点击显示 收起
1 From the Institute on Nutraceuticals and Functional Foods (SD, PYC, LC, JD, BL, and NB) and the Departments of Food Science and Nutrition (SD, IG, LC, BL, and NB) and Animal Sciences (PYC and JD), Laval University, Québec, Canada; the College of Pharmacy, Touro University–California, Vallejo, CA (NB); and the Lipid Research Center, Centre Hospitalier de l’Université Laval Research Center, Québec, Canada (BL and PYC)
2 Supported by the Dairy Farmers of Canada, Natural Sciences and Engineering Research Council of Canada, FCAR Fund, Novalait Inc, the Quebec Ministry of Agriculture, Fisheries and Food, and the Canada Research Chair in Nutrition, Functional Foods and Cardiovascular Health. 3 Reprints not available. Address correspondence to N Bergeron, College of Pharmacy, Touro University–California, Vallejo, CA 94592. E-mail: nbergeron{at}touro.edu.
ABSTRACT
Background: Dietary conjugated linoleic acid (CLA) is known to reduce atherosclerosis, plasma total and LDL-cholesterol concentrations, and body fat accumulation in several animal species. Of the few studies that investigated the effects of CLA supplementation in humans, all used commercially formulated oral supplements made from a mixture of CLA isomers.
Objective: We compared the effects on plasma lipoproteins and body composition of the consumption of a modified butter naturally enriched with CLA (CLA-B: 4.22 g CLA/100 g butter fat) by the addition of sunflower oil to the diet of dairy cows with the consumption of a control butter (CON-B) that was low in CLA (0.38 g CLA/100 g butter fat).
Design: In a crossover design study including an 8-wk washout period, 16 men [
Conclusion: These results suggest that a 10-fold CLA enrichment of butter fat does not induce beneficial metabolic effects in overweight or obese men.
Key Words: Conjugated linoleic acids • lipid profile • plasma lipids • lipoproteins • body composition • obesity • functional foods • C-reactive protein • LDL size
INTRODUCTION
Conjugated linoleic acid (CLA) is a term used to describe positional and geometric derivatives of linoleic acid containing conjugated double bonds. CLA is a group of naturally occurring fatty acids that are mainly present in foods from ruminant sources. In contrast to other fatty acids, which are usually present in gram quantities, CLA are present only in milligram quantities in meats and dairy products (1). Most milk-fat CLA is synthesized endogenously via -9 desaturase from trans-vaccenic acid, an intermediate in the biohydrogenation of linoleic and linolenic acids in the rumen. The remainder of the CLA in milk fat arises directly from CLA absorbed from the digestive tract after being produced in the rumen as an intermediate in linoleic acid biohydrogenation (2). Interest in the potentially therapeutic effects of CLA can be traced back to in vitro investigations by Pariza et al (3) that showed the presence of CLA’s mutagenesis-inhibitory activity in extracts from fried ground beef. Subsequent studies established that the extract exhibited anticarcinogenic activity as well and that CLA was the active component responsible for these effects (4). Those first studies of CLA led to hundreds of investigations and the recognition that CLA has multiple biological effects, such as reductions in atherosclerosis (5–7), plasma lipoproteins and lipids (5–8), and body fat accumulation (9–13). CLA has also been reported to have anticarcinogenic activity (14), antiinflammatory effects (15), and antidiabetic effects (16).
To the best of our knowledge, the few studies that have investigated the effects of CLA in humans were conducted by using commercially formulated oral supplements made from a mixture of synthetic CLA isomers (17–23), which have the advantage of providing high doses of CLA (range: 3–7 g/d). In contrast, naturally occurring CLA in common foods are present only in small quantities. The extent to which various amounts of CLA provided in the form of food can affect the health profile of humans is unknown. Therefore, the current study was conducted to test the hypothesis that the incorporation of a modified butter that is naturally rich in CLA into a well-controlled experimental diet induced beneficial effects on the blood lipid profiles of healthy overweight and obese men. In addition, because consumption of CLA has been found to induce a wide range of beneficial health effects in animals, the effects of the experimental diets on variables related to body composition and to nontraditional cardiovascular disease risk factors such as LDL peak particle diameter and C-reactive protein (CRP) were evaluated as secondary outcome measures. To test our hypothesis that a modified butter naturally rich in CLA benefits the lipid profile, we strictly controlled dietary intake by providing all food to the subjects for the duration of the 4-wk dietary interventions.
