Olestra (Olean, Procter & Gamble, Cincinnati, OH) is a mixture of hexa-, hepta- and octaesters of sucrose formed from long-chain fatty acids derived from edible oils. Olestra has physical properties and taste and cooking characteristics similar to those of dietary fats and oils (Bernhardt 1988, Jandacek and Webb 1978, Kester 1993). However, olestra is not hydrolyzed by gastric enzymes (Mattson and Volpenhein 1972) and is not absorbed intact from the gastrointestinal (GI)³ tract (Miller et al. 1995). Because of these unique properties, olestra can serve as a replacement for conventional fats and oils, contributing no calories or fat to the diet. Olestra is approved for use in replacing up to 100% of the fat used in the preparation of savory snacks (Federal Register 1996).

Olestra is lipophilic and has the potential to interfere with the absorption of lipophilic nutrients or other dietary components (Jandacek 1982). When olestra and a fat-soluble nutrient come in contact in the GI tract, a portion of the nutrient partitions into the olestra. This portion is then unavailable to the mixed intestinal micelles and is removed from the body with the nonabsorbed olestra. Key factors controlling this partitioning mechanism include the following: 1) the lipophilicity of the nutrient  the more fat-soluble it is, the greater the amount that will partition into the olestra; (2) the relative amounts of olestra and nutrient in the GI tract  the larger the amount of olestra relative to the amount of nutrient, the greater the amount of nutrient that will partition into the olestra; and (3) the time between consumption of olestra and nutrient  the olestra and nutrient must be in the GI tract simultaneously for partitioning to occur.

The potential effects of olestra on the availability of a number of essential water- and fat-soluble nutrients have been assessed in studies in humans and pigs under a variety of olestra-nutrient consumption patterns (Cooper et al. 1997a-c, Daher et al. 1997b and 1997c, Koonsvitsky et al. 1997, Schlagheck et al. 1997a and 1997b). These studies, described elsewhere in this issue, showed that olestra does not affect the absorption of water-soluble nutrients but has the potential to affect the absorption of fat-soluble nutrients. They also showed that the effects of olestra on the availability of fat-soluble vitamins can be offset by adding extra amounts of the vitamins to olestra foods.

In addition to vitamins and minerals, there are other components of the diet for which essential nutritional functions have not been established. However, these dietary components may provide beneficial health effects. For example, studies of associations between diet and health have shown that increased consumption of fruits and vegetables is consistently associated with reduced risk of certain chronic diseases such as cancer and heart disease (Block et al 1992, Committee on Diet and Health 1989, U.S. Department of Health and Human Services 1988, Willett 1994). The beneficial effects of diets high in fruits and vegetables have been hypothesized to come from fiber as well as phytochemicals. The phytochemicals fall into several broad classes such as the carotenoids, phytosterols, terpenoids, flavonoids, polyphenols and indoles, as well as others (Institute of Food Technologists 1993, Steinmetz and Potter 1991, Tanka 1994).

One of the purposes of the study described here was to assess the potential for olestra to affect the availability (i.e., absorption efficiency) of the major classes of phytochemicals. The phytochemicals considered included those for which the major dietary sources are fruits and vegetables and those that have been identified in the nutrition and medical literature as being potentially beneficial. To make the assessment, the lipophilicity of the phytochemicals was compared with the lipophilicity of molecules for which it has been shown that olestra does or does not affect availability. In making the comparison, octanol-water partition coefficients (log₁₀ p_c) were used as a measure of lipophilicity (Jandacek 1982).

It is possible to assess the potential for olestra to affect the availability of a molecule from knowledge of its lipophilicity because the mechanism by which olestra affects absorption of other substances is a physical interaction between olestra and the substances in the GI tract. In this interaction, the GI tract serves primarily as a mixing vessel. The physiological processes responsible for nutrient digestion and absorption are not affected by olestra (Cooper et al. 1997a-c, Daher et al. 1997b and 1997c, Koonsvitsky et al. 1997, Schlagheck et al. 1997a and 1997b); nor does olestra affect GI structure, function or physiology (Bergholz et al. 1991).

Among the phytochemicals, the carotenoids have been the most thoroughly investigated for potentially beneficial health effects. Epidemiological data have shown associations between high carotenoid intake, resulting from the consumption of fruits and vegetables, and lowered risk of diseases such as lung cancer (Doll and Peto 1981, Peto et al. 1981) and cardiovascular disease (Gaziano 1996). A plausible mechanism supporting the potential protective effects of carotenoids has been proposed, namely, that carotenoids, particularly -carotene, act as an antioxidants (Burton and Ingold 1984). However, several large-scale intervention studies failed to show beneficial effects for -carotene (Alpha-Tocopherol Beta-Carotene Cancer Prevention Group 1994, Greenberg et al. 1990 and 1994, Hennekens et al. 1996, Omenn et al. 1996).

