Previous - Next
Chapter 10 : Isomeric fatty acids
Occurrence
in diets
Metabolic effects in animals
Human
studies
Pregnancy
and lactation
Conclusions
Under conditions of partial hydrogenation, a double bond may change from a cis to a trans configuration (geometric isomerization) or move to other positions in the carbon chain (positional isomerization). Both types of isomerization frequently occur in any one fatty acid undergoing hydrogenation.
Trans fatty acids are unsaturated fatty acids with at least one double bond in the trans configuration. The most common trans fatty acids are monounsaturated, but various diunsaturated cis, trans and trans, cis isomers also occur. In trans monounsaturated fatty acids of partially hydrogenated oil, the double bond tends to be normally distributed around positions 9 to 11 with a range of 5 to 15 (Dutton, 1979; Marchand, 1982).
The most common sources of isomeric fatty acids are margarines and shortenings containing partially hydrogenated vegetable or fish oil. Dairy products and meat from ruminant animals derive their isomeric fatty acids from the hydrogenation process in the rumen where bacteria carry out anaerobic fermentation. In milk fat, a trans double bond may occupy positions 6 to 16, with a preference for position 11. The trans fatty acids comprise approximately 5 percent of the total fatty acids in products from cows and sheep, whereas those in commercially hydrogenated fats can range to more than 50 percent (Gurr, 1990).
Whenever partially hydrogenated fats are consumed, both cis and trans isomers are involved, however, the presence of trans fatty acids is used to detect hydrogenated fat. Hydrogenation of polyunsaturated fatty acids enhances the assortment of isomers. Although the levels of cis, trans and trans, cis isomers of linoleic acid in stick margarines comprise as much as 7.6 percent of the fatty acids, and fatty acids with two trans double bonds as much as 2.8 percent, most are under 1 percent (Ratnayake, Hollywood and O'Grady, 1991). In partially hydrogenated canola oil, the major polyunsaturated isomer was found to be cis-9, trans-13 octadecadienoic acid (Ratnayake and Pelletier, 1992). The cis bond at position 9 prevailed in most isomers of linoleic acid. The occurrence of isomers of polyunsaturated fatty acids can be controlled in the processing of oil. Good oil processing practices provide oils with low levels of cis/trans and trans/trans isomers.
Estimates of trans fatty acid intakes have been controversial. The average intake for the population in the USA, based on disappearance data, was given as 7.6 g/day and later revised to 8.1 g/day (Hunter and Applewhite, 1986; 1991). Other calculations suggested an average intake of 13.3 g/day; this was based on such assumptions as food service and industrial shortenings having 40 percent trans fatty acids and salad and cooking oils having 25 percent (Enig et al., 1990, Steinhart and Pfalzgraf 1992, 1994). In West Germany, the intake of trans-octadecenoic acid was estimated to be 4.5 to 6.4 g/caput/day with 35-45 percent from ruminant products (Heckers et al., 1979). For the British population, the average intake was given as 7 g/day with a possible range of 5 to 27 g/day, depending on the selection of foods (British Nutrition Foundation, 1987). In India, the average intake of hydrogenated fat having 55 percent trans fatty acids was 2.04 g/day (National Council for Applied Economic Research, 1991). An adult living in the Indian states with highest levels of consumption could have an intake of approximately 11 grams or 4 percent of energy per day of trans fatty acids. Apparent average levels of intake may not be as relevant to health as higher levels and estimates of the 90th percentile of intake of trans fatty acids in each geographica1 area would be useful.
The consumption of isomeric fatty acids is expected to decrease further as soft margarines and low-fat spreads continue to replace the stick or print-type of product. The replacement of lard or tallow by partially hydrogenated oil, however, would have the opposite effect. Care must be taken when fats are substituted so that a reduction of isomers in one type of food does not lead to an increase in another food.
An examination of the early literature showed that many results which originally had been attributed to trans fatty acids were in fact confounded by essential fatty acid deficiency (Beare-Rogers, 1983; Gurr, 1983). Isomeric monounsaturated acids, like saturated fatty acids, tend to occupy the 1-position and polyunsaturated acids the 2-position of animal phosphoglycerides. Studies of the general pattern of fatty acids in membrane constituents imply that, with adequate polyunsaturated acids, the transmonounsaturated acids do not accumulate in the 2-position of phosphoglycerides nor do they influence eicosanoid production. A level of 2 en % of linoleic acid in the diet was sufficient to prevent an effect on eicosanoid synthesis (Zevenbergen and Haddeman, 1989). When a high dose of transfatty acids was fed to rats, an assessment of the level of linoleic acid needed to prevent metabolic changes gave the figure of 5 percent of fat or 2 en % (Verschuren and Zevenbergen, 1990). An adequate level of essential fatty acids is, therefore, critical to prevent specific effects of isomeric fatty acids.