SUBJECTS AND METHODS
Subjects
Seventeen overweight and obese (BMI: 26.0–29.9 and >30.0, respectively; 24) men in good health were recruited in the Quebec City area to participate in the nutrition study, and 16 men completed the study. The subject who withdrew completed the first phase of the study but could not come back for the second phase because of relocation for his job. Of the 16 men who completed the study, all were white, and most were Canadian. Subjects were initially screened on the basis of a complete physical examination and medical history. Subjects recruited for the study had to be nonsmokers, to be between 18 and 55 y of age, and to have a BMI >26 and a waist circumference >90 cm. We selected both overweight and obese men as a target population that would benefit from CLA supplementation because of studies showing that 90% of men with a BMI >27 also have adipose tissue areas >100 cm2, a threshold value above which moderate alterations in metabolic variables predictive of coronary heart disease are thought to begin occurring (25). None of the participants took medication. Exclusion criteria included the presence of a monogenic dyslipoproteinemia; use of medication known to affect lipid metabolism; chronic, metabolic, or acute disease; and significant weight change in the 6 mo before the experiment. Subjects with regular alcohol intake (>1 drink/d or >7 drinks/wk), unusual dietary habits such as vegetarianism, food allergies, or a dislike for foods included in the experimental diets were also excluded.
The study protocol was fully explained to the participants, who gave their written informed consent. The protocol was approved by the Clinical Research Ethics Committee of Laval University.
Test fats
The 2 experimental fats compared in the present study were a modified butter (CLA-B) that is enriched in CLA and a control butter (CON-B) that is low in CLA. Both butters were manufactured from milk produced by dairy cows from a Laval University dairy herd. To produce the experimental butters, cows were first fed a total mixed diet composed of concentrates plus corn and grass silages as the roughage source. After 3 wk, milk samples were obtained, and the CLA content of milk fat was determined. Cows with the lowest concentrations of CLA were identified, and their milk was collected to make CON-B. Once the production of CON-B was completed, the same cows were fed a similar diet to which 5% sunflower oil was added. Milk samples were taken after 3 wk, and cows with the greatest concentration of CLA continued consuming the diet for the purpose of milk collection and the manufacture of the CLA-B, which was rich in CLA. For both butters, raw milk was separated into cream and skim milk (Westfalia Separator AG, Oelde, Germany). Cream was then immediately pasteurized at 75 °C for 16 s by using a plate heat exchanger (type P20-HB; Alpha-Laval, Lund, Sweden) and stored at 14 °C for 24 h. The control and high-CLA creams were churned at 14 °C until butter granules were formed. The buttermilk was then drained off, and the butter was salted (2%), transferred to 0.5-kg plastic containers, and kept at –20 °C until use.
Dietary intervention
Before the beginning of the 2 experimental phases, a registered dietitian instructed the selected participants in how to complete a 3-d (2 weekdays and 1 weekend day) food intake record to estimate their usual energy intakes and thereby set proper energy levels for the experimental period. Subjects began the study at the energy level closest to their usual energy intake. Because fluctuations in body weight can affect lipoprotein metabolism, weight was monitored every weekday before lunch, and the energy level was adjusted if body weight was found to have fluctuated >2.0 kg from baseline. During the daily attendance of the subjects at the metabolic kitchen, staffers were present to monitor consumption of the meals and to strongly encourage continued dietary compliance. Subjects were also provided with a daily reminder sheet on which they were required to check all food items consumed and to report any deviation from the diet, as well as any illness or use of medication. Any use of over-the-counter medications required prior approval by the study physician.
The 2 experimental diets were designed to be identical in food composition, except for the test fat, which was either CLA-B (rich in CLA) or CON-B (low in CLA). Both diets provided 40% of energy as fat, >60% of which was derived from experimental fats (Table 1). The experimental butters were incorporated in a variety of recipes, such as those for muffins, cakes, and sauces, and they were also used as a spread. Protein sources included boneless chicken breast (4 meals/wk), extra-lean pork (5 meals/wk), fish (2 meals/wk), veal (1 meal/wk), eggs (1 meal/wk), lean ham (1 meal/wk), and tofu (1 meal/wk). Of possible milk products, only fat-free milk, non-fat yogurt, and 1% fat cottage cheese were used, so that almost no milk fat sources besides the butter were present in the experimental diets. Seven-day cycle menus were developed for each experimental diet and were designed to supply the daily recommended allowances for essential nutrients of the Institute of Medicine, National Academy of Sciences (26). The nutritional composition of the diets and the dietary intake from the 3-d food intake records were assessed with the use of the Canadian government’s Nutrient File database (Health Canada, Ottawa, Canada, 1997) and the NUTRITION DATA SYSTEM FOR RESEARCH (NDS-R) software (version 4.03_31; Nutrition Coordinating Center, Minneapolis, MN).
