Effects of Dietary Coconut Oil, Butter and Safflower Oil on Plasma Lipids, Lipoproteins and Lathosterol Levels

Effects of dietary coconut oil, butter and safflower oil on plasma lipids, lipoproteins and lathosterol levels

C Cox1, W Sutherland2, J Mann1, S de Jong2, A Chisholm1 and M Skeaff1

Departments of 1Human Nutrition and 2Medicine, University of Otago, Dunedin, New Zealand

Objective: The aim of this present study was to determine plasma levels of lathosterol, lipids, lipoproteins and apolipoproteins during diets rich in butter, coconut fat and saf¯ower oil. Design: The study consisted of sequential six week periods of diets rich in butter, coconut fat then saf¯ower oil and measurements were made at baseline and at week 4 in each diet period. Subjects: Forty-one healthy Paci®c island polynesians living in New Zealand participated in the trial. Interventions: Subjects were supplied with some foods rich in the test fats and were given detailed dietary advice which was reinforced regularly. Results: Plasma lathosterol concentration (P < 0.001), the ratio plasma lathosterol/cholesterol (P 0.04), low density lipoprotein (LDL) cholesterol (P < 0.001) and apoB (P < 0.001) levels were signi®cantly different among the diets and were signi®cantly lower during coconut and saf¯ower oil diets compared with butter diets. Plasma total cholesterol, HDL cholesterol and apoA-levels were also signi®cantly (P 0.001) different among the diets and were not signi®cantly different between buffer and coconut diets. Conclusions: These data suggest that cholesterol synthesis is lower during diets rich in coconut fat and saf¯ower oil compared with diets rich in butter and might be associated with lower production rates of apoB-containing lipoproteins. Sponsorship: The study was supported by the Health Research Council of New Zealand and the Anderson Telford Trust. Descriptors: diet; fat type; lathosterol; lipoproteins; polynesians

Introduction SAFA in the diet. A suggested (Glatz & Katan, 1993) explanation for these varying ®ndings is that the gas Diets rich in saturated fatty acids (SAFA) raise plasma chromatographic (GC) separation of faecal plant sterols cholesterol and low density lipoprotein (LDL) cholesterol and endogenous sterols, a vital step in the cholesterol levels compared with diets rich in polyunsaturated fatty balance method of measuring cholesterol synthesis, may acids (PUFA). Saturated fatty acids with different chain not have been achieved in the earlier studies before the lengths vary in their hypercholesterolemic effect in the advent of capillary GC. Cholesterol balance estimates order of increasing potency myristate, palmitate and laurate whole body cholesterol synthesis as the difference between (Denke & Grundy, 1992; Zock et al, 1994; Cox et al, dietary intake and excretion of cholesterol and its metabo- 1995). Differences in plasma LDL cholesterol when these lites in the faeces in the steady state. In view of the fatty acids are predominant in the diet are theoretically due technical dif®culties and the requirement for steady-state to differences in the rates of very low density lipoprotein conditions in the cholesterol balance technique, alternative (VLDL) production and conversion to LDL or the catabol- methods for measuring cholesterol synthesis have been ism of LDL by hepatic receptors or both. There is limited developed. evidence that diets rich in palmitic acid stimulate the Plasma lathosterol levels and the ratio plasma latho- production of very low density lipoproteins (VLDL) and sterol/cholesterol have been established as excellent indices LDL (Cortese et al, 1983). The rate of VLDL secretion of cholesterol synthesis. They correlate closely with human from the liver is closely linked to cholesterol synthesis rates hepatic 3-hydroxy-3-methylglutaryl Coenzyme A (HMG in normolipidemic subjects (Watts et al, 1995). Whether CoA) reductase activity, the rate limiting step in cholesterol cholesterol synthesis varies during diets rich in SAFA of biosynthesis (BjoÈrkhem et al, 1987) and with whole-body different chain lengths has not been widely studied. cholesterol synthesis rates measured by cholesterol balance The effect of changing the quality of dietary fat on (Kempen et al, 1988). Also, levels of the sterol are cholesterol synthesis is controversial. While several studies decreased when cholesterol synthesis is reduced during have indicated that whole-body cholesterol synthesis is treatment with drugs that inhibit the HMG CoA reductase unchanged (Shepherd et al, 1980; Spritz et al, 1965; enzyme (Kempen et al, 1988) and are increased under Grundy & Ahrens, 1970; Nestel et al, 1976) others have conditions of increased cholesterol synthesis (Miettinen, documented decreased (Glatz & Katan, 1993) and 1985). Lathosterol is a cholesterol precursor that is formed increased (Jones et al, 1994) rates when PUFA replace after the rate-limiting step in cholesterol biosynthesis and `leaks' from cells at rates which are proportional to Correspondence: Dr W Sutherland, Department of Medicine, Dunedin School of Medicine, PO Box 913, Dunedin, New Zealand. rates of cellular cholesterol synthesis. Plasma lipoproteins Received 4 December 1997; revised 24 April 1998; accepted 9 May 1998 are acceptors for precursor sterols released from cells. Dietary fat type and lathosterol levels C Cox et al 651 Consequently, the ratio plasma lathosterol/cholesterol has naire. Body weight was measured and fasted venous blood been frequently used to correct for the effect of varying samples were taken on two occasions during the baseline numbers of lipoprotein acceptors on plasma lathosterol period and at weeks four and six of each diet period. levels. When PUFA are substituted for SAFA in the diet Plasma lipids and HDL cholesterol levels were measured in plasma latho-sterol levels and the ratio lathosterol/choles- all blood samples collected. Plasma lathosterol, very low terol are reduced (Glatz & Katan, 1993) suggesting that density lipoprotein (VLDL) cholesterol and apolipopro- cholesterol synthesis is also reduced. teins A±I, A±II and B were measured on one occasion The aim of the present study was to determine the effect during the baseline period and at the four-week point of of diets rich in coconut fat, butter and saf¯ower oil on each diet period. An earlier study from our laboratories plasma levels of lathosterol. The study was conducted in (Cox et al, 1995) suggested that plasma lipid and lipo- Paci®c Islanders because these types of diets have parti- protein levels had attained equilibrium by four weeks after cular relevance in that population. Traditionally, they eat a change in the type of fat in the diet. diets rich in coconut fat but on migration to New Zealand they increase their intake of other types of fat (Stanhope et al, 1981). Diets The experimental diets were designed to provide approximately 17% protein, 47% carbohydrate and 36% fat as a Subjects and methods fraction of total energy intake. Target values for fat intake based on an 8.4 MJ/d diet were 84 g total fat for all diets Subjects and including; 39 g fat from butter plus fat from palmitic Forty-seven Paci®c Islanders from two Paci®c Island acid-rich foods to give an intake of approximately 17 g churches in Dunedin, New Zealand were recruited. Six palmitic acid (butter diet); 39 g fat from coconut oil to give subjects dropped out immediately after initial screening and an intake of approximately 17 g lauric acid (coconut diet); 41 subjects started the dietary phase of the study. The 41 and 24 g fat from saf¯ower oil to give an intake of subjects included 24 men and 17 women ages 19±72 y (31 approximately 17 g linoleic acid (saf¯ower diet). Calcu- Western Samoans and 10 Cook Islanders) with plasma lated intakes of myristic acid and linoleic acid were cholesterol levels between 4.2 mmol/l and 7.5 mmol/l and approximately similar in the coconut and butter diets. To plasma triglycerides less than 3 mmol/l. None was receiv- encourage dietary compliance, participants were supplied ing