Purified eicosapentaenoic and docosahexaenoic acids have
`differential effects on serum lipids and lipoproteins, LDL particle
`size, glucose, and insulin in mildly hyperlipidemic men 1- 3
`
`Trevor A Mori, Valerie Burke, Ian B Puddey, Gerald F Watts, David N O'Neal, James D Best, and Lawrence J Beilin
`
`ABSTRACT
`Background: Regular consumption of n -3 fatty acids of marine
`origin can improve serum lipids and reduce cardiovascular risk.
`Objective: This study aimed to determine whether eicosapen(cid:173)
`taenoic (EPA) and docosahexaenoic (DHA) acids have differ(cid:173)
`ential effects on serum lipids and lipoproteins, glucose, and
`insulin in humans.
`Design: In a double-blind, placebo-controlled trial of parallel
`design, 59 overweight, nonsmoking, mildly hyperlipidemic
`men were randomly assigned to receive 4 g purified EPA,
`DHA, or olive oil (placebo) daily while continuing their usual
`diets for 6 wk.
`Results: Fifty-six men aged 48.8 ± 1.1 y completed the study.
`Relative to those in the olive oil group, triacylglycerols fell by
`0.45 ± 0.15 mmol/L (=20%; P = 0.003) in the DHA group and
`by 0.37 ± 0.14 mmol/L (=18%; P = 0.012) in the EPA group.
`Neither EPA nor DHA had any effect on total cholesterol. LDL,
`HDL, and HDL2 cholesterol were not affected significantly by
`EPA, but HDL 3 cholesterol decreased significantly (6.7%;
`P = 0.032). Although HDL cholesterol was not significantly
`increased by DHA (3 .1 % ), HDL2 cholesterol increased by =29%
`(P = 0.004). DHA increased LDL cholesterol by 8% (P = 0.019).
`Adjusted LDL particle size increased by 0.25 ± 0.08 nm
`(P = 0.002) with DHA but not with EPA. EPA supplementation
`increased plasma and platelet phospholipid EPA but reduced
`DHA. DHA supplementation increased DHA and EPA in plasma
`and platelet phospholipids. Both EPA and DHA increased fasting
`insulin significantly. EPA, but not DHA, tended to increase fast(cid:173)
`ing glucose, but not significantly so.
`Conclusions: EPA and DHA had differential effects on lipids,
`fatty acids, and glucose metabolism in overweight men with mild
`Am J Clin Nutr 2000;71:1085-94.
`hyperlipidemia.
`
`KEY WORDS
`Eicosapentaenoic acid, docosahexaenoic acid,
`EPA, DHA, hyperlipidemia, fish oil, n-3 fatty acids, lipids, LDL
`particle size, glucose metabolism, insulin metabolism, men
`
`INTRODUCTION
`There is considerable evidence to support a protective effect
`of dietary n-3 polyunsaturated fatty acids against atheroscle(cid:173)
`rotic heart disease (1). The 2 principal n-3 fatty acids in marine
`oils, eicosapentaenoic acid (EPA; 20:5n -3) and docosa-
`
`hexaenoic acid (DHA; 22:6n-3), have a wide range of biolog(cid:173)
`ical effects (1-3). Those relevant to heart disease include
`influences on lipoprotein metabolism (4, 5), platelet and
`endothelial function, vascular reactivity, neutrophil and mono(cid:173)
`cyte cytokine production, coagulation, fibrinolysis, and blood
`pressure (1-3, 6, 7). In addition, the effect of n-3 fatty acids
`may be dependent, to some extent, on the presence of underly(cid:173)
`ing disorders such as dyslipidemia, hypertension, diabetes
`mellitus, and vascular disease.
`n - 3 Fatty acid supplementation in animals and humans
`results in substantial increases in plasma and tissue EPA and
`DHA as well as variable incorporation in different phospholipid
`classes in different tissues (8-10). These differences may be
`important to the subsequent utilization and metabolism of EPA
`and DHA. Although both fatty acids are considered to be bio(cid:173)
`logically active, most studies have focused on the relative impor(cid:173)
`tance and effects of EPA, primarily because of its predominance
`in marine oils and fish species. The recent availability of purified
`EPA and DHA, however, has enabled studies of the independent
`biological effects of these fatty acids.
`Evidence from in vitro studies suggests differential effects of
`EPA and DHA (11, 12). In vitro (13) and animal (10, 14, 15)
`studies have also suggested that EPA may be primarily responsible
`for the hypotriglyceridemic effect of n -3 fatty acids. Rambjor et
`al (16) concluded that EPA is responsible for the triacylglycerol(cid:173)
`lowering effect of fish oils in humans, but their study had small
`numbers of subjects and was of short duration. In contrast, a
`hypotriglyceridemic effect of DHA was shown in healthy sub(cid:173)
`jects (17) and in patients with combined hyperlipidemia (18).
`
`1 From the Department of Medicine, The University of Western Australia
`and The West Australian Heart Research Institute, Perth, and the Department
`of Medicine, University of Melbourne and St Vincent's Hospital, Melbourne.
