Effects of Fish Oil Supplementation on Apolipoprotein B100 Production

Home , Triolein

Effects of ®sh oil supplementation on apolipoprotein B100 production and lipoprotein metabolism in normolipidaemic males

P Bordin1, OAF Bodamer2, S Venkatesan3, RM Gray3, PA Bannister3 and D Halliday2

1Istituto Clinica Medica, 34100 Trieste, Italy, 2Medical Unit, Institute of Child Health, London WC1N 1EH, 3MRC Human Metabolism Research Group, Imperial College School of Medicine at St. Mary's, London W2 1PG

Objective: To evaluate the effects of four weeks of ®sh oil supplementation on apolipoprotein B100 production and lipoprotein metabolism in normolipidaemic males. Design and subjects: Very low density lipoprotein (VLDL) apolipoprotein B100 (apoB100) kinetics in ten healthy, white males, aged 22±43 y (mean 32 y) were investigated using a 13C-leucine technique and gas chromatography-mass spectrometry before and after ®sh oil supplementation. Intervention: All subjects received 10 g (1.8 g EPA, 1.2 g DHA)=d of ®sh oil concentrate for four weeks. Results: Fish oil supplementation resulted in a decrease of total plasma VLDL (mean Æ s.d. 1.11 Æ 0.41 vs 0.87 Æ 0.28 mmol=l, P < 0.05) and triacylglycerol concentrations (0.74 Æ 0.27 vs 0.48 Æ 0.21 mmol=l, P < 0.01). VLDL apoB100 pool size was decreased without alteration of the fractional synthetic rate but a signi®cant decrease of apoB100 production (2.23 Æ 0.90 vs 1.54 Æ 0.52 mg=dl=h, P < 0.02). Following ®sh oil supplementation plasma concentrations of glucose and insulin as well as lipoprotein and hepatic lipase activities were unchanged. Fasting plasma concentrations of non-esteri®ed fatty acid (NEFA) were decreased (0.45 Æ 0.12 vs 0.33 Æ 0.10 mmol=l, P < 0.05). Conclusions: Dietary supplementation with ®sh oil in healthy males results in decreased VLDL-triacylglycerol concentrations through a decrease in VLDL particle synthesis. The decrease in NEFA substrate supply also contributes. Sponsorship: Fish oil (Maxepa1) was supplied by Seven Seas Ltd, Hull, England, free of charge. Dr OAF Bodamer was funded by the `Deutsche Forschungsgemeinschaft' (Bo 1193=1-2). Descriptors: ®sh oil; n-3 fatty acids; apoB100; lipoprotein; mass spectrometry; stable isotopes

Introduction cerol concentrations and also reverse the fall of HDL cholesterol which is typically seen in subjects on a low- Several epidemiological studies have revealed an inverse fat diet (Mori et al, 1994). relationship between ®sh oil consumption and the incidence The effects of ®sh oil on apolipoprotein B100 (apoB100) or mortality of coronary heart disease (CHD) (Norell et al, metabolism are not fully understood. ApoB100 production 1986; Gibson, 1988; Berg-Schmidt et al, 1993). The and pool size were reduced in two studies in which lipoprotein bene®cial effects of ®sh oil are related to their content of kinetics were studied using radiolabelled lipoproteins, where long chain n-3 polyunsaturated fatty acids (n-3 PUFA) the production rate was obtained indirectly from the catabolic (Nordoy et al, 1994). rate (Illingworth et al, 1984; Nestel et al, 1984). The effects of n-3 PUFA on lipid metabolism are The aim of this study was to analyse the short term complex. A constant ®nding is a dose-dependent decrease effects of a high dose ®sh oil supplementation on VLDL in triacylglycerol (TG) plasma concentrations (Harris, composition and apoB100 production, measured directly, 1989). Plasma concentrations of total cholesterol (TC), in healthy, male subjects using stable isotope tracer. high density lipoprotein cholesterol (HDL-C) and low density lipoprotein cholesterol (LDL-C) typically tend to increase following high doses of ®sh oil (10 g=d) (Harris, Subjects and methods 1989; Mori et al, 1994). However, contradicting responses have also been reported, which may be explained by Subjects variations in the quantity and quality of dietary fat intake The study was approved by the St. Mary's Hospital Local as well as plasma lipid concentrations (Harris, 1989). Research Ethics Committee in accordance with the Hel- However any potential adverse effects of n-3 PUFA on sinki Declaration of 1975 as revised in 1983. Written, cholesterol metabolism are reversed by a 30%-fat diet. informed consent was obtained from each subject. Ten Regular ®sh meals signi®cantly reduce plasma triacylgly- healthy Caucasian, male volunteers, aged 22±43 y (mean 32 y) were included in the study. None of the subjects was on any medication or had any metabolic disorder.