View this table:
TABLE 1. Nutritional composition of habitual and experimental diets1
Because the diets were administered under conditions in which body weight was to be held constant, subjects were asked to eat all the food and only the food that was provided to them by the metabolic kitchen staff for the duration of the 4-wk experimental periods. Consumption of alcohol 1 wk before and during each experimental period was forbidden. Consumption of caffeinated beverages such as soft drinks, coffee, and tea was limited to a maximum of 2 drinks/d; the portion size for soft drinks was a 16-oz can (474 mL), and that for coffee or tea was 8 oz (237 mL). On weekdays, subjects came to the metabolic kitchen daily to consume their lunch meal, and they were then given their next dinner and breakfast meals in a package to take home. Weekend meals were distributed to the participants on Fridays.
Experimental design
A crossover design was used to compare the effects of the CLA-B diet that was rich in CLA with those of the CON-B diet that was low in CLA on plasma lipids, lipoproteins, and body composition. Eight subjects were randomly assigned to the CLA-B diet for the first 4-wk period and to the CON-B diet for the second 4-wk period. The other 8 subjects were assigned to the nutritional treatment sequence in reverse. Experimental periods were separated by an 8-wk washout period, during which the subjects resumed their usual diets to remove the residual effects of the preceding experimental diet on the tested variables. Subjects were blinded to dietary assignments and were not informed of their lipid or body-composition responses until the study was completed. Principal investigators and the laboratory technicians were also blinded to dietary assignments. Throughout the study, participants were asked to maintain their usual level of physical activity, which was evaluated by a weekly questionnaire completed by the subjects.
Fatty acid analyses
The fatty acid composition of the experimental fats was analyzed by using a gas chromatograph (HP 5890 chromatograph; Hewlett-Packard Co, Palo Alto, CA) equipped with a 60-m DB-23 capillary column (internal diameter: 0.32 mm; film thickness: 0.25 µm film thickness; J and W Scientific, Folsom, CA) and a flame ionization detector. At the time of the sample injection, the column temperature was 150 °C, and it was then ramped up at 5 °/min to 200 °C. Inlet and detector temperatures were 240 and 250 °C, respectively. The split ratio was 100:1. The flow rate for hydrogen carrier gas (Praxair Inc, Vanier, Canada) was 2.8 mL/min. Peak area was measured using a NELSON ANALYTIC SYSTEM 2600 (version 5; PE Nelson, Cupertino, CA). Each peak was identified with the use of pure methyl ester standards (Alltech, Deerfield, IL) on the basis of their retention times. The area-to-concentration ratio for all identified fatty acids was used to determine their respective concentrations after adjustment for the difference in molecular mass between the fatty acids and their methyl esters (27).
The CLA isomers in milk fat were analyzed with silver ion-HPLC according to published procedures by using 3 ChromShper 5 Lipids columns in series (ChromPack, Bridgewater, NJ) (28). The trans 18:1 isomers were separated with the use of silver ion-TLC and analyzed by using gas chromatography with a 100-m CP Sil 88 capillary column (ChromPack, Middelburg, Netherlands) (29, 30).
Composition of test fats
CLA-B was characterized by having more than 10 times as much CLA (4.22 g CLA/100 g fatty acids) as did CON-B (0.38 g CLA/100 g fatty acids). Cholesterol concentrations in CON-B and CLA-B were 1.9 and 2.5 mg/g, respectively. A similarly greater cholesterol content was previously observed in butter made from the milk of cows fed sunflower seeds than in butter made from the milk of cows fed a low-fat control diet (31). This effect has been attributed to a greater amount of small fat globules due to the depression of fat in the milk from cows receiving sunflower seeds. A large proportion of cholesterol in milk fat is found at the level of the fat globule membrane. Reducing the size of the fat globule may have increased the membrane surface area and, consequently, the concentration of cholesterol in butter. The size of the fat globules was not ascertained in the current experiment, but the milk-fat content was actually 17% less in cows fed the control diet than in those fed the sunflower oil diet. Egg yolk powder was therefore added to the CON-B diet to make the 2 diets more similar in their cholesterol content.
