHA ratio (tuna oil, salmon oil ration in the diet and the risk of CHD. acids and CHD derived from with added palmitic acid) reduced apoB Therefore, it: is not possible for FDA to and LDL cholesterol, and increased LDL correlations between low rates of CHD draw any conclusions about these ratios triglyceride to a smaller extent than the and high consumption of fish oils. and their possible modifying effects on pollock oil in comparison to the butter- Similarly, most of the intervention the omega-3farry acids CHD health rich diet ( Ref. 17). However, the effects studies have used fish oil or fish as a claim. of the two major oniega-3 fatty acids source of EPA and DHA, not plant oils However, because the fish oils used have not yet been systematically rich in linolenic acid. The comment contained high concentrations of omega- investigated. FDA recognizes that provided no evidence that linolenic acid 3 fatty acids, FDA believes that the purified EPA and DHA are now has biochemical effects comparable to amounts of fish oils supplemented to available for research; such supplements EPA or DHA, nor has FDA found any the various test diets would have will enable the study of the individual evidence of a relationship between affected the AA:omega-3 fatty acid ratio effects of these fatty acids. linolenic acid and CHD. Moreover, only and the omega-6:omega-3 fatty acid ratio 6. One comment stated that a limited amount of linolenic acid is of the diets to some extent. FDA advises conservation of omega-3 fatty acids in

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2684 the body calls into question the additional clinical trials will be done data were no stronger than the data importance of the amounts of omega-3 due to their expense, and that, therefore, supporting the link between omega-3 fatty acids used in scientific studies. FDA should rely more heavily on fatty acids and CHD. The comments However, the comment did not suggest epidemiologic studies, animal studies, asserted that, by basing its decision on any alternate method to describe intake. and biochemical and physiological the relationship between the nutrient FDA recognizes that fish is not interventions that -suggest an effect of and a surrogate marker for the disease, ordinarily consumed daily. However, omega-3 fatty acids on risk of CHD. or for a susceptible subpopulation, FDA the 1990 amendments require that FDA has no basis upon which to agree held other claims to a less restrictive health claims on foods be stated in such or disagree with the comments’ standard. One comment stated: “The a way as to enable the public to assertion that further clinical trials are FDA statement is internally consistent understand the relative significance of unlikely. FDA disagrees that its in denying health claims for omega-3 such information in the context of a emphasis on clinical trials was fatty acids, but this is only in the total daily diet (section 403(r)(3)(B)(iii) misplaced. Because the 1990 context of holding these food of the act). Thus, a reasonable estimate amendments addressed nutrient-disease components to essentially impossible of daily dietary intake of omega-3 fatty relationships, FDA considered human standards not required for other, acids is needed when assessing the studies that used CHD as the endpoint allowable, claims.” relationship between omega-3 fatty to be the most directly relevant studies, Specific comparisons were made to acids and the risk of CHD. Most of the although these studies do not the proposed claims on fat and CHD, fat studies reviewed by the agency used demonstrate that the effects are and cancer, calcium and osteoporosis, daily supplementation with a known specifically due to omega-3 fatty acids. and sodium and hypertension. Other amount of omega-3 fatty acids, but Human studies in which a surrogate comments indicated that qualified others estimated intake of omega-3 fatty marker for CHD risk was measured as claims, such as that for calcium and acids from foods consumed in the dally the endpoint of the treatment were also osteoporosis, were appropriate models diet. Both types of intake estimates are considered carefully. The advantage of for the claim relating omega-3 fatly important. Daily supplementation is these studies is that they are able to acids to CHD. useful to relate changes to a carefully demonstrate specificity of the effects FDA disagrees with these comments. controlled amount of omega-3 fatty due to omega-3 fatty acids. However, the FDA believes that for these other claims acids. The average daily intake of relationships between many of these there is significant scientific agreement omega-3 fatty acids in nonintervention surrogate markers and risk of CHD are among qualified experts regarding the studies provides a basis upon which to not well established, making it difficult relationship between the nutrient and determine whether the amounts of to relate changes in these endpoints the disease, whereas there is not such omega-3 fatty acids fed in brought about by omega-3 fatty acids to agreement regarding the relationship supplementation studies are reasonable the risk of disease. Biologic markers can between omega-3 fatty acids and CHD, in the context of the total daily diet serve as markers of a developing or between omega-3 fatty acids and disease, but the relevance of such agreed surrogate markers for risk of 2. Criteria used in evaluating studies evidence depends directly on the CHD. For example, based on the totality strength of the association between the of the publicly available scientific In the proposed rule, FDA listed some marker and the disease (Ref. 115). evidence, FDA determined that there is of the criteria used in evaluating FDA agrees that there are considerable epidemiological studies on the significant scientific agreement about additional data in animal studies, in relationship of omega-3 fatty acids to the role of calcium in maintaining bone CHD: (1) The reliability and accuracy of vitro studies, and biochemical and mineral density (the relationship of the the methods used in food intake physiological interventions regarding nutrient to the intermediate marker for the effects of omega-3 fatty acids. the disease), and about the relationship analysis and measurements of disease However, it is not clear that the results between peak (maximal) bone mass and endpoints, (2) the choice of control of such studies are relevant to the risk the risk of developing osteoporosis and subjects, (3) the representativeness of related bone fractures later in life (the the subjects, (4) the control of of human disease. Thus, FDA believes relationship between the intermediate confounding factors in data analysis, (5) that these other types of data are of marker and the disease itself) (see 56 FR the potential for misclassification of secondary importance compared to 60689; see also the final rule on calcium individuals with regard to dietary clinical data that measure either CHD exposure or disease endpoints, (6) the per se or established surrogate markers and osteoporosis published elsewhere presence of bias, and (7) the degree of for CHD. in this issue of the Federal Register). compliance and how compliance was However, in response to the Similarly, FDA relied on a long history assessed (56 FR 60667). comments, FDA has provided a more of Federal Government and other However, FDA stated that it thorough description of animal and in consensus statements to conclude that considered randomized, double-blind, vitro studies that suggest a role for there is significant scientific agreement placebo-controlled trials to be more omega-3. fatty acids in reducing the risk about the role of sodium as a causal valuable than other types of human of CHD, particularly with respect to the factor in hypertension for a segment of studies because they were less effects of omega-3 fatty acids on the the population. (See 56 FR 60825; see susceptible to bias, and because they development of atherosclerosis and with also the final rule on sodium and allowed inference about the specific respset to the responsiveness of blood hypertension, published elsewhere in effects of omega-3 fatty acids. Studies in vessels to ischemia (see comments 38 this issue of the Federal Register.) FDA which the endpoint was CHD, by and 49 and section II.C.3.a. of this also recognized the history of significant definition, provide the most persuasive document). scientific agreement about the type of evidence, but studies measuring 8. Many comments stated that the relationships between fat and cancer CHD to date have not provided the agency’s position on omega-3 fatty acid and between fat and CHD evidenced by specificity to show that the observed and CHD was inconsistent with its statements in reports issued by Federal effects were due to omega-3 fatty acids. position on other health claims, and Government and other authoritative 7. Some comments expressed the argued that for each of the four claims bodies. (See 56 FR 60764, 56 FR 60726; concern that it was unlikely that proposed to be allowed by FDA, the see also final rules on fat and cancer and

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2685

fat and CHD, published elsewhere in evidence published before 1988. The relationship between omega-3 fatty this issue of the Federal Register.). reports were also used as a way of acids and CHD. Placed in chronological Thus, these other nutrient-disease determining whether there was order, the concluding sections from the relationships have a history of being significant scientific agreement among cited review articles exemplify the lack recognized in Federal Government and qualified experts that the evidence of certainty as to the relationship authoritative reports, indicating supports a relationship between omega- between omega-3 fatty acids and risk of significant scientific agreement, whereas 3 fatty acids and CHD. The agency’s CHD. the relationship between omega-3 fatty reliance on these reports is consistent The review of the relationship acids and CHD has not been so with the 1990 amendments, which between omega-3 fatty acids and CHD recognized. For two of these other require the agency to consider reports by Leaf and Weber (Ref. 91) was nutrient-disease relationships, the data from authoritative scientific bodies of considered in the National Academy of relate to the disease itself, rather than to the United States in assessing health Sciences’ “Diet and Health: Implications markers for the disease. In the other claim petitions and to justify any for Reducing Chronic Disease Risk” two, calcium and osteoporosis and fat decision rejecting the conclusions of report (Ref. 115). FDA elected to include and CHD, there is significant scientific such reports (section 403(r)(4)(C) of the the Leaf and Weber review in its agreement that the dietary factors are act). citations because it covered, in the most related to surrogate markers for the Recognizing, however, that comprehensive manner of all available diseases, and that the surrogate markers considerable research had been reviews, the state of scientific are related to the diseases. published since these reports, and that knowledge about omega-3 fatty acids in There is significant scientific these reports had not been updated, CHD at the time the Federal agreement that serum cholesterol and FDA also reviewed the available studies Government and other comprehensive blood pressure are risk factors for CHD, on humans published since 1988. FDA reviews were published. Leaf and Weber as indicated by the emphasis on these relied on its own review of individual wrote: “Despite claims that n-3 fatty factors in Federal Government and other studies rather than review articles, acids can help prevent atherosclerosis, authoritative documents (Refs. 34 because review articles generally reflect recommendations to the public on diet through 36, 100, 115, and 169). Data the bias of the author and may not have been conservative; people have regarding the effects of omega-3 fatty consider the totality of the evidence. been advised to increase their acids on these endpoints have been FDA focused its independent review on consumption of fish by replacing two or carefully reviewed. However, the other primary papers published between three meals a week containing red meat endpoints measured in studies of the January 1988 and August 1991. Surveys with meals containing fish.” Their effects of omega-3 fatty acids, e.g., in and cross-sectional or prospective concluding sentence was: “If vitro platelet aggregation, various studies that were published before 1988 prospective double-blind, placebo- growth factors, fibrinogen, have not and used to generate the hypothesis of controlled clinical trials were to show achieved the same extent of scientific a relationship between omega-3 fatty that n-3 fatty acids helped to prevent agreement. acids and CHD were also reexamined. atherosclerosis, these agents apparently Where authorized health claims Thus, by utilizing the two reports in would represent one of the most benign include qualifications, the qualifications question, supplemented with an interventions in our pharmacopeia.” are intended to assure that the wording independent review of the subsequently (Emphasis added.) of allowed claims reflects those published research, FDA was able to Bonaa (Ref. 10) wrote in his particular aspects of the substance- assess the totality of the scientific conclusion that the data on blood disease relationship for which there is evidence on omega-3 fatty acids and pressure: significant scientific agreement, not to CHD in compliance with the statutory standard. qualify the extent of agreement. * * * provide some support for the 9. Some comments stated that FDA 10. One comment suggested that FDA hypothesis that dietary marine lipids relied heavily on material published in was inconsistent with the conclusions influence blood pressure in man. the National Academy of Sciences 1989 of the major reviews of this area, Supplementation of n-3 PUFAs report, “Diet and Health: Implications published after the Federal Government [polyunsaturated fatty acids] to Western diets for Reducing Chronic Disease Risk” and other comprehensive reports. They consistently lowered systolic blood pressure, while results for diastolic blood pressure (Ref. 115) and the Surgeon General’s stated that of the nine major reviews (excluding Kinsella, and Connor and were conflicting * * *, There is no evidence 1988 report (Ref. 34), and did not place of any substantial hypotensive response to enough emphasis on information Connor), eight concluded that omega-3 marine lipids and further studies should be published since that time. fatty acids played a beneficial role with designed to detect small effects. FDA acknowledges that the two factors affecting heart disease. Lands (Ref. 89) did not review the reports in question were important to its Although FDA did not rely on review relationship between omega-3 fatty assessment of the scientific evidence. articles to assess the strength of acids and any specific disease, but However, the agency does not agree that association between omega-3 fatty acids presented the hypothesis that the it failed to give appropriate weight to and CHD, each review was read, and the balance of omega-3 and omega-6 fatty subsequently published research. The agency interprets these reviews as acids in the diet may be related to 1990 amendments required the agency supporting the hypothesis in concept. diseases associated with overproduction to consider the totality of publicly However, each review contained of eicosanoids from AA. He indicates in available scientific evidence in reservations about the extent to which the introduction that, “We are now in the relationship betweenomega-3 fatty assessing nutrient-disease relationships. an uncertain time of evaluating the acids and CHD was established. The Given the time constraints imposed by benefits and risks of dietary n-6 and n- cautionary statements suggest general the 1990 amendments for developing 6 polyunsaturated fats.” agreement that the area of omega-3 fatty and publishing proposed regulations. Weber (Ref. 161) concluded: FDA depended on Federal Government acids and CHD holds promise for further research along a number of lines, but The promise of n-3 fatty acids deduced reports and reports of authoritative from biochemical and functional effects will bodies (e.g. the National Academy of that, at present, there are not sufficient have to be evaluated in ongoing and future Sciences) for assessment of the scientific data to have certainty about the carefully designed and conducted studies. So

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2686 far, published data of controlled clinical become available, “clinicians should be patients, and epidemiologic trials incorporating clinical endpoints after n- advised to follow the dietary observations are grouped with other 3 PUFAs are available only in abstract form. recommendations of the National comments on these topics and discussed Therefore, the gap between biochemical and in this document. functional effects of dietary fatty acids Cholesterol Education Program’s expert assumed to be of clinical benefit in the panel,” which is silent on omega-3 fatty 12. One comment considered FDA’s prevention of atherothrombotic and allergic/ adds and limits the total : caution against extrapolation of results inflammatory disorders is only beginning to polyunsaturated fat to 10 percent of from studies conducted in at-risk be closed. (Emphasis added.) calories. These reviewers added their populations to the general population to Connor and Connor (Ref. 21) wrote in own recommendation that omega-6/ be questionable, and possibly biased their summary: omega-3 ratio of the PUFA’s be against hypertensives. The comment The exact place ofomega-3 fatty acids from approximately 3/1, with the omega-3 stated that the health claim should be fish and fish oil remains to be defined. fatty acids from marine sources. allowed, based on data showing that However, this much seems certain. Fish Weber and Leaf (Ref.162) stated: omega-3 fatty acids reduce blood provides an excellent substitute for meat in Despite all the laboratory, human, animal, pressure among hypertensives. the diet. Fish is lower in fat, especially and epidemiologic studies suggesting an anti- FDA disagrees with this comment. saturated fat, and contains the omega-3 fatty atheromatous action ofw-3 fatty acids, we FDA stated that, although it considered acids. Fish oil may have promise as a have been lacking adequate clinical trials therapeutic agent In certain hyperlipidemic studies in the healthy population to be which will determine in prospective, the most relevant, it also considered states, especially the chylomicronemia of placebo-controlled, randomized studies, type V hypolipidemia. Fish oil has logical whether all the above experimental and studies in a subpopulation with CHD or and well-defined antithrombic and epidemiologic evidence adds up to a risk factors for CHD, in part because antiatherosclerotic activities since it demonstrable effect of fish oils to prevent high risk populations may be more depresses thromboxane A2 production and atherosclerosis, e.g., coronary heart disease in sensitive to showing a nutrient-disease inhibits cellular proliferation responsible for humans at high risk for heart attacks. relationship than the general population the progression of atherosclerosis. As the The Burr paper (Ref. 16) was (56 FR 60663 at 60667). FDA stated that years pass and more experiments are reported, it seems reasonable to place the described in Weber and Leafs review, it extrapolated positive results from at- omega-3 fatty acids from fish oil in a and thus was considered in the above risk populations cautiously, and that prominent position for specific summary statement. comparable findings in the general hypolipidemic, antithrombotic and In summary, these reviews indicate population were needed to support a antiatherosclerotic activity. that what is agreed is that there is a health claim. Kinsella et al. (Ref. 82) wrote: plausible biochemical basis for a 13. Two comments discussed FDA’s relationship between omega-3 fatty criteria for weighing various types of The cumulative findings concerning fish acids and CHD, and that there are some data. One comment stated that oils suggest that further amelioration of data supporting some of the epidemiologic data are the “most coronary heart disease may be feasible by hypothesized mechanisms by which significant class of evidence,” and that dietary manipulation and by optimizing the omega-3 fatty acids might be related to FDA should give priority to various intake of n-6 and n-3 PUFAs, not only to CHD. What is not agreed, as indicated types of data in the same order that reduce plasma lipids but to ensure balanced by the cautious tones of these various types of date were reviewed in eicosanoid metabolism—a prospect that the proposal. One comment stated that deserves more research * * *. Overall, in concluding statements, is that such a view of the prevalence of coronary heart relationship already has been FDA should not have considered disease, consumption of n-3 PUFA oils established by the evidence. epidemiological studies separately from should be considered as a useful 11. A concern raised by many clinical trials. complementary option for the amelioration of comments was that FDA’s conclusions FDA considered the totality of coronary vascular diseases. were different from the conclusions publicly available scientific evidence in Knapp (Ref. 84) introduced his paper reached in the report from the Life its assessment of the relationship of stating: “The role of dietary Sciences Research Office (LSRO) of the omega-3 fatty acids to CHD. However, polyunsaturated fats in the prevention Federation of American Societies for some types of evidence were weighted of human vascular disease has not been Experimental Biology (Ref. 100), the more heavily than others because they defined, but population and most recent comprehensive review, and were more useful in establishing intervention studies have suggested that that FDA did not explain why it reached whether or not the scientific basis of the w-3 fatty acids (FAs) from marine lipids a different conclusion from that reached claim was valid. In particular, the may have a number of potentially in the LSRO report. agency was concerned that both the beneficial effects.” (Emphasis added.) The LSRO report was contracted for substance (omega-3 fatty acids) and the And in conclusion he wrote: “The proof by FDA as an independent review of the disease (CHD) be carefully of our hypotheses must be derived from scientific evidence about the characterized. FDA also considered it increasingly ambitious clinical trials, relationship between omega-3 fatty important that the amount of omega-3 which assess the potential benefits of acids and CHD. A draft of the tentative fatty acids tested was reasonably related dietary polyunsaturates in particular final report was received immediately to normal dietary intake, and that the clinical settings, the recent prior to the publication of the proposed findings apply to the general demonstration that three helpings of rule. Thus, there was insufficient time population. FDA agrees that oily fish per week prolongs survival for the agency to prepare a detailed epidemiologic studies in which the after Ml (Ref. 16) is an example of this.” discussion of the report. The final report endpoint was CHD provide persuasive (Emphasis added.) was submitted to FDA as a comment to evidence for a relationship between fish Nestel (Ref. 111) concluded: “More consumption and CHD, but these the proposal. The LSRO report’s basic understanding of the actions of studies did not provide the specificity to conclusions on hypertension, fish oils is necessary before fish oils can show that the observed effects were due thrombosis, the development of be recommended widely to the public.” to omega-3 fatty acids. Intervention atherosclerotic plaque and intimal Nordoy and Goodnight (Ref. 112) trials using fish oil supplements often hyperplasia, plasma lipids and cautioned that until additional data showed that the effects were specific to lipoproteins, diabetic and prediabetic omega-3 fatty acids (by controlling with

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2687

other types of fatty acids) but typically may be resolved by further research. consumption of fish inversely to CHD to did not measure the primary endpoint, Thus, FDA’s analysis should provide be sufficient to support a health claim, CHD. Thus, these different types of data guidance for additional research rather but did not supply any new information complement each other and must be than inhibit it. or arguments to support their position. considered together in assessing the B. Relationship Between Omega-3 Fatty FDA disagrees with the comments totality of the scientific evidence. Acids and CHD FDA found that: 14. One comment offered the services In the proposed rule, FDA tentatively of the International Society for the concluded that the totality of the Only a few studies found an association Study of Fatty Acids and Lipids for the between, fish intake and CHD, while others scientific evidence does not provide a evaluation of the relationship between have found no association. Thus there was basis upon which to authorize a claim omega-3 fatty acids and CHD. not consistency of findings. None of the that omega-3 fatty acids are associated FDA appreciates this offer. In the final studies that reported a relationship with the risk of CHD (56 FR 60663). rule on general requirements for health distinguished fish consumption from other FDA noted that. factors associated with fish consumption, claims published elsewhere in this issue the epidemiological research on this topic and therefore they did not demonstrate of the Federal Register, FDA advises specificity. Even in those studies reporting a that it welcomes the input of any revealed that the available studies applied only to the consumption of fish, which relationship between fish consumption and professional organization that can contain omega-3 fatty acids, and * * * it was CHD, it was not clear that the effects were provide expertise in reviewing data and not possible to ascribe any effects specifically because of the omega-3 fatty acids in fish. in developing a thoughtful and well- to the omega-3 fatty acids. Examination of Also, the omega-3 fatty acid content of the organized petition for a health claim on data from clinical studies revealed that the fish diet associated with reduced CHD was so a particular topic. In fact, FDA has effects on blood lipids of fish oils containing low that the importance of omega-3 fatty added to § 101.70(b) the provision that omega-3 fatty acids were primarily a acids is questionable * * * information submitted with petitions reduction of blood triglycerides, a blood lipid (56 FR 60663 at G0672.) may include any findings, along with variable not considered to be an independent 17. One comment described the risk factor for CHD, but they had no effect on the basis of the findings, of an outside results of the Dolecek and Grandits serum cholesterol, low-density lipoprotein panel with expertise in the subject area (LDL) cholesterol, or high density lipoprotein analysis of multiple risk factor at issue. FDA, however, retains the (HDL) cholesterol, the blood lipid variables intervention trial (MKFIT) data (Ref. 38) authority to review such petitions and, most closely associated with risk of CHD. as indicating a greater protective effect through rulemaking, to decide whether The scientific data are ambiguous on the against CHD due to consumption of 0.6 or not to authorize the claim. effects of omega-3 fatty acids on blood gram (g) omega-3 fatty acids than all 15. Two comments stated that it was pressure and other risk factors for CHD. other conventional efforts combined contradictory for the U.S. Government (56 FR 60663.) (reducing saturated fat, cholesterol, to contract for research on the omega-3 A number of comments supported the cigarette smoking, and hypertension). fatty acids through the biologic test agency’s position on this health claim, FDA agrees with this comment that materials program but not to allow a but without any specific reasons for that the association between omega-3 fatty health claim. Another comment pointed support. One comment agreed with the acid consumption and CHD mortality out that the Food and Agriculture agency’s position in principle, but reported in this study has the potential Organization and the World Health contested the agency’s interpretation of to make a very important public health Organization support research on the scientific information in some areas. impact. Notably, the results were omega-3 fatty acids. Other comments Other comments disagreed with the obtained on data adjusted for age, race, slated that the tone of FDA’s proposed agency’s review of the scientific smoking at entry to the study, diastolic rule was unduly negative and that, by information and its conclusion blood pressure, and high-density taking such a position, FDA may retard regarding the strength of the evidence lipoprotein (HDL) and LDL further research. supporting the proposed health claim. concentrations. Furthermore, the omega- FDA disagrees that Federal Specific comments are summarized 3 fatty acids were obtained in the Government sponsorship of a program below. normal diet, providing evidence that the to provide test materials for research on amount of omega-3 fatty acids the effects of omega-3 fatty acids and the 1. Epidemiologic evidence consumed in a normal dietary intake is sufficient for the effect. denial of the omega-3 fatty acid-CHD In the proposed rule, FDA reviewed The researchers’ adjustments for health claim are contradictory actions. correlational and cross-sectional The purpose of the biologic test lipoprotein measurements should studies, prospective studies, and control for some other dietary variables materials program is to develop, and intervention studies available since standardize a source of omega-3 fatty that have been associated with CHD 1988. (See 56 FR 60663 at 60667 acids and enable carefully controlled through their effects on these through 60668). Except for the research on the effects of particular lipoproteins, e.g., saturated fat, but other intervention studies (which were omega-3 fatty acids. In the proposed dietary variables associated with CHD typically clinical trials) these studies rule, FDA’s intent was to examine the were not controlled, e.g., alcohol. The used fish as a source of omega-3 fatty total available scientific evidence, some association between omega-3 fatty acid acids. FDA concluded that those studies of which was generated using omega-3 consumption and CHD mortality that used fish as the source of omega-3 fatty acids from the biologic materials described in this study is among the fatty acids were: “ambiguous, because test program, and to state its most provocative findings to date in this they are not capable of distinguishing conclusions about the relationship area, and merits additional study using the effects that are specific to omega-3 between omega-3 fatty acids and CHD, a design that will document that the fatty acids from those that are related to In its proposal and in this final rule, active dietary component’ is or is not the fish consumption.” (56 FR 60663 at FDA has identified a number of areas omega-3 fatty acids (i.e., specificity of 60668.) where agreement is lacking that an the effect). 16. A number of comments observed effect of omega-3 fatty acids is 18. One comment pointed out that the considered the evidence from related to the risk of CHD, or where Burr paper (Ref. 16) deserved close there are ambiguities in the data that epidemiologic studies that relates the consideration, because, in contrast to

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2688

trials on lipid lowering drugs, it showed by the displacement by fish of other behavioral factors (e.g., smoking, that consumption of fish containing atherogenic foods from the diet. alcohol) must be controlled before omega-3 fatty acids or dietary FDA is not persuaded by these meaningful conclusions may be drawn supplements of omega-3 fatty acids may comments. The limitation in these about the effects of omega-3 fatty acids. reduce the risk of heart disease. One studies is that they did not control for FDA notes that alcohol consumption comment stated that it considered the dietary factors associated with CHD, not was also a confounding factor in a study Burr paper to be a positive finding, but that fish consumption displaced other that reported an association between gave no reason for this conclusion. The atherogenic foods. FDA noted in its fish consumption and CHD (Ref. 87). LSRO final report, submitted as a proposal that the Zutphen study found 22. A few comments stated that many comment, also recognized the Burr significant correlations between fish of the reported effects come from paper as a very important trial. LSRO consumption and other dietary factors studies on fish consumption, but that all pointed out that, although separate (i.e., alcohol, polyunsaturated fats) measured biochemical changes related results were not shown for those related to CHD. Comparable correlations to CHD that are produced by fish have consuming fatty fish and those were not addressed in the Burr paper also been produced with fish oil consuming supplemental fish oil, the because dietary intake data were not concentrates. results were dramatic, especially since reported. Also, the design of the Burr FDA agrees in part with this all-cause mortality was reduced, in paper was to encourage consumption of comment. The fact that the same contrast to results from trials of plasma fish, which would likely have resulted biochemical results have been obtained lipid-lowering drugs. LSRO concluded in a reduction in the consumption of red using fish oils rather than fish provides that “future research will be needed to meat (and, therefore, saturated fat). strong evidence that particular define the amount and duration of w-3 20. Two comments discounted the biochemical markers are affected fatty acid supplementation required to Curb et al. study (Ref. 25), which specifically by omega-3 fatty acids. produce the beneficial effects.” showed no association between fish Also, since most studies have used fish FDA agrees that the Burr paper consumption and CHD mortality among oils, these results add consistency to the provides valuable evidence consistent subjects in Hawaii. The comments effects reported for studies that used with the hypothesized relationship stated that the dietary source of fish was fish. However, FDA disagrees that the between omega-3 fatty acids and CHD. likely tropical fish, and since tropical comparable findings in studies that used However, FDA noted in its proposal (56 fish feed on coral they have a high fish oils and fish are sufficient to FR 60663 at 60668) that there are two content of AA, which would counteract support the health claim that omega-3 specific shortcomings in this paper: the the effect of omega-3 fatty acids. fatty acids reduce the risk of CHD, absence of separate data for subjects FDA disagrees with the comments. No because the particular biochemical who consumed fish and those who data regarding the AA content of the markers affected by both fish and fish consumed fish oil capsules, and the diet in this study, or in other oils are not recognized with significant absence of dose-response data. These correlational studies, have been scientific agreement as useful surrogate data would have provided evidence for reported. Indeed, most epidemiologic risk factors for CHD in the general a specific effect of omega-3 fatty acids. correlation studies have not quantified population. Ideally, other data regarding the the intakes of omega-3 fatty acids, a 23. One comment argued that the fact subjects diet would also show that fundamental measurement to establish that Greenland Eskimos ate diets with there was no difference in consumption an association between omega-3 fatty half the saturated fat and more of other dietary factors related to CHD. acids and CHD. Finally, there is an polyunsaturated fat than Danes and had The study design specifically included abstract reporting that the omega-3 fatty much less CHD than Danes strengthens two such dietary factors, dietary fat and acid to omega-6 fatty acid ratio of the case for fish oil-derived omega-3 dietary fiber, and the lack of significant tropical fish is comparable to or greater fatty acids. effects of these components argues than that of fish m higher latitudes (Ref. FDA agrees with the comment that against dietary factors other than omega- 237). Thus, the comments’ explanation diets lower in saturated fat are 3 fatty acids as responsible for the for a negative finding must be consistent with reduced CHD mortality association. considered theoretical. (see the final rule on “Dietary Lipids FDA does not consider the Burr paper 21. One comment argued that the lack and Coronary Vascular Disease” to have established a beneficial effect of of an association between fish published elsewhere in this issue of the omega-3 fatty acids, although its results consumption and CHD in two Federal Register). The differences in are consistent with such an action. The populations in Canada, a prairie saturated fat intake, however, do not LSRO conclusion indicates that neither province and a coastal province (Ref. strengthen the case for omega-3 fatty the amount of omega-3 fatty acids 74), was because the prairie population acids, because they do not distinguish necessary for beneficial effects nor the consumed more alcohol and the coastal omega-3 fatty acids from duration of their intake has been population smoked more. This comment polyunsaturated fats. Rather, the established. The specificity of the criticized FDA for not pointing out the differences in dietary fat intakes substance responsible for the beneficial cautions raised by the authors about strengthens the argument that saturated effects, the quantitative amount needed potential confounders like the fat is associated with CHD mortality. to produce the effect and the duration difference in alcohol consumption. The numerous dietary differences of intake needed to produce the effect FDA believes it presented the results between the Greenland Eskimos and need to be established before FDA can of this paper fairly. While the authors Danes make it difficult to ascribe to any authorize a claim linking omega-3 fatty reported small differences in smoking single dietary factor the differences in (more in the coastal population) and acids to reduction of risk of CHD. CHD. alcohol consumption (more in the 19. Some comments stated that the 24. One comment pointed out that, of prairie population), they stated, “It amount of fish in the Zutphen and Burr the ten prospective studies cited in the seems unlikely that these differences are studies was so low that the association proposal (including three in Table 1 of sufficiently large to offset any strong between fish consumption and reduced the proposal), six support an inverse effect of fish consumption.” FDA is CHD mortality could not be explained relationship between fish consumption keenly aware that dietary and and CHD. The comment noted that one

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2689 study only showed a relationship in mortality were due to differences in 29. One comment provided new dose- men under 45, and argued that this dietary omega-3 fatty acids, response data from additional analyses result is promising because one might 26. Two comments stated that FDA of data of the Dart study, previously expect to find positive effects of long- had erred in stating that no biochemical reported in part by Burr et al. (Ref. 16), term fish consumption on CHD in data were reported in the Burr paper that related the dietary intake of EPA at younger, relatively healthy men rather (Ref. 16). 6 months into the trial to the risk of than in older men. FDA agrees with this comment, and CHD events (heart attacks, or Ml’s) or FDA agrees that some but not all stands corrected. Bun et al. (Ref. 16) did CHD mortality. The 947 subjects for reports find an inverse relationship report that the geometric mean whom dietary data were obtained were between fish consumption and CHD. percentages of EPA were 0.59 percent grouped according to EPA intake; 114 FDA does not agree, however that this and 0.40 percent in men given advice to consumed less than 1 g per week (1 g/ effect would more likely be noticed consume more fish and those not so week), 373 consumed 1 to 2 g/week, and among younger, relatively healthy men. advised, respectively, a highly 460 consumed 2 or more g/week. The Older men would have had a greater significant difference (p <0.01). The fact percentage of subjects that experienced duration of intake of omega-3 fatty acids that a geometric mean rather than an either a nonlethal heart attack or died and a greater incidence of CHD than arithmetic mean was reported implies from a heart attack decreased as dietary younger men. Both these factors would that there was substantial skewing of the EPA increased. For heart attacks the favor finding an effect in older men data. rates were 7.9 percent, 7.0 percent and rather than in younger men. It is not clear from the article whether 6.7 percent, for the less than 1 g/week, 25. One comment noted that only two these differences were for the 6-month 1 to 2 g/week and 2 or more g/week studies found a positive dose-response. groups, respectively. The percentages in time into the trial, or for the end of the Another comment stated that the studies each group who died were 6.1 percent, trial. The authors did not correlate that show no effect are those where the 5.1 percent, and 4.1 percent, plasma EPA concentrations directly base group already consumed fish, respectively. There were no statistical with myocardial infarction (MI) or CHD whereas in studies that showed an analyses of these data reported. deaths. effect, the base group did not consume FDA notes some limitations in these 27. One comment argued that it was fish. A third comment stated that there data as reported that caution against highly misleading to state in Table 1 were data that described on inverse strong conclusions. Most notably, the that Kromhout et al. (Ref. 67) reported analysis excluded the events and deaths dose-response relationship between that, “lean fish, low in omega-3 fatty serum EPA and CHD deaths among during the first 6 months of the trial, acids, had some protective effect against when about half of all events and deaths Japanese, but did not identify a CHD.” because Kromhout did not particular study. occurred. This clearly diminishes the distinguish between the effects of lean sensitivity of the analysis, and may FDA disagrees that only two studies and fatty fish. found a dose response correlation. Each result in an underestimation of the true FDA disagrees with this comment. effect, since the difference in survival study that response a relationship The authors made two statements about between fish consumption (or, in one between the group advised to eat more lean fish that imply that additional data study, the calculated intake of omega-3 fish and the group not advised to eat analyses were conducted, although (as fatty acids) and CHD found a dose- more fish was most pronounced, during the comment correctly notes) results of response relationship. FDA agrees that the first 6 months. Alternatively, if the these analyses were not included in the most of the dose-response relationships healthiest subjects were also the most paper. The authors wrote, “Lean fish reported suggest that the primary compliant subjects, the reduced death was also inversely related to mortality difference in rate of CHD is between rate in the highest EPA-consumption from coronary heart disease,” and subjects may reflect the underlying those who consume no fish at all and “Thus, the inverse relation between lean those who consume a small amount of health of those subjects, and the fish and coronary heart disease cannot importance of dietary EPA may be fish, and that there appears to be little be explained by eicosapentaenoic acid.” overestimated. additional benefit from consumption of FDA interprets these comments as a Also, the unequal group sizes for this large amounts of fish (Ref. 88). An caution to the reader against assuming alternate way of describing these data is analysis places a greater weight on each that EPA was the active component subject in the smallest group (less than that these who consume no fish have an responsible for the observed reduction 1 g/week) than in the other groups. This increased risk of CHD. These data merit in CHD among fish-consuming subjects. may be particularly important because followup, because a showing that the 28. LSRO included in its report two the smallest group includes those who relationship is due to the omega-3 fatty studies that correlated plasma omega-3 consume no fish, arid who may differ acids may provide evidence that the fatty acids with dietary intake of these from fish consumers in other dietary or long-chain omega-3 fatty acids are fatty acids (Refs. 213 and 225). Two behavioral factors associated with CHD essential in the diet. other papers reviewed by LSRO but not risk. The sensitivity of the results to In its proposal, FDA reviewed a study included in the FDA proposal were small changes in outcomes is shown by that described on inverse dose-response correlation studies of mortality from example: one fewer death (6/114 rather relationship between serum EPA and different diseases among Greenlanders than the reported number, 7/114) makes CHD mortality among two groups of and Danes (Ref. 170) and diet-disease the CHD death rate of the less than 1 g/ Japanese (Ref. 70). FDA concluded that correlations in Japan (Ref. 284). week group equal to the rate in the 1 to these cross-sectional, correlational data FDA agrees with LSRO’s descriptions 2 g/week group. were useful in generating a hypothesis. of these studies. FDA notes that the Finally, although the dietary intake Other notable dietary factors (including authors of the studies that correlated data at 6 months are useful, this study a difference in salt intake of 50 percent) intake and plasma levels of omega-3 also assayed plasma fatty acids. Use of and risk factors for CHD (prevalence of fatty acids did not relate their data to plasma EPA (or EPA plus DHA) in the hypertension) also differed between the CHD. The correlation studies of dose-response analysis would have been two groups, so it is not possible to Mortality did not provide any specific a more powerful analysis, because it conclude that differences in CHD data regarding omega-3 fatty acids. eliminates errors in the diet record data

