Olestra, a mixture of hexa-, hepta- and octaesters of sucrose formed from long-chain fatty acids isolated from edible oils, has cooking characteristics similar to those of traditional dietary fats and oils (Bernhardt 1988). Olestra and triglycerides have similar thermal stability profiles (Kester 1993). Olestra does not undergo lipolysis in the gut and is not absorbed intact (Mattson and Nolen 1972, Mattson and Volpenhein 1972, Miller et al. 1995); therefore it contributes no energy to the diet. Because of its unique properties, olestra can serve as a zero-calorie replacement for conventional fats and oils.

The evaluation of Olestra (Olean, Proctor & Gamble, Cincinnatti, OH) for use as a replacement for dietary fat in foods includes determination of its effect on the availability of dietary nutrients. Olestra has been shown to reduce tissue concentrations of the lipophilic nutrients cholesterol and vitamins A, D and E, but not vitamin K (Jandacek et al. 1990, Jones et al. 1991a and 1991b, Mattson et al. 1976, Mellies et al. 1983 and1985). It has been shown in animals that the effect of olestra on the availability of the fat-soluble vitamins can be offset by introducing additional amounts of the vitamins into the diet (Lafranconi et al. 1994).

Jandacek (1982) proposed that the absorption of highly lipophilic substances would be affected by olestra because such substances partition into the olestra in the gastrointestinal (GI)⁴ tract and thus are not available to the intestinal micelles. This partitioning would not be expected to occur with water-soluble nutrients; therefore olestra would not be expected to affect the absorption of these nutrients.

The domestic pig was chosen as the model in which to investigate the potential nutritional effects of olestra. The pig has been used extensively in nutrition research (Miller and Ullrey 1987, Tumbleson 1986), including a recent evaluation of the nutritional effects of a poorly absorbed fat substitute (Hayes et al. 1994a and 1994b). Pigs and humans have similar GI anatomy, morphology and physiology. The GI tract of a 30- to 40-kg pig is similar in total length to that of an adult human (Leigh-Browne and Harpur 1975) and the relative sizes of sections of human and pig GI tracts are similar (Kurihara-Bergstrom et al. 1986).

Pigs and humans have similar gastric cell types and similar gastric villi (Kurihara-Bergstrom et al. 1986). The composition and control of gastric secretions are similar in the two species (Kidder and Manners 1978) as are ingesta transit times (Miller and Ullrey 1987).

Table 1. Design of a 4-wk study to evaluate the pig as a model for assessing the nutritional effects of olestra (OA) [View Table]

Digestive and absorptive abilities of the pig are similar to those of humans (Bjorkman et al. 1984, Forsum et al. 1981, Graham and Aman 1987, Juste et al. 1983, Kidder and Manners 1978). Both species depend on dietary quality because gut microflora play a relatively minor role in modifying nutrients in the diet (Miller and Ullrey 1987). The requirements of the weanling pig for nutrients are similar to or greater than human requirements (Miller and Ullrey 1987, Pond and Houpt 1978); thus the pig is a good model for subgroups of the population with high nutrient demands such as children and pregnant or lactating women (Middleton et al. 1997). Pigs reach sexual maturity and display reduced growth by 6-8 mo of age (Anderson 1974); thus it is possible to conduct studies over reasonable periods of time covering the major growth and developmental phases of the pig's life.

Pigs have been used extensively in pediatric nutritional research (Cooper 1975, Glauser 1966, Mellor and Cockburn 1986, Moughan et al. 1992, Schneider and Sarett 1966). A contributing reason is that the weanling pig and the young human child have similar GI tracts and nutritional requirements. The pig's GI tract matures by about 4 wk of age (Leary and Lecce 1976), whereas digestive and absorptive processes in children mature by 9-12 mo of age (Grand et al. 1976, Hamosh 1987). The micronutrient requirements of the weanling pig are similar to or greater than those of children 1-3 y of age (Middleton et al. 1997). Finally, the sizes of the various digestive organs of the two species at birth are markedly similar (Moughan et al. 1992) and the weanling pig's body size, 10-12 kg, is similar to that of a child.

