Olestra (Olean, Procter & Gamble, Cincinnati, OH) is a mixture of hexa-, hepta- and octaesters of sucrose prepared by esterifying sucrose with long-chain fatty acids isolated from edible fats and oils (Rizzi and Taylor 1978). The physical properties of olestra are similar to the those of a triglyceride with the same constituent fatty acids (Jandacek and Webb 1978). Olestra made from highly unsaturated fatty acids is liquid at room temperature; olestra made from highly saturated fatty acids is solid. A schematic representation of the structure of olestra, in this case an octaester, is shown in Figure 1, along with a schematic triglyceride structure for comparison.

Olestra has the organoleptic and thermal properties of fat (Bernhardt 1988, Kester 1993) but is not hydrolyzed by gastric or pancreatic enzymes (Mattson and Volpenhein 1972) and is not absorbed intact from the gastrointestinal (GI)³ tract (Daher et al. 1996, Miller et al. 1995). For this reason, olestra does not provide calories to the diet. Because of its unique properties, olestra can serve as a zero-calorie replacement forfat in a variety of foods. The U.S. Food and Drug Administration (FDA) has approved the use of olestra as a replacement for fats and oils currently used in the preparation of savory snacks such as potato and corn chips (Federal Register 1996).

Traditional food additives are generally present in the diet in small (i.e., milligram) quantities. The safety of such food additives is evaluated in animal tests conducted with doses of the additive that are substantial multiples of expected human intakes. The FDA has published guidelines for establishing the safety of such additives (USFDA 1982). In contrast, a macronutrient replacement such as olestra may make up a significant portion of the diet (i.e., gram amounts). Testing such additives, especially nonnutritive ones, at doses that represent several multiples of expected human intake can produce results that may be confounded by nutritional imbalances in the animals and, for this reason, uninterpretable. Therefore, special considerations must be included in the design of a program to assess the safety of such an additive (Borzelleca 1992, Gershoff 1996, Vanderveen and Glinsmann 1992). An evaluation of nutritional effects becomes a critical part of the safety assessment of such additives because large quantities of an additive in the GI tract might affect the absorption or digestion of other dietary components, especially if the additive is not absorbed.

The safety of olestra has been investigated in a variety of traditional animal and human studies (Bergholz 1992), including long-term studies in rats (Wood et al. 1991), mice (Lafranconi et al. 1994) and dogs (Miller et al. 1991), and multigeneration studies in rabbits (Nolen et al. 1987). In those studies, olestra was fed in daily amounts as high as judged possible without complicating the results by introducing nutritional imbalances. In most tests, the daily amounts fed were severalfold greater than expected human intakes. No evidence of toxicity was found.

To evaluate the potential for olestra to affect the absorption and utilization of other dietary components and to determine how any observed effects might be offset, a number of nutritional studies have been conducted in an animal model, the domestic pig, and in humans. This introductory paper describes the general approach of this nutrition program; experimental details and results of the individual studies are presented in subsequent papers in this issue.

INTERACTION OF OLESTRA WITH OTHER DIETARY COMPONENTS IN THE GASTROINTESTINAL TRACT

The introduction of a nonabsorbable, inert substance such as olestra into the GI tract raises the possibility that the digestion and absorption of other substances might be affected. For example, it is known that high dietary levels of insoluble fibers can affect the bioavailability of minerals (Reinhold et al. 1976, Rossander 1987). From knowledge of olestra's properties and on the basis of physiochemical principles, Jandacek (1982) proposed a mechanism by which olestra might affect the absorption of other dietary components. The basis of this mechanism is an interference with the micelle-mediated absorption process resulting from a partitioning of dietary components between olestra and the mixed micelles. Some portion of the dietary components, especially those that are fat soluble, could partition into the olestra in the GI tract and not be incorporated into the intestinal micelles. That portion would be carried out of the body with the nonabsorbed olestra, possibly resulting in a decrease in absorption.

Several characteristics or consequences of this partitioning mechanism can be predicted from physiochemical principles. First, highly lipophilic molecules should be affected more strongly than less lipophilic molecules because olestra has a greater affinity for strongly lipophilic molecules than for weakly lipophilic molecules. At the other extreme, water-soluble molecules will not partition into olestra and thus olestra should have a negligible effect, if any, on the absorption of water-soluble substances.

