The Journal of Nutrition Vol. 128 No. 5 May 1998, pp. 875-880

Poor Digestibility of Fully Hydrogenated Soybean Oil in Rats: A Potential Benefit of Hydrogenated Fats and Oils^1,2,3

Randall J. Kaplan^{*, 4} and Carol E. Greenwood^{*, , 5}

^* Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, Toronto, Ontario M5S 3E2, Canada and  Kunin-Lunenfeld Clinical Research Unit, Baycrest Centre for Geriatric Care, Toronto, Ontario M6A 2E1, Canada

	ABSTRACT

Abstract Introduction Methods Results Discussion References

The digestibility and absorption of dietary triacylglycerols are dependent on a number of factors including their fatty acid profile. Data demonstrating poor bioavailability of dietary stearic acid would suggest that hydrogenated oil sources would have lower digestibility coefficients compared with their native oils. To test this hypothesis, postweanling rats were fed one of four diets, formulated to contain 40% of energy as fat (assuming complete bioavailability), for 14 d. The diets only differed by fat type, containing soybean oil (SBO), fully hydrogenated soybean oil (HSB), medium-chain triglyceride oil (MCT), or hydrogenated coconut oil (HCO). Rats fed HSB consumed more food during the last 6 d (155.2 ± 2.7 g) than those in each of the other groups (MCT: 118.9 ± 2.2 g; HCO: 124.7 ± 3.2 g; SBO: 123.8 ± 2.3 g), yet, they did not gain more weight. Two-day fecal excretion was almost three times greater in HSB-fed rats than in rats fed any other diet (P < 0.0001) because HSB was very poorly available. The digestibility coefficients (a measure of bioavailability) of the four fats were: HSB (30.9 ± 1.3%) < HCO (94.5 ± 0.4%) < SBO (97.0 ± 0.4%) < MCT (98.7 ± 0.2%) (P < 0.0007). All rats compensated for the incomplete availability of the fats, as apparent absorbable energy consumed did not differ among diet groups. The present data suggest that HSB only contributes 11.6 kJ/g (most fats contribute ~37.7 kJ/g) and that not only manufactured fat substitutes, such as olestra, but also more conventional fats are incompletely available to the body. Foods that currently contain HSB may contribute much less utilizable fat and energy than presently realized.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

The current nutrition recommendation for fat consumption in both Canada (Health and Welfare Canada 1990) and the United States (National Research Council 1989) is that the diet should contain no more than 30% of energy as fat and no more than 10% as saturated fat. In an attempt to help North American consumers meet these recommendations and decrease their risk of developing various chronic diseases, without compromising palatability, several manufacturers have developed fat-based fat substitutes that mimic the flavor and texture of conventional fats but that are inefficiently available to the body and therefore provide reduced energy. [The availability or bioavailability of a fat or fat substitute refers to the portion that has been digested, absorbed and metabolized such that it can be utilized by the body (U.S. Department of Health and Human Services 1996)].

In addition to these fat substitutes, certain conventional fats (fats that do not require extensive processing), such as cocoa butter, also have been shown to be incompletely available in rats (Apgar et al. 1987, Chen et al. 1989, Monsma et al. 1996) and to a lesser extent in humans (Mitchell et al. 1989). A number of studies suggest that the poor availability of conventional fats such as cocoa butter is related to their content of stearic acid (18:0) and specifically tristearin (Bergstedt et al. 1990, Kamei et al. 1995, Mattson 1959). Generally, the higher the content of stearic acid and tristearin in a fat, the poorer the digestion and absorption (for review see Kritchevsky 1994). Other evidence also suggests that the presence of long-chain saturated fatty acids at the 1 and 3 positions of the triacylglycerol are partly responsible for the poor absorption of cocoa butter (Bracco 1994).

Another potential source of tristearin is in highly hydrogenated fat sources. The hydrogenation of natural fats and oils that contain a high content of 18-carbon unsaturated fatty acids will result in an increased content of stearic acid and tristearin, presumably making these fats less available. Fully hydrogenated soybean oil (HSB),⁶ a fat source that currently is consumed in North America as a component of liquid shortening, would fall into this category. This oil is soybean oil (SBO) that essentially has been hydrogenated to completion (>99% saturated).