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2690 corrects for losses during food oils or fish containing omega-3 fatty during the fish diet phase. There was no preparation and individual differences acids on these blood lipid measures; difference in total fat of the two diets. in bioavailability of the fatty acids, and and (2) there is general agreement that However, there was significantly less integrates intake of omega-3 fatty acids certain blood lipids are strongly saturated fat in the fish diet, so it is not over a longer period than the diet record associated with the risk of CHD. certain that the omega-3 fatty acids were data. FDA undertook to evaluate findings in responsible for the decrease in the Therefore, FDA finds these dose- these studies whether or not fish or fish cholesterol measures. response data to be consistent with the oils were used as the source of omega- FDA is concerned that the studies hypothesis that omega-3 fatty acids 3 fatty acids and whether or not the abstracted by LSRO do not accurately reduce the risk of CHD, but the outcome measures were generally represent the totality of publicly shortcomings discussed above limit recognized as predictive of CHD. This available scientific evidence. For their usefulness in establishing a approach allowed the identification of example, in its proposal, FDA included relationship between omega-3 fatty biologic activities of omega-3 fatty acids five studies among normal subjects acids and risk of CHD. that may be related at some point to risk (Refs. 2, 6, 24, 73, and 143) and three 2. Evidence relating omega-3 fatty acids of CHD, and identified areas where studies among subjects with preexisting to intermediate or surrogate markers of additional research is needed. FDA lipid or lipoprotein abnormalities (Refs. CHD included in the summary of its 18, 73, and 93) not included in the proposed rule (56 FR 60663) only In the proposed rule (56 FR 60663 at LSRO report that had data for effects of measures of generally recognized risk 60668), FDA stated that most omega-3 fatty acids on plasma lipids or factors. FDA did not intend to imply information about the effects of omega- lipoproteins. FDA determined that that data on alternate markers were not 3 fatty acids on CHD has been derived seven studies that reported changes in considered in its decision. from clinical trials using concentrated total cholesterol had the most rigorous fish oils enriched in EPA and DHA, and FDA agrees that other factors designs and the largest numbers of contribute to blood lipid measures, but in some cases purified methyl or ethyl normal subjects. None of these seven believes that randomization should esters of EPA and DHA. FDA concluded studies (Refs. 6, 9, 14, 49, 54, 73, and control for these factors. In that: 166) in normal subjects found a nonrandomized studies, these sources of significant change in total cholesterol

potential bias limit the conclusions that after fish oil supplementation. FDA * * * there are a few established effects of can be inferred from the data. This is an found similar results with regard to omega-3 fatty acids from fish oils on important reason that data from hyperlipidemic subjects. thrombosis and hemostasis. Standardized Only two of these seven strongest bleeding times are increased, and platelet correlation studies do not conclusively aggregation and function are reduced. establish a relationship between omega- studies in normal subjects were However, direct relationships between the 3 fatty acids and risk of CHD. abstracted in the LSRO text, and two changes in bleeding times or platelet function 31. The LSRO report paid others were not cited at all by this and risk of CHD have not been established. considerable attention to changes in report. LSRO did not distinguish While there is an established relationship blood lipids after increased between normal and hyperlipidemic between blood pressure and CHD, it has not consumption of omega-3 fatty acids, and subjects in its summary or conclusions. been shown that omega-3 fatty acids reached some conclusions about effects LSRO summarized the evidence on total specifically affect blood pressure in. normal subjects in a way that would provide a of omega-3 fatty acids on blood lipids cholesterol by stating, “Decreases in protective benefit toward the risk of CHD. that differed from those reached by total cholesterol * * * have also been Effects of omega-3 fatty acids on other FDA. LSRO abstracted three studies reported,” (emphasis added), without markers linked with CHD, e.g., fibrinogen or from before the period covered by FDA mentioning that the predominant lipoprotein (a) have not been established. review (Refs. 284a, 267, and 2.57). These finding is that there is no effect on total (56 FR 60663 at 60671). studies were considered, in Federal cholesterol. a. Atherosclerosis Government and other comprehensive Similarly, FDA stated that the reports reviewed in the proposed, rule strongest studies among normal subjects

and not discussed further by FDA. (Refs. 6, 9, 14, 49, 54, 73, and 166) i. Blood lipids LSRO also included three studies not found no change in LDL cholesterol, 30. Numerous comments criticized reviewed by FDA in its proposal (Refs. and one reported an increase in LDL FDA’s focus on blood cholesterol as a 168, 226, and 301). Agren et al. (Ref. cholesterol (Ref. 54). Indeed, most surrogate marker for risk of CHD, 168) studied healthy students randomly studies on hypertriglyceridemic or although one comment noted that such assigned to their normal diet (one fish hypercholesterolemic subjects reported an emphasis would be expected, given meal per 2 weeks), a fish diet, or a fish an increase in LDL cholesterol, following the importance of cholesterol in CHD. diet low in saturated fat for 15 weeks. fish oil supplementation (56 FR 60663 Another stated that the focus on There was no change in total serum et. 60669). Consequently, FDA disagrees cholesterol ignores other factors that cholesterol on the control diet or fish strongly with the summary statements determine blood cholesterol such as diet, but the low-fat fish diet produced in the LSRO report: heredity, exercise, and stress, etc. a reduction in total cholesterol. Jensen Effects of fish oil upon LDL have been FDA disagrees with the comments et al. (Ref. 226) studied 18 healthy variable, in part because of different doses. In normolipidemic individuals, LDL has that it overemphasized the importance subjects supplemented sequentially for generally declined significantly. In some of studies in which cholesterol was 4-week periods with 4-week washouts patients with primary hypercholesterolemia, measured as a surrogate marker for between, with fish oils containing 1, 3, consumption of fish (sic) has not resulted in CHD. The considerable attention given and 6 g EPA plus DHA, and found no altered plasma cholesterol levels; other to blood cholesterol measurements (and change in total or LDL cholesterol. studies have shown decreased cholesterol measurements of other blood lipids and Wolmarans studied healthy subjects fed and LDL levels. (Emphasis added.) the protein components of blood a meat diet or fish diet containing 6.1 g 32. Two comments stated that FDA lipoproteins) was the consequence of EPA plus DHA for 6 weeks each in a had not considered all relevant data on two factors: (1) There was a large crossover design, and found reduced HDL2 cholesterol, and cited additional number of studies on the effects of fish total cholesterol and LDL cholesterol

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2691

studies that reported increased HDL2 FDA in other contexts, but data from FDA disagrees with all but the last cholesterol after fish oil these papers regarding HDL2 cholesterol comment. FDA is aware that there has supplementation. One comment stated levels were not discussed (Refs. 1, 32, been, and still is, substantial interest in that overall HDL cholesterol tends to and 148). the potential role of triglycerides in the rise, and cited a review paper by Harm FDA reexamined those papers that it etiology of CHD (e.g., Ref. 208). Because (Ref 62). The LSRO report also cited but from which it did not present of the continued interest, the concluded that HDL was increased by data regarding HDL2, together with the relationship between triglycerides and fish oil supplementation. newer papers. When fractions of HDL CHD was the topic of a consensus FDA disagrees with the comment cholesterol have been reported, an development conference sponsored by regarding the overall HDL cholesterol increase has generally been found in the NHLBI on February 26 through 28, change after fish oil supplementation. HDL2 fraction (Refs. 1, 32, 148, 235, 251, 1992. NHLBI had previously addressed The agency considered HDL changes and 291), with a comparable decrease in this topic in 1983 and concluded at that separately for normal, healthy subjects the HDL3 fraction (Refs. 1, 235, and time that the relationship was and for hyperlipidemic subjects (56 FR 251). This represents a shift within the controversial. The recent conference 60663 at 60669). Nearly all studies on HDL fractions toward a lipid-rich (Ref. 255) concluded, “For triglyceride, normal subjects found no significant lipoprotein, and away from a protein- the data are mixed; although strong change in HDL cholesterol level. Some rich lipoprotein, similar to that reported associations are found in some studies, investigators reported increased HDL2, for LDL, below. This shift has been the evidence on a causal relation is still but the data on HDL2 were equivocal. reported when there is (Refs. 32, 148, incomplete.” FDA also disagrees with the 235, and 291) or is not (Refs. 1 and 251) FDA agrees that the statistical conclusions of the LSRO report a change in total HDL cholesterol. This methods previously used to study the regarding HDL cholesterol, because it raises the possibility that a shift relationship between triglycerides and does not represent the totality of occurred in other studies where total CHD have lessened the likelihood that publicly available scientific evidence. HDL was reported as not changed. triglycerides would be found to be a The LSRO summary states, “In some However, the importance of the shift significant, independent predictor of studios HDL concentrations have in subfractions of HDL is not clear. FDA CHD. Furthermore, the agency believes actually increased with consumption of noted in its proposal (56 FR 60663 at that study design and analytic fish oil” (emphasis added), not 60869) that there is evidence that the measurement methods have contributed acknowledging that the balance of HDLs fraction is the one most closely to variation in triglycerides that may available scientific evidence on HDL linked to risk of CHD. However, the have resulted in reducing the statistical indicates no change. In the review by agency was unable to find evidence that association between triglycerides and Harris cited in the comment, the there was significant scientific CHD. FDA believes that these sources of changes in HDL cholesterol in each agreement that HDL2 was the fraction of variation in triglycerides can be reduced study were weighted according to the HDL most closely associated with CHD. by careful study design and number of subjects in the study, giving The National Institutes of Health’s standardized analytical measurement a per-subject change. This method of National Heart, Lung, and Blood techniques, and also that clinical pooling data from different studies does Institute (NHLB1) consensus studies designed to lower triglycerides not account for the variation of the development conference on could provide a basis upon which to response of subjects in each study, the Triglyceride, High Density Lipoprotein reconsider the importance of amount of omega-3 fatty acids fed, the and Coronary Heart Disease (Ref. 255), triglycerides in CHD. duration of feeding, or the source of the anticipated in the proposal (56 FR 34. Some comments stated that some omega-3 fatty acids. Therefore, it must 60663 at 60664), concluded that, “It is very recent evidence from the Helsinki be considered only an estimate of the not known to what extent these Heart Study supports a protective effect effects of omega-3 fatty acids on HDL alterations of HDL contribute to of lowering triglycerides, at least for a cholesterol. Harris calculated the atherogenesis,” selected subpopulation of people with a average HDL cholesterol change for Therefore, data on changes in HDL high ratio of LDL cholesterol/HDL normal subjects to be an increase of subtractions after increased cholesterol and very high triglycerides. approximately 3 percent, a net change consumption of omega-3 fatty acids do FDA agrees that fish oils reduce smaller than the usual variability in the not provide a sufficient basis for a plasma triglycerides. In its proposal test used to measure HDL. health claim, because there is not FDA wrote, “The predominant blood The agency agrees with the comment significant scientific agreement that lipid effects offish oils * * * are decreased plasma triglycerides and that not all HDLs data were considered HDL2 is directly related to risk of CHD. in the proposed rule, although FDA If the risk of CHD becomes linked with VLDL.” (56 FR 60663 at 60669.) In this regard FDA and LSRO were in noted (56 FR 60663 at 60669) that some HDL2, these findings in normal subjects studies among normal subjects found may be of great importance. agreement. The LSRO summary states, increases in the HDLs fraction of HDL 33. Many comments indicated that “The most striking effect is lowering of cholesterol, and that these reports were high triglycerides are causally related to plasma triglyceride and VLDL the most promising changes in blood decreased HDL, that triglycerides are an concentrations.” lipids. Of the six references cited by the independent risk factor for CHD, or that FDA disagrees, however, that comments as not Included among statistical manipulations of data and triglycerides have been established as an studies showing increased HDL imprecise measurements of triglycerides independent risk factor for CHD. The cholesterol after onwge-3 fatty acids, obscure the importance of triglycerides recent results from the Helsinki Heart two were published after the time as a risk factor for CHD. One comment Study (Ref. 242) were discussed at period covered in the proposed rule provided additional citations regarding length at the NHLBI consensus (Refs. 235 and 252). One other paper not the relationship between triglycerides development conference (Ref. 255 ). cited by FDA in its proposal, although and HDL, but these did not bear on risk While the reduction in CHD mortality it was published during 1988 (Ref. 291), of CHD. One comment stated that it was following drug intervention was dealt with insulin-dependent diabetics. generally agreed that triglycerides were dramatic (i.e., approximately 7-fold) for The other three papers were cited by not independently associated with CHD. a particular subgroup with both elevated

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2692 triglycerides and elevated LDL to HDL quantities of w-3 fatty acids. This may be a notes that patients in these studies are ratio, this result was obtained by a post most important anti-atherogenic action. under a physician’s care. FDA’s hoc analysis of earlier results. Because FDA agrees that fish oils do not description of the Dehmer study points the combination of factors used to generally lower HDL. FDA also agrees out a limitation of the data that is connote the high-risk group (i.e., high that major blood lipid effects of omega- common in other reports of no effect of LDL cholesterol to HDL cholesterol ratio 3 fatty acids are reductions of omega-3 fatty acids in restenosis (Refs. and high triglycerides) was determined triglyceride and VLDL. The role of 56,106, and 121), that the studies have after the data were collected, those omega-3 fatty acids in the reduction of not controlled for generalized effects of results are not the results of the testing postprandial triglycerides was described PUFA’s that are not specific to omega- of a hypothesis, but are the origins of a in three papers abstracted by LSRO 3 fatty acids. A better balanced new hypothesis. The authors indicate (Refs. 15, 59, and 163). While the first experimental design would be that the cut-off points for the ratio of two papers used high levels of omega- comparison of drugs plus omega-3 fatty LDL to HDL and triglycerides chosen 3 fatty acids (30 and 9 g of EPA plus acids to drugs plus alternate PUFA’s were to some extent arbitrary. The DHA/day, respectively), the recent (e.g., corn oil). actual number of cardiac events in the paper used only 5 g of fish oil, FDA agrees that the new studies study was small (e.g., 18 events among containing 1.7 g EPA plus DHA. These provide some support for the role of 138 subjects in the highest risk studies showed that the concentration of omega-3 fatty acids in prevention of subgroup), and the reduction in all- plasma chylomicrons after a high-fat test restenosis, although neither was cause mortality due to the lipid- meal was significantly less if the designed to distinguish effects of omega- lowering drug, gemfibrozil, was not subjects had been consuming a fish oil 3 fatty acids from effects of omega-6 significant. Finally, independent of LDL diet than if they had been consuming a PUFA’s. to HDL ratio, increased triglycerides saturated fat or olive oil supplemented Nye et al. (Ref. 259) studied 79 men alone were not associated with an diet. Thus, FDA agrees that fish oils and 29 women who were referred for increased risk of heart attack. reduce postprandial lipomia. angina and underwent coronary However, FDA disagrees that there is The dramatic reduction of percutaneous transluminal angioplasty significant scientific agreement that triglycerides by omega-3 fatty acids has (PCTA), i.e., a mechanical opening of a VLDL and triglycerides are atherogenic, resulted in their use in the treatment of closed heart blood vessel. The subjects or that the reduction in postprandial were randomly assigned to one of three a rare genetic hypertriglyceridemia (type hyperlipemia is a most important anti- treatments: (1) A combination of aspirin V) to prevent noncardiovascular effects atherogenic action. Neither the Federal plus dipyridamole (an anti-platelet of high triglycerides (i.e., pancreatitis), Government nor other authoritative combination of drugs), (2) olive oil but the usefulness of lowering reports have included these blood lipid placebo, or (3) 12 milliliters (mL) fish triglycerides as a general strategy in measures among those they consider to oil containing 3.2 g EPA plus DHA/day. prevention of CHD is not generally be independent risk factors associated Subjects were restudied 1 year later or agreed. Therefore, FDA believes that the with CHD (Refs. 34 through 36, and before if symptoms recurred, and 93 triglyceride-lowering effect of fish oils 115). Furthermore, postprandial lipemia percent of all subjects were followed for for some at-risk persons does not was discussed at the February 1992 the year. Although there was no provide a basis for a health claim at this NHLBI consensus development significant difference in angina among time. conference. The summary of that the groups, the rate of restenosis, 35. Numerous comments indicated conference stated, “Postprandial defined in this study as a loss of 50 that postprandial triglyceridemia is a triglyceride may be more important than percent or more of the luminal diameter mechanism of action in the the fasting triglyceride levels [to CHD], increased by PCTA, was significantly development of atherosclerosis. Some but little is known about this at the less in the fish oil group (11 percent) comments indicated that the present time.” (Ref. 255). than in the placebo group (33 percent). relationship of elevated triglycerides to FDA notes that the only paper in the The use of olive oil as the placebo did risk of CHD would be discussed at the LSRO report cited in support of this not control for effects due to PUFA’s NHLBI consensus development hypothesized mechanism of action of (omega-6). Also, it is notable that the conference (Ref. 255). Others pointed omega-3 fatty acids in the prevention of restenosis rate in the aspirin group was out that LSRO had concluded that CHD was a review paper published in somewhat higher (17 percent) than in elevated very low density lipoproteins 1979 (Ref. 305). Therefore, FDA believes the fish oil group, because aspirin is a (VLDL) and triglycerides were that there is not significant scientific much more potent inhibitor of platelet atherogenic. LSRO stated that the agreement at this time that postprandial function than EPA in fish oil. reduction of postprandial triglycerides are related to the risk of Nonetheless, these results are consistent hyperlipidemia is a “most important CHD. with an effect of omega-3 fatty acids in anti-atherogenic action.” LSRO wrote in reducing restenosis. the summary that, “Since postprandial ii. Vessel wall effects The full “Quebec study” was published after the receipt of the lipemia has been identified as an 36. One comment indicated that two atherogenic risk factor, its prevention by comment, but because it was cited in new studies support the use of omega- w-3 fatty acids would be a most the comment it will be discussed here. 3 fatty acids to prevent restenosis, the desirable effect” (emphasis added), and closing of a mechanically opened blood In this study, Bairati et al. (Ref. 172) in its conclusions LSRO wrote: vessel (Refs. 172 and 259). This conducted a double-blind, randomized Fish oil has a generally accepted comment suggested that FDA intervention with either fish oil hypolipidemic effect without depressing discounted the findings of the Dehmer containing 4.5 g EPA plus DHA/day, or HDL. This applies most to VLDL and study (Ref. 30) on the basis that it olive oil placebo in 205 patients triglyceride, lipids now believed to be employed simultaneous treatment with undergoing first PCTA. The treatments atherogenic. There is little doubt that there is drugs and fish oils. were started 3 weeks before the a reduction of postprandial hyperlipidemia FDA considered the use of omega-3 procedure, and continued for 6 months following the ingestipn of dietary fat if the fatty acids to prevent restenosis to be a after, Restenosis was assessed background diet contains relatively small drug usage (56 FR 60663 at 60670), and angiographically, using a quantitative

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2693

computer analysis program. Restenosis electro-physiologic measurement) in decreased platelet-derived growth factor was reduced in the fish oil group persons from a Japanese fishing village (PDGF); compared to the olive oil group compared to those from a farming decreased superoxide; according to 3 of 4 definitions of village. Other data showed the decreased interleukin-1(TNF); restenosis. It was not reduced according populations differed in their intake of increased endothelium-derived to the clinical definition used by Nye et omega-3 fatty acids. Rapp et al. (Ref. relaxation factor (EDRF); al. (Ref. 259), above, of a loss of 50 268) measured the amount of omega-3 decreased lipoprotein (a) (Lp(a)); percent or more of the luminal diameter fatty acids in the atherosclerotic lesion reduced inflammatory response; and increased by PCTA. after consumption of omega-3 fatty acids increased fibrinolytic activity. This study also collected dietary data. at a high level (6 percent of calories, 16 The LSRO report stated that other The third of the subjects with the to 21 g EPA plus DHA/day) for 6 to 120 mechanisms, such as cellular growth highest consumption of omega-3 fatty days prior to planned surgical factors, interleukin-1 and cytokins, and acids (0.15 g/day) and the third of the intervention, and found that the amount EDRF may be important in the subjects with intermediate consumption of omega-3 fatty acids in the lesion development of atherosclerosis, and be ofomega-3 fatty acids (0.033 to 0.15 g/ continued to increase throughout the affected by omega-3 fatty acids. day) had significantly lower rates of time of ingestion. Force et al. (Ref. 200) However, except for a single in vitro restenosis than the third consuming the studied the effects of fish oils and study on PDGF, no data are described in least amount of omega-3 fatty acids. In aspirin on the production of urinary the report regarding these factors, nor is fact, dietary omega-3 fatty acids (other metabolites of AA and EPA. Fish oil their relevance to human CHD than the supplement) were associated feeding resulted in a slight decrease in discussed. with a greater reduction in chance of the amount of thromboxane A2 made in FDA addresses fibrinogen, Lp(a), and restenosis than was the supplement. the platelet, a decrease in the amount of fibrinolytic activity in comment 46 and This result was somewhat surprising, AA-derived prostacyclin made in the in section II.C.2. of this document. FDA since the supplement contained 30 endothelial cell, and an increase in the does not agree that omega-3 fatty acids times the amount of omega-3 fatty acids amount of EPA-derived prostacyclin produce changes in all of the listed in the diet. No differences in rate of made in the endothelial cell. Schmidt et parameters. FDA has determined that restenosis were found according to al. (Ref. 277) described decreased for some of these endpoints the changes intake of total fat, polyunsaturated fat, Bionocyte chemotaxis among have not been shown to be specific to monounsaturated fat, saturated fat, hypertensive patients after fish oil omega-3 fatty acids, but may be due to cholesterol, or total seafood feeding. The Nye et al. study is polyunsaturated fats instead. FDA consumption. These results suggest that discussed m comment 36 of this disagrees that the changes brought about chronic consumption of low amounts of document. by omega-3 fatty acids will prevent omega-3 fatty acids may be as useful in FDA considers these studies to be atherosclerosis. Most of the data preventing restenosis as much larger observational, not clearly associating regarding changes in these endpoints amounts consumed for a few weeks omega-3 fatty acids with risk of CHD. brought about by omega-3 fatty acids prior to and after PCTA. The correlation data of Hamakazi et al. have been derived from tissue culture or In general, the results of Bairati et al. do not indicate a specific role for animal experiments, and the relevance (Ref. 172) and Nye et al. (Ref. 259) are omega-3 fatty acids. The Rapp et al. data to human atherosclerosis has not been consistent, even though they obtained verify that it is possible to incorporate demonstrated. different results accenting to one omega-3 fatty acids into preexisting Thromboxanes and prostacyclins are identical definition of restenosis. The atherosclerotic plaque, but the relevance compounds derived from omega-3 fatty Bairati et al. study, like Nye et al. 1990, of incorporated omega-3 fatty acids has acids and omega-6 fatty acids that affect used olive oil as the control. If the not been established. The studies of the relaxed state of the blood vessels. mechanism of action of omega-3 fatty Force et al. and Schmidt et al. relate to Thromboxanes are produced primarily acids in restenosis is through a potential mechanism of action of in platelets, and prostacyclins are competition with AA, this control is omega-3 fatty acids, but the importance produced primarily in the endothelial suitable, and an omega-6 fatty acid oil of these actions in reducing risk of CHD cells of the blood vessels. The would have made the difference due to has not been established. thromboxane made from an omega-6 omega-3 fatty acids even more 38. Many comments stated that the fatty acid called AA. thromboxane A2 is pronounced. If, however, the biochemical and physiological actions a potent vasoconstrictor. EPA competes mechanism of action is through of omega-3 fatty acids are anti- with AA for the enzyme that makes nonspecific effects of highly unsaturated atherogenic because they favor thromboxane A2, and thereby fatty acids, then a control of a PUFA vasodilatation and inhibit diminishes the rate of production of (e.g., corn oil) might have reduced the vasoconstriction. One comment by a thromboxane A2; the thromboxane made apparent effect of omega-3 fatty acids. It manufacturer of omega-3 fatty acids from EPA is a much less potent is notable that the only study of considered these actions have potential vasoconstrictor. The prostacyclins made restenosis that has used a for future significance. Two comments from AA or EPA in the endothelial cells polyunsaturated fat control (an olive oil- cited a list of effects of omega-3 fatly are vasodilators. Thus, the relative corn oil mix) did not find an effect (Ref. acids, suggesting that each of the effects amounts of AA and EPA in platelets and 56). in the list was anti-atherogenic, and endothelial cells play a role in 37. Five studies in humans relevant to other comments referred to one or more determining the form and amounts of the action a of omega-3 fatty acids on the of the components in the list. The listed the prostaglandins and thromboxanes vessel wall were referenced in changes were: that affect the tension of the vessel wall. comments (Refs. 200, 213, 259, 268, and decreased thromboxane; Excessive constriction may lead to an 277), including two published since the increased prostacyclin and leukotriene occlusion, resulting in a heart attack. time period covered by FDA’s review in (LTB4); While there is general recognition that its proposed rule (Refs. 200 and 268). decreased fibrinogen, these vasoactive compounds may play a Hamakazi et al. (Ref. 213) found a decreased platelet activating factor role in the formation of clots and slower aortic pulse wave velocity (an (PAF); thereby in heart attacks, there is no

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2694 agreement about the extent of changes agreement that changes in these markers biochemical basis for the hypothesized needed in the concentrations of the produced by omega-3 fatty acids are stabilization of cardiac arrhythmias by vasoactive compounds in order to have causally related to CHD. omega-3 fatty acids. Although, this study an effect on heart disease. Changes in b. Thrombosis and hemostasis (Ref. 212) was performed, in vitro on heart cells from rats, it showed that the the amounts of these vasoactive 39. A few comments stated that the protective effect was specific to omega- compounds, produced by consumption mode of action of omega-3 fatty acids of fish oil, are only useful as marker for 3 fatty acids (EPA) because a similar may be through stabilization of CHD only insofar as there is significant effect was not obtained when a highly arrhythmia, and noted the reduced rate scientific agreement that the magnitude unsaturated omega-6 fatty acid (AA) was of death after heart attacks (MI’s) in the of the changes is related to CHD. FDA used instead. Dart study (Ref. 16). This comment also is not aware of any such agreement, nor FDA also regards the evidence stated that certain animal data were did the comments provide any evidence the Burr study of reduced death of agreement that particular changes In consistent with. this hypothesis. The following a heart attack among the levels of these vasoactive comments stated that the fibrillation men, advised to increase fish compounds were related to a reduction mechanism suggested by DART was consumption as consistent with a in risk of disease. Furthermore, the compelling, because 60 percent of stabilization of arrhythmias (Ref. 16). amount of omega-3 fatty acids needed to sudden deaths are caused by ventricular FDA agrees that this postulated produce these changes in humans is not fibrillation following reperfusion. Many mechanism of action is of great potential known. commented that data from nonhuman public health significance. However, the For PDGF the evidence is confined to primate models show that omega-3 fatty agency finds the clinical data available- acids abolish arrhythmias, whereas animal studies (Ref. 201), and the at this time are not in agreement with polyunsaturated fat (safflower oil) had a relevance to human disease has only animal and in vitro data. Because the lesser effect. been suggested, not demonstrated. The clinical data are not in agreement with FDA disagrees with these comments. animal studies on PDGF also did not these other types of data and because of FDA’s review of the literature regarding show that the effect was specific to the limitations in the animal studies, the usefulness of omega-3 fatty acids in ornega-3 fatty acids. For example, the FDA concludes that there is not arrhythmia and ventricular fibrillation PDGF effect was observed also after sufficient basis for protective effect found only one study on arrhythmias in polyunsaturated fats, and was abolished specific to omega-3 fatty acids on humans, and it reported no significant by anti-oxidants, suggesting that any arrhythmias, and, therefore, CHD in effect of omega-3 fatty acids (Ref. 58). A highly un saturated fatty acids prone to humans. review in 1989 also concluded that, oxidation would have the effect. The 40. One comment criticized the 6- even among the animal studies, there experiments on EDRF (Ref. 181) also did was no significant difference between week clinical study by Hardarson et al. not show that the effects were specific omega-3 fatty acids and other that found no effect of omega-3 fatty to omega-3 fatty acids, since the polyunsaturated fats on arrhythmias acids on arrhythmias (Ref. 58), arguing experiments were earned out in the (Ref. 269). that the time for incorporation of omega- presence of indomathacin, which, blocks 3 fatty acids into heart phospholipids The data from the studies in the eicosanoid effects of EPA. In fact, was too short for an effect to be Nonhuman primates (i.e., the marmoset the authors consider changes in observed. monkey) were published only as a membrane fluidity to be a reasonable FDA agrees in part mid disagrees in explanation for the effects. In yet other Nonpeer-reviewed paper in a book (Ref. 188). Two papers by the same author on part with this comment. Generally, the cases, e.g., TNF, there are conflicting time needed for incorporation of omega- results depending on the species (Refs. the same topic were cited in 1990 as in press in a peer-reviewed journal, but 3 fatly acids into heart phospholipids 41 and 236), and the findings must be is short; studies in animals show such have not yet been published. Therefore, considered preliminary. incorporation, in a period of weeks (Ref. FDA considered the effect of omega- FDA regards the data on nonhuman primates as preliminary only. 249). In the Hardarson study (Ref. 58), 3 fatty acids on chemotaxis, one aspect a substantial amount of cod liver oil was of inflammatory response (56 FR 60663 Furthermore, the data for the marmoset monkey were obtained after prolonged fed (20 mL/day) and a 230 percent at 60670). A complete discussion of the increase in plasma phospholipid EFA role of fish oils in inhibition of the feeding for 12 or 24 months with a supplement of DHA-rich fish oil at a was found. There was no trend toward inflammatory process is outside of the reduced arrhythmias. Other data, scope of this rulemaking, but the level of 8 percent of the diet by weight. FDA calculates that this would provide however, show that although plasma relationship between omega-3 fatty phospholipids increase the ornega-3 acids and inflammatory response could 2.5 g of omega-3 fatty acids from fish oil/kilogram (kg), over 50 times the fatty acid content during the first few be the subject of a petition for a health weeks of supplementation, the claim that includes the necessary usual rate of supplementation in human studies (10 g fish oil or 3 g omega-3 fatty incorporation of omega-3 fatty acids in information about this relationship. human atherosclerotic plaque continues FDA agrees that the biochemistry of acids/day for a 70 kg subject), and over to increase through 120 days (Ref. 268), the products formed from the omega-3 300 times the amount of omega-3 fatty Therefore, FDA agrees with the fatty acids in vivo (i.e., eicosanoids) acids associated with reduced risk of comment the supplementation period in have been shown under experimental CHD in the epidemiologic literature the Hardarson study (Ref. 58) may have conditions, usually in vitro, to have (Refs. 16, 38, and 87 report 300 to 660 been too short to find an effect of fish pronounced effects on the vessel wall. milligrams (mg)/day). Thus, the However, demonstration of isolated relevance of these studies to omega-3 oils on occurrence of arrhythmias. Also, biochemical effects is not a sufficient fatty acids in the human diet is the agency notes that the absence of a difference in CHD mortality during the basis upon which to make a claim questionable. first 6 weeks of the Burr study (Ref. 16) regarding the outcome of a FDA is aware of in vitro data that is consistent with the hypothesis that multifactorial process. Intermediate show a specific protective effect of EPA prolonged intake of omega-3 fatty acids markers of CHD are useful only insofar against toxicity of heart muscle cells in (longer than 6 weeks) is needed to as there is significant scientific culture. These results provide a observe an effect on arrhythmias or