Whether or not olestra might affect the absorption of a substance can be predicted from knowledge of the lipophilicity of the substance. The lipophilicity of a molecule can be determined from its equilibrium distribution, expressed as a partition coefficient, between an aqueous phase and an oil phase. Partition coefficients can be measured or calculated from a knowledge of molecular structure (Hansch and Leo 1979). When the distribution is measured, octanol is normally used as the oil phase and the partition coefficient is generally expressed in log units (log₁₀ p_c). Water-soluble molecules have log₁₀ p_c values < 0; lipophilic molecules have values > 0. Knowledge of the octanol-water partition coefficient allows a prediction to be made about whether olestra will affect absorption or not but, among substances lipophilic enough to be affected, does not allow the precise magnitude of the effect to be predicted. That must be determined experimentally and depends on other factors.

An important characteristic of the partitioning mechanism is that olestra will affect the absorption of another substance only if the two are present in the GI tract together because the mechanism is a physical interaction that occurs in the lumen of the gut. Separating the time between the ingestion of olestra and a dietary component will decrease or eliminate the opportunity for the two to interact in the GI tract and thus decrease or abolish the potential for olestra to affect absorption. This eating-pattern factor is critically important when trying to extrapolate results of clinical studies to real-life situations if the studies use dietary patterns for ingestion of olestra and nutrients different than those that occur in real life.

Another predictable characteristic of the partitioning mechanism is that the greater the amount of olestra in the GI tract to compete with the intestinal micelles, the greater the olestra effect on absorption (i.e., the effect should be dose responsive with respect to olestra). This factor is of course important when extrapolating results of animal and clinical studies to real life.

Finally, because olestra simply reduces the availability of the nutrient to the intestinal micelles, it in effect lowers the amount available from the diet, and it would be predicted that any effect of olestra on nutritional status can be offset by supplying an extra amount of the affected nutrient to the diet.

Further evidence that olestra does not affect the absorption of water-soluble molecules is found in a study in which the absorption of glucose (log₁₀ p_c = 2.4) was investigated in the rat (FAP 1987). The effect of olestra on glucose absorption was assessed by measuring blood glucose concentration after rats were dosed with olestra and glucose. Rats were intubated with 0.35 g of olestra immediately followed by 4.5 mL of a solution containing 10g/100 mL glucose. Blood samples were taken by tail clip and glucose concentrations measured with a commercial glucose analyzer. Olestra affected neither total glucose absorption, which is proportional to the area under the blood glucose concentration-time curve, nor the peak blood glucose concentration.

The absorption of the weakly lipophilic drugs propanolol (log₁₀ p_c = 2.6) and diazepam (log₁₀ p_c = 2.7) in humans was unaffected by a single 18-g dose of olestra (Roberts and Leff 1989). The absorption of the contraceptive steroids norethindrone (log₁₀ p_c = 3.0) and estradiol (log₁₀ p_c = 3.9) also was unaffected in this test. These steroids are more lipophilic than diazepam, for example, but considerably less lipophilic than dietary constituents such as cholesterol or the fat-soluble vitamins. A 28-d crossover study in premenopausal women taking the oral contraceptive steroids ethinyl estradiol (log₁₀ p_c = 3.7) and norgestrel (log₁₀ p_c = 3.5) confirmed that the absorption of these moderately lipophilic molecules was unaffected by olestra (Miller et al. 1990).

A human study in which five inpatients ate 50 g/d of olestra for 10 d provided evidence that olestra does not significantly affect the absorption and utilization of dietary fat (Fallat et al. 1976). In this study, the level of dietary nonolestra lipids excreted in the feces was used as a measure of fat absorption. During the 10-d period in which olestra was eaten, 97.9 ± 0.7% of dietary fat was absorbed. In comparison, a value of 98.5 ± 0.5% was measured during a 10-d basal period in which no olestra was eaten. Further evidence that olestra does not affect the absorption or utilization of fat, or other macronutrients, comes from long-term animal feeding studies (Lafranconi et al. 1994, Miller et al. 1991, Wood et al. 1991). Olestra, fed at levels of up to 10% of the diet, had no effect on growth or feed efficiency of the animals in these studies.