The purpose of the present investigation was to determine the bioavailability of HSB compared with three other fat sources and to determine the influence of these fats on body weight and food intake.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Rats and diets.  Postweanling male Wistar rats (Charles River, St. Constant, Quebec), initially weighing 75-100 g, were housed singly in stainless-steel wire-mesh cages (necessary to measure food spillage and fecal excretion) in a temperature-controlled environment (22 ± 1°C). They were maintained on a 12-h light:dark cycle and were given free access to all diets and water. Diets were fed from food cups that were secured with a spring to minimize spillage. The protocol was approved by the University of Toronto Animal Care Committee.

All diets were nutritionally adequate purified granular mixtures formulated to be isoenergetic (17.9 kJ/g diet) containing 40% of energy as fat (assuming complete bioavailability of all fat sources), with no deficiencies in linoleic or -linolenic acid (Bourre et al. 1990, 1993) and similar linoleic:-linolenic acid ratios. This was accomplished by adding a safflower-flaxseed oil (Baldwin's Natural Foods, Toronto, Ontario) mixture to each diet, except for the SBO diet. The four diets contained either medium-chain triglyceride oil (MCT; provided by Mead Johnson Canada, Toronto, Ontario), hydrogenated coconut oil (HCO; Teklad, Madison, WI), fully hydrogenated soybean oil (HSB; provided by Thomas J. Lipton Inc., Toronto, Ontario) or soybean oil (SBO; Noah's Natural Foods, Toronto, Ontario). Diet composition is shown in Table 1.

Fatty acid composition of each diet (Table 2) and of each specific fat source are shown in detail (Table 3). Fatty acid composition of HSB, SBO, flaxseed oil and safflower oil were determined by gas chromatography as described later. Fatty acid compositions of MCT and HCO are based on the values provided by the manufacturers.

Because HSB consists of almost 100% long-chain saturated fatty acids, it is extremely hard in texture and difficult to incorporate into a homogenous diet. In a pilot study, HSB was melted before incorporation into the diet mixture, however, the fat formed into small pellets such that the rats could avoid ingestion of the fat if preferred. To avoid this problem, the HSB fat source was ground into a fine powder before incorporation into the diet such that the mixture was completely homogeneous and the rats could not avoid eating the fat.

Experimental design.  Rats were assigned to one of four dietary treatment groups (n = 10 or 11 per group) such that initial body weights did not differ among groups. All rats consumed an experimental diet, containing either MCT, HCO, HSB or SBO, for 14 d. The total concentrations of saturated fatty acids were similar among the MCT, HCO and HSB diets (Table 2).

All rats were weighed and food intake was recorded every 2nd d at the same time during the light cycle throughout the experiment. Two-day fecal excretion was collected from all rats at the end of the experimental period to determine the digestibility coefficient for each fat source. [The digestibility coefficient is a measure of bioavailability expressed as the percentage of ingested fat that was not excreted in the feces (Apgar et al. 1987, Mitchell et al. 1989, Monsma et al. 1996)]. The final 2 d of feeding and of fecal excretion were used for this calculation. The experiment was performed in two identical phases (5 or 6 rats per group per phase). The second phase started 1 wk after the first phase started.

Fatty acid analysis.  After methylation with boron trifluoride in methanol, samples were dried under a stream of nitrogen and redissolved in saline and hexane (Morrison and Smith 1964). Fatty acid composition of each dietary fat source was analyzed on a gas chromatograph (Hewlett Packard 5890A GC with 7673A autosampler and 3393A integrator) using a DB-23 fused silica capillary column (30 m × 0.25 mm × 0.25 µ) (J & W Scientific, Folsom, CA) and flame ionization detection. Retention times were verified with purified standards (NuChek Prep Inc., Elysian, MN). Results are expressed as percent of total fatty acids.