Study design Correspondence: Dr OAF Bodamer, Department of Molecular and Human Genetics, Baylor College of Medicine, One Baylor Plaza, T607, Houston, Subjects received 10 g of ®sh oil concentrate (1.8 g eico- TX, 77030, USA. sapentanoic acid, 1.2 g docosahexanoic acid) (Maxepa1, Received 30 June 1997; revised 1 October 1997; accepted 5 October 1997 Seven Seas Ltd, Hull, England) daily for four weeks with Effects of ®sh oil supplementation P Bordin et al 105 no changes to their diet and physical activity. Dietary Quanti®cation of apoB100 intake of each subject was recorded using food scales for ApoB100 in VLDL was estimated from the mean of 5 d before starting and during the last week of ®sh oil samples taken at three time points in the study (0, 2 and supplementation. Before and at the end of the four week 4 h) using an ELISA technique with antihuman apoB trial the subjects came to the Metabolic Day Ward after an antiserum (Boehringer) and anti-sheep IgG alkaline con- overnight fast for 12 h. At approximately 09.00 h a cannula jugate (Sigma). ApoB calibration serum (Boehringer) was was inserted into a dorsal vein of each hand; one cannula used for the calibration curve for each plate. The apoB100 was used for the administration of L-[1-13C]leucine content of VLDL aliquots was corrected for the plasma (99%13C, Cambridge Isotope Laboratories, Woburn, MA, volume from which it had been derived. USA) and the other for blood sampling. Following a priming dose (1 mg=kg body weight) of the tracer, a Measurement of stable isotope enrichment constant infusion (1 mg=kg body weight=h) was com- Selective ion monitoring of the derivatized samples at a menced for 6 h. At the end of the infusion, a bolus of mass to charge ratio (m=z) of 302 for unlabelled norleucine 100 U=kg body weight sodium heparin (C.P. Pharmaceu- and leucine and m=z 303 for labelled leucine was used to ticals Ltd., Wrexham, UK) was given intravenously in determine isotopic abundance. The atom percent excess order to measure lipoprotein and hepatic lipase activities. (APE) enrichment was calculated using the following Subjects fasted throughout the study with only mineral formula water being allowed. IR IR Blood samples (12 ml) were taken into EDTA tubes at 0 Enrichment APE t 0 Â 100 and 1 h and thereafter every 30 min throughout the infu- IRt IR0 1 sion for VLDL apoB100 isolation and for plasma lipopro- where IRt is the isotope ratio of the sample at time t, and teins and NEFA concentrations at 0 h and 1 h. Plasma was IR is the isotope ratio of sample at time zero. The immediately separated by low-speed centrifugation for 0 enrichment of plasma aKIC values (precursor pool) was 30 min at 4 C. An equal mixture (50ml) of 5% NaN3 and similarly obtained. Mole percent excess (MPE) was calcu- 5% gentamycin was added to each of these plasma samples. lated from enrichment calibration curves that were deriva- Blood (5 ml) was taken into heparinized chilled tubes at tised and run with the study samples. This method of times 0, 2 and 3 h and thereafter at 30 min intervals for the calculation is equivalent to the method of Cobelli et al, remainder of the infusion. The plasma was stored at 1987. 