`2 Supported by a grant (Studies in Hypertension and Cardiovascular Dis(cid:173)
`ease) from the National Health and Medical Research Council of Australia.
`Purified eicosapentaenoic and docosahexaenoic acids and olive oil capsules
`were kindly provided by the Fish Oil Test Materials Program and the US
`National Institutes of Health.
`3 Address reprint requests to TA Mori, University Department of Medi(cid:173)
`cine. Medical Research Foundation Building. Box X 2213 GPO. Perth. West(cid:173)
`ern Australia 6847. E-mail: tmori@cyllene.uwa.edu.au.
`Received August 10, 1999.
`Accepted for publication October 20, 1999.
`
`Am J Clin Nutr 2000:71 :1085-94. Printed in USA.© 2000 American Society for Clinical Nutrition
`
`1085
`
`0
`~ :,
`0
`~
`(D
`a.
`cl
`3
`?.
`?.
`?.
`.Je!.
`0
`::,
`0
`c.d
`0-
`'<
`(C
`C
`(D
`~
`0
`:,
`0
`(D
`£
`3
`0-
`~
`
`~
`0,
`
`I\)
`0
`
`~
`
`ICOSAPENT DFNDTS00011026
`
`Hikma Pharmaceuticals
`
`IPR2022-00215
`
`Ex. 1004, p. 1 of 10
`
`

`

`1086
`
`MORI ET AL
`
`Grimsgaard et al (19) reported that EPA and DHA have similar
`triacyglycerol-lowering effects compared with corn oil placebo.
`However, DHA significantly increased HDL cholesterol, whereas
`EPA significantly lowered both total cholesterol and apolipopro(cid:173)
`tein (apo) A-I concentrations (19). Neither fatty acid altered
`LDL-cholesterol concentrations significantly.
`Possible benefits of n -3 fatty acids have to be weighed
`against the potential for impairment of glycemic control, partic(cid:173)
`ularly in patients with type 2 diabetes (20-23). However, studies
`in healthy subjects, in patients with dyslipidemia (24), and in
`patients with untreated hypertension (25) showed no adverse
`effects of n -3 fats on plasma glucose concentrations. To our
`knowledge, there have been no studies in which the effects of
`pure EPA were compared with those of DHA on indexes of glu(cid:173)
`cose and insulin metabolism in humans.
`In view of the increasing use of n -3 fatty acids in the diet as
`food additives or as therapeutic substances, it is important to
`determine the extent of any differential effects of EPA and DHA.
`This study examined the independent effects of EPA and DHA
`on fatty acid and lipid metabolism, as well as on fasting glucose
`and insulin concentrations. The study also aimed to determine
`whether EPA and DHA differ in their effects on HDL-cholesterol
`subfractions and LDL particle size.
`
`SUBJECTS AND METHODS
`
`Study population
`
`Mildly hypercholesterolemic but otherwise healthy, non(cid:173)
`smoking men aged 20-65 y were recruited from the general
`community by media advertising. Entry criteria included a serum
`cholesterol concentration >6 mmol/L, a triacylglycerol concen(cid:173)
`tration> 1.8 mmol/L, or both; a body mass index (BMI; in kg/m2
`)
`between 25 and 30; and no recent (previous 3 mo) symptomatic
`heart disease, diabetes, or liver or renal disease (plasma creatinine
`> 130 µmol/L). None of the subjects were regularly taking nons(cid:173)
`teroidal antiinflammatory, antihypertensive, or lipid-lowering
`drugs or other drugs known to affect lipid metabolism. All of the
`men had a usual weekly consumption of not more than one fish
`meal and drank <210 mL ethanol/wk. Fifty-nine of the 136
`subjects screened satisfied the entry criteria. The study was
`approved by the ethics committee of the Royal Perth Hospital and
`all subjects gave written consent.
`
`Dietary education and intervention
`
`All subjects maintained their usual diets and alcohol intakes
`during a 3-wk familiarization period. Baseline measurements
`were collected and the men were stratified for age and BMI
`before being randomly assigned to 1 of 3 groups: 4 g daily of
`EPA, DHA, or olive oil (placebo) capsules for 6 wk. Capsules
`contained either purified preparations of EPA ethyl ester
`(=96%), DHA ethyl ester (=92%), or olive oil (=75% oleic acid
`ethyl ester). All participants were instructed to maintain their
`usual diets, alcohol intakes, and physical activities, and not to
`make any changes to their lifestyle throughout the intervention
`period.
`At an initial interview, subjects were given written and verbal
`instructions by a dietitian on how to keep diet records, with food
`weighed or measured. The same dietitian monitored the dietary
`intake of all the volunteers at 2-wk intervals and ensured that
`usual eating habits were maintained. A 3-d diet record (2 week-
`
`days and 1 weekend day) was completed by the volunteers at
`baseline and postintervention.
`
`Lifestyle assessment and anthropometry
`
`Alcohol intakes, physical activities, and any medications
`taken were monitored every second week during the intervention
`by using 7-d retrospective diaries. Weight was measured every
`second week with an electronic scale.