The fatty acids provided by the 2 butters also were slightly different (Table 2). CLA-B contained more oleic acid (18:1) than did CON-B, and CON-B contained more saturated fatty acids (12:0, 14:0, and 16:0) than did CLA-B. To make the experimental diets containing these butters more comparable in their content of fatty acids (other than CLA), olive oil was added to the CON-B diet, and coconut oil and palm stearin were added to the CLA-B diet. The resulting fatty acid composition of both diets is shown in Table 2.
View this table:
TABLE 2. Fatty acid composition of the 2 experimental butters, the added oils, and the resulting experimental fats1
The relative isomeric distribution of CLA isomers in both experimental butters is shown in Figure 1. As could be expected from available data in the literature, the most abundant CLA isomer present in CON-B and CLA-B was cis-9, trans-11. Increasing the CLA content of the butter resulted in a smaller proportion of the trans-10, cis-12 isomer. The relative distribution of the trans octadecenoic acid (trans 18:1) isomers shown in Figure 2 indicates that the 18:1 trans-11 isomer (trans-vaccenic acid), a suspected precursor of cis-9, trans-11 CLA isomer, accounted for >50% of the 18:1 trans fatty acids in CLA-B.
FIGURE 1.. Relative isomeric distribution of conjugated linoleic acid (CLA) isomers in the control butter () and CLA butter ().
FIGURE 2.. Relative isomeric distribution of trans octadecenoic acid (trans 18:1) isomers in the control butter () and conjugated linoleic acid butter ().
Anthropometric measurements and computed tomography
At the beginning (day 1) and at the end (day 28) of each experimental period, waist circumference was measured midway between the lowest rib and the iliac crest by using a standard tape measure (32). Height and body weight were measured according to the procedures recommended at the Airlie Conference (33). Measures of variability for waist circumference, height, and body weight were ± 0.1 cm, ± 0.1 cm, and ± 0.1 kg, respectively. In the week after each experimental period (day 30 or 31), visceral adipose tissue accumulation (measure of variability: ± 0.1 mm) was assessed by computed tomography, which was performed on a Siemens Somatom DRH scanner (Erlangen, Germany) and analyzed as described previously by Després and al (34).
Separation and analysis of plasma lipoproteins
Fasting blood samples (12-h fast) were collected from an antecubital vein into evacuated tubes containing disodium EDTA at the beginning (day 1) and at the end (day 28) of each experimental period. Samples were then immediately centrifuged at 4 °C for 10 min at 1500 x g to obtain plasma samples, which were then stored at 4 °C until processed.
Triacylglycerol-rich lipoproteins (VLDL; density <1.006 g/mL) were isolated by ultracentrifugation. HDL particles were obtained after precipitation of LDL in the infranatant fluid (density >1.006 g/mL) with heparin and manganese chloride (35). HDL2 and HDL3 subfractions were separated by dextran-sulfate precipitation (36). Fasting concentrations of plasma and lipoprotein cholesterol and triacylglycerols were measured enzymatically by using an RA-500 analyzer, as previously described (37). LDL apolipoprotein (apo) B and HDL apo A-I were measured by rocket immunoelectrophoresis, as previously described by Laurell et al (35). The CV for cholesterol, triacylglycerols, and apolipoprotein measurements was <5% (38).
Distinct subpopulations of LDL particles in whole plasma were separated by size with the use of nondenaturing 2–16% gradient gel electrophoresis (39). Particle size was quantified by densitometric scanning of Sudan Black–stained gels, using IMAGE MASTER 1 D Prime software (version 3.01; Amersham Pharmacia Biotech, Baie d’Urfé, Canada). LDL peak particle diameter was identified as the predominant subclass of LDL in each subject and was calculated from calibration curves by using plasma standards of known diameter. The CV of the calculated particle diameters was estimated as <0.6%.