In contrast to the lack of effect of olestra on water-soluble and slightly or moderately lipophilic molecules, there is evidence that olestra can affect the absorption of highly lipophilic molecules. An increase in fecal cholesterol (log₁₀ p_c = 8.7) excretion was observed in rats fed olestra (Mattson et al. 1976). Decreases in serum cholesterol (Crouse and Grundy 1979, Fallat et al. 1976, Glueck et al. 1979) and increases in fecal cholesterol (Jandacek et al. 1980 and 1990) were observed in humans consuming olestra. These molecules are about one million times more lipophilic than the oral contraceptives.

Human studies have shown that olestra can affect the absorption of fat-soluble vitamins such as vitamin D (log₁₀ p_c = 10.4) and vitamin E (log₁₀ p_c = 12.2). Serum 25-hydroxyergocalciferol concentration, that component of serum total 25-hydroxyvitamin D that comes from the diet, was reduced by ~19% in normal healthy subjects who ate 20 g/d olestra for 6 wk (Jones et al. 1991b). A number of early clinical studies showed that olestra could affect serum tocopherol concentration (Crouse and Grundy 1979, Fallat et al. 1976, Glueck et al. 1979 and 1982, Mellies et al. 1983 and 1985). The observed effects ranged from a 17% reduction for subjects who ate 27 g/d olestra for 16 wk (Mellies et al. 1985) to a 44% reduction for subjects who ate 50 g/d olestra for 10 d (Glueck et al. 1979).

Serum phylloquinone (log₁₀ p_c = 11.7) concentration, measured at biweekly intervals for 6 wk, was reduced at wk 2 and 4 but not at wk 6 in the study in which the subjects ate 20 g/d olestra for 6 wk (Jones et al. 1991a). However, there was no effect on vitamin K function as measured by the Simplastin-Ecarin assay, a measure of under--carboxylated prothrombin (Suttie et al. 1988), or by prothrombin and partial thromboplastin times.

Olestra, admixed in the diet at high concentrations, has been shown to reduce liver vitamin A stores in the rat (Mattson et al. 1979); however, variable effects on plasma retinol (log₁₀ p_c = 7.6) have been reported for humans (Crouse and Grundy 1979, Fallat et al. 1976, Glueck et al. 1979 and 1982, Mellies et al. 1983 and 1985). Because plasma retinol levels do not reflect changes in vitamin A intake until liver stores of the vitamin become depleted (Olson 1984, Sauberlich et al. 1974), changes in plasma retinol concentration would not be expected to occur for many months or longer. The apparent reductions seen in some studies at some time points may have resulted from the analytical technique used in most of the studies, a fluorometric assay. Measurement of plasma retinol by this assay is subject to interference by the carotenoid phytofluene (Thompson et al. 1971). In a study in which HPLC was used to measure plasma retinol concentration, no effect on serum retinol was observed (Mellies et al. 1985).

OLESTRA NUTRITION PROGRAM

The pig has been used extensively in nutrition research (Miller and Ullrey 1987, Tumbleson 1986), including evaluation of the nutritional effects of a poorly absorbed fat substitute (Hayes et al. 1994a and 1994b). There are several factors that make the pig a suitable animal in which to conduct nutritional studies. Its GI tract is physiologically and anatomically similar to the human GI tract (Kidder and Manners 1978, Kurihara-Bergstrom et al. 1986, Leigh-Browne and Harpur 1975, Miller and Ullrey 1987); at the weanling stage, the GI tract of the pig closely resembles that of a 2- to 5-y-old child (Juskevich and Guyer 1990, Leary and Lecce 1976). The requirements of the growing weanling pig for macro- and micronutrients meet or exceed those of most subgroups of the human population (NRC 1979, 1988 and 1989). For example, its demands exceed those of children or pregnant or lactating women, subgroups who might be at nutritional risk because of high demand for certain nutrients. The rapid growth rate of the pig markedly exaggerates nutrient demands in relation to those of humans, thus making the pig a sensitive model. Pigs reach sexual maturity and display a reduced growth rate by 6-8 mo of age (Anderson 1974), allowing feasible and practical studies to be conducted over periods covering the major growth and developmental phases of the life of the pig. Finally, the pig is able to ingest, tolerate and metabolize fat at a level commonly found in the human diet.

Specific studies comprising the olestra nutrition program are briefly described in the following sections. References are provided to papers in this issue that describe the studies in detail.

