This document reviews evidence that fish oils may prevent or slow the development of atherosclerosis. It discusses epidemiological studies showing lower rates of heart disease in populations that consume fish. It also describes in vitro studies demonstrating that fish oil incorporation changes eicosanoid production in a way that could inhibit platelet aggregation and reduce inflammation. Animal studies and some clinical trials further support potential anti-atherosclerotic effects of fish oil supplementation.

1. i74
REVIEWS
Fish Oils in the Prevention of Atherosclerosis
DOUGLAS H, ISRAEL, MD . RICHARD GORLIN, MD . FACC
New Yo:k, New York
The hypohes8 that a ds dervvcd loans the flesh of fish and marble
mammals inl sit the atherosclerotic process h ei,tcalb reviewed.
Populations consuming a dirt rich la fish have law vie, at
coronary heart disease. Dietary fish oil is associated with changes
N serum lipids, prealaglandio and leukotriente metpWi el, en.
hatted t itdolhelial fcaetlen and effects a growth rac:on released
from platelets, knineyles and endothelial cells. Dietary fish ant
In the past decade remarkable advances have been made in
the treatment or eoronrwy artery disease, particularly in
relation to the advanced thrombohc complications of ather-
oselerosis . Whereas basic research is providing greater
insight into the pathaphysiology of atherosclerosis itself, to
date our progress in preventing and treating the early ath-
erosclerotic lesion has been less spectacular . In this review
we will examine the basic, experimental and clinical evi-
dence contenting the hypothesis that fish oils may prevent
or retard the de•nelepment of atherosclerosis .
Support for a positive role for fish oils in inhibiting
atherosclerosis is based on four lines of evidence:
1 . Epidemiologic roadies on the effect offish consumption
on morbidity and mortality from coronary heart disease in
humans.
2- In Wins studies showing that Ash oils aher biochemical
processes implicated in atherogenesis.
3 . Anima l studies showing an inhibitory effect of fish oil
supplementation on the development of atherosclerosis .
4. Studies with clinical endpoints in humans.
Epidemiologic Evidence
Much of she interest in the possible role of fish oils in
preventing atherosclerosis derives from epidemiologic ob-
servations of Greenland Eskimos . Despite the average 60-
year life span of these Eskimos, they had only a 3 .5%
mortality rate from coronary heart disease (1,2) . Original
inves:igatione (3,4) showed that the Eskimos, crmpred with
Fwm the Depatnrcnl of Mediim. Division OF Codiobst, Moot Stem
Medical Center. New Yovk, New York.
Matwsabt received November 7,1990 : eesksd soeswsipt waived key
30. 1971, a srpled Ocmber 9. 1991.
,&,..s to,.aeinr : khsard C-li, MD. Chv.-aan, Drpvnawm or
Medkhe, Mount Sisal Medical Center, Box I I lt, I Costars L . Levy place .
New Yek, New York 10029,
01992 by the American Coltear. of Cardlulupy
JACC Vol . I9 . No . 1
ran., I592:17e-s'~
sunk eseMntlou has been associated with kmlMtrvn atatheraude&
rash expestmemidly induced by dietary II)p"Holdemills anal bell.
Man Injury . Results ofshades of the use of fish dl to Inhibit
pmtmsgloplagy resteaonit la human subjects have been facaacta-
0555 .
Ii Am Cola Cwrdiol 199249:174-85)
Danish subjects, had low blood levels of cholesterl and
triglycerides but elevated levels or high dens . y :,p'nrotein
(HDL) cholesterol despite a verb st , intake of dietary fa+ .
Fatty acid content afflub and marine mammals. A alysis
of the Eskimos' diet showed that most of the at and catories
were derived from the flesh of oily fish and marine mammals.
Biochemically the fatty acid profile of fish amp marine
mammals differs from that of vegetables or land animals.
There is a high content of long chain fatty acids eontauiog 20
or 22 carbons with up to 5 or 6 unsatumted eaeoons.
Furthermore. in fatty acids common in vegetable fat, the last
unsaturated carbon is usually located sixth from the methyl
end, whereas in fatty acids coiamon in fish, the final unsat-
urated site is most commonly located third from the methyl
end . Thus. fatty acids common in fish have been termed N3
or omega-3 fatty acids, whereas vegetable fatty acids are
tatted 146 or omega-6 fatty acids (Fig. 1). The two most
abundant N3 fatty acids found in fish are a 211-carte„ fnny
acid with five unsaturated carbons called eicosapentaenoic
acid and a 22-carbon chain fatty acid with six unsaturated
sites called dtcosahexacnoie acid (Fig. 1) . Although many of
tale important biochemical effects of N3 rang acids discussed
hem have been associaled wide eieosapemaennic acid, do-
cosahexaenoic acid may also be important. Docosa-
hexaenoic acid tends to concentrate in human phospholipid
and can he slowly metabolized to eicosapentaenoie acid .
Thus . it may serve as a depot form of eicusapentacooic acid
Ad perhaps has other, as yet unknown, specific functions .
In addition, docosahexaenoic acid inhibits platelet aggrega-
tion in vitro and may contribute to the overall platelet
inhibitor actions of N3 fatty acids .
Effects of Ash and whale meat on platelets and blood
cholesterol. Bang and Dyerberg (5) reported that it was
common for Eskimos to consume 400 to 500 g of fish or
whale meat daily containing up to 7 g of N3 fatty acids and
that dietary N3 fatty acids, particularly eicosapemaenoic
acid, were incorporated into the phospholipid of the Eski-
1n5-t097 .T51 .1o

2. IACC Vol . 19, No. I
ranuary 1992:174-85
:tools Acid (n4) Family
H5C
VA .
COON
linnlerc acid
18 .a.-8
H st~/~~ y , . •. CODH arachidnnic arid
to 4 9
a -Llnolenle Aela (n .3) Fnmay
Ha ,
Count
e - hnmenic acid
H af- /, J . ./
~~,~"J` . CCCH ricnsaoertaennic acid
no s..s
coo"
alid
Flgnre I . h .^ iml formulas far the N3 and N6 families of fatty
acids. f.inoleic acid. derived tram vegetable sdurees, undergoes
elongation and desaturauon in its conversion to arachidomc acid
(top of puod). Alpha-linolenie a rid is synthesised by plankton and
serves as the precurior for eicosapeotaenoic acid and ducrsa-
hexaenoicacid!haine,tn ase ) the areconcentratedinthemarine
food chain . The two families differ in the length of the carbon
skeleton, the number of unsaturated bonds and the position of the
last unsaturated bond. Maine fatty acids have theirfirst unsaturated
carbon third from the methyl eud and are Ihue termed N3 oremcga-3
fatty acids: vegetable fatty acids have their first double bond
between the sixth and seventh carbon from the methyl end and are
terw-d W or omegafi fatty acids. Reprinted with permission from
van Schacky C . Propbytaxis of atherosderosis with marine omega-3
fatty acids: a compreueusF,.: Ana laces Med 1902,192 :
59(5-9 .
mos' platelet membranes leading to a prolonged bleeding
time and decreased in vitro platelet aggregation (6,7). The
low incidence of ischemic heart disease was attributed to
inhibition of platelet function and low blood cholesterol
levels that were dietary rather than genetic in nature, be-
cause Eskimos on a standard Danish diet had serum lipid
values similar to those of the Danes (4).
This hypothesis :s supported by other epideiniologic
studies. As reported by Keys (8), there is a low incidence of
coronary heart disease in Japan, where per capita fish
consmnphon averages NO g daily (91. The inverse relation
between coronary heart disease and fish consumption in
Japan appears to be dose dependent (9) .
A strong inverse relation between coronary heart dihare
morteliry and fish eonsumprfan was also found in the 7.ut-
phen study (10) . However, protective effects correlated with
a much lower average d; dy buttconsumption and did not
appear to be dose dependent . Compared with non-fish eat-
ers, men who consumed an average of only 30 glday of fish
had n `.O decrease in coronary mortality .
A reexamfnarfan of epldemfnlagic data from be Wectera
Electric study (11,12) also detected an inverse relation be-
tween fish consamr •lion and coronary heart disease mortality
ISRAEL AND GORLIN
175
FISH OILS IN THE PREVENTION OF ATHEROSCLEROSIS
and showed no positive association between fish intake and
mortality from other crises . Despite these results, other
stallies have failed to demonstrate any protective effect of
fish consumption en mortality due to coronary heart disease
(1),141 .
Summary of epideoialogic dun . In perspective, the epi-
dumiolegic data provide suggestive, but inconclusive. evi-
dence that fish roncumption protects against coronary
deaths, Considerable caution is sdNistable when interpreting
these data. Ear example, it is possible that the striking
absence of atherosclerosis in Greenland Eskimos is attribut-
able . at (cast in purl, hr genetic rather than dietary influ-
ences . For example . although adoption of a western diet by
Cirectand Eskimos results in adverse charges in their serum
lipid profiles, an follow-up data show that this population
then experiences a higher incidence at atherosclerotic dis-
casc . Studying a different Eskimo population, kuhinowitch
(15) believed that he had shown that Canadian Eskimos
living in more southerly latitudes (and consuming a more
western diet) did show clinical evidence of aiilerosclerosis .