Trans polyunsaturated fatty acids. In studies with deuterated isotopes, the trans-9, cis-12-isomers which accumulated in mouse liver were 2 to 4 times more than the cis-9, trans-12-isomers (Beyers and Emken, 1991). The action of the trans, cis-isomer in increasing 9 desaturation was similar to that observed in animals which were deficient in essential fatty acids. The position of the trans double bond appears to influence a variety of enzyme activities.
The trans-9, trans-12-octadecadienoic acid has been extensively studied in animals (Kinsella et al., 1981). This isomer has the greatest potential for interfering at the enzyme level with the metabolism of essential fatty acids and eicosanoids. The production of this isomer is usually kept to a negligible amount and it is expected to remain so.
Expenmental evidence. The deposition of isomeric fatty acids in human tissue has been of considerable interest. In the United Kingdom, the trans fatty acid content of adipose tissue was reported to be 5 percent (Thomas et al., 1981). Analysis of tissues during routine autopsies in the USA showed that the levels of trans-octadecenoic acid varied from 0.4 to 5 percent of the total fatty acids, and reflected the distribution of isomers in the dietary fat (Ohlrogge, Emken and Gulley, 1981). The highest level of trans isomers occurred in adipose, cardiac and aortic tissues, with triacylglycerol having similar isomeric patterns to the dietary oils, and phosphatidylcholine showing a selective incorporation, albeit less dramatic, than that found in animal and in vitro experiments (Ohlrogge, Gulley and Emken, 1982). There was a tendency for isomers having a double bond near the terminal methyl group to be retained without an accumulation of any one isomer.
Studies have demonstrated that adult males showed no discrimination in absorption or oxidation of various octadecenoic isomers labelled with deuterium and fed as triacylglycerol (Emken et al., 1979, 1980 a, b; Emken, 1984). After a single meal containing these triacylglycerols, the incorporation of the different isomers was measured in various plasma lipids. The 12-cis isomer was incorporated into the 2-position of phosphatidylcholine and could consequently, replace arachidonic acid.
The most prevalent trans-polyunsaturated fatty acid in human adipose tissue appeared to be cis-9, trans-13 octadecadienoic acid (Hudgins, Hirsch and Emken, 1991; Ratnayake and Pelletier, 1992). Its effect on the metabolism of essential fatty acids and eicosanoids is unknown.
Cholesterol. For more than three decades, various results came from attempts to relate the trans fatty acids to changes in serum or plasma total cholesterol. These have been reviewed by Hunter (1992) and Wood (1992). Two studies with liquid formula diets published in the same year came to opposite conclusions. When white, Trappist monks and nuns in the Netherlands were provided with a diet containing trans monounsaturated fatty acids, serum cholesterol levels were elevated in the presence of dietary cholesterol but not in its absence. Serum cholesterol levels at the end of 4 weeks of experimental feeding were highest in the group receiving the saturated fatty acids (lauric and myristic), followed by the group receiving fat with 38 percent trans fatty acids, and then the group receiving oleic acid. Compared to the values at the beginning of the experimental feeding period, cholesterol values of the group receiving oleic acid fell and those of the group receiving the saturates rose slightly (Vergroesen, 1972; Vergroesen and Gottenbos, 1975).
The other major study with liquid formula diets containing cholesterol involved male prison inmates in the USA who were predominantly black and from cities. The dietary fat containing 34 percent trans monounsaturated fatty acids did not produce different results than the control fat which was high in oleic acid (Mattson, Hollenbach and Kligman, 1975). A later study of college students indicated that lightly hydrogenated soy oil was less effective than unhydrogenated soy oil in lowering total and low density lipoprotein cholesterol (Laine et al., 1982).
Lipoproteins. The first measurements on lipoproteins in a study of trans fatty acids were reported in 1990 (Mensink and Katan, 1990). Normocholesterolemic men and women consumed a diet of mixed foods: one group received a diet high in oleic acid, another had a diet high in trans monounsaturated fatty acids, while the third consumed a diet high in saturated fatty acids to provide 10 percent of total energy. The dietary fatty acids were balanced to provide the same amount of stearic acid in each group while the total saturated and polyunsaturated fatty acids (4.6 en %) were similar for the diets high in oleic or trans fatty acids. Compared to the oleic diet, the trans diet increased LDL-cholesterol and decreased HDL-cholesterol, while the saturated diet increased LDL and produced no change in HDL-cholesterol after 3 weeks. Both the trans and the saturated diet increased triacylglycerol. The results on the apoproteins paralleled the LDL and HDL results. The ratio of LDL/HDL was higher for the trans diet than for the saturated diet. The data indicated that trans fatty acids had no advantage over saturated fatty acids.