Fecal analysis.  Total 2-d freeze-dried fecal excretion from each rat was ground with a mortar and pestle into a homogenous mixture. Fecal lipid was extracted using a modification of the method of Folch et al. (1957). A small amount of water (0.5 ml) was added to 0.5 g of freeze-dried feces. Methanol (5 ml) followed by chloroform (10 ml) was added to each tube, which was shaken and heated in a hot-water bath (60°C, 1 h) (Kamei et al. 1995). Heating was necessary to completely extract HSB because this fat does not readily dissolve in the normal chloroform-methanol Folch solution. The sample was centrifuged and reheated to dissolve any particles that may have precipitated during cooling in the centrifuge. The solution was transferred to a second test tube. No filtering was used to avoid any cooling of the solution, which would allow HSB to solidify and remain on the filter. The remaining solid residue was reextracted with methanol and chloroform and heated (60°C, 1 h). Potassium chloride in water (8.8 g/L) was added to the combined filtrates in a proportion of one-quarter of the total volume. The tube was shaken and centrifuged at 265 × g for 5 min to obtain two distinct layers. The bottom layer containing the purified lipid was transferred to a preweighed test tube and dried under nitrogen until only a solid lipid residue remained. Finally, the tube was freeze-dried overnight to ensure complete drying of the sample. The weight of total lipid was determined by difference.

By using samples of diet containing known concentrations of fat and comparing the results to another extraction method (Solvent extraction with Tecator Soxtec System HT; Hoganas, Sweden), it was determined that ~95% of lipids were recovered using the modified Folch extraction method. However, another conventional technique for fecal fat extraction (Jeejeebhoy et al. 1970) incompletely extracted HSB and was therefore unsuitable for use in this experiment.

Statistical analysis.  Statistical analyses were conducted with SAS 6.08 (SAS Institute, Inc., Cary, NC) for the microcomputer. For all analyses, the acceptable level of significance (Type I error) was P  0.05. Bartlett's test for homogeneity of variance (Rosner 1990) was used to compare group variances. When all variances were equal, comparisons of group means were made by analysis of variance followed by Student-Newman-Keuls' test for multiple comparisons (Steel and Torrie 1960). When at least two group variances were unequal, Student's t tests for samples with equal or unequal variances were used to compare each pair of means separately. Where body weights differed among groups, covariate analyses were performed to remove these differences as confounding variables.

	RESULTS

Abstract Introduction Methods Results Discussion References

A summary of the body weight, food intake, fecal excretion and bioavailability data for each diet and for each fat source are shown in Table 4. Because an adjustment period to the diets may have been necessary, food intake and body weights were analyzed during both the entire 14-d feeding period and during the last 6 d of feeding.

Initial body weights did not differ among the four dietary groups. Rats fed the HSB diet consumed more total food during the 14-d experimental period than rats fed the other three diets (P < 0.0001). However, rats in the HCO and SBO groups weighed more than MCT- and HSB-fed rats (P < 0.0002) by the end of the feeding period. The initial decrease in growth rate in rats fed MCT was anticipated from previous results (Harkins and Sarett 1968).

A comparison of total food intake and total body weight gain during the final 6 d of feeding shows that rats in the HSB group consumed more total food than rats in each of the other three dietary groups (P < 0.0001), yet these rats did not gain more weight. Rats fed MCT gained less weight than rats fed HCO or SBO (P < 0.03) but did not differ from rats fed HSB. Similar differences among groups were observed during the final 2 d of feeding (HSB consumed more than rats in the other groups; P < 0.0001), suggesting that feeding behavior was stable during the final 6 d.

Two-day fecal excretion (dry weight) for the HSB-fed rats was almost three times greater than the excretion for the HCO- and SBO-fed rats; this was in turn greater than the excretion for the MCT-fed rats (P < 0.0001). The lower fecal weight excreted by MCT-fed rats may be explained partly by the lower fiber concentration (75.0 g/kg) in the MCT diet (75.0 g/kg) compared with HCO, HSB and SBO diets (88.5 g/kg).

The feces excreted by rats fed HSB contained a higher percentage of fat than the feces excreted by HCO-fed rats, which in turn, contained a higher percentage of fat than the feces excreted by MCT- and SBO-fed rats (P < 0.0001). Percent fecal moisture (difference between wet and dry fecal weights) did not differ among groups (20.5-25.3%; data not shown).

Almost 60% of the dietary fat consumed by HSB-fed rats was excreted in the feces, and therefore not available to these rats (41.5% digestibility coefficient). The digestibility coefficients for the other three diets were much higher than the coefficient for the HSB diet. The MCT diet had the highest coefficient, followed by the SBO diet and the HCO diet. The coefficient was significantly different for all four diets (P < 0.002).