Leptin and C-reactive protein concentrations
Plasma leptin concentrations were determined with the use of a highly sensitive, commercial, double-antibody radioimmunoassay (Human Leptin Specific RIA Kit; Linco, St Louis, MO) that detects leptin concentrations 0.5 ng/mL. The CV for the repeated assays ranged from 4.0% to 5.5% and from 6.5% to 8.5% for lower and higher plasma leptin concentrations, respectively (40). Plasma CRP concentrations were measured by using a commercially available, highly sensitive immunoassay with a monoclonal antibody coated with polystyrene particles (Behring Latex Enhanced on the Behring Nephelometer BN-100; Behring Diagnostic, Westwood, MA) as described previously (41). The run-to-run CV at CRP concentrations ranging from 1.0 to 10 µg/mL was <5%.
Statistical analysis
Data were analyzed by using SAS software (version 8.2; SAS Institute Inc, Cary, NC). Repeated-measures analysis of variance, adjusted for crossover designs, using the general linear model was performed to identify differences between experimental treatments (42). In the analysis of variance model, the carryover effects for each of the variables studied were tested by introducing a term referring to the sequence in which the dietary treatments were given. Although no significant sequence effect was found, a significant interaction was observed between treatment and sequence for VLDL-triacylglycerol and LDL apo B, which suggested a different response to the 2 diets across time. Therefore, only the results of the first period were analyzed for those 2 variables. Paired t tests were used to identify differences within the experimental diets only when significant differences were found between dietary treatments. Group averages are reported as means ± SDs. CRP values were logarithmically transformed before statistical analysis to achieve normal distribution. Two subjects who were found, during the course of the study, to have CRP concentrations >10 mg/L, which suggested the presence of bacterial infection or inflammation, were excluded from statistical analysis for that variable (43). Differences were considered significant at P < 0.05.
RESULTS
Characteristics of subjects at baseline
The 16 participants were overweight or obese, but their mean blood lipid profiles were within the normal range (44); that is, the HDL-cholesterol concentrations were slightly low and triacylglycerol concentrations were slightly high (Table 3). Data from the 3-d food record completed by the subjects before study onset were compiled and compared with the nutritional composition of the experimental diets (Table 1). That some subjects lost weight despite the fact that their experimental energy intake was higher than their reported habitual energy intake (Table 1) suggested that, on average, subjects underestimated their energy intake, a behavior that has been reported in 50% of obese persons (45).
View this table:
TABLE 3. Physical characteristics and plasma lipid profile of the subjects at baseline1
Anthropometric measures and body composition
The diet-induced variations in body weight, waist circumference, and plasma leptin concentrations did not differ significantly between the 2 groups (Table 4). These findings are consistent with those for body fat distribution, which, although taken only at the end of each experimental diet to minimize the subjects’ exposure to radiation, showed that there was no significant difference in the accumulation of visceral and subcutaneous adipose tissue between the diet groups.
View this table:
TABLE 4. Body-composition variables and plasma leptin concentrations at the beginning (day 1) and at the end (day 28) of each diet intervention1
Plasma and lipoprotein cholesterol
When compared with the CON-B diet, the CLA-B diet resulted in a significantly smaller reduction in plasma total cholesterol, but the changes in VLDL, LDL, and HDL cholesterol and HDL subfractions did not differ significantly between the 2 diets (Table 5). The greater reduction in total cholesterol after the CON-B diet was accompanied by a significantly greater decrease in total:HDL and LDL:HDL cholesterol after the CON-B diet than after the CLA-B diet.
View this table:
TABLE 5. Fasting plasma and lipoprotein cholesterol concentrations at the beginning (day 1) and at the end (day 28) of each diet intervention1
Plasma and lipoprotein lipids and apolipoproteins
The magnitude of the changes in plasma total cholesterol and VLDL, LDL, and HDL triacylglycerols did not differ significantly between the 2 diets (Table 6). The CON-B diet resulted in a significantly greater reduction in plasma apo B concentrations than did the CLA-B diet. The magnitude of the variation in VLDL apo B, LDL apo B, and HDL apo A-I concentrations did not differ between the 2 groups (Table 7).
View this table:
TABLE 6. Fasting plasma and lipoprotein triacylglycerol concentrations at the beginning (day 1) and at the end (day 28) of each diet intervention1
View this table:
TABLE 7. Mean concentrations of apolipoproteins at the beginning (day 1) and at the end (day 28) of each diet intervention1
LDL peak particle diameter and plasma CRP concentrations
Neither diet affected the LDL peak particle diameter or the concentration of plasma CRP, a marker of low-to-moderate systemic inflammation (Table 8).