Although the fat used in this work was catalytically isomerized rather than commercially hydrogenated, it did contain the trans isomers that are common in "Western" diets. The greatest difference between the test fat and the dietary fat of the USA appeared to be a positional cis isomer with a double bond in the 8 position (Mensink and Katan, 1991).
A second experiment from the same laboratory had a lower level of trans fatty acids, 7.7 percent, instead of 10 or 11 percent of energy (Zock and Katan, 1992). The trans diet was compared to one rich in linoleic acid and one rich in stearic acid, a fatty acid that is not known to raise blood cholesterol levels. Relative to linoleic acid, the isomeric fatty acids increased LDL-cholesterol and decreased HDL-cholesterol.
A study in the USA compared saturated fatty acids, oleic acid and two levels (3.8 and 6.6 en %) of trans isomers (Judd et al., 1994). All diets contained at least 10 en % of the saturated fatty acids, lauric, myristic and palmitic acids and 3 en % stearic acid. Linoleic acid was maintained at 6 en % of the diet. The tightly controlled experiment demonstrated differences in the effects of cis and trans monounsaturated fatty acids, even when they constituted less than 4 en %. After 6 weeks on the test diets, the "moderate trans", "high trans" and saturated diets raised cholesterol levels. In agreement with the results of Mensink and Katan (1990), a diet containing trans fatty acids raised LDL-cholesterol and apoprotein b and depressed HDL-cholesterol and to a lesser extent apoprotein A- 1, compared to the oleic diet. Statistically significant changes at a lower level of intake were observed since the saturated diet increased both the LDL-cholesterol and the HDL-cholesterol. The ratio of LDL-to HDL-cholesterol was less favourable with the 6.6 en % trans diet than with the saturated diet. The diet containing 6.6 en % of trans fatty acids, compared to the other diets studied, also elevated blood triacylglycerol. Although the overall changes were small, the trans fatty acids and the saturated fatty acids had somewhat similar effects.
The LDL/HDL ratios obtained in the experiments of Mensink and Katan (1990 -Exp. 1), Zock and Katan (1992 - Exp. 2), and Judd et al. (1994) are indicative of a dose response to trans-fatty acids (Figure 10.1).
FIGURE 10.1 : Influence of trans fatty acids on LDL and HDL cholesterol levels
A study on mildly hypercholesterolemic males showed that an increase of 4 en % in trans monounsaturated acids did not negate the LDL-lowering potential of vegetable oils when the concentration of linoleic acid was approximately 3 times greater than that of palmitic acid (Nester et al., 1992). Considering most known dietary patterns, such a distribution of fatty acids is highly unlikely.
Other studies involved test fats containing trans fatty acids in the diet of free-living normocholesterolemic men (Wood et al., 1993 a, b). A soft, zero-trans margarine which was high in linoleic acid significantly reduced total cholesterol, LDL-cholesterol and apolipoprotein b relative to other test diets. There was no change in HDL-cholesterol with the diet of 1:1:2 in saturates, monounsaturates and polyunsaturates. The butter diet elevated total serum cholesterol and LDL-cholesterol, while the hard margarine produced similar results to butter-sunflower oil and butter-olive oil blends. The difference in response to the hard and soft margarines was attributed largely to trans fatty acids and supported the findings of Mensink and Katan (1990).
Margarine containing hydrogenated corn oil which was used to replace unhydrogenated corn oil in a diet designed to reduce plasma cholesterol, led to less reduction in total cholesterol, LDL-cholesterol and apoprotein B, but to no difference in HDL-cholesterol nor Lp(a) (Lichtenstein et al., 1993). The diets were already low in saturated fatty acids and cholesterol.
In laboratories in Australia and the Netherlands, lipoprotein (a) [Lp(a)] which is regarded as an independent risk factor for cardiovascular disease, increased in individuals receiving trans fatty acids, however, this was not found in the USA. An "elaidic-rich" diet, which might have contained trans isomers other than the 9-trans isomer, was fed at a trans intake of 7 percent of energy as a substitute for oleic acid (Nester et al., 1992). The resulting LDL levels were similar to those produced by saturated fatty acids and were higher than those obtained with oleic acid. Only the diet containing trans isomers led to significant elevations of Lp(a). Similarly, an examination of samples from previous studies on dietary fat and lipoprotein demonstrated that isomeric fatty acids were associated with increased Lp(a) (Mensink et al., 1992). Oleic acid had the effect of decreasing Lp(a) and LDLcholesterol. It appeared unlikely that dietary effects on LDL-cholesterol and Lp(a) were mediated by the same pathway.