A second purpose of this study was to estimate the effect of olestra on the absorption of the carotenoids when olestra snacks and carotenoid-containing foods are eaten in free-living dietary patterns. Effects of olestra on the availability of carotenoids have been observed (Koonsvitsky et al. 1997, Schlagheck et al. 1997a and 1997b, Weststrate and van het Hof 1995). However, the studies in which the effects were observed used olestra dietary patterns that were exaggerated relative to what would be expected when individuals eat snack foods made with olestra and carotenoid-containing foods in free-living patterns, in both the frequency of co-consumption of olestra and carotenoids and the daily amount of olestra consumed. Because of the exaggerated dietary conditions, the measured effects on carotenoid absorption were also exaggerated.

METHODS

The calculation of partition coefficients from molecular fragments was first proposed by Hansch and Leo (1979) and computerized by Chou and Jurs (1980). The KOWWIN program improves on early methods by calculating partition coefficients by summing the contributions from molecular fragments and then applying correction factors for substructure orientations. The contributions from fragments or atoms in the molecule were determined by multiple linear regression of values for simple molecules for which atom/fragment summation produces partition coefficients that are in good agreement with measured values. The correction factors (e.g., steric interactions, hydrogen bonding or effects from polar substructures) were derived from differences between measured partition coefficients and coefficients calculated by including only the fragment contributions for molecules with common substructures, again using multiple linear regression. Partition coefficients calculated by this method agree closely with measured values. An r² of 0.94 was obtained when the method was tested on a separate validation set of 6055 measured partition coefficients (Meylan and Howard 1995).

From the measured effects of olestra on the absorption of molecules ranging in lipophilicity from water soluble to highly lipid soluble and the partition coefficients (log₁₀ p_c) of those molecules, it was determined what log₁₀ p_c must be for the absorption of a molecule to be affected by olestra.

Phytochemicals included in the comparison were those for which the major dietary sources are fruits and vegetables and for which potential beneficial effects have been hypothesized or demonstrated experimentally. In addition, molecules that belong to the same class or have structures similar to those meeting these criteria were included. The phytochemicals were identified from publications in the nutrition and medical literature.

The above process was then repeated; for those eating occasions when a snack food and a carotenoid-containing food were eaten together, the availability of -carotene from the carotenoid-containing food was reduced to reflect the effect of olestra on carotenoid absorption. Again it was assumed that all snacks eaten contained olestra and 100% of the fat in the snacks had been replaced with olestra. The reduction in the amount of -carotene intake for these occasions was determined by using the following algorithm, developed from the results of the studies in which olestra and -carotene were always eaten together (Schlagheck et al. 1997a and 1997b): where Y(x ) = intake of -carotene (µg) when eaten with olestra, Y(0) = intake of -carotene (µg) when eaten without olestra, and X = amount (g) of olestra co-consumed with -carotene.

This algorithm allowed the olestra effect on -carotene absorption to be calculated at olestra intakes between and greater than those tested in the clinical studies (i.e., 2.7, 6.7 and 10.7 g/meal).

The intake of -carotene eaten at a time different from the time at which olestra was eaten was not reduced. Virtually all of the olestra eaten with a solid-liquid meal empties from the stomach within 2 h (Cortot et al. 1982). This means that carotenoids eaten ~2 h after olestra is eaten had little opportunity to mix with and interact with the olestra, at least in the proximal portion of the digestive tract where most nutrient absorption occurs. Data from pigs fed a diet in which vitamin A was provided as a 3:1 mix of retinyl palmitate and -carotene showed that olestra eaten by the pig in potato chips between feedings of the diet had no significant effect on liver vitamin A concentration (Daher et al. 1997a). In contrast, liver vitamin A concentration was reduced by 44% when olestra was co-consumed with the diet. Pigs and humans have similar GI anatomy, morphology and physiology, including ingesta transit times (Miller and Ullrey 1987).