770C for analysis of 13C-enrichment of plasma a-keto isocaproic acid (aKIC), the deamination product of leucine. Calculation of apoB100 production Blood (4 ml) was taken at 0, 2, 4 and 6 h for measurement Fractional synthetic rate (FSR) of VLDL apoB100 (as of plasma lipids and glucose; insulin was measured at 0 and pools=h) was calculated from the rate of incorporation of 6 h. Blood (5 ml) was taken into tubes without additives 13C-leucine into the circulating protein and application of 15 min after the heparin injection to assess lipase activity. monocompartmental kinetics and a monocompartmental model (Parhofer et al, 1991) VLDL and apoB100 separation and quanti®cation VLDL was separated by ultracentrifugation and tube sli- A t Ap 1 e k t d cing, and stored at 770C for apoB100 isolation (Venka- tesan et al, 1993). A 0.6 ml aliquot was delipidated twice where A(t) is the tracer=tracee (labelled=unlabelled leucine) with centrifugation at 4C and decantation of the super- ratio at time t, Ap is the precursor tracer=tracee ratio (taken natant, ®rstly by adding 24 ml of a 10:30 vol=vol mixture as the plasma aKIC enrichment), d is the delay time until of methanol and diethyl ether and centrifugation for the appearance of labelled apoB100 in the VLDL fraction 10 min, and secondly by adding 6 ml of diethyl ether and and k represents the FSR of VLDL apoB100. The absolute centrifugation for 2 min. The precipitated protein was dried production rate (APR) of VLDL apoB100 was calculated at 30C in a fume cupboard, then reconstituted in 60 m l as: TRIS-buffer containing mercaptoethanol and glycerol, APR FSR (pools=h) 6 apolipoprotein pool size (mg) incubated at 100C for 3 min and subjected to polyacryla- mide gradient (5±16%) gel electrophoresis. After overnight Other assays electrophoresis, the gel was stained with Cromassie blue Laboratory glucose measurements were performed by a (0.5% in methanol=acetic acid solution) and destained for glucose oxidase method. Insulin was determined with an several hours (Venkatesan et al, 1993). The apoB100 gel ELISA method (Anyaoku & Johnston, 1995). Cholesterol, band was cut and transferred to 25 ml tubes with Te¯on1- triacylglycerol and NEFA were measured enzymatically lined screw caps (Phillip Harris Ltd, London, England). using a centrifugal analyser (COBAS, Baker, Windsor, 2 ml of 6N HCl were added and the tubes sealed under UK). HDL cholesterol was isolated by the dextran-sulphate nitrogen and incubated at 110C for 24 h to completely MgC12 method and LDL cholesterol calculated with Frie- hydrolyze the apoB100. dewald's formula (Friedewald et al, 1972). Plasma and A known amount of norleucine, a non-physiological VLDL apoB100 were measured by an ELISA technique amino acid used as internal standard, was added to each using Boheringer antibodies and Nunc plates. Plasma lipase sample. The amino acids were puri®ed by passing through a activity was quanti®ed on the basis of fatty acid release cationic resin column and derivatized as described pre- from a radiolabelled triolein substrate, and assessed by viously (Venkatesan et al, 1993). Brie¯y, plasma aKIC was liquid scintillation counting (Baynes et al, 1991). reacted with pyridine and N,O-bis(trimethylsilyl)tri¯uoro- acetamide to form the O-trimethylsilylquinoxalinol deriva- Body impedance tive. Enrichment of both leucine