View this table:
TABLE 8. LDL peak particle diameter and plasma C-reactive protein (CRP) concentrations at the beginning (day 1) and at the end (day 28) of each diet intervention1
DISCUSSION
Improvements in the lipid profile after CLA supplementation were reported in several animal species (7, 8, 46), but little is known about the effect of CLA, particularly when CLA is provided in the form of naturally occurring foods, on the lipid profile of humans. The purpose of our study thus was to compare the effects of a CLA-B, naturally enriched in CLA by the addition of sunflower oil to the diet of dairy cows, and a CON-B, low in CLA, on body composition, the lipid profile, and other cardiovascular disease risk factors in overweight and obese men.
We found that, in overweight and obese men, in contrast with observations in animals, the ingestion of the CLA-B diet for 4 wk did not improve the lipid profile more than did the ingestion of the CON-B diet. If anything, the CLA-B diet resulted in a significantly smaller reduction in plasma cholesterol and in total:HDL cholesterol, presumably in part because of the tendency of the CLA-B diet to induce a smaller reduction in VLDL cholesterol than did the CON-B diet. It is noteworthy that a diet that contained 40% of energy as fat and that was relatively high in saturated fatty acids did not induce the negative effects so often associated with this type of dietary habit. It is interesting that obese men have been found not only to underreport their daily energy intakes but also to selectively underreport their fat intakes (47). This suggests that the habitual dietary intakes of our study participants, calculated to be 34.9%, may have been underreported, which could partially explain why the consumption of an experimental diet providing 40% of energy as fat did not lead to a worsening of their lipoprotein or lipid profile. Although studies by others of the effects of CLA in humans (20, 48) remain scarce, their findings are globally consistent with those in our current study.
The lack of amelioration in the health profile in the current study could be related in part to the CLA isomers used. We used a modified butter naturally rich in CLA, in which nearly 80% of total CLA was in the form of cis-9, trans-11 CLA (Figure 1). In more recent animal studies, the desirable effects of CLA on blood lipids and body composition were associated with the trans-10, cis-12 isomer rather than with the cis-9, trans-11 isomer (8, 12, 49–51). Data on the potential benefit of the trans-10, cis-12 CLA isomer for the lipid profile in humans remain scarce, but human studies reported no improvement in blood lipids after supplementation with either trans-10, cis-12 or cis-9, trans-11 CLA isomers (52, 53).
The lack of improvement in the lipid profile after the CLA-B diet could also be partly explained by the stereospecific distribution of fatty acids on the triacylglycerol, which has been suggested to affect the processing of dietary fats. Indeed, the hydrolysis of triacylglycerols by lipoprotein lipase and the uptake of remnant particles by the liver were reported to be slower when the fatty acid in the sn-2 position is saturated than when it is unsaturated (54). Furthermore, fatty acids seem to be absorbed better when they are in the sn-2 position than when they are in the sn-1 or sn-3 position (55, 56). In the current study, palm stearin was added to the CLA-B diet to increase its palmitic acid content to make it more comparable to that of the CON-B diet. However, up to 59% of palmitic acid in palm stearin is in the sn-2 position (57, 58), whereas only 42% of the palmitic acid in milk fat takes up that position (59). Incorporating palm stearin, which contains more cholesterol-raising palmitic acid (60) in the sn-2 position, to the CLA-B may thus have rendered it less likely to improve plasma and LDL-cholesterol concentrations. One could argue that the absence of a decrease in plasma and LDL cholesterol after the CLA-B diet could in part be due to that diet’s greater content of 18:1 trans-11 (trans-vaccenic acid; Figure 2) compared with the content in the CON-B diet. It has indeed been suggested that the consumption of trans fatty acids increases LDL-cholesterol and lowers HDL-cholesterol concentrations (61). In that regard, Willett et al (62) found a positive significant association between the intake of trans fatty acids formed by the partial hydrogenation of vegetable oils and the incidence of coronary heart disease. However, there was no significant association between the intake of trans isomers from animal sources and the incidence of coronary heart disease. This lack of association was thought to be due to the distinct structure of the predominant trans isomer in ruminant fat, trans-vaccenic acid (62). In addition, 18:1 trans-11 appears to be partly converted to cis-9, trans-11 18:2 (CLA) via the action of the -9 desaturase present in the intestinal epithelium, liver, and adipose tissue of humans (63, 64). Hence, in the current study, it is unlikely that the greater amount of trans fatty acids in the CLA-B was responsible for the lack of improvement in the lipid profile, because these trans fats more likely provided an additional source of CLA when subjects consumed the CLA-B diet.