Some key aspects of human nutrition have not been modeled in the pig. These include the micronutrient content of the diet, the level of fat in the diet, the chemical forms and concentrations of nutrients, the response of vitamin D status to UV light exposure and meal-eating pattern. These conditions were investigated in this study. The general purpose of the study was to evaluate the appropriateness of the pig as a model for determining the effects of olestra on nutrient availability. Specific objectives were as follows: 1 ) to evaluate the pig's tolerance for dietary fat levels similar to those in the human diet; 2 ) to evaluate the adequacy of a purified diet to produce good growth; 3 ) to determine whether tissue concentrations of vitamins A, D, E and K respond to changing dietary concentrations of the vitamins and to olestra; and 4 ) to determine how much exposure to UV light is required to produce a vitamin D₃ contribution to total vitamin D status representative of the normal human situation.

MATERIALS AND METHODS

Table 2. Composition and digestible energy content of diets fed to pigs for 4 wk [View Table]

Upon receipt, the pigs were housed in groups of six or seven in 0.9 × 2.5 M pens for 2 wk. During the first week of this acclimation period, the standard starter diet was provided. During the second week, the pigs were fed a purified basal diet containing the NRC requirements for micronutrients for 5- to 10-kg pigs (NRC 1988) and 14% (wt/wt) fat, which provided 30% of the total digestible energy.

During the acclimation period, the pigs were given a detailed physical examination and a limited clinical pathology screening, and were observed daily for abnormalities indicating ill health. Only animals with normal clinical pathology were selected for the treatment phase of the study.

At the start of treatment, the pigs were placed into individual pens in a sunlight-free facility with a 12-h light:dark cycle maintained by incandescent lights. The temperature of the facility was kept above 18°C; humidity was not regulated. The pens were cleaned once or twice daily to reduce the potential for coprophagy.

To evaluate the adequacy of a purified diet to support growth, three groups of pigs were fed purified diet containing 1, 1.3 or 1.6 times the NRC requirement levels of micronutrients (1 NRC, 1.3 NRC and 1.6 NRC groups) and 30% of energy from fat. The dietary concentrations of micronutrients and protein were based on the NRC requirements for 5- to 10-kg swine (NRC 1988), the body weight range of the pigs at the start of the study. The 1.3 NRC diet provides nutrients at levels similar to those in a human diet containing the Recommended Dietary Allowances (RDA) of nutrients because the RDA for most nutrients is set above the average requirement by an amount sufficient to meet the needs of almost all members of the population, often 2 SD above the average (NRC 1989). For most nutrients, the SD is ~15% of the mean (Guthrie 1989). The 1 NRC and 1.6 NRC diets provide nutrients at levels that bracket the levels in the 1.3 NRC diet.

The composition of the 1 NRC diet is shown in Table 2. The compositions of the 1.3 and 1.6 NRC diets were the same as that of the 1 NRC diet except that the amounts of the vitamin-mineral premix were increased from 8.0 to 10.4 or 12.8 g/100 g diet, respectively. In these diets, the amounts of Alphacel were decreased proportionally. The total amounts of Alphacel did not differ significantly among the three diets, however, because it was the major component of the vitamin-mineral mix. Table 3 shows the compositions of the vitamin-mineral premixes used in preparing the diets.

Table 3. Amounts of vitamin/mineral premix ingredients per kilogram of diet fed to pigs for 4 wk1 [View Table]

The amount of each micronutrient added to the purified basal diet was calculated from the NRC recommended dietary concentration by adjusting for the metabolizable energy content of the basal diet (4172 kcal/kg, or 17.5 MJ/kg) relative to a standard pig feed providing 3240 kcal (13.6 MJ) of metabolizable energy per kilogram, on which the NRC requirement is based. The digestible energy content of the diets was calculated by summing the digestible energy contributed per kilogram of diet by each dietary ingredient. The targeted amounts of nutrients in the purified diets are shown in Table 4.