Before the initiation of the human and pig studies, a menu census survey was conducted among nationally representative households using methodology originally developed by the National Academy of Sciences (NAS) generally regarded as safe (GRAS) review committee (Abrams 1992) and refined in continued work with the NAS GRAS Review Committee and with the FDA. The survey uses diaries to track foods and beverages eaten by ~5000 individuals over a 14-d period. The objective of the survey was to estimate the amounts of olestra various subgroups of the population might ingest from eating savory snacks (Webb et al. 1997). To do this, estimates of olestra intake were computed for each individual in the survey from the number of times snacks were eaten over a 14-d period, the average amount of the snack consumed at each eating occasion and the amount of olestra in the snacks, assuming 100% replacement of fat. Intakes by each individual in the survey were accumulated for each day to arrive at a single-day (acute) intake. Intakes were also accumulated for each individual over the 14-d period to provide a 14-d average (chronic) intake. The pattern and frequency of consumption of savory snacks were also determined from this survey. The estimated olestra intakes were considered when choosing daily intakes of olestra to be tested in the pig and human studies.

Pig studies. A total of six studies were conducted in the domestic pig. The first of two 4-wk studies was conducted to establish 1 ) that the pig responds in a predictable manner to consumption of purified diets with and without olestra, 2 ) that the pig can tolerate a high fat diet, 3 ) that measurable changes in body stores of fat-soluble vitamins can be produced in reasonable time, and 4 ) that husbandry procedures and analytical methodology be validated (Cooper et al. 1997d). The study also demonstrated that exposure to UV light produced an amount of cutaneously synthesized vitamin D in the pig, typical of that found in humans. The second 4-wk study quantified the extent to which feeding olestra admixed in the diet exaggerated the olestra effects in relation to feeding it in food forms (e.g., snack foods such as potato chips) fed along with the main diet at meals or as between-meal snacks (Daher et al. 1997a).

The first of two 12-wk studies was conducted 1 ) to establish the dose response of olestra on body stores or circulating concentrations of the fat-soluble vitamins A, E and D, and on vitamin K function (Cooper et al. 1997c) and 2 ) to determine if olestra affected the status of selected water-soluble micronutrients or the absorption and digestion of macronutrients.

The second 12-wk study was conducted to establish that the effects of olestra on tissue concentrations of the fat-soluble vitamins A, D and E could be offset by adding extra amounts of the vitamins to the diet and to determine the amounts of extra vitamins required to offset the effects (Cooper et al. 1997a). An important question asked in both 12-wk studies was whether different body pools of vitamin E responded similarly to olestra and to the addition of extra amounts of the vitamin to the diet. This question was addressed to provide assurance that measurements of serum vitamin E can be used to assess the effect of olestra on vitamin E status in humans.

Two long-term pig studies were conducted, one for 26 wk and one for 39 wk (Cooper et al. 1997b). These studies were conducted to confirm that the effects measured in the shorter studies did not worsen as the pigs aged and that no additional effects occurred. In these studies, olestra intakes representative of expected human intakes, as well as exaggerated intakes, were tested, and the studies were conducted over periods extending beyond the major growth and development stages of the pig's life, from weaning through sexual maturity. These studies were also used to confirm that the extra amounts of vitamins A and E required to offset the effect of olestra on these vitamins could be predicted and were effective at olestra intakes representing expected human intakes over periods covering major growth and developmental stages.

The pig studies were designed to determine the nature of the olestra effects under exaggerated dietary conditions. The olestra intake was exaggerated relative to estimated chronic and acute human intake, and the frequency of consumption was exaggerated relative to human consumption under free-living dietary patterns. The diets were formulated to provide essential nutrients at requirement levels to increase the possibility that any potential effect of olestra on nutrient availability would be seen. Olestra was admixed in the diet to maximize the opportunity for olestra to interact with the nutrients.

Human studies. Studies were conducted with healthy, adult human volunteers 1 ) to confirm and extend the data from the pig studies in the species of interest under relevant as well as exaggerated olestra exposures and 2 ) to obtain data on some nutrients not possible or practical in the pig studies.

Five individual human studies were conducted. A 16-wk study was conducted in free-living male and female adults eating habitual self-selected diets 1 ) to determine the effect of a single daily intake, 18 g/d, of olestra on the status of the fat-soluble vitamins D, E and K and the absorption of carotenoids and 2 ) to assess the feasibility of offsetting the effects of olestra on vitamin E absorption by adding extra amounts of vitamin E to olestra (Koonsvitsky et al. 1997).