Different protocols were used in the various studies to determine the effect of dietary trans fatty acids on plasma lipoproteins. Whether subjects received foods for a complete diet or selected their non-test foods may have influenced the precision attained in an experiment. The part of the diet that was not controlled may have great variations in fatty acid composition, particularly in the content of fatty acid isomers. It may be noteworthy that the studies showing definite changes in both LDL- and HDL-cholesterol had foods prepared for the subjects.
To date, only those trans fatty acids found in partially hydrogenated vegetable oils were studied in controlled human investigations. It is not known what effect partially hydrogenated fish oils with C20 and C22 fatty acids have on human lipoprotein profiles.
Epidemiologic evidence. Epidemiological studies have included prospective, case-control and crosssectional designs. Persons dying of ischaemic heart disease in the UK had a lower concentration of shorter chain fatty acids and a higher concentration of trans fatty acids in adipose tissue fat, and were therefore judged to have consumed less ruminant animal fat and more commercially hydrogenated fat (Thomas and Scott, 1981; Thomas, Winter and Scott, 1983). An attempt to estimate trans fatty acid intake in adult men from semi-quantitative food frequency questionnaires indicated that the 10th and 90th percentiles of intake were 2.1 and 4.9 g/day, respectively (Troisi, Willett and Weiss, 1992). These low estimates, compared with the average of 8.1 g/day (Hunter and Applewhite, 1991), showed a weak positive correlation with LDL-cholesterol and an inverse correlation with HDLcholesterol. In the Nurses' Health Study, the quintiles varied from 1.3 to 3.2 percent of energy or 2.4 to 5.7 g/day and were below the expected average intake (Willett et al., 1993). Such foods as margarine, cookies, biscuits, cake and white bread were used for the calculation of trans fatty acids in hydrogenated vegetable fat. The relative risk of fatal and non-fatal myocardial infarction for women in the highest category of trans fatty acid intake was 1.6 compared to those with the lowest intakes. In a cross-sectional study of patients undergoing coronary angiography (Siguel and Lerman, 1993), somewhat higher values for plasma trans fatty acids were observed in patients than in reference subjects (1.38 percent versus 1.11 percent). Other data used to support the hypothesis that the intake of partially hydrogenated vegetable oils may contribute to the risk of cardiovascular disease came from a case-control study where the relative risk for the increasing quintiles of trans fatty acid intake were 1, 0.74, 0.43, 0.63 and 1.94 (Ascherio et al., 1994). It is interesting that the third quintile represented less than half the risk of the first quintile and only the highest quintile was associated with increased risk.
One of the great handicaps in such epidemiological studies is the measurement of exposure to isomeric fatty acids. Even within one food group, there is great variation in the concentration of trans fatty acids (Ratnayake et al., 1993).
Depending on the diet of mothers, human milk may contain various amounts of trans fatty acids. Koletzko (1991) detected the transfer of trans fatty acids across the human placenta and reported an inverse correlation between trans fatty acid exposure and birth weight in premature infants. Further study indicated that the trans fatty acids in the plasma lipids of infants were negatively correlated to arachidonic and docosahexaenoic acids (Koletzko, 1992). More studies are required in this area.
If there is a possibility of a deficiency of essential fatty acids during pregnancy and lactation, as observed by Holman, Johnson and Ogburn (1991), the level of intake of trans fatty acids could be of concern. On the basis of animal experiments, the intake of linoleic acid is of critical importance when diets contain partially hydrogenated oils.
In controlled metabolic studies using oleic acid as a comparison, trans fatty acids from partially hydrogenated vegetable fats raise plasma LDL cholesterol in a manner which is similar to that observed with saturated fatty acids. In contrast to saturated fatty acids, however, trans fatty acids do not increase plasma HDL cholesterol and may lower this lipid fraction compared to oleic acid. Thus, the ratio of total cholesterol to HDL appears to be more unfavorable for trans fatty acids compared with equivalent amounts of either oleic acid or saturated fatty acids. In two studies, trans fatty acids increased levels of lipoprotein (a), another reputed factor for coronary heart disease.
A number of conclusions can be drawn. First, isomeric fatty acids in partially hydrogenated vegetable oils appear to be hypercholesterolemic, however, the interpretation of epidemiological studies is uncertain. Second, where the intake of saturated fatty acids is curtailed, it seems appropriate to curb trans fatty acids to improve the profile of plasma lipoproteins. Consumers should be encouraged to substitute liquid oils, soft margarines and spreads for hard fats, wherever feasible. Third, where there is a possibility of a deficiency of essential fatty acids during pregnancy and lactation, a high level of intake of trans fatty acids should be avoided. Fourth, food processors are encouraged to reduce the isomerization of fatty acids as much as possible. It is unacceptable to use marketing claims such as "low in saturates" when the product is high in trans (unsaturated) isomers. Fifth, governments should monitor the intake of fatty acid isomers and regulate the claims being made with respect to products containing them.