drugs known to in¯uence lipid metabolism. Character- with butter and cream during the butter diet, coconut istics of the subjects at baseline are shown in Table 1. All cream, coconut oil and monounsaturated margarine subjects gave informed consent and the study was approved during the coconut diet and saf¯ower oil and polyunsatu- by the Ethics committee of the Otago Area Health Board. rated margarine during the saf¯ower diet. Specially baked bread including similar amounts of butter, coconut oil or Study design saf¯ower oil were supplied to the study participants during The study included a run in period of six weeks followed the appropriate diet period. Egg yolk was added to the by three consecutive six week periods when diets rich in coconut diet and the saf¯ower diet to maintain approxi- butter (butter diet), coconut fat (coconut diet) then saf- mately similar cholesterol intake (250±300 mg/d) during ¯ower oil (saf¯ower diet) were consumed without washout the experimental diets. Subjects assigned to the butter diet periods between the experimental diet periods. A sequen- were instructed to eat butter, cream, butter bread, cheese, tial study design rather than the preferred randomized meat, ®sh, chicken, fruit and vegetables, rice, noodles, full crossover design was used to encourage compliance with cream milk and milk products and were instructed not to the diets because participants frequently ate communally eat coconut cream or products, margarine and cooking oils. within the church group. During the run in period partici- Those assigned to the coconut diet were instructed to eat pants completed a food frequency questionnaire to assess coconut cream, coconut bread, monounsaturated margarine, their usual food and nutrient intake and a health question- coconut, ®sh, meat cheese, chicken, fruit and vegetables, rice, noodles, trim milk and other low fat dairy products Table 1 Characteristics of the subjects at baseline and not to eat butter, cream, other margarines, chicken skin, deep-fried takeway foods and vegetable oils. Those Men Women assigned to the saf¯ower diet were instructed to eat poly- Variable n 21 n 16 unsaturated margarine, saf¯ower oil, saf¯ower bread, Age (y) 35 (9) 39 (9) cheese, meat, ®sh, chicken, fruit and vegetables, rice, Body mass index (kg/m2) 29.7 (3.5) 26.8 (5.4) noodles, trim milk, low fat dairy products and not to eat Fasting blood glucose (mmol/l) 5.12 (0.82) 5.66 (1.56) butter, cream, coconut cream and products, cream, reduced Plasma lipids (mmol/l) Total cholesterol 5.47 (0.89) 5.55 (1.05) fat cream and deep-fried takeaway foods. Prior to commen- VLDL cholesterol 0.38 (0.27) 0.30 (0.13) cing the experimental diets and during each appointment LDL cholesterol 3.50 (0.85) 3.64 (0.95) for blood sampling a dietitian provided detailed dietary HDL cholesterol 1.07 (0.26) 1.13 (0.17) instructions which were reinforced during group sessions Triglycerides 1.98 (1.61) 1.50 (0.49) throughout the study. Dietary instruction included ways to Plasma apolipoproteins (g/l) A±I 1.19 (0.19) 1.24 (0.17) incorporate the test fats in the diet and foods to eat and A±II 0.32 (0.09) 0.34 (0.05) foods to avoid during each experimental diet. Menus were B 0.76 (0.22) 0.78 (0.19) suggested and recipes were provided. Dietary compliance Lathosterol (mmol/mmol cholesterol) 1.14 (0.41) 1.14 (0.47) was assessed from the plasma triglyeride fatty acid com- Values are mean (s.d.). position at week 4 of each diet period. All contact with Abbreviations: VLDL, very low density lipoproteins; LDL, low density study participants took place on Sunday mornings in the lipoproteins; HDL, high density lipoproteins. church hall after church service. Dietary fat type and lathosterol levels C Cox et al 652 Laboratory methods study group, the ratio was signi®cantly different among the Subjects fasted overnight and at approximately midday the diet periods and was signi®cantly lower during diets rich in following day (Sunday) venous blood was taken into tubes coconut fat and saf¯ower oil