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2695

other mechanisms that reduce CHD of death and/or nonfatal heart attack in regarding changes in platelet function mortality. FDA agrees that effects of patients with previous infarction or following omega-3 fatty acid long-term consumption of omega-3 fatty unstable angina pectoris as a consumption. LSRO concluded that acids on arrhythmias, other platelet or professional labeling indication omega-3 fatty acids prevented platelet vessel wall functions, and even some (provided to health professionals, but aggregation. blood lipid measures have not been not to the general public), in the In its proposal, FDA stated: “Platelet sufficiently studied. tentative final monograph for over-the- aggregation is generally considered to be i. Bleeding times counter internal analgesic, antipyretic, decreased by fish oil consumption.” (56 and antirheumatic drug products FR 60663 at 60670.) The agency also 41. Two comments stated that there is (November 16, 1988, 53 FR 46204 at stated: “* * * platelet aggregation and no evidence of increased bleeding even 46259). However, FDA does not function are reduced: However, direct among patients who had invested 6 to consider the effects of aspirin in the relationships between the changes in 8 g of EPA plus DHA/day and Physicians’ Health Study sufficient to * * * platelet function and risk or CHD underwent emergency surgery, coronary establish that dietary omega-3 fatty have not been established.” (56 FR artery bypass surgery or angioplasty. acids would have the same effect in the 60633 al 60671.) Thus, FDA agrees with The comments argued that increased general population. The Physicians’ the conclusions of LSRO about effects of bleeding has not a safety concern. Health Study did not evaluate omega-3 omega-3 fatty acids on platelet FDA agrees that there are few reports fatty acids. The study population was aggregation. of excessive bleeding after ingestion of highly selected; the rate of heart attacks Two of the studies described by LSRO omega-3 fatty acids. However, FDA was approximately 10-fold lower than were not considered by FDA in its notes that the cited reports are for in the general population, and review, because they were published subjects with CHD, and evidence of the cardiovascular mortality was only 15 before 1988, and had been considered lack of excessive bleeding complications percent of that expected for the general by Federal Government and other in this population is not sufficient to population of white men of the same authoritative reports. One study (Ref. assure safety of omega-3 fatty acids in age. Also, the results of the Physician’s 227) used a large amount of fish oil (50 the general population. FDA believes Health Study are not as straightforward mL/day) not reasonably related to that changes in bleeding due to as presented in the comments. The normal dietary intake. The other study consumption of omega-3 fatty acids chairman of the Physicians’ Health (Ref. 211) involved 13 insulin- remains a valid safety concern (see Study reported that there was a reduced dependent diabetics, and therefore is of comment 52 of this document). risk of MI in the aspirin group, questionable relevance for the general ii. Platelet aggregation predominantly in nonfatal MI, but that population. In the proposal, FDA stated: there was no significant effect on overall In its proposed rule, FDA considered The relationship between platelet cardiovascular mortality (a 2 percent 13 of the other 17 studies that were aggregation and the risk of heart attacks or reduction, not statistically significant) abstracted by LSRO. One of the four CHD death in the general population is an (Ref. 66). In addition, the aspirin group studies not addressed by FDA was a important line of evidence that would in this study had a greater number of study on the effects of added vitamin E support drug claims and perhaps health sudden deaths (Ref. 282). to fish oil on fibrinogen and fibrinolysis claims for omega-3 fatty acids. Although In the other primary prevention trial there is some evidence that changes in (Ref. 210). Two papers (Refs. 234 and platelet aggregation may help prevent second (Ref. 265), aspirin did not have any 244) were published after the time heart attacks * * * it has not been shown that significant effect on heart attacks, on period covered by FDA review. changes in platelet aggregation in the general stroke, or on total vascular mortality. Marckmann et al. (Ref. 244) compared population will reduce the risk of CHD. There was a significant increase in the effects of a fish diet and a lean meat (56 FR 60663 at 60670.) disabling stroke in the group taking diet on plasminogen activator (t-PA), The agency added: “What has not been aspirin. plasminogen activator inhibitor (PAI-1) established, however is that platelet On the basis of these studies there has and the activity of the inhibitor (PAI-1 aggregation is a bona fide surrogate risk not been an endorsement of the use of activity), Li and Sterner (Ref. 234) factor for CHD in the general aspirin as a prophylactic measure described changes in in vitro platelet population,” (56 FR 60663 at 60672.) against CHD by the general population adhesion after fish oil supplementation. 42. Many comments argued that by the American Heart Association or by The fourth paper was an uncontrolled platelet aggregation is completely the Canadian Medical Association (Ref. observation study that found a high substantiated as a marker for risk of 187). Notably, “1992 Heart and Stroke frequency of nosebleeds in adolescents CHD, based on the results of the Facts” published by the American Heart supplemented with fish oils (Ref 189). Physicians’ Health Study (Ref. 66). One Association (Ref. 169) makes no Six other papers on thrombosis were comment qualified this conclusion reference to platelet aggregation as a risk not described in the LSRO text, but were stating that the primary effect of omega- factor for heart attacks (although sticky included in the table (Refs. 203, 204, 3 fatty acids in vivo was to reduce platelets are mentioned to be a 209, 226, 245, and 254). Of these six, platelet deposition at sites of aortic consequence of cigarette smoking in the one was not relevant to the nutrient- lesions. section on stroke), nor is aspirin disease relationship (Ref. 245) because it FDA acknowledges that aspirin discussed as an option for CHD did not study EPA and DHA. Jensen et studies provide evidence that platelet prophylaxis, even though other drug al. (Ref. 226) found no significant aggregation is a risk factor for CHD. The and surgical treatments are discussed. change in bleeding times in normal effect of aspirin in inhibiting platelet Therefore, FDA concludes that there subjects after 1, 3, or 6 g EPA plus DHA/ function has been shown. Among is not significant scientific agreement at day in healthy subjects. Green et al. persons who have already had an MI, this time that platelet aggregation is a (Ref. 209) found no change in platelet aspirin is effective in preventing a surrogate marker for CHD in the general aggregation or platelet count in 27 second infarction. FDA has proposed population. hyperlipidemic subjects in a that aspirin be used to reduce the risk 43. The LSRO report, submitted as a randomized double-blind placebo comment, contained abstracts of 19 controlled crossover trial. The studies in humans that contained data treatments were 15 g/day fish oil

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2696 containing 4.3 g EPA plus DHA and a regard to the development of found a decrease compared to olive oil 50:50 mix of corn:olive oil, with each atherosclerosis and the risk of CHD. (Ref. 49) and one found a decrease treatment lasting 8 weeks, and a 4-week However, as for platelet aggregation, compared to soybean oil only when 30 washout between. Blood viscosity was FDA does not believe that there mL of fish oil were consumed, but not decreased by fish oil, Gazso et al. (Ref. currently is significant scientific when 15 mL were consumed (Ref. 57). 204) found decreased platelet agreement that platelet adhesion is an Additional well-designed studies not aggregation in healthy subjects after accepted risk factor for CHD in the cited by LSRO, but considered in the consumption of EFAmol-marine general population. FDA proposal, reported no change (Ref. compared to olive oil in a double-blind iii. Regulators of bleeding 24) and an increase (Ref. 131) in randomized crossover study. These In its proposal (56 FR 60663 at 60670 fibrinolytic activity. results of studies confirm others cited through 60671), FDA reviewed data on Therefore, FDA stands by its earlier by FDA. The other studies (Refs. 203 the effects of omega-3 fatty acids on conclusion that the publicly available and 254) pertained to regulators of other factors that are involved in the scientific evidence does not support a bleeding and are discussed below. regulation of bleeding—fibrinogen, relationship between omega-3 fatty Eight papers on platelet fraction were fibrinolytic activity and Lp(a)—and that acids and decreases in fibrinogen or reviewed by FDA but not by LSRO have been associated with CHD. increases in fibrinoiysis. This (Refs. 2, 6, 18, 24, 73, 93, 131, and 143). 45. Some comments, including the conclusion is supported by findings that Three studies were uncontrolled (Refs. consumption of other PUFA’s have LSRO report, stated that omega-3 fatty 18, 93, and 143), while two were acids increase fibrinolysis. effects comparable to those produced by randomized (Refs. 2 and 131). Three consumption of omega-3 fatty acids. FDA disagrees with these comments. were randomized, double-blind, FDA found that there was no clear 46. Two comments cited unpublished placebo-controlled trials that used data by Kostner and Herrmann, relationship between omega-3 fatty saturated vegetable oil (Ref. 6), vitamin reporting reduced Lp(a) after fish oil acids and factors involved in dissolving E (Ref. 24) or wheat germ oil (Ref. 73} consumption. blood clots (56 FR 60663 at 60671). FDA as the placebos. The two studies that FDA was unable to find the full paper noted that the data did not establish that used vegetable oil or vitamin E as by these authors showing the decrease controls found a reduction m platelet omega-3 fatty acids reduced fibrinogen, in Lp(a). FDA did find a paper by these aggregation after omega-3 fatty acids, because most studies did not control for researchers published in 1991 (Ref. 241) where no difference was reported in the other factors that might have reduced that reported no effect of fish oils on trial that used a wheat germ oil placebo, fibrinogen, e.g., other PUFA’s. Lp(a) and did not cite conflicting work although the data were not provided in FDA has reviewed the relevant from their laboratory. studies again, as well as studies brought this paper. iv. Blood pressure FDA and LSRO reached the same to its attention in the LSRO report. In its proposal, FDA considered the conclusions with regard to the effects of LSRO cited three papers on fibrinogen or fibrinolysis not cited by FDA. One relationship between omega-3 fatty omega-3 fatty acids on platelet function. acids and blood pressure, one of the LSRO also concluded that platelet placebo (vitamin E) controlled study found no change in fibrinolytic activity recognized risk factors for CHD. FDA survival is also enhanced, but the only stated: two studies published since 1987 that (Ref. 210). Mullertz et al. (Ref. 254) reported increased platelet survival supplemented seven healthy adults with These results for effects: of omega-3. fatty acids on blood pressure of normal subjects (Refs. 94 and 144) were both 0.55 g ERA plus DBA/day for 21 days and found increased levels of PAI-1, but are ambiguous. Some studies found a uncontrolled, so the effect cannot be reduction in. systolic blood pressure after attributed specifically to omega-3 fatty no change in t-PA, suggesting that fish consumption of fish oils containing omega-3 acids. oil decreased fibrinolytic capability. fatty acids, whereas others did not. None of 44. One comment agreed in principle Cans et al. (Ref. 203) reported no change the studies found a significant reduction in with the agency’s assessment of the in fibrinogen concentration after diastolic blood pressure. Therefore, it also effects of omega- 3 fatty acids on EFAmol-marine compared to corn, oil, remains to be established that the normal, bleeding times, platelet aggregation, which is rich in polyunsaturated fat. healthy population will reduce their risk of CHD via a reduction in blood pressure regulators of bleeding and blood These studies do not support the conclusion that omega-3 fatty acids following consumption of omega-3 fatty pressure. The comment, pointed out the acids. extent of inhibition of platelet adhesion reduce fibrinogen, or increase (56 FR 60663 at 60671.) is as much, as 60 percent, a marked fibrinolysis. reduction in adhesion. The selection of studies abstracted by FDA also stated that it was not known FDA agrees that the reported extent of LSRO may not have represented the whether or not the magnitude and reduction of platelet adhesion publicly available scientific evidence. duration of the effect would persist after omcga-3 fatty acid intake is remarkable For example, five papers abstracted longer term supplementation. FDA (Ref. 234). The agency notes that this found either a decrease in fibrinogen or recognized that studies among effect appears specific to omega-3 fatty an increase in fibrinolytic activity (Refs. hypertensives found an effect more acids at reasonable intake levels. FDA 57, 71, 98, 104, and 117). In contrast, consistently than studies among normal notes that animal studies (Refs. 230 two studies found no change in subjects, although sometimes large amounts of fish oils were used. through 233) published since the fibrinolytic activity (Refs. 150 and 166), proposed rule provide evidence of and only one found increased fibrinogen 47. Some comments considered the reduced platelet adhesion to blood (Ref. 144), leaving the impression that effects of omege-3 fatty adds on vessel endothelium in viva in response omega-3 fatty acids usually have been hypertension as evidence, of a reduction to agents that provoke such adhesion. reported to enhance fibrinolysis. in CHD risk. Other comments called for Because of the magnitude of the However, three other studies not FDA to reassess the studies on blood of omega-3 fatty acids on platelet abstracted by LSRO but included in Pressure. One of these comments adhesion, FDA considers this action of their tables reported no effect of omega- that the results of studies on blood omega-3 fatty acids on blood platelet 3 fatty acids on fibrinogen compared to pressure are not “completely function to have great potential with corn oil (Refs. 10, 118, and 203). One

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2697

ambiguous. One comment agreed in (Ref. 299), where subjects remained on did not show that omega-3 fatty acids principle with the agency’s assessment blood pressure lowering medications were more effective than other of the blood pressure studies. One and no effects of added fish oils were polyunsaturated fats. comment considered a number of observed. 48. Comments stated that other lines animal models to be relevant for FDA disagrees with the LSRO of evidence were not discussed in the hypertension. The LSRO report also descriptions of the Singer and Cobiac proposal. Examples given were changes considered the evidence relating omega- studies. The placebo in the Singer study in plasma viscosity, increased vascular 3 fatty acids to blood pressure to be (Ref. 285) was olive oil, but this was not compliance, and reduced white blood important in relation to CHD. The LSRO pointed out in the LSRO text. The cell (WBC) count. report concluded that, “Fish oil reduction of blood pressure observed FDA disagrees with the comment with probably has a mild hypotensive effect, after fish oil, therefore, may have been respect to plasma viscosity and vascular especially in high doses.” due to a general unsaturated fatty acid compliance. In its proposed rule, FDA FDA disagrees that the publicly effect not specific to omega-3 fatty acids. acknowledged that plasma viscosity was available scientific evidence supports a In the description of the Cobiac et al. decreased and red cell deformability relationship between omega-3 fatty study (Ref. 190), LSRO did not note that was increased by omega-3 fatty acids, acids and hypertension. At best, as fish oil treatment alone (without but that the importance of these effects stated in the proposal, the data are simultaneous reduction of salt) had no on the risk of CHD had not been ambiguous. Qualifiers are needed to effect on blood pressure. established (56 FR 60663 at 60670). indicate that the reductions in blood Two other studies were cited in the The agency agrees that it did not pressure have not generally been shown LSRO tables but not in the text and were systematically consider WBC count to be specific to omega-3 fatty acids. not included in the FDA review. Neither among the effects produced by omega- Also, many valid studies have reported of these found an effect on blood 3 fatty acids. WBC count was not no effect. pressure. Gans et al. (Ref. 203) used a included among the actions of omega-3 LSRO reported a total of 13 studies on randomized double-blind, placebo- fatty acids considered in major reviews. hypertension. Four were published controlled design and found a reduction FDA notes that WBC count has only before 1988, and were not reviewed by in diastolic blood pressure for both fish recently been identified as associated FDA in the proposed rule. Three of oil and corn oil (placebo). Meland et al. with risk of CHD by the Caerphilly these studies used fish as the source of (Ref. 247) carried out a randomized, Collaborative Heart Disease Study (Ref. omega-3 fatty acids and therefore did double-blind multicenter trial among 40 301a). Only two papers among literature not show the effect specifically to be mildly hypertensive subjects, using 6.8 from 1988 to present have reported a due to omega-3 fatty acids. In fact, in g EPA plus DHA/day, but found no reduction of WBC count after fish oil one study (Ref. 292), the control diet of difference in blood pressure compared supplementation (Refs. 183 and 253). meat produced a decrease in blood to a 50:50 olive:corn oil control. pressure comparable to that of the fish Three other large and appropriately 3. Other relevant information controlled studies not in the text of the diet. The study that used fish oils (Ref. a. Animal studies 271) used an olive oil control, rather LSRO report but included in its table than an oil high in PUFA’S. This study were also reviewed by FDA. Two 49. Numerous comments asserted that is the only study to show an effect of randomized studies on normal subjects animal studies did not receive an omega-3 fatty acids on diastolic blood (Refs. 9 and 49) and one controlled appropriate amount of discussion. One pressure in normal subjects. study among mildly hypertensive of these same comments stated that Of the other 10 studies on subjects (Ref. 20) reported no animal studies are not sufficient to hypertension described in the LSRO differences in blood pressure support the claim, and that clinical report, 6 were also reviewed by FDA attributable to omega-3 fatty acids. trials on effects of omega-3 fatty acids (Refs. 11, 57, 80, 85, 95, and 101). The FDA reviewed in its proposal three directly on CHD are needed. One LSRO and FDA interpretations of the other randomized, double-blind, comment criticized FDA’s review of results from these papers did not differ placebo-controlled studies among animal studies because the negative in any significant regard, except that healthy subjects that were not included findings have been in inappropriate FDA specifically noted that two of these in the LSRO review. Two of these models and should not have been studies (Refs. 85 and 95) used very high studies were on normal, healthy discussed. Another comment stated that amounts (50 mL) of fish oil to show the subjects (Refs. 6 and 24) and found a they did not believe that there is an effect. In fact, FDA singled out the decrease in systolic blood pressure appropriate animal model for human Bonaa et al. (Ref. 11) and Kestin et al. compared to a saturated vegetable oil or cardiovascular and CHD. The LSRO (Ref. 80) studies as well-designed vitamin E, respectively. The third study report considered animal studies to studies that showed an effect specific to (Ref. 73) found that omega-3 fatty acids provide important evidence for an anti- omega-3 fatty acids in hypertensive and did not affect blood pressure in atherogenic effect of fish oils, stating, normal subjects, respectively (56 FR hypertensives or normal men compared “Omega-3 fatty acids have been shown 60663 at 60671). to wheat germ oil. to retard the development of the The LSRO report reviewed four Therefore, FDA concludes that the atherosclerotic plaque in experimental papers not originally reviewed by FDA evidence of an effect of omega-3 fatty animals including the pig and rhesus (Refs. 190, 285, and 299), including one acids on blood pressure in normal monkey.” study on linolenic acid outside of the subjects is ambiguous, because some FDA agrees that the evidence from scope of the definition of omega-3 fatty studies reported a blood pressure studies in animals warrants additional acid, as used in this regulation (Ref. lowering effect, whereas other equally discussion. FDA has reviewed here 262). Two other papers that appeared in well-designed studies found no specific those animal studies that were cited in the LSRO table but not in the text (Refs. effect. Studies among hypertensives its proposed rule and those that were 203 and 247) were also not reviewed by found an effect more consistently than cited in the LSRO report that were FDA in its proposal. studies among normal subjects, relevant to the development of FDA agrees with the LSRO although sometimes large amounts of atherosclerosis. Other animal studies interpretation of the Wing et al. study fish oils were used, and many studies relevant to the development of

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2698 atherosclerosis, and animal studies on were scored, including various human diets do not alter serum aspects of CHD other than measures of the extent of plaque, loss of cholesterol concentrations, the results atherosclerosis are reviewed under endothelium, intimal thickening, and from these animal experiments are of section II.C.3.a. or this document. inflammation. Overall there was no questionable importance regarding In its proposal, FDA cited eight benefit of fish oil; in some cases, the human atherosclerosis. Finally, LSRO animal studies and one abstract on the atherosclerotic measure indicated more did not review numerous animal studies development of atherosclerosis that disease in the fish-oil fed animals. that found no effect or an adverse effect were not included in the LSRO review The LSRO report considered the of supplementation with fish oils, and (Refs. 19, 51, 65, 81, 97, 123, 126, and amount offish oil in the diets in this therefore the LSRO conclusion does not 151); seven of which reported either no experiment (1.3 to 1.8 percent of represent the totality of publicly beneficial effect or an adverse effect in calories) too low to observe an effect. In available scientific information. fish oil supplemented animals. Only fact, FDA calculates a lower percent of With these qualifications in mind, one animal study on the effects of calories from fish oil (1.0 to 1.5 percent) FDA notes that some of the reported omega-3 fatty acids on restenosis was than calculated by LSRO. This level is effects of the dietary interventions with abstracted by LSRO, although the others about half the amount used in short- fish oils on the development of cited by FDA were described in the term human studies (i.e., 10 mL/day), atherosclerosis have been dramatic. LSRO tables. and FDA agrees that the low level makes Also, FDA recognized that animal Three studies in nonhuman primates it less likely that an effect would be studies are of great importance for study have been reported (Refs. 27, 47, and observed than if a higher amount had of long-term effects on chronic diseases 116). In the Davis et al. (Ref. 27) and been used. However , diets were of consumption of amounts of omega-3 Parks et al. (Ref. 116) studies, the supplemented for a prolonged period of fatty acids, particularly in amounts that polyunsaturated fat intake was higher in time (20 months) and in the therapeutic might be obtained in a reasonable diet. the fish oil groups, and polyunsaturated diet other dietary factors were also Therefore, FDA encourages further fat is known to lower total plasma changed that might have made the research in this area using rigorous cholesterol. Also, the control diet of the effects of omega-3 fatty acids more study designs and amounts of omega-3 Davis et al. study had more saturated fat noticeable (e.g., ratio of polyunsaturated fatty, acids reasonably available in a than the fish oil diet. Thus, the effects to saturated fatty acids). Finally, the fact normal diet to elucidate any effects on atherosclerosis may not have been that there were differences between the specific to these fatty acids. specific for omega-3 fatty acids. fish oil-supplemented group and the After closer scrutiny of the animal In these studies, the total cholesterol polyunsaturated fat group while on studies cited in FDA’s proposal and in values for the fish oil groups were either the atherogenic or therapeutic the LSRO report, the agency has reached substantially lower than for the control regimens suggests that there was the same conclusion that it reached in group, which may explain the observed sufficient sensitivity in the experimental its proposed rule: there are some data in differences in atherosclerosis. In design to detect protective effects of studies from animals which, suggest the support of this interpretation, Parks et omega-3 fatty acids. possibility of a beneficial effect of al. (Ref. 116) noted that one of monkeys In some animal studies that showed a omega-3 fatty acids on CHD; however, fed the fish oil diet responded protective effect of fish oils, an invasive the data are equivocal. (56 FR 60663 at differently than the other 11 monkeys procedure was used to accelerate 60671.) fed fish oil. This one monkey had a atherosclerosis, either mechanical injury b. Safety considerations plasma cholesterol level comparable to (Ref. 122) or vein grafts (Refs. 90 and 50. One comment stated that the that of the lard-fed control monkeys, 186). These studies may be most increases in LDL cholesterol observed and also had atherosclerosis comparable relevant to the late stages of to the lard-fed monkeys. atherosclerosis, and to CHD in humans were a chance occurrence, and another Changes in total cholesterol levels following invasive procedures. All of stated that increased LDL should not be were noted by authors of another study the animal studies cited in the LSRO considered an adverse finding in light of the results of the Burr study. in pigs that showed a reduction in report except Kim et al. (Ref. 81) and FDA disagrees with this comment. atherosclerosis concomitant with a Fincham et al. (Ref. 47) did not control FDA found that increased LDL reduction in time-weighted total for PUFA’s, so the effects observed have cholesterol was ordinarily found when cholesterol (Ref. 81). Since cholesterol not been shown to be specific to omega- hyperlipidemics or diabetics were given concentrations are not changed by fish 3 fatty acids. Additionally, the level of fish oil supplements. This may be due oils in humans, animal studies where use of fish oils has been high, e.g., 22 in part to the fact that fairly large fish oil treatment lowered total percent of calories in Parks et al. (Ref. amounts of omega-3 fatty acids (i.e., 5 g cholesterol levels are of questionable 116) and 25 percent of the diet in Davis relevance to the role of omega-3 fatty et al. (Ref. 27), which limits the EPA plus DHA/day or more) were used in these studies. Increased LDL is not acids in the development of human extrapolation of findings in these atherosclerosis. studies to levels that might be ordinarily seen in the studies on normal A recent study (Ref. 47) in nonhuman reasonably consumed by humans. subjects. primates (vervet) compared the effects FDA disagrees with the summary of FDA does not consider the Burr study offish oil supplementation to sunflower the literature in animals, as expressed in (Ref. 16) to have established that omega- oil supplementation in either an the LSRO-report, because that report 3 fatty acids reduce the risk of CHD, and atherogenic diet (high fat, low fails to mention important limitations in therefore remains concerned that polyunsaturated fat to saturated fat the data. FDA notes that most studies increases in LDL cholesterol could be ratio, high cholesterol) or, following the did not have an adequate design to adverse for some subjects. FDA notes atherogenic diet, in a therapeutic diet show specificity of effects as due to that concern about increased LDL (low fat, high polyunsaturated fat to omega-3 fatty acids. Furthermore, cholesterol was expressed in the report saturated fat ratio and low cholesterol). reductions in total cholesterol in the of the NHLBI concensus development Animals in each diet group were fish oil fed animals may explain the conference (Ref. 255). matched for serum cholesterol. Sixteen reported reductions in atherosclerosis. 51. One comment stated that it was separate measures of atherosclerosis Since reasonable amounts of fish oils in inappropriate to consider adverse effects

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2699 in subpopulations without describing oxidized products, may have adverse Also, the possibility of increased rates of the advantages of omega-3 fatty acids in effects. stroke are raised by the data from those same populations. FDA agrees that oxidation of omega- studies on aspirin (Ref. 66). FDA disagrees with this comment. As 3 fatty acids is a concern. In fact, there Thus, FDA considers these potential noted in the final rule on general are many studies that have been adverse reactions to be legitimate requirements for health claims, reported since the publication of the concerns, primarily in the context of published elsewhere in this issue of the proposed rule, or that were not included very high intakes of omega-3 fatty acids. Federal Register, it would be a violation in FDA’s literature review, that indicate 57. One comment stated that even if of the agency’s responsibility under the such a concern (Refs. 184, 210, 217, 229, adverse effects were only suspected in act to authorize a health claim about a 240, 248, 263, 283, and 290). a medical disorder, pronounced substance without being satisfied that Antioxidants have been successfully the use of the substance was safe. The added to supplements and may be warnings or contradictions would be agency attempted to examine ail adequate to protect the omega-3 fatty required. available scientific evidence regarding acids in foods. It may be necessary to As noted above, the agency must be the effects of omega-3 fatty acids. FDA establish conditions that protect against satisfied that the use of a substance is separated out the potential adverse oxidation of omega-3 fatty acids and safe before it will authorize a health effects discovered during its review, incorporate those conditions into any claim about the substance. Thus, because it wanted to draw attention to future regulation authorizing health suspicions about potential adverse these issues as impediments to a health claims for omega-3 fatty acids. effects would need to be resolved prior claim for omega-3 fatty acids and CHD. 55. A related comment indicated that to the authorization of a claim. Certain Such potential adverse effects must be the majority of the fish oil preparations health claims may require appropriate resolved, and may be important in that have been used are severely qualifications as a way of minimizing setting the conditions under which FDA oxidized, including National Institutes potential safety concerns. of Health Fish Oil Test Materials. would allow a health claim to appear on C. New Scientific Data the label and labeling of foods and food However, no data regarding the extent of supplements. oxidation, the nature of the oxidation To determine whether or not new 52. Two comments stated that the products, or the physiologic action of scientific data published since the safety issues raised in the Mitre Corp. these products was provided. proposed rule provided a basis for report (Ref. 72) were outdated but did FDA agrees with this comment. Many modifying FDA’s conclusions regarding not indicate which issues, or suggest of the biologically active products of the relationship between omega-3 fatty why they were outdated. omega-3 fatty acids are oxidation acids and risk of CHD, FDA conducted FDA recognizes that there has been products. Oxidation of test materials a search of the scientific literature for considerable debate regarding the may explain some contradictory relevant studies. Reviews published clinical importance of bleeding times findings in the literature. since the period covered in the since the publication of the Mitre Corp. 56. One comment pointed out that literature review in FDA’s proposed rule report (Ref. 72). However, the agency increased prothrombin times and were used to identify recently published believes that the issues raised in that possibility of increased stroke were not studies. report have been restated in subsequent discussed. 1. Epidemiologic studies literature, and that all issues of safety FDA agrees with flits comment. FDA a. Cross-sectional studies and surveys are important in deciding whether or did not specifically review data on (Table 1) not to authorize a health claim. prothrombin times, although data on Bulliyya et al. (Ref. 185) found lower bleeding times as a measure of 53. A few comments recommended total serum cholesterol and higher HDL hemostasis were discussed for both that FDA balance the benefits of cholesterol in a fish-consuming coastal reduced risk of CHD against the risk of normal subjects and for subjects with village population than in a nonfish- reduced glycemic control among risk factors for CHD. The importance of consuming population from the interior diabetics when deciding whether or not the increase in bleeding time brought of India. These correlational data are about by supplemental fish oils or to authorize a health claim. One consistent with a beneficial effect of increased fish consumption is not clear. comment stated that physicians could omega-3 fatty acids on blood lipids, but adjust the dose of insulin if omega-3 FDA noted in the proposal that most reports suggest that serious bleeding is many possible confounding variables fatty acids reduced their glycemic prevent strong conclusions regarding a control, but another comment stated not an issue in patients supplemented with omega-3 fatty acids, and also that specific role for omega-3 fatty acids. that glycemic control must be In a retrospective study, Popeski et al. considered a real adverse effect. standardized bleeding times do not closely correlate with clinically (Ref. 266) found that women from FDA agrees that limitations on the use significant bleeding. However, concerns communities with higher marine oil of a substance by a subpopulation (e.g., about untoward bleeding after consumption had significantly lower diabetics) do not necessarily exclude a supplemental fish oils have been raised diastolic pressure in the last 6 hours of substance from bearing a health claim in the literature (Refs. 106,120, and pregnancy than women from for the general population, because the communities with low fish oil 189). claim may be appropriately restricted. consumption. Pregnancy associated However, FDA agrees that the loss of FDA did not discuss the possibility of increased occurrence of stroke as a hypertension was 2.6 times more glycemic control is a potentially serious common in communities with low fish consequence of increased consumption adverse effect that must be fully consumption. These correlational data addressed before a health claim could of omega-3 fatty acids. The papers that reported a correlation between high are consistent with an effect of omega- be authorized. 3 fatty acids on blood pressure in this 54. Another comment stated that a consumption of omega-3 fatty acids from fish and other marine animals and particular situation. Again, many major concern about omega-3 fatty acids possible confounding variables prevent low rate of CHD mortality also noted an not mentioned in the proposed rule is strong conclusions regarding a specific increase rate of stroke, particularly that they may be oxidized and, as role for omega-3 fatty acids. hemorrhagic stroke (Refs. 8 and 84).

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2700 b. Prospective studies (other than cholesterol after switching from a meat The relative importance of various intervention studies) (Table 1) diet to a fish diet, but the fish diet had subfractions of LDL partlicles (and the Bjerregaard and Dyerberg(Ref. 176) significantly less saturated fat than the associated composition of the particles), reported age-standardized mortality meat diet. One study (Ref. 283) found a however, is still controversial. While rates per 10,000 person-years for CHD in slight increase after 5.4 g EPA plus Homma et al. (Ref. 224) suggest that men in Greenland settlements (5.3) as DHA/day from MaxEPA (with 30 large, light LDL are the fraction half of that reported for men in Denmark percent saturated fatty acids), and one associated most closely with (10.0). There was an increasing rate of study (Ref. 224) found a slight reduction atherosclerosis, Austin et al. (Ref. 171) CHD from settlements to towns in in total cholesterol after 2.7 g purified report that the phenotype of small, Greenland. The difference in rates of EPA/day, but neither study was placebo dense LDL is the fraction most closely CHD in women were less apparent, with controlled for effects of polyunsaturated related to increased CHD risk. The lower rates in Denmark than in towns in and saturated fat contained in the February 1992 NHLBI consensus Greenland. These studies do not have supplements. development conference (Ref. 255) sufficient specificity to identify omega- Similarly, nearly all of the 17 studies included among its recommendations 3 fatty acids as causal in reducing CHD, on subjects in at-risk subpopulations. for further research the identification of but are consistent with the hypothesis including all of the studies that the atherogenic and anti-atherogenic that they are. controlled for PUFA’s (Refs. 203, 209, subfractions that may be present in Van Houwelingen et al. (Ref.294) 247, and 258), found no effect of VLDL and HDL; the uncertainty about found that, while men from a high fish supplemental omega-3 fatty acids on the relevance of changes in the amounts consumption group had higher total cholesterol (except for a post hoc of subtractions of these two lipoproteins concentrations of plasma phospholipid analysis of a subgroup in one study; similarly applies to LDL. EPA and DHA than men from a low fish (Ref. 209)). One study in diabetics (Ref. In at risk populations, there have been consumption group, there was no 252) found an increase in serum some additional reports of increased significant difference in collagen- cholesterol, but the statistical LDL cholesterol after fish oil induced platelet aggregation, cutaneous significance of the result may have been supplementation (Refs. 170,191, and bleeding time, ATP-release in whole due in part to a change in the opposite 251), a concern raised in the proposal. blood, or platelet number between the direction in the control group. One However, most studies have found LDL two groups. This study suggests that the study among hyperlipidemics (Ref. 191) cholesterol not changed by fish oils outcome measures found commonly to found decreased cholesterol after (Refs. 174, 189, 205, 209, 219, 258, and be affected in clinical studies may not relatively high doses (4.6 to 6 g EPA 278). Moreover, each of the studies that be related to consumption of omega-3 plus DHA/day) but not after 3.6 g EPA used a polyunsaturated fat placebo fatty acids in the free-living population. plus DHA/day, and did not control for control group found no change in LDL c. Intervention studies PUFA effects of the supplements. The cholesterol (Refs. 203, 209, and 258). other study that reported decreased Therefore, FDA concludes that these There were no new prospective cholesterol after supplemental omega-3 most recently reviewed studies support intervention studies measuring fatty acids (Ref. 268), similarly, found the conclusion reached in the proposed occurrence of heart attacks or CHD the effect after a high level (6 percent of rule, that for the general population, mortality. calories, 16 to 21 g EPA plus DHA/day) there is no significant effect of omega- 2. Evidence relating omega-3 fatty acids and did not control for the 3 fatty acids on LDL cholesterol. The to intermediate or surrogate markers of polyunsaturated fat effects of the results of recent studies among at-risk CHD (Table 2) supplement. subjects, however, are not in complete a. Atherosclerosis These studies support the conclusion agreement with the conclusions in the reached in the proposed rule, that i. Blood lipids proposed rule, and suggest that omega- among normal, healthy subjects there is 3 fatty acids may not uniformly increase Through its own literature review, no significant effect of omega-3 fatty LDL cholesterol. Additional study is FDA has found another 34 studies not acids from fish or fish oils on total needed to determine the conditions reviewed in the proposal that report serum cholesterol. under which LDL cholesterol is data for serum cholesterol after FDA concluded in the proposed rule increased by omega-3 fatty adds. consumption of fish containing omega- that the best studies among normal Among more recent studies in normal 3 fatty acids or fish oil concentrated in subjects found no effect of fish oils on healthy subjects found in FDA’s omega-3 fatty acids. Among these, 25 LDL cholesterol. All of the additional updated literature review, about half found no change in blood cholesterol studies among normal healthy subjects have found no effect of fish oils or fish levels, 3 found an increase, and 6 found obtained in FDA’s updated literature on HDL cholesterol (Refs. 202, 206, 217, a decrease. search have reported no change in LDL 219 (after 1.25 and 2.5 g/day EPA plus Studies among normal healthy cholesterol (Refs. 220, 253, 253, and DHA), 226 (after 1 and 3 g/day EPA plus subjects generally reported no change in 277). DHA), 241, and 244), but about half total cholesterol (Refs. 168, 196, 202, One study (Ref. 224) reported that have found increased HDL (Refs. 210, 210, 217, 220, 235, 241, 253, 254, and purified EPA produced a significant 219 (after 3.75 and 5 g/day EPA plus 277), although none of these studies was decrease in a subtraction of large, light DHA), 220, 226 (after 6 g/day EPA plus controlled for nonspecific effects of the LDL cholesterol (LDL1), and a DHA), 235, 253 (compared to baseline, omega-3 fatty acids as polyunsaturated significant increase in small, dense LDL significant compared to olive oil fats. cholesterol (LDL2), but FDA calculates control), 278, 283, and 301), including One study among normal subjects no change for the sum of these two a metabolic ward study that very found that feeding a high fish diet did fractions of LDL cholesterol. Some carefully controlled for total fat and not change total cholesterol, unless clinical studies reviewed in the saturated fat intake (Ref. 253). combined with a low total fat and low proposed rule (Refs. 1, 43, 53, and 129) Weintraub et al. (Ref.298) found saturated fat diet (Ref. 168). Another described changes in the composition of decreased HDL after fish oil compared study (Ref. 301) reported decreased total LDL particle after consumption of fish to saturated fat diet. oil.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2701