Because the MCT, HCO and HSB diets were supplemented with flaxseed and safflower oil to eliminate the possibility of a linoleic or -linolenic acid deficiency, the digestibility coefficients for the respective diets do not show the bioavailability of each actual fat source. Therefore, the specific digestibility coefficient for each fat component, assuming that the flaxseed-safflower oil mixture was absorbed to the same extent as SBO also is shown in Table 4. This assumption is reasonable because flaxseed oil, safflower oil and SBO all contain similar concentrations of saturated (~9, 9, and 14%, respectively) and unsaturated fatty acids (U.S. Department of Agriculture 1979). These data show that ~70% of the HSB component of the HSB diet was not available (30.9% digestibility coefficient) to the rats that were fed this fat source. The MCT, SBO and HCO fat components were all much more available than the HSB component. The digestibility coefficient was significantly different for all four fat components; MCT had the highest coefficient, followed by SBO and HCO (P < 0.0007).

Taken together, the fecal fat analysis demonstrated that most of the dietary HSB was not digested or absorbed and therefore passed into the feces. Furthermore, HCO was significantly less available than SBO, which was in turn, less available than MCT. Because the nonabsorbed fats did not contribute any energy to the various diets, the present data suggest that although all of the diets were designed to be isoenergetic (17.9 kJ/g diet), the rats in different dietary groups did not actually absorb diets of the same energy density.

The apparent total energy absorption for each diet group during the final 6 d of the experimental period was calculated assuming that all diets were isoenergetic (17.9 kJ/g diet) and with adjusted values based on fecal fat excretion (Table 5). (Because fecal content of protein and carbohydrate were not determined, the actual absorption of each diet is not known, however, only a minimal amount of nonabsorption of these macronutrients would be expected and no differences among groups would be expected). To recalculate the apparent absorbed energy density of each diet, all available fats (ingested and not excreted in feces) were given the Atwater value of 37.7 kJ/g except MCT, which was given a value of 34.7 kJ/g (Harkins and Sarett 1968). The adjusted absorbed energy densities of each diet were determined to be (kJ/g diet): MCT, 17.8; HCO, 17.5; HSB, 13.7; and SBO, 17.7. 

The greater intake of total food by HSB rats (and apparent total energy) during the experimental period appears to be completely explained by fecal fat content (Table 5). That is, when the nonabsorbed dietary fat components are considered, it appears that rats in all groups consumed the same amount of total energy. Thus, complete compensation for the nonabsorbed fat appears to have occurred.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

The results reported here suggest that HSB (a fat that is currently consumed in North America) is poorly digested and absorbed by rats and therefore contributes very little energy. These findings are consistent with the notion that various conventional fat sources are available to the body to different extents, and do not actually provide similar amounts of fat and energy.

Rats fed the HSB diet consumed a significantly greater amount of total food than rats fed diets containing MCT, HCO or SBO, yet they did not gain more weight. These rats excreted ~70% of the fully hydrogenated soybean oil component of the HSB diet that they ingested (30.9% digestibility coefficient). Hence, HSB acted more like a low-energy substance or fat substitute than a normal dietary fat source. Assuming that the absorbed component of HSB contributed the Atwater value of 37.7 kJ/g, the HSB fat source only contributed about 11.6 kJ/g. As a result of the poor availability of HSB, rats in this dietary group consumed more food to obtain the same amount of energy as the other rats and grow normally. That is, during the final 6 d of feeding, the apparent energy absorbed and the body weight gain of rats fed HSB was similar to that of rats fed HCO and SBO. Thus it appears that during the final 6 d of feeding these rats could completely compensate for the lower absorbable-energy-density of their diets.

It is important to note that the specific digestibility coefficients that have been determined are estimates calculated from the feces excreted and food consumed during the final 2 d of feeding. We cannot be certain that the feces collected exactly represented the beginning and end points of food consumption because a fecal marker was not used. Furthermore, the lower fiber concentration of the MCT diet and the higher volume of food consumed by the HSB-fed rats may have contributed somewhat to the observed differences in digestibility coefficients among the fat sources. Nevertheless, the huge difference in the digestibility coefficient between HSB and the other fat sources is unmistakably due to a difference in the actual digestibility of the fats.