compared with diets rich in containing EDTA. Plasma was separated by low speed butter. Similarly, mean ( Æ s.d.) plasma lathosterol concen- centrifugation of blood at 4C. VLDL was separated by tration was signi®cantly (P 0.0006, repeated measures ultracentrifugation of plasma according to the Lipid ANOVA) different among the diet periods (butter: Research Clinics' protocol (Lipid Research Clinics' Pro- 6.86 Æ 2.72 mmol/l; coconut: 5.83 Æ 1.86 mmol/l; saf¯ower gram, 1974). HDL cholesterol was measured in the super- oil: 5.62 Æ 2.38 mmol/l). natant after precipitation of apoB-containing lipoproteins Plasma levels of LDL cholesterol and apoB were sig- with phosphotungstate and magnesium ions (Assman et al, ni®cantly lower and plasma triglyceride concentration was 1983). Cholesterol (coef®cient of variation 1.6%) and lower at a marginal level of signi®cance at four weeks triglycerides were measured in plasma and plasma frac- during the coconut diet compared with the butter diet. tions using commercial enzymatic kits (Boehringer Man- Plasma cholesterol, LDL cholesterol, HDL cholesterol, nheim, Germany). Plasma apoA±I, and B were measured apoA±I and apoB were signi®cantly lower during the saf- by immunoturbidimetry using commercial kits (Boehrin- ¯ower diet compared with both the butter diet and the ger Mannheim, Germany) and apoA±II was measured by a coconut diet. These results are summarised in Table 3. modi®cation of a published method (Siedel et al, 1988). Plasma cholesterol and triglycerides levels did not change Plasma lathosterol levels were measured by gas±liquid signi®cantly between weeks four and six in the study. The chromatography as described previously (Sutherland et pattern of plasma lipid, lipoprotein and apolipoprotein al, 1991) and values were also expressed as mmol/mmol levels was similar in men and women during the diets. plasma cholesterol to correct for varying numbers of Age was not correlated signi®cantly with plasma choles- lipoprotein acceptor particles for the sterol in plasma. terol (r 0.231) and log ratio plasma lathosterol/choles- The coef®cient of variation for the lathosterol measure- terol (r 0.173) at baseline and during the study. ment was typically 4%. Plasma samples for lathosterol Plasma triglyceride lauric acid content was signi®cantly measurement were stored at 780C until the end of the (P < 0.05) higher during the coconut diet and the content of study and all samples from an individual were measured linoleic acid was signi®cantly higher during the saf¯ower on the same day. Plasma triglyceride fatty acid composi- diet compared with other diets. Plasma triglyceride content tion was also measured by gas±liquid chromatography of myristic acid and palmitic acid were not signi®cantly (Cox et al, 1995). different between the diet periods. These results are summarised in Table 4. Statistical analysis Analyses were carried out on values from the four week points in the diet periods. Data from the three diet periods Discussion were compared by ANOVA and when a signi®cant effect of diet was detected paired t-tests were used to compare In an earlier randomised study in moderately hypercholes- values between pairs of diets. Two-tailed tests of signi®- terolemic subjects we reported that plasma cholesterol and cance were used and a P value of less than 0.05 was LDL cholesterol levels were lower during diets rich in considered to be statistically signi®cant. coconut fat and were lower still during diets rich in saf- ¯ower oil compared with diets rich in butter (Cox et al, 1995). The present data con®rm these ®ndings for plasma Results LDL cholesterol in healthy subjects. Plasma cholesterol Forty-one subjects completed the study and four subjects levels were not clearly lower in spite of reduced LDL did not have plasma lathosterol levels measured for tech- cholesterol levels because levels of plasma HDL choles- nical reasons and their data were excluded from the present terol and VLDL cholesterol tended to increase when coco- report. Body weight (mean Æ s.d.) did not change signi®- nut fat replaced butter in the diet. We also report that cantly during the study (baseline: 91.7 Æ 15.2 kg; 4-week plasma lathosterol and the ratio plasma lathosterol/ values: butter: 92.5 Æ 16.1 kg; coconut: 92.6 Æ 16.4 kg, saf- ¯ower: 92.2 Æ 16.2 kg). Table 3 Plasma lipids, lipoproteins and apolipoproteins in the subjects Table 2 shows plasma lathosterol/cholesterol ratio in during butter, coconut and saf¯ower diets subjects during the dietary intervention periods. In the total n 37 ANOVA Table 2 Plasma lathosterol to cholesterol ratio on subjects during butter, Butter Coconut Saf¯ower P coconut, and saf¯ower oil diets TC (mmol/l) 5.61 (0.96)a 5.47 (0.91)a 5.10 (0.93)b 0.0001 All subjects Men Women VLDL-C (mmol/l) 0.34 (0.18) 0.41 (0.36) 0.45 (0.45) 0.31 n 37 n 21 n 16 LDL-C (mmol/l) 4.08 (0.89)a 3.79 (0.75)b 3.50 (0.84)c 0.0001 HDL-C (mmol/l) 1.16 (0.24)a 1.21 (0.27)a 1.06 (0.21)b 0.001 Diet mmol/mmol cholesterol TG (mmol/l) 1.86 (0.89) 1.61 (0.93) 1.77 (1.25) 0.18 ApoA±I (g/l) 1.23 (0.18)a 1.33 (0.28)a 1.15 (0.14)b 0.0003 Butter 1.24 (0.24)a 1.22 (0.49) 1.27 (0.39) ApoA±II (g/l) 0.34 (0.08) 0.35 (0.08) 0.35 (0.08) 0.31 Coconut 1.09 (0.34)b 1.04 (0.39) 1.11 (0.36) ApoB (g/l) 1.00 (0.22)a 0.87 (0.38)b 0.76 (0.18)b 0.0001 Saf¯ower 1.14 (0.49)b 1.11 (0.57) 1.19 (0.41) ANOVA Pc 0.04 0.08 0.56 Values are mean (s.d.). Abbreviations: TC, total cholesterol; VLDL-C, very low density Values are mean (s.d.). lipoprotein cholesterol; LDL-C, low density lipoprotein; HDL-C, high a,bValues with different superscripts are signi®cantly different by paired density lipoprotein; TG, triglycerides; Apo, apolipoprotein. t-test (P < 0.05). a,bValues in the same row with different superscripts are signi®cantly cRepeated measures analysis of variances on log-transformed data. different by paired t-test (P < 0.01). Dietary fat type and lathosterol levels C Cox et al 653 Table 4 Plasma triglyceride fatty acid composition in subjects during during diets rich in coconut fat compared with those rich in butter, coconut and saf¯ower diets butter. The mechanism underlying a lower rate of choles- Fatty acid Butter Coconut Saf¯ower terol synthesis in response to dietary coconut fat is unre- (mol%) (n 18) (n 18) (n 18) solved by our data. Conceivably, synthesis and secretion of VLDL may be decreased which could lead to reduced b a b C12:0 0.29 (0.36) 0.80 (0.47) 0.51 (0.31) requirement for hepatic cholesterol and in turn decreased C14:0 3.18 (1.12) 2.95 (0.78) 2.74 (0.85) C14:1 0.21 (0.18) 0.52 (0.87) 0.32 (0.37) cholesterol synthesis. In rhesus monkeys, the production C16:0 29.95 (2.89) 29.20 (2.87) 29.90 (3.26) rate of VLDL-apoB was 3-fold lower in animals fed a diet C16:1 4.77 (0.87) 5.35 (0.85)c 4.44 (0.92) rich in coconut fat compared with those fed a diet rich in C18:0 6.34 (1.40) 5.30 (1.32) 5.67 (2.47) palmitic acid in the form of palm oil ((Khosla & Hayes, C18:1 n-9 42.21 (4.34)b 42.60 (2.98)b 38.90 (5.02)a C18:2 n-6 7.46 (3.14)b 6.44 (1.89)b 10.42 (4.60)a 1991). However, the effect of a diet rich in coconut fat on C18:3 n-6 0.11 (0.10)b 0.15 (0.09)a,b 0.17 (0.08)a VLDL secretion in humans has not been reported. C18:3 n-3 0.08 (0.14)b 0.12 (0.18)b 0.34 (0.19)a On the other hand, it is possible that a lower rate of C20:4 n-6 0.18 (0.09)b 0.23 (0.14)b 0.43 (0.14)a cholesterol synthesis during diets rich in coconut fat may C22:6 n-3 0.27 (0.42) 0.15 (0.23) 0.34 (0.30) reduce hepatic secretion of VLDL-apoB rather than the Values are mean (s.d.). reverse. There is evidence that cholesterol synthesis may be a,bValues in the same row with different superscripts are signi®cantly an important determinant of VLDL-apoB production. The different by paired t-test (P < 0.05). correlation between cholesterol synthesis and VLDL-apoB secretion reported previously may indicate that cholesterol synthesis is a determinant of VLDL production (Watts et cholesterol were lower which may indicate lower choles- al, 1995). Also, Arad and coworkers (1992) have reported terol synthesis during