FDA attempted to ascertain how those uniformly associated with increased under which fish oils reliably increase studies that reported an increase in HDL HDL. Comparable results were found (total) HDL cholesterol have not been cholesterol after increased intake of after inspection of data on phospholipid established, either in a specific omega-3 fatty acids differed from those DHA after supplementation, however, subpopulation, or in the general studies in which no effect was found, because not all studies reported population. however, there was no apparent phospholipid fatty acid values, no When fractions of HDL cholesterol difference between the studies that conclusion can be drawn about the have been reported, an increase has reported that omega-3 fatty acids relationship between phospholipid generally been found in the HDL2 reduced HDL cholesterol and those that DHA and IIDL concentration. Notably, fraction (Refs. 1, 9, 54, 148, 191, 202, reported no change. Most of the studies recent data suggest a direct correlation 203, 220, 235, 251, and 286), with a that found a change used supplements between plasma EPA and HDL, but an comparable decrease in the HDL3 containing substantial amounts (e.g., 30 inverse relationship between plasma fraction (Refs. 202, 235, 251, and 286). percent) of saturated fatty acids, raising DHA and HDL (Refs. 177 and 178), Interestingly, the two recent reports that the possibility that the saturated fatty underscoring the importance of failed to find increased HDL2 both used acids in the supplements were reporting these data in future studies. esterified omega-3 fatty acids rather responsible for the increase in HDL (Ref. Among subjects with risk factors for than the fish oil triglyceride (Refs. 191 17). However, some supplements had CHD fewer reports found increased HDL and 224), although others using ethyl low amounts of saturated fatty acids (Refs. 191 (for type IV on SuperEPA esters have found increased HDL2 (Refs. (Ref. 278) or saturated fat in the diet was only), 195, 203, 209 (for type lib), and 9 and 286). specifically controlled (Ref. 235), and in 219) than found no change (Refs. 170, These studies suggest that fish oils one study the control diet was reported 174, 189, 191 (for type lIb and type IV produce a shift within the HDL fractions to have significantly more saturated fat on MaxEPA), 209 (type IV), 224, 258, toward a lipid-rich, and away from a than the fish diet (Ref. 301), so the 263, 277, and 299). protein-rich lipoprotein, as well as saturated fat intake during omega-3 fatty Few studies have controlled for within the LDL fractions. This shift may acid supplementation cannot be the effects of PUFA’s by giving a PUFA occur whether or not there is any factor responsible for increased HDL.. supplement. Two papers found no change in total HDL cholesterol. FDA The amounts of omega-3 fatty acids change in HDL in normal subjects fed noted (56 FR 60663 at 60669) that some used in those studies that reported fish oil as Promega (Ref. 73) or MaxEPA studies among normal subjects found increased HDL tended to be high (e.g., (Ref. 166) compared to wheat-germ oil increases in the HDL2 fraction of HDL more than 5 g EPA plus DHA/day), but or safflower oil (Refs. 73 and 166, cholesterol, and that these reports were some studies that found a change used respectively). Cobiac et al. (Ref. 20) the most promising changes in blood lower amounts (Ref. 278) and some reported increased HDL for mildly lipids. New studies published after the studies that used high amounts found hypertensive subjects fed salmon and period covered in FDA’s review of the no change (e.g., Ref. 241 (used 6.7 g EPA sardines in sild oil compared to those literature, however, found that both plus DHA/day) and 253 (used 8 g/day)). given a safflower-olive oil mix, but in HDL2, and HDL3 were correlated with Some studies in which fish was fed, comparable subjects, Meland et al. (Ref. reduced risk of MI (Refs. 185a and rather than fish oil, found an effect 247) found no change in HDL 287a), and the NHLBI consensus (Refs. 235 and 301), but others did not cholesterol after MaxEPA fish oil conference (Ref. 255) concluded that, (Refs. 206 and 244). There was no compared to when the subjects were “The current studies of HDL2 and HDL3 systematic difference in sample sizes of given a corn-olive oil mix. Very recent levels have not shown consistent the studies that found an effect and results, also for a mildly hypertensive associations with CHD.” Therefore, data those that did not; seven of the negative population, found increased HDL after on changes in HDL subtractions after studies reviewed in the proposed rule or either, ethyl esters of EPA and DHA, or increased consumption of omega-3 fatty in the present document had 30 or more after corn oil (Ref. 177). Thus, for acids do not provide a sufficient basis subjects, compared to only one of the normal and hypertensive subjects, the for a health claim, because there is not positive studies. Small studies (n = 10 change in HDL appears to not be a significant scientific agreement that the or fewer) may not have observed a specific effect of omega-3 fatty acids, but endpoints are directly related to risk of significant difference because of small may be related nonspecifically to CHD. If the risk of CHD becomes linked increased PUFA’s, either omega-3 fatty sample size, but larger studies did not with particular subtractions of these acids or omega-6 fatty acids. find a significant difference, even lipoproteins, these findings in normal In contrast, there are two reports of though some found a trend toward subjects may be of great importance. increased HDL cholesterol in subjects increased HDL after fish oil However, FDA also notes that recently with type lib hyperlipidemia fed fish oil supplementation (Ref. 217). published data from a prospective study compared to safflower oil (Ref. 166) or Finally, the enrichment of plasma demonstrate an effect of aspirin a corn-olive oil mix (Ref. 209), and one phospholipids with EPA and DHA consumption in reducing the incidence tended to be higher for subjects in report of increased HDL in type IIa of first heart attacks among women (Ret studies where increased HDL was found hyperlipidemics after fish oil or olive oil 243). Another study shows a compared to corn oil (Ref. 286). Others than that for subjects in the studies relationship between spontaneous found fish oil did not change HDL in where no change in HDL was found, platelet aggregation in vitro and type IV hyperlipidemics (Refs. 166 and reflecting the tendency of higher doses incidence of CHD (Ref. 288). Both 209) or patients with CHD (Ref. 258) to produce increased HDL. In particular, studies were conducted in the general compared to PUFA controls. all studies in which the plasma population and their results support the Therefore, at this time, FDA phospholipid EPA value was 3.9 hypothesis that platelet aggregation is a percent or more found increased HDL. concludes that there is some evidence useful marker for CHD risk in the However, the studies that fed the that omega-3 fatty acids, in some form general population. Additionally, highest amounts of EPA but that did not and amount and in some selected preliminary data from the Caerphilly find an effect on HDL did not report populations, may increase HDL Collaborative Heart Disease Study (Ref. data for phospholipid EPA, so it is not cholesterol, but that current data are 302) supports a relationship between clear whether high phospholipid EPA is ambiguous because the conditions platelet aggregation and the incidence of

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2702 ischemic heart disease; final data from The only new study among healthy substances released from platelets. this study will be available in the near subjects that reported no difference in Neither of these studies used a future. These recently published and responsiveness to ADP used EPA ethyl nonomega-3 PUFA control. forthcoming studies may provide the esters as the source of omega-3 fatty iv. Regulators of bleeding basis for significant scientific agreement acids (Ref. 179). Furthermore, the data Two recent studies in normal subjects regarding the use of platelet function as were not shown in this brief report, so have reported that omega-3 fatty acids a surrogate marker for CHD risk among it is not clear if there was a trend toward have no effect on the clotting protein the general population. an effect that might not have been fibrinogen (Refs. 183 and 210), although ii. Vessel wall effects. statistically significant due to small in one of these studies a large New human studies on the effects of number of subjects (eight per group). supplement of vitamin E was associated omega-3 fatty acids on vessel wall Those studies in healthy subjects with a decrease (Ref. 210). An effects were discussed in response to reviewed in the proposed rule that did uncontrolled study in normal subjects comments 35 through 37 of this not find statistically significant found a decrease in fibrinogen after fish document. A recent meta-analysis of differences in platelet responsiveness to oil supplementation (Ref. 278). studies on use of fish oils in the ADP did have trends in the direction of Studies on subjects at risk for CHD prevention of restenosis concluded that reduced responsiveness (Refs. 24 and have reported no change (Ref. 203), a the most plausible interpretation of the 54). decrease (Refs. 276 and 277), and an results was that there was a small to Other studies found no effect of fish increase in fibrinogen (Ref. 287). In moderate beneficial effect of fish oils, oils on platelet aggregation in response agreement with its tentative conclusion- to collagen (Refs. 179, 256 and 277). but that chance could not be ruled out in the proposed rule, PDA finds that the Each of these studies had a relatively as a cause of the results (Ref. 260). The data on the effects of omega-3 fatty acids small number of subjects, and there was authors noted a significant on fibrinogen level-are ambiguous, a trend toward decreased sensitivity because they do not distinguish effects heterogeneity in the findings and toward collagen at a high dose of omega- concluded that data from a large clinical due to PUFA’s from effects specific to 3 fatty acids in one study (Ref. 277). omega-3 fatty acids. study are necessary to confirm their However, in the recent metabolic ward interpretation. No study of restenosis to Plasminogen is an enzyme that study (Ref. 256) there was no trend date has compared fish oil to an dissolves clots. Plasminogen activator is toward decreased sensitivity toward alternate polyunsaturated oil to control a substance that increases clot collagen or thrombin. These findings for nonspecific effects of PUFA’s. dissolving; plasminogen activator is contrast with the results described specifically inhibited by another b. Thrombosis and hemostasis above (Refs. 218, 241, 251, and 297) and substance, the PAI-1. Thus, a high level i. Bleeding times with studies in healthy subjects of PAI-1 decreases the capability to A number of studies have reported described in the proposed rule (Refs. 2, dissolve-clots. data that show no significant effect of 24, 54, 86,143, and 166). Three recent studies reported fish oils on standardized bleeding time Studies reporting no effect of fish oils increased concentrations of PAI-1 after tests (Refs. 179, 218, 253, 268, and 277). on PAF or AA-indueed platelet fish oil supplementation (Refs. 254, 278, However, others have found a aggregation (Refs. 179 and 218) may not and 287), which would appear significant increase in bleeding time due have had sufficient power to find a inconsistent: with a clot-dissolving effect to fish, oil (Refs. 195, 219, 220, and 278) statistically significant difference; where of fish oil. Two of those investigators or salmon (Ref. 297) or have reported the data were reported there was a trend also found no change in the amount of increased bleeding as a side effect of toward decreased sensitivity for both plaminogen activator (t-PA) after treatment (Refs. 189 and 295). agents (Ref. 218). supplemental fish oil (Refs. 254 and ii. Platelet aggregation. iii. Platelet adhesion 287) including one who used a very specific immunologic assay (Ref. 254), Consistent with the literature A provocative study by Li and Steiner (Ref. 234) showed a 60-percent decrease suggesting that fish oils do not increase previously reviewed, recent, studies in the extent to which platelets prepared clot dissolution by increasing the show that fish oil tends to decrease from subjects fed fish oils adhered to amount of this protein. The third group platelet aggregations to numerous stimuli substrates in a laminar flow chamber. however, found an increase in the including AA (Refs. 179 and 256), activity of tissue plasminogen activator adenosine diphosphate (ADP) (Refs. The high flow rates used in this experiment showed that the change in (Ref. 277), which suggests that fish oils 204, 256, and 297), collagen (Refs. 218, adhesiveness of the platelets was due to might increase clot dissolution by a 241, 251, and 297), thrombin (Ref. 241), changes on the platelet surface, and not different mechanism than affecting the and PAF(Ref. 251). Only one of these due to a difference in the amount of amount of activator. Another group studies controlled for effects due to material released from platelets that found no. effect of cod liver oil on t-PA PUFA’s (Ref, 204). The importance of subsequently caused adhesion (i.e., AA). activity or fibrinolysis measured the polyunsaturated fat control is less Also, a dose-response relationship was directly (Ref, 216). These reports are in critical for studies on platelet function observed, and the time to return to pre- contradiction to a report of increased than for studies on blood lipids, because fish oil adhesion values was related to fibrinolytic activity after a fish or fish nonomega-3 PUFA’s (i.e., omega-6 fatty the amount consumed. plus fish oil diet (Ref. 183). FDA has not acids derived from plant oils) produce been able to find a reason for this rather effects in the opposite direction in However, another study found no marked contradiction. Therefore, in platelets as omega-3 fatty acids (whereas effect on fish oils on in vitro platelet agreement with the conclusion in its many of the blood lipid effects of these adhesion to everted rabbit aorta, proposed rule, FDA finds there is no two classes of fatty acids are in the same although there was a trend toward clear relationship between omega-3 fatty direction). Thus, the effects of omega-3 increased adhesion after 2 and 4 weeks acids and factors involved in dissolving fatty acids on platelet responsiveness of supplementation (Ref. 264). The blood clots, or clot dissolution activity. are not likely to be produced by PUFA’s reperfusion assay used in this study Numerous investigators (Refs. 174, in general. does not distinguish platelet membrane effects from effects mediated by 191, 210, 220, 235, 241, and 279) have

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2703 recently reported that fish oils do not oil both reduced diastolic blood vein into an artery, a procedure affect the concentration of Lp (a), a pressure comparably, but systolic blood associated with an accelerated lipoprotein correlated with the risk of pressure was only reduced by the corn development of atherosclerosis. Two CHD. One investigator reported that oil treatment papers (Refs. 275 and 303) each used a very high levels of fish oils (9 g EPA The results of these studies support polyunsaturated fat control and studied plus DHA/day) gave a trend toward the tentative conclusions reached in the fish oil effects after vein allografts in lower values, but the response may have proposed rule, that omega-3 fatty acids animals treated with the been due to the PUFA’s (Ref. 279). One reduce blood pressure to a small degree immunosuppressant cyclosporin. In one study reported no effect overall of fish in hypertensive people, but that it is not study (Ref. 303), six groups of rabbits oils on Lp(a) among clear if there is any specific effect received one of three amounts of fish oil hypertriglyceridemics, but Lp(a) was among normal subjects. (giving 29. 87 and 174 mg ERA plus reduced in those whose initial values 3. Other relevant information DHA/kg. respectively, similar amounts were high (Ref. 174). On the basis of to those used in most human studies) or these reports and those reviewed in the a. Animal studies comparable amounts of safHower oil. In proposed rule, FDA concludes that Animal studies are especially this study, safflower oil was more omega-3 fatty acids do not affect the risk important for studying effects of long- effective at reducing cholesterol than of CHD by lowering Lp(a). term consumption of omega-3 fatty fish oil. and there was a trend toward v. Blood pressure acids, where there are few data from more protection from atherosclerosis in Most of the studies not reviewed in human intervention studies. The animal the safflower oil-fed group. In the other the proposed rule that report data on studies cited in the proposed rule study (Ref. 275), rats received, in blood pressure after consumption of fish related to the ability of omega-3 fatty addition to cyclosporin, either fish oil oils have not found a significant change. acids to inhibit the development of (containing 210 mg EPA plus DHA/kg). One study of 50 elderly, healthy atherosclerosis, an area not readily or safflower oil with aspirin, or subjects reported that fish oils in available for study in humans. A more safflower oil only. The fish oil group combination with a salt-restricted diet complete discussion of the previously had remarkably less atherosclerosis than decreased both systolic and diastolic cited studies, with emphasis on those the other two groups. The contradictory blood pressure, but that fish oil alone studies in nonhuman primates, is given results in these two studies, both of had no effect (Ref. 190). There was a in response to comment 47 of this which used the same model of vein reduction in blood pressure during the document Other recent animal studies allografts with cyclosporin run-in period, when the cited in the comments or found during immunosuppression and the same polyunsaturated fat placebo, sunflower FDA’s updated literature search that polyunsaturated fat control, may be oil, was fed. provide data on the development of related to dose and species differences. Most studies on subjects with mild atherosclerosis (where atherosclerosis is A third study of vein allografts in hypertension also have reported no measured directly) are reviewed here. dogs (Ref. 274) found significantly less change (Refs. 247, 277, and 289), Also reviewed are studies on effects of atherosclerosis in fish oil-fed animals including one large, randomized, omega-3 fatty acids during experimental either fed the fish oil alone or in placebo-controlled, multicenter trial of ischemia, obviously not available for - combination with aspirin or a various behavioral changes and dietary human study. thromboxane synthetase inhibitor. Other supplements (Ref. 289). One study in i. Atherosclerosis animals were treated with aspirin only or a thromboxane synthetase inhibitor hypertensives found reduced systolic One recent study in rabbits found less only. There was no difference among and diastolic blood pressure comparable atherosclerosis in fish oil-supplemented groups for blood lipids, platelet function to reductions after the hypertension animals, but there was no control for medication propranolol (Ref. 285), and or eicosanoid metabolism. This study PUFA’s, and the fish oil-treated animals in some cases the combined treatment of suggests that mechanisms of also had reduced serum cholesterol (Ref. fish oil plus propranolol gave a greater atherosclerosis other than those 192). Because humans do not have decrease than either treatment alone. involving blood lipids and platelet reduced serum cholesterol after fish oil This study was controlled by olive oil function may be affected by omega-3 consumption, these results are of (which is predominantly fatty acids. questionable relevance to humans. monounsaturated fatty acids), and These animal models are most Furthermore, the effect cannot be therefore does not distinguish effects of relevant to comparable surgical omega-3 fatty acids from other PUFA’s. attributed specifically to omega-3 fatty procedures or other invasive procedures Another double-blind randomized, acids rather than to polyunsaturated fats (e.g., angioplasty) that would be in general. placebo-controlled study in expected to activate platelets in Fish oil feeding has also been hypertensives whose blood-pressures humans. Use of omega-3 fatty acids in associated with reduced binding of LDL were maintained by medications found these settings is a drug usage, but to the blood vessel endothelium in comparable blood pressure lowering provides information on the extent to compared to pretreatment values by fish monkeys (Ref. 193), and purified EPA which omega-3 fatty acids may modify oil or olive oil placebo (Ref. 299): ethyl ester was reported to reduce platelet response in vivo. The very One uncontrolled study among susceptibility of LDL to oxidation (Ref. different results of omega-3 fatty acids hyperlipidemics also found reduced 273), but these studies did not control in modifying the response to vein systolic and diastolic blood pressure for PUFA’s. The antioxidant levels in allografts in immune-suppressed (Ref. 263), but no effect was found in the diets with respect to the amount of animals indicates that the actions of uncontrolled trials in subjects with end- omega-3 fatty acids may be as important omega-3 fatty acids in these settings are stage renal disease (Ref. 207) or in determining whether or not there is not yet well established. diabetics (Ref. 215). In a any effect of omega-3 fatty acids on the ii. Response to ischemia oxidation of LDL. polyunsaturated fat (corn oil) controlled One major line of research on omega- study on subjects with stable Three recent papers describe effects of 3 fatty acids in animals is experimental claudication (Ref. 203) fish oil and corn fish oils fed before surgical grafting of a ischemia (deficiency of blood flow to

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2704 the heart). Force et al (Ref. 199) found proposed to be due to production of an feeding hamsters fish oil at 5 percent of that rats fed a diet containing a high active metabolite, rather than due to the diet for 4 weeks greatly reduced amount of fish oil (20 percent of the direct effects of EPA on the cell (over 60 percent) the adhesion of diet.) for 6 to 12 weeks had much greater membranes. This study suggests a leukocytes to the vessel wall (Ref. 233). blood flow once the occlusion was specific effect of EPA in stabilizing the The reduced adhesion could be relevant removed than rats fed diets enriched in heart myocytes during stress. Prevention for both the conditions during which corn oil or lard. There were no of arrhythmia by stabilization of these atherosclerosis develops (indeed, the differences among diet groups in the heart cells has been proposed as a stimulus used to elicit leukocyte amount of tissue damaged by ischemia. mechanism by which omega-3 fatty adhesion was oxidized LDL, a candidate Increased blood flow after ischemia has acids may increase the chances of for promoting atherosclerosis in also been reported in a pig model (Ref. survival following a heart attack as humans), and the acute response to 221). This study did not control for reported in the Dart study (Ref. 16). coronary ischemia. polyunsaturated fat and used a lower These studies indicate that, in various These animal data suggest amount of omega-3 fatty acids, and the animal models, dietary fish oils promote mechanisms by which omega-3 fatty differences in blood flow were not as greater reestablishment of blood flow in acids could affect the development of pronounced as in the Force study. heart tissues following a transient blocks atherosclerosis or the response of heart Another study found evidence of less as occurs in an acute heart attack. tissue after a transient occlusion of its tissue damage during reperfusion when Importantly, the results are consistent blood flow. Both modes of action could rats had been fed a diet with 1.2 percent across many animal species, and in make important contributions to the risk fish oil compared to other rats fed the some cases have been shown to be of CHD and, therefore, merit additional same level of corn oil (Ref. 223). specific for omega-3 fatty acids rather study. The reperfusion studies and the Another study in yet a third animal than simply due to any PUFA. Finally, myocyte toxicity study have species (Ref. 230) Showed that the the experimental designs included demonstrated specificity of the effect as functional capillary density was coronary occlusions in otherwise to omega-3 fatty acids. However, the preserved during reperfusion in healthy animals who were not suffering increase in reperfusion volume is not hamsters fed 5 percent fish oil for 4 from heart disease, a model relevant for sufficient to ensure a reduced risk of weeks prior to experimental ischemia. use of omega-3 fatty acids in reducing CHD death. Omega-3 fatty acids may not A reperfusion study in dogs the risk of CHD rather than in therapy affect the extent of tissue damaged determined the effects of fish oil on the for persons with preexisting heart during an occlusion, or the tendency for duration of time needed for drug- disease. The studies remain limited in a second, spontaneous occlusion. induced reperfusion following an that ischemia was produced by an acute Additionally, omega-3 fatty acids may electrically induced blockage, and on blockage produced by mechanical or not affect tissue vulnerability during the occurrence of spontaneous electrical means rather than by chronic reperfusion. Those studies where CHD reocclusion in the reopened vessel (Ref. dietary means, and the response to these deaths or second occlusions have been 182). High amounts of fish oil (one-third different types of block may not be the recorded used large amounts of fish oils, of total calories) for 3 weeks before the same. and do not indicate whether amounts of surgery resulted in a shorter time Other studies have attempted to learn omega-3 fatty acids found in the diet needed for the drug-induced the mechanisms by which the platelet- would have the same effects. Thus, reperfusion, but did not affect the time vessel wall interactions are modified by there are many possible avenues necessary for the electrically mediated omega-3 fatty acids. One study (Ref. suggested by these animal studies for occlusion to occur, the occurrence of 240) found that aortas from rats fed fish beneficial effects of omega-3 fatty adds second occlusion, or the time it took for oil or corn oil did not contract as much on the development of CHD, but there the second occlusions to appear. This in response to agents that cause are also important limitations in the study did not have a polyunsaturated fat contraction as aortas from rats fed beef study designs and models used that control, and even at the high intake only tallow (saturated fat). This was true both prevent drawing conclusions from these a modest effect of oroega-3 fatty acids on before and after oxygen deprivation. The data about the importance of omega-3 platelet function was seen, that being aortas from fish oil-fed rats were more fatty acids in reducing the risk of human primarily an enhancement of the effects responsive to one of three tested CHD. of the fibrinolytic drug. chemical relaxers than aortas from corn Another possible consequence of oil-fed or beef tallow-fed rats. Another b. Safety concerns Ischemia is arrhythmia, when the heart study found, that EPA potentiated the i. Diabetes fails to maintain its normal rhythmic release of an EDRF (Ref. 181), but the beating. The effects of fish oils on Three additional papers (Refs. 170, effect was thought to be related to the 222, and 304) and one major review arrhythmia in monkeys are discussed in unsaturation of the EPA, because the response to comment 39 of this (Ref. 238) on the effects of fish oils in experiments were carried out in the document. Similarly, data have been diabetics were published after the time presence of inhibitors of EPA reported for experimental ischemia in period reviewed in the proposal. All metabolism. rats that show that, both fish oil and three new studies found increased LDL One research group has recently cholesterol-after fish oil consumption in sunflower oil reduced the occurrence of shown that leukotrienes, chemicals type II diabetics. However, effects oil arrhythmia during occlusion and produced from AA, are important in the reperfusion compared to a saturated fat fasting glucose varied, with no change tissue injury that accompanies (Ref. 170), a transient increase (Ref. 222) diet (Ref. 246). Another study, done on reperfusion (Refs. 230 and 232). Since or an increase (Ref. 304) reported. isolated, cultured rat heart cells EPA competes with AA for the enzyme Although fasting insulin concentration (myocytes) showed that EPA, but not that makes leukotrienes from AA, EPA AA, prevented a known toxin (ouabain) was unchanged after fish oil (Refs. 170- could potentially reduce the amount of from disturbing the rhythmic and 304), postprandial, insulin response leukotriene formed from. AA. This same contractions and killing the cells (Ref. usually, but not always (Ref. 170), has group has shown that leukotrienes 212). The effective amount of EPA was been reported as reduced (Refs. 238 and promote the adhesion of leukocytes to so low that the mode of action was 304). the vessel wall (Ref. 231), and that

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2705

These effects of fish oils on blood III. Overall Summary and Conclusions might well have a protective effect glucose appear to depend on the amount FDA concludes that there is some against CHD in such a subpopulation. of fish oils fed. One study found no evidence that supports a relationship There are other areas where change in fasting blood glucose levels between omega-3 fatty acids and CHD, additional research is needed to show, among type n diabetics treated with 3.0 but that the totality of scientific for agreed endpoints, that the effects are g/day EPA plus DHA for 2 weeks (Ref. evidence available at this time does not consistently produced, or are 170). Two other studies that used 3 or provide an adequate basis for a health specifically due to omega-3 fatty acids. 2.7 g/day EPA plus DHA for 6 and 8 claim. In some cases, there is not These areas require additional data to weeks (Refs. 222 and 79) only found significant scientific agreement that the establish that the effect of omega-3 fatty transient increases in blood glucose changes that are specific to omega-3 acids is specific, or to further define the halfway through their respective fatty acids will reduce the risk of CHD. conditions under which omega-3 fatty supplementation periods. A fifth study In other cases, the data do not acids have their effects. For example, (Ref. 12) that used 3.0 g/day EPA plus demonstrate that the effect is data are needed to show a reduction in DHA for 3 weeks found comparable specifically due to the omega-3 fatty MI or CHD mortality among individuals increases in fasting blood glucose when acids and not due to other dietary fed supplemental omega-3 fatty acids either fish oil or safflower oil was fed, variables. For yet other cases, the data (specifically) compared to a group fed so the increase cannot be attributed are ambiguous because effects of omega- omega-6 PUFA’S. The critical failing of specifically to the omega-3 fatty acids. 3 fatty acids are not consistently some recent studies associating omega- Similarly. Vessby and Boberg (Ref. 157) observed, which suggests that other 3 fatty acids and CHD is that specificity fed 3 g/day EPA plus DHA and did not variables are important in determining was not obtained. Future studies should find a difference in fasting glucose or whether or not an effect is seen. carefully control for known glycosylated hemoglobin after fish oil Therefore, the agency does not consider confounders, particularly dietary supplementation compared to baseline; the evidence sufficiently strong to draw variables. they did find a significant difference a firm conclusion about the relationship Finally, the available data suggest that compared to the olive oil treatment that between omega-3 fatty acids and risk of some set of conditions or population produced changes in the opposite CHD, and therefore is not authorizing may exist for which omega-3 fatty acids direction from fish oil. Studies on type the claim at this time. will increase HDL. Additional research II diabetics that reported increased In the course of developing this should be able to define the conditions glucose used higher amounts (4.5 to 8 g/ regulation. FDA has identified some under which omega-3 fatty acids have day) of omega-3 fatty acids (Refs. 52, 55, areas where greater agreement is needed this effect. 128, and 304). Thus, FDA concludes that the effects produced by omega-3 Interested parties may choose to that glycemic control among diabetics fatty acids are directly related to the risk petition the agency for approval of other remains a valid safety concern, but of CHD. Many surrogate markers have health claims about omega-3 fatty acids. notes that restriction of the amount of been hypothesized, on the basis of For example, additional data may be supplemental omega-3 fatty acids may limited evidence, to be related to developed to support an omega-3 fatty suitably address this concern. specific diseases, including CHD, but acids/hypertension health claim ii. Increase in LDL cholesterol few have withstood the continued petition. Because the blood pressure- scrutiny of scientific investigation. Also, lowering effect of omega-3 fatty acids Many studies published after 1987 appears most marked against a with data for LDL or apoB report some markers may have scientific validity, but may not be applicable for background of very low dietary intakes increased LDL cholesterol or apoB after of omega-3 fatty acids, the role of fish oils, in hypercholesterolemic or use in the general population, because of technical limitations. Thus, FDA omega-3 fatty acids in the total diet hypertriglyceridemic subjects (Refs. 1, would need clarification before such a 26, 32, 55, 60, 61, 63, 73, 75, 94, 114, asserts that only when a surrogate marker for a disease has been accepted petition could be adequately supported. 119, 120, 129, 146, and 166). Virtually Similarly, limitations of the effects of all the studies with 10 or more subjects as a risk factor for the general population, as indicated by a statement omega-3 fatty acids on the magnitude supplemented with 5 g/day EPA plus and duration of change in blood DHA or more found increased LDL. by an unbiased, nationally representative authoritative scientific or pressure, the quantitative amounts of Some studies on normal subjects (all of medical body, can the agency authorize omega-3 fatty acids required for the which were reviewed in the proposed a health claim based on the relationship effects, and characterization of the rule) also report increased LDL or apoB of a nutrient to the surrogate marker of sensitive subpopulation would require after fish oil consumption (Refs. 54,127, the disease. Examples of potential discussion in a petition. A petition and 156). Many studies that found no surrogate measures for which validation should also address apparent conflicting effect may not have had sufficient is needed are in vitro platelet pieces of information, e.g., high blood sample size to detect a difference due to aggregation, in vitro platelet adhesion. pressure among populations that have treatment. FDA concludes that elevated fasting triglycerides high intakes of omega-3 fatty acids. increased LDL cholesterol among some postprandial triglycerides (recently Safety concerns raised in this final rule populations already at increased risk of considered at the NHLBI consensus will, of course, require resolution prior CHD remains a valid safety concern, but development conference. Ref. 255), and to the authorization of any petitioned because most reports of increased LDL subtractions of LDL and/or HDL. claim. are in studies where large amounts of IV. Decision Not to Authorize a Health fish oils are given, it is possible that In some cases additional research is needed to determine whether Claim Relating Ingestion of Omega-3 restriction of the amount of Fatty Acids to Reduced Risk of CHD hypothesized subpopulations, e.g., those supplemental omega-3 fatty acids and/ In evaluating the scientific evidence. or changes in the fatty acid composition with high LDL:HDL ratio and high triglycerides, are at increased risk of FDA considered: (1) The strength of the of omega-3 fatty acid supplements may association of omega-3 fatty acids with suitably address this concern. disease. The pronounced triglyceride lowering effects of omega-3 fatty acids CHD or surrogate markers for CHD, (2) the consistency of findings among the

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2706 many studies, (3) the specificity of the generally agreed to be risk factors for cholesterol in normal, healthy persons, outcome to omega-3 fatty acids, (4) the CHD. Therefore, FDA finds numerous or among persons at risk for CHD. presence or absence of a dose-response reasons for not accepting all of the Numerous studies, including some large relationship, and (5) the biologic findings of the LSRO report. FDA’s or multicenter studies, have reported plausibility of an association. FDA has conclusions regarding the relationship these results, demonstrating consistency determined that there is inadequate between omega-3 fatty acids and CHD in the findings and providing the agency evidence to show that increased rely instead on FDA’s independent confidence that they were not spurious. consumption of omega-3 fatty acids will review of the publicly available The data on HDL cholesterol are reduce the risk of CHD. scientific information, and are ambiguous. There appear to be other FDA sought to determine whether consistent with the Federal Government factors in the dietary interventions there was significant scientific and other comprehensive and besides the omega-3 fatty acids that agreement among qualified experts that authoritative reports except for the determine whether or not the totality of publicly available LSRO report. supplementation with fish or fish oils scientific evidence supported the claim The surveys, cross-sectional studies, raises HDL. that omega-3 fatty acids reduce the risk and nonintervention prospective studies An increase in bleeding times and a of heart disease. FDA reviewed the do not provide adequate support for a decrease in platelet aggregation have position taken in numerous Federal relationship between consumption of been observed frequently, but not Government reports and other omega-3 fatty acids and CHD. Only a always, after supplemental omega-3 authoritative scientific reports and few studies found an association fatty acids in normal healthy evaluated the publicly available between fish intake and CHD, while individuals as well as in diseased scientific evidence that has become others have found no association; thus, persons. Additionally, there is evidence available since those reports were there was no consistency of findings. that platelet adhesion is reduced by written. The decision to deny a health None of the studies that reported a omega-3 fatty acids. The effects of claim is based on the conclusions relationship distinguished fish decreased platelet aggregation and reached following review of the consumption from other factors platelet adhesion appear to be related to following sources of information: (1) associated with fish consumption, and the intake of omega-3 fatty acids in a “The Surgeon General’s Report on therefore none demonstrates specificity. dose-response relationship. What has Nutrition and Health;” (2) the National Even in those studies that reported a not been established, however, is that in Research Council’s “Diet and Health: relationship between fish consumption vitro platelet aggregation or platelet Implications for Reducing Chronic and CHD, it is not clear that the adhesion are bona fide surrogate risk Disease Risk,” and (3) the National observed effects were due to the omega- factors for CHD in the general Cholesterol Education Program’s Report 3 fatty acids in fish. Also, the omega-3 population. of the Expert Panel on Detection. fatty acid content of the fish diet Omega-3 fatty acids have been shown Evaluation and Treatment of High Blood associated with reduced CHD in these to reduce blood pressure in Cholesterol in Adults. Each of these studies was so low that the importance hypertensive people to a small degree, reports concluded that there was of omega-3 fatty acids is questionable, which may bear on a relationship inadequate evidence of a relationship thus weakening the biologic plausibility between omega-3 fatty acids and CHD. between consumption of omega-3 fatty of the relationship. This effect was not of large magnitude, acids and CHD. FDA has reviewed again The data from intervention studies but it is specific to omega-3 fatty acids, all of the relevant cross-sectional data also do not establish a relationship it has been reported by a number of from which the hypothesized between omega-3 fatty acids and risk of investigators, a dose response was relationship between omega-3 fatty CHD. The most compelling type of found, and the effect is biologically acids and CHD originated, and all evidence to support a diet-disease plausible through at least two clinical intervention data published relationship is a prospective, double- mechanisms. However, it has not been since these Federal Government and blind. placebo-controlled intervention established that omega-3 fatty acids other authoritative reports were issued study, with CHD morbidity and reduce blood pressure in normal to determine whether the additional mortality as endpoints. To date, there is subjects (lack of consistency, weak evidence is adequate to support a health only one such trial (Ref. 16). The results effect, absence of dose-response claim for omega-3 fatty acids. of that study showed that increased relationship). Additionally, it has not The LSRO report reached a different consumption of fish does not reduce the been demonstrated that the magnitude conclusion from the other authoritative risk of a second heart attack but may and duration of changes in blood pressure observed in short-term studies reports by finding a relationship reduce the risk that the attack will be will persist during long-term between omega-3 fatty acids and CHD. fatal. This study provides evidence for consumption of omega-3 fatty acids, or The report used only selected evidence, a protective effect of fish consumption that these changes result in a reduced and often did not distinguish effects against second heart attacks. However, specifically due to omega-3 fatty acids as with the nonintervention study data, risk of CHD. from effects due to PUFA’s in general. this study did not provide evidence to Finally, the potential that omega-3 The description of international attribute the benefit to omega-3 fatty fatty acids may increase LDL cholesterol epidemiologic findings of a relationship acid intake rather than some other factor and/or apoB among diabetics and between fish consumption and CHD, associated with fish consumption hyperlipidemics, and the potential that omega-3 fatty acids may worsen control similarly, was not shown to be specific (specificity). of blood glucose in diabetics are to omega-3 fatty acids. In some Less persuasive than prospective significant safety concerns that must be instances, FDA disagreed with the studies in which CHD per se is addressed before a claim may be made interpretation of the studies reviewed by measured, but still very useful, are that consumption of omega-3 fatty acids LSRO, or with LSRO’s conclusions. prospective clinical trials in which by the general population will reduce Finally, the LSRO report also based its surrogate markers for CHD are the risk of CHD. conclusions about the usefulness of measured. Recent studies have not In conclusion, there are numerous omega-3 fatty acids, in part, on changes found beneficial effects from omega-3 effects of omega-3 fatty acids that may in blood lipid parameters that are not fatty acids on total cholesterol or LDL