`
`Serum lipids, glucose, and insulin
`
`Fasting serum lipids, lipoproteins, glucose, and insulin were
`measured twice at baseline and twice at the end of the interven(cid:173)
`tion. Serum glucose was measured with an automated Technicon
`Axon Analyzer (Bayer Diagnostics, Sydney, Australia) by using a
`hexokinase method within 12 h of collection. The assay precision
`for serum glucose at 4.9 mmol/L was 3.1 %. Serum insulin was
`measured by radioimmunoassay with an automated immunoassay
`analyzer (Tosoh Corporation, Tokyo). The CV for serum insulin
`at 21 and 102 pmol/L was 14.0% and 8.0%, respectively. The pre(cid:173)
`cision in the range of 234-720 pmol/L was 7.0%.
`Serum total cholesterol and triacylglycerols were determined
`enzymatically on the Cobas MIRA analyzer (Roche Diagnostics,
`Basel, Switzerland) with reagents from Trace Scientific (Mel(cid:173)
`bourne). The assay CVs were 2.2% at 4.2 mmol/L and 1.4% at
`10.5 mmol/L for total cholesterol, and 1.6% at 4.0 mmol/L and
`2.5% at 1.2 mmol/L for triacylglycerol. HDL cholesterol was
`determined on a heparin-manganese supernate (26); the CV at
`1.1 mmol/L was 1.9%. HDL2 and HDL 3 cholesterol were deter(cid:173)
`mined by using a single precipitation procedure (27). LDL cho(cid:173)
`lesterol was calculated by using the Friedewald formula (28).
`Serum for the analyses of lipids, lipoproteins, and insulin was
`snap-frozen in liquid nitrogen and stored at - 80°C. Samples
`obtained at baseline and at the end of the intervention were
`measured in a single assay to minimize interassay variation.
`
`LDL particle size
`
`LDL particle size was determined from LDL isolated by verti(cid:173)
`cal density-gradient ultracentrifugation of 4 mL plasma collected
`into EDTA (29). LDL particle diameter was determined by using
`a previously published method (30, 31) with use of commercially
`available 3-13% nondenaturing native gels (Gradipore, Sydney,
`Australia). Markers used were 28-nm latex beads (Duke, Palo
`Alto, CA) and high-molecular-weight standards (Pharmacia, Pea(cid:173)
`pack, NJ). Gels were scanned by Tracktel video densitometry
`(Vision System Ltd, Adelaide, Australia) to provide a quantitative
`estimate of the dominant peak size. Particle diameter was
`obtained from a standard curve of the logarithm of the diameter
`of the standards (latex beads, 28 nm; thyroglobulin, 17 nm; and
`ferritin, 12.2 nm) against their positions on the scanned gel. Asta(cid:173)
`tistical package was used to derive a regression equation that
`allowed test samples to be sized. The CV of a 26.1-nm quality(cid:173)
`control sample run on every gel was 0.8%.
`
`Plasma and platelet phospholipid fatty acids
`
`Plasma (1 mL) and washed platelets prepared from blood col(cid:173)
`lected into EDTA were extracted with chloroform:methanol (2: 1
`by vol, 5 mL). The phospholipid fraction was obtained from total
`lipid extracts by thin-layer chromatography by using a solvent
`system of hexane:diethyl ether:acetic acid:methanol (170:40:4:4,
`by vol) on silica gel 60 F254-precoated aluminum sheets (Merck,
`Darmstadt, Germany). Fatty acid methyl esters were prepared by
`
`0
`~ :,
`0
`~
`(D
`a.
`cl
`3
`?.
`?.
`?.
`.Je!.
`0
`::,
`0
`c.d
`0-
`'<
`(C
`C
`(D
`~
`0
`:,
`0
`(D
`£
`3
`0-
`~
`
`~
`0,
`
`I\)
`0
`
`~
`
`ICOSAPENT DFNDTS00011027
`
`Hikma Pharmaceuticals
`
`IPR2022-00215
`
`Ex. 1004, p. 2 of 10
`
`

`

`EFFECTS OF EPA AND DHA ON LIPIDS, GLUCOSE, AND INSULIN
`
`1087
`
`TABLE 1
`Characteristics of participants in the 3 groups at baseline1
`
`EPA
`(n = 19)
`
`Olive oil ( control)
`(n = 20)
`
`DHA
`(n = 17)
`Age (y)
`48.9 ± 1.7
`48.4 ± 2.0
`49.1 ± 2.2
`88.7 ± 2.0
`Body weight (kg)
`90.8 ± 2.8
`89.1±2.3
`BMI (kg/m2
`28.9 ± 0.7
`29.0 ±0.7
`28.4 ± 0.5
`)
`Waist-to-hip ratio
`0.94 ± 0.01
`0.93 ±0.01
`0.94 ± 0.01
`1x ± SEM. There were no significant differences by one-way ANOVA.
`EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid.