and aKIC was quanti®ed Lean body mass (LBM) was assessed by using a body by using a Finnigan INCOS XL mass spectrometer (MS) impedance meter (Bodystat 1500, Douglas, Isle of Man, run in electron impact mode. UK), which measures bioelectrical impedance following a Effects of ®sh oil supplementation P Bordin et al 106 50 Hz, 800 mA current through the body. LBM was calcu- Table 2 Plasma lipids, lipase activities, glucose and insulin levels before lated using standard equations with bioelectrical impe- and after ®sh oil dance, height and body weight (Houtkooper et al, 1996). Pre ®sh oil Post ®sh oil Lean body mass is expressed as a percentage of total body weight. Cholesterol (mmol=l) Total 4.77 Æ 0.66 4.85 Æ 0.94 VLDL 0.20 Æ 0.13 0.10 Æ 0.06* Statistical analyses LDL 3.43 Æ 0.65 3.48 Æ 0.99 The mean and standard deviation were calculated for each HDL 1.12 Æ 0.19 1.20 Æ 0.23 parameter. Differences in the same individuals before vs Triacylglycerol (mmol=l) Total 1.11 Æ 0.41 0.87 Æ 0.28* after ®sh oil were examined by two tailed paired t-test. Data VLDL 0.74 Æ 0.27 0.48 Æ 0.21** which were not normally distributed (insulin values) were NEFA (umol=l) 0.45 Æ 0.12 0.33 Æ 0.10* log-transformed before analysis. These calculations and the Lipase (mmol=h=l) monoexponential function for apoB kinetics were run with Lipoprotein 11.91 Æ 2.69 13.33 Æ 5.17 Hepatic 20.21 Æ 5.92 19.19 Æ 6.43 the Statistical Package for Social Sciences (SPSS 6.0 for Glucose 0 h (mmol=l) 5.17 Æ 0.56 4.94 Æ 0.30 Windows). Glucose 6 h (mmol=l) 4.73 Æ 0.28 4.65 Æ 0.30*** Insulin 0 h (pmol=l) 1.49 (0.53±1.80) 1.50 (0.41±1.66) Insulin 6 h (pmol=l) 0.87 (0.23±1.72) 0.69 (0.10±1.50)*** Results All values except insulin are expressed as mean Æ s.d. Insulin values are Table 1 summarises the characteristics of the subjects log-transformed and expressed as median (range); NEFA non esteri®ed before begin of the study. Body mass index (before, fatty acids, *P < 0.05, **P < 0.01; lipase activities are expressed as mmol of free fatty acid released=h=l of post heparin plasma; ***P < 0.05 at 0 vs mean Æ s.d., 23.0 Æ 2.4; after, 23.0 Æ 2.5) and lean body 6 h. mass (before, 83.8 Æ 3.8; after, 83.9 Æ 3.0%) remained constant during the study. Dietary energy intake did not alter, nor was there any alteration in intake of other fats 13 during the study period. Fish oil was well tolerated and the C-leucine enrichment curve was shifted to the left follow- compliance good. Three subjects experienced mild ¯atu- ing four weeks ®sh oil ingestion (Figure 1). The observed lence, belching and mild epigastric discomfort. Table 2 apoB100 kinetic data are shown in Table 3. There was no shows the effects of ®sh oil on lipids, glucose, insulin and signi®cant change in FSR. The VLDL apoB100 pool size lipase values. There was a signi®cant decrease in plasma decreased after ®sh oil ingestion (43%, P < 0.05). The NEFA (27%, P < 0.05) and total triacylglycerol (22%, apoB100 production rate was reduced (29%, P < 0.05). P < 0.05) concentrations due to a reduction in VLDL The ratio of apoB100 to triacylglycerol (Table 4) did not triacylglycerol. HDL cholesterol concentrations did not change signi®cantly after ®sh oil; the ratio was unique to change signi®cantly and nor did total or LDL cholesterol. individual