Although >94% of each of the HCO (94.5%), SBO (97.0%) and MCT (98.7%) fat components were available to the rats, the digestibility coefficients significantly differed from one another. Nevertheless, the apparent energy absorbed during the final 6 d of feeding did not differ among these groups, suggesting that compensation for the different energy densities of the diets had occurred. MCT-fed rats may not have fully compensated because these rats gained less weight than the SBO- and HCO-fed rats. Previous literature suggests that MCT-fed rats initially may gain less weight than rats fed other fat diets, but after 8 wks, they gain a similar amount of weight (Harkins and Sarett 1968). Thus a longer adjustment period to MCT may be required before rats fed this diet can consume the proper amount of food to grow at a normal rate.

The incomplete availability of HSB in this study is supported by the results of one recent study that suggested that a highly hydrogenated soybean oil fed at both 25 and 50% of 200 g fat/kg diet (as opposed to 84% of 190 g fat/kg diet in the present experiment) was poorly absorbed (Kamei et al. 1995). Thus it appears that this fat is poorly absorbed at both high and low fat levels and when it is blended with other oils.

The explanation for the incomplete availability of HSB may relate to its high content of long-chain saturated fatty acids. Evidence suggests that stearic acid and tristearin are poorly digested and absorbed in rats (Bergstedt et al. 1990, 1991, Kamei et al. 1995, Mattson 1959). Because 84% of HSB is composed of stearic acid, a large component of this fat must be made up of tristearin [the highly hydrogenated soybean oil used by Kamei et al. (1995) contained 57% tristearin]. Furthermore, in the present study, the higher the content of stearic acid in a fat source, the poorer was the availability of the fat. Although MCT, HCO and HSB all contain >98% saturated fatty acids, MCT had the highest availability, followed by HCO and HSB, which mirrors the relative contents of stearic acid in these fats (Table 3). The availability of SBO, which was in between that of MCT and HCO, also reflects the intermediate content of stearic acid in this oil. In fact, the coefficient of determination between the stearic acid content of the four diets and the digestibility coefficient of the diets was very strong (r² = 0.99, P < 0.003, determined from group means).

Despite this strong correlation, the content of tristearin in HSB may be more important in explaining the extremely low digestibility of this fat. In an early study, Mattson (1959) found that the digestibility coefficient of a fat was inversely proportional to its content of tristearin and not to stearic acid. In fact, the digestibility coefficient for hydrogenated linseed oil (91% tristearin) when fed to rats as the only fat source was found to be 15.4%.

Although the availability of a high tristearin-containing fat source has not been examined in humans, the present findings, along with other research that indicates that cocoa butter is less available to the body than other conventional fats (Apgar et al. 1987, Chen et al. 1989, Mitchell et al. 1989, Monsma et al. 1996), raises the issue of whether nutrition labels and health claims on food products containing such fats should reflect this information. Currently, the U.S. Food and Drug Administration (FDA) requires that nutrition labels and nutrient content claims on all products containing fat must reflect total fat content based on "all fatty acids obtainable from a total lipid extraction" (U.S. Department of Health and Human Services 1996). The one exception to this rule is that the manufactured fat-based fat substitute, olestra [a nonabsorbed sucrose polyester composed of six to eight fatty acids bound to sucrose by ester bonds (Bergholz 1992)], does not have to be reported as a source of fat or energy. In addition to olestra, however, other fat-based fat substitutes (e.g., salatrim and caprenin) have been developed that are also incompletely available. As a result, the U.S. FDA is currently considering a proposal that would allow manufacturers of such fat substitutes to report the fat and energy content of foods containing these products based on their availability rather than on the actual fat content of the foods (U.S. Department of Health and Human Services 1996). The rationale behind implementing such a change in the regulations is based on the assumption that the benefits of foods with less available fat are similar to the benefits of foods with less total fat.