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2707

be related to the risk of CHD, e.g., comments, that the overall food labeling Eicosapentaenoic and Docosahexaenoic reduction in fasting and postprandial reform initiative constitutes a major rule Acids on Blood Pressure in Hypertension,” as defined by Executive Order 12291. New England Journal of Medicine, 322:795- triglicerides, reductions in platelet 801, 1990. aggregation and adhesion, changes in Further, the agency has concluded that although the costs of complying with 12. Borkman, M., et al., “Effects of Fish Oil the compostion of lipoproteins. Supplementation on Glucose and Lipid the new food labeling requirements are Metabolism in NIDDM,” Diabetes, 38:1314- However, at this time these endpoints substantial, such costs are outweighed are not generally agreed to be closely 1319, 1989. by the public health benefits that will be 13. Bowles, M. H., et al., “ EPA in the related to the risk of CHD. In other realized through the use of improved Prevention of Restenosis Post PTCA,” areas, additional data are needed to show that effects nutrition related information to fish provided by food Angiology—The Journal of Vascular consumption are specifically due to the labeling. Diseases. omega-3 fatty acids in the fish, and to VII. References 14. Brown, A. J., D. C. Roberts, J. E. Pritchard, A. S. Truswell, “A Mixed define the conditions under which The following references have been Australian Fish Diet and Fish-oil omega-3 fatty acids consistently placed on display in the Dockets Supplementation: Impact on the Plasma increase HDL. These avenues may Management Branch (address above) Lipid Profile of Healthy Men,” American provide a reasonable basis for a future and may be seen by interested persons Journal of Clinical Nutrition, 52:825-833, petition for a health claim relating between 9 a.m. and 4 p.m., Monday 1990. through Friday. 15. Brown, A. J., D. C. K. Roberts, omega-3 fatty acids to the risk of CHD. 1. Abbey, M., P. Clifton, M. Kestin, B. “Moderate Fish Oil Intake Improves Lipemic V. Environmental Impact Belling., P. Nestel, “Effect of Fish Oil on Response to a Standard Fat Meal,” The agency has determined under 21 Lipoproteins, Lecithin: Cholesterol, Arteriosclerosis and Thrombosis, 11:457- CFR 25.24(a)(11) that this action is of a Acyltransferase, and Lipid Transfer Protein 466, 1991. type that does not individually or Activity in Humans,” Arteriosclerosis, 10:85— 16. Burr, M. L., J. F. Gilbert, R. M. Holliday, cumulatively have a significant effect on 94,1990. P. C. Elwood, “Effects of Changes in Fat, the human environment. Therefore, 2. Agren, J. J., 0. Hanninen, A. Hanninen, Fish, and Fibre Intakes on Death and Myocardial Reinfarction: Diet and neither an environmental assessment K. Seppanen, “Dose Responses In Platelet Fatty Acid Composition, Aggregation, and Reinfarction Trial (DART),” Lancet, ii:757- nor an environmental impact statement Prostanoid Metabolism During Moderate 761, 1989. is required. Freshwater Fish Diet,” Thrombosis Research, 17. Childs, M. T., L B. King, R. H. Knopp, VI. Economic Impact 57:231-28, 1990. “Divergent Lipoprotein Responses to Fish In its food labeling proposals of 3. Al-Mahtaseb, N., N. Hayat, M. Al- Oils with Various Ratios of Eicosapentaenoic Acid and Docosahexaenoic Acid,” American November 27, 1991 (56 FR 60366 et Khafaji, “Lipoproteins and Apolipoproteins in Young Male Survivors of Myocardial Journal of Clinical Nutrition, 52:632-9, 1990. seq.), FDA stated that the food labeling Infarction,” Atherosclerosis, 77:131-138, 18. Clark, W. F., A. Parbtani, M. W. Huff, reform initiative, taken as a whole, 1989. B. Reid, B. J. Holub, P. Faladeau, “Omega-3 would have associated costs in excess of 4. Austin, M. A., “Plasma Triglyceride and Fatty Acid Dietary Supplementation in the $100 million threshold that defines Coronary Heart Disease,” Arteriosclerosis Systemic Lupus Erythematosus,” Kidney a major rule. Thus, in accordance with and Thrombosis, 11:1-14, 1991. International, 36:653-660, 1989. Executive Order 12291 and the 5. Azar, R., F. Dequiedt, J. Awada, P. 19. Clubb, F. J., J. M. Schmitz, M. M. Deouiedt, A. Tacquet, “Effects of Fish Oil Butler, L. M. Buja, J. T. Willerson, W. B. Regulatory Flexibility Act (Pub. L. 96- Campbell, “Effect of Dietary Omega-3 Fatty 354), FDA developed one Rich in Polyunsaturated Fatty Acids on Hyperlipidemia of Hemodialysis Patients,” Acid on Serum Lipids, Platelet Function, and comprehensive regulatory impact Kidney International, 36:26-27, 1989. Atherosclerosis in Watanabe Heritable analysis (RIA) that presented the costs 6. Bach, R., U. Schmidt, P. Jung, H. Hyperlipidemic Rabbits,” Arteriosclerosis, and benefits of all of the food labeling Kiesewetter, “Effects of Fish Oil Capsules in 9:529-537, 1989. provisions taken together. That RIA was Two Dosages on Blood Pressure, 20. Cobiac, L., P. M. Clifton, M. Abbey, G. published in the Federal Register of Piatelet Functions, Haemorrheological, and B. Belling, P. J. Nestel, “Lipid, Lipoprotein, November 27, 1991 (56 FR 60856), and Clinical Chemistry Parameters in Apparently and Hemostatic Effects of Fish vs. Fish-oil N- along with the food labeling proposals, Healthy Subjects,” Annals Nutrition 3 Fatty Acids in Mildly Hyperlipidemic the agency requested comments on the Metabolism, 33:359-367, 1989. Males,” American Journal of Clinical 7. Bagdade, J. D., W. E. Buchanan, R. A. Nutrition, 53:1210-1216, 1991. RIA. Levy, P. A. Subbaiah, M. C. Ritter, “Effects 21. Connor, W. E., S. L. Connor, “Diet, FDA has evaluated more than 300 of W-3 Fish Oils on Plasma Lipids, Atherosclerosis, and Fish Oil,” Advances comments that it received in response to Lipoprotein Composition, and Postheparin Internal Medicine, 35:13 9-71, 1990. the November 1991 RIA. FDA’s Lipoprotein Lipase in Women with IDDM,” 22. Council for Responsible Nutrition, discussion of these comments is Diabetes, 39:426-431, 1990. “Proceedings and Conclusions of the Round- contained in the agency’s final RIA 8. Bang, H. O., J. Dyerberg, “Lipid table Discussion on Fish Oils and Bleeding Metabolism and Ischemic Heart Disease in published elsewhere in this issue of the Time,” Chester, England, Council for Greenland Eskimos,” Advances in Responsible Nutrition, Washington, DC, Federal Register. In addition, PDA will Nutritional Research, 3:1-21,1986. 1990. prepare a final regulatory flexibility 9. BIonk, M. C., H. J. G. Bilo, J. J. P. Nauta, 23. Crombie, I. K., P. McLoone, W. C. S. analysis (RFA) subsequent to the C. Popp-Snijders, C. Mulder, A. J. M. Donker, Smith. M. Thomson, H. T. Pedoe, publication of the food labeling final “Dose-response Effects of Fish-oil “International Differences in Coronary Heart rules. The final RFA will be placed on Supplementation in Healthy Volunteers,” Disease Mortality and Consumption of Fish file with the Dockets Management American Journal of Clinical Nutrition, and Other Foodstuffs,” European Heart Branch (HFA-305), Food and Drug 52:120. Journal, 8:560-563, 1987. Administration, rm. 1-23, 12420 10. Bonaa, K., “Epidemiological and 24. Croset, M., et al., “Functions and Parklawn Dr., Rockville, MD 20857, and Intervention Studies on the Effect of Marine Tocopherol Content of Blood Platelets from a notice will be published in the Polyunsaturated Fatty Acids on Blood Elderly People After Low Intake of Purified Pressure,” Journal of Internal Medicine Eicosapentaenoic Acid,” Thrombosis Federal Register announcing its Supplement, 225:105-10, 1989. Research, 57:1-12, 1990. availability. 11. Bonaa, K. H., K. S. Bjerve, B. Straume, 25. Curb. J. D., D. M. Reed, “Fish In the final RIA, FDA has concluded, I. T. Gram, D. Thelle, “Effect of Consumption and Mortality from Coronary cased on its review of available data and Heart Disease,” New England Journal of Medicine, 313:821, 1985.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2708

26. Dart, A. M., R. A. Riemersma, M. F. 42. Ernst, E., “Effects of N-3 Fatty Acids Journal of Internal Medicine, 227:347-53 Oliver, “Effects of Maxepa on Serum Lipids on Blood Rheology,” Journal of Internal 1990. in Hypercholesterolemic Subjects,” Medicine, 225:129-132, 1989. 58. Hardarson, T., et al., “Cod Liver Oil Atherosclerosis, 80:119-124, 1989. 43. Failor, R. A., M. T. Childs, E. L. Does Not Reduce Ventricular Extrasystoles 27. Davis, H. R., R. T. Bridenstine, D. Bierman, “The Effects of W3 and W6 Fatty after Myocardial Infarction,” Journal of Vesselinovitch, R. W. Wissler, “Fish Oil Acid-enriched Diets on Plasma Lipoproteins Internal Medicine, 226:33-37, 1989. Inhibits Development of Atherosclerosis in and Apoproteins in Familial Combined 59. Harris, W. S., W. E.Connor, N. Alam, Rhesus Monkeys,” Arteriosclerosis, 7:441- Hyperlipidemia,” Metabolism, 11:1021-1028, D. R. Iillingworth, “Reduction of Postprandial 449, 1987. 1988. Triglyceridemia in Humans by Dietary N-3 28. DeCaterina, R., et al., “Vascular 44. Fasching, P., K. Ratheiser, W. Fatty Acids,” Journal of Lipid Research, Prostacyclin is Increased in Patients Ingesting Waldhausl, M. Rohac, “Metabolic Effects of 29:1451-1460, 1988. Omega-3 Polyunsaturated Fatty Acids Before Fish-oil Supplementation in Patients with 60. Harris, W. S., C. A. Dujovne, M. Zucker, Coronary Artery Bypass Graft Surgery,” Impaired Glucose Tolerance,” Diabetes, B. Johnson, “Effects of Low Saturated Fat, Circulation, 82:428-438, 1990. 40:583-589, 1991. Low Cholesterol Fish Oil Supplement in 29. Deck, C., K. Radack, “Effects of Modest 45. FDA, Taylor J. M., letter to J. D. Hypertriglyceridemic Patients,” Annals of Doses of Omega-3 Fatty Acids on Lipids and Cordaro, June 1, 1988. Internal Medicine, 109:465-470, 1988. Lipoproteins in Hypertriglyceridemic 46. FDA, Chesemore R. G., letter to R. 61. Harris, W. S., M. L. Zucker, C. A. Subjects, a Randomized Controlled Trial,” William Seller, Ph.D., June 20, 1990. Dujovne, “W-3 Fatty Acids in Archives of Internal Medicine, 149:1857- 47. Fincham, J. E., et al., “Chronic Effects Hypertriglyceridemic Patients: Triglycerides 1862, 1989. of Fish Oil and a Therapeutic Diet in vs Methyl Esters,” American Journal of 30. Dehmer, G. J., J. J. Popma, E. K. Van Nonhuman Primates,” Arteriosclerosis and Clinical Nutrition, 48:992-997, 1988. Den Berg, E, J. Eichhorn, “Reduction in the Thrombosis, 11:719-732, 1991. 62. Hams, W. S., “Fish Oils and Plasma Rate of Early Restenosis After Coronary 48. Fisher, M., et al., “Dietary N-3 Fatty Lipid and Lipoprotein Metabolism in Angioplasty by a Diet Supplemented with N- Acid Supplementation Reduces Superoxide Humans: A Critical Review,” Journal of Lipid 3 Fatty Acids,” New England Journal of Production and Chemiluminescence in a Research, 30:785-807.1989. Medicine, 319:733-740, 1988. Monocyte-enriched Preparation of 63. Harris, W. S., et al., “Fish Oils in 31. DeLany, J. P., V. M. Vivian, J. T. Snook, Leukocytes,” American Journal of Clinical Hypertriglyceridemia: A Dose-response P. A. Anderson, “Effects of Fish Oil on Nutrition, 51:804-808, 1990. Study,” American Journal of Clinical Serum Lipids in Men During a Controlled 49. Flaten, H., et al., “Fish-oil Concentrate: Nutrition, 51:399-406, 1990. Feeding Trial,” American Journal of Clinical Effects on Variables Related to 64. Harris, W. S., S. Silveira, C. A. Dujovne. Nutrition, 52:477-485, 1990. Cardiovascular Disease,” American Journal “Combined Treatment with Fish Oils and 32. Demke, D. M., G. R. Peters, O. I. Linet, of Clinical Nutrition, 52:300-306, 1990. Aspirin Affects Bleeding Times and Platelet C. M. Metzler, K. A. Klott, “Effects of a Fish 50. Fox, P. L., P. E. Di Corleto, “Fish Oils Function,” World Reviews of Nutrition and Oil Concentrate in Patients with Inhibit Endothial Cell Production of Platelet- Dietetics, 66:530, 1991. Hypercholesterolemia,” Atherosclerosis, devised Growth Factor-like Protein,” 65. Hearn, J. A., D. S. Sgoutas, K. A. 70:73-80, 1988. Robinson, S. B. King, R. J. Siegel, G. S. Science, 241:453-456, 1988. 33. Department of Health and Human Koubin, “Marine Lipid Concentrate and 51. Foxall, T. L., G. T. Shwaery, “Effects of Services (DHHS), Public Health Service Atherosclerosis in the Rabbit Model,” Dietary Fish Oil and Butterfat on Serum (PHS), “The International Classification of Atherosclerosis, 75:39-47, 1989. Lipids and Monocyte and Platelet Diseases, 9th Revision,” DHHS, Washington, 66. Hennekens, C. H., J. E. Buring, P. Interactions with Aortic Endothelial Cells,” DC, 1980. Sandercock, R. Collins, R. Peto, “Aspirin and Atherosclerosis, 80:171-179, 1990. 34. DHHS, PHS, “The Surgeon General's Other Antiplatelet Agents in the Secondary 52. Friday, K. E., et al., “Elevated Plasma Report on Nutrition and Health,” DHHS, and Primary Prevention of Cardiovascular Glucose and Lowered Triglyceride Levels Washington, DC, 1988. Disease,” Circulation, 80:749-756, 1989. from Omega-3 Fatty Acid Supplementation 35. DHHS, National Cholesterol Education 67. Herold, P. M., J. E. Kinsella, “Fish Oil in Type II Diabetics,” Diabetes Care, 12:276- Program (NCEP): Report of the Expert Panel Consumption and Decreased Risk of 281, 1989. on Detection, Evaluation, and Treatment of Cardiovascular Disease: A Comparison of 53. Friday, K. E., R. A. Failor, M. T. Childs, High Blood Cholesterol in Adults, DHHS, Findings from Animal and Human E. L. Bierman, “Effects of N-3 and N-6 Fatty Washington, DC, 1989. Feedingtrials,” American Journal of Clinical Acid-enriched Diets on Plasma Lipoproteins 36. DHHS, National Cholesterol Education Nutrition, 43:566-598, 1986. and Apolipoproteins in Heterozygous Program (NCEP): Population Panel Report, 68. Hirai, A., et al., “Eicosapentaenoic Acid Familial Hypercholesterolemia,” DHHS, Washington, DC, 1990. and Platelet Function in Japanese,” Lancet, Arteriosclerosis and Thrombosis, 11:47-54, 37. DHHS, PHS, “Healthy People 2000 ii:1132, 1980. 1991. National Health Promotion and Disease 69. Hollander, et al., 1987 (to be added). Prevention Objectives,” DHHS, Washington, 54. Fumeron, F., et al., “N-3 70. Holub, B. J., “Dietary Fish Oils DC, 1990. Polyunsaturated Fatty Acids Raise Low- Containing Eicosapentaenoic Acid and the 38. Dolecek, T. A., G. Grandits, “Dietary Density Lipoproteins, High-Density Prevention of Atherosclerosis and Polyunsaturated Fatty Acids and Mortality in Lipoprotein 2, and Plasminogen-activator Thrombosis,” Canadian Medical Association the Multiple Risk Factor Intervention Trial,” Inhibitor in Health Young Men,” American Journal, 139:377-381, 1988. Journal of Clinical Nutrition, 54:118-122, World Review Nutrition Dietetics, 66:205- 71. Hostmark, A., T. Bjerkedal, P. Kierulf, 216, 1991. 1991. H. Fiaten, K. Ulshagen, “Fish Oil and Plasma 39. Dyerberg, J., H. O. Bang, “A Hypothesis 55. Glauber, H., P. Wallace, K. Griver, G. Fibrinogen,” British Medical Journal, on the Development of Acute Myocardial Brechtel, “Adverse Metabolic Effect of 297:180-181, 1988. Infarction in Greenlanders,” Scandinavian Omega-3 Fatty Acids in Non-insulin- 72. Hsia, M. T. S., R. D. Mavis, J. M. Journal of Clinical and Laboratory Dependent Diabetes Mellitus,” Annals of DeSesso, “Health Effects of Refined Investigation, 42:7, 1982. Internal Medicine, 108:663-668, 1988. Menhaden Oil Mitre Corporation, McLean 40. Emeis, J. J., A. C. Houwelingen, C. M. 56. Grigg, L. E., et al., “Determinants of VA, 1989. Hoogen, G. Hornstra, “A Moderate Fish Ecstenosis and Lack of Effect of Dietary 73. Hughes, G. S., et al., “Fish Oil Produces Intake Increases Plasminogen Activator Supplementation with Eicosapentaenoic and Atherogenic Lipid Profile in Inhibitor Type-1 in Human Volunteers,” Acid on the Incidence of Coronary Artery Hypertensive Men,” Atherosclerosis, 84:227- Restenosis After Angioplasty,” Journal of Blood, 74:233-237, 1989. 237, 1990. American College Cardiology, 13:665-667, 41. Endres, S., et al., “The Effect of Dietary 74. Hunter, D. J., I. Kazda, A. 1989. Supplementation With N-3 Polyunsaturated Chockalingam, J. G. Fodor, “Fish Fatty Acids on the Synthesis of Interleukin- 57. Haghmd, O., R. Wallin, R. Luostarinen, Consumption and Cardiovascular Mortality 1 and Tumor Necrosis Factor by T. Saldeen, “Effects of a New Fluid Fish Oil in Canada: An Inter-regional Comparison,” Mononuclear Cells,” New England Journal of Concentrate, ESKIMO-3, on Triglycerides, American Journal Preventive Medicine, 4:5- Medicine 320:265-271, 1989. Cholesterol, Fibrinogen and Blood Pressure,” 10, 1988.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2709

75. Inagaki, M., W. S. Harris, “Changes in Dissonance Between Expected and Observed Controlled Double Blind Study,” Clinical Lipoprotein Composition in Events,” Journal of Intern Medicine Chemical Action, 178:251-260, 1988. Hypertriglyceridemic Patients Taking Supplement, 225:11-20, 1989. 106. Milner, M. R., et al., “Usefulness of Cholesterol-free Fish Oil Supplements,” 90. Landymore, R. W., M. MacAulay, B. Fish Oil Supplements in Preventing Clinical Atherosclerosis, 82:237-246, 1990. Sheridan, C. Cameron, “Comparison of Cod- Evidence of Restenosis After Percutaneous 76. Iso, H., et al., “Serum Fatty Acids and liver Oil and Aspirin-dipyridamole for the Transluminal Coronary Angioplasty,” Fish Intake in Rural Japanese, Urban Prevention of intimal Hyperplasia in American Journal of Cardiology, 64:294-299, Japanese, Japanese American and Caucasian Autologous Vein Grafts,” Annals of Thoracic 1989. American Men,” Internal Journal of Surgery, 41:54-57, 1986. 107. Molgaard, J., H. Von Schenck, C. Epidemiology, 18:374-81, 1989. 91. Leaf, A., P. C. Weber, “Cardiovascular Lassvik, T. Kuusi, A. G. Olsson, “Effect of 77. Jenson, T., S. Slender, K. Goldstein, G. Effects of N-3 Fatty Acids,” New England Fish Oil Treatment on Plasma Lipoprotems Holmer, T. Deckert, “Partial Normalization Journal of Medicine, 318:549-557, 1988. in Type III Hyperlipoproteinaemia,” by Dietary Cod-liver Oil of Increased 92. Lehtonen, A., I. Raiha. R. Puumalainen, Atherosclerosis, 81:1-9, 1990. Microvascular Albumin Leakage in Patients A. Seppanen, J. Mamiemi, “The Effect of the 108. Mori, T. A., R. Vandongen, J. R. L. with Insulin-dependent Diabetes and Short-term Administration of Fish Oil on Masarei, D. Dunbar, K. G. Stanton, “Serum Albuminuria,” New England Journal of Serum Lipoproteins in Old People,” Lipids in Insulin-dependent Diabetics are Medicine, 321:1572-1577, 1989. Gerontology, 35:311-314, 1989. Markedly Altered by Dietary Fish Oils,” 78. Kagawa, Y., et al., “Eicosapolyenoic 93. Lempert, K. D., J. S. Rogers, M. J. Clinical Experimental Pharmacology and Acids of Serum Lipids of Japanese Islanders Albrink, “Effects of Dietary Fish Oil on Physiology, 15:333-337, 1988. with Low Incidence of Cardiovascular Serum Lipids and Blood Coagulation in 109. Mori, T. A., R. Vandongen, J. R. L. Diseases,” Journal Nutritional Sciences and Peritoneal Dialysis Patients,” American Masarei, K. G. Stanton, D. Dunbar, “Dietary Vitaminology, 28:441-453, 1982. Journal of Kidney Diseases, 11:170-175, Fish Oils Increase Serum Lipids in Insulin- 79. Kusion, S. E., et al., “Effect of Omega- 1988. dependent Diabetics Compared with Healthy 3 Fish Oils on Lipid Metabolism Glycemic 94. Levine, P. H., et al., “Dietary Controls,” ‘Metabolism, 38:404-409, 1989. Control and Blood Pressure in Type II Supplementation with Omega-3 Fatty Acids 110. Mori, T. A., R. Vandongen, J. R. Diabetic Patients,” Journal of Clinical Prolongs Platelet Survival in Hyperlipidemic Masarei, “Fish Oil-induced Changes in Endacrinohgy and Metabolism, 67:5, 1988. Patients with Atherosclerosis,” Archives Apolipoproteins in IDDM Subjects,” Diabetes 80. Kestin, M., P. Clifton, G. B. Belling, P. Internal Medicine, 149:1113-1116, 1989. Care, 13:725-732, 1990. J. Nestel, “N-3 Fatty Acids of Marine Origin 95. Levinson. P. D., A. H. losiphidis, A. L. 111. Nestel, P. J., “Effects of N-3 Fatty Lower Systolic Blood Pressure and Saritelli, P. N. Herbert, M. Steiner, “Effects of Acids on Lipid Metabolism,” Annual Triglycerides but Raise LDL Cholesterol N-3 Fatty Acids in Essential Hypertension,” Reviews of Nutrition, 10:149-67, 1990. Compared with N-3 and N-6 Fatty Acids American Journal of Hypertension, 3:7 54- 112. Nordoy, A., S. H. Goodnight, “Dietary From Plants,” American Journal of Clinical 760, 1990. Lipids and Thrombosis,” Arteriosclerosis, Nutrition, 51:1028-1034, 1990. 96. Li, X., Steiner, M., “Fish Oil: A Potent 10:149-163, 1990. 81. Kim, D. N., H.T. Ho, D. A. Lawrence, Inhibitor of Platelet Adhesiveness,” Blood, 113. Noreli, S E., et al., “Fish J. Schmeo, W. A. Thomas, “Modification of 76:938-945, 1990. Consumption and Mortality from Coronary Lipoprotein Patterns and Retardation of 97. Lichtenstein, A. H., A. V, Chobanian, Heart Disease,” British Medical Journal, Atherogenesis by a Fish Oil Supplement to “Effect of Fish Oil on Atherogenesis in 293:426, 1985. a Hyperlipidemic Diet for Swine,” Watanabe Heritable Hyperlipidemic Rabbit,” 114. Nozaki. S., A. Garg, G. L. Vega, S. M. Atherosclerosis, 76:35-54, 1989. Arteriosclerosis, 10:597-606, 1990. Grundy, “Postheparin Lipolytic Activity and 82. Kinselia, J. E., B. Lokesh, R. A. Stone, 98. Lox, C. D., “The Effects of Dietary Plasma Lipoprotein Response to W-3 “Dietary N-3 Polyunsaturated Fatty Acids Marine Fish Oils (Omega-3 Fatty Acids) on Polyunsaturated Fatty Acids in Patients with and Amelioration of Cardiovascular Disease: Coagulation Profiles in Men,” General Primary Hypertriglyceridemia,” American Possible Mechanisms,” American Journal of Pharmacology, 21:241-6, 1990. Journal of Clinical Nutrition, 53:638-642, Clinical Nutrition, 52:1-28, 1990. 99. Lox, C. D., “Effects of Marine Fish Oil 1991. 83. Klasing, S. A., S. M. Pilch, “Review of (Omega-3 Fatty Acids) on Lipid Profiles in 115. National Research Council/National the Epidemiological and Clinical Evidence Women,” General Pharmacology, 21:295-8, Academy of Science, National Research on the Role of Omega-3 Fatty Acids in Health 1990. Council/National Academy of Sciences, Diet and Disease,” LSRO, Rockville, MD, 1986. 100. LSRO, “Cardiovascular Effects from and Health, Implications for Reducing 84. Knapp, H. R., “Polyunsaturates, Omega-3 Fatty Acids,” Federation of Chronic Disease Risk, National Academy Endogenous Eicosanoids, and Cardiovascular American Societies for Experimental Biology, Press, Washington, DC, 1989. Disease,” Journal of the American College of Rockville, MD, 1991. 116. Parks, J. S., J. Kaduck-Sawyer, B. C. Bullock, L. L. Rudel, “Effect of Dietary Fish Nutrition, 9:344-351, 1990. 101. Margolin, G., et al., “Blood Pressure Oil on Coronary Artery and Aortic 85. Knapp, H. R., G. A. FitzGerald, “The Lowering in Elderly Subjects: a Double-blind Atherosclerosis in African Green Monkeys,” Antihypertensive Effects of Fish Oil,” New Crossover Study of W-3 and W-6 Fatty Arteriosclerosis, 10:1102-1112, 1990. England Journal of Medicine, 320:1037-1043, Acids,” American Journal of Clinical 1989. Nutrition, 53:562-572, 1991. 117. Radack, K., C. Deck, G. Huster, “Dietary Supplementation with Low-dose 86. Kristensen, S. D., E. B. Schmidt, J. 102. Meade, T. W., et al., “Haemostatic Fish Oils Lowers Fibrinogen Levels: A Dyerberb, “Dietary Supplementation with N- Function and Ischaemic Heart Disease: Randomized, Double-blind Controlled 3 Polyunsaturated Fatty Acids and Human Principle Results of the Northwick Park Heart Study,” Annals of Internal Medicine, Platelet Function; A Review with Particular Study,” Lancet, ii:533-537, 1986. 111:757-758, 1989. Emphasis on Implications for Cardiovascular 103. Mehta, J. L., L. M. Lopez, D. Lawson, 118. Radack, K., C. Deck, G. Huster, “The Disease,” Journal Internal Medicine T. J. Wargovich, L. L. Williams, “Dietary Supplement, 225:141-150, 1989. Supplementation with Omega-3 Comparative Effects of N-3 and N-6 87. Kromhout, D., E. B. Bosschieter, C Polyunsaturated Fatty Acids in Patients with Polyunsaturated Fatty Acids on Plasma Coulander, “The Inverse Relation Between Stable Coronary Heart Disease,” American Fibrinogen Levels: A Controlled Clinical Fish Consumption and 20-year Mortality Journal of Medicine, 84:45-52, 1988. Trial in Hypertriglyceridemic Subjects,” From Coronary Heart Disease,” New England 104. Mehta, J., D. Lawson, T, Saldeen, Journal American College Nutrition, 9:352- Journal of Medicine, 312:1205-1209, 1985. “Reduction in Plasminogen Activator 357, 1990. 88. Kromhout, D., “N-3 Fatty Acids and Inhibitor-1 (PAI-1) with Omega-3 119. Radack, K. L., C. C. Deck, G. A. Coronary Heart Disease: Epidemiology from Polyunsaturated Fatty Acid (PUFA) Intake,” Huster, “N-3 Fatty Acid Effects on Lipids, Eskimos to Western Populations,” Journal American Heart Journal, 116:1201-1206, Lipoproteins, and Apolipoproteins at Very Internal Medicine Supplement, 225:47-51, 1988. Low Doses: Results of a Randomized 1989. 105. Miller, J. P., et al., “Triglyceride Controlled Trial in Hypertriglyceridemic 89. Lands, W. E., “N-3 Fatty Acids as Lowering Effect of MaxEPA Fish Lipid Subjects,” American Journal of Clinical Precursors for Active Metabolic Substances: Concentrate: A Multicentre Placebo Nutrition, 51:599-605, 1990.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2710