Two 8-wk studies were conducted with healthy 18- to 44-y-old male and female volunteers fed preselected and controlled diets. Individuals of this age were chosen for the study population because that age group has the greatest estimated intake of olestra from savory snacks (Webb et al. 1997). The first study was conducted to determine 1 ) the dose response of olestra on the absorption of the fat-soluble vitamins and the carotenoids, 2 ) the effect on sensitive measures of vitamin K function, and 3 ) the effect on selected water-soluble micronutrients (Schlagheck et al. 1997b).

The second 8-wk study was conducted to confirm that restoration of vitamins D and E status, demonstrated in the pig studies, could be accomplished in humans, to determine the amounts of the affected vitamins required to restore status of the vitamins to control levels and to confirm that olestra does not affect the availability of water-soluble micronutrients (Schlagheck et al. 1997a). In these studies, the olestra was eaten in food forms typical of its intended use, and the daily olestra intake was evenly divided among the three daily meals, again to maximize the potential interaction with other dietary components.

Two studies were conducted with adult male volunteers to determine the dose response of olestra on the absorption of retinyl palmitate and to confirm that olestra had no nutritionally significant effect on the absorption of dietary triglyceride (Daher et al. 1997b and 1997c). These studies were conducted in metabolic wards using radiochemical techniques to measure absorption.

As with the pig studies, the human studies employed dietary conditions that were exaggerated in relation to expected real-life dietary conditions. Dietary conditions used in the two 8-wk studies were the most exaggerated. These studies tested olestra daily intakes that were up to four times the 8.1 g/d estimated chronic 90th-percentile intake by 18- to 44-y-old males, the age group with the greatest estimated intake of snacks. Olestra was eaten with every meal, three times a day over the 56 d of the studies (i.e., 168 consecutive meals), and the diets were controlled where possible to provide 80-120% of the recommended dietary allowance (RDA) of the micronutrients of interest. All meals were eaten under supervision of study personnel and the subjects were not allowed to eat anything between meals. Under these conditions, all nutrients were eaten simultaneously with olestra.

Data from the population-based menu census survey showed that 90th-percentile snack eaters (both gender, all ages) eat savory snacks about 10 times in a 14-d period and potentially consume an estimated 6.9 g/d of olestra from those snacks (Webb et al. 1997). The 50th-percentile savory snack eater eats snacks about four times in a 14-d period, and the average potential intake of olestra from those snacks is estimated to be 3.1 g/d. A substantial amount of nutrients are eaten at times other than when olestra would be eaten. At the 50th percentile, only 7% of main meals include a savory snack; at the 90th-percentile intake level, 18% of main meals contain savory snacks. Because of the exaggerated conditions used in the 8-wk studies (42 eatings in 14-d repeating periods, at every meal), the effects measured are substantially exaggerated in relation to effects that might occur when olestra is eaten as an ingredient of savory snacks foods in real life.

In the 16-wk study, the subjects ate 18 g/d olestra, a daily intake almost three times the estimated 90th-pecentile intake for the total population of snack eaters. The olestra food items were to be consumed with meals at the subject's discretion throughout the day. The subjects were not specifically requested to evenly divide the daily allocation among the meals nor were they restricted from eating between meals. All other food items were self-selected and were consumed ad libitum. Under this protocol, the subjects ate olestra up to 42 times in repeating 14-d periods, but were free to eat other foods at times when the olestra foods were not eaten.

In the retinyl palmitate and triglyceride absorption studies, the subjects ate olestra at daily intakes that were four times the estimated chronic 90th-percentile intake by 18- to 44-y-old eaters of savory snacks, and the absorption of the nutrients eaten simultaneously was measured.

In both the pig and human studies, the olestra was heated before being incorporated into the diet or food items under conditions equivalent to or more severe than those used in the commercial preparation of savory snacks. This was done to ensure that the olestra eaten would contain the same kinds and amounts of degradation products as those that might be found in olestra used to prepare savory snacks commercially. Comprehensive analytical analyses have shown that olestra undergoes the same oxidative and hydrolytic processes during heating as do triglycerides (Gardner et al. 1992, Henry et al. 1992).