`
`treating phospholipid extracts with 4% H2SO4 in methanol at
`90°C for 20 min and analyzed by gas-liquid chromatography
`with a model 5980A gas chromatograph equipped with a 3393A
`computing integrator (Hewlett-Packard, Rockville, MD). The
`column was a BPX70 (25 m X 0.32 mm, 0.25-µm film thick(cid:173)
`ness; SGE, Ringwood, Australia) with a temperature pro(cid:173)
`grammed from 150 to 210 °C at 4 °C/min and with nitrogen as the
`carrier gas at a split ratio of 30:1. Peaks were identified by com(cid:173)
`paring them with a known standard mixture. Individual fatty
`acids were calculated as a relative percentage with the evaluated
`fatty acids set at 100%.
`
`Statistical analysis
`
`Diet records were analyzed by using DIET/1 (version 4;
`Xyris, Brisbane, Australia), which is based on the Australian
`Food Composition Database NUTTAB 1995A (32). Data were
`analyzed by using SPSS (SPSS Inc, Chicago) with general lin(cid:173)
`ear models to assess the effects of EPA or DHA relative to the
`olive oil group. Significance levels were adjusted for multiple
`comparisons by using the Bonferroni method. Values are
`reported as means± SEMs.
`
`RESULTS
`
`Study population
`
`Fifty-six of the 59 subjects completed the study. Two subjects
`withdrew because they were unable to maintain the schedule of
`laboratory visits and one subject withdrew because of gastroin(cid:173)
`testinal symptoms. Baseline characteristics of the 3 groups con(cid:173)
`firmed that they were well matched for the entry criteria (Table 1
`and Table 2).
`
`Energy and macronutrient intakes
`
`Evidence of adherence to the diets was from analysis of diet
`records and capsule counts. There was no significant difference
`in body weight between the groups at baseline (Table 1) and no
`significant change during the intervention. Weight changes in the
`3 groups were as follows: 0.2, 0.2, and 0.3 kg in the control,
`EPA, and DHA groups, respectively. Analysis of diet records
`indicated that total energy and major macronutrient intakes were
`not significantly different between groups at baseline (Table 3)
`and did not change significantly in any of the groups during the
`intervention. Alcohol drinking and physical activity were
`unchanged during the intervention in all groups.
`
`Plasma and platelet phospholipid fatty acids
`
`At baseline, there were no significant differences between
`groups in plasma and platelet phospholipid fatty acid composi-
`
`tion. The changes in plasma (Figure 1) and platelet (Figure 2)
`phospholipid fatty acids in each group indicated compliance
`with capsule intake. There were no significant changes in fatty
`acid composition in the control group.
`
`Plasma fatty acids
`
`In plasma phospholipids, EPA supplementation increased EPA
`by 494% (P < 0.01) and docosapentaenoic acid (DPA; 22:5n-3)
`by 87% (P < 0.01), without significantly changing DHA (9%
`change; NS). In the DHA group, DHA and EPA increased by 167%
`(P < 0.01) and 52% (NS) respectively, whereas DPA was not
`affected significantly. Oleic acid (18: ln -9) concentrations were
`significantly decreased by both EPA (by 11 %; P < 0.01) and DHA
`(by 11 %; P < 0.01) supplementation. There was a significantly
`larger (P < 0.01) decrease in linoleic acid (18:2n-6) in the EPA
`group (by 21 %; P < 0.01) than in the DHA group (by 12%;
`P < 0.01). EPA and DHA decreased arachidonic acid (20:4n-6)
`by 25% (P < 0.01) and 22% (P < 0.01), respectively, and decreased
`20:3n-6 by approximately the same extent, 36% (P < 0.01) and
`28% (P < 0.01), respectively.
`
`Platelet fatty acids
`
`EPA supplementation significantly increased platelet phospho(cid:173)
`lipid EPA by 370% (P < 0.01) and DPA by 56% (P < 0.01), but
`also significantly decreased DHA by 28% (P < 0.01). DHA sup(cid:173)
`plementation significantly increased DHA by 155% (P < 0.01)
`and EPA by 54% (NS). DPA, however, unlike in plasma phos(cid:173)
`pholipids, decreased significantly by 34% (P < 0.01). Both EPA
`and DHA significantly decreased stearic acid (18:0) (P < 0.01),
`whereas only EPA decreased 20:3n-6 (by 25%; P < 0.01). Sim(cid:173)
`ilar to plasma phospholipids, 20:4n-6 decreased significantly
`more (P < 0.01) after EPA (by 15%; P < 0.01) than after DHA (by
`7%; P < 0.01).