subjects. Lipid concentrations did not change over the 6 h infusion (data not shown). Hepatic and lipoprotein lipase activities were not modi®ed by ®sh oil. Blood glucose and insulin Discussion concentrations did not change signi®cantly. Fish oil has several effects on lipid metabolism, most notably a reduction of triacylglycerol concentrations due VLDL apoB100 kinetics to diminished hepatic synthesis (Mackness et al, 1994; Plasma aKIC 13C-enrichment increased rapidly following Eritsland et al, 1995). Data on apoB kinetics are limited: the start of 13C-leucine infusion, reaching a plateau by the Nestel et al, 1984 reported that ®sh oil supplementation second hour. The pattern was similar before and after ®sh reduced VLDL apoB synthesis using a radioisotopic tech- oil. Enrichment of 13C-leucine in the VLDL apoB100 nique. These studies have both practical and theoretical demonstrated monoexponential kinetics (Figure 1). The limitations: the use of radioactive tracer excludes certain enrichment plateau was reached at 5±6 h (MPE 5.01± groups of patients and makes repetitive studies on the same 5.4% before and MPE 5.22±5.57% following ®shoil). The patient ethically dif®cult. Structural modi®cations of lipo-

Table 1 Basal characteristics of 10 normolipidaemic males

Age BMI LBM Glucose Total C Triacylglycerol HDL C LDL C Subject years kg=m2 % mmol=l mmol=l mmol=l mmol=l mmol=l

1 32 23.7 82.8 5.00 4.56 0.99 1.02 3.34 2 43 19.2 80.4 5.30 4.35 1.56 0.80 3.24 3 30 21.8 83.4 5.20 5.08 1.18 1.10 3.74 4 37 24.0 82.5 4.50 4.34 0.73 1.14 3.05 5 28 25.7 78.9 5.00 5.79 1.71 1.09 4.36 6 37 22.3 85.2 5.30 3.65 0.55 1.32 2.22 7 22 27.4 81.7 4.70 4.74 1.69 1.01 3.39 8 24 23.4 87.7 5.50 4.33 0.87 1.23 2.93 9 32 21.1 92.1 6.50 5.30 0.99 1.49 3.61 10 38 20.9 83.2 4.70 5.56 0.85 1.00 4.39 mean 32.3 23.0 83.8 5.17 4.77 1.11 1.12 3.43 s.d. 6.6 2.4 3.8 0.56 0.66 0.41 0.19 0.65

BMI body mass index; LBM lean body mass; C cholesterol; HDL high density lipoprotein; LDL low density lipoprotein. Effects of ®sh oil supplementation P Bordin et al 107

Figure 1 1-13C leucine enrichment (MPE%, mean Æ s.d.) of apoB 100 before (open symbols) and after (closed symbols) ®sh oil supplementation.

Table 3 VLDL apoB-100 production before and after ®sh oil supple- Table 4 VLDL composition before and after ®sh oil supplementation mentation Pre ®sh oil Post ®sh oil Before ®sh oil After ®sh oil TG=apoB CT=apoB TG=apoB CT=apoB FSR Pool Production FSR Pool Production Subject mol=mol mol=mol mol=mol mol=mol Subject pool=h mg=dl mg=dl=h pool=h mg=dl mg=dl=h 1 3163 1284 3504 1051 1 0.48 7.12 3.38 0.49 4.70 2.29 2 2770 1421 2017 896 2 0.17 15.06 2.59 0.33 4.90 1.62 3 6506 1215 8845 1597 3 0.34 9.04 3.03 0.28 4.47 1.27 4 12262 863 13253 722 4 0.39 3.18 1.25 0.37 2.28 0.84 5 7031 2207 4945 1545 5 0.25 6.77 1.71 0.41 3.55 1.44 6 6555 698 8163 583 6 0.55 3.94 2.15 0.56 1.89 1.06 7 10252 3613 12626 3703 7 0.27 6.54 1.77 0.34 3.26 1.10 8 8629 1418 9759 1555 8 0.36 4.65 1.66 0.53 3.88 2.07 9 4733 1420 5283 1383 9 0.49 4.64 2.26 0.49 4.36 2.15 10 10675 3378 10615 2307 10 0.45 4.09 1.84 0.45 3.58 1.60 mean 7258 1752 7901 1534 mean 0.37 6.50 2.17 0.42 3.69 1.54 s.d. 3196 1003 3834 915 s.d. 0.12 3.50 0.66 0.09 1.00 0.49 P 0.085 0.014 0.012 TG triacylglycerol; CT cholesterol.