However, this conclusion was based on a greater frequency
of sclerotic radial and temporal arteries, as detected by
palpation, and a higher mean blood pressure !ban found in
Canadian Eskimos living in more northerly latitudes and
consuming a more traditional diet. It is difficult to account
for the very positive results of the Zutphen study (I0) in a
population whose mean consumption of 20 glibly of lean fish
would supply only 0.4 to 0 .5 g of N3 fatty acids daily . It is
probable that the variable results of the different epidemio-
logic _fcs may toe explained a i peoet :e differences is tire
study populations . differences in the amount and the type of
fish consumed and differences in study design and accuracy
of reporting . It should be recognized that there may be
serious weaknesses in epidemiologic studies, particularly
when investigators use dietary recall methods or patient
questionnaires to estimate average daily fish consumption
over long periods of time and then attempt to relate this
information to coronary events, rometimes without adjust-
ing fe. (he effect of other relevant risk facors .
to Vitro Studies
Fatty acids released from cellular membrane phnspholip-
ids function as precursors for misty biologically significant
products. When fish fat is a major component of the diet,
eicosapentaendic and docosahexaenoic acids are incorpo-
rated into the membrane lipids of platelets, leukocytes,
erythrricytes and endothelial cells. Because of their struc-
tural similarity to arachidonic acid, these fatty acids may
provide an alternative substrate for the erreymes cyctooxy-
genase and lipoxygenase and perhaps others . leading to the
synthesis of distinct prostarroids with properties very dif-
ferent from those chat are derived from arachidoate (Fig . 2).
Thus . it is net surpri .,lng that fish oils have diverse biologic
ellects .

3. 1 76
ISRAEL AND riORLIN
Plsv OILS IN TOTE NREVEN .'ION OF ATHERUSCI FROSTS
11-HPEPE
o. ~ eau„
I
.5501„
~PGG3,PGH3
PGE,.D,,P,~ PGI,
TXA.
PROSTAGLANDINS ANO THROMBOXANES
PHOSPHOLIPIDB
1
EICOSAPeNTAENOIC ACID
5-HPEPE
1
o
coon
5t's
L.utaN»n . As O.TA,r
~Hs e/ `.rarnuo .
LToO LTC. -, To, .-CL LTO
LEUROTRIENES
Meets on prostaglandin metabolism (Fig . 2). As men-
tioned, Dverherg and Bang !ol first demonstrated that di-
Mary eicosapentoetloic acid is incoTpnrated into platelet
membrane phospholipid by performing chemical analysis on
platelets obtained from Eskimos and Danish control s •_-~
jc s. Compared with Danes, Eskimos with a diet rich in fish
had decreased in vitro platelet aggregation and a longer
bleeding time, findings reproduced in other studies (16 .171
and demonstrated to he dose dependent (18) . Since then,
studies using platelets isutaled from volunteers fed a diet rich
in fish oil have shown that the eicosapentaenoic acid incer.
porated into the platelet membrane is convened to throm-
boxane AJ (19-22) (Fig. 2), a structural analogue of throm-
5ar-aae A, with no plaieici agoai l aciivily (23,24) . In
endothelial cells, cicosapenraenoic acid is convened into
prostoghandm I, . a toeaoic analogue of prostacycli 1 (25 .26).
Unlike the biologically inert thrombo .eane Aa , prostaglandin
1, retains potent vasodilator and platelet antagonist proper.
tics .
Effect of N3 fatty acid an human prostnglandin lnetabo.
hsm . Knapp et at-122) assessed the effects of N3 fatty acids
on prostaglandin metabolism in atherosclerotic subjects and
normal control subjects using doses that approximate those
ingested by Eskimo populations (10 gldny eicosapentaenoic
acid). Baseline urinary excretion of thromboxane A, and
prostacyclin metabolites was significantly elevated in athero-
sclerotic subjects, probably reflecting increased platelet-
vessel wall interactions and generalized platelet activation
(27), Excretion of urinary metabolites Ol thromboxane A,
decreased markedly both in patients with atherosclerosis
and in heakhy subjects. whereas the excretion of urinary
metabolites of prostacyclin was significantly depressed only
in subjects with atherosclerosis . Increases in urinary metab-
olites of thromboxane A3 and pross4andin i3 were noted in
both groups . Thus, supplementation with fish oil normalized
elevated prostacyclin synthesis in subjects with atherosele-
rosis, suggesting a decrease in abnormal platelet-vessel wall
interactions and partially impaired the synthetic capacity for
thromboxane A 7 in all subjects. Thromboxane A3 synthesis
was only partly suppressed, probably because eicusapen .
JACC vd. 19, No. I
Jsnunry 1997 :!]4-r5
figure 2. Eicusanoid products of the metabolism of
nsapcn/aenulc acid, When ingested ie the diet,
eiaosapnslaenuic acid is incarporafed into platelet,
leukocyte and endothelial cell membrane phospholipid
and leads to the biosynthesis of a family of trienaic
prrrstagiandins (PG) including thromboxane (TX) A,
and prostaglandin l, . and tetraenoie leukotricnes II.T)
ir.Jsfing lenkmrienc B,. See text for deta'! : . Adapted
with permission from von Schacky C . Pruphylaxis of
atbaassrderusis with marire omega-) fatty acids- a
rcmprehensive strategy . Ann Intern MM 19f7 ;107:
Old- 9 .
tac.ioic acid compete ; inefficiently with araehidonic acid for
cyclaoxygenase (1;) . Recent data showing an inurease in
platelet survival time in hyperlipidemic patients with docu-
mented atherosclerosis consuming a diet supplemented with
fish oil provide further evidence that fish oil may cause a
decrease In abnormal platelet-vessel wall interactions I291 .
Thus, there is evidence that high doses of dietary fish oils
engender a net change in the hemostatic balance, and thus
would appear to protect against thrombosis .
'Reels an kuketrbene metabolism (Fig . 2) . Current evi-
dence suggests that leukotrienes contribute to the acute
inflammatory response produced by vascular injury and
myocardial infarction 130) . In neutrophils and monocytes,
membrane -bound arachuionate is converted through inter-
illediates, to a family or leukutrienes including leukotriene
B., a potent chemoattractant forncutrophils and monoeyres.
Leukc,trienes C„ D, and E4)31), synthesized by monocytes.
have been implicated in coronary vasoconstriction (32).
ischcmic myocardial depression (331 and ventriculararrhyth-
mia (34).
Role of dietary N3 ratty acids in kukatriene metabolism .
In experimental animals exogenous administration ofdictary
eicosapenlaenoic acid suppresses the formation of teukatri-
ene B, in a dose dependent manner and results in the
syntbesis of a. biologically much less active molecule, Icuko-
triene B,(35 .36) (Fig . 2). incorporation of ekosapentaenoie
acid into neutrophil membrane phospholipid and in vitro
generation of leukotriene B s have also been demonstrated in
humans after dietary supplementation with N3 fatty acids
(37,38) • Using uemrophils collected from normal subjects
ingesting 3 .2 g of acosapentaenoic acid daily, Lee et W . 139)
demonstrated in vitro synthesis of leukotriene B, and 48%
suppression of leukorriene B, production . Monocyte synthe-
sis of leukotriene B, was also decreased by 58% after 6
weeks. Even the moderate doses of fish oil used in their
study led to major impairment in the neutrophi' chemotactic
response to leukotriene B„ and to decreased neutrophil
adhesion to cndoLhalial monolayers pretreated with leuko-
triene B 4 .Monocyte chemutaxis may also be impaired by
fish oils (40) .

4. 1ACC Vol. 19, No. I
lrnumy 1992:171-85
Although the significance of these findings is speculative
at present . altered leukatriene metabolism could potentially
have important modulating effects in myocardial infarction
;ud possibly chronic atherogeresis.
Role he myocardial Inrnrerion, Neutrophi! adhesion to
endothelium may be noted adhin 15 to 24 min of coronary
occlusion (41). There ~. evidence that neutrophil-donee(i
mediators produce alterations in microvascular rare, pro-
mote interstitial edema and the "no reflow" phenomenon
and c.nntribute to myocardial stunning and arrhythmogene-
s . Abhough the mechanisirt remains to he clurdtocl, pre-
treatmcsl with N3 fatly acids in expermental coronary
occlusion and reperfusion in dogs, cats and rats (42-441
resulted it sienificandy smaller infarcts, fewer arrhythmias
and fawn sudden deaths than in control animals. Similar
csaults wen obtained in a feline carotid artery occlusion
model 145); cats pretreated with N3 tally acids had signifi-
cantly smaller cerebral in€ards 17% versus 19% of the
territory at risk). In addition, dietary tuna oil was protective
against ventricular fibrillation during both ischemia and
reperfusion in a rat model of coronary occlusion (46) .
Protective effects have not been associated with impaired
thrombogenesis (42.43) or ddlerenees in regional myocardial
blood How (42,43) and it is possible that altered leukotriene
metabolism and impaired leukocyte responses may have
contributed in part to the beneficial effects noted .
Role in atherugenesis . The important role of leukocyles
and leukocyte-derived growth factors in atherosclerosis is
being increasingly recognized. In the hypercholesterolemic
primate model (4)) monocyles adhere to endothelium, Iro- .
Irate the imima and begin to accumulate lipid within 2
weeks of beginning an atherogenic diet. Continued choles-
terol feeding results in further monocyte accumulation and
smooth muscle cells migrate from the media toward the
intima where they proliferate and synthesize connective
tissue (47,481 . The myointimat orolilerafion that follows is
dependent on many factors including mimgenic and chemo-
tactic factors derived from monocytes. platelets . endothelial
cells and smooth muscle cells, as tee]] as toxic release
products from macrophages . The overall effects of N3 fatty
acids on monoeyielmacrophage participation in atheroscle-
rosis remains to he etarihed . bat it is known that dietary fish
oil impairs monocyle chemotaxis (40) and decreases synthe-
sis of interleukin-I and tumor necrosis factor (discussed
later), platelet-activating factor (49) and toxic oxygen-
derived free radicals (50) .