Table 4. Target concentrations of olestra (OA) and nutrients in purified diets fed to pigs for 4 wk [View Table]

The concentrations of the nutrient were kept at the requirements for 5- to 10-kg pigs throughout the study. This was done to avoid a change in nutrient intake during the study that might prevent tissue concentrations of the fat-soluble vitamin from reaching steady-state values. Further, the slight excess of nutrients fed to the pigs at the end of the study, relative to the requirements for 5- to 10-kg pigs, would not prevent detection of an effect of olestra on nutrient status because nutrient stores were being measured and stores are proportional to intake.

To determine whether tissue concentrations of the fat-soluble vitamins respond to olestra and to changes in vitamin intake, two groups of pigs were fed either the 1 NRC or the 1.6 NRC purified diet containing 4.8% olestra [1 NRC(OA) and 1.6 NRC(OA) groups]. Olestra was added to, not substituted for, triglyceride in these diets; therefore the total digestible fat content of the diets was the same in all groups.

To determine the effect of UV light on vitamin D status, two groups of pigs were fed purified diet containing 1.6 times the NRC requirements of micronutrients and were exposed to either 15 or 45 min/d UV light [1.6 NRC(15 UV) and 1.6 NRC(45 UV) groups]. The UV lamps (FS-40 T12, National Biological, Twinsburg, OH) were suspended lengthwise over the pen, 1.2 m above the floor and controlled by timers.

Vitamin A was provided in the diets as a 3:1 mixture of retinyl palmitate and -carotene, in terms of retinol equivalents; this ratio is typical of sources of vitamin A in the U.S. diet (Olson 1987). The amount of -carotene was based on the equivalency of 1 µg retinol and 6 µg -carotene, a conversion factor determined for the pig (Olson 1983). Vitamin D was provided as ergocalciferol; vitamin E was provided as dl--tocopheryl acetate. Vitamin K was provided as phylloquinone because this is the form in the human diet. The amount of phylloquinone was based on the NRC requirement for menadione for swine and on the equivalency of menadione and phylloquinone on a weight basis (Griminger and Donis 1960, Nelson and Norris 1960).

The concentrations of nutrients and olestra and the homogeneity of the diets were confirmed by analysis. Stability of retinol, -tocopherol, -carotene, ergocalciferol, phylloquinone and olestra in the diets was confirmed by analysis of samples stored under the same conditions used to store the study diets.

Dietary concentrations of retinol and -tocopherol were measured by HPLC (Cort et al. 1983, Thompson and Duval 1989). After saponification and extraction with anhydrous ethyl ether, the vitamins were separated by using a silica HPLC column (Zorbax, 5-µm, Dupont, Wilmington, DE) and UV detection. Retinol was detected at 325 nm; -tocopherol was detected by fluorescence excitation at 292 nm and emission at 325 nm. Quantitation was by measurement of peak areas and linear regression from standard curves.

The dietary concentration of -carotene was measured colormetrically at 440 nm by the AOAC method (Deutsch 1990). Diet samples were saponified and extracted as described above. Before analysis, the extract was evaporated to dryness, reconstituted in hexane and purified by passing through an open column containing aluminum oxide.

The dietary concentration of ergocalciferol was determined from the concentration of ergocalciferol measured in the vitamin-mineral premix and the amount of premix added to the diet. Ergocalciferol in the premix was measured by HPLC (Thompson et al. 1982). Samples were saponified with ethanolic pyrogallol and KOH, and extracted with hexane. After the extract was purified by passing it through a column of aluminum oxide, ergocalciferol was isolated by using a 5-µm silica column (Zorbax, Dupont) and UV detection (Model 783A, Kratos Analytical, Ramsey, NJ). Quantitation was performed on a HPLC system equipped with a reverse-phase C-18 column (Dupont) and UV detector, using peak height measurement and linear regression from a standard curve.

The dietary concentration of phylloquinone was measured by HPLC, following the method of Haroon et al. (1986). Dietary lipids were extracted by the procedure of Bligh and Dyer (1959). After polar components were removed with a silica Sep-Pak column (Waters, Milford, MA), the lipids were reductively extracted with an acidic solution of hexane, acetonitrile and zinc chloride. Phylloquinone was converted to the hydroquinone with zinc metal and then extracted with acetonitrile. Chromatography was performed with dichloromethane:methanol as the mobile phase, a silica column (5-µm ODS Zorbax, DuPont), and a fluorescence detector (Kratos). Dihydro-vitamin K₁ was used as an internal standard. Phylloquinone concentration was determined by peak height measurement and linear regression from a standard curve.

The concentration of olestra in the diet was measured by reversed-phase liquid chromatography (Tallmadge and Lin 1993). Dietary lipids were extracted with 50:50 hexane/ethyl ether, dissolved in methylene chloride and injected onto an octadecylsilane column (Zorbax ODS, MAC-MOD Analytical, Chadds Ford, PA) equipped with an evaporative light-scattering detector (Applied Chromatography Systems, MacClesfield, England). Olestra and triglycerides were separated by gradient elution using methylene chloride/acetonitrile as the mobile phase. Concentration was determined by peak area measurements and linear regression from a standard curve.

The diets were prepared by ICN Biomedicals (Cleveland, OH) and were shipped and stored at 20°C in plastic-lined, lightproof containers under nitrogen until offered to the pigs.

The daily allotment of feed offered to the pigs was calculated from the pig's body weight at the beginning of the week and the predicted weight gain over the week, using growth curves for the same strain of cross-bred pigs fed standard corn-soy-based swine feed (Martin and Crenshaw 1989). To facilitate complete consumption of the daily feed allotment, the pigs were fed 95% of the NRC requirement for digestible energy based on body weight (NRC 1986).

Prothrombin time (PT) was measured weekly. Prothrombin time was chosen as a measure of vitamin K status because it is the standard measurement used in the pig. More sensitive measures are available for assessing vitamin K status in humans; those measures were used to develop data about the effects of olestra on vitamin K status in human studies (Schlagheck et al. 1997a and 1997b). Serum concentrations of 25-hydroxyergocalciferol [25(OH)D₂] and 25-hydroxycholecalciferol [25(OH)D₃] were used as measures of vitamin D status. Measurements were made at base line, wk 2 and wk 4. Liver concentrations of vitamin A (total retinol and retinyl esters) and vitamin E (-tocopherol) were used as measures of the status of these vitamins. Measurements were made for the group of pigs killed at base line and for all groups killed at the end of the study.

The concentrations of vitamin A (total retinol and retinyl esters) and vitamin E (-tocopherol) in liver tissues were measured by HPLC using a method that allowed both vitamins to be determined simultaneously (Kayden et al. 1983). The samples were saponified with ethanolic KOH, and the vitamins extracted with hexane. Separation and quantitation were performed with a silica HPLC column (Zorbax, 5-µm, Dupont) and UV detection. Retinol was detected at 325 nm; -tocopherol was detected by fluorescence excitation at 292 nm and emission at 325 nm. USP standards of all-trans-retinol and -tocopherol were used to generate standard curves.

Serum concentrations of 25(OH)D₂ and 25(OH)D₃ were measured simultaneously by HPLC (Kao and Heser 1984). The serum samples were acidified with concentrated hydrochloric acid, and the 25-hydroxy metabolites extracted with prepacked octadecylsilano silica cartridges (C-18 Bond Elute, Analytichem International, Harbor City, CA). Further purification was accomplished by reextracting on aminopropyl cartridges (NH₂ Bond Elute, Analytichem International). The metabolites were separated and quantified with a silica column (Zorbax, 5-µm, Dupont) and UV detection (Kratos Analytical). Concentrations of the two metabolites were calculated from a single calibration curve established with a 25(OH)D₃ standard (Duphar, Amsterdam, The Netherlands). Recovery was determined for each serum sample by adding a 25-hydroxy-(25[27]-methyl-³H)cholecalciferol standard (Amersham, Arlington Heights, IL) to the sample before extraction.

PT was measured weekly as part of the clinical chemistry battery.

Values of serum 25(OH)D₂ or 25(OH)D₃ concentrations that fell below the limit of detection (~4 nmol/L) were assigned a value of one half the detection limit, rather than zero, for the purpose of statistical analysis (Helsel 1990).