`
`Serum lipids
`
`There were no significant differences in fasting serum lipids
`at baseline between groups (Table 2). Changes in fasting lipids
`and lipoproteins are shown in Figures 3 and 4. There were no
`significant changes in lipids with olive oil supplementation. Nei(cid:173)
`ther EPA nor DHA supplementation had an effect on serum total
`cholesterol concentrations. After adjustment for baseline values,
`fasting triacylglycerols decreased significantly by 18.4% with
`EPA (P = 0.012) and by 20% with DHA (P = 0.003), relative to
`the placebo group. Serum LDL cholesterol increased signifi(cid:173)
`cantly with DHA (by 8%; P = 0.019), but not with EPA (by
`3.5%; NS). In the EPA group, the nonsignificant 3% decrease in
`HDL cholesterol was attributable to a significant 6.7% reduction
`in HDL 3 cholesterol (P = 0.032) and no change in HDL 2 choles(cid:173)
`terol. A small, albeit nonsignificant increase (3.1 % ) in HDL cho(cid:173)
`lesterol after DHA supplementation was due to a significant
`increase (29%) in the HDLz-cholesterol sub fraction (P = 0.004)
`with no significant change in the HDL 3-cholesterol subfraction.
`
`LDL particle size
`
`LDL particle size was not significantly different between
`groups at baseline (Table 2). Neither olive oil nor EPA had a
`significant effect on LDL particle size, whereas DHA supple(cid:173)
`mentation significantly increased LDL particle size (P = 0.002)
`after adjustment for baseline values (Table 2 and Figure 5). At
`baseline, LDL particle size was inversely correlated with triacyl(cid:173)
`glycerol (r = -0.58, P < 0.0001) and positively correlated with
`
`0
`~ :,
`0
`~
`(D
`a.
`cl
`3
`?.
`?.
`?.
`,!,!,
`0
`::,
`0
`c.d
`0-
`'<
`(C
`C
`(D
`~
`0
`:,
`0
`(D
`£
`3
`0-
`~
`
`~
`0,
`
`I\)
`0
`
`~
`
`ICOSAPENT DFNDTS00011028
`
`Hikma Pharmaceuticals
`
`IPR2022-00215
`
`Ex. 1004, p. 3 of 10
`
`

`

`1088
`
`MORI ET AL
`
`TABLE2
`Fasting serum lipids, glucose, and insulin at baseline and postintervention in the 3 groups1
`
`Olive oil ( control)
`(n = 20)
`
`EPA
`(n = 19)
`
`DHA
`(n = 17)
`
`Treatment effect (P)2
`EPA
`DHA
`
`Cholesterol (mmol/L)
`Baseline
`Postintervention
`
`Triacylglycerols (mmol/L)
`Baseline
`Postintervention
`
`LDL cholesterol (mmol/L)
`Baseline
`Postintervention
`
`LDL particle size (nm)
`Baseline
`Postintervention
`
`HDL cholesterol (mmol/L)
`Baseline
`Postintervention
`
`HDL2 cholesterol (mmol/L)
`Baseline
`Postintervention
`
`HDL3 cholesterol (mmol/L)
`Baseline
`Postintervention
`
`Glucose (mmol/L)
`Baseline
`Postintervention
`
`6.47 ± 0.21
`6.22 ± 0.10
`
`6.20 ± 0.20
`6.16 ± 0.11
`
`6.18 ± 0.18
`6.34 ± 0.11
`
`2.04 ± 0.19
`1.95 ± 0.10
`
`2.01 ± 0.19
`1.58 ± 0.10
`
`2.25 ± 0.40
`I.SO± 0.11
`
`4.41 ± 0.19
`4.31 ± 0.09
`
`4.28 ± 0.19
`4.46 ± 0.10
`
`4.27 ± 0.17
`4.64 ± 0.10
`
`25.68 ± 0.14
`25.72 ± 0.05
`
`25.64 ± 0.09
`25.69 ± 0.05
`
`25.69 ± 0.11
`25.96 ± 0.06
`
`1.12 ± 0.07
`1.02 ± 0.02
`
`1.00 ± 0.04
`0.99 ± 0.02
`
`0.96 ± 0.04
`I.OS± 0.02
`
`0.33 ± 0.05
`0.26 ± 0.02
`
`0.25 ± 0.02
`0.27 ± 0.02
`
`0.24± 0.03
`0.33 ± 0.02
`
`0.80 ± 0.03
`0.76 ± 0.01
`
`0.74 ± 0.03
`0.72 ± 0.02
`
`0.72± 0.03
`0.72± 0.02
`
`4.95 ± 0.12
`5.03 ± 0.08
`
`5.03 ± 0.09
`5.24 ± 0.08
`
`5.15 ± 0.13
`5.08 ± 0.09
`
`-0.06 ± 0.15
`(NS)
`
`0.11 ± 0.15
`(NS)
`
`-0.37 ± 0.14
`(0.012)
`
`-0.45 ± 0.15
`(0.003)
`
`0.15 ± 0.13
`(NS)
`
`0.34 ± 0.14
`(0.019)
`
`0.03 ± 0.07
`(NS)
`
`0.25 ± 0.08
`(0.002)
`
`-0.03 ± 0.03
`(NS)
`
`0.03 ± 0.03
`(NS)
`
`0.01 ± 0.02
`(NS)
`
`0.07 ± 0.02
`(0.004)
`
`-0.05 ± 0.02
`(0.032)
`
`-0.04 ± 0.02
`(NS)
`
`0.21 ± 0.11
`(0.062)
`
`0.05 ± 0.12
`(NS)
`
`0
`~ :,
`0
`~
`(D
`a.