Of its On Growth Farf :wrg
Ia addition to the cffc :.s an prostaglandin aid leukoiri-
ene metabolism noted previously, N3 fatty acids may affect
the synthesis of growth factors by platelets, endothelial cells
and monocytes .
Platelets. Platelets synthesize and release several growth
factors during their adhesion to the injured blood vessel wall .
These include platelet-derived growth factor, epidermal
ISRAEL AND GORLIN
177
FISH OILS IN THE PREVENTION OF ATHEROSCLEROSIS
growth factor, transforming growth factor beta, platelet
factor 4 and betadhromboglubutin . Platelet factor 4 and
beta.thromboglebu!in are chemotactic for smooth muscle
cells and aunlccytes. During the platelet veasel wall interac-
tion stimulated by vascular injury, platelet factor 4 rapidly
penetrates the intima and media and may play an important
contributing role in the subsequent hyperplastic response
(51) . In a study by Hay et al . (52), administration or 3 .5 g of
eicusapentaeaac acid daily to 13 patients with ischemic
heart disease led to a 75% decrease in bloc sAevels of platelet
factor 4 and a 70% reduction in beta thromboilobulin,
whereas platelet survival time (29.52) . a marker of platelet
activation and consumption, was increased by 100 . A
decrease in serum berg-rhromhngltsbulin was also shown by
Knapp ci al . (22) after dietary supplementation with high
dose eicosapentaenoic acid . Neither study elucidated the
rrrxhanism whereby fish oil led to decreased blood levels of
these substances . To date, no study has examined the effect
of dietary fish oil on release of platelet-derived growth factor
from platelets, although there is evidence regarding release
of this mitogen from endothelial cells .
Endotttoliat cells. Endothelial cells are known to synthe-
size raitogens for vascular smooth muscle cells and fibro-
blasts (531 . including platelet-derived growth factor (54) .
Recent in vitro evidence (55) has demonstrated nearly com-
plete soppression of endothelial cell platelet-derived growth
factor activity when cultured bovine endothelial cells were
incubated with fish al concentrate, an effect that apparently
depends on oxidative processes because it was i,thibited
when antiasidants were added to the culture medium . This
effect may be important in inhibiting imitnal hyperplasia in
the setting of denuding cndoibelial injury as found alter
angionlasty because endothelial release of platelet-derived
growth factor appears to be markedly increased in the setting
of endothelial injury or cell death (56) .
Meawyles, A recent study (57) demonstrated that di-
etary supplementation with 4 to 5 g of N3 fatty acids daily
ran suppress the synthesis of intereukin-I and tumor necro •
its factor by matrotytes and other mornnuelear cells ob .
tained from normal human volunteers. Theoretically, im-
paired interleukin-I and tumor necrosis factor release could
inhibit atherogenesis in several ways because these sub-
stances are known to have a direct toxic effect on the
endothelium, promoting leukocyte adherence and inducing a
procoagulant state (58.59) . In addition, interleukin • : has
direct mitogenic activity for smooth muscle cells (60) and
stimulates fibroblast growth and metabolism resulting in
increased biosynthesis of collagen and other connective
Tissue elements (61) .
Effects on Blood Lipids
Despite the vast number of experimental animal and
human feeding trials, much confusion persists regarding the
effects of fish orconcent ated fish oil consumption on serum
lipids. T.wc comprehensive recent reviews (62,63) summa-

5. 178
ISRAEL AND r,ORLIN
FISH OILS IN THE PRFVENIION OF A711EROSCLEROSIS
me what is currently known, based on human and animal
feeding trials.
Triglyceride levels, These are generally decreased in a
dose-dependent manner with fish or fish oil supplementation .
At very high doses of up to 20 to 30 glday of oil, fish oils are
among the most powerful hypotriglyceridemic agents avail-
able, although the effect may not always be sustained (64).
Total cholesterol levels. These are variably affected by
fish or fish oil supplementation. Results from animal studies
are inconsistent, probably because of differences :R the dose
of fish oil used, species differences and differeraes in the
remainder of the diet, Similar variability is noted in trials
involving humans. In many cases decreases in total choles-
terol levels are due to a fall in very low density lipoprotein
(LDL)-cholesterol in patients with a marked hypottiglyceri-
demic response to fish oil .
LDL cholesterol levels. These are variably affected by N3
fatty acids. In trials involving fish or fish oil administration to
normolipidemic humans (62,63), the major factor determin-
ing its effect on LDL cholesterol has been the amount of
saturated fat in the diet. When fish oil was added to the diet
(and saturated fat intake held constant), LDL cholesterol
levels tended to rise. When fish oil was substituted for
saturated fat, LDL cholesterol levels decreased in 30% to
40% of cases and remained unchanged in the rest ; this effect
is similar in magnitude to that seen when polyunsaturated
vegetable oils were substituted for saturated fat (62 .63). In
studies with hyperlipidemic patients, saturated fat intake
was usually not a variable and LDL cholesterol levels rose in
most of the studies.
HDL cholesterol levels. High density lipoprotein (HDL)
cholesterol levels hove been reported to be increased, de-
creased or unchanged after dietary supplementation with 143
fatly acids. Most placebo-controlled crossover studies (63(
have shown an increase in HDL cheiesterol of 5% to 10%
with small to moderate doses of fish oils ; these effects may
reflect a r ciprocai relation with trigiycendes, When very
high doses of N3 fatly acids arc used . HDI. cholesterol may
fall and this effect may also be seen with polyunsaturated
vegetable oils in high doses . Some studies (65,66) ineicate
that N3 fatty acids in high doses increase HDL levels with
short- but not long-term supplementation . A study in non-
human primates (67) suggests that when HDI . levels are
decreased, this effect may be due to the synthesis of smaller
HDL particles with a lower cholesterol content per particle .
Finally, there is some evidence (68) that N3 fatty arias may
increase HDL turnover or up-regulate HDL receptors .
Corrrlatian o inhibitory effects on atherosclerosis and
sent∎ lipoproleia caneentrat(ons. Although a number of
experimental studies showing inhibitory effects of N3 fatty
acid supplementation on atherosclerosis failed to correlate
these effects with favorable changes in serum lipoprotein
concentrations, it should be noted that N3 fatty acids ma)
alter the atherogenicity of lipoprotein particles by affecting
their size, apoprotcin content and physical properties (63) or
alter their uptake and turnover by affecting receptor mech .
r ce sat 19 . No . I
Jonuarr 1993:179-x5
anisms (68). These effects are potentially clinically relevant
and would not be reflected in simple measurements of blood
lipid concentration. Concerning the inconsistent results of
fish oil feeding on Iolal . LDL and HDL cholesterol concen-
tralions, these reflect very large differences in study design,
dusing, patient-related variables and especially the overall
dietary composition and total ratio of polyunsaturated co
saturated fatty acids . In addition. heterogeneity of the fatty
acid composition and cholesterol content of the fish ail
supplements themselves may explain some of the variability
o;effects of fish oils en human lipoprotein levels. Most 60
oil capsules contain only 12% eicosapentaenoic acid and
18% docosahexacnoic acid-the remaining 70% of fatty
acids are composed of a variable mixture of saturated and
maooasssaturated and polysatsaturated fatty acids . In a re
cent study (69), hypereholeslerolemic men fed 2 .7 g of highly
purified eicosapenlaenoic acid daily had a major red uctiun io
total and LDL cholesterol with an increase in apoprotein A-I
levels, suggesting a change in the composition of the HDL
panicle.
Effects on Other Relevant Variables
N3 fatty acids also influence a variety of other clinical and
laboratory variables relevant to the development of athero .
sclerosis (Table I),
Endutheliam-derived relaxing factor . It has been demon-
strated (70) that dietary supplementation with N3 fatty acids
may enhance in vitro enothelium-dependent vasodilation in
a dose-dependent manner in hyperehnlesterolemic and ath-
erosclermic pigs. Recently the relevance of this effect has
been confirmed angiographically in patients with coronary
artery disease (71) . Patients with significant disease of the
left circumflex or right coronary artery but only mild lumen
irregularities of she left anterior descending artery had a
significant dose-dependent vasoconstrictor effect to acetyl-
choline i^;ected selectively into the left anterior descending
artery, indicating endothelial dysfunction . Patients were
retested after 6 months of high dose fish oil administration-
Improved endothelial function was noted in all patients, with
resumption of a normal vasodilator response in 75% . Similar
results were recently demonstrated (72) in heart transplant
recipients after dietary supplementation with N3 fatty acids .
Hil rinolylic responses. A reduction in the concentration
of plasminogen activator inhibitor-I levels was found by
Mehta et al . (73) in the blood of normal subjects as well as
those with coronary artery disease after 4 weeks of dietary
supplementation with N3 forty acids . Although the mecha-
nism of this effect is unclear, it may be related to the
hypotriglyeeridemic effect of N3 fatty acids. Recently two
studies (74,751 demonstrated a fibrinogen-lowering effect of
N3 fatty acid supplementation . Although this effect was dose
dependent (75) . a clinically significant effect (21.6% decrease
in fibrinogen) was seen after 20 weeks of supplementation
with only a moderate dose of fish oil concentrate supplying
2.2 glday of N3 fall) acids . Preliminary evidence (76) sug-

6. JACC Vol. 19. No . I
January 1992 :174--as
FISH OILS IN THE PREVENTION OF ATHEROSCLEROSIS
Table 1. Effects of N3 Fatty Acids on Blood Cells, Lipids and Coagulation- Clinical Implications
Inflammation
CNgnId s
Dec,hbnoogen
Oee .llberogen :dec.