FSR fractional synthetic rate; SD standard deviation; P t-test after vs before ®sh oil supplementation. re¯ect the enrichment of the direct precursor of apoB-100 (leucyl-RNA); (c) newly synthesized apoB is homoge- nously labelled and secreted. As some newly synthesized proteins during their isolation and=or radiolabelling may VLDL are degraded intracellularly, the production rate of alter their kinetics. Furthermore, synthetic rates are indir- apoB100 measured in turnover studies, either using endo- ectly obtained from the measured catabolic rates assuming genous (mass spectrometry) or exogenous labelling (radio- steady-state conditions throughout the study (Halliday et al, labelling) re¯ects not the level of absolute hepatic 1993). These problems are overcome by using an intrinsic synthesis, but rather the level of secretion or appearance labelling technique (Waterlow et al, 1978). We therefore of apoB100 from the liver into the circulation. employed a 13C-leucine technique in which the incorpora- This is the ®rst time the effect of n-3 PUFA on VLDL tion rate of the stable isotope labelled aminoacid into apoB100 synthesis has been studied by direct measurement VLDL apoB was measured. The validity of the method using a stable isotope approach. We chose a sample of depends on certain assumptions (Venkatesan et al, 1993): healthy normolipidaemic people homogeneous for sex, age (a) there must be equilibrium between plasma and intrahe- and race, who consumed a high dose (3 g=d) of n-3 PUFA patic aKIC; (b) the level of plasma aKIC enrichment must for four weeks, although control subjects without dietary Effects of ®sh oil supplementation P Bordin et al 108 intervention were not included. This obvious limitation in reported in subjects with type II diabetes (McManus et al, study design is overcome by the good reproducibility of 1996). this technique which has recently been demonstrated (Demant et al, 1996; Thornton et al, 1996). The reported AcknowledgementsÐWe are indepted to the statistician, Ms. J. Woods- coef®cient of variation was 10.5% (Thornton et al, 1996). worth, for help in the design of the study protocol and to Seven Seas Ltd The most striking effect of n-3 PUFA observed was a for supplying the ®sh oil. We thank Dr. Paul Pacy and Prof. D.G. Johnston, decrease in VLDL apoB100 pool size (43%, P < 0.05) and Head of the Unit of Metabolic Medicine, for their support and scienti®c production rate (29%, P < 0.05). This was observed with- advice. We also thank Dr. V. Anyaoku and Miss L. Nimmo for technical out any change in BMI, lean body mass or diet. The results help. are similar to those observed by Nestel et al, 1984 who found a 73% decrease in pool size and a 55% decrease in production rate using a considerably higher dose of ®sh oil References (up to 30% of daily energy needs). Hepatic and lipoprotein Anyaoku V & Johnston D (1995): Rapid, speci®c and sensitive enzyme- lipase activities did not change after n-3 PUFA supplemen- linked immunosorbent assay for intact human insulin. Diabetologia 38 tation, consistent with other studies (Desager et al, 1989; (S1), A157. Bagdade JD, Buchanan WE, Levy RA, Subbaiah PV & Ritter MC (1990): Bagdade et al, 1990). Our own and other data therefore do Effects of w-3 ®sh oils on plasma lipids, lipoprotein composition, and not suggest a major alteration in VLDL catabolism. Pre- postheparin lipoprotein lipase in women with IDDM. Diabetes 39, 426± vious studies have suggested that ®sh oil supplementation 431. results in the production of small VLDL, with a reduced Baynes C, Henderson AD, Anyaoku V, Richmond W, Hughes CL, Johnston DG & Elkeles RS (1991); The role of insulin insensitivity lipid content, which are rapidly converted to LDL (Harris, and hepatic lipase in the dyslipidaemia of type 2 diabetes. Diabet. Med. 1989; Berg-Schmidt et al, 1993). 