diets in coconut fat and saf¯ower oil that inhibition of HMG CoA reductase activity, the rate- compared with diets rich in butter. limiting step in cholesterol biosynthesis, by lovastatin Since this study was conducted in free-living individuals treatment appears to reduce the assembly and secretion of reduced compliance with experimental diet regimen could VLDL in subjects with hyperlipidemia. Lastly, rates of have potentially in¯uenced results. Participants were given cholesteryl ester synthesis and apoB secretion are closely detailed dietary advice and supplies of the test fats and coupled in cultured HepG2 cells incubated with fatty acids special breads baked using these fats. Furthermore, they (Cian¯one et al, 1990). were contacted frequently to reinforce dietary instructions. It is unlikely that lathosterol in the diet in¯uenced Plasma triglyceride fatty acid composition which showed appreciably the differences in plasma lathosterol and the highest levels of lauric acid during the coconut diet, highest ratio lathosterol/cholesterol during this study. Eggs, which levels of linoleic acid during the saf¯ower oil diet and a appear to be the main source of lathosterol in the diet trend toward highest levels of stearic acid during the butter (Duane, 1995), were added to the diets rich in coconut fat diet suggests that on average subjects complied satisfacto- and saf¯ower oil during which plasma lathosterol levels rily with dietary instructions. and the ratio lathosterol/cholesterol were lower compared Glatz & Katan (1993) have reported a higher ratio with the butter diets. Furthermore, cholesterol (egg) intake plasma lathosterol/cholesterol and a higher rate of whole was maintained low and in a range which probably has body cholesterol synthesis in subjects consuming diets rich little in¯uence on plasma lathosterol levels. in saturated fat compared with polyunsaturated fats. Our Since LDL is derived from the intravascular hydrolysis of data in the present study and in an earlier study (Cox et al, VLDL, any decrease in VLDL±apoB secretion could, in 1996) are in accord with this ®nding. Also, the magnitude theory, contribute to the lower plasma LDL cholesterol of the differences in the ratio plasma lathosterol/cholesterol levels during diets rich in cocount fat compared with diets (0.10 mol/mol) and mean plasma cholesterol concentration rich in butter. However, Watts and coworkers have reported (0.52 mmol/l) between the butter diet and the diet rich in that plasma LDL cholesterol concentration and VLDL±apoB saf¯ower oil are approximately in proportion to the corre- secretion rates in normolipidemic men are not correlated sponding differences (ratio plasma lathosterol/cholesterol, (Watts et al, 1995) which suggests that VLDL±apoB secre- 0.21 mol/mol; mean plasma cholesterol, 0.93 mmol/l) tion rate may not be a major determinant of plasma LDL reported between SAFA and PUFA diets (Glatz & Katan, cholesterol levels in man. An increase in LDL receptor 1993). These ®ndings suggest that the average magnitude number is probably an important mechanism by which of the decrease in cholesterol synthesis may predict the reduction of dietary saturated fat reduces plasma LDL average magnitude of the decrease in plasma cholesterol cholesterol levels (Mustad et al, 1997). While hepatic LDL levels when PUFA replaces SAFA in the diet. receptor activity (and plasma cholesterol levels) is similar in It has been postulated previously that the higher rate of hamsters fed diets rich in lauric acid or palmitic acid cholesterol synthesis during diets rich in SAFA may be due (Woollett et al, 1992), the effect of substituting butter with to increased synthesis and secretion of VLDL from the liver coconut fat in the diet on LDL receptor numbers in man has (Glatz & Katan, 1993). Synthesis of VLDL is a major drain not been determined. Reduced intake of SAFA may lead to on the hepatic cholesterol pool (Shepherd & Packard, 1988) higher numbers of LDL receptors (Mustad et al, 1997) which and apparently