Pal-1
Deer = decreased; EDRF = •..d.11- 4-deri-cd mincing facto; t41 = 'nnerleukin.l : ='n - rcased; LTB, = leukoaieoe B4; LTB, = It-limit 115:
PIE - plamlcl-edivet ng labs; PAM - plaani.nnged anti asw inhibit-1. PG1, = prostacydm : Pill, = I, izgtendm 13; TNF = moan necrosis fur;
TxA; = Ibrnmhoaane A,. TXA, -&-h- ,A, . ' lopeM, nn amrn+ition of diet crA I oral pdynnsatumrod to satualed fatty acid satin.
Bests that docosehexaenoic acid tray have a more powerful
effect on fibrinogen levels and platelet funciiuii Lines :wt of
eicosapentaecoic acid ; thus . further research will be needed
to elucidate which ratty acidss should be given and in which
doses to produce optimal inhibition of hemostatic function .
Bind pressure. Some studies (77,78) have shown a mild
hypotccnsive effect of N3 fatty acids that appears to he dose
related. Studies (79,80) using low doses of N3 fatly (3
to 3 .5 glday)have had inconsistent effectson blood p •essure.
Animal Studies
Antiatherogenk effects of N3 fatty acids in pins and mon-
keys (Table 2). Experimental studies have examined the
potential antiatherogenic effects of N3 fatty acid supplemen-
tation in several animal species, with the most notable
positive results in pigs and monkeys . Clear decreases in
coronary lesion area in hyperlipidemic swine giver, fish oil
supplements have been observed (70,81,82) . A relatively low
dose of ..:id liver oil (30 mLiday) was added to an atherogenic
diet in each of these studies. In two studies (70,81) balloon
abrasion of the left anterior descending artery was also
performed m stimulate lesion development . Each study
showed a significant decrease in coronary lesion area as-
sessed morphomettically in serial sections of the arteries but
no apparent beneficial effects on scrum lipids . In another
study (83), normolipiaemic pigs, fed a lard or lard-mackerel
diet, had a nonobsumctive Teflon constrictor surgically
implanted on the left anterior descending coronary artery .
After a 4-month period, the pig, mete killed and lumen
encroachment was evaluated by morphometric techniques
ISRAEL AND GOBLIN
179
Des . LTR,; inc . LTB,
Des. LTB. ; inc . LTB, ;
dec . adhesim and
ti-t-,
similar to those used by others (91,92) . Lumen encroach-
tneil! was increased less in the lard-mackerel oil group (I1%)
Than in the group given lard only (62%) .
A follow-up study by this group (84) demonstrated appar-
ent regression of atherosclerosis after N3 fatty acid feeding
in the pig model. Coronary atherosclerosis was induced by
hafloon abrasiem or the left anterior descending artery in ags
fed a cholesterol-lord diet for 4 months. After this Induction
period. control pigs was killed; bal€oua-abraded arteries
had 11% lumen eneroachmem compared with 1 .3% in non-
abraded coronary arteries . Other groups of pigs subject to
the same induction protocol were then fed diets containing
lard, fish oil or both for 3 additional months- After this time,
fish oil-fed pigs had significantly less lumen encroachment in
the balloon-abraded arteries (3%) than did the control pigs
(11%) or those given a lard diet (13%) . Lumen encroachment
in nonabraded arteries progressed in the lard-fed group
(I1%) but not in the fish oil-fed group (L1%).
In rhesus monkeys fed a 2% cholesterol and coconut oil
diet for 12 manlhr (85), cholesterol levels rose to approxi-
mately 900 mgJdl and up to 80% of the inllmal surface of the
aorta demonstrated atherostlerotic changes. Monkeys re-
ceiving thin diet but also receit.ing supplements of either a
low or a high dose of menhaden v9, had a 40% to 55%
dose-dependent reduction in the involved aortic intimat
surface area despite a corresponding dose-dependent do-
crease in HDL concentrations. Surface involvement was
also decreased in the carotid and femoral arteries . Both
cholesterol content and lesion thickness were shown to ho
decreased by menhaden oil in the; study . i n another study
(86) in hvperlipidemic monkeys, animals with fish oil-
Atherngenesis and Inamat Hypetplasia Thrombdgenesis Vasnmotion Ischemia and M ite rfasion
Hateiels
Dec . factor.1; den. e .thrombogk blr ee . T'XA, ; me . THA,- lkc . TSA, : tee
tnc . platelet survival platelet court : inc, 'I'5A,
ing time
Fndothelium
INc . POOP Ire . PGI,; pte,etsed its Poll, inc.
POt,; me . EORF effect I7DRF err-
M-,[.
Dec. IL-1,7NF; ^ doe. chemmacis; rlus. RAF Des . free radical synthesis;
dec . five tailed synthesis; ^ dec . dec . PAP
mortnepe-gbrinogen interuelioss
Neulrephil
Dec . LTB,: ire . LIE, Dec . LTB,; inc . LIB,. dec .
adhesim xd cheawtaais
Lipds
Variable eaeaO Dcc . to Iyecrides

7. ISO ISRAEL AND GORLIN
1ACC v0:. 1e N
FISH OILS IN THE pREVENJJON OF ATHEROSCLEROSIS
January 7EC 74-Ct
TAW 2, Studies of Fish Oil Supplementatio nn in Experirrsensal Adwrosclerai.: .
FReet en
Type and A-nu r1Fivh Uil
Ads-learns
Effea on Lipid
Cod lire o3 (w['doy(
Markedly deceased
LlildlydsreascdHDL-
C, no charge LDL-C
Cod I:eu it 130 nrUdayl
Markedly dcncased
MiWSy decrrvud tmd C.
LD),C
a .5it no : S-ral oil • 4.3"i Ia'J Mildly decreased
Parts. low C; 1117i:
C . LDL.C mt
assessed
2% choederol - cod liver it
Moderately decreased
No different in LDL :
ris milday )
HDL Ialio in treated
versus timbals
Mtnhaleo rat Vlf))
I . 2511 MO/ducadal ail 1 :1
Mildly decreased
Mildly decreased total
2 . ItT MO/cescono l it 3:1
Markedly decreased
C . esarkedly deceased
HDL-C
Fish oil nmspeeifcedy 2 Eday
Mildly decreased; mildly
No signdic aut elect
increased at carotid
bifurcation
Macepa ronc!dmte. • 8 .6m of Slight increase
faaease toed C and
diet by 50910
LDL-C
Muepa coesetenate. • 751 of
Mildly decreased
Markedly decreased
diet by w0ghr
anal C
Prplahol ranccmrae'
t .L.w dose it mVdayl
htilay Jacreused
Blamed due in local C-
ndldiy decreased
HDL-C
Markedly dcercased
Bloraed rise in 10501 [-
mildly deceased
HDL{
Mildly decreased
Increased last C .
markedly decreased
HDL-C
Mildly decreased
May decreased total C
Markedly increaned
No change in goal C
No change
No change
Markedly Jeerused
Mildly rnereased f
Mildly increased coral C.
cells
mddlysleereased
Mildly increased
HDLC
m0005010 adlsudon
and mildly increased
foam cells
Lmdy'mcre 196)
Dog
Aulologoas grabs and
Cod liver oiL EPA (L 3 gldayl
Markedly decseuu cd
No difference
hypedipidereia
oossnal lbidmessl
Protochot std Maxepacomccnuntrs supply ISS _irnsacenaenoic acid and 12Gi,laaselesaenake acid by weigh : C n crnleslerd: DHA = docosaleetoendc
acid ; EPA = esosapenaenoic acid', HDL-C = high density roprouin chokslerol LAD = left 01551501 descending s erallary sorry : LDLC = low density
Spcgrorein chol-sterol . NZ = New Zedcod ; For = reference stuns' en ts''HHL , ode otrbe H0eilahle Hypcdipidenic rabcR-
supplemented diets demonstrated a 50% to 6600 reduction in
the surface area of the aorta, coronary arteries and common
carotid arteries with atherosclerotic changes compared with
animals not receiving N3 fatty acids, However, there was a
signiScuayt but unexplained increase in atheros clerosis at :is'
carotid bifurcation .
Studies In tower species . These have yielded conflicting
results. Moderate doses offish oil have had inhibitory effects
on atherosclerosis in Japanese quail (87) . but these effects
were not observed at higher doses in spontaneously athera-
sclerotic strains (88) . In rabbits, studies with different de-
signs have shown a decrease (89,90), increase (91) or no
Rel-(Is! author) Species Model
Weiner(aII Pig Uolaryhypedipidemia
and ball- abr -
Kim (82) Rg Dietary hyperlipidemia
Harlot (III Pig Normolipidemia ; Teaon
consrridor on LAD
Shinlokarea 001 ME Dlelary, hyperlipidemia
said balloon b- i.
V* is 185) Rheeos Dietary hypcdipndema
monkey
HoaanderInMI Cynoelulgto Droraryhypcdlprdoxa
Chamberlain (871 Japanese quad Oidary hyperlipidemia
Chambedain 188) Japanese quail Dietary hypalipidemia
fpenemlly susceptible
Slash)
Zhu 1991 NZ rabbit Dreary hyperlipidemia
2. Medium dose (2, sl L.v1
r .High dose 0 m .',Iay)
Bolton-Smith NZ mbbil Semm sickness Fish oil (type aospeerliedl
(70) 120 gykg)
Thiery (911 NZ rabbit Dlelary hyperlipidcmia Motepa cancrnlrsS •
12 mpday I
Rich (92l WHHLrabbil Gcneochyperlipidemir Maaepaonax,n •a,es
12 .5 mIdayl
Stare' (951 Rut Hewn transplant Super Mascps :nc.rtate
I2mpkgpe 1,, ., Lea,
130 rt 1sg on it I) DHA
Icy mg4,g per ivy(
Rogers 194) Re I . Dsteely Matepa cmxen. e r I W mglkg
hyperlipidemia per dayl
2. + Hypothyroidism

8. J- N 1592 . Ira-sn
change (92) in tote[ acnic atheroscIerosis with dietary Osh oil
suipydementation. In one study 1931 . either fish oil or veta-
pa.rlil appeared to cause regression of atherosclercsis in
tdhhi returned to a normal diet after a period ofcrdekstcrol
fe. ding but them was no additive effect when the two agents
curie contained .