The stability of dietary components was determined by regression analysis of mean values vs. time.

All analyses were conducted with SAS Version 5.18 or 6.06 software (SAS Institute, Cary, NC). All comparisons were made at the two-tailed 0.05 significance level.

RESULTS

Table 5. Concentrations of retinyl palmitate, -tocopherol, -carotene, phylloquinone and olestra (OA) in purified diets measured over 4 wk [View Table]

The concentration of olestra decreased slightly in the 1 NRC(OA) diet, determined from a single measurement at each time point. Subsequent measurements made on the same purified diet over a period of 8 mo showed no change in olestra concentration (Cooper et al. 1997b), an indication that the trend observed here is unlikely to represent instability of olestra in these purified diets.

No significant differences were found between males and females with respect to their pattern of growth; therefore the data were combined for intergroup comparisons. All groups grew at the same rate. Cumulative weight gains at the end of the study are shown in Table 6, expressed as absolute gain, gain relative to starting body weight or gain relative to body surface area, for all groups. For the group fed purified diet providing 1.6 of the NRC requirements for nutrients, without olestra, the absolute cumulative weight gain was significantly lower than that for the group fed the standard feed containing 14% fat; however, the significant difference disappeared when the gain was expressed relative to starting weights or body surface area. Other significant differences were found among these parameters at other time points. The differences were randomly distributed among the groups and there was no consistent trend to indicate that any group grew at a significantly different rate.

Table 6. Cumulative weight gains for pigs fed standard feed or purified diets with or without 4.8% olestra for 4 wk1 [View Table]

Cumulative feed consumption, expressed in terms of digestible energy, and cumulative feed efficiency, calculated as cumulative weight gain in kilograms per cumulative energy consumed, were the same for all groups (Table 7).

Table 7. Cumulative energy consumption and feed efficiency for pigs fed standard feed or purified diets with or without 4.8% olestra for 4 wk1 [View Table]

Table 8. Concentrations of vitamin A and vitamin E in liver of pigs fed purified diets with or without 4.8% olestra for 4 wk1 [View Table]

Olestra produced significant reductions in the liver concentrations of vitamins A and E (Table 8). Addition of 4.8% olestra to the diet providing the NRC requirements for the two vitamins reduced liver vitamin A and vitamin E concentrations by 55 and 66%, respectively.

The serum concentration of 25(OH)D₂ , measured at the end of the study, increased in a dose-responsive manner as the concentration of ergocalciferol in the purified diet was increased from 1 to 1.6 times the NRC requirement (Table 9). The trend was also present at wk 2. Relative to base line, serum 25(OH)D₂ concentration increased by 2.5, 4.3 and 5.9 times over the 4 wk of the study for pigs fed the 1 , 1.3 or 1.6 NRC diet, respectively.

Table 9. Serum concentrations of 25-hydroxycholecalciferol [25(OH)D3] and 25-hydroxyergocalciferol [25(OH)D2] for pigs fed purified diets with or without 4.8% olestra and with or without daily exposure to UV light for 4 wk1 [View Table]

Olestra significantly reduced the serum 25(OH)D₂ concentration when added to the 1 and 1.6 NRC diets. The mean reduction was 33% for the pigs fed the 1 NRC diet.

For the pigs that were not exposed to UV light, serum 25(OH)D₃ concentration in most animals was essentially at the detection limit of the analytical method; this was an expected result because the animals also had received no sun exposure. Measurements of serum 25(OH)D₂ and 25(OH)D₃ at wk 2 showed the same trends as the measurements made at the end of the study.