RESULTS

Table 1. Octanol-water partition coefficients (log10 pc) for substances for which the effects of olestra on absorption have been measured in human or animal studies1,2,3 [View Table]

A log₁₀ p_c value > 0 means that the substance has a greater equilibrium concentration in the oil phase than in the water phase. Conversely, a log₁₀ p_c value < 0 means that the equilibrium concentration in the water phase is greater than that in the oil phase. Substances that have been tested range from highly lipophilic molecules (e.g., carotenoids) to weakly lipophilic molecules (e.g., ethinylestradiol), to water-soluble molecules (e.g., amino acids). Because the coefficients are expressed in log units, a difference in log₁₀ p_c of 1 unit indicates a 10-fold difference in lipophilicity.

The absorption of molecules with log₁₀ p_c values 7.6 was affected to some extent by olestra in humans. In the studies in which these effects were noted, olestra and the affected molecules were eaten together. Some of the same molecules were tested in animal studies and the same results were found. In the animal studies, olestra was mixed in the diet and, as in the human studies, always eaten at the same times as the affected substances. In addition, animal studies showed an effect of olestra on the absorption of pentachlorin (DDT), which has a log₁₀ p_c value of 6.9. Both human and animal data show that olestra had no effect on the absorption of molecules with log₁₀ p_c values < 6.9 even when eaten together.

Table 2. Ranges of octanol-water partition coefficients (log10 pc ) for the major classes of phytochemicals and the numbers of molecules within the ranges [View Table]

Estimated -carotene intake and frequency of co-consumption of foods containing carotenoids and olestra snacks.

Table 3. Effective -carotene intake among all snack eaters and among the top 10% of snack eaters with and without olestra in the diet and the reduction in the effective intake resulting from the co-consumption of olestra snack foods and carotenoid-containing foods [View Table]

The average effective -carotene intake (i.e., the mount that might be available to absorb) was estimated to be 2.0 mg/d for the total population of snack eaters (all ages, males and females) when olestra snacks were not included in the diet. This value ranged from ~1 mg/d for 2- to 5-y-old children to ~2.9 mg/d for adults > 64 y of age. The 5th-percentile intake of -carotene estimated for the total population of snack eaters was 0.434 mg/d. The average intake of -carotene estimated for heavy snack eaters (all ages, males and females) was 2.1 mg/d.

These data show that when olestra snacks are co-consumed with carotenoid-containing foods, the average availability of -carotene will be reduced by 5.9% for the total population of snack eaters (all ages, males and females). The calculated reduction ranged from 5.1 to 6.2% among the age groups. For individuals with low -carotene intake (5th-percentile), the reduction was calculated to be 6.0 % and for heavy snack eaters (95th-percentile) 9.6%.

In a separate analysis it was conservatively assumed that the effect of olestra eaten at occasions adjacent to those occasions when carotenoid-containing foods are eaten (e.g., mid-morning or mid-afternoon snacks) is 50% of the effect predicted from the algorithm developed from the human clinical studies. The average reduction in -carotene availability was calculated to be 8% for the total population of snack eaters and 14% for heavy snack eaters (data not shown).

The menu census survey showed that the average snack consumer eats snacks about five times in a 14-d period and eats -carotene-containing foods about 34 times. The 90th-percentile consumer eats snacks about 10 times in a 14-d period and carotenoid-containing foods about 46 times. Snacks and foods containing -carotene are eaten together only about four times in 14 d by the average snack consumer and about six times by the 90th-percentile snack consumer.

DISCUSSION

The absorption of only one molecule with a log₁₀p_c value < 7.6 has been shown to be affected by olestra; that occurred in an animal study in which olestra was fed in ways and amounts not relevant to manner in which olestra will be eaten by humans, in savory snacks. That molecule was pentachlorin (DDT), which has a log₁₀ p_c value of 6.9 (Volpenhein et al. 1980). The data on DDT absorption were obtained by dosing thoracic duct-cannulated rats by stomach tube with [¹⁴C]DDT in an emulsion diet (6 g) containing 22% sucrose, 20% nonfat milk solids, 2% salt mix, 36% water, 10% soybean oil, 10% olestra and 30 ppm [¹⁴C]DDT, or by dosing intact rats with [¹⁴C]DDT in the emulsion diet (minus the olestra) by stomach tube followed by feeding 10% (wt/wt) olestra mixed in a casein-based diet for 72 h. Absorption was determined from the recovery of ¹⁴C from lymph, adipose tissues and liver.

Feeding 10% olestra emulsified in a liquid diet is not representative of the manner in which olestra will be eaten in snack products, nor is mixing olestra at 10% (wt/wt) directly into the complete diet. It has been shown that the effect of olestra on the absorption of dietary constituents is exaggerated severalfold by mixing olestra in the diet relative to the situation in which olestra is eaten at the same time, but in a separate food such as potato chips (Daher et al. 1997a).