`cl
`3
`?.
`?.
`?.
`,!,!_
`0
`::.,
`0
`c.d
`0-
`'<
`(C
`C
`(D
`~
`0
`:,
`0
`(D
`£
`3
`0-
`~
`
`~
`0,
`
`I\)
`0
`
`~
`
`9.59 ± 0.99
`11.38 ± 0.55
`
`Insulin (pmol/L)
`Baseline
`Postintervention
`
`9.79 ± 1.24
`8.76 ± 0.51
`
`8.78 ± 0.83
`10.34 ± 0.52
`
`2.62 ± 0.74
`1.58 ± 0.73
`(0.001)
`(0.035)
`1x ± SEM. There were no significant differences between the 3 groups at baseline (by one-way ANOVA). Postintervention values were not adjusted.
`EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid.
`2 A general linear model was used to assess treatment effects on postintervention values adjusted for baseline values. Data represent the change and
`P values for each of the 2 treatment groups relative to the olive oil group.
`
`HDL-cholesterol (r = 0.62, P < 0.0001) concentrations. The
`change in LDL particle size from baseline to postintervention
`was inversely correlated with the change in triacylglycerols
`(r = -0.35, P = 0.009) and positively correlated with the change
`in HDL cholesterol (r = 0.28, P = 0.035). In regression analysis,
`DHA supplementation remained the strongest independent pre(cid:173)
`dictor of postintervention LDL particle size (adjusted R 2 = 0.776,
`P = 0.002) after adjustment for baseline values and for the change
`in LDL particle size (adjusted R 2 = 0.216, P = 0.023). DHA sup(cid:173)
`plementation remained significant in regression models, which
`included changes in serum triacylglycerols and other lipids.
`
`Glucose and insulin
`
`At baseline there were no significant differences between
`groups in fasting serum glucose or insulin concentration (Table 2).
`Postintervention, however, there were significantly different
`responses between the EPA and DHA groups (Figure 6). Olive
`oil did not change either fasting glucose or insulin. After
`
`adjustment for baseline values, there was a trend toward
`increased fasting glucose concentrations with EPA (P = 0.062),
`but not with DHA (NS), relative to the control group. Both EPA
`and DHA significantly increased fasting insulin, by 18%
`(P = 0.035) and 27% (P = 0.001), respectively. DHA supple(cid:173)
`mentation also significantly decreased the glucose-insulin ratio
`by 0.13 ± 0.05 (P = 0.018).
`
`DISCUSSION
`This study addressed whether purified EPA and DHA have
`different effects on serum lipids and lipoproteins, LDL particle
`size, glucose, and insulin in mildly hyperlipidemic men. We
`found that DHA, but not EPA, improved serum lipid status, in
`particular a small increase in HDL cholesterol and a significant
`increase in the HDL2-cholesterol subfraction, without adverse
`effects on fasting glucose concentrations. Neither EPA nor
`DHA affected total cholesterol and both fatty acids reduced
`
`ICOSAPENT DFNDTS00011029
`
`Hikma Pharmaceuticals
`
`IPR2022-00215
`
`Ex. 1004, p. 4 of 10
`
`

`

`EFFECTS OF EPA AND DHA ON LIPIDS, GLUCOSE, AND INSULIN
`
`1089
`
`TABLE3
`Total energy and macronutrient intakes at baseline and changes during the
`intervention in the 3 groups1
`
`Olive oil (control)
`(n = 20)
`
`EPA
`(n = 19)
`
`DHA
`(n = 17)
`
`10441 ± 588
`-471 ± 497
`
`9516 ± 677
`82 ± 844
`
`10550 ± 588
`-188 ± 421
`
`34.2 ± 1.2
`-0.4 ± 1.3
`
`30.9 ± 1.6
`2.8 ± 1.6
`
`32.6 ± 1.6
`2.3 ± 1.2
`
`Total energy intake (kJ/d)
`Baseline
`Change
`Total fat(% of energy)
`Baseline
`Change
`Fatty acids (% of energy)
`Saturated fat
`Baseline
`Change
`Monounsaturated fat
`Baseline
`Change
`Polyunsaturated fat
`Baseline
`Change
`Protein (% of energy)
`Baseline
`Change
`Carbohydrate(% of energy)
`Baseline
`Change
`Fiber (g/d)
`26.1 ± 1.1
`27.4 ± 1.8
`30.8 ± 2.9
`Baseline
`-0.1 ± 1.7
`-2.8 ± 2.1
`-3.3 ± 2.3
`Change
`1x ± SEM. Baseline measures were compared by one-way ANOVA. A
`general linear model was used to test for treatment effects on postinterven(cid:173)
`tion values adjusted for baseline value. There were no significant differ(cid:173)
`ences between the groups in any of the dietary nutrients at baseline and no
`significant changes during the intervention. EPA. eicosapentaenoic acid;
`DHA. docosahexaenoic acid.