To determine the amount of UV exposure needed to provide 50-80% of serum total 25-hydroxyvitamin D concentration from the synthesis of vitamin D₃ for pigs fed the NRC requirement of ergocalciferol, a linear regression of the inverse of serum 25(OH)D₃ concentration against UV exposure time was performed. The inverse of the serum concentration, [25(OH)D₃], and the exposure time, T_exp , were used for the regression because the expression 1/[25(OH)D₃] = A + B(T_exp) is equivalent to a hyperbolic dose response. The equation resulting from the regression was as follows:

This relationship predicts that 1-2 min/d exposure is needed to produce a 50-80% contribution of 25(OH)D₃ to serum total 25-hydroxyvitamin D concentration for pigs when the inhibitory effect of UV light on serum 25(OH)D₂ concentration is considered. A 50-80% contribution of vitamin D₃ to total vitamin D status models the human situation (Arnaud et al. 1977, Delvin et al. 1979).

No significant differences in PT were found among the groups fed the standard feed diets or the purified diets with or without 4.8% olestra (Table 10).

Table 10. Prothrombin time (PT) for pigs fed standard feed or purified diets with or without 4.8% olestra for 4 wk1 [View Table]

DISCUSSION

Pigs fed a purified diet in which all micronutrient contents met NRC requirements, with 30% of energy from fat, grew at a rate equivalent to that produced by the standard swine feed with or without 30% of energy from fat. Olestra had no effect on growth. The pigs fed the purified diets, with or without olestra, grew at the same rate and had the same feed efficiency. All of the pigs grew at essentially the same rate observed by other researchers for the same crossbred strain fed a corn-soybean meal. During the last week of the study, for example, when the pigs weighed 25-30 kg, the growth rate for all groups was about 5 kg/wk, the same rate that Martin and Crenshaw (1989) reported for 11- to 60-kg pigs of the same cross breed.

The purified diet providing vitamins A, D, E and K at or near the NRC requirements for 5- to 10-kg swine produced adequate but marginal stores of these vitamins. The pigs' vitamin stores were responsive to changes in dietary concentrations of the vitamins and were affected by olestra in the diet. The serum concentration of 25(OH)D₃ responded to UV exposure.

This is the first study in swine to show the adequacy of a diet in which all micronutrients were set to the NRC requirement level for swine. Purified diets composed of similar ingredients have been developed and used previously for swine nutrition research, but those diets contained multiples of the NRC requirements of some micronutrients (Cunha et al. 1994, Heinemann et al. 1946, Miller et al. 1965). The adequacy of the 1 NRC diet fed in this study indicates that the micronutrient requirements, determined under conditions in which a single micronutrient is limiting in the diet, are valid when all micronutrients in the diet are limited to their requirement level. This is significant in light of the numerous interactions known to occur among micronutrients (Machlin and Langseth 1988). Experience in other species indicates that deficiencies in micronutrients present at close to required levels are sometimes manifested when the levels of other nutrients are manipulated. This was the case for vitamin K in the AIN-76 diet for rats (Bieri 1980) and for potassium in domestic cats fed purified diets with variable protein concentration (Hills et al. 1982).

Body weight gain and digestible feed efficiency were used as key measures of the adequacy of the dietary micronutrient levels. These parameters are the most reliable overall indicators of micronutrient status when protein intake is sufficient and when energy intake is controlled, as in this study. Casein, a complete and easily digested source of amino acids and nitrogen, was included in the purified diets at a level of 25%, enough to provide sufficient protein intake. Energy intake was controlled at 95% of the NRC energy requirement for growing swine. Feeding at this level ensured that the pigs ate all or nearly all of their daily feed allotment without experiencing a significant limitation in growth rate.

Because these pigs grew rapidly, they are a sensitive model in which to detect effects of substances on nutrient bioavailability inasmuch as nutrient pools are expanding rapidly and nutrient demands to support growth are high. Growth was measured for only 4 wk in this study, a length of time adequate to test the nutritional adequacy of a diet; however, a subsequent study has shown that the same pigs fed the 1 NRC diet used here maintained similar growth rates for up to 26 wk (Cooper et al. 1997b). Therefore this diet is adequate for use in both short- and long-term studies.

The use of a diet that provides micronutrients at or near the NRC requirement levels, as in this study, offers a rigorous test of factors that affect nutrient availability because nutrient stores are proportional to intake in this range of dietary nutrient levels. The 1 NRC diet provided intakes that were adequate but not excessive. Feeding pigs a diet that provides nutrients at the NRC level of requirements, the 1 NRC diet, increases the possibility that an effect on any given micronutrient will be detected as a decrease in growth. Depressed growth is a common sign of nutrient deficiency.