24
One rmusual sends• in rats 19ili examined the effects offish
a,l supplementation on monocyte adhcsinr nd foam cell
formation. In normocholesterolemic rats seep : . .mented with
fish oil, no difference was found in mane,[° adhesion or
foam cell formation compared with control - :s Rats made
hyperchoie oerolemic by cholesterol and choic ;,,:to feeding
exhibited a similar degree of monocyte adhesrcrr bet those
-.-hose diet was supplemented with fish oil hat a twofold
ncrease in foam cell formation . Another group of rats was
.made severely hyperchoteste,c iemic by feeding s . ith choies-
Ierdl and eholic acid and the addilien of the anlitt yroid agent
-Ihiouracil . In this group animals on a Fish oil-suTplemcntcd
diet exhibited a fourfold increase in monocyte acheston and
Withal foam cell density . Allhuugh pruvocalis° . this ~ tfecl
may be species specific or mediated in part by an Interaction
with fish oil and hypothyroidism induced by 2-thios.racil . In
the previously discussed study by K :ni et al. 182 , in the pig
motto! ° s PI-fact animals evhibned signiiicam!v fewer
manocytes within the atherusclerolic lesions than did the
control animals.
Another radii 195J sita„vd Henefi, ial ifesfr of NJ far(y
acids in the rat cardiac transplant made! . Fish oil-fed rats
were computed with rats fed saftowcr oil plus aspirin and
dipyridamole and control rays fed a standardr laho-atory 3ict.
When compared with the other groups, fish ml-fed rats
demonstrated fewer diseased vessels and a lower mean
arterial disease grade but no difference in transplant rejec-
tion, thus suggesting a nonimmunologic mechanism . The
safflower oil/aspirinidipyridarnde group showed a degree of
Atherosclerosis similar to that seen in control rats.
Finally. in a hypertiptdomic caaiae model, several studies
M-OF) have demonstrated that the addition offish or fish oil
to the diet resulted in significantly less intimal hyperplasia in
autotogous femoral artery vein grafts than that seen in
control grafts . Although one study 196) found a similar but
smaller protective effect of aspirin and dipyridamohr in
preventing intimal hyperplasia, another study (901) compared
fish oil, aspirin, an experimental thromboxane synthctase
inhibitor (CGS-12970) and the cembina(ion of fish oil with
aspirin or CGS-12970 and found that only fish oil prevented
intimal thickening perhaps because of a weak reduction in
serum growth factor activity. There was no additive Ienefit
when fish oil was combined with platelet inhibitors . Using
synthetic femoral artery vein grafts 4 mm in diameter. Canal i
et al. (99) demonstrated significantly less neaintimal thick-
ening in dug% fed a mackerel duel supplemented with men-
haden oil than in dogs co0SUmillg a rcgutar diet with or
without aspirin and dipyridamole . Craft pat nncy rates were
similar in the groups receiving suppleme,itat fish oil and
platelet inhibitors and significantly hurler thin those in dogs
ISanFJ .tNOCURLIN
191
FISH OILS IN THE PREVENTION OF ATHEROScLEROSIS
receiving neither; however, neointimal growth was inhibited
city in the dogs receiving the fish diet .
Studies With Ciiufcai End Points In Humans
Postangiopfasty Hestenosis
Gala are now available for analysis from five studies
(i00-1041 of N3 fatty acids in human pustangioplasty reste-
nosis. Restenosis was significantly lower in two studies
1100.701) and unchanged in two others (102.103). For unex-
plained reasons, the final study (104) showed a significant
iimitat!on in restenosis in a group with single-vessel coro-
nary disease but no difference in patients with multivessel
diseak;, In some ways angoplasty serves as an ideal model
for the study of intmal hyperplasia . but it it important to
emphasize three concepts before drawing conclusions from
these studies :
I .From moiecafar biologic study, it has recently been
sMwa that activation of genes by platelet-derived growth
factor occurs within 4 h o: angioplasty . ribonuctele acid
and protein synthesis begin within 24 h and peak within
the tsr week (Tauhmen -f, persona ; communication).
Tuns . restenosis represents a heading response to injury
dial begins immediately.
2. N3 fully acids exert muay of their rffects try becoming
incorporated into rdtldinrmembranes in competition with
other dietary fatty acids, thus requiring a loading period .
A reasonable approach to ensure maximal effectiveness
would he to begin the administration of N3 fatty acids
about 2 weeks before angioplasty whenever the proce-
dare is elective.
3- Re.aezmnia rare., can testy he compared benvera studies
using a standardized definilion of restenosis and with
angiographic follow-up. Use of clinical criteria or stress
testing may underestimate or overestimate the netted
incidence.
Design of posta-.aJngt4 y strdles oft N3 fatty acdds. Of
the poslangioplasty studies cited earlier, only two (100 .102)
included angiographic follow- , 1p in the majority of patients.
In the study of Debmeret al. (1 W 682 men with 103 coronary
lesions were randomly assigned to receive 3 .2 glday of N3
fatty acids of placebo, in addition to aspirin and dipyrhda-
mo!e. 'therapy was begun I week before angioplasty and
continued for 6 months. Compliance with lhcmpy was
assessed by detmninalion of platelet membrane fatty acid
ccmposilien and restenosis identified by repeat coronary
arplogmphy . This study showed a significant decrease in
reslcnosis in the tnealment group whether analyzed per
lesion (36% vs. 1690) or per patient (46% vs . 19%) .
In contrast. Grigg rt al . (102) also employed routine
ongiograr File follow-Up at 4 months but found no difference
in restcnosis in patients treated with 1 .8 g of N3 fatty acids
daily beginning the day before or the day of angioplasty.

9. !a2
ISRARL ANDCORLIN
FISH OILS fN THE I VCt YION OF ATHEROSCLEROSIS
Although their study failed to confirm a beneficial effect of
N3 fatly acids. the results emphasize the importance of
angiographic follow-up in accurately doteeting restcnosis
that was angiographirally present in only 51% of patients
with recurrent symptoms and in only 55% of those with a
positive exercise test . In contrast, 38% of patients with
restcnosis were asymptomatic . N3 fatty acids were begun
before angioplasty in two other studies (101 .103) ; in one
(101), there was no benefit, and the other showed a reduction
in rc=-tenosis from 35% in control subjects to 22% inpatients
receivwg supplement-. o(4 .3 giday of N3 fatty acids. How-
ever, the results or bath these trials are weakened because
each employed a stepwise fallow-up in which repeat angiog-
raphy was performed only for clinical symptoms or an
abnormal exercise test .
Rcsr metdations. Because o5 the inadequacies in design
of the current studies in light of the considerations we have
outlined, the present data are inconclusive. The heteroge-
neous fatly acid composition and cholesterol content of the
fish oil supplements used in the various trials may be an
additional confounding factor (67) . To test adequately
whether N3 fatty acid suppementation inhibits postangio-
ptasty restenosis, further randomized, double-blind, place-
ho-controlled clinical trials will need to be conducted, in-
cluding studies in patients with both stable and unstable
anginal syndromes Adequately designed trials must
allow n, sufficient wash-in period for the study drug ; this
may be assessed by biochemical analysis of fatty acid
content of platelet membrane phospholipid Determination
of the restcnosis rate should be based on angiographic
follow-up .
Secondary Prevention of Myocardial fnfarcrkun
MI. and Retduretion Trial (DARI). A very interesting
recent trial (105) suggests that steal increases in dietary fish
intake may prevent secondary mortality after myocardial
infarction. In a trial involving more than 2 .0(10 patients, men
were advised to increase their dietary intake offish . increase
their dietary intak of fiber or decrease their consutttpuioo of
total fat while emphasizing the r-se of polyunsaturated veg-
etable fat. Because the trial involved a factorial design, all
possible combinations of these interventions were tested.
After 2 years all cause mortality v;,:s 29% lower in the "fish
advice" grnnn and this decrease wa< entirely altr sutable to
a reduction in dnaht from ischemic heart disease . Curi-
ously, no decrease in nonfatal Ischemic heart disease
events occurred in any group. 9 rvival in the "fat advice"
group was unchanged despt; a small decrease in total
cholesterol and an increase in HDL cholesterol . Survival in
the "fiber advice" group tended to be worse than in other
groups, an effect that was not statistically significant and
thus possibly due to chance. Compliance was monitored
primarily by dietary questionnaire, although subsets of pa-
tients had 7-day weighted dietary intake records and mea-
surements of plasma fatly acids that suggested adequate
5ACC vol . 19 . No . I
tatwa7 1992:174-67
compliance. This DART trial (1051 involved only modest
increments in fish intake (about 301 g of fatly fish or 2 .5 g of
eicosapentacnoiC acid weekly) and so corroborates results
from the Zutphen (10) and Western Electric (11,12) .md :es,
which also noted beneficial effects of relatively small quan-
tities of fish.