`
`13.6 ± 0.8
`0.0 ± 0.6
`
`12.1 ± 0.9
`1.2 ± 0.9
`
`13.6 ± 0.9
`0.8 ± 1.0
`
`12.2 ± 0.6
`-0.2 ± 0.6
`
`11.1±0.8
`1.4 ± 0.8
`
`11.5±0.7
`0.3 ± 0.6
`
`5.0 ± 0.3
`0.0 ± 0.4
`
`4.2 ± 0.3
`0.2 ± 0.4
`
`4.6 ± 0.2
`0.9 ± 0.5
`
`18.0 ± 0.5
`-0.5 ± 0.6
`
`19.8 ± 0.9
`-0.9 ± 0.8
`
`18.1 ± 0.6
`0.1 ± 0.7
`
`42.1 ± 1.7
`1.9 ± 1.6
`
`44.6 ± 1.7
`-1.0± 1.5
`
`41.8 ± 1.8
`-1.2 ± 1.4
`
`Ill Olive oil
`~ EPA
`
`■ DHA
`
`10
`9
`8
`~ 7
`~ 6
`UI
`"O 5
`'i3
`m
`4
`.?'
`~ 3
`"O 2
`·c.
`0 1
`..c: 0
`C.
`UI
`-1
`0
`..c:
`C. -2
`m
`E -3
`UI m -4
`ii:
`<l -5
`-6
`-7
`-8
`
`triacylglycerols and increased fasting insulin concentrations to
`a similar extent. DHA supplementation significantly increased
`LDL cholesterol; however, this was associated with an increase
`in LDL particle size, which may represent a shift to a less
`atherogenic LDL particle.
`Although numerous studies have examined the effect of n - 3
`fatty acids on serum lipids, glucose, and insulin (1-5, 20-22),
`few have assessed the independent effects of EPA and DHA. In
`vitro, both EPA (13, 33-35) and DHA (13, 35, 36) inhibit tri(cid:173)
`acylglycerol synthesis and secretion. In rats, EPA lowered tri(cid:173)
`acylglycerols, whereas DHA lowered cholesterol (10, 14, 15).
`These studies, however, used very high doses (1-2 g-kg- 1 -d- 1)
`of fatty acids, equalling 12-24 g/d in humans.
`In humans, n -3 fatty acids reduce triacylglycerols ( 4, 5, 37),
`with more variable effects on total cholesterol, LDL cholesterol,
`and HDL cholesterol (4, 5). These contradictory findings may be
`explained, in part, by variations in the amount of n - 3 fatty acids
`consumed, the manner in which they are presented (fish, fish oils,
`or purified oils), and the lipoprotein phenotype of the patients.
`Our own studies have shown that the background dietary fat
`intake influences serum lipid responses to n -3 fatty acids (37).
`Trials in humans using mixtures enriched in EPA and DHA
`have suggested different effects of the 2 fatty acids on serum
`lipids (38, 39). In a placebo-controlled study, 4 g EPA/d reduced
`triacylglycerols by 35% (40). It was also shown in a single-blind
`crossover study that EPA reduced triacylglycerols and VLDL
`cholesterol, increased LDL cholesterol and HDL cholesterol, but
`had no effect on total cholesterol (16). DHA did not affect cho(cid:173)
`lesterol, triacylglycerols, VLDL cholesterol, LDL cholesterol, or
`HDL cholesterol, but increased the HDL2-cholesterol subfrac(cid:173)
`tion and reduced the HDL 3-cholesterol subfraction (16). That
`study, however, had only a small number of subjects in the DHA
`group, was short in duration, and included only a 2-wk washout
`period between treatments (16).
`
`* t
`
`t
`
`*
`
`t
`
`0
`~ :,
`0
`~
`(D
`a.
`cl
`3
`?.
`?.
`?.
`,!,!.
`0
`::.,
`0
`c.d
`0-
`'<
`(C
`C
`(D
`~
`0
`:,
`0
`(D
`£
`3
`0-
`~
`
`~
`0,
`
`I\)
`0
`
`~
`
`•,t
`
`Treatment effect (P)
`
`NS
`
`0.010
`
`NS
`
`0.001
`
`<0.0001
`
`<O 0001
`
`<0.0001
`
`<0.0001
`
`NS
`
`< 0.0001
`
`<0.0001
`
`16:0
`
`16:1n-7
`
`18:0
`
`18:1n-9 18:2n-6 20:3n-6 20:4n-6 20:Sn-3 22:4n-6 22:Sn-3 22:6n-3
`
`FIGURE 1. Mean (±SEM) changes in plasma phospholipid fatty acids from baseline to the end of the intervention in the olive oil (control; n = 20),
`eicosapentaenoic acid (EPA; n = 19), and docosahexaenoic acid (DHA; n = 17) groups. ANOVA was used to assess treatment effects. 'Significantly
`different from the olive oil group, P < 0.01. 1Significantly different from the EPA group, P < 0.01.