In contrast to the proposed amendment to the food labeling regulations for manufactured fat-based substitutes, the U.S. FDA tentatively has concluded that the use of similar information would not be permitted for conventional fats (U.S. Department of Health and Human Services 1996). The U.S. FDA has suggested that the lack of testing and the potential confusion to both consumers and health professionals would outweigh any benefits that providing fat-availability information about conventional fats might have. Experiments with rats have suggested that the digestibility coefficient for cocoa butter is between 53 and 78% (Apgar et al. 1987, Chen et al. 1989, Monsma et al. 1996), whereas one study has suggested that the coefficient is 89% when consumed by humans (Mitchell et al. 1989). Hence, much more testing would be necessary to determine a proper coefficient, and if a value close to 89% is accurate, then the benefits of using availability rather than actual fat content may indeed be too small to warrant a change in nutrition labeling regulations.

However, the present findings demonstrate that more extreme differences in the availability of conventional fat sources may exist. Clearly, a fat source such as HSB, which may only contribute 11.6 kJ/g, should not be considered equivalent to a fat that contributes 37.7 kJ/g. The health implications of a product containing HSB in place of another conventional fat would be very different with respect to both total energy content and available fat content. Hence, although more testing on humans is needed, the tentative conclusion by the U.S. FDA to not allow labeling of conventional fats to be based on availability may be premature. Avoiding the fact that conventional fat sources greatly differ in their availability to the body actually may be more misleading and harmful than taking the steps to implement a change in policy. Allowing manufacturers of fat products to use availability as a measure of fat content for all fat-containing foods may enable health-conscious consumers to accurately monitor their fat intake and may give such companies more options to develop cheaper and healthier low-fat foods.

Fully hydrogenated soybean oil appears to be a safe ingredient that could be used more commonly in place of other fats. One concern about hydrogenated fats is that some of the remaining double bonds are in a trans configuration rather than their natural cis configuration because recent evidence suggests that trans fatty acids increase serum cholesterol levels (for review see Anonymous 1996) and may be associated with cancer [see Ip and Marshall (1996) for review]. However, this is not a concern with completely hydrogenated fats because they are virtually devoid of trans fatty acids. Another concern with HSB is its high content of saturated fatty acids, which is often linked to high cholesterol levels. However, most of HSB contains stearic acid, which has been shown in numerous studies to be less cholesterolemic and less atherogenic than other saturated fatty acids (for review see Kritchevsky 1994). Nevertheless, it is possible that the absorbed portion of HSB may contain more cholesterolemic fatty acids, such as palmitate. However, it has been reported that a highly hydrogenated soybean oil prevented an expected dietary cholesterol-induced rise in plasma total cholesterol and reduction in HDL-cholesterol (Kamei et al. 1995). Finally, there is some concern that a poorly digestible fat such as HSB could potentially decrease the absorption of fat-soluble vitamins. If this is the case, foods containing a high concentration of HSB could be supplemented with vitamins as are products containing olestra.

In summary, these results suggest that HSB is incompletely available to rats and only provides ~11.6 kJ/g. Foods that contain HSB that currently are being consumed may actually contain less energy and less available fat than presently realized. These findings add support to the view that the bioavailability of conventional fats, rather than actual fat content, should be considered in the determination of the energy and fat content of foods.

	ACKNOWLEDGMENTS

The authors wish to thank Mary Ann Ryan and Stephen C. Cunnane for their advice with the fatty acid and fecal analyses and for the use of their equipment.