Adverse Effects
Relatively few adverse effects have been related to the
consumption offish or concentrated fish ails . Mild decreases
in the platelet count are common but clinically significant
thrombocytopenia is very rarely seen, and then only at very
high doses (106).
Increased Mood sager In dia!relk patlentx . One study
(1071 showed an increase in blood sugar levels associated
with N3 fatty acids in patients with naninsulindependenl
diabetes. This effect has not been described in patients with
insulin-dependent diabetes, in whom fish oil administration
has been associated with some favorable metabolic effects,
such as partial normalization of the transcapillary albumin
escape rate (108) . However, pending further study, caution
is warranted in advising heavy ceostn .plion of either fish or
concentrated fish oils in nonimmlin-dependent diabetic pa-
tients .
Carcinogenic effects . Concerns about the possible card .
nogenic effects of fish oils have not been supported by
experimental or clinical evidence (106), although some ex .
perimental evidence (106) suggests a possible antneoplastic
effect .
Coatamhalbn with (omen and heavy Ineta s. Finally, in
advising a long-term increase in dietary fish consumption,
possible contamination with heavy metals and organic toxins
is of concern . Recent evidence (109) has documented swat
fish coughs in contaminated waters may be an important
source of human exposure to dbenzafitrans aid dioxins .
Further study will be needed to catalogue the levels of these
and other toxins in various species of fish and their oils and
to assess the biologic consequences topersons ^:ho consume
large amounts offish,
Cnnrlus:ons. Epidemio'ogic studies provide suggestive
evidence that fish consumption may decrease mortality rates
from coronary heart disease. This is supported by a wealth
of experimental evidence showing a beneficial effect of N3
fatty acid consumption on clinical and biochemical variables
implicated in the pathephysiology of vascular disease . Ex-
perimental studies of atherosclerosis in pigs and monkeys
point to the clinical relevance of the in vitro results but this
has not been consisteraly observed in other animal models .
Studies of the impact of N3 fatty acid sup ;lllenentation on
postangioplasty restenrisis are inconclusive bcccuse of the
heterogeneity of study design .
The hypothesis that N3 fatty acids inhibit the wherosCle-
rotic process is sufficiently supported to justify its further
study in humans. Although a rc :ommendation may be made
for moderate increases in dietary fish consumption on the

10. IACC Vol . I9, No. I
JW 0111992 174 -95
basis of the available dam, particularly in individuals with
known or suspected coronary disease ur after myocardial
infadclion, it is not ycljustified to recommend supplementa-
lion with Concenlrateu fish oils pending fat 1 :lor cvalaati-)n of
safety and efficacy.
References
I . ThoSmteof Hahh in GrernlanA . .Snn ie! re;nn from ho CNet Medical
Often in Greenland for the years 19'3,1974 . 1975 and 1976. MinisHy of
(1reenland, 1978.
2. Gronlad 1911. Ministry cf Greenland . 1919.
3. Bart H0 . DectNmg 1 . Ni-hen All . iota It,1id and lipoprotein paern
in GrceN18,1w WCiI C0ON Edumnx . 1„Ooel 1971 :1 :1143 .6.
4. Bang Ho. 7yefnerg J . Plums 5pd, anti lipoprolnas in Gmeni ,oIaa
West Cwsl rokirros . Acct Med Stand 1972:142 :P5-94.
5. hang110 .Dprbergl,SiadurHI.CT'hecomnwhiunoftheEskoaofood
in North Wes!Tn Greenland. Am 1 Con Null 1980 :334657 .61 .
6. DyerbergJ, Ban ; HU . lfemoslauc funeiionand platelet pllycnsaluated
tarry acids in EnF. .' .m. Larxa 19792:431-5.
7. Dymbecg 1. scars Ha . SloTerxn E . Monacda S. Vane 1R. Eicosepen-
lamoic acid and pevoo Lion of Ihrrmbosis and aahetdxlerosis . Lancet
19784n 17x1.
a. Key, A. Sewn Countries: A Mull-- A,aly- of 13-1, ,it
Coronary H, . Disease Ca)btidge, Man,. : riarvud Umveailp Press •
1980 .
!a. K0awa Y. NisNaawn N. 5uzuki M . Eicoupolyenaic acid of umm
11,101 a laparase islanders with low incidera of cxdiovauuar dis-
eases.1 Icar SO Vilaminot tTokyo)19Hl:28:441-53 .
10. Krtnshout D. Bouchkler ES. Coalaader COL. The inverse relation
baween 6sh conaarpton and 20 yea rlwrtality tram coronary howl
disease . N Engli Med 19X5:312:1265-9 .
I I. 91.1111e RB. Shryodc AM. Paid 0. Din . serum cholesmmi amt d- h
f wranry heart direase: the Western Electric sadv . N Engl 3 Mod"n,
;301'65-70-
12. Sholidle HE . Paul 0, Shryock AM, Slander J. Fish conmmpion and
-.talky from Coronary heart disease (khan. N Engl I Med 1985313:
X20.
i3. Cuts ID. Reed DM . Lenorlo the editor . N Engl . 'Aid 1985913 :821 .
14. Vo1ba SE, Iknch 1, B;elke E Latter to the editor . N Engi 1 Mad
1985.313:X20-I .
l5. Rahaewileh III Ciirnccal and other ob,ersahnns on Canadian Eskim•
in the 9451e7n Arctic. Can Med Asses 119316 :)4:497-501 .
16- Si- W, Sdtaer B . Bodily B, Rah P. Kurzmann I . Weber PC,
P6ndet-ibi,ane fatty aids, plaekr ag(Bepation and thromboaane
famari0n dsrng a mackerel dkl . Ia000t 15X0;1 :44744.
17. Gnodtdghl SH . Haul, W5, Fsmar WE, Pffens ofdietary omega-3 fatty
acid, on Ilatald compo,ilion and tanctinn in nun: a prospective
-milled study. IRood 1981L57tXi0-5 .
18, Fataa Y . Hhai A, Terano T. a a1 . Cloud and epdemabgic nudxs
of eirrnpedeernic arid (EPA) in lawn . Prig Lipid Re, 19X6 :25 :461-6 .
19 . Fxk! 5 . Weber PC . TiuomMxane AJ ITU3) io fooled in human
palelms Zen dietary eicosapemamek aid 1C20 J oino o S1 . Btochem
Bilphys Res Canon 1983:116:1091-p.
20. Von Schacky C, Fodrcr S, Weber PC . Long-teal effects of dietary
moan-3 faly acids upon plasromld eelluarttpidn, platelet funclion 7.M
ronnandd ftxmatom in man . 1 Olin invest 1995;74 :1626-31 .
21 . Van Soacky C, Sicm W, Fhcher S, Wetter PC . A comparative goody of
eicosapemamoie acId metabolism by human plmakls In -vivo and In-
"tro, I U* Re, 1985Trr.457-64 .
22 . Ktupp HR, Rtey IAG. Al sadiiini P- FiloGe.ld GA. Jr-
vi-d-- plarrktard .. .ha fitnniarduri9fish udl adar id unaGon in patients
with atkraxlaosis N Eno I Mod 198&314 :937-42.
23 . Nndknrmr P . Ran A, Minkes MS. Per(enddll IA. Sprecher H . Tittle,
prnsapandas: prmlocydn and UranboxaM biosynthesis and unique
biologod pmparfes . Proc Nail dead Sci USA 1979 :76:944-8.
14. Gygkwski RI . Sainloc IA, Uhtuber Ps, Wemkdy BC. Moncada S .
Vane JR . ERatsof allots 5-9-11-14-17 eirrnapemaeauc acid and PGH3
on plaak1 aggrtgdion. Pyoslagtandins 19918:453-78 .
ISRAEL AND GORLIN
1 8 3
FISH OILS IN THE PREVENTION OF ATHEROSCLEROSIS
25 . Dyerborg J . Jagencon KA . The egad of arachidotic and eico,apem
taeaob: acid on the synthesis of prostaglmdin-like maleral in huan
umbilical vasculature . Angry 1900bd2-7 .
26 . Oukcr S. Wetter PC. Prom nalandin 13 ie fumed in-vial in man lrr
dietary riciwpena-nnir mid Nature 1984 ;307:165 .8 .
27 . Fi!zGerald GA. Smith H. Peat- AK. Btash AR . inaeascd prosla<:y-
din insymhesi, in patients with severe athemsdrrons and plaulel
:,aiivauun . N Pro I Mad 19X4210 :1055-Il
2E. Hamberg M . Transformations d 5.6.11 .14 .17 ekosapentuundc acid in
human patdets. Btochim Biophys Solo 1980 :618 :)19-9X .
29. Ixvine PS, Fi,her M, Schaeider PB, d at. Dietay stippknrenaii n will.
omcgad fatty acid pmlnngs p1melel survival in hvptrlipiaemic patients
.suh alhemsckmsis. Arch Inrem Med 19896491113-A.
30, Mull : :neKM .Salmon IA.Knemerl.Leillowtederivedmelaholitesof
arac6tdonic mid in iuhemia'induced myocantiol injury . Fed Pree 1987;
46:2422-33 .
31 . Lawis LA . Austin KFThe biologically active twkmrienes. I Clin
Invest 19r473:1019-97.
12 . t-rn 1) . Herd LM, F-16 . G, Gddstein RE. Cormay cmetrat'
b1' Ieukoiriu,e C4 . D4 and EA in the in,acl p g heart . Am I Cardiol
1183 .51 :!451-4,
33 . Michelassi F. Laana L . Hill RD . a d . Leukmnene D4: a plaint
nn.,ornery v7.
rated wkb impaimd ssngkubn
cannaaion. $knee 1942',217A41-3,
34. nurke IA . Levi R. Guo ZG. Corey El, Leukarierres Cl. D4 and E4:
effect, on human and guiuee pig cadac preparacaa in.vi!ro. I Pharnn-
wI topTher 1982:221:235-41 .