No absorption data on a molecule having exactly the same lipophilicity as DDT (log₁₀ p_c = 6.9) have been collected in humans. However, the available human data strongly support the conclusion that the absorption of molecules with log₁₀ p_c values less than ~7.5 will not be reduced in any meaningful way by olestra, even when olestra is consistently eaten at every meal. The reasoning behind this conclusion follows.

The effect of olestra on the absorption of nutrients such as retinol and oleic acid, which have log₁₀ p_c values of 7.6 and 7.7, respectively, was minimal when olestra and the nutrients were eaten at the same meal (Daher et al. 1997b and 1997c). No effect on the absorption of either nutrient was observed with either 8 or 20 g olestra; 32 g of olestra reduced the absorption of oleic acid by only 1.2% and had no significant effect on the absorption of retinol, although the area under the ¹⁴C-retinyl esters plasma concentration-time curve was decreased by 13-19%.

Information on the potential for olestra to affect the absorption of moderately lipophilic nutrients, those with log₁₀ p_c values in the 9-12 range, comes from the 16-wk study in free-living subjects reported elsewhere in this issue (Koonsvitsky et al. 1997). In that study, the subjects were requested to eat 18 g/d olestra (in foods prepared with olestra) with their meals, but were not required to evenly divide the daily dose among the three meals. Further, and importantly, they were not restricted from eating other foods at other times throughout the day. In that study, the plasma concentration of cholesterol (log₁₀ p_c = 8.7) was not affected and the serum concentration of tocopherol (log₁₀ p_c = 12.2) was reduced by only 6%. Cholesterol is at least 10 times more lipophilic than retinol and oleic acid; tocopherol is more than 10,000 times more lipophilic than those nutrients.

It is the frequency at which olestra and a nutrient are eaten together over time that is the primary determinant of olestras effect on the absorption of the nutrient. Because olestra interferes with absorption through a physical interaction, interference can occur only when the interaction is allowed to take place (i.e., when olestra and the nutrients are in the GI tract at the same time). How the frequency of co-consumption influences the potential of olestra to alter the absorption of lipophilic nutrients is illustrated by the following findings. When subjects ate 20 g/d olestra in snack foods evenly divided among the three daily meals every day for 56 d and were not allowed to eat any other foods between meals, serum -carotene concentration was reduced by about 62% and serum vitamin E concentration was reduced by about 18% (Schlagheck et al. 1997b). When subjects ate 18 g/d olestra at meals, not necessarily evenly divided among the three daily meals, and were permitted to eat any other foods they desired between meals, in the 16-wk study discussed above, -carotene absorption was reduced by about 27% and vitamin E absorption by 6%, two- to threefold less than the effects measured when olestra and the nutrients were always eaten together. Even the dietary pattern used in the 16-wk study was exaggerated relative to the free-living pattern in which snack foods prepared with olestra are expected to be eaten. It is estimated from current snack use patterns that snacks prepared with olestra will be eaten only about five times in a 14-d period by the average snack eaters, with about 8% of these occurring with meals (Webb et al. 1997). From the same population data, it is also estimated that the intake of olestra from snack consumption by the average snack eater will be 3.1 g/d and that by the 90th-percentile eater will be 6.9 g/d. Because of this dietary pattern, the effects of olestra on the absorption of nutrients with log₁₀ p_c values between ~7.5 and 12 are unlikely to be measurable in individuals eating olestra snack foods in real life. In addition, the real-life effects of olestra on the absorption of more lipophilic molecules (e. g., those with log₁₀ p_c values > 12) would be expected to be considerably less that the effects measured in the clinical studies discussed above as well as other studies from which the correlation between lipophilicity and absorption was developed.

Phytochemicals in fruits and vegetables may be responsible for the inverse association between diets high in fruits and vegetables and the risk of certain chronic diseases such as cancer and heart disease. From the assessment of the lipophilicity of the major classes of phytochemicals, it can be concluded with some certainty that olestra will not affect the availability of phytochemicals with the exception of the carotenoids and the phytosterols because the vast majority of phytochemicals have log₁₀ p_c values < 7.6 (Table 2 and Appendix A).

Phytosterols have log₁₀ p_c values ranging from 7.9 to 12.3. On the basis of the effects of olestra on the absorption of nutrients that have log₁₀ p_c values in this range, such as retinol, oleic acid, cholesterol and vitamin E, it would be expected that the absorption of the phytosterols would be reduced by <10% if olestra was eaten consistently with every meal. The effect would be expected to be even less when olestra snack foods are eaten under free-living dietary patterns.