120. Reis, G. J., et al., “Effects of Two and Leukocyte Chemotaxis—Effects in After ingestion of Omega-3 Fatty Acids in Types of Fish Oil Supplements on Serum Hyperlipidemia and Dose-response Studies Human Subjects,” Thrombosis Research, Lipids and Plasma Phospholipid Fatty Acids in Healthy Men,” Arteriosclerosis and 54:573-82, 1989. in Coronary Artery Disease,” American Thrombosis, 11:429-435, 1991. 151. Thiery, J., D, SeideI, “Fish Oil Feeding Journal of Cardiology, 66:1171-1175, 1990. 137. Shekelle, R. B., et al., “Fish Results in an Enhancement of Cholesterol- 121. Reis, G. J., et al., “Randomized Trial Consumption and Mortality from Coronary induced Atherosclerosis in Rabbits,” of Fish Oil for Prevention of Restenosis After Heart Disease,” New England Journal of Atherosclerosis, 63:53-58, 1987. Coronary Angioplasty,” Lancet, 11:177-181, Medicine, 313:820, 1985. 152. Trip, M. D., V. M. Cats, E. J. L. Van 1989. 138. Shimokawa, H., P. M. Vanhoutte, Capelle, J. Vreeken, “Platelet Hyperreactivity 122. Rich et al., 1987 [to be added]. “Dietary Omega-3 Fatty Acids and and Prognosis in Survivors of Myocardial 123. Rich, S., et al., “Development of Endothelium-dependent Relaxations in Infarction,” New England Journal of Atherosclerosis in Genetically Hyperlipidemic Porcine Coronary Arteries,” American Medicine, 322:1549-1554, 1990. Rabbits During Chronic Fish-oil Ingestion,” Journal of Physiology, 256:H968-H973, 1988. 153. Urakaze, M., T. Hamazaki, H. Arteriosclerosis, 9:189-194, 1989. 139. Silverman. D. L., J. A. Ware, F. M. Kashiwabara, K. Omori, “Favorable Effects of 124. Rillaerts, E. G., G. J. Engelmann, K. M. Sacks, R. C. Pasternak, “Comparison of the Fish Oil Concentrate on Risk Factors for Van Camp, I. De Leeuw, “Effect of Omega-3 Absorption and Effect on Platelet Function of Thrombosis in Renal Allograft Recipients,” Fatty Acids in Diet of Type I Diabetic a Single Dose of N-3 Fatty Acids Given as Nephron, 53:102-109, 1989. Subjects on Lipid Values and Fish or Fish Oil,” American Journal of 154. Vacek, J. L., W. S. Harris, K. Haffey, Hemorheological Parameters,” Diabetes, Clinical Nutrition, 53:1165-1170, 1991. “Short-term Effects of Omega-3 Fatty Acids 38:1412-1416, 1989. 140. Simons, L. A., A. Parfitt, J. Simons, S. on Exercise Stress Test Parameters, Angina 125. Rodgers, R. P. C., J. Levin, “A Critical Balasubramaniam, “Effects of an Ethyl Ester and Lipoproteins,” Biomedicine Reappraisal of the Bleeding Time,” Seminars Preparation of Fish Oils (Himega) on Lipids Pharmacotherapeutics, 43:375-379, 1989. in Thrombosis and Hemostasis, 18:1-20 and Lipoproteins in Hyperlipidaemia,” 155. Valdini et al. [to be added]. 1990. Australia New Zealand Journal Medicine, 156. Van Houwelingen, R., H. 126. Rogers, K. A., M. J. Karnovsky 20:689-694, 1990. Zevenbergen, P. Groot, A. Kester, G. “Dietary Fish Oil Enhances Monocyte 141. Simonsen, T., A. Vartun, V. Lyngmo, Hornstra, “Dietary-fish Effects on Serum Adhesion and Fatty Streak Formation in the A. Nordoy, “Coronary Heart Disease, Serum Lipids and Apolipoproteins, a Controlled Hypercholesterolemic Rat,” American Lipids, Platelets, and Dietary Fish Study,” American Journal of Clinical Journal of Pathology, 132:382-388, 1988. Consumption in Two Communities in Nutrition, 51:393-398, 1990. 127. Sanders, T. A., A. Hinds, C. C. Pereira, Northern Norway,” Acta Medica 157. Vessby, B., M. Boberg, “Dietary “Effects of Dietary Fish Oil on Blood Lipids Scandmavica, 222:237-245, 1987. Supplementation with N-3 Fatty Acids May in Normal Subjects,” Journal of Internal 142. Simonsen, T., A. Nordoy, C. Impair Glucose Homeostasis in Patients with Medicine Supplement, 225:99-104, 1989. Sjunneskog, V. Lyngmo, “The Effects of Cod Non-insulin-dependent Diabetes Mellitus,” 128. Schectman, G., S. Kaul, A. H. Liver Oil in Two Populations with Low and Journal of Internal Medicine, 228:165-171, Kissebah, “Effect of Fish Oil Concentrate on High Intake of Dietary Fish,” Acta Medica 1990. Lipoprotein Composition in NIDDM,” Scandanavica, 223:491-498, 1988. 158. VoIIset, S. E., I. Heuch, E. Bjelke, Diabetes, 37:1567-1573, 1988. 143. Skeaff, C. M., B. J. Holub, “The Effect “Fish Consumption and Mortality from 129. Schectman, G., S. Kaul, A. H. of Fish Oil Consumption on Platelet Coronary Heart Disease,” New England Kissebah, “Heterogeneity of Low Density Aggregation Responses in Washed Human Journal of Medicine, 313:820, 1985. Lipoprotein Responses to Fish-oil Platelet Suspensions,” Thrombosis Research, 159. Von Schacky, C. “Prophylaxis of Supplementation in Hypertriglyceridemic 51:105-115, 1988. Atherosclerosis with Marine Omega-3 Fatty Subjects,” Arteriosclerosis, 9:345-354, 1989. 144. Smith, P., H. Arnesen, T. Opstad, K. Acids,” Annals of Infernal Medicine, 130. Schectman, G., S. Kaul, G. D. Cherayil, H, Dahl, J. Eritsland, “Influence of Highly 107:890-899, 1987. M. Lee, A. Kissebah, “Can the Concentrated N-3 Fatty Acids on Serum 160. Wahlqvist, M. L., C. S. Lo, K. A. Hypotriglyceridemic Effect of Fish Oil Lipids and Hemostatic Variables in Survivors Myers. “Fish Intake and Aterial Wall Concentrate Be Sustained?”, Annals of of Myocardial Infarction Receiving Either Characteristics in Healthy People and Internal Medicine, 110:346-352, 1989. Oral Anticoagulants or Matching Placebo,” Diabetic Patients,” Lancet, 944-946, 1989. 131. Schmidt, E. B., S. D. Kristensen, J. Thrombosis Research, 53:467-474, 1989. 161. Weber, P. C., “Clinical Studies on the Dyerberg, “The Effect of Fish Oil on Lipids, 145. Solomon, S. A., et al., “A Placebo- Effects of N-3 Fatty Acids on Cells and Coagulation and Fibrinolysis in Patients with controlled, Double Blind Study of Eicosanoids in the Cardiovascular System,” Angina Pectoris,” Artery, 15:316-329, 1988. Eicosapentaenoic Acid-rich Oil in Patients Journal of Internal Medicine Supplement, 132. Schmidt, E, B., S, D. Kristensen, P. J. with Stable Angina Pectoris,” Current 225:61-68, 1989. Sorensen, J. Dyerberg, “Antithrombin III and Medical Research and Opinion, 12:1-11, 162. Weber, P. C., A. Leaf, “Cardiovascular Protein C in Stable Angina Pectoris-influence 1990. Effects of W3 Fatty Acids,” World Review of of Dietary Supplementation with 146. Stacpoole, P. W., J. Alig, L. Ammon, Nutrition Dietetics, 66:218-232, 1991. Polyunsaturated Fatty Acids,” Scandinavian S. E. Crockett, “Dose-response Effects of 163. Weintraub, M. S., R. Zechner, A. Journal Clinical Laboratory Investigation, Dietary Marine Oil on Carbohydrate and Brown, S. Eisenberg, J. L. Breslow, “Dietary 48:469-473, 1988. Lipid Metabolism in Normal Subjects and Polyunsaturated Fats of the W-6 and W-3 133. Schmidt, E. B., E. Ernst, K. Varming, Patients with Hypertriglyceridemia,” Series Reduce Postprandial Lipoprotein J. O. Pederson, J. Dyerberg, “The Effect of N- Metabolism, 38:946-56, 1989. Levels,” Journal of Clinical Investigation, 3 Fatty Acids on Lipids and Haemostasis in 147. Sterner, A., et al., “Effect of Fish Oil 82:1884-1893, 1988. Patients with Type IIa and Type IV on Blood Pressure and Serum Lipids in 164. Wilt, T.J., et al., “Fish Oil Hyperlipidemia,” Thrombosis and Hypertension and Hyperlipidaemia,” Journal Haemostasis, 62:797-801, 1989. Supplementation Does Not Lower Plasma Hypertension Supplement, 7:S73-76, 1989. Cholesterol in Men with 134. Schmidt, E. B., et al., “The Effects of 148. Subbaiah, P. V., M. H. Davidson, M. N-3 Polyunsaturated Fatty Acids on Lipids, Hypercholesterolemia,” Annals of Internal C Ritter, “Effects of Dietary Supplementation Medicine, 111:900-905, 1989. Haemostasis, Neutrophil and Monocyte with Marine Lipid Concentrate on the Plasma 165. Zhu, B-Q., D. L. Smith, R. E. Sievers, Chemotaxis in Insulin-dependent Diabetes Lipoprotein Composition of W. M. Isenberg, W. W. Parmley, “Inhibition Mellitus,” Journal of Internal Medicine, Hypercholesterolemic Patients,” of Atherosclerosis by Fish Oil in Cholesterol- 225:201-206, 1989. Atherosclerosis, 79:157-166, 1989. fed Rabbits,” Journal of American College of 135. Schmidt, E. B., J. O. Pederson, S. 149. Sweny, P., et al., “Dietary Fish Oil Cardiology, 12:1073-1078, 1988. Ekelund, N. Gmnnet, C. Jersild, J. Dyerberg, Supplements Preserve Renal Function in 166. Zucker, M. L., D. S. Bilyeu, G. M. “Cod Liver Oil Inhibits Neurtophil and Renal Transplant Recipients with Chronic Helmkamp, W. S. Harris, C. A. Dujovne, Monocyte Chemotaxis in Healthy Males,” Vascular Rejection,” Nephrology Dialysis “Effects of Dietary Fish Oil on Platelet Atherosclerosis, 77:53-57, 1989. Transplantation, 4:1070-1075, 1989. Function and Plasma Lipids in 136. Schmidt. E. B., J. O. Pederson, K. 150. Takimoto, G., J. Galang, G. K. Lee, B. Hyperlipoprotenemic and Normal Subjects,” Varming, E. Ernst, “N-3 Fatty Acids A. Bradlow, “Plasma Fibrinolytic Activity Atherosclerosis, 73:13-22, 1988.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2711

167. Beswick, A. D., A. M. Fehily, D. S. Exposure of Cultured Endothelial Cells to 194. Elwood, P. C.,, et al., “Ischemic Heart Sharp, S. Renaud, and J. Giddings, “Long- Eicosapentaenoic Acid Potentiates the Disease and Platelet Aggregation: The Term Diet Modification and Platelet Release of Endothelium-Derived Relaxing Caerphilly Collaborative Heart Disease Activity,” Journal of Internal Medicine, 229, Factor(s),” British Journal of Pharmacology, Study,” Circulation, 83, 38-44, 1991. 511-515, 1991. 99, 176-180, 1990. 195. Eritsland, J., H. Araesen, P. Smith, I. 168. Agren, J. J., O. Hanninen, M. Laitinen, 182. Braden, G. A., H. R. Knapp, D. J. Seljeflot, and K. Dahl, “Effects of Highly K. Seppanen, I. Bernhardt, L. Fogelholm, J. Fitzgerald, and G. A. FitzGerald, “Dietary Concentrated Omega-3 Polyunsaturated Fatty Herranen, and I. Penttila, “Boreal Freshwater Fish Oil Accelerates the Response to Acids and Acetylsalicylic Acid, Alone and Fish Diet Modifies the Plasma Lipids and Coronary Thrombolysis With Tissue-Type Combined, on Bleeding Time and Serum Prostanoids and Membrane Fatty Acids in Plasminogen Activator: Evidence for a Lipid Profile,” Journal of Oslo City Hospital, Man,” Lipids, 23, 924-929, 1988. Modest Platelet Inhibitory Effect In Vivo,” 39, 97-101, 1989. 169. American Heart Association, “Heart Circulation, 32, 178-187, 1990. 196. Fahrer, H., F. Hoeflin, B. H. and Stroke Facts,” 1992. 183. Brown, A. J. and D. C. K. Roberts, Lauterburg, E. Peheim, A. Levy, and T. L. 170. Annuzzi, G., A. Rivellese, B. Capaldo “Fish and Fish Oil Intake: Effect On Visher, “Diet and Fatty Acids: Can Fish L. Di Marino, C. lovine, G, Marotta, and G. Haematological Variables Related To Substitute for Fish Oil?”, Clinical and Riccardi, “A Controlled Study on the Effects Cardiovascular Disease,” Thrombosis Experimental Rheumatology, 9, 403-406, of N-3 Fatty Acids on Lipid and Glucose Research, 64, 169-178, 1991. 1991. Metabolism in Non-Insulin-Dependent 184. Brown, J. E. and K. W. J. Wahle, 197. Fenetti, A., et al., “Effect of Fish Oil Diabetic Patients,” Atherosclerosis, 87, 65- “Effect of Fish-Oil and Vitamin E Supplementation on the Excretion of the 73, 1991. Supplementation on Lipid Peroxidation and Major Metabolite of Prostaglandin E in 171. Austin, M, A., M. C. King, K. M. Whole-Blood Aggregation in Man,” Clinica Healthy Male Subjects,” Lipids, 26, 500-503, Vranizan, and R. M. Krauss, “Atherogenic Chimica Acta, 193, 147-156, 1990. 1991. Lipoprotein Phenotype A Proposed Genetic 185. Bulliyya, G., et al., “Lipid Profiles 198. Feskens, E. J. M., C. H. Bowles, and Marker for Coronary Heart Disease Risk,” Among Fish-Consuming Coastal and Non- D. Kromhout, “Inverse Association Between Circulation, 82, 495-506, 1990. Fish-Consuming Inland Opulations,” Fish Intake and Risk of Glucose Intolerance 172. Bairati, I., L. Roy, and F. Meyer, European Journal of Clinical Nutrition, 44, in Normoglycemic Elderly Men and “Double-Blind, Randomized, Controlled Trial 481-485, 1990. Women,” Diabetes Care, 14, 935-941, 1991. of Fish Oil Supplements in Prevention of 185a. Buring, J. E. et al., “Decreased HDL2 199. Force, T., et al., “N-3 Fatty Acids Recurrence of Stenosis After Coronary and HDL3 Cholesterol, ApoA-I and Apo A- Increase Postischemic Blood Flow But Do Angioplasty,” Circulation, 85, 950-956, 1992. II, and Increased Risk of Myocardial Not Reduce Myocardial Necrosis,” American 173. Bairati, I., L. Roy, and F. Meyer, Infarctlon,” Circulation, 85, 22-29, 1992. Physiology Society, 26, H1204-H1210, 1989. “Effects of a Fish Oil Supplement on Blood 186. Cahill, P. D., et al., “Inhibition of Vein 200. Force, T., et al., “Aspirin-Induced Pressure and Serum Lipids in Patients Graft Intimal Thickening by Decline in Prostacyclin Production in Treated for Coronary Artery Disease,” Eicosapentaenoic Acid: Reduced Patients With Coronary Artery Disease Is Due Canadian Journal of Cardiology, 8, 41-46, Hromboxane Production Without Change in to Decreased Endoperoxide Shift,” 1992. Lipoprotein Levels or Low-Density Circulation, 84, 2286-2293, 1991. 174. Beil, F. U., W. Terres, M. Orgass, and Lipoprotein Receptor Density,” Journal of 201. Fox, P. L. and P. E. DiCorleto, “Fish H. Greten, “Dietary Fish Oil Lowers Vascular Surgery, 7, 108-118, 1988. Oils Inhibit Endothelial Cell Production of Lipoprotein(a) in Primary 187. Canadian Task Force on the Periodic Platelet-Derived Growth Factor-Like Hypertriglyceridemia,” Atherosclerosis, 90, Health Examination. “Periodic Health Protein,” Science, 241, 453-456, 1988. 95-97, 1991. Examination, 1991 Update: 6. Acetylsalicylic 202. Franceschini, G., L. Calabresi, P. 175. Bhathena, S. J., et al., “Effects of W3 Acid and the Primary Prevention of Maderna, C. Galli, G. Gianfranceschi, and C. Fatty Acids and Vitamin E on Hormones Cardiovascular Disease,” Canadian Medical R. Sirtori, “W-3 Fatty acids Selectively Raise Involved in Carbohydrate and Lipid Association Journal, 145, 1091-1095, 1991. High-Density Lipoprotein 2 Levels in Healthy Metabolism In Men,” American Journal of 188. Charnock, J., “Antiarrhythmic Effects Volunteers,” Metabolism, 10, 1283-1286, Clinical Nutrition, 54, 684-688, 1991. of Fish Oils,” World Reviews of Nutrition and 1991. 176. Bjerregaard, P. and J. Dyerberg, Dietetics, 66, 278-291, 1991. 203. Cans, R. O. B., et al., “Fish Oil “Mortality From Ischaemic Heart Disease and 189. Clarke, J. T. R., G. Cullen-Dean, E. Supplementation in Patients with Stable Cerebrovascular Disease in Greenland,” Regelink, L. Chan, and V. Rose. “Increased Claudication,” The American Journal of International Journal of Epidemiology, 17, Incidence of Epistaxis in Adolescents With Surgery, 160, 490-495, 1990. 514-519, 1988. Familial Hypercholesterolemia Treated With 204. Gazso, A., J. Kaliman, D. Horrobin, 177. Bonaa, K. H., K. S. Bjerve, and A. Fish Oil,” Journal of Pediatrics, 116, 139- and H. Sinzinger, “Effects of Omega-3 Fatty Nordoy, “Docosahexaenoic and 141, 1991. Acids on the Prostaglandin System in Eicosapentaenoic Acids in Plasma 190. Cobiac, L., P. J. Nestel, L. M. H. Wing, Healthy Volunteers,” Prostaglandins in Phospholipids are Divergently Associated and P. R. C. Howe, “Effects of Dietary Clinical Research: Cardiovascular System, with High Density Lipoprotein in Humans,” Sodium Restriction and Fish Oil 517-521, 1989. Arteriosclerosis and Thrombosis, 12, 675— Supplements on Blood Pressure in the 205. Gerasimova, E., et al., “The Effect of 681, 1992. Elderly,” Clinical and Experimental Dietary N-3 Polyunsaturated Fatty Acids on 178. Bonaa, K. H., K. S. Bjerve, and A. Pharmacology and Physiology, 18, 265-268, HDL Cholesterol in Chukot Residents vs Nordoy, “Habitual Fish Consumption, 1991. Muscovites,” Lipids, 26, 261-265, 1991. Plasma Phospholipid Fatty Acids, and Serum 191.Davidson, M.H., J. H. Burns, P.V. 206. Gerhard, G. T., B. D. Patton, S. A. Lipids: The Tromso Study,” American Subbaiah, M. E. Conn, and K. B. Drennan, Lindquist, and R. C. Wander, “Comparison of Journal of Clinical Nutrition, 55, 1126-1134, “Marine Oil Capsule Therapy for the Three Species of Dietary Fish: Effects on 1992. Treatment of Hyperlipidemia,” Archives of Serum Concentrations of Low-Density- 179. Bordet, J. C., M. C. Trzeciak, S. Internal Medicine, 151, 1732-1740, 1991. Lipoprotein Cholesterol and Apolipoprotein Durbin, and M. Dechavanne, “Dose-Response 192. Demiroglu, C., et al., “Suppression of in Normotriglyceridemic Subjects,” of Platelet Aggregation and Bleeding Time Atherogenesis by N-3 Fatty Acids in the American Journal Clinical Nutrition, 54, 334- During Eicosapentaenoic Acid Intake,” Cholesterol-Fed Rabbit,” Angiology—The 339, 1991. Thrombosis and Haemostasis, 65, 110, 1991. Journal of Vascular Diseases, 323-330, 1991. 207. Goren, A., H. Stankiewicz, R. 180. Boudreau, M. D., et al., “Lack of Dose 193. Edwards, I. J., A. K. Gebre, W. D. Goldstein, and A. Drukker, “Fish Oil Response by Dietary N-3 Fatty Acids at a Wagner, and J. S. Parks, “Reduced Treatment of Hyperlipidemia in Children and Constant Ratio of N-3 to N-6 Fatty Acids in Proteoglycan Binding of Low Density Adolescents Receiving Renal Replacement Oppressing Eicosanoid Biosynthesis From Lipoproteins From Monkeys (Macaca Therapy,” Pediatrics, 88, 265-268, 1991. Arachidonic Acid,” American Journal of fascicularis) Fed a Fish Oil Versus Lard 208. Gotto, A. M., J. Patsch, and A. Clinical Nutrition, 54, 111-117, 1991. Diet,” Arteriosclerosis and Thrombosis, 11, Yamamoto, “Postprandial Hyperlipidemia,” 181. Boulanger, C., V. B. Schini, H. 1778-1785, 1991. American Journal of Cardiology, 68, 11A- Hendrickson, and P. M. Vanhoutte, “Chronic 12A, 1991.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2712

209. Green, P., et al., “Effects of Fish-Oil Fatty Acids on Myocardial Ischemia and Macrophages is Enhanced by Dietary N-3 Ingestion on Cardiovascular Risk Factors in Reperfusion,” American Journal of polyunsaturated Fatty Acids,” Immunology Hyperlipidemic Subjects in Israel: A Physiology, 259, H1518-H1526, 1990. Letters, 23, 281-286, 1989/1990. Randomized, Double-Blind Crossover 224. Homma, Y., K. Ohshima, H. 237. Lytle, J. S. and T. F. Lytle, “W3 Fatty Study,” American Journal of Clinical Yamaguchi, H. Nakamura, G. Araki, and Y. Acids in 40 Species of Warm Water Finfish,” Nutrition, 52, 1118-1124, 1990. Goto, “Effects of Eicosapentaenoic Acid on World Review of Nutrition and Dietetics, 66, 210. Haglund, O., R. Luostarinen, R. Plasma Lipoprotein Subtractions and 513-514, 1991. Wallin, L. Wibell, and T. Saldeen, “The Activities of Lecithin: Cholesterol 238. Malasanos, T. H. and P. W. Stacpoole, Effects of Fish Oil on Triglycerides, Acyltransferase and Lipid Transfer Protein,” “Biological Effects of W-3 Fatty Acids in Cholesterol, Fibrinogen and Atherosclerosis, 91, 145-153, 1991. Diabetes Mellitus,” Diabetes Care, 14, 1160- Malondialdehyde in Humans Supplemented 225. Innis, S. M.,H. V. Kuhnlein, and D. 1179, 1991. with Vitamin E,” Journal of Nutrition, 121, Kinloch, “The Composition of Red Cell 239. Malis, C. D., P. C. Weber, A. Leaf, and 165-169, 1991. Membrane Phospholipids in Canadian Inuit J. V. Bonventre, “Incorporation of Marine 211. Haines, A. P., et al., “Effects of a Fish Consuming a Diet High in Marine Lipids into Mitochondrial Membranes Oil Supplement on Platelet Function, Mammals,” Lipids, 23, 1064-1068, 1988. Increases Susceptibility to Damage by Haemostatic Variables and Albuminuria in 226. Jensen, C. D., G. A. Spiller, V. J. Calcium and Reactive Oxygen Species: Insulin-Dependent Diabetics,” Thrombosis Wookey, L. G. Wong, J. H. Whittam, and J. Evidence for Enhanced Activation of Research, 43, 643-655, 1986. Scala, “Plasma Lipids on Throe Levels of Phospholipase A2 in Mitochondria Enriched 212. Hallaq, H., A. Sellmayer, T. W. Smith, Fish Oil Intake in Healthy Human Subjects,” with N-3 Fattyacids,” Proceedings of the and A. Leaf, “Protective Effect of Nutrition Reports International, 38, 165-172, National Academy of Sciences, 87, 8845- Eicosapentaenoic Acid on Ouabain Toxicity 1988. 8849, 1990. in Neonatal Rat Cardiac Myocytes,” 227. Knapp, H. R., I. A. G. Reilly, P. 240. Malis, C. D., et al., “Effects of Dietary Proceedings of the National Academy of Alessandrini, and G. A. FitzGerald, “In Vivo w3 Fatty Acids on Vascular Contractility in Sciences, 87, 7834-7838, 1990. Indices of Platelet and Vascular Function Preanoxic and Postanoxic Aortic Rings,” 213. Hamakazi, T., M. Urakaze, S. During Fish-Oil Administration in Patients Circulation, 84, 1393-1401, 1991. Sawazaki, K. Yamazaki, H. Taki, and S. Yano, with Atherosclerosis,” New England Journal 241. Malle, E., W. Saltier, E. Prenner, H. J. “Comparison of Pulse Wave Velocity of the of Medicine, 314, 937-942, 1986. Lets, A. Hennetter, A, Gries, and G. M. Aorta Between inhabitants of Fishing and 228. Kremer, J. M., et al., “Dietary Fish Oil Kostner, “Effects of Dietary Fish Oil Farming Villages in Japan,” Atherosclerosis, and Olive Oil Supplementation in Patients Supplementation on Platelet Aggregabllity 73, 157-160, 1988. With Rheumatoid Arthritis,” Arthritis and and Platelet Membrane Fluidity in 214. Hamakazi, T., S. Fischer, M. Urakaze, Rheumatism, 33, (1990):810-820. Normolipemic Subjects with and without S. Sawazaki, S. Yano, and T. Kuwamori, 229. Lamers, J. M. J., J. M. Hartog, C. High Plasma Lp(a) Concentrations,” “Urinary Excretion of PGI 2/3-M and Recent Guamieri, I. Vaona, P. D. Verdouw, and J. F. Atherosclerosis, 88, 193-201, 1991. N-6/N-3 Fatty Acid Intake,” Prostaglandins, Koster, “Lipid Peroxidation in Normoxic and 242. Manninen, V., L. Tenkanen, P. 37, 417-424, 1989. Ischemic-Reperfused Hearts of Fish Oil and Koskinen, J. K. Huttunen, M. Manttari, O. 215. Hamazaki, T., E. Takazakura, K. Lard Fed Pigs,” Journal of Molecular and Heinonen, and M. H. Frick, “Joint Effects of Osawa, M. Urakaze, and S. Yano, “Reduction Cellular Cardiology, 20, 605-615, 1988. Serum Triglyceride and LDL Cholesterol and in Microalbuminuria in Diabetics by 230. Lehr, H. A., A. Guhlmann, D. Nolte, HDL Cholesterol Concentrations on Coronary Eicosapentaenoic Acid Ethyl Ester,” Lipids, D. Keppler, and K. Messmer, “Leukotrienes Heart Disease Risk in the Helsinki Heart 25, 541-545, 1990. as Mediators in Ischemia-Reperfusion Injury Study,” Circulation, 85, 37-45, 1992. 216. Hansen, J. B., J. O. Olsen, L. Wilsgard, in a Microcirculation Model in the Hamster,” 243. Manson, J. E., et al., “A Prospective and B. Osterud, “Effects of Dietary Journal of Clinical Investigation, 87, 2036- Study of Aspirin Use and Primary Prevention Supplementation with Cod Liver Oil on 2041, 1991. of Cardiovascular Disease in Women,” Monocyte Thromboplastin Synthesis, 231. Lehr, H. A., C. Hubner, B. Finckh, D. Journal of the American Medical Association, Coagulation and Fibrinolysis,” Journal of Nolle, U. Beisiegel, A. Kohlschutter, and K. 266, 521-527, 1991. Internal Medicine, 225, 133-139, 1989. Messmer, “Dietary Fish Oil Reduces 244. Marckmann, P., J. Jespersen, T. Leth, 217. Harats, D., et al., “Fish Oil Ingestion Leukocyte/Endothelium Interaction and B. Sandstrom, “Effect of Fish Diet Versus in Smokers and Nonsmokers Enhances Following Systemic Administration of Meat Diet on Blood Lipids, Coagualtion and Peroxidation of Plasma Lipoproteins,” Oxidatively Modified Low Density Fibrinolysis in Healthy Young Men,” Journal Atherosclerosis, 90, 127-139, 1991. Lipoprotein,” Circulation, 84, 1725-1731, of Internal Medicine, 229, 562-572, 1991. 218. Harris, W. S., S. Silveria, and C. A. 1991. 245. McDonald, B. E., J. M. Gerrard, V. M. Dujovne, “The Combined Effects of N-3 Fatty 232. Lehr, H. A., et al., “Role of Bruce, and E. J. Cornoer, “Comparison of the Acids and Aspirin on Hemostatic Parameters Leukotrienes in Leukocyte Adhesion Effect of Canola Oil and Sunflower Oil on in Man,” Thrombosis Research, 57, 517-526, Following Systemic Administration of Plasma Lipids and Lipoproteins and on In 1990. Oxidatively Modified Human Low Density Vivo Thromboxane A2 and Prostacyclin 219. Harris, W. S., S. L. Windsor, and C. Lipoprotein in Hamsters,” Journal of Clinical Production in Healthy Young Men,” A. Dujovne, “Effects of Four Doses of n-3 Investigation, 88, 9-14, 1991. American Journal of Clinical Nutrition, 50, Fatty Acids Given to Hyperlipidemic Patients 233. Lehr. H. A., M. D. Monger, D. Nolle, 1382-1388, 1989. for Six Months,” Journal of the American and K. Messmer, “Favorable Effects of 246. McLennan, P. L., M. Y. Abeywardena, College of Nutrition, 10, 220-227, 1991. Dietary Fish Oil on Capillary Perfusion and J. S. Chamock, “Reversal of the 220. Harris, W. S. and S. L Windsor, “N- Homogeneity After Ischemia and Arrhythmogenic Effects of Long-Term 3 Fatty Acid Supplements Reduce Reperfusion,” Transplantation Proceedings, Saturated Fatty Acid Intake by Dietary N-3 Chylomicron Levels In Healthy Volunteers,” 23, 2356-2358, 1991. and N-6 Polyunsaturated Fatty Acids,” American Journal of Clinical Nutrition, 51, Journal of Applied Nutrition, 43, 5-15, 1991. 234. Li, X. and M. Steiner, “Dose Response 221. Hartog, J. M., J. M. J. Lamers, C. E. of Dietary Fish Oil Supplementation on 53-58, 1990. Essed, W. P. Schalkwijk, and P. D. Verdouw, Platelet Adhesion,” Arteriosclerosis and 247. Meland, E., P. Fugelli, E. Laemm, R. “Does Platelet Aggregation Play a Role in the Thrombosis, 11, 39-46, 1991. Ronneberg, and L. Sandvik, “Effect of Fish Oil on Blood Pressure and Blood Lipids in Reduction in Localized Intimal Proliferation 235. Lindgren, F. T., G. L. Adamson, V. G. Men with Mild to Moderate Hypertension,” in Normolipidemic Pigs With Fixed Coronary Shore, G. J. Nelson, and P. C. Schmidt, Scandinavian Journal of Primary Health Artery Stenosis Fed Dietary Fish Oil?”, “Effect of a Salmon Diet on the Distribution Care, 7, 131-135, 1989. Atherosclerosis, 7B, 79-88, 1989. of Plasma Lipoproteins and Apolipoproteins 248. Meydani, M., et al., “Effect of Long- 222. Hendra, T. J., et al., “Effects of Fish in Normolipidemic Adult Men,” Lipids, 26, Oil Supplements in NIDDM Subjects: Term Fish Oil Supplementation on Vitamin 97-101, 1991. E Status and Lipid Peroxidation in Women,” Controlled Study,” Diabetes Care, 13, 821- 236. Lokesh, B. R., T. J. Sayers, and J. E. 829, 1990. Journal of Nutrition, 121, 484-491, 1991. Kinsella, “Interleukin-1 and Tumor Necrosis 249. Mohammed, K. S., T. A. Nagve, and 223. Hock, C. E., L D. Beck, R. C. Bodine, Factor Synthesis by Mouse Peritoneal and D. K. Reibel, “Influence of Dietary N-3 H. Sprecher, “The Metabolism of 20- and 22-

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2713

Carbon Unsaturated Acids in Rat Heart and 261. Oh, S. Y., J, Ryue, C. H. Hsieh, and Arteriosclerosis by Fish Oil,” Journal of Myocytes as Mediated by Feeding Fish Oil,” D. E. Bell, “Eggs Enriched in W-3 Fatty Thoracic Cardiovascular Surgery, 97, 841- Lipids, 25, 854-858, 1990. Acids and Alterations in Lipid 855, 1989. 250. Molvig, J., F. Pociot, H. Worsaae, L. D. Concentrations in Plasma and Lipoproteins 276. Saynor, R. and T. Gillott, “Changes in Wogensen, L. Back, P. Christensen, T. and in Blood Pressure,” American Journal of Blood Lipids and Fibrinogen with a Note on Mandrup-Poulsen, K. Andersen, P. Madsen, Clinical Nutrition, 54, 689-695, 1991. Safety in a Long Term Study on the Effects J. Dyerberg, and J. Nerup, “Dietary 262. Olaf, A., “Effects of Linolenic and of N-3 Fatty Acids in Subjects Receiving Fish Supplementation with W-3 Polyunsaturated Alpha Linolenic Acids Intake on Blood Oil Supplements and Followed for Seven Fatty Acids Decreases Mononuclear Cell Pressure in Man,” Progress in Clinical Years,” Lipids, 27, 533-538, 1992. Proliferation and Interleukin-1 B Content but Biologic Research, 301, 523-528, 1989. 277. Schmidt, E. B., K. Vanning, E. Ernst, Not Monokine Secretion in Healthy and 263. Olivieri, O., et al., “Effects of Dietary P. Madsen, and J. Dyerberg, “Dose-Response Insulin-Dependent Diabetic Individuals,” Fish Oil on Malondialdehyde Production and Studies on the Effect of N-3 Polyunsaturated Scandinavian Journal of Immunology, 34, Glutathione Peroxidase Activity in Fatty Acids on Lipids and Haemostasis,” 399-410, 1991. Hyperlipidaemic Patients,” Scandinavian Thrombosis and Haemostasis, 63, 1-5, 1990. 251. Mori, T. A., R. Vandongen, F. Journal of Clinical Laboratory Investigation, 278. Schmidt, E. B., L. K. Nielsen, J. O. Mahanian, and A. Douglas, “The Effect of 48, 659-665, 1988. Pedersen. H. J. Kornerup, and J. Dyerberg, Fish Oil on Plasma Lipids, Platelet and 264. Owens, M. R. and W. T. Cave, “The Effect of N-3 Polyunsaturated Fatty Neutrophil Function in Patients with “Dietary Fish Lipids Do Not Diminish Acids on Lipids, Platelet Function, Vascular Disease,” Advances in Platelet Adhesion to Subendothelium,” Coagulation, Fibrinolysis and Monocyte Prostaglandin, Thromboxane, and British Journal of Haematology, 75, 82-85, Chemotaxis in Patients with Hypertension,” Leukotriene Research, 21, 229-232, 1990. 1990. Clinica Chimica Acta, 189, 25-32, 1990. 252. Mori, T. A., R. Vandongen, J. R. L. 265. Peto, R., et al., “Randomized Trial of 279. Schmidt, E. B., I. C. Klausen, S. D. Maserei, I. L. Rouse, and D. Dunbar, Prophylactic Daily Aspirin in British Male Kristensen, H. H. Lervang, O. Faergeman, and “Comparison of Diets Supplemented with Doctors,” British Medical Journal, 296, 313- J. Dyerberg, “The Effect of N-3 Fish Oil or Olive Oil on Plasma Lipoproteins 316, 1988. Polyunsaturated Fatty Acids on Lp(a),” Clinica Chemica Acta, 198, 271-278, 1991. in Insulin-Dependent Diabetics,” 266. Popeski, D., L. R. Ebbeling, P. B. Metabolism, 40, 241-246, 1991. Brown, G. Homstra, and J. M. Gerrard, 280. Scientific Review Committee, Health 253. Mueller, B. A., R. L. Talbert, C. H. “Blood Pressure During Pregnancy in and Welfare, Canada, “Nutrition Tegeler, and T. J. Prihoda, “The Bleeding Canadian Inuit: Community Differences Recommendations,” 1990. Time Effects of a Single Dose of Aspirin in Related to Diet,” Canadian Medical 281. Seidelin, K. N., B. Myrup, and B. Subjects Receiving Omega-3 Fatty Acid Association Journal, 145, 445-454, 1991. Fischer-Hansen, “N-3 Fatty Acids in Dietary Supplementation,” Journal of 267. Phillipson, B. E., D. W. Rothrock, W. Adipose Tissue and Coronary Artery Disease are Inversely Related,” American Journal of Clinical Pharmacology, 31, 185-190, 1991. E. Connor, W. S. Harris, and D. R. Clinical Nutrition, 55, 1117-1119, 1992. 253a. Muller, A. D., A. C. van Illingworth, “Reduction of Plasma Lipids, Houwelingen, M. C. E. van DamMieras, B, M. Lipoproteins, and Apoproteins by Dietary 282. Sempos, C. T. and R. S. Cooper, “The Bas, and G. Hornstra, “Effect of a Moderate Fish Oils in Patients with Physician’s Health Study: Aspirin for the Primary Prevention of Myocardial Fish Intake on Haemostatic Parameters in Hypertriglyceridemia,” New England Journal Infarction,” New England Journal of Healty Males,” Thrombosis and Haemostasis, of Medicine, 312, 1210-1216, 1985. Medicine, 318, 924-925, 1988. 61, 468-473, 1989. 268. Rapp, J. H., W. E. Connor, D. S. Lin, 283. Shapiro, A. C., et al., “The Effect of 254. Mullertz, A., G. Holmer, and J. and J. M. Porter, “Dietary Eicosapentaenoic Fish Oil Supplementation on Plasma a- Grondahl-Hansen, “Increased Concentration Acid and Docosahexaenoic Acid From Fish Tocopherol, Retinol, Lipid and Lipoprotein of Plasminogen Activator Inhibitor Type-1 in Oil. Their Incorporation Into Advanced Levels in Normolipidemic Subjects,” Plasma After Intake of Fish Oil,” Fibrinolysis, Human Atherosclerotic Plaques,” Nutrition Research, 11, 539-548, 1991. 4, (suppl 2), 86-88, 1990. Arteriosclerosis and Thrombosis, 11, 903- 284. Shimamoto, T., et al., “Trends for 255. NHLBI, Consensus Development 911, 1991. Coronary Heart Disease and Stroke and Their Conference on Triglyceride, High Density 269. Riemersma, R. A. and C. A. Sargent, Risk Factors in Japan,” Circulation, 79, 503- Lipoprotein, and Coronary Heart Disease, “Dietary Fish Oil and Ischaemic 515, 1989. February 26 through 28, 1992. Arrhythmias,” Journal of Internal Medicine, 284a. Simons, L. A., J. B. Hickie, and S. 256. Nelson, G. J., P. C. Schmidt, and L. 225, (Suppl), 111-116, 1989. Balasubramaniam, “On the Effects of Dietary Corash, “The Effect of a Salmon Diet on 270. Rogers, K. A. and M. J. Karnovsky, N-3 Fatty Acids (MaxEPA) on Plasma Lipids Blood Clotting, Platelet Aggregation and “Dietary Fish Oil Enhances Monocyte and Lipoproteins in Patients with Fatty Acids in Normal Adult Men,” Lipids, Adhesion and Fatty Streak Formation in the Hyperlipidemia,” Atherosclerosis, 54, 75-88, 26, 87-96, 1991. Hypercholesterolemic Rat,” American 1985. 257. Nestel, P. J., “Fish Oil Attenuates the Journal of Pathology, 132, 382-388, 1988. Cholesterol-Induced Rise in Lipoprotein 271. Rogers, S., et al., “Effects of Fish Oil 285; Singer, P., S. Melzer, M. Goschel, and Cholesterol,” American Journal of Clinical Supplement on Serum Lipids, Blood S. Augustin, “Fish Oil Amplifies the Effect of Nutrition, 43, 752-757, 1986. Pressure, Bleeding Time, Haemostatic and Propranolol in Mild Essential Hypertension,” 258. Nikkila, M., “Influence of Fish Oil on Rheological Variables: A Double Blind Hypertension, 16, 682-691, 1990. Blood Lipids in Coronary Artery Disease,” Randomised Controlled Trial in Healthy 286. Sirtort C. R., et al., “Olive Oil, Corn European Journal of Clinical Nutrition, 45, Volunteers,” Atherosclerosis, 63, 137-143 Oil, and N-3 Fatty Acids Differently Affect 209-213, 1991. 1987. Lipids, Lipoproteins, Platelets, and Superoxide Formation in Type II 259. Nye, E. R., M. B. Ablett, M. C. 272. Rose, G., “Review of Primary Hypercholesterolemia,” American Journal of Robertson, C. D. J. Ilsley, and W. H. F. Prevention Trials,” American Heart Journal, Clinical Nutrition, 56, 113-122, 1992. Sutherland, “Effect of Eicosapentaenoic Acid 114, 1013-1023, 1987. 287. Spannagl, M., et al., “Plasmatic on Restenosis Rate, Clinical Course and 273. Saito, H., K. J. Chang, Y. Tamura, and Factors of Haemostasis Remain Essentially Blood Lipids in Patients after Percutaneous S.Yoshida, “Ingestion of Eicosapentaenoic Unchanged Except for PAI Activity During Transluminal Coronary Angioplasty,” Acid-Ethyl Ester Renders Rabbit LDL Less n-3 Fatty Acid Intake in Type I Diabetes Australian New Zealand Journal of Medicine, Susceptible To Cu2+-Catalyzed-Oxidative Mellitus,” Blood Coagulation and 20, 549-552, 1990. Modification,” Biochemical and Biophysical Fibrinolysis, 2, 259-265, 1991. 260. O’Connor, G. T., D. J. Malenka, E. M. Research Communications, 175, 61-67, 1991. 287a. Stampfer, M. J., P. M. Sacks, S. Olmstead, P. S. Johnson, and C. H. 274. Sams, G. E., et al., “Mechanisms Salvini, W. C. Willett, and C. H. Hennekens, Hennekins, “A Meta-Analysis of Randomized Responsible for Inhibition of Vein-Graft “A Prospective Study of Cholesterol, Trials of Fish Oil in Prevention of Restenosis Arteriosclerosis by Fish Oil,” Circulation, 80, Apolipoproteins, and the Risk of Myocardial Following Coronary Angioplasty,” American I-109-I-123, 1989. Infarction,” New England Journal of Journal of Preventive Medicine, 8, 186-192, 275. Sarris, G. E., et al., “Inhibition of Medicine, 325. 373-381, 1991. 1992. Accelerated Cardiac Allograft