The CLA-induced reduction in adiposity and weight gain observed in animals (9–13) could also be a mechanism for mediating the improvements in the lipid profile often reported with CLA supplementation (5, 7). However, consistent with the lack of an improvement in the lipid profile with the CLA-B diet in the current study, the reductions in body weight and waist circumference, as well as in the visceral and subcutaneous adipose tissue areas after the diet intervention, were comparable with the CLA-B and CON-B diets. Another study, conducted in healthy women ingesting 3.9 g CLA/d for 63 d and in which only 23% of the CLA was in the form of the trans-10, cis-12 isomer, also showed that CLA had no significant effect on body composition. This lack of effect of CLA was interpreted to be due to the type of CLA isomer used; the most effective in reducing body fat in preadipocytes and animal models reportedly is the trans-10, cis-12 isomer (8, 12, 49–51). In support of this hypothesis, supplementation of 3 g of an isomeric blend containing primarily the cis-9, trans-11 CLA isomer had no effect on the body weight or BMI of nonobese (BMI <25; 23) or overweight and obese subjects (BMI 27–35; 53).
We also failed to observe significant changes in the concentration of CRP, a systemic marker for inflammation and, thereby, for processes leading to atherosclerosis, after either the CON-B or CLA-B diet. In contrast, a study reported that, compared with ingestion of a placebo containing olive oil, trans-10, cis-12 CLA supplementation led to significantly greater plasma CRP concentrations in overweight men (65). This difference from the current study may be related in part to the difference in CLA isomer used, because our butter fat was mostly rich in cis-9, trans-11 CLA.
Whereas our results seem to be in agreement with most of the studies aimed at unraveling the effect of CLA in humans, it must be stressed that some aspects of our study design—such as the number of subjects used, their health profile, and the length of the experimental periods—may not have been optimal for observing desirable effects of CLA on health. First, 16 subjects took part in the current crossover study; this design is known to have more power because each subject is his or her own control. In a power analysis with an of 0.05 and a ß of 0.8, this sample size was found to be appropriate to detect a 10% significant difference in our primary outcomes represented by the cholesterol variables studied, which have relatively small variability. However, the possibility that the sample size was insufficient to detect significant differences in other variables, such as triacylglycerols, that vary considerably more over time should not be ruled out. Another possible limitation of the current study is that, despite being overweight or obese, our subjects were healthy and had normal blood lipid profiles. Hence, it has been suggested that hypercholesterolemic persons tend to respond to nutritional interventions more than do normocholesterolemic subjects (66). As for the length of the experimental periods, it has been shown that blood lipid variables normally stabilize within 4 wk of being subjected to strictly controlled dietary intervention conditions (67). However, the beneficial effects of CLA on blood lipids in animal studies were usually observed after nearly 2 mo, and thus it is possible that a longer experimental period may have been necessary to replicate comparable results in humans.
In conclusion, our results indicate that, in overweight or obese men with only a slight deterioration of their lipid profile, a 10-fold CLA enrichment of butter fat via the addition of sunflower oil to the feed of dairy cows does not induce significantly greater metabolic effects than are observed with the consumption of a control butter diet low in CLA. It remains unclear whether this lack of improvement in the lipid profile and body composition can be attributed to the nature of the CLA isomer found in butter, to the amount of CLA ingested, or to the length of administration of the experimental diets. Further human studies are needed to evaluate the individual effects of different CLA isomers on lipoprotein metabolism and in persons with a more detrimental lipid profile at study onset.
ACKNOWLEDGMENTS
We are indebted to John KG Kramer for his help in the determination of the conjugated linoleic acid isomers and of the trans octadecenoic acid isomers within the experimental butters. We express our gratitude to the metabolic kitchen staff and participants for their dedication and cooperation.
NB and PYC were the principal investigators for the study; IG was a co-investigator; PC was responsible for the screening and medical supervision of the study participants; LC is a research assistant to NB; JD is a research assistant to PYC; BL was responsible for the measurement of the nontraditional risk factors; and SD coordinated the study, performed statistical analyses, analyzed the data, and wrote the manuscript. None of the authors had a personal or financial conflict of interest.
REFERENCES