8, 560±566. The hypotriacylglycerolemic effect of n-3 PUFA Berg-Schmidt E, Dalby Kristensen S, De Caterina R & Illingworth DR (22%) was almost exclusively due to the decrease of (1993): The effects of n-3 fatty acids on plasma lipids and lipoproteins VLDL-TG (35%). A reduction in VLDL-C (750%) and other cardiovascular risk factors in patients with hyperlipidemia. Atherosclerosis 103, 107±121. and plasma apoB100 was also observed in accordance Cobelli C, Toffolo B, Bier DM & Nosadini R (1987): Models to interpret with the literature (Nestel et al, 1984; Harris, 1989; kinetic data in stable isotope studies. Am. J. Physiol. 253, E551±E564. Berg-Schmidt et al, 1993; Mori et al, 1994). The Demant T, Packard CJ, Demmelmair H, Stewart P, Bedynek A, Bedford D, reduction of VLDL apoB100 means a diminished number Seidel D, Shepherd J (1996): Sensitive methods to study human apolipoprotein B metabolism using stable isotope-labeled amino of circulating VLDL particles, given that each lipoprotein acids. Am. J. Physiol. 270, E1022±E1036. particle contains only one apoB100 molecule. In the pre- Desager JP, Dricot J & Harvengt C (1989): Hypotriglyceridemic action of sent study the VLDL ratio of components apoB100 to omega-3 fatty acids in healthy subjects does not occur by enhanced triacylglycerol, was unaltered. Large inter-subject variation lipoprotein lipase and hepatic lipase activities. Res. Commun. Chem. in VLDL composition were observed in the current Pathol. Pharmacol. 65, 269±272. Eritsland J, Arnesen H, Selje¯ot I & Hostmark AT (1995): Long-term study suggesting that this parameter may represent a metabolic effects of n-3 polyunsaturated fatty acids in patients with unique statement of VLDL assembly in normolipidaemic coronary artery disease. Am. J. Clin. Nutr. 61(4), 831±836. individuals. Fasching P, Ratheiser K, Waldhausl W, Rohac M, Osterrode W, Nowotny Although VLDL-TG and VLDL apoB100 decrease after P & Vierhapper H (1991): Metabolic effects of ®sh-oil supplementation in patients with impaired glucose tolerance. Diabetes 40, 583±589. ®sh oil supplementation, LDL-C did not change (or has Friedewald WT, Levy RI & Fredrickson DS (1972): Estimation of the even been reported to increase in hypertriacylglycerolemic concentration of low density lipoprotein cholesterol in plasma without subjects) (Nestel et al, 1984; Nenseter et al, 1992). The use of the preparative ultracentrifuge. Clin. Chem. 18, 499±502. absence of change of LDL-C in the present study re¯ects Gibson RA (1988): The effect of diets containing ®sh and ®sh oils on the small pool of VLDL relative to LDL in circulation. disease risk factors in humans. Aust. N. Z. J. Med. 18(5), 713±722. Halliday D, Venkatesan S & Pacy P (1993): Apolipoprotein metabolism: a HDL cholesterol did not change signi®cantly in the present stable-isotope approach. Am. J. Clin. Nutr. 57, suppl., 726S±731S. study and results in the literature are con¯icting (Rudel et Harris WS (1989): Fish oils and plasma lipid and lipoprotein metabolism al, 1991; Thornburg & Rudel, 1992). in humans: a critical review. J. Lipid. Res. 30, 785±807. Houtkooper LB, Lohman TG, Going SC & Howell WH (1996): Why bioelectrical impedance analysis should be used for estimating adipos- Conclusions ity. Am. J. Clin. Nutr. 64, (suppl), 436S±448S. Illingworth DR, Harris WS & Connor WE (1984): Inhibition of low The signi®cant decrease in NEFA (27%, P < 0.05) follow- density lipoprotein synthesis by dietary omega-3 fatty acids in man. ing ®sh oil