The NRC requirement produces a lower level of nutrient status in the pig than the RDA produces in humans because the RDA for most nutrients is set at a value sufficiently above the average requirement, usually 2 SD, to meet the needs of almost everyone in the population (NRC 1989). Pigs fed a diet that provides micronutrients at the NRC requirement level thus model the situation in which humans consume marginal but adequate levels of micronutrients. Such individuals are most at risk of nutritional deficiency if nutrient availability is further restricted.

The 1 NRC diet provided adequate but not excessive fat-soluble vitamin status in ranges responsive to changes in availability. The mean liver vitamin A concentration in the group fed the 1 NRC diet was 55.5 nmol/g liver. This liver concentration is sufficient to maintain normal plasma retinol concentrations in pigs (Hennig et al. 1985) but is below the 70 nmol/g liver considered adequate in humans (International Vitamin A Consultative Group 1993). The mean liver vitamin E concentration in this group of pigs was 11.6 nmol/g liver. This is similar to values that other researchers have measured in pigs fed the NRC requirement of vitamin E in a purified diet, and higher than required to prevent clinical deficiency (Hoppe et al. 1993). The serum 25(OH)D₂ concentration for pigs fed the 1 NRC diet was adequate to prevent clinical deficiency and similar to values measured by other researchers in pigs fed the NRC requirement of vitamin D (Engstrom and Littledike 1986). Vitamin K status, as measured by PT, was normal in the pigs fed the 1 NRC diet.

The NRC requirement levels used for these and all other micronutrients were those for 5- to 10-kg pigs, the size of the pigs at the start of the acclimation period. Even though NRC guidelines indicate that the requirements for micronutrients decrease as pigs age, the stores and tissue levels of vitamins A, D and E did not change substantially from the beginning to the end of the study. This finding supports the use of the 5- to 10-kg pig requirements for pigs up to at least 30 kg, the weight of the pigs at the end of this study. Use of a single level of nutrients prevents possible perturbations of steady-state vitamin concentrations in tissues.

The chemical forms of the fat-soluble vitamin in the purified diets were chosen to be as relevant to human nutrition as was practical. Vitamin A was provided as a 3:1 mixture of retinyl palmitate and -carotene, typical of the average proportions and the forms of vitamin A in the U.S. diet (Olson 1983). Including -carotene as a dietary source of vitamin A ensures that the full effect of a nonabsorbed lipophilic substance, such as olestra, on vitamin A status will be measured. The interaction between olestra and lipophilic nutrients increases as the lipophilicity of the nutrient increases (Jandacek 1982). The availability of -carotene should be more strongly affected by olestra, or any other nonabsorbed lipophilic substance, than the availability of retinyl palmitate because -carotene is more lipophilic. A measure of the lipophilicity of a molecule is its octanol-water partition coefficient (p_c). The value of the octanol-water partition coefficient is generally expressed as log₁₀ p_c . Water-soluble molecules have log₁₀ p_c values 0; lipid-soluble molecules have log₁₀ p_c values >.0. The log₁₀ p_c for -carotene is 17.6; the value for retinol, the species absorbed when retinyl palmitate is eaten, is 7.6 (Cooper et al. 1997d). The greater effect of olestra on the absorption of -carotene, in relation to the effect on the absorption of retinyl palmitate, has been shown in human studies (Daher et al. 1997b, Schlagheck et al. 1997a and 1997b).

Vitamin E was provided as dl--tocopheryl acetate, a source of vitamin E in the U.S. diet and a model compound for the various isomers of the vitamin (Farrell 1988). Vitamin D and vitamin K were provided as ergocalciferol and phylloquinone, respectively, major sources of these vitamins in the human diet. Menadione is the usual source of vitamin K in swine diets. If menadione were used instead of phylloquinone, however, the potential for olestra to affect vitamin K status in humans could not be detected because menadione is poorly soluble in lipids and thus unlikely to be affected by olestra. The log₁₀ p_c for menadione is 2.2.