To determine whether any olestra effects observed might be nutritionally or physiologically meaningful, parameters that reflect overall status were measured when possible. These included serum concentrations of 25-hydroxyvitamin D, 1,25-dihydroxyvitamin D, parathyroid hormone; liver and serum concentration of -tocopherol; liver stores of total vitamin A; serum concentrations of carotenoids, iron, zinc and calcium; plasma concentrations of folate and prothrombin; bone concentrations of calcium, zinc and phosphorus; and the urinary concentration of zinc. The absorption of preformed vitamin A was measured directly as was the absorption of vitamin B₁₂ and triglyceride. In addition, sensitive indicators of vitamin K function, such as the plasma concentration of des--carboxyprothrombin and the urinary excretion of -carboxyglutamic acid, were measured in humans.

Other studies. In addition to the pig and human studies, the following studies or assessments were conducted. Results from the menu census survey used to estimate olestra intake for savory snacks and snack eating patterns were used to assess the potential effect of olestra on carotenoid absorption under real-life dietary patterns (Cooper et al. 1997e). The frequency of co-consumption of savory snacks and carotenoid-containing foods determined from the survey was used in combination with the effect on carotenoid absorption observed when the two were eaten together to estimate the olestra effect on carotenoid absorption in real life.

Another assessment was made to evaluate the potential for olestra to affect the availability of dietary phytochemicals, potentially beneficial substances found in fruits and vegetables (Cooper et al. 1997e). Octanol-water partition coefficients, a measure of lipophilicity, were collected for a large number of molecules covering the major classes of phytochemicals. These partition coefficients were then compared with partition coefficients of molecules for which an olestra effect, or lack of effect, had been experimentally determined. From this comparison, the potential for olestra to affect the absorption of numerous classes of phytochemicals and specific individual phytochemicals was estimated.

Finally, the results of the pig and humans studies were assessed to determine if they were applicable to subgroups of the population such as children and pregnant and lactating women. These subgroups may be at greater risk of nutrient insufficiency than the general population because of greater nutritional needs or because of unique dietary patterns that produce greater olestra-to-nutrient intake ratios than those found in the general population (Middleton et al. 1997). This was done by comparing the metabolic needs (requirements) and dietary patterns (olestra-to-nutrient ratios) of the subgroups to those of the pigs and humans in the studies. If the nutritional needs and dietary pattern of a subgroup were less severe than those covered in the studies, it is reasonable to conclude that the effects of olestra on the nutritional status of the subgroup would be no larger or different than the effects measured in the studies.

Table 1 shows the kinds of information about the potential effects of olestra on the selected nutrients that were provided by the various studies.

Table 1. Information on the effects of olestra on the status of fat-soluble and water-soluble nutrients from pig and human studies1 [View Table]

On the basis of these estimates, olestra doses of 8, 20 and 32 g/d were chosen for the human studies, with the exception of the 16-wk study, which tested 18 g/d. The 8 g/d amount approximates the estimated 90th-percentile chronic intake of the 18- to 44-y-old subjects used in the studies and is about twice the average intake for this group, 3.7 g/d. The 20 g/d amount approximates the estimated 90th-percentile acute intake of that group, 21.7 g/d. The highest dose, 32 g/d, about 10 times the estimated average chronic intake or about four times the 90th-percentile intake of the test subjects, was chosen as an exaggerated dose.

The 18 g/d dose used in the 16-wk study was simply chosen as an exaggerated dose. It is about six times the estimated mean intake of the total population (male and females, all ages) of snack consumers, 3.1 g/d.

The lowest amount (% wt/wt) of olestra fed in the early pig studies was chosen to be comparable to the estimated 90th-percentile chronic human intake. To determine what that would be, the estimated potential olestra intakes for humans were calculated on the basis of grams per day, grams per kilogram diet per day, grams per kilogram body weight per day, and grams per 4184 kJ (1000 kcal) metabolizable energy per day; the resulting values were converted into percentage dry weight of pig diet. These calculations produced a range of values. It was decided that grams per kilosgram diet per day and grams per 4184 kJ metabolizable energy per day were the most relevant bases for extrapolating human intakes to intakes in pigs because they best account for intraluminal factors such as volume of chyme and the ratio of olestra to other dietary components in the chyme. These two methods of comparison produced about the same value, 1.1% of the diet (wt/wt), which was used. None of the other methods produced significantly greater values. The higher intakes were simply chosen as multiples of 1.1%.