3$. Murphy RC, PRkett WC . Cap BR . Lands WE, Teinede and peolaatt
kukotnenes: selective production Craw mane ntaslocytotracells after
dietary manipulation. Prosaglandin 1981 ;22:613-25.
36. Inn TH . Isrml E, Drarrn JM . Enhancement of plasma kveh of
biologically active kuharkra B compounds during anaphylaxis to
guinea pigs pretreated by hidnmelhaoc or bye MS dl enrkkd diel .
J immanol 19X6:1362577-82 .
37 . SrrasserTH, Fischer S . Weber PC. Leukotneoe 85 is cued in human
neulrophils after distaly supplemenlation with elrosaperaeuok acid .
Proc Natl Alan Sci USA 19X5 ;82,1540-3 .
38 . Pr-on SM,Denouement GA.Morriso.AR .Theeffects ofadielrkhin
fish oil on human 6euamplily identification of LIPS as a metabolite.
Prostaglandin 1985:30:209-27.
39. Len TO . HuoverILL, WiikamsJD.ttd .17,oofeaafdiaary endchmem
with eicosapenta micanddocosoltexaentdcacidsenin-vilmneutrophl
nod moaocylo kukarime gemrabn ad lea . 0pkl 100x1107.14 2agl 1
Men 1986:312:1217-24,
40. Schmidt El, Pcderson JO, Ianild C. Diced 1, Gunael N, Dyerherg 1 .
The effect d n-3 polyunauureed (arty midi an 1ip11, hatmontHik .
neutrophil and mnnoyre chemaaxia in osurn dependent nubile,
mellhis, I IMem Mad foppll 19109225 :201-6.
41 . Mull... KM, Read N . Sdomn IA. M .-La S. Role of leakneynt, in
rate myocardial infldian in aneslhaieed door : rcialionclip to noyo-
eardial salnge by anal-6Aamnrafay dal, . I Pharmecol Gap TMr
1984:229 :510-22.
42 . Culp BR, L otUs WEM, Luscka BR, Pin B . Runo 1. The efev: of
dietary wappiemeromiin elfish it on eaperirn .. .uiui air,5.1
lion. Proniaglatdina 1980 ;20:1021-31.
43. Hock CE. Holalm1 MA, oetbd DK. Effect of dienry gab al an
myocardial pcspholipds and myaeardial sekndut damage, And Phys-
id 1987 :25PH554-60 .
a. MatsukurT.Miyagaw7.M,Sab0K,MlneoS .YaeagaoraA,hhikawa
K . Efu ofdismay flrl old an tofarit size in 10Bne myaesrdal iaehemia
labstrl Circulation 1W11aNsupp 1111111.39-
45, Black KC . Culp B, Mail" D, Radall OS. Lands WBM . The prdne-
five effects of dkary Mail on fad ardrd id-tan . Pronaglatdins
1979;3.257-6X,
46. Me-ma, PL, Akynoadena MY. Chamock IS. Marl, fib dl pre .
were1 ventncu100 lbn5minn f411-ins coronary afar,. oalasan ad
reperfuston. Am Him 1 199A M, 709-17.
47. Fnggiouo A. Ross R. Harks L Studies of hypemhokctmokitpa in
ran-human primates. 1 . Changes that lad to fatty streak fanasnn .
Aneriosdernis 1984:4:323-9.

11. 194
ISRABL ANDGORLIN
FISH O1L61N THE PREVENTION OF ATHEROSCLEROSIS
48 . Ghwn AM, Tsukada T. Ross R . Human athem,derosi, II• Immunocyto-
chemical analysis or the cellular compasision of human aherswdercnk
Icrians, Am J Pnlhol 198625191-7 .
49 . Sprdino 111 Pol,, II Kylamh , KA, Lee TN, Louis RA, A-,wit Y.F.
The effects of N3 pnlyunnnwrred rally acid, on the V .- h- of
platelet ecloaling fatter 6y human morocvtes . J Immunol 1987 :139
4186-91 .
50. Fisher M. L, ., . PH . Wriner BH . Dietary NJ Noy dad suppkmertla .
It. reduces superoxide production and cftemiWminescence in a mono-
eym.enneaed poyWauon of lyephncytcs. Ain 1 Chn Nun 1990 :51 :
gut-0 .
$1. Goldberg ID. Slemermmr MB- V.achlar permeafnn of plat Cctor 4
After endothelial injury . Science 1940 :20ifiH-2.
'2. HayCPM .DurberAP .SaynorO .ERectufhitlnilonpanletkinetics in
patients with iscbmc bean die- I.m•ce1 I9w •.a . :3i,v-72 .
55 . Calluses; C . DiCarlum P£. Rnu R. 5010,10 501 . An endothelial
cell d0000d growth,aelor- 1 Cell flint IM. PISa67-7!,
54 . DiCorkto PE . Bnwen.Fope DF. Cul and endothelial eelis prm)ua a
plaleler-0erived growth fxsordike protein . Pros Nall Acrid Si USA
1983 :80:1919-23.
55 . Fox PL, DCndnn PF. Fob nil, inhibit endolnolml cell production of
fdalelet-de,ved krmvth faelord)ke protein. Seers 1980:2411455-6-
56 . Fox PL, D)Codelo PE. R,goIrlion of puodnclion of a plalekhderisrned
gonwlh fsmr-tike pontoon 'no caa :ond henine antic endothelial cells,
i Cell Phssin119S4;1I1:7S8-1178.
57 . EndresS.GhorhamR.Kell-,VF.elal.The .Redof.Lerarys, pknsen .
anon with .1d polyunsawreted W.) a crd. as the syathe,i, ofailerleu-
kin-I and mma, ton--osa factor by .noa I r cells. N Fngl 1 Med
ISO 1 520:165-71 .
58 . fSnarello CA. M- W . Commit co-pro: lymphokmes . N Foul I Met
1987 :317:948-5 .
59. 8.1k, B . Cerami A. Coehettin : more than a laimnr neenmn 1001,11.
N Engl1 Mrd I957;11b:J79-85.
0) . LibbyP.skimmer SIC . Friedm :mlB.INerkukin-I :amitagenforhumsn
?anal,, smooth mule ..It, that inlnret the rd"-f growth inh)bi-
Wry prentermds.! Clin layout 1990;01 :417-91 .
hl . E1,115 A. cpstilu N . Border w, Frundlich B . synenpsue stimulation 1,F
fibroblast prostaglandin production by mcomMnant lnInIelkm •I and
turner recresis fades I hnmunol 19W-139'381?-6 .
62 . Herold PM, Kinsella 0E. Fish oil consumplian and decreased risk of
a:dionn.eulardi.ease arompnr .nnnoffindingstramaninmandhumanp,
ding trials. Am I C 9 Nut A 45 :5611-98 .
63 . Harts WS . Fish al, and plasma lipid and Iipomorein memlmlism in
humans: a etilieal oeview . 1 Lip ., He, 198990985-007.
67 . 5chectman G, ..a: S, C :-_ ;1 GD . Lc,. M . Ki,ebt A . Can :he
hypotriglyceridemk effect of fish al concentrate t_ ...mined . Ann
Intern Med 1989 :710946-52.
65. Sanders TAB . kbstry M . Controlled trial, of fish oit supplemen, on
plasma lipid cancrntration, . Br I Clin Pmt 191'1.1978-81 .
66, Saynor P . Verei D. Eicosapenlaennic arid . bkeArg lime and semen
lipid, . s .ancd 198•_ :2 271-4 .
67. Babe k 1, Lindgren FT. Rudel LL, Effects of atoned and polyuncnl
orated dietary fat on the concentration of 11DL subpopulat .ons in
Afncan green monkeys . Anenoscleonis 19858'. 77-32.
60. Roach P . Kamhouris A. Tumble R. ToNI no DL . Newel P1. The effects
of ditlaty fish oil on hecallc HDL and LDL Iipoprokin ree :ptm
activities in the m . PEGS Loll 19RT,222:159-62 .
69. Nm4ki S . Kb. M . Takeem . K. Matsurawa Y . Tort S. Effects of
purified eimsapanl'nennic Aid dbyl ester (EPA) on lippr0teia canto
,inon.lab,rrl-Cnulullon 1YJge2lsupp11111 :111-077.
70. Shimdcawa H. Vanhouue PSI. DSetary cod-list, Oil irpro,es endmhe-
Bum dependent responses in hyperchdnclonokmi and athems ;leronc
po:cax ""nary nouns,. Circulation 196'9:78:34'_!-3^ .
71, Vekshae)n VI . Ymng AC, Vna J A. cl al . Fish 611 impmsesendolhelium-
deperrdent nelaaation in patients with comrury nanny disease lah,1r,
Circulation 1999:80tnuppl111:11-a34.
72 . Fleishhaam Fl . Lee TC, Nrllessen U. Fischell LL Fish oil improves
erlnnhehium-drpendem corona,,.• -dRaaion in eardiae transpbut nr
eipknts iabsn1. Cnulatinn 1990 :elsupp11111:111-468.
1ACC V.I . 19, No . I
lanuany 1992:174-85
73, McMo J. Law- D, Snldeen T. Reduclion in plasminogen adivabr
inhibitor) IPAI-11 with omega-3 polyunsaturated fatty acids 1PUFAs
intake. Am Head 1 1%88:116-181-6.