Other conditions used in the study model human dietary conditions. For example, the fat content of the diet was set at 30% of energy. This level was based on current dietary recommendations for good health (U.S. Department of Health and Human Services 1988) and contrasts with the very low fat levels typically used in swine nutrition studies. A dietary fat level similar to that in the human diet is important when investigating the availability of fat-soluble vitamins because of the role of fat in the absorption of these vitamins and in gastric emptying. Olestra had no effect on the digestion or utilization of fat based on the absence of an effect on growth and digestible feed efficiency in the olestra-fed groups. Olestra has been shown not to significantly affect triglyceride absorption in humans (Daher et al. 1997c).

The pigs were fed three meals per day to mimic the major eating pattern in humans. Pigs given free access to feed eat 7-12 times during the day and usually do not consume large amounts at any one time (Houpt and Houpt 1991). In this study, the pigs adapted readily to the three-meals-a-day regimen. Use of a meal-eating pattern similar to the human pattern is important inasmuch as eating patterns theoretically can affect the digestion and metabolism of nutrients.

The pigs were exposed to UV light to stimulate the cutaneous production of vitamin D₃. Including both vitamin D₂ and vitamin D₃ as contributors to vitamin D status models the human situation, in which vitamin D₃ is the primary contributor. Generally, dietary vitamin D₂ provides only 10-20% of vitamin D status in humans (Arnaud et al. 1977, Haddad and Hahn 1973, Jones 1978, Poskitt et al. 1979). This study showed that a vitamin D₃ contribution to total vitamin D status typical of the human situation can be obtained in pigs by 1-2 min/d exposure to UV light.

Exposure to UV light caused a reduction in the serum concentration of 25(OH)D₂ . The pigs fed the 1.6 NRC diet and exposed to UV light had lower serum 25(OH)D₂ concentrations than the pigs fed the 1.6 NRC diet and not exposed to UV light. A plausible explanation for this observation is competition between vitamin D₃ and vitamin D₂ for the liver 25--hydroxylase enzyme. In pigs, vitamin D₃ is the preferred substrate for the liver enzyme (Horst et al. 1982). As a result, an increase in the amount of vitamin D₃ available for hydroxylation might be expected to cause a decrease in serum 25(OH)D₂ concentration. Because of this interaction, the effect of a substance such as olestra on the availability of dietary vitamin D₂ must be determined in the pig under conditions in which the amount of UV exposure is the same for all test groups.

In agreement with studies in the rat (Mattson et al. 1979) and in humans (Jones et al. 1991a and 1991b, Koonsvitsky et al. 1997, Schlagheck et al. 1997a and 1997b), olestra reduced vitamin A, D and E status, but did not affect vitamin K status. Subsequent studies in the pig confirmed these results (Cooper et al. 1997b and 1997c, Daher et al. 1997a) and showed that the effects could be offset by introducing additional amounts of the vitamins to the diet (Cooper et al. 1997a and 1997b). The effects of olestra on vitamin A, D and E status were the same whether the pigs were fed the NRC requirements of the vitamins or 1.6 times the requirements. This finding shows that the effect of olestra is not affected by dietary nutrient concentrations in this range.

The results from this study confirm that the domestic pig, fed a purified diet providing the NRC requirements of nutrients and 30% of energy from fat, can be used to assess the effects of olestra on nutrient availability. In addition to the important similarities between pigs and humans in GI physiology and in nutritional needs and utilization, other considerations make the pig a useful model for developing relevant information about potential nutritional effects of olestra in humans, both adults and children. For example, studies in the pig can be conducted with olestra intakes that are exaggerated in relation to human intakes, and invasive methods can be used to measure nutrient stores. In addition, use of the weanling pig allows studies to be conducted over periods of time that cover the major growth and developmental phases of the pig's life, including the development of nutrient stores, which occurs early in life. This last-mentioned ability is extremely important in assessing the potential effects of olestra on the nutritional status of young children.

ACKNOWLEDGMENTS