It was found that pigs fed 1.1% olestra ate about 18-20 g/d, about three times the 90th-percentile human intake of 6.9 g/d. Comparison of the effects of olestra on common measures of fat-soluble vitamin status in humans and pigs, such as serum vitamin E, also suggested that the 1.1% dietary level provided an exaggerated olestra intake in relation to the 90th-percentile human consumption. Therefore, olestra dietary concentrations of 0.25 and 0.5% (wt/wt) were included in the later pig studies to more closely approximate expected human intakes. Also, in certain studies, some of the higher doses (>1.1%) were not fed.

Existing data indicated that olestra did not interfere with the absorption of water-soluble nutrients, as would be predicted on the basis of the partitioning mechanism. To provide further assurance that olestra does not interfere with the absorption or digestion of water-soluble nutrients by means other than the partitioning mechanism, the status of selected water-soluble nutrients was monitored in the pig and human studies. Folate and vitamin B₁₂ were selected as examples of a class of water-soluble nutrients that are digested and absorbed by complex multistep processes. By assessing the effect of olestra on the status of these nutrients, it can be determined if olestra can interfere with one or more of the steps involved in their uptake by a mechanism other than the partitioning mechanism.

Vitamin B₁₂ is released from polypeptide linkages in food by gastric acid and enzymes. The free molecule is bound to a protein in the stomach and subsequently released in the small intestine by the action of pancreatic trypsin. The free molecule then is bound to intrinsic factor. The vitamin B₁₂-intrinsic complex attaches itself to absorptive sites on the brush border of the ileum in the presence of ionic calcium (Herbert and Colman 1988). Folate is present in food primarily in the form of polypterolyglutamates, which must be cleaved by pteroylpolyglutamate hydrolases before absorption (Herbert and Colman 1988).

Calcium, zinc and iron were selected as examples of water-soluble nutrients that are limiting in the U.S. diet (NRC 1989). Because of this, any reduction in absorption of these micronutrients could be nutritionally significant. Also, for such nutrients, the amount of olestra in the diet would be large, on a mass scale, in relation to the amounts of these micronutrients, thus increasing the possibility that olestra might affect their absorption.

Although available data indicate that olestra does not interfere with the absorption and utilization of macronutrients, additional data on the macronutrients were collected. Proteins and carbohydrates are cleaved to smaller molecules before absorption; however, these cleavage processes occur in an aqueous environment and would not be expected to be subject to interference by olestra. In contrast, fat digestion and absorption involve lipolysis and solubilization of the resulting products into intestinal micelles; olestra might potentially interfere with these processes. To assess the potential of olestra to affect the digestion or absorption of macronutrients, growth parameters and feed intake were measured in the pig studies. Because each macronutrient provides a significant proportion of total energy intake and because data on growth and feed efficiency were collected over the pigs' most rapid growth period, and therefore the period of greatest macronutrient need, any meaningful effect of olestra on the absorption or utilization of any of the macronutrients would be evident. In the pig diets, fat was kept at 30% of energy to mimic the desired level in the human diet. In humans, the effect of olestra on the absorption of dietary fat was assessed by directly measuring triglyceride absorption in the absence and presence of olestra.

The human and pig studies were designed to induce and maximize any potential effects of olestra on the nutrients. Olestra was fed in exaggerated daily amounts and in a continuum of exaggerated dietary patterns, ranging from the situation in which pigs were fed up to 7.7% (wt/wt) olestra mixed directly in the diet at every eating occasion, to the 8-wk human studies in which olestra was eaten at every meal every day at doses as high as 32 g/d and no other foods allowed, to the situation in the 16-wk human study in which 18 g/d olestra was eaten every day but other foods could be eaten at other occasions. In the pig studies and in most of the human studies, dietary nutrients were maintained at minimal requirements. Because of these conditions, effects measured in these studies on the fat-soluble nutrients are probably exaggerated relative to effects that may occur in real life.

The findings from the various studies and assessments are described in detail in subsequent papers in this supplement. These findings define the effects of olestra on essential nutrients and other dietary components and, for those nutrients affected, show how the olestra effects will be compensated for or why the effects do not place consumers at meaningful nutritional risk.