74 Hnstmad. AT H'rr' .odal T, Kicrulf P. Flown H. Ulshsaeo K. Fish oil
amt plasma fib...noun . Br sA.d 1 196A9n•l en_I
75 . Radack K . Heck C . Huster G. Dt.tavy suppkrronlannn coal, baJ,.n
h :h oil lowers P.brinogen levels: o randanh ad, doable-blind controlled
sway. Ann Intern hied 1135:1 H^57-R.
76. Clifton PM. Coboc L. Abbey M, Ne,tcl Pl . 178fermg'8em,al 1.11 and
fish oil in hypnbpldemic mcnlabsbl . Cin,Iah,an 1995,81IIsonpl IID:III-
476
77. Knapp HR. FitoGelald GA . 1 k antihype0ensose offal, of fish alt a
controlled study po[yumaturaled fatty acid supdemems In estiem)d
hypertension. N £ngl l Me. Pl1Aa321911197-a3 .
7K Bona. KH. Hit- KS. So-aume B . Great IT . 77rnte D. EOec, of
eicosapentaenac and docosuhenaeno :e ae:ds an tried pressure in hyr
pert-on, apopulatwn-based inlen'eelion trial from t6eT .omso Study .
N EnidiMod 1990:322:793-801 .
79. Morlennn JZ.5cl nIl ER, Nielsare Ad, Dyerberp J . The effectoaf n-h
neat 1,-1 p,lyoustra ed ratty acids as henpslasis, blood lipids and
b'-alpressure.ThrombHoe- 1901,95211-6,
00. Von Housdiraten R, Nods A. under Seek E, Haul.mulier U. duhletz
M . Hamstr (I LT., M a moiemre fish )make on blood prrssuon,
Meedlng slate hemaroiogy and clinical chemislry in Whiny teaks . A . J
01n Net, 14M .46,414-36.
NI . Went 11111,1kkeneIS .IxsinePH .01ol lihrhllronafalhemsderosia
by cod-11- nil in a bypetlipidemk swore model . N Engl 1 MM
1116:513lilt_6
87. is r Dia . Ho HT. Lonlo rice DA, Sdunne I. Thnvnas WA . Modification
or Npoprote.in pallems and reran nnian M arherogenesis by a fish 1
supplement to a hypedipidemk diet for switn . .AllKmukroais (989 :76:
35-34 .
83 . Hmtng 1M. (amen, 1MJ . Enrol CE . 5chalkwijk WP . Venbuw PD,
Does platelet aggretaliun play a onk in the nedn otion in luakaed imimal
probft moan in rormolipidemic pigs with fixed canocur9 aneiry sunosis
fed dinary fish 08n Askaoxlerasi, 1389.76.79-a8.
04, Sassen LIMA, Hadog JM . 3amers JMJ, K!ompe M• Volt WcOrknA, U,
Vandeuw PD. Mackerel oil and atherosderosn in r `; . En' Hewn !
1989:11038-16 .
85 . Davis HR. Brldeneline RT. Vesulinovnch D. Wisskr RW. Irish cid
inhibilsdeselopnent of atheroselerosis in rhesus monkeys. Artuiosek•
ro,i, 1967 ;?;44)-9.
96. Hollander W . 1idrg S. Kiekpalrkk 1.1. Lee A. Colombo M . Pasty S.
Differeaiul 617.01, of Sob oil sapplemmds on arherosclerosis lab-.
Circulation 1987:76)soppl IVIIV•3 13 .
87. Chamberlain JO . Falkel M, Piazn T- Rcdaorien in chc!eaerol. trigly1-
crides and early coronary artery plasma, formation is ader :nckmsb
prnle Japanese glad fad a menhaden fish oil ,.ppleormics! dim. Anal
Rise 1906 :20:214-6.
88. Chamberlain 1G. Bolton C . Effects of btm Iesm cowvmplio„ offish oil
IMaSEPAI on semen lipids and arterial uhrasnuemne m lapmtcnu Wail .
Athmasderosis 1997 :61 :95-IS) .
89. An 0.Q, Smith DL . Sieven RE . Isenberg WA1, i+-1m, lit 0 .1 . hih :"
lion of athemscterosis by fish o)1 in cholecmmHed rabbi . J Am Coll
Cardta I5OS :1*_f0T1-g .
90 BohooSmiihC.GibnoyMl.Csallutt!n,P1,Jewe1R.HillinnK.Effeclor
Polyunsaturated fanyea)dc at then-3 and n-S comics an lipMmmposat)an
and eiwsaooid syndtesu of Plat and aorta, and nn immunobgic
blue ies of athemsdmasis in rabbits- AlnnmsckloMS 1985 ;72:29-35-
91 . Thierry J. Seldei D. Fish oil feedirg rcsulls in an enhaacemeel of
choks-l.irdueed athemsderosis in rabbits . Albensxkmsis 1987 :61
53-56.
92 . Riot 5, %IilkrJF, Charaus 5 . et rd- tkvelopmere of alhemsei.asis in
fnen)cally hypcnliplJcmi rabbits dining ebmnic fish oil ingestion .
cLklonlerasis 1909:9:109-9a.
93. Zhu B Q . She- RE, Isenberg WM . Smith DL. Pamadey WW. Regrea
con of arhs no-li nosis m chdearenol-fed rbbim: effects d' fish al and
s empamil .I Am Cell Candid 1990:15:231-7.
94. Rep. KA, Karossks Ml, Dklary 111b oil enhaloes nmnocyte adhe-
sion and rally st :na§ formation m the hypercboles :eroleni nt . Am 1
P01171 199 :112:752-N .

12. IACC Val . 19, No. I
January 1992:174-05
95, Sands GE, Milchell RS . Billingham ME . Glassun 28. Cahill PD. Milky
DC . luhibifion of o_eel 'ed card oo A!o0ruft atrrv:c!er^lip by fish nil,
I'Iloac Cardio- Sung 1999 .97 :841-55.
96 . landymoec RW . MacAulay M . Sheridan B . Cameron C . Comparimo of
cod Iiva it and avpinrvapyridovo4n fur roe prevemiun of tCt:tttt
hyper era in am-luas vein crafts. Ana hour Surg 1986,41 54--7.
97 . Cahd1 PD, Saris GE, Cooper AD. ei al. Inhibition of celn grab Idimat
thickening by eicwapenranoie acid : red-d ahrdmbanane production
witl.nal change rn lipoprotein levels cv low.dersiry lipoprotein receptor
Jcnnly . J Vas c_.ng 1915,7aOg-Il.
95 . 50006 GE. M09hep :4S . Biflinghare ME . Ma.hanisme rcepunsibk f r
lnhihibon of seingrft anenorolerosis by 6sh oils . Cinculalon 1989:
SIXsuppl 11:1 .109-27 .
99 . Casa11 RE . Hale IA. LIN., L . IF- F . Mi., MD . Improved grill
potency associated with akered plaldel fane0en induced by marmot folly
cony in du*. I Surg Res 1956.406-12 .
Ill De rover Gl . Pp-J1, ven den Rerg EK. eI al. Rdaclian In the rate of
,,dy reaenasis after coronary anyaopdasry b, a did supplemented wkh
a 1 fan? acid ;. N Eel I Med 1959:119 - 3 "D
101 . a.ilner MR. Galling RA. Ldhngnnll A. cr d. Usefulness of fish ail
tupplerornn is
paavmdng clinical eviJence of reslenvnv elm, percna'
morn Ransluminal coronary angiodaety, Ant J Cardinl 1469 ;64:29aA .
102. 0, go LE. Kay TWH, Valentine PA. n al . 1%eeennin:ms m reslenous
and lack of nBul o1-dietary yu ppkn#nlariod v ith eicmapentaeroic acid
ISRAEL AND GORLIN
185
FISH OILS IN THE PREVENTION OF ATHEROSCLEROSIS
on the incidence of coron ary satery nan04s after angtoplanly. J Am
Cull Cardiol 1989:13:665-72.
IUJ . RoivGJ .SippedyME .BouchrrTM .Ital.Rarrdomisedtrial of6,hnilfor
preventiod of MOO- alter -nary angiaplasly. Lever 19892 :177-
51 .
184 . Slack 10 . Pimkenan CA. Van Txsel1I
it al. On orrI fish of supplement
t
leansis after percmaneous lnnsluminl axunary angio-
plasryrab mleans i
s
srrl7 J Am Call Carwol 1987 ;9:64A.
10J . B•_ rrMi .,FmfiyAM,Gi!f--rIF,dal .Ededaor,hangvuinfat.fish,and
fibre Intakes on death and my-dial rinfarclion: diet and reinfaredon
Ira /DART, LaoceR 1989:7:757_61 .
106 Gaodnighl SH, Fisher M, FdlGercId GA . Levrne P. Assessmenl of
therapeutic use of dlelary fist dl In alberoskrork uascalar dictate and
thrombosis Ches11W9 ;9SIScppl25:J9S-255 .
107 . Glauber H . Wallace P. Gunny K . &echlel G . Adveelt me .atloiie eS'ecl
of ome&.3 fuly odds in Imndnsulm dependent dlabnes me5incs . Arn
Intern Med 1988:105:663-0,
luI. Jeeren 'I', Slender S, Gajslem D, Halmer G, Decked T . Pineal
nmmalimlien by d etary rod-lienroi of inceeased micravasuler cilia .
-leakage it patieatswith inmGa-0ependenr dicbeles and alhuro mre&.
N EMI I Med 1959 :321 :1572-7,
109. Svensson B ., Nil- A, Ilarwon M, happy C. Akr,non B, Skeriving
S . Eapoccrc ro dionds and dibenynranns I eonih d:a consumption of
fisb . N Erg] I Med 1991 ;724 :5-12.