activates cholesterol synthesis to maintain a may contribute substantially to lower plasma cholesterol and suf®cient supply of the sterol for lipoprotein synthesis (Goh LDL cholesterol levels during diets rich in saf¯ower oil & Heimberg, 1979). A close correlation between choles- compared with those rich in coconut fat (and butter) in the terol synthesis rates and VLDL-apoB secretion in normo- present study. However, the ratio plasma lathosterol/choles- lipidaemic subjects (Watts et al, 1995) is in line with a link terol was similar during these diets which suggests that the between VLDL production and cholesterol synthesis. decrease in plasma cholesterol levels when dietary saf¯ower In the present study, the lower ratio plasma lathosterol/ oil replaced coconut fat is not linked to cholesterol synthesis cholesterol suggests that cholesterol synthesis is reduced or associated metabolic pathways. Dietary fat type and lathosterol levels C Cox et al 654 There are limitations to the present ®ndings. Subjects Glatz JFC & Katan MB (1993): Dietary saturated fatty acids increase were not randomized into diet groups and a sequential cholesterol synthesis and fecal steroid excretion in healthy men and women. Eur. J. Clin. Invest. 23, 648±655. study design was used. Therefore, it could be argued that Goh EH & Heimberg M (1979): Relationship between the activity of 3- order of the diets and changes with time including regres- hydroxy-3-methylglutaryl Coenzyme A reductase and secretion of very sion to the mean could affect the results. However, it is low density lipoprotein cholesterol in isolated perfused liver and in the unlikely that these factors were in¯uential because similar mutant rat. Biochem J 184, 1±6. changes in plasma lipids, lipoproteins and apolipoproteins Grundy SM & Ahrens EH Jr (1970): The effects of unsaturated dietary fats on absorption, excretion, synthesis and distribution of cholesterol in were recorded in our previous study in subjects who man. J. Clin. Invest. 49, 1135±1152. consumed diets rich in coconut fat, butter and saf¯ower Jones PJH, Lichtenstein AH, Schaefer EJ & Namchuk GL (1994): Effect oil in a randomised cross-over design (Cox et al, 1995). In of dietary fat selection on plasma cholesterol synthesis in older, the present study, the switch from diets rich in butter to moderately hypercholesterolemic humans. Arterioscler. Thromb. 14, coconut fat diets (the main comparison) without a washout 542±548. Kempen HJM, Glatz JFC, Gevers Leuven JA, van der Voort HA & period would tend to minimize the effects of this change in Katan MB (1988): Serum lathosterol concentration is an indicator of dietary fat on measured variables. whole-body cholesterol synthesis in humans. J. Lipid Res. 29, 1149±1155. Khosla P & Hayes KC (1991): Dietary fat saturation in rhesus Conclusions monkeys affect LDL concentrations by modulating the independent production of LDL apolipoprotein B. Biochim. Biophys. Acta 1083, Our data suggest that cholesterol synthesis is reduced 46±56. leading to lower plasma lathosterol levels when butter is Lipid Research Clinics' Program (1974): Manual of Laboratory Opera- replaced by coconut fat in the diet. The mechanism under- tions. Lipids and Lipoproteins Analysis. DHEW Publication No. (NIH) lying this change and its relationship to changes in lipo- 75±628, Bethesda, MD, pp. 51±56. Miettinen TA (1985): Cholesterol precursors and their diurnal rhythm in protein metabolism when the composition of dietary fat is populations of patients with jejuno-ileal bypass and ileal dysfunction. altered remains to be determined. Metabolism 34, 425±430. Mustad VA, Etherton TD, Cooper AD, Mastro AM, Pearson TA, AcknowledgementsÐThe authors greatly appreciate the cooperation of Jonnalagadda SS & Kris-Etherton PM (1997): Reducing saturated participants in the study and the excellent technical and research assistance fat intake is associated with increased levels of LDL receptors on mononuclear cells in healthy men and women. J. Lipid Res. 38, of Ashley Duncan, Dean