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2714

288. Thaulow, E., J. Erikssen, L. Sandvik, Immunology and Immunopathology, 61, 161- Insulin-Dependent Diabetes Mellitus,” H. Stormorken, and P. F. Cohn, “Blood 176, 1991. American Journal of Clinical Nutrition, 56, Platelet Count and Function Are Related to 296. Vohwinkel, M., M. Hanisch, H. 447-454, 1992. Total and Cardiovascular Death in Wieland, and C. Luley, “Effect on Glucose 305. Zilversmit, D. B., “Atherogenesis: A Apparently Healthy Men,” Circulation, 84, Metabolism in Healthy Volunteers Upon Postprandial Phenomenon,” Circulation, 60, 613-617, 1991. Treatment With Omega-3-Fatty Acids,” 473-485, 1979. 289. Trials of Hypertension, “The Effects of Klinische Wochenschrift, 68, 124, 1990. List of Subjects in 21 CFR Part 101 Nonpharmacologic Interventions on Blood 297. Wander, R. C. and B. D. Patton, Pressure of Persons With High Normal “Comparison of Three Species of Fish Food labeling. Reporting and Levels,” Journal of the American Medical Consumed as Part of a Western Diet: Effects recordkeeping requirements. Association, 267, 1213-1257, 1992. on Platelet Fatty Acids and Function, Therefore, under the Federal Food, 290. Van den Berg, J. J. M., et al., Hemostasis, and Production of Drug, and Cosmetic Act and under “Increased N-3 Polyunsaturated Fatty Acid Thromboxane,” American Journal of Clinical authority delegated to the Commissioner Content of Red Blood Cells from Fish Oil-Fed Nutrition, 54, 326-333, 1991. of Food and Drugs, 21 CFR part 101 is Rabbits Increases In Vitro Lipid Peroxidation, 298. Weintraub, M. S., R. Zechner, A. amended as follows: but Decreases Hemolysis,” Free Radical Brown, S. Eisenberg, and J. L. Breslow, Biology and Medicine, 11, 393-399, 1991. “Dietary Polyunsaturated Fats of the W-6 PART 101—FOOD LABELING 291. Vandongen, R., T. A. Mori, J. P. and W-3 Series Reduce Postprandial 1. The authority citation for 21 CFR Codde, K. G. Stanton and J. R, L. Masarei, Lipoprotein Levels: Chronic and Acute “Hypercholesterolaemic Effect of Fish Oil in Effects of Fat Saturation on Postprandial part 101 continues to read as follows: Insulin-Dependent Diabetic Patients,” Lipoprotein Metabolism,” Journal of Clinical Authority: Secs. 4, 5, 6 of the Fair Medical Journal of Australia, 148, 141-143, Investigation, 82, 1884-1893, 1988. Packaging and Labeling Act (15 U.S.C. 1453, 1988. 299. Wing, L. M. H., et al., “Effect of Fish 1454, 1455); secs. 201, 301, 402, 403, 409, 292. Van Houwelingen, R., A. Nordoy, E. Oil Supplementation on Blood Pressure in 701 of the Federal Food, Drug, and Cosmetic van der Beek, U. Houstsmuller, M. de Metz, Treated Hypertensives,” Journal of Act (21 U.S.C. 321, 331, 342, 343, 348, 371). and G. Homstra, “Effect of a Moderate Fish Hypertension, 8, 339-343, 1990. 2. Section 101.71 is amended by Intake on Blood Pressure, Bleeding Time, 300. Wojenski, C. M., M. J. Silver, and J. adding new paragraph (f) to read as Hematology, and Clinical Chemistry in Walker, “Eicosapentaenoic Acid Ethyl Ester follows: Healthy Males,” American Journal of Clinical as an Antithrombotic Agent: Comparison to Nutrition, 46, 424-436, 1987. an Extract of Fish Oil,” Biochimica et § 101.71 Health claims: claims not 293. Van Houwelingen, A. C., A. D. M. Biophysica Ada, 1081, 33-38, 1991. authorized. Kester, D. Kromhout, and G. Hornstra, 301. Wolmarans, P., et al., “Plasma * * * * * “Comparison Between Habitual Intake of Lipoprotein Response to Substituting Fish for Polyunsaturated Fatty Acids and Their Red Meat in the Diet,” American Journal of (f) Omega-3 fatty acids and coronary Concentrations in Serum Lipid Fractions,” Clinical Nutrition, 53, 1171-1176, 1991. heart disease. European Journal of Clinical Nutrition, 43, 302. Yarnell, J. W. G., et al., “Fibrinogen, 11-20, 1989. Viscosity and White Blood Cell Count are Dated: October 30,1992. 294. Van Houwelingen, A. C., G. Homstra, Major Risk Factors for Ischemic Heart D. Kromhout, and C. D. L. Coulander, Disease,” Circulation, 83, 836-844, 1991. David A. Kessler, “Habitual Fish Consumption, Fatty Acids of 303. Yun, K. L., et al., “Dose Response of Commissioner of Food and Drugs. Serum Phospholipids and Platelet Function,” Fish Oil Versus Safflower Oil on Graft Louis W. Sullivan, Atherosclerosis, 75, 157-165, 1989. Arteriosclerosis in Rabbit Heterotopic Secretary of Health and Human Services. 295.Virella, G., et al., Cardiac Allografts,” Annals of Surgery, 214, “Immunosuppressive Effects of Fish Oil in 155-167, 1991. Normal Human Volunteers: Correlation with 304. Zambon, S., et al., “Effect of Glyburide Note: The following tables will not appear the In Vitro Effects of Eicosapentaenoic Acid and W3 Fatty Acid Dietary Supplements and in the annual Code of Federal Regulations. on Human Lymphocytes,” Clinical Lipid Metabolism in Patients with Non- BILLING CODE 4160-01-F

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2715

TABLE 1 Epidemiological Studies: Omega 3 Fatty Acids and CHD

Study Study Design Subjects Methods Results Comments

Beswick et al. Dietary Counseling compared 56 men, ages 36 to 71 Questionnaire to estimate Significant inverse correlation Study inconclusive on role of (Ref. 167) with survey data and platelet (mean=58). 13 exclueded for polyunsaturated to saturated between fat ratio and platelet EPA in platelet activity. Small activity 6 monts later. Men medication, 1 for inadequate fat ratio, percent calories from activity. Nonsignificant trend sample. In vitro assay not from Cardiff, Wales with blood sample. fat, and EPA intake. Platelet for decreased platelet repeated for accuracy. Nutrient history of myocardial 18 smokers, 34 patients on activity based on ADP added for aggregation with increased EPA intake estimated. Health infarction. cardiovascular medication, 19 aggregation of plateleta in intake. behaviors, medications, and on B blockers, 8 on blood and plasma. other factors which may have antihypertensive drugs, and 24 influenced results not on angina medication. presented. Study population ill.

Bulliyya et al. Dietary and serologic survey in 100 individuals in fish Dietary survey. Total serum Total serum cholesterol was Dietary survey methods not (Ref. 185) Nellore and Chittor districts of consuming populations and 109 and HDL—cholesterol, TG, and lower in the coastal village presented in paper. India. individuals in nonfish phospholipid measured. (152.7 mg/dL) than in the Dietary intake was presented as consuming populations. Clotting and bleeding times nonfish consuming village correlational data. The observed. Age, weight, pulse (214.9 mg/dL). HDL— potential confounding effect of rate, blood pressure tabulated. cholesterol was higher in the other components of diet on coastal village. cholesterol level warrants explanantion. Lower total cholesterol level in the fish eating population was observed.

Bjerregaard and Dyerberg, Analysis of CHD mortality All deaths due to CHD in A register of deaths by cause Age standardized mortality Genetic protection, exercise (Ref. 176). rates, Greenland. Greenland from 1968 to 1983. was developed using rates per 10,000 person-years and confounding factors such information from death for CHD in men in Greenland as tobacco use cannot be certificates, parish registers, settlements (5.3) were half eliminated as factors in this and civil registration records. those of men in Denmark observation. These data The register was computerized (10.0). There was an increasing support earlier observations of and brought under the rate of CHD from settlements lower CHD rates in male management of the Danish to towns in Greenland. The Greenlanders versus Danes, but Board of Health from 1975 to difference in rates of CHD in not for females. Note, the 50 1983. women were less apparent, percent difference is with lower rates in Denmark approximately the same as than in towns in Greenland. reported by Kromhout for comparing men eating no fish and those eating approximately 30 g/day.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2716

TABLE 1--CONTINUED

Study Study Design Subjects Methods Results Comments

Bonaa et al. Dietary and 156 subjects selected Food consumption Fish consumption was Divergent relationships (Ref. 178) serologic subjects from a survey questionnaire, with inversely correlated between EPA and HDL, and survey of population of 21,826 separate questions for with TG’s, but no DHA and HDL, may explain residents of for an intervention fatty and lean fish, significant discrepancies in the Tromso, Norway trial. also alcohol. Also, 2 correlations between literature regarding the unannounced 24-hour fish and cholesterol effect of various diet recalls. Analysis or apoprotein supplements on HDL of covariance and measures. EPA multiple linear regression. correlated with HDL until TG’s were controlled. EPA correlated with TG even after HDL was controlled. DHA inversely correlated with HDL and apoA1

Gerasimova et Dietary and Randomely selected men; HDL by Chukot residents had Correlation between al. (Ref. 205) serologic Moscow n= 650 ultracentrifugation. A ↓ cholesterol, TG, increased consumption of survey of Chukot n= 261. subsample was used for LDL; ↑ HDL, ↑ omega-3 fatty acids and residents of HDL phospholipids and consumption of omega- serologic measures is Moscow and the apoproteins. 24-hour 3 fatty acids, plasma consistent with the Chukot recall for dietary lipid EPA. hypothesis of a peninsula. data. relationship between omega-3 fatty acids and CHD, but many other dietary and behavioral factors could also be correlated and were not examined in this survey.

Innis et al. Dietary and Sample was selected as Phospolipid fatty acid Mean chain length and Observational data. (Ref. 225) serologic part of an unrelated analysis. unsaturation index of Supports a dietary origin survey. dietary survey. 185 the lipids in the two of phospholipids. Not Canadian Inuit and 24 populations was very directly relevant to CHD. Vancouver residents. similar. The Inuit had greater EPA and lower AA than the Vancouver population. NS cholesterol

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2717

TABLE 1—CONTINUED

Study Study Design Subjects Methods Results Comments

Popeski et al. Retrospective Hypertension study: Retrospective study: Communities with Ecologic data. Generates (Ref. 266) survey of 300 medical records Blood pressure higher marine oil hypothesis for a pregnancy from Inuit women in 7 measurements 6 hours consumption have relationship between the induced communities in the before delivery, significantly lower consumption of fish and hypertension in Northwest Territories. incidence of pregnancy diastolic pressure in blood pressure during Inuit women and induced hypertension. their women in the pregnancy. Prospective diet survey of Diet survey 27 Inuit Diet survey: Reported last 6 hours of study or clinical trial of women in these women in the 7 diet, hunter pregnancy. Pregnancy diet and pregnancy needed. communities communities above. interviews, cord serum associated phospholipid from 16 hypertension was 2.6 infants born in 6 times more common in months. communities with low fish consumption.

Seidelin et al. Correlation 40 consecutive Coronary artery disease No correlation Limited data are presented (Ref. 281) study of autopsies from was quantified by semi between extent of for a limited number of adipose tissue subjects age 52 to 90 automatic image stenosis and 16:10, subjects, i.e., no data fatty acids and years. analysis, Degree of 18:0, 18:2n-6, 16:1n- for other fatty acids of extent of stenosis was used to 9 or 18:1n-9, but a interest are presented, coronary artery divide subjects into significant inverse e.g., EPA, AA. stenosis. three groups for correlation with correlations. 22:6n-3. Stenosis Umbilical adipose correlated with lipids were extracted extent of body with methanol- weight. chloroform, trans- methylated esters by gas chromotagraphy.

Van Houwelingen Sample of Men in low (n-15) and Individuals in this There was Small same reflects et al. (Ref. clinical high (n=25) fish study where interviewed significantly higher inconsistency of fish 294) parameters from consumption groups. 4 times in 25 years serum phospholipid consumption over time. 40 healthy men Low consumption group using cross-check concentration of Dose of omega-3 in high selected from ate an average of 2 g dietary methods for omega-3 fatty acids fish consumption group cohort in fish/day while high habitual fish EPA and DHA acid and lower than in most longitudinal consumption group ate consumption. Of 79 men no significant clinical trials. study, Zutphen, an average of 33 g selected for the study, difference in Holland. fish/day. 40 completed it. Blood collagen-induced collected for fatty platelet aggregation, acid composition, PAI cutaneous bleeding activity, collagen- time, ATP-release in induced platelet whole blood or aggregation, ATP platelet number release, and TXB2. between the two groups.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2718

TABLE 1--CONTINUED

Study Study Design Subjects Methods Results Comments

Van Houwelingen Sample of 61 healthy elderly Cross-check dietary Dietary history The discrepancy between et al. (Ref. clinical male volunteers ages history was taken to seemed to correlate dietary history and serum 293) . parameters from 67 to 82. Lowest assess the habitual with serum linoleic level of fish-related 61 elderly male quartile of fish intake of acid. The fatty acid may be the volunteers from consumption ate 0 polyunsaturated fatty correlation between result of large variation longitudinal g/day. Highest acids. Blood was dietary history using in the level of these cohort. quartile ate 27 g/day. collected under the cross check fatty acids in the same Zutphen, controlled conditions method for fish foods. Holland. for measurement of consumption and serum phospholipids, TG’s, level of fish-related and cholesterol esters. fatty acids, however, seems less reliable.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2719

TABLE 2 Intervention Trials of Omega-3 Fatty Acids and CHD

Study Study Design Subjects Methods Results Comments

Agren et al. Randomized 62 Normal, Plasma lipids and Controls: NS TG, This study shows the (Ref. 168) parallel trial healthy female lipoproteins at cholesterol, apoA1, apoB. importance of the balance of of fish, fish students. baseline, 7 and 15 Fish eaters: NS versus the diet, particularly plus reduced fat weeks. baseline TG, cholesterol, regarding saturated fat, in (especially apoB; ↓ apoA1; ↓ TG versus determining the blood lipid saturated fat) controls. resoponse to omega-3 fatty or usual diet Fish plus low fat: ↓ TG, acids. Changes were slight (one fish meal cholesterol, apoA1, apoB. unless a low fat low per 2 weeks) for saturated fat diet was used. 15 weeks. Study would have been stronger with a control group eating a low fat low satrurated fat intake to the nonfish control.

Annuzzi et al. Randomized Eight female One of the patients was ↓ TG, VLDL; ↑ LDL; NS Duration period was very (Ref. 170). double-blind NIDDM, without treated by diet only; cholesterol, HDL FFA; NS short and no fatty acid crossover trial liver, kidney or other on Glibenclamide LDL-TG, HDL-TG; NS fasting analysis on neither fish oil of 10 g fish any other disease (4) and metformin (3). glucose, average glucose nor olive oil diet. There oil/ day (3.0 g known to The patients were under insulin response, was no washout period between EPA plus DHA, influence lipid dietary control and they sensitivity. cross-over of the study. MaxEPA) versus and/or were in the metabolic olive, no wash carbohydrate ward. out, 2 weeks metabolism. each treatment.

Bairati et al. Randomized, 119 subjects Angiographic assessment, By three of four definitions Comparable compliance. Olive (Ref. 172) double-blind undergoing first, quantified by computer. of restenosis, and oil control doesn’t control placebo- successful, diet by food frequency. multivariate analysis to for polyunsaturated fatty controlled trial computer Restenosis defined in control for exclusions, fish acids. The comparable of 15 g fish quantified four was for analysis. oil reduced restenosis. NS magnitude of the effects of oil/day (MaxEPA, percutaneous ECG evidence of myocardinal dietary omega-3 fatty acids 4.5 g EPA plus transluminal ischema, but trend toward ↓ and fish oil supplementation DHA) versus coronary fish oil, also on exercise suggest long-term, low dose olive oil from 3 angioplasty, testing. Significant effects are as strong as weeks pre evaluated also at difference also according to short-term, larger amounts. angioplasty to 6 6 months. dietary omega-3 fatty acids No associations of restenosis months; (highest, middle versus with other dietary variables, total concurrent lowest tertile) although fat, classes of fat, cholesterol, or aspirin. highest tertile intake was dietary only 0.15 g/day, and dietary seafood. Differences between odds ratio comparable to groups use of blood pressure fish oil odds ratio. medications (see Bairati et al. Ref. 173)) was not discussed.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2720

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Bairati et al. Randomized 125 patients Recumbent blood Blood pressure increased in Multiple linear regression (Ref. 173). double-blind, undergoing first pressure, blood lipids the olive oil control group, analysis used to control use placebo- percutaneous by commercial kits. and was unchanged in the of blood pressure medications controlled trial transluminal fish oil group, possibly reduced differences in LDL with 15 g/day coronary because greater number of and HDL to borderline MaxEPA (4.5 g angioplasty, patients in the olive oil significance (p= 0.06), and EPA plus DHA) evaluated also at group discontinued inclusion of TG resulted in versus olive 6 months. concurrent blood pressure NS change in HDL. oil. medications. Fish oil ↓ TG’s, NS in cholesterol, LDL or HDL versus control, but the change from baseline was different between groups with ↑ LDL and HDL in fish oil.

Beil et al. Randomized 30 Patients with Lipids by standard ↓ TG in high fish oil group Although randomized, there were (Ref. 174) double-blind primary hyper- methods, Lp(a) by versus placebo; NS large differences in trial with 10.5 triglyceridemia. commercial kit. cholesterol, LDL, HDL apoB; NS initial Lp(a) levels, with g/day fish oil Patients off Lp(a), post-hoc analysis shows ↓ only 1 of 10 in the placebo group (3.15 g EPA plus lipid lowering Lp(a) in those over 6 mg/dL versus 6 and 8 of DHA, MaxEPA) drugs for 6 initially greater than 10 10 in the low and versus 5.25 weeks, no beta mg/dL. high fish oil groups, respectively. g/day fish oil blockers, Therefore, on plus 5.25 g diuretics or the basis of Lp(a) the oleic acid (low hormones. Fat randomization was inadequate. fish oil) versus restricted diet Oleic acid does not control 10.5 g oleic (30 percent, for polyunsaturated fatty acid (placebo) 6 polyunsaturated acids. weeks. fat: saturated fat ratio 1.5:1 cholesterol <250 mg/day).

Bhathena et al. Nonblinded, longitudinal design, 40 Healthy females, Diety for the placebo and test Fish oil ↑ fasting glucose; ↓ TG, Unusual source and high level of (Ref. 175) 10 week run-in on placebo (15 g no history of group was controlled, fish was insulin, glucagon, GH, omega-3 fatty acids. Study mixed fat); 10 weeks with 15 g metabolic eliminated from the menu. Forty somatomedin-C; NS cholesterol, design doesn’t allow conclusions fish oil (anchovy oil, 6.5 g EPA disease, no percent of energy comes from cortisol, dihydroepiandrosterone about omega-3 fatty acid-specific plus DHA); 8 weeks of 15 fish oil medication, no dietary fat which was more than sulphate. effects. plus 200 mg vitamin E smoking. the level of current dietary Fish oil plus vitmain E gave no guideline recommended. further change in glucose, TG, glucagon cortisol or cholesterol, but ↓ GH, insulin, increased somatomedin-C to placebo levels, and produced a ↓ in DHEA-S.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2721

TABLE 2—CONTINUED

Study Study Design Subjects Methods Results Comments

Bonaa et al. Randomized, 146 healthy Fasting blood samples at Fish oil ↓ TG; In both No change in dietary fat, (Ref. 1?8). double-blind subjects. beginning and end of groups NS cholesterol, LDL, alcohol or protein. Both placebo- intervention, standard, groups had a small, apoB, ↑ HDL; ↑ apoA1 in controlled trial commercial assays for corn oil group. In fish oil significant weight gain. The of 6 g/day Norek lipids and apoproteins. divergent results underscore group ↑ phospholipid EPA Hydro (4.5 g Multiple linear the need for studies on correlated with HDL, but ethyl esters of regression. individual omega-3 fatty not in corn oil group, EPA plus DHA) acids, that may help explain whereas in corn oil group versus corn oil, inconsistencies in results of DHA was inversely 6 month fish oil effects on HDL. correlated to HDL. observational run in, 10-week intervention.

Bordet et al. Randomized, 32 healthy Bleeding times by NS platelet aggregation Differences from other (Ref. 179). dose-response females. Simplate II. sensitivity to ADP, studies may be due to absence study to 300, collagen, PAF; ↑ of DHA, and need for longer 900, 2700 mg EPA sensitivity to ate studies to allow DHA (ethyl ester, NS bleeding time. incorporation. MND-21, Mochida, Japan) 4 weeks plus 4-week washout.

Brown and Randomized 3 X 3 12 healthy Clotting times on 2 ↓ leukocytes on fish, fish The authors review other Roberts (Ref. crossover of females. samples taken 4 days plus fish oil diets. reports on fibrinolytic 183). fish (0.6 g EPA apart at the end of each Platelet count ↓ on fish activity and note the plus DHA/day) diet treatment. inconsistency in findings. oil; NS fibrinogen, ↑ versus fish plus Fibrinolytic activity in fibrinolytic activity on fish oil (2.0 g pooled samples from each both fish and fish plus EPA plus individual assayed by fish oil. DHA/day) versus euglobulin activity control (fish versus fibrin. free) 6 weeks each with 6-week washout.

Brown and Crossover trial 11 healthy Thiobarbituric acid NS conjugated dienes, Small number of subjects; Wahle, (Ref. of 15 g fish oil females. reactive substances, creatine kinase, or TBARS considerable variance in many 184). (MaxEPA, 4.5 total tocopherol by in the platelet poor measures with the exception g/day EPA plus fluorometric assay, whole plasma; ↑ total plasma of plasma TBARS. The DHA) with or blood aggregation by TBARS on both, but less interaction between glucose without 400 IU electrival impedance with added vitamin E; ↑ suggests a mechanism to vitamin E for 4 after collagen stimulus. glucose on fish oil without address potential adverse weeks each with vitamin E. effect in diabetics. 2-week washout between.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2722

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Clarke et al. Noncontrolled 7 male and 4 Usual low cholesterol, NS TG’s cholesterol, LDL, Numbers of observation at (Ref. 189). supplementation, female low saturated fat diets, HDL. Increased nose bleeds each time per subjects not 1 g fish oil/day adolescents with 2 received colstipol. during fish oil treatment given for blood lipids (0.3 g EPA plus FHL type II (5 versus before and after. measurements, so fish oil DHA, MaxEPA) type Iia, 6 type amounts for “after” treatment increasing in 1 Iib). not clear. Platelet count and g increments other biochemestries monthly to 5 described as normal, but no g/day for months data of effects if fish oil 5 and 6. are described.

Cobiac et al. Randomized 50 Elderly, Blood pressure in sitting ↓ sys, dias blood pressure Note that sodium restriction (Ref. 190). double-blind healthy position. on Na restricted diet plus alone had no effect. 55 placebo- subjects. fish oil; NS on fish oil elderly subjects started the controlled, 2 only; NS on sunflower oil expiriment but only 50 week run-in on (the run-in treatment). completed the study. No restricted Na explanation was given why intake plus 8 to some subjects were dropped 1 g sunflower out. Dietary intake and oil capsules and compliance were not 8 to 600 mg slow controlled. No fatty acid release NaCl. analysis of the diet and/or Then 4 weeks on the control and test oils. No fish oil (8 g washout period in between HIMEGA, 4.2 g cross-over. EPA plus DHA) on either lo or normal Na, and crossover to alternate Na for 4 weeks.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2723

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Davidson et al. Three 1. 16 Type II-B 1. Stable AHA phase 1 1. ↓ TG, cholesterol on Olive oil control increased (Ref. 191). experiments: 1. hyperlipidemic diet for 3 months prior 7.2, 5.6 g EPA plus TG cholesterol versus run-in Dose response patients. to and during study. DHA/day, NS cholesterol on diet. Design doesn’t control study with fish 3.6 g EPA plus DHA/day; NS for polyunsaturated fatty oil, (7.2, 5.4, 2. 12 Hyper 3. Stable on AHA phse I HDL; cholesterol ↑ on olive acids. 3.6 g EPA plus triglyceridemic or Phase II dietas. oil. DHA/day, type IV. SuperEPA) for 6 Larger decrease in 2. SuperEPA ↓ cholesterol weeks 3. 148 Patients cholesterol by SuperEPA may ↑ sequentially, 6- of different more than MaxEPA; MaxEPA be due to its reduced content week washout hyperlipidemias. LDL, HDL, HDL, versus of saturated fatty acids. between doses; SuperEPA; NS HDL2. highest dose crossover to 3. Each MaxEPA, SuperEPA, olive oil Promega ↓ TG’s, Capsule counts were used to 2. Crossover of cholesterol; ↑ HDL in Type assess compliance. MasEPA at 4.8 g Iib on SuperEPA only; ↓ LDL EPA plus DHA/day in familial each , 6 weeks, hypercholesterolemia after 6-week washout SuperEPA, NS Lp(a). 3. Uncontrolled supplementation of cases from a 4 year period, 148 subjects treated for 6 week periods with various fish oil.

Eritsland et al. Randomized to 22 female 4 week run-in. Beta ↓ TG’s by fish oil; ↓ 2 minor bleeding episodes on (Ref. 195). aspirin (300 patients with blockers used by 9 cholesterol by fish oil aspirin, none on fish oil. mg/g) for 1 stable CHD. patients, nitrates plus aspirin; ↑ HDL in fish No was out between week, followed allowed. Usual diet. oil only. treatments, small number of by 4 week on subjects makes differences in fish oil (Norak, response to fish oil only and 85 percent ethyl fish oil plus aspirin esters of EPA questionable. plus DHA, 3.4 g/day) or 4 week on fish oil followed by 1 week on aspirin.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2724

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Fahrer et al. Self selected to 21 female, 21 No run-in, fish was pink ↓ TG’s in both fish and fish Baseline blood lipid values (Ref. 196). treatment of male healthy salmon, tuna, herring, oil groups; NS cholesterol, were comparable in the self fish (1.5 to 2.0 volunteers. mackerel, pilchard. HDL; decrease in TG selected groups. g EPA plus Fish consumption recorded correlated with increase in DHA/day), fish in daily diary. EPA. oil (Sanomega s18, 3.1 g EPA plus DHA/day), normal diet for 2 months.

Ferretti et al. Nonblinded, 40 nonsmoking PGE-M by a stable isotope Fish oil alone and fish oil PGE-M is the sum of PGE1 plus (Ref. 197). longitudinal females. dilution method developed supplemented with vitamin E PGE, derived from AA. The design, 10 week in the authors’ produced comparable results EPA-derived PGE3 was not run-in on laboratories. on average, with the mean measured. Dietary intake was placebo (15 g values µg PGE-M excreted per well controlled. Prolonged mixed fat); 10 24 hours in control, fish oil use of fish oil weeks with 15 g and fish oil plus vitamin E supplementation was not fish oil of 15.41 ± 2.12, 12.51 ± 1.78 recommended. The range of (anchovy oil, baseline values was from 3.8 and 12.77 ± 1.85, 6.5 g EPA plus respectively. Paired t-test to 60.9 µg/24 hours. There DHA); 8 weeks of was no apparent relationship showed a significant (p < 15 fish oil plus between initial values and 0.002) reduction between 200 mg vitamin magnitude of the change, and baseline and fish oil E. there were many individuals treatment. who had substantial differences for the two fish oil treatments. Thus, the significance of an average change of about 20 percent is not clear.

Force et al 12 g fish oil 14 females, 2 Gas chromatography-mass Fish oil ↓ TXA, 38 percent, This is a study on the (Ref. 200). (n=8)(6 g EPA males, spectroscopy with ASA ↓ 97 percent at each mechanism of action of fish plus DHA, clinically oil. No concurrent placebo dose; fish oil and ASA each ↓ Promege) or 16 g stable but control. Dietary intake was PGI, (ASA more than fish fish oil (n=6) advanced CHD. not controlled. ↑ 3 (8 g EPA plus oil); fish oil PGI , but DEA), 6 weeks. ASA did not increase PGI3. After 6 weeks on fish oil only, concurrent increasing daily dosages of ASA (50, 100, 225, 1,300 mg) 2 weeks each.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2725

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Franceschini et Uncontrolled 5 healthy subjects. HDL subfractions assayed by NS cholesterol, HDL, ↑ HDL2, ↓ Dietary intake was not al. (Ref. 202). supplementation nondenaturing HDL3, ↑ HDL2/HDL3 mass ratio. controlled. Compliance was study of 6 g/day polyacrylamide gel checked by plasma PL fatty acid fish oil (Norsk electrophoresis. composition. Small study. hydro, 2.8 g EPA plus 1.7 g DHA), 6 weeks.

Gans et al. (Ref. Randomized double- Stable claudication Supine blood pressure at ↓ Diastolic by both groups, Wide variations of age (18 to 203). blind study (fish patients; 37 rest and 1, 6, 10 min post ↓ systolic only in CO group; ↑ 80 years), Dietary intake was oil, source not enrolled, 16 per exercise. RBC deformability; not controlled. Compliance was specified, 1.8 g group completed. Fibrinogen by commercial NS cholesterol, LDL; checked by plasma PL fatty acid EPA plus 1.2 g kit. composition. fibrinogen; ↑ HDL, HDL2; ↓ TG’s; DHA/day) versus Blood pressure was lowered in NS pain, walking distance. corn oil, 4 both groups. months.

Gazso et al. Randomized double- 17 normal healthy 6 ic conversion, MDA, ↓ Platelet aggregation by Groups differ in the endpoints (Ref. 104). blind placebo- males, 11 females 6- platelet aggregation to Efamol-marine; ↓ MDA in all at beginning of the expiriment, controlled Efamol- Efamol marine, 5- ADP. groups. so it is difficult to interpret marine (1.2 g EPA) Efamol, 6-olive oil. changes. The MDA ↓ may be due versus Efamol (v to the vitamin E added. olive oil) 12 capsules/day 6 weeks.