supplementation has not been previously Atherosclerosis 4, 270±275. Mackness MI, Bhatnagar D, Durrington PN, Prais H, Haynes B, Morgan J reported. Although the present study did not include control & Barthwick L (1994): Effects of a new ®sh oil concentrate on plasma subjects, plasma NEFA concentrations decreased consis- lipids and lipoproteins in patients with hypertriglyceridaemia. Eur. J. tently in all subjects studied (individual data not shown), Clin. Nutr. 48(12), 859±865. while dietary intake and the level of physical activity were McManus RM, Jumpson J, Finegood DT, Clandinin MT, Ryan EA (1996): unchanged. Samples for plasma NEFA concentrations have A comparision of the effects of n-3 fatty acids from linseed oil and ®sh oil in well-controlled type II diabetes. Diabetes Care 19(5), 463± been collected under exactly the same conditions (fasting, 467. day time, resting) before and after ®sh oil supplementation. Mori TA, Vandongen R, Beilin LJ, Burke V, Morris J & Ritchie J (1994): It is therefore most likely that the decrease in plasma NEFA Effects of varying dietary fat, ®sh and ®sh oil on blood lipids in a concentrations is secondary to ®shoil supplementation. This randomized controlled trial in men at risk of heart disease. Am. J. Clin. Nutr. 59(5), 1060±1068. could be a factor in the TG decrease, limiting substrate Nenseter MS, Rustan AC, Lund-Katz S, Soyland E, Maelandsmo G, supply for VLDL-TG synthesis. It may also re¯ect Phillips MC & Drevon CA (1992): Effect of dietary supplementation increased insulin sensitivity induced by ®sh oil, as hypothe- with n-3 polyunsaturated fatty acids on physical properties and meta- sized by some authors (Fasching et al, 1991). In the present bolism of low density lipoprotein in humans. Atheroscl. Thromb. 12, study there was no signi®cant change in basal insulin and 369±379. Nestel PJ, Connor WE, Reardon MF, Conner S, Wong S & Boston R glucose plasma concentrations, suggestive of increased (1984): Suppression by diets rich in ®sh oil of very low density insulin sensitivity. This is consistent with ®ndings recently lipoprotein production in man. J. Clin. Invest. 74(1), 82±89. Effects of ®sh oil supplementation P Bordin et al 109 Nordoy A, Hatcher L, Goodnight S, Fitzgerald GA & Conner WE (1994): Thornburg JT & Rudel LL (1992): How do polyunsaturated fatty acids Effects of dietary fat content, saturated fatty acids, and ®sh oil on lower lipids? Curr. Opin. Lipido. 3, 17±21. eicosanoid production and hemostatic parameters in normal men. J. Thornton CA, Welle S, Griggs RC & Abraham GN (1996): Human IgG Lab. Clin. Med. 123(6), 914±920. production in vivo: determination of synthetic rate by nonradioactive Norell SE, Ahlbom A, Feyehting M & Pedersen NL (1986): Fish tracer incorporation. J. Immunol. 157(2), 950±955. consumption and mortality from coronary heart disease. Br. Med. J. Venkatesan S, Cullen P, Pacy P, Halliday D & Scott J (1993): Stable 293, 421±426. isotopes show a direct relation between VLDL apoB overproduction Parhofer KG, Hugh P, Barrett R, Bier DM & Schonfeld g (1991): and serum triglyceride levels and indicate a metabolically and bio- Determination of kinetic parameters of apolipoprotein B metabolism chemically coherent basis for familial combined hyperlipidemia. Ather- using amino acids labeled with stable isotopes. J. Lipid. Res. 32, 1311± oscl. Thromb. 13, 1110±1118. 1323. Waterlow JC, Garlick PJ & Millward DJ (1978): Protein Turnover in Rudel LL, Sawyer JK & Parks JS (1991): Dietary fat, lipoprotein structure, Mammalian Tissues and in the Whole Body. Amsterdam: North-Holland and atherosclerosis in primates. Atherosclerosis Rev. 23, 41±50. Publishing Company.