Hackett, Barbara McSkimming, and Margaret 459±468. Waldron. Nestel PJ, Homma Y, Scott TW, Cook LJ & Havenstein N (1976): Effect of dietary unsaturated pork on plasma lipid and sterol excretion in man. Lipids 11, 42±48. References Shepherd J, Packard CJ, Grundy SM, Yeshurun D, Gotto, AM & Taunton Arad Y, Ramakrishnan R & Ginsberg HN (1992): Effects of lovastatin OD (1980): Effects of saturated and polyunsaturated fat diets on the therapy on very-low density lipoprotein triglyceride metabolism in chemical composition and metabolism of low density lipoproteins in subjects with combined hyperlipidemia: Evidence for reduced assembly man. J. Lipid Res. 21, 91±99. and secretion of triglyceride-rich lipoproteins. Metabolism 41, 487±493. Shepherd J & Packard J (1988): Pharmacological control of plasma Assman G, Schriewer H, Schmidt G & Hagele E (1983): Quanti®cation of cholesterol: Mechanisms of action of hypocholesterolaemic agents. In high density lipoprotein cholesterol by precipitation with phosphotungs- Atherosclerosis Reviews, eds. J. Stokes III & M Mancini, p. 109. New York: Raven Press. tic acid/MgCl2. Clin. Chem. 29, 2026±2030. BjoÈrkhem I, Miettinen T, Reihner E, Ewerth S, Angelin B & Einarsson K Siedel J, Schlefer S, Rosseneu M, Bergeaud R, De Keersgieter W, Pautz B, (1987): Correlation between serum levels of some cholesterol precur- Vinaimont N & Ziegenhorn J (1988): Immunoturbidimetric method for sors and activity of HMG CoA reductase in human liver. J. Lipid Res. routine determinations of apolipoproteins A±I, A±II, and B in normo- 28, 1137±1143. and hyperlipidemic sera compared with immunonephelometry. Clin. Cian¯one KM, Yasruel Z, Rodriguez MA, Vas D & Sniderman AD Chem. 34, 1821±1825. (1990): Regulation of apoB secretion from HepG2 cells: Evidence for Spritz M, Ahrens EH Jr & Grundy SM (1965): Sterol balance in man as a critical role of cholesterol ester synthesis in the response to a fatty acid plasma cholesterol concentrations in man are altered by exchange of challenge. J. Lipid Res. 31, 2045±2055. dietary fat. J. Clin. Invest. 44, 1482±1493. Cortese C, Levy Y, Janus ED, Turner PR, Rao SN, Miller NE & Lewis B Stanhope JM, Sampson VM & Prior AM (1981): The Tokelau island (1983): Modes of action of lipid-lowering diets in man: studies of migrant study: serum lipid concentrations in two environments. apolipoprotein B kinetics in relation to fat consumption and dietary fat J. Chronic. Dis. 34, 45±55. composition. Eur. J. Clin. Invest. 13, 79±85. Sutherland WHF, Robertson MC, Williamson SA & Nye ER (1991): Cox C, Mann J, Sutherland W, Chisholm A & Skeaff M (1995): Effects of Plasma noncholesterol sterols in male distance runners and sedentary coconut oil, butter and saf¯ower oil on lipids and lipoproteins in persons men. Eur. J. Appl. Physiol. 63, 119±123. with moderately elevated cholesterol levels. J. Lipid Res. 36, Watts GF, Naoumova R, Cummings MH, Umpleby AM, Slavin BM, 1787±1795. SoÈnksen PH & Thompson GR (1995): Direct correlation between Cox CMA, Sutherland WHF, Ball MJ & Mann JI (1996): Response of cholesterol synthesis and hepatic secretion of apolipoprotein B-100 in plasma lathosterol concentration to change in the quality of dietary fat normolipidemic subjects. Metabolism 44, 1052±1057. in men and women. Eur. J. Clin. Nutr. 50, 358±363. Woollett LA, Spady DK & Dietschy JM (1992): Regulatory effects of the Denke MA & Grundy SM (1992): Comparison of effects of lauric acid and saturated fatty acids 6:0 through 18:0 on hepatic low density lipoprotein palmitic acid on plasma lipids and lipoproteins. Am. J. Clin. Nutr. 56, receptor activity in the hamster. J. Clin. Invest. 89, 895±898. 1133±1141. Duane WC (1995): Serum lathosterol levels in human subjects re¯ect Zock PL, de Vries HM & Katan MB (1994): Impact of myristic acid versus changes in whole-body cholesterol synthesis induced by lovastatin but palmitic acid on serum lipid and lipoproteins in healthy men and not dietary cholesterol. J. Lipid Res 36, 343±348. women. Arterioscler. Thromb. 14, 567±575.