A small reduction in absorption of the phytosterols is unlikely to be of concern because the hypothesized beneficial effect of these molecules is to reduce cholesterol absorption when consumed in large amounts (Linscheer and Vergrosen 1988). This is also a potential benefit of olestra (Jandacek et al. 1990). Absorption of large amounts of phytosterols themselves is undesirable inasmuch as it may result in increased risk of hypercholesterolemia (Linscheer and Vergrosen 1988).

The carotenoids are the most lipophilic class of phytochemicals. With a few exceptions, carotenoids have log₁₀ p_c values that range from ~11 to 18. As discussed above, the availability of carotenoids is affected by olestra when the two are eaten together (Koonsvitsky et al. 1997, Schlagheck et al. 1997a and 1997b, Weststrate and van het Hof 1995). However, from the assessment of eating patterns of snack foods and foods containing -carotene, the reduction in the availability of -carotene from eating snack foods prepared with olestra is estimated to be 6-10%. These estimates are still exaggerated relative to the real-life situation because it was assumed in making the calculations that all snacks eaten by the individuals were snacks containing olestra. For this to be true, olestra snacks would have to comprise 100% of the savory snack market. Any decline in the rate of building body stores of vitamin A resulting from the effect of olestra on the availability of the pro-vitamin A carotenoids can and will be offset by adding extra amounts of vitamin A to snack foods prepared with olestra (Cooper et al. 1997a, Federal Register 1996).

The reduction of carotenoid absorption from a meal by olestra snacks eaten with that meal is akin to other "meal effects" on carotenoid absorption. For example, -carotene absorption from a high fiber meal is reduced by ~50% (Rock and Swendseid 1992), and -carotene absorption is reduced from meals containing only small amounts of fat by >70% (Dimitrov et al. 1988). Because such dietary interactions are sporadic unless dietary patterns are such that these interactions occur at a large fraction of total meals, they have no significant nutritional effects over time. Such would be the effect of olestra snacks on carotenoid availability.

The MRCA methodology used to estimate the average -carotene intake among snack eaters when olestra snacks are not eaten produced a value of 2.0 mg/d, virtually the same as the value of 2.1 mg/d that can be calculated for the total U.S. population from the data in the 1987-88 Nationwide Food Consumption Survey (U.S. Department of Commerce 1988), assuming that 90% of the 383 µg retinol equivalents of carotene intake comes from -carotene. The close agreement between the two values indicates that the MRCA survey provides a reasonable estimate of -carotene eating occasions and provides an appropriate base for estimating the potential olestra effect on -carotene availability under realistic dietary patterns.

Fruits and vegetables contain other hydrocarbon carotenoids such as lycopene and -carotene in addition to -carotene. The effect of olestra on the availability of these carotenoids is likely to be similar to the effect calculated for -carotene because they have the same log₁₀ p_c values (17.6) as -carotene. Furthermore, most clinical studies have shown that olestra affects the absorption of -carotene and lycopene to the same degree that it affects the absorption of -carotene (Koonsvitsky et al. 1997, Schlagheck et al. 1997a and 1997b). In one study (Weststrate and van het Hof 1995), sucrose polyester affected the absorption of lycopene to a somewhat greater extent than -carotene. Possibly this was because olestra was eaten only at the evening meal in that study and the frequency at which lycopene was eaten at that meal, in that study, may have been greater than the frequency at which -carotene was eaten at the same meal.

Lutein and zeaxanthin are other carotenoids widely distributed among fruits and vegetables (Mangels et al. 1993). These carotenoids and -cryptoxanthin (known as xanthophylls) are less lipophilic than -carotene by about a factor of 100 and therefore would be expected to be less affected than -carotene by olestra. That has been demonstrated for lutein and zeaxanthin (Schlagheck et al. 1997a and 1997b, Weststrate and van het Hof 1995).

Fruits and vegetables contain many chemical constituents. No single constituent has been shown to be responsible for the observed inverse correlation between fruit and vegetable intake and risk of certain chronic diseases. Olestra will not affect the absorption of the vast majority of the constituents of fruits and vegetables, including the phytochemicals, the essential water-soluble nutrients that have been hypothesized to protect against chronic diseases, such as vitamin C and folate, or fiber. The potential effects of olestra on the availability of carotenoids from fruits and vegetables, when eaten in the context of the typical diet, will be limited. Therefore it is anticipated that the consumption of olestra snacks will not decrease the beneficial effects of diets high in fruits and vegetables.

ACKNOWLEDGMENTS