Gerhard et al. 3-period crossover 21 normo- ApoB standardized to Salmon, sablefish diets ↑ Diets were comparable for total (Ref. 206). of three fish triglyceridemic Centers for Disease Control cholesterol, apoB, LDL. fat, saturated fat. Study diets; Dover sole females. standards. VLDL and LDL Sole ↓ HDL, ↑ LDL; Sablefish ↓ design doesn’t allow (2 g EPA plus precipitated with magnesium HDL. conclusions about omega-3 fatty DHA), Salmon (4 g phosphotungstate, HDL acid-specific effects. EPA plus DHA), or enzymatically in the sablefish (3.4 g supernatant. EPA plus DHA), 18 d each with 3-week washout between.

Goren et al. Uncontrolled 16 Patients with end Blood lipids before and ↓ TG; NS cholesterol, ↓ Fish oil dosage was adjusted to (Ref. 207). supplementation stage renal disease, after supplementation. cholesterol/HDL in a subset of the body weight of chronic study, 150 to 200 6 type Iib, 1 type cholesterol by excessively hyperlipidemic renal failure young patients mg/kg (EPAGIS) Iia, 8 type IV. microenzymatic method. subjects; NS blood pressure, (7 to 18 years). (0.86 to 2.3 g EPA Apoproteins by turbidity platelet counts, apoA1:apoB. plus DHA/day) 8 assays. weeks.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2726

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Green et al. Randomized 27 Hyperlipidemic Cholesterol, HDL, TG by Overall NS cholesterol, LDL. Substantial amount of other (Ref. 209). double-blind subjects, 15 type commercial kits. omega-3 fatty acids in this placebo- Iib, 12 type IV. Apoprotein by Type IIb: ↓ TG, supplement (0.50 g 18:3, 0.45 controlled immunoturbidity assay. cholesterol, LDL, HDL. g 18:4 and 0.42 g 22:5) per crossover 15 g day. Patients were on weight- fish oil Type IV: ↓ TG; NS maintenance diet but no cal (EPAGIS, 4.3 g cholesterol, LDL, HDL. percent or wt percent of each EPA plus DHA) component was given. versus Both: NS platelet count, Compliance was shown on blood corn:olive oil platelet aggregation, RBC lipid analysis. mix, 8 weeks deformability, apoB, apoA,; ↓ each treatment blood viscosity. with 4 week wash out.

Haglund et Double-blind 12 Normal healthy Lipids, glucose, Lp(a), Both fish oil ↓ TG’s; ↑ HDL, High vitamin E produced some al. (Ref. crossover study subjects. fibrinogen. glucose; NS cholesterol, effects independently of the 210). 30 mL/day of low Lp(a), aopB, insulin. fish oil, underscoring the vitamin E fish need the control for oil (9.6 g EPA Low vitamin E fish oil ↑ MDA, polyunsaturated fatty acids, plus DHA/day, and have adequate vitamin E fructosamine; ↓ vitamin E. Eskimo-3 or in the test substances.

Inuit-3) or same ↓ oil supplemented High vitamin E fish oil with 1.5 IU TG’s, fibrinogen. vitamin E, 3 week.

Hamazaki et Uncontrolled 16 Diabetics, 5 Albumin by NS TG, cholesterol, glucose, NS body weight. Compliance al. (Ref. supplementation IDDM, 11 NIDDM. radioimmunoassay with a HbA10, systolic diabolistic was checked by blood EPA 215). study, 1.8 g EPA commercial kit. HbA10 by blood pressure, blood level.Dietary intake was not ethyl ester/day, high performance liquid viscosity; ↓ albumin controlled. In IDDM fasting (laboratory chromatography. excretion, hematocrit. glucose was reduced, and purified), 6 barely missed statisticaly months. significance. Study design doesn’t allow conclusions about omega-3 fatty acid- specific effects.

Hansen et al. Crossover trial 20 healthy female Whole blood clot NS fibrinogen, fibrinolytic Study design doesn’t allow (Ref. 216). of 25 mL cod and 20 healthy male (produced by added activity, t-Pa; ↓ TXB2. conclusions about omega-3 liver oil/day subjects. thrombin) lysis time to fatty acid-specific effects. (6.25 g EPA plus complete lysis. t-PA by DHA), 8 weeks commmercial ELISA kit. versus no Fibrinogen by supplement. spectrophotometric assay.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2727

TABLE 2-CONTINUED

Study Study Design Subjects Methods Results Comments

Harats et al. Parallel a. Smokers 6 40 hour smoking ↓ TG; NS cholesterol, HDL; TBARS in plasma, and LDL, more (Ref. 217). untreated MaxEPA, 5 abstinence and overnight fish oil ↑ TBARS in plasma responsive to cigarettes than controls versus controls. fast prior to blood draw; fish oils. Most, but not all presmoking ↑ TBARS after 10 g MasEPA/day 90 minutes 4 to 6 of the increase due to fish smoking. (3 f EPA plus b. Smokers 3 cigarettes smoked, for oil alone could be blocked by DHA), 4 weeks. control, 4 fish second blood draw. the added vitamin E. oil only, 4 fish oil plus 400 mg smokers.

c. Nonsmokers.

Harris et al. Uncontrolled 8 healthy males. Bleeding times by Bleeding NS on fish oil; Medications were controlled (Ref. 218). supplementation Simplate II. fish oil plus aspirin same but diet was not controlled. trial of fish Platelet aggregation to as aspirin only; fish oil Short-term study with a small oils plus AA, collagen, PAF and AA and fish oil plus aspirin number of subjects may explain aspirin; 325, 80, in combination with the NS on platelet sensitivity inconsistencies with other 80 mg aspirin for other agonists. to AA, collagen, PAF, but comparable studies. 3 days; 4 day fish oil ↓ extent of The study may not have wash out; 2 weeks aggregation to collagen. adequate statistical power to on 4.5 g EPA plus determine whether bleeding DHA (SuperEPA); time increases of aspirin and 325, 80 80 mg fish oil are additive or aspirin plus greater than additive. SuperEPA for 3 days.

Harris and Uncontrolled 12 male and 4 Bleeding times by ↓ TG’s VLDL; NS No medications. No difference Windsor (Ref. supplementation female healthy Simplate. Background cholesterol, LDL, apoB, between capsules and emulsion 220). study on normolipidemic diets had 32 to 36 apoA1, HDL3, vitamin E, in test meal, possibly because postprandial subjects. percent fat with 12 to 14 Lp(a); ↑ HDL, HDL2. most fat was from other diet lipemia with of percent as saturated, 12 components. fish oil (2.2 g to 13 percent as ↑ Bleeding time; NS RBC EPA plus DHA, monounsaturated, and 6 deformability. Dale Alexander percent as

Omega-3), random polyunsaturated. Test Postprandial lipemia assignment or meal provided 1 g fat/kg reduced about 40 percent. emulsion for 4 (61 percent of total weeks. calories: 32 percent of calories from saturated fatty acids; 13 percent from monounsaturated fatty acids, 7 percent from omega-3 fatty acids. Two hour blood samples through 10 hours post meal.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2728

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Harris et al. Randomized dose- 28 Hyperlipidemic Blood lipids, Simplate II ↓ TG in dose-related manner 1 Discrepancies among studies, (Ref. 219). response, 1.25 to patients. for bleeding times. month and 6 months, except lowest methodologies were discussed. 5 g EPA plus dose NS at 6 months, ↓ VLDL on Irregularities may be due in part DHA/day (Promega) all but lowest dose; NS to small sample size. 6 months. cholesterol, LDL, HDL, HDL2, 4-week washout returned most values to pretreatment levels. except 2.5 g/day at 6 months ↑ LDL, HDL). ↑ Bleeding times on 2.5, 5 g/g RBC deformability largely unaffected.

Hendra et al. Randomized, 80 Noninsulin- Fibrinogen by Clauss, HDL Transient ↑ glucose; NS HbA1. Large, carefully controlled study (Ref. 222). double-blind, dependent diabetic by precipitation, LDL by in an at-risk population. Olive placebo- subjects. calculation. ↓ TG; NS cholesterol, HDL; ↑ LDL oil control doesn’t allow controlled trial (versus baseline). conclusions about omega-3 fatty of 10 g/day MaxEPA acid- specific effects. (3 g EPA plus DHA) ↓ Spontaneous platelet versus olive oil, aggregation, but NS response to 6 weeks. induced aggregation. ↓ blood pressure in both treatments.

Homma et al. Uncontrolled 15 outpatients. Ad libitum diets. Blood ↓ cholesterol, TG, apoB, small Authors state the relative (Ref. 224). supplementation samples after 12 hour dense LDL; ↑ large light LDL, atherogenicity of large light LDL test of 2.7 g/day fast. Plasma lipids by lipid transfer protein activity; and small dense LDL is purified EPA ethyl ultracentrifugation every NS HDL, HDL, HDL3, apoA1, apoC, controversial. ester (source not 4 weeks. apoE. specified), 12 weeks.

Jensen et al. Sequential dose- 14 healthy males and 4 1 month run-in on fish ↓ TG; VLDL on 6 g dose; ↑ HDL and Changes in baseline HDL not (Ref. 226). response with 1, healthy females. free diet, otherwise diet HDL/LDL ratio on 6 g dose, but shown. 3, 6 g EPA plus not controlled. Bleeding baseline HDL changed; NS DHA (Shaklee EPA), times by Simplate II. cholesterol, LDL. 4 weeks each with Lipids by auto laboratory 3-week washout method. between.

Kremer et al. Randomized double- 49 with rheumatoid IL-1 by bioassay. ↓ IL-1 38 percent 41 percen and Actual doses were adjusted per kg (Ref. 228). blind placebo arthritis completed 55 percent in olive oil, low and body weight. (olive oil) this study. high fish oil groups, controlled trial respectively; NS IL-2 in both of 3.25 or 6.5 g fish oil groups. EPA plus DHA/day (ethyl esters, Pharmacaps), 24 weeks.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2729

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Li and Steiner Dose-response, 5 Normal healthy Platelet adhesion ↓ Platelet adhesion to This procedure reduces (Ref. 234). parallel design: subjects each measured ex vivo in collagen I and fibrinogen, formation of thrombi, dilutes 4.8, 9.6, or dose. laminal flow chamber, near maximal response at 3 g platelet derived factors. 14.4 g EPA plus using purified EPA/day; speen of return to Measures direct interaction DHA/day, (source substrates. baseline values in the of platelets with surface not specified), washout was directly related matrix. Study design doesn’t 3 weeks. to dose. allow conslusions about omega-3 fatty acid-specific effects.

Lindgren et al. Metabolic ward 9 normolipidemic Plasma proteins NS cholesterol, LDL, HDL2a, Carefully designed metabolic (Ref. 235). crossover of females. measured by competitive HDL3a, apoB, apoE, Lp(a); ↓ ward study, using practical salmon versus ELISA except apoA, by TG, HDL3a, HDL3b, apoA1, apoA2; level of omega-3 fatty acids, prudent diet (30 radioimmunodiffusion. ↑ HDL, HDL2b. and Fat (saturated fatty percent fat). 20 acids and omega-6 day run-in, 40 polyunsaturated fatty acids) days each diet. carefully controlled. Salmon diet Details on effects of provided 2.1 lipoprotein subfractions. Two percent of assay methods for Lp(a) gave calories as same result. omega-3 fatty acids, (approximately 5 g/day EPA plus DHA).

Mallo et al. Uncontrolled Normolipemic 4 week run-in. ↓ TG, platelet aggregation; Comparable lipid and platelet (Ref. 241). supplementation subjects with very normal diets. NS cholesterol, LDL, HDL, responses for the low Lp(a) with fish oil, high (?) and very Platelet aggregation to Lp(a). and high Lp(a) groups. (6.7 g EPA plus low i.e., collagen, thrombin. DHA/day, EPAX- undetectable Lp(a) 5000), 6 weeks levels (7). and 4-week washout.

Marckmann et al. observational, 12 healthy Clause fibrinogen NS cholesterol, HDL, TG, Since both diets produced (Ref. 244). sequential diets females. assay, t-PA and PAI-1 fibrinogen; TG ↓ on both changes with respect to the of fish (3.4 antigens by ELISA. diets; ↓ PAI-1 and t-PA initial diets (that were g/day EPA plus uncontrolled) it is difficult antigen , PAI activity and ↑ DHA, 10 d) to ascribe any change to the t-PA activity on meat, but NS uncontrolled (18 omega-3 fatty acids. on fish. d), and meat (10 However, the changes on the d). meat diet are more in line with reduced CHD risk than the lack of change on the fish diet.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2730

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Meland et al. Double-blind, 40 females mild Calibrated instruments at NS blood pressure, cholesterol; Power to detect a 5 mm blood (Ref. 247). randomized multi- hypertension. 8 centers. Time of day for ↓ TG;s on fish oil; ↓ pressure difference was 96 center placebo measurements was cholesterol/HDL ratio in both percent; a 10 percent (corn oil- olive controlled. groups. cholesterol difference was 61 oil mix) percent. Cholesterol/HDL ratio controlled trial, decrease in placebo was nearly 20 mL MaxEPA/day more than that after fish oil (6.8 g EPA plus (p<0.07). 11 of 14 subjects on DHA), 6 weeks. fish oil guessed their assignment correctly.

Meydani et al Uncontrolled 25 males. Blood at 1, 2, 3 months. ↓ TG’s; ↑ lipid peroxides. 6 IU vitamin E may not be (Ref. 248). supplementation adequate. study 2.4 g/day EPA plus DHA (Promega), 3 months.

Molvig et al. Randomized, 25 Healthy subjects Isolated monocyte cell ↓ IL-1B immunoreactivity on Placebo had 20 percent (Ref. 250). double-blind and 8 IDDM subjects. cultures. TNF and IL-1 by high dose only; NS after low polyunsaturated, 38 percent Placebo- commercial ELISA kits. dose. NS TNF-α; ↓ proliferative monounsaturated fatty acidsl controlled trial response. Spontaneous and LPS- stimulated of 1.6, 3.2 g EPA leucotriene B, and PGE, plus DHA secretion did not differ among (Pikasol) versus groups at baseline or after 7 fatty acid blend, weeks of treatment. IL-1 7 weeks. returned to baseline with 3- week washout.

Mori et al. Matched (age, 32 females with No aspirin for 14 days ↑ choleseterol, LDL, HDL2; ↓ TG Compliance by capsule count. (Ref. 251). weight) and peripheral vascular prestudy platelet by fish oil, (but olive oil ↓ Changes in control make randomized to 15 disease. aggregation to PAF, cholesterol, LDL); interpretation difficult. Olive g MaxEPA/day (4.5 collagen. oil does not control for ↓ platelet aggregation by fish g EPA plus DHA) polyunsaturated fatty acids. ↑ or olive oil, 4 oil, but olive oil weeks. aggregation.

Mori et al. Matched groups 27 normolipidemic HDL by heparin, manganese NS cholesterol, LDL, HDL; ↑ Study design doesn’t allow (Ref. 252). randomly assigned insulin-dependent chloride precipitation, HDL2, ↓ HDL3, TG. conslusions about omega-3 fatty to 15 g/day female diabetics. followed by separate acid-specific effects. MaxEPA (4.5 g EPA precipitation of plus DHA), olive subfractions. LDL by oil, or olive oil calculation. plus cholesterol.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2731

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Mueller et al. Randomized 12 Healthy Bleeding times by Fish oil-NS bleeding time Trend toward ↑ HDL versus (Ref. 253). double-blind adults. Simplate II before and versus baseline but ↑ versus baseline, but olive oil in crossover trial after administration of olive oil both before (p < same direction, some order of 8 g EPA plus 325 mg aspirin. Excludes 0.02) and after (NS) effects confound results. DHA (Promega) subjects with platelet or aspirin. versus olive coagulation disorders, ↓ TG on fish oil, platelet oil, 21 d. thrombocytopenia, count, WBC count; NS ethanol. cholesterol, LDL, HDL.

Muller et al. Multicenter 84 healthy Published methods for NS fibrinogen, other blood Compliance by lithium (Ref. 253a). observational females. facror X, antiplasmin, coagulation measures (only ↑ excretion. Study design trial of 135 g plasminogen. Fibrinogen factor X), or fibrinolysis doesn’t allow conclusions canned mackerel by Clause. measures; meat ↓ plasminogen. about omega-3 fatty acid- paste (4.7 g/day specific effects. EPA plus DHA) or meat paste.

Mullertz et al. Uncontrolled 7 Healthy Normal diets, PAI-1 and ↓ α-Tocopherol; NS Suggests that differences (Ref. 254). supplementation, adults. u-PA by ELISA kits. cholesterol, TG’s; ↑ PAI-1; reported for PAI-1 are due to 0.55 g EPA plus NS t-PA, u-PA. the assay used, with the DHA/day double antibody assay used in (Pikasol), 21 this study, and the days. monocional antibody used by Emeis et al. providing specificity. Concludes that fish oil diecreases fibrinolytic activity.

Nelson et al. Metabolic ward 9 normolipidemic Platelet aggregation to NS bleeding time; salmon diet Carefully designed metabolic (Ref. 256). crossover of females. ADF, AA, collagen, ↓ platelet counts NS platelet ward study, using practical salmon, prudent thrombin: threshold and response to collagen, level of omega-3 fatty acids, diet (30 percent maximum response. thrombin but ↓ sensitivity to and Fat (saturated fatty fat). 20 day Bleeding time by ADP. acids and omega-6 run-in, 40 days Simplate II. polyunsaturated fatty acids) each diet. carefully controlled. Salmon diet gave 2.1 percent of calories as omega-3 fatty acids, (approximately 5 g/day EPA plus DHA).

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2732

TABLE 2—CONTINUED

Study Study Design Subjects Methods Results Comments

Nikklila (Ref. Randomized, 32 females with Lipid lowering diet for NS cholesterol, LDL, HDL, HDL inversely related to TG 258). double-blind, CHD, increased TG 4 weeks prior to study. HDL/cholesterol ratio, apoA1, in study group pretreatment. placebo (corn and decreased HDL, apoB; ↓ TG’s; During open oil) controlled, 62 percent were phase those with severe crossover, 2.4 g overweight. hypertriglyceridemia had ↑ EPA plus DHA/day HDL/cholesterol. as ethyl ester (EPA X 6000EE, Almarin), two 4- week periods with a 4-week washout between, followed by open study of 3.6 g EPA plus DHA/day for 4 weeks.

Nye et al. Randomized, fish 79 females, 29 Angiography (blind) at NS angina (trend toward less No deaths in any group (Ref. 259). oil and its males post PTCA one year or before in in A/D and fish oil groups D through 1 year, 93 percent placebo were referred for those with anginal restenosis by EPA (11 percent follow-up rate. Results double-blind angina, none had symptoms; restenosis versus 30 percent for olive suggest that MaxEPA is as 1. Aspirin 300 grafts. defined as a loss of 50 oil) useful or moreso than mg plus percent or more of the MaxEPA not different in any aspirin/dipyridamole. dipyridamole 75 gain produced by PTCA. regard versus A/D mg.’ NS in any blood lipids in a 2. 3.6 g EPA subset (n= 42). plus DHA (MaxEPA). 3. Olive oil, up to 1 year.

Oh et al. (Ref. Randomized 9 female and 3 Recumber blood Omega-3 fatty acid-enriched One of the groups used butter 261). crossover of 4 male healthy pressure; VLDL by eggs did not ↑ cholesterol, to prepare eggs, changing the normal eggs volunteers. untracentrifugation, but regular eggs did. Omega- P:S ratio. Pooled data were versus 4 omega-3 HDL by manganese- 3 fatty acid-enriched eggs ↓ not given despite absence of fatty acid- heparin precipitation. TG in one group. order effects for most enriched variables. ↓ LDL in one eggs/day (4.5 g group; NS HDL in either EPA plus treatment. Systolic blood DHA/day), 2 week pressure ↓ in both groups, run-in, 4 weeks diastolic only in one. each treatment.

Olivieri et al. Uncontrolled 20 hyperlipidemic No hypolipidemic drugs ↓ Systolic diastolic blood Design doesn’t allow (Ref. 263). supplementation 16 female, 4 male. for 15 days pre trial. pressure, TG; NS cholesterol, conclusions about omega-3 trial of 20 mL blood pressure by HDL, vitamin E; ↑ glutathione fatty acid-specific effects. fish oil/day blinded nurse. peroxidase activity in RBCs (3.0 g EPA plus and platelets, ↓ MDA. DHA, source not specified), 8 weeks.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2733

TABLE 2—CONTINUED

Study Study Design Subjects Methods Results Comments

Owens and Cave, Observational 6 normal females. Simplate II for bleeding NS WBC, prothrombin time, Trend toward increased (Ref. 264). study, 15 g/day time. Platelet adhesion platelet adhesion, bleeding adhesion with duration of MaxEPA (4.5 g in Baumgartner chamber time. feeding. Assay method EPA plus DHA) 4 using everted rabbit measures platelet changes, weeks. aorta. Prothrombin time, but does not assay vessel WBC and platelet count by wall changes. Study design automated methods. doesn’t allow conclusions about omega-3 fatty acid- specific effects.

Rapp et al. (Ref. Uncontrolled 11 patients, 9 Excludes subjects with Fish oil increased content Shows incorporation of omega- 268). supplementation female, 2 male habitual fish intake. of omega-3 fatty acids in 3 fatty acids into plaque, study of MaxEPA with obstructive 15 endarterectomy atherosclerotic lesion especially DHA. Relevance to at 6 percent of atherosclorosis specimens. Control linearly with duration of CHD not known. Not a specific calories (16 to scheduled for specimens from 18 nonfish feeding, although plasma effect of omega-3 fatty 21.3 g EPA plus peripheral consuming subjects enrichment of omega-3 fatty acids, but would be expected DHA/day), 6 to vascular surgery. undergoing vascular acids plateaued by 2 to 3 to polyunsaturated fatty 120 days. reconstruction. weeks; ↓ cholesterol; NS acids. High amount of omega-3 TG’s, platelet counts, fatty acids. bleeding times.

Saynor and Gillot Uncontrolled 365 During 1 month Total cholesterol by ↓ TG; ↓ cholesterol only Large attrition makes it (Ref. 276). long-term to 40 at 84 enzymatic assay. HDL for initial high difficult to ascribe changes supplementation months. 47 percent after precipitation. cholesterol; ↑ HDL for to fish oil (responders to with 20 mL/day had survived a treatment are more likely to total group; NS LDL; ↓ MaxEPA during heart attack, 49 stay in the study). Lack of fibrinogen. year 1, 10 percent had blinding also may have mL/day angina. contributed to bias. Some thereafter. data were presented for all subjects only, other data only for subsets. Estimates of deviation from mean values not shown. Lack of control prevents conclusions regarding effects to omega-3 fatty acids.

Schmidt et al. Dose-response 10 healthy Simplate II for bleeding NS cholesterol, LDL, Design doesn’t allow (Ref. 277). study 1.3, 4, 9 females. times; t-PA, PAI by platelet aggregation; ↑ conclusions about omega-3 g EPA plus commercial kits; HDL, bleeding time on 4 and fatty acid-specific effects. DHA/day fibrinogen by thrombin 9 g/day, PAI and t-PA (Pikasol), 3 clotting time. antigen after 9 g/day; ↓ periods of 6 TG, fibrinogen. weeks/amount. Randomized to increasing or decreasing dose. 6-week washout.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2734

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Schmidt et al. Uncontrolled 10 Untreated Supine blood pressure. NS cholesterol, LDL, HDL, Design doesn’t allow (Ref. 270). supplementation hypertensives. TG, platelet aggregation to conclusions about omega-3 with 4 g EPA collagen, ADP, systolic, fatty acid-specific effects. plus DHA/day diastolic blood pressure, Absence of significant change (Pikasol), 6 blooding time; ↓ in plasma TG despite 25 weeks. cholesterol/HDL ratio, percent decrease suggests fibrinogen, monocyte inadequate sample size. chemotaxis. Before and after compared by Pratt’s test.

Schmidt et al. Uncontrolled Various at-risk Normal diets. NS Lp(a) in any group. Reports Lp(a) data for (Ref. 279). supplementation subjects with Lp(a) by two-site subjects from 5 previous with 1.3 to 9.0 angina (14), immunoradiometric test Schmidt reports (Refs. 133 g EPA plus IDDM (10), kit. through 135), and the current DHA/day hyperlipidemia refs above. Design doesn’t (Pikasol, MaxEPA (17), allow conclusions about or cod liver hypertension omega-3 fatty acid-specific oil), most for 6 (10), and effects. weeks, angina healthy subjects subjects for 12 (46). weeks.

Shapiro et al. Uncontrolled 10 3 Samples per time ↑ cholesterol, LDL, HDL, Multiple samples per (Ref. 203). supplementation normolipidemic point, 2 to 3 days vitamin E, retinol versus treatment reduces day-to-day with 10 g healthy females. apart. presupplementation and fluctuations, magnitude of MaxEPA/day (5.4 washout; ↓ TG versus changes: Cholesterol 6 g EPA plus washout. percent; LDL 9 percent; HDL DHA/day), 6 11 percent versus average of weeks, 10-week pretreatment and washout washout. values.

Singer et al. Randomized to 47 female Two baseline blood P ↓ systolic, diastolic Olive oil control doesn’t (Ref. 205). propranolol (P), patients with pressure measure 4 blood pressure, recumbent control for polyunsaturated or fish oil (2.9 mild essential weeks apart, blood and upright; fatty acids. Study duration g EPA plus hypertension. pressure measured in fish oil ↓ systolic, and multiple measure (each 12 DHA/day, source triplicate at fixed diastolic blood pressure in weeks) shoos blood pressure not specified), time and post 2 hours recumbent and upright; lowering effect is for 36 weeks, or of rest each 12 weeks. Some additive effects of P persistent. (P) only (12 plus fish oil. weeks) then P plus fish oil (12 weeks) then P plus olive oil placebo (12 weeks). Each followed by 4 week washout.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2735

TABLE 2--CONTINUED

Study Study Desing Subjects Methods Results Comments

Sirtori et al. Randomized, three- 12 Type Iia Lipids by enzymatic Fish oil ↓ cholesterol, LDL, Excellent design. Divergent (Ref. 286). arm crossover of 6 hyperlipidemics. assays, apoproteins by ↑ HDL; Olive oil ↓ LDL, ↑ results from another recent g fish oil (Norsk immunoturbidity assay. study with comparable design HDL; Corn oil ↓ HDL, apoB. Hydro, 4.5 g EPA Platelet aggregation (Bonaa et al. (Ref. 178) that Platelet aggregation ↓ by all plus DHA ethys versus AA. TXB2 by used the same amount and form esters) versus radioimmunoassay. three oils. Fish oil ↓ of fish oil supplement and same olive oil (middle Superoxide by superoxide is monocytes. control (corn oil), suggesting arm for each spectrophotometry. that responses to supplemental sequence) versus oils may be different for corn iol for 6 different sub-populations. weeks each. 1 Absence of change in platelet month run-in and 4 aggregation may also be a week wash-out population specific finding. between each arm with low saturated fat diet.

Spannagl et al. Uncontrolled 13 (3 male 10 Fibrinogen by turbidity NS clotting tests, t-PA; ↑ Design doesn’t allow (Ref. 287). supplementation female) near assay. t-PA, PAI by test PAI, fibrinogen; ↓ TG. conclusions about omega-3 fatty with 8.1 g EPA normoglycemic kits. acid-specific effects. plus DHA/day (PGE- type I diabetics. Fairly high amount of omega-3 technology), 4 fatty acids in this nonnormal weeks. population.

Trial of Randomized life 2182 female and Sitting blood pressure NS systolic, diastolic blood Large, multicenter design with Hypertension style male with after 5 minute rest. pressure. many internal comparisons. Prevention interventions and diastolic blood Measurements for Collaborative double-blind, pressure 80 to 89 baseline, 3 and 6 months Research Group placebo-controlled mm Hg. were made in triplicate (Ref. 289). nutritional on 3 different days. On supplement the fish oil arm there interventions were 161 active and 157 including 3.0 g control subjects. omega-3 fatty acids/day (source not specified), 6 months.

Vandongen et 2 week run-in, 22 female Double precipitations for ↑ cholesterol, LDL, HDL, Study design doesn’t allow al. (Ref. 291). observational insulin-dependent HDL subfractions. HDL2; ↓ TG, HDL3. conclusions about omega-3 fatty trial of 15 g/day diabetics. acid-specific effects. MaxEPA (4.5 g EPA plus DHA) versus no supplement.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2736

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Virella et al. Double-blind 6 normal subjects. Blood samples at ↓ TG in 3/4 on fish oil; Small number of subjects (Ref. 295). placebo (olive baseline, 3, 6, 14, 20, variable in vivo response to limits conclusions. Increase oil) controlled 28. tetanus toxoid booster; ↓ in in bleeding reported in 2/4. study of 2.4 g Humoral immune response. vitro response to tetanus EPA (DHA not toxoid, ↓ IL-2 release. given) 6 weeks.

Vohwinkel et al. Randomized, 48 Healthy subjects. Glucose tolerance tests ↑ Fasting glucose, insulin at Complex results according to (Ref. 296). double-blind to 100 g 4 hours post load. insulin resposiveness. Olive crossover of 6 g oligosaccharides. Response of glucose to load oil doesn’t control for EPA plus DHA/day affected differently by fish effects of polyunsaturated (source not oil, depending on initial fatty acids. specified) insulin response; among low versus olive responders fish oil increased oil, 3 weeks. insulin response and decreased glucose; among high insulin responders, fish oil reduced insulin response and lowered glucose response.

Wander and 3-period 23 normo- Bleeding time by Salmon ↑ bleeding time. Diets were comparable for Patton, (Ref. crossover of triglyceridemic Simplate. Platelet Sablefish ↓ platelet total fat, saturated fat. 297). three fish females. aggregation versus aggregation to collagen; Study design doesn’t allow diets; Dover collagen, ADP. TXB2 by conclusions about omega-3 Both sablefish and salmon ↓ sole (2 g EPA radioimmunoassay. fatty acid-specific effects. ↓ 2 plus DHA), aggregation to ADP, TXB . Salmon (4 g EPA plus DHA), or sablefish (3.4 g EPA plus DHA), 18 day each with 3-week washout between.

Weintraub et al. Metabolic ward 8 normolipidemic Vitamin A fat load test Omega-3 diet ↓ cholesterol, Excellent design studies both (Ref. 298). study. Crossover females. in fasted subjects; HDL TG, LDL, HDL, platelet count chronic and acute fat to 3 isocaloric by precipitation, LDL by versus saturated fat. effects. Many postprandial diets:saturated calculation. Lipolysis Omega-6 diet ↓ cholesterol, fat effects were larger on fat; omega-6 assay using human milk TG, LDL the omega-3 than the omega-6 fatty acids and lipoprotein lipase. NS fasting glucose, diets, but not statistically omega-3 fatty postprandial insulin. significant. acids (3.4 g/day Both omega-3 and omega-6 EPA plus DHA), reduced postprandial lipemia 25 days each versus saturated fat. NS in with 5 to 7 day lipemia between omega-6 and wash-out. omega-3.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2737

TABLE 2—CONTINUED

Study Study Design Subjects Methods Results Comments

Wing et al. Double-blind, 20 Treated Supine and standing Blood pressure lower Study design doesn’t allow (Ref. 299). placebo (olive hypertensives blood pressure. HDL comparable on olive oil and conclusions about omega-3 oil) controlled maintained on after manganese fish oil. ↓ TG in fish oil; fatty acid-specific effects. crossover trial blood pressure chloride/heparin NS HDL on either treatment. of 15 g fish oil medications. precipitation. (Lipitac), 4.5 g EPA plus DHA, 8 weeks each.

Wojenski et Sequential 9 healthy female Bleeding time by Bleeding time ↑ on ethyl No aspirin or ibuprofen. al. (Ref. treatments with volunteers. Simplate II; HDL by EPA; Platelet count ↓ on Evidence for a greater effect 300). ethyl oleate hospital automated Res-Q1000 and ethyl EPA; by the ethyl ester than for a (placebo), 6 g method, platelet comparable amount of omega-3 Ethyl EPA ↓ cholesterol, Res-Q1000 (3.6 g aggregation to ADP, fatty acids in a mixed TG TG, platelet aggregation; EPA plus collagen. TXB2 by supplement. NS fibrinogen binding. DHA/day), or 4.0 radioimmunoassay. g ethyl EPA. Fibrinogen binding by Washouts between (125)I-Fibrinogen phases of 5 versus saline. weeks, 4 months, respectively, and 8 weeks posttreatment.

Wolmarans et Crossover Healthy Habitual diet. Fish diet ↓ cholesterol, NS total fat but ↓ saturated al. (Ref. comparison of subjects, 12 LDL, VLDL; ↑ HDL. fat on fish diet. EPA was 301). red meat to females, 16 1.91 g/day versus 0.06 g/day fish, (6.1 g EPA males. in baseline and 0.01 g/day on plus DHA/day) 3 meat; total omega-3 fatty week baseline, 6 acids were 6.1 g/day on fish, week treatment, and 0.9 g/day otherwise. 6 week posttreatment and 3-month washout.

Federal Register / Vol. 58, No. 3 / Wednesday, January 6, 1993 / Rules and Regulations 2738

TABLE 2--CONTINUED

Study Study Design Subjects Methods Results Comments

Zambon et al. Randomized 10 females with Regular diets. Insulin Fish oil ↑ fasting glucose, High amount of omega-3 fatty (Ref. 304). crossover trial NIDDM. by radioimmunoassay. NS fasting insulin, ↓ acids may produce effects on of fish oil 15 Automated glucose postprandial insulin. Fish glucose metabolism not seen g/day SuperEPA analysis. Cholesterol with lower amounts. Effects oil ↑ LDL, NS cholesterol, (8 g EPA plus by enzymatic methods. are consistent with other HDL DHA ethyl reports, but absence of esters), with polyunsaturated fat control and without limits inferences about glyburide 8 specificty of the effects. weeks on fish oil, 4 weeks on and 4 weeks off glyburide. Baseline treatment was glyburide alone, 4 weeks.

Abbreviations used: apoA, apoprotein A (a protein in high-density lipoprotein) apoE, apoprotein E (a protein in many lipoproteins, most notable VLDL and HDL); ASA, acetylsalicylic acid; ATP, adenosine triphosphate; CDC, Centers for Disease Control; CHD, coronary heart disease; ELISA, enzyme-linked immunosorbant assay; MDA, malondialdehyde; NIDDM, noninsulin dependent diabetes mellitus; NS, not statistically significantly different; PGE-M, prostaglandin-M; TEARS, thiobarbituric acid reactive substances; TG’s, triglycerides; TXB, thromboxane.