`
`ICOSAPENT DFNDTS00011030
`
`Hikma Pharmaceuticals
`
`IPR2022-00215
`
`Ex. 1004, p. 5 of 10
`
`

`

`1090
`
`6
`
`MORI ET AL
`
`5
`...... 4
`-;!!_
`~
`Ill 3
`"C
`·13
`Cll 2
`
`>, -1ii -"C
`
`"ii 0
`0 .c
`-1
`C.
`Ill
`0 ..c -2
`C.
`"iii -3 -
`ai
`
`Cll
`
`0:: -4
`<1
`-5
`
`-6
`
`m!1Jil Olive oil
`~ EPA
`■ DHA
`
`*
`
`1/j
`
`* t
`
`*
`1//
`
`Treatment effect ( P)
`
`NS
`
`0.011
`
`NS
`
`NS
`
`< 0.0001
`
`*,t
`
`*
`< 0.0001
`
`<0.0001
`
`NS
`
`<0.0001
`
`<0.0001
`
`16:0
`
`18:0
`
`18:1n-9
`
`18:2n-6
`
`20:3n-6 20:4n-6 20:Sn-3 22:4n-6
`
`22:Sn-3
`
`22:6n-3
`
`FIGURE 2. Mean (±SEM) changes in platelet phospholipid fatty acids from baseline to the end of the intervention in the olive oil (control; n = 20),
`eicosapentaenoic acid (EPA; n = 19), and docosahexaenoic acid (DHA; n = 17) groups. ANOVA was used to assess treatment effects. 'Significantly
`different from the olive oil group, P < 0.01. 1Significantly different from the EPA group, P < 0.01.
`
`Several reports have described the effects of DHA supple(cid:173)
`ments on serum lipids in humans. Nelson et al (17), in a single(cid:173)
`blind study of healthy men, compared the effects of 6 g DHA/d
`with those of a control diet. They reported that after 90 d total
`cholesterol, LDL cholesterol, apo A-I, apo B, and lipoprotein(a)
`were unchanged, whereas triacylglycerols decreased and HDL
`cholesterol increased. Similarly, in patients with combined
`hyperlipidemia, Davidson et al (18) compared the effects of 1.25
`and 2.5 g DHA/d with those of a vegetable-oil control and
`showed that DHA reduced triacylglycerols significantly. The
`higher dose of DHA was also associated with a significant
`increase in LDL cholesterol.
`In another study, healthy, nonsmoking men were supple(cid:173)
`mented daily with 4 g EPA, DHA, or corn oil for 7 wk (19). Both
`EPA and DHA reduced triacylglycerols, by 21 % and 26%,
`respectively. In the present study, the same dose of EPA and
`DHA for 6 wk reduced triacylglycerols by 18% and 20%, respec(cid:173)
`tively. We observed no significant effect of EPA or DHA on total
`cholesterol. In contrast, Grimsgaard et al (19) reported increased
`total cholesterol with EPA. The difference in results between the
`2 studies may have been due to differences in the baseline serum
`lipid concentrations of the subjects.
`It has been suggested that serum HDL cholesterol is better
`maintained with DHA-enriched than with EPA-enriched oils
`(38). The present data and previous findings (19) support this
`hypothesis. We observed that the increase in HDL cholesterol
`was due to a 29% increase in HDL2 cholesterol. Increased HDL2
`cholesterol was reported previously by our group after daily con(cid:173)
`sumption of fish or fish oils by subjects with type 2 diabetes or
`at risk of heart disease (23, 37). In contrast, Grimsgaard et al
`(19) surmised that both EPA and DHA increase HDL 2 choles-
`
`terol because both fatty acids increased the ratio of HDL choles(cid:173)
`terol to apo A-I. DHA increased HDL cholesterol and EPA
`decreased apo-AI, suggesting an increased surface-to-core
`ratio of the HDL particle and a redistribution of the HDL sub(cid:173)
`classes toward the larger HDL 2 particles (41). The mechanisms
`by which DHA increases HDL cholesterol are not known, but
`may be related to alterations in lipid transfer protein activity,
`which decreases after n-3 fatty acid supplementation (41). In
`epidemiologic terms, the increase in HDL2 cholesterol could
`have a marked effect on the incidence of cardiovascular dis(cid:173)
`ease, given that HDL 2 cholesterol may be the subfraction of
`HDL cholesterol that may be most protective against coronary
`heart disease (42).
`Although the LDL-cholesterol concentration increased after
`EPA and DHA intakes, the increase was significant only after
`DHA. The increased LDL-cholesterol concentration may relate
`to the hypotriglyceridemic effects of these fatty acids (43). n-3
`Fats reduce hepatic VLDL synthesis, VLDL secretion, or both
`with the result that the smaller VLDL particles formed are more
`readily converted to LDL than are the larger VLDL particles
`(44). Smaller VLDL particles can also compete with LDL for
`uptake by LDL receptors. A down-regulation of the LDL recep(cid:173)
`tor has been reported in some but not all studies (43).
`LDL particle size increased significantly with DHA supple(cid:173)
`mentation, a result that mig