	FOOTNOTES

Manuscript received 27 March 1997. Initial reviews completed 9 June 1997. Revision accepted 8 January 1998.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Anonymous Position paper on trans fatty acids. ASCN/AIN Task Force on Trans Fatty Acids. American Society for Clinical Nutrition and American Institute of Nutrition. Am. J. Clin. Nutr. 1996; 63:663-670 ^{[Abstract/Free Full Text]}
• Apgar J. L., Shively C. A., Tarka S.M.J. Digestibility of cocoa butter and corn oil and their influence on fatty acid distribution in rats. J. Nutr. 1987; 117:660-665
• Bergholz C. M. Safety evaluation of olestra, a nonabsorbed, fatlike fat replacement. Crit. Rev. Food Sci. Nutr. 1992; 32:141-146 [Medline]
• Bergstedt S. E., Bergstedt J. L., Fujimoto K., Mansbach C., Kritchevsky D., Tso P. Effects of glycerol tripalmitate and glycerol trioleate on intestinal absorption of glycerol tristearate. Am. J. Physiol. 1991; 261:G239-G247 [Abstract/Free Full Text]
• Bergstedt S. E., Hayashi H., Kritchevsky D., Tso P. A comparison of absorption of glycerol tristearate and glycerol trioleate by rat small intestine. Am. J. Physiol. 1990; 259:G386-G393 [Abstract/Free Full Text]
• Bourre J. M., Dumont O., Pascal G., Durand G. Dietary a-linolenic acid at 1.3 g/kg maintains maximal docosahexaneoic acid concentration in brain, heart and liver of adult rats. J. Nutr. 1993; 123:1313-1319
• Bourre J. M., Piciotti M., Dumont O., Pascal G., Durand G. Dietary linoleic acid and polyunsaturated fatty acids in rat brain and other organs. Minimal requirements of linoleic acid. Lipids 1990; 25:465-472 [Medline]
• Bracco U. Effect of triglyceride structure on fat absorption. Am. J. Clin. Nutr. 1994; 60:1002S-1009S [Abstract/Free Full Text]
• Chen I. S., Subramaniam S., Vahouny G. V., Cassidy M. M., Ikeda I., Kritchevsky D. A comparison of the digestion and absorption of cocoa butter and palm kernel oil and their effects on cholesterol absorption in rats. J. Nutr. 1989; 119:1569-1573
• Folch J., Lees M., Stanley G.H.S. A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 1957; 226:497-509 ^{[Free Full Text]}
• Harkins R. W., Sarett H. P. Nutritional evaluation of medium-chain triglycerides in the rat. J. Am. Oil Chem. Soc. 1968; 45:26-30 [Medline]
• Health and Welfare Canada (1990) Nutrition Recommendations: The Report of the Scientific Review Committee. Minister of Supply and Services Canada, Ottawa, Canada.
• Ip C., Marshall J. R. Trans fatty acids and cancer. Nutr. Rev. 1996; 54:138-145 [Medline]
• Jeejeebhoy K. N., Ahmad S., Kozak G. Determination of fecal fats containing both medium and long chain triglycerides and fatty acids. Clin. Biochem. 1970; 3:157-163 [Medline]
• Kamei M., Ohgaki S., Kanbe T., Niiya I., Mizutani H., Matsui-Yuasa I., Otani S., Morita S. Effects of highly hydrogenated soybean oil and cholesterol on plasma, liver cholesterol, and fecal steroids in rats. Lipids 1995; 30:533-539 [Medline]
• Kritchevsky D. Stearic acid metabolism and atherogenesis: history. Am. J. Clin. Nutr. 1994; 60:997S-1001S [Abstract/Free Full Text]
• Mattson F. H. The absorbability of stearic acid when fed as a simple or mixed triglyceride. J. Nutr. 1959; 69:338-342
• Mitchell D. C., McMahon K. E., Shively C. A., Apgar J. L., Kris-Etherton P. M. Digestibility of cocoa butter and corn oil in human subjects: a preliminary study. Am. J. Clin. Nutr. 1989; 50:983-986 [Abstract/Free Full Text]
• Monsma C. C., Gallaher D. D., Ney D. M. Reduced digestibility of beef tallow and cocoa butter affects bile acid excretion and reduces hepatic esterified cholesterol in rats. J. Nutr. 1996; 126:2028-2035
• Morrison W. R., Smith L. M. Preparation of fatty acid methyl esters and dimethylacetals from lipids with boron trifluoride-methanol. J. Nutr. 1964; 5:600-608
• National Research Council (1989) Diet and Health: Implications for Reducing Chronic Disease Risk. National Academy Press, Washington, DC.
• Rosner, B. (1990) Fundamentals of Biostatistics. PWS-Kent Publishing Company, Boston, MA.
• Steel, R.G.D. & Torrie, J. H. (1960) Principles and Procedures of Statistics. McGraw-Hill, New York, NY.
• United States Department of Agriculture (1979) Composition of Foods: Fats and Oils, Handbook No. 8-4.
• United States Department of Health, Human Services: Food, Drug Administration Food labeling: nutrient content claims pertaining to the available fat content of foods. Fed. Reg. 1996; 61:67243-67260

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences