The possible relationship between consumption of trans fatty acids (TFAs) and risk of insulin resistance or development of diabetes mellitus type II has been considered by a number of human and animal studies over the past decade. This review evaluates the evidence, and concludes that there is limited evidence for a weak association at high TFA intakes, but very little convincing evidence that habitual exposure as part of a standard western diet has a significant contribution to risk of diabetes or insulin resistance. The possibility of increased risk for individuals with particular genotypes (such as the FABP2 Thr54 allele) is of interest, but further work would be required to provide sufficient evidence of any association.
The global prevalence of diabetes mellitus type II is on the rise. The World Health Organization estimated there were 171 million cases in 2000, and their projections suggest there may be 366 million cases by 2030 (Wild et al., 2004). It is well established that dietary and lifestyle factors have an important part in the development of this disease (Steyn et al., 2004), and it seems likely that the consumption of high levels of heart disease than saturated fatty acid (SFA) has a deleterious effect on glucose and insulin metabolism, whereas long-chain polyunsaturated fatty acid consumption appears to offer a beneficial effect (Hu, 2001, Steyn et al., 2004).
Consumption of trans fatty acids (TFA) have a more significant effect on risk of coronary heart disease than SFA intake (Mozaffarian and Clarke, 2009), but the evidence for associations between TFA and other health issues, and mechanisms which might underlie such associations, have not been fully elucidated. In late 2007, the Scientific Advisory Committee on Nutrition (SACN) of UK produced an updated position statement on the health impacts of TFA (SACN, 2007). In this document, they concluded that ‘the evidence relating TFA intakes to risk of diseases other than coronary heart disease is limited, and no reliable risk assessments can be made. Future reports on these associations should be monitored, particularly the effect of TFA on insulin sensitivity and diabetes’.
This review is produced by the researchers who undertook the SACN review and provides an updated assessment of the relationship between TFA, insulin resistance and diabetes risk.
Trans fatty acids
TFA are a particular class of mono-unsaturated fatty acid (MUFA), in which the double bond is in a different configuration (trans) than the configuration that is present in most naturally occurring MUFA (cis). There are a small percentage of TFA that are naturally found in dairy products and the meat of ruminants, which are formed by microbial action in the rumen. The other source of TFA is from the partial hydrogenation of oils containing unsaturated fatty acids (FAs), usually with the intent of producing a product that is more solid or stable at room temperature than the original oil. In milk fat or ruminant meat, the predominant (30–60%) TFA is vaccenic acid (also referred to as trans 18:1, n-7; trans 18:1, ω-11; (E)-11-octadecenoic acid or (E)-octadec-11-enoic acid), whereas in partially hydrogenated vegetable oils, the main isomer (30%) is elaidic acid (trans 18:1, n-9; trans 18:1, ω-9; (E)-9-octadecenoic acid or (E)-octadec-9-enoic acid; European Food Safety Authority, 2004). The TFA isomers from either source are chemically identical, and the considerable overlap in isomers common to both food sources (European Food Safety Authority, 2004) prevents current analytical techniques from being able to reliably distinguish between ruminant and industrial TFA simply on the basis of vaccenic versus elaidic acid composition. It has been suggested that the TFA from different sources may not have the same biological effects, with some evidence that animal sources of TFA do not share the adverse effects of isomers of vegetable oil origin. However, relatively few studies have specifically compared the effects either of TFA from animal sources with those of vegetable oil origin, or of isolated vaccenic acid with isolated elaidic acid. Where such comparisons can be made using the results of studies discussed in this review, they are highlighted in the text.
To evaluate the relevance of evidence provided by the studies covered in this review, it is useful to consider the mean population TFA intakes of various populations. The estimated mean TFA intake in the UK is 1% of energy (E) (1.7–2.4 g/day; SACN, 2007), in western Europe it is 0.9–1.0% of E (2.0–2.4 g/day; van de Vijver et al., 2000), and in Australia and New Zealand it is 0.5–0.6% of E (1.2–1.6 g/day; Food Standards Australia New Zealand, 2009). Canadian intake has been estimated as 2.2% of E (4.9 g/day; Health Canada, 2006), with the USA intake ∼1.8–2.2 g/day (Lemaitre et al., 1998).
Relevant papers were identified through keyword searches in the Medline and Pubmed database, supplemented by a careful inspection of citations in the various reviews and research papers. The SACN framework for the evaluation of evidence (SACN, 2002) was used as the basis to assess the strength of the scientific evidence. Prospective cohort, case–control, cross cultural studies and randomized controlled trials (RCTs) reported between 1997 and 2009 were reviewed, along with appropriate cell and animal studies published between 1995 and 2009.
Cell and animal studies: design and methodology
Most of the cell studies have investigated the effects of varying levels of exposure to TFA on glucose uptake in insulin sensitive targets cells (adipose tissue) or on the secretion of insulin from pancreatic islet cells. Animal studies have investigated a wide range of TFA exposure levels on in vivo, and in vitro/ex vivo, measures of insulin secretion and/or responsiveness. In some cases, the interpretation of the data is limited by the fact that the studies have failed to adequately control the levels of intake of other dietary FAs, which might also influence the processes under investigation. However, when there is agreement between human and animal or cell data, it can provide useful confirmation of an effect and potential elucidation of mechanisms.
The data from the cell and animal studies has been summarized in Table 1, to allow an overview of the different studies and their findings.
Rat adipocytes incubated for 2 h in media containing purified 18:1 FA isomers showed that both trans 18:1, n-9 and trans 18:1, n-7 FA reduced the amount of glucose converted to cell lipid (P<0.01) and inhibited the oxidation of glucose to carbon dioxide (P<0.05) compared with cis 18:1, n-9 (Cromer et al., 1995). In a study which compared the effects of acute and chronic exposure to FA, acute exposure of isolated mouse pancreatic islets to different cis and TFA resulted in an increase in glucose-stimulated insulin secretion (GSIS) for all FA, although the trans isomers elicited a higher level of insulin secretion than their cis counterparts (P<0.05; Alstrup et al., 1999). The rate of glucose oxidation at high glucose concentrations was suppressed by cis FA (P<0.05) but not affected by the trans isomers. The authors also considered the effect of long-term exposure of the islet cells to a range of 18:1 isomers (Alstrup et al., 2004). Basal insulin release was higher in cells exposed to the cis 18:1, n-7 (P<0.05), but there was no difference between cis and trans 18:1, n-9. GSIS was not altered by chronic exposure to either cis or trans 18:1, n-7 or by cis 18:1, n-9, but was stimulated by 0.3–0.4 mmol/l trans 18:1, n-9. In contrast to the observations at the shorter exposure times, the cells incubated with the FA isomers for 3 days showed no differences in glucose oxidation, but FA oxidation was higher in the presence of the trans isomers (P<0.05).
Animal studies: rodent models
Insulin action and adipocyte plasma membrane FA composition and fluidity were compared in Winstar/NIN rats fed diets containing different FA (0–3 % TFA; Ibrahim et al., 2005). Statistically significant increases in levels of insulin secretion and reduced membrane fluidity were observed in the groups receiving dietary TFA (P<0.05). In further studies, the mRNA expression of resistin was upregulated, and peroxisome-proliferative-activated receptor-γ and lipoprotein lipase were downregulated in adipocytes of the animals receiving the diets containing 3% TFA (Saravanan et al., 2005b). It was also reported that the TFA diet caused an increase in the mRNA levels of resistin and downregulated the production of peroxisome-proliferative-activated receptor-γ, but had no effect on adoponectin or GLUT4 (Saravanan et al., 2005a). However, because of the variations in the levels of other FA (SFA, MUFA, polyunsaturated fatty acid) in the diets, it is not possible to conclude that these differences were specifically because of the presence of TFA.
The addition of 1.5% (wt/wt) vaccenic acid to the diets of lean and obese JCR:LA-cp rats for 3 weeks had no effect on fasting glucose or insulin for either group (Wang et al., 2009). A meal test also showed no differences in postprandial area under curve for glucose or insulin between the vaccenic acid and control MUFA diets.
The addition of trans 16:1 at 4% of E resulted in 7.6 times higher levels of GSIS than cis 16:1 in the perfused pancreas of Sprague–Dawley rats. However, the difference did not quite reach statistical significance (P=0.07; Stein et al., 1997). There was no difference between the effects of cis or trans 18:1 FA. A study in Wister rats compared isoenergetic diets containing 4.5 % TFA (derived from margarine) and 0% TFA (lipid content provided by corn oil), and reported no differences in fasting plasma glucose and insulin levels or insulin sensitivity index (Huang et al., 2009). This is in agreement with a study that reported male Wistar rats fed a diet enriched in either ruminant or industrial TFA at 4% of E for 8 weeks did not demonstrate significantly different insulin or glucose responses to a glucose tolerance test when compared with a control MUFA diet (Tardy et al., 2008). Similarly, a euglycemic-hyperinsulinemic clamp was performed to measure insulin sensitivity of male Sprague–Dawley rats after 8 weeks on a 10% SFA diet (LF-SFA), a 45% SFA diet (HF-SFA) or a LF-trans diet containing 4.6% TFA (and 5.4% other FA; Dorfman et al., 2009). There was no significant effect of the LF-trans diet on whole-body insulin action, and whereas there was a reduction in glucose disposal compared with the LF-SFA group, this was not different in any individual skeletal muscles that were tested.
Quantitative insulin-sensitivity check index was used to assess the effect of TFA consumption on male AKR/J mice (Koppe et al., 2009). The animals were fed isoenergetic diets containing either lard (0% E from TFA) or shortening (20% E from TFA). After 4 and 8 weeks, the mice fed the high TFA diet had significantly higher insulin levels and lower quantitative insulin-sensitivity check index values, with increased IL-1β expression (P<0.05). However, a further study evaluated the glucose metabolites and gluco-regulatory enzyme activities in skeletal muscle and liver of Wister rats fed diets that also contained high levels of TFA (17% (wt/wt) cis MUFA or TFA), and no significant differences in glucose metabolism or enzyme activity were observed (2006).
Overall, the rodent studies provide some conflicting results, with a number of studies suggesting that high levels of TFA accentuate pancreatic GSIS, but in general those that have used TFA at levels similar to those seen in human diets have not shown any significant differences between groups. Those that have shown negative effects of TFA have either been confounded by variations in other fat types (Ibrahim et al., 2005; Saravanan et al., 2005a, 2005b) or used extremely high levels of TFA (Koppe et al., 2009).
Animal studies: primate models
A well-designed 6-year study fed 42 male African green monkeys diets containing fat blends composed either primarily of cis MUFA or with the substitution trans MUFA isomers at a level of 8% of E (Kavanagh et al., 2007). There were no significant differences between fasting insulin and glucose or homeostatic model assessment (HOMA) index, although the authors report a trend towards higher fasting glucose in the monkeys on the TFA diet (P=0.15). A subset of 12 monkeys underwent an assessment of postprandial metabolism, and there was a significantly higher insulin response in the animals fed TFA than those on the cis FA diet (P=0.015). Fructosamine levels were also higher in the TFA group (P=0.002), but lost significance when the model controlled for the higher level of intra-abdominal fat present in the TFA animals. TFA-fed animals also had lower insulin-stimulated Akt activation (P=0.02), with no differences between insulin receptor activation or TNF-α levels, suggesting that the effect was post-receptor.
Epidemiological studies and RCTs: study design and methodology
The odds ratio and relative risk stated throughout this review are those that have been adjusted for potential confounding factors. The specific factors that were adjusted for in each model have been shown within the tables, and usually include age, body mass index (BMI), % E intake and family history. The ‘P for trend’ values refer to the level of difference between the highest and lowest tertile/quartile/quintile of intake.
One of the problems in interpreting outcomes from the RCTs is that many use TFA levels far higher than would be normally consumed. The argument behind this approach is that it is necessary to have a high exposure in order that effects observed in response to habitual diets over long time scales can be observed over the shorter timescales that operate in most experimental diet studies. This approach assumes that there will be a linear association between the size of any effect and intake, that the same physiological systems and pathways will be involved at all levels of intake and they will be accelerated in response to increasing levels of the relevant compound. These assumptions need to be treated with care with respect to TFAs, particularly at extremely high levels of TFA intake, and it should be noted that our previous work has shown a potential non-linear relationship between TFA intake and coronary heart disease risk (SACN, 2007).
The three prospective cohort and five cross-cultural studies determined TFA exposure by dietary assessment, with two cross-cultural studies also measuring TFA levels in serum. These studies used a range of outcome measures, including risk assessments for diabetes or poor glycemic control and correlations between TFA intake and metabolic measurements. Eight RCTs investigated the effects of dietary FA substitution for between 3 and 6 weeks, whereas a meal study considered the effects of acute exposure during a single meal. A variety of fasting and postprandial markers of glucose tolerance/insulin sensitivity were used as outcome measures.
A group of 38 adults with a range of ages, BMIs and levels of glucose tolerance (five with type 2 diabetes) had fasting glucose measured and underwent a 2 h oral glucose tolerance test (Lovejoy et al., 2001). There was no association between either self-reported TFA intake or TFA enrichment in serum lipids and any of the insulin resistance parameters measured. A similar study in obese and normal-weight Spanish children found no correlation between measurements of insulin resistance (HOMA index and quantitative insulin-sensitivity check index index) and TFA intake or plasma TFA levels (Larque et al., 2006). Data from 1284 American Indians, who had been diagnosed with diabetes initially appeared to show a trend toward higher TFA consumption and poor glycemic control (Xu et al., 2007). However, this did not reach statistical significance, and was attenuated after adjustment for additional diabetic risk factors.
A significant relationship was reported between hydrogenated vegetable oil consumption and insulin resistance in adult Iranian women (Esmaillzadeh and Azadbakht, 2008). However despite stating that this study was initiated because of an interest in the relationship between TFA intake and diabetes, the authors did not quantify TFA intakes so that the exposure levels of subjects within the study cannot be ascertained with confidence. It is stated that the average TFA intake in Iran is high (4.2% of E), and so it is difficult to establish the relevance of their findings in the context of the diet in western countries where the intake is much lower.
Prospective cohort studies
The largest prospective study to report on the relationship between TFA intake and diabetes risk has been the Nurses' Health Study, which administered food frequency questionnaire at 2–4 years intervals. The data from 84 204 women with a follow-up of 14 years found that E-adjusted TFA intake was positively associated with risk of developing type 2 diabetes (relative risk 1.31; P for trend=0.02; Salmerón et al., 2001). A 2% increase in E from TFA was associated with a relative risk of 1.39, and substituting 2% of E from TFA with carbohydrate or polyunsaturated fatty acid was associated with a 28% and 40% lower risk of developing diabetes, respectively (P<0.001). A validation study found that the diagnosis of diabetes was confirmed in 98% of cases in this cohort (Colditz et al., 1992).
Another large study used the Iowa Women's Health Study cohort, and included 35 988 older women with 11 years follow-up (Meyer et al., 2001). This study initially found a small inverse association between TFA intake and risk of diabetes (P for trend 0.03), which strengthened through adjustment for magnesium and cereal fiber intake (P for trend 0.004), but was no longer significant when adjusted for different dietary FA (P for trend 0.20). The food frequency questionnaire was only administered at baseline in 1986. This lack of dietary follow-up is of concern as changes in fats formulations and individual food habits over the period of the study may lead to significant misclassification of TFA intake. The authors also carried out a small validation study among 85 participants, in which 44 had self-reported a diagnosis of diabetes. Only 28 of these diagnoses were confirmed, suggesting a possible misdiagnosis in 36% of cases, and raising the possibility of a large number of false positives having been included in the analysis of the whole cohort.
An analysis of data from the Health Professionals Follow-up Study used 4-year dietary and 2-year health questionnaires from 42 504 men (van Dam et al., 2002). There was a positive correlation between TFA intake and risk of developing diabetes when adjusted for age and E intake (P for trend=0.0004). The trend remained significant after further adjustment for physical activity, smoking, alcohol consumption, hypercholesterolemia, hypertension and family history (P for trend=0.009); but lost significance after adjustment for cereal fiber, magnesium and BMI (P for trend=0.33). No adjustments for other types of dietary fats were undertaken.
A prospective case–control study in Australia using 3737 men and women with 4 year of follow-up found a strong inverse correlation between TFA intake and diabetes risk (P for trend <0.0001) after adjustment for most standard factors, but no adjustment for diet was included (Hodge et al., 2007). The authors did not report the TFA intake for the different quintiles, but gave mean intakes of 0.10 and 0.12 g/day for the control and case group, respectively. These levels are significantly lower than the mean of 1.3 g/day reported for the Australian population by a recent government review of TFA intake (Food Standards Australia New Zealand, 2009). If the reported figures are correct, this would suggest that the participants in the study do not reflect the normal population, and that they consumed very low levels of products that contain ruminant TFA, such as dairy and meat products. In Australia, it is estimated that these products account for ∼63–75% of TFA intake, which would usually equate to intakes around 1.0 g/day TFA (Food Standards Australia New Zealand, 2009). As no adjustment for dietary factors was included in the model, it may be that the apparent protective effect of TFA intake reflects a lower risk in those individuals who consumed more dairy and animal protein.
Data from prospective epidemiological studies that evaluated dietary TFA intake (and reported quintile levels) are summarized in Figure 1.
Randomized control and meal studies
A randomized cross-over trial in obese subjects with type 2 diabetes used three 6 week diets, each containing 20% of E from SFA, MUFA or TFA (Christiansen et al., 1997). It should be noted that this is an extremely high TFA intake. Insulin and glucose levels were determined in fasting and postprandial blood samples. Postprandial insulin levels were 59% higher after the TFA diet than after the MUFA diet (P<0.05), and 26% higher than after the baseline diet (P<0.05). C-peptide concentrations (a marker of insulin secretion) showed a similar pattern during the TFA diet, and were 42 and 32% higher than the MUFA and baseline diets, respectively (both P<0.05). There were no significant differences between the SFA and the TFA diets. Another study used 63 subjects with abdominal obesity but provided diets containing TFA at levels consistent with North American TFA intakes (Tardy et al., 2009). Each participant was assigned to one of three diets: low TFA diet (low TFA, 0.23±0.04% of E), ruminant TFA-rich diet (enriched in vaccenic acid, 2.04±0.27% of E from TFA) or industrial TFA rich (enriched in elaidic acid, 2.59± 0.48% of E from TFA). Volunteers were given spreads, desserts and biscuits to be consumed in set amounts each day for 4 weeks, with euglycemic hyperinsulinemic clamps used to measure insulin sensitivity before and after the diet period. There was no difference between any group for any measure of insulin sensitivity (fasting glucose, fasting insulin, HOMA or any parameter from the clamp).
A study involving 25 healthy subjects of normal (BMI <25 kg/m2) and increased (BMI 25–30 kg/m2) body weight compared three 4-week diets, using very high TFA levels in one diet phase (Lovejoy et al., 2002). Each diet contained ∼7% of E from trans 18:1, cis 18:1 or 16:0 FA. There was no significant effects of diet on any markers of insulin resistance, but whole body fat oxidization was lowest when the diet was enriched with MUFA (26.0±1.5 g/day) and highest when enriched with TFA (31.4±1.5 g/day; P=0.02). The overweight subjects had reductions in insulin sensitivity of 11 and 24% on the TFA and SFA diets, respectively, when compared with the MUFA diet. However, this difference did not reach statistical significance. Slightly lower TFA levels were used in a trial in 14 healthy young women, which compared a high-TFA diet (5.1% of E from TFA, 80:20 trans 18:l and 18:2) with a high-oleic diet (5.2% of E from cis 18:1; Louheranta et al., 1999). Each diet phase lasted 4 weeks, and was preceded by a 2-week basal diet. There was no significant difference between the diets for insulin sensitivity index or acute insulin response; although there was a small non-significant increase in fasting insulin levels on the TFA diet compared with the MUFA diet (8.1±0.6 mU/l and 7.4±0.5 mU/l, respectively; P=0.089).
The effect of six different types of dietary fats on glucose homeostasis was investigated using soybean oil, semi-liquid margarine, soft margarine, partially hydrogenated soybean oil (PHSO), traditional stick margarine and butter (Lichtenstein et al., 2003). Each diet lasted for 35 days and provided between 0.05–5.2% of E as TFA. There were small differences between the diets for some measurements, but these were not consistent with the levels of TFA present. The authors concluded that there was no association between TFA and glucose metabolism. A cross-over study compared diets containing either PHSO, soybean oil, palm oil or canola oil in 15 volunteers who were over the age of 50 years and had raised serum low-density lipoprotein-cholesterol (⩾130 mg/100 ml; Vega-Lopez et al., 2006). Each diet phase lasted for 35 days, and the PHSO diet provided 4.15% of E as TFA. There were no differences in fasting glucose levels between the diets, but the PHSO diet resulted in significantly higher fasting insulin levels and HOMA scores than either the soybean or canola oil diets (P<0.05). The effects of diets containing palm olein (POL), PHSO or interesterified soybean oil (IE) on glucose and insulin metabolism were compared in 30 participants (Sundram et al., 2007). The IE oil had the same FAs present as standard soybean oil, but their position on the triglyceride backbone had been modified. The diet containing PHSO provided 3.2% of E as TFA, with minimal TFA present in any of the other diets. Each diet was consumed in a random order, and lasted for 4 weeks. Fasting glucose increased modestly after the PHSO diet relative to POL (P<0.05), but was highest in the IE diet. Using the POL diet as a reference, fasting insulin decreased by 10 and 22% on the PHSO and IE diets, respectively (P>0.05 and P<0.001). Postprandial glucose was higher in the IE diet, but similar in the PHSO and POL diets. Postprandial 2 h insulin was lower in both IE and PHSO compared with the POL diet, but C-peptide levels were only significantly lower for the IE diet. Overall, the authors concluded that adverse effects on glucose metabolism were most significant in the diet containing IE relative to the dietary treatments that were higher in SFA or TFA.
A double-blind parallel study compared a control butter and a butter high in vaccenic acid (ruminant TFA; Tholstrup et al., 2006). Male volunteers consumed 115 g/day of the test fat for 3 weeks, providing a vaccenic acid intake of 3.6 g/day for those on the ruminant TFA diet. There was no difference in fasting insulin or glucose between groups, although volunteers on the control diet also had significantly lower MUFA intake during the study period, which may have confounded the results.
Consumption of a single meal containing 10% of E from either 18:1 cis MUFA or TFA resulted in a significantly greater increase in insulin levels after the TFA meal (P<0.05) for a group of overweight but healthy participants (Lefevre et al., 2005). An index of relative insulin sensitivity (product of postprandial changes in insulin and glucose concentrations) was also greater after the consumption of the TFA meal than the cis MUFA meal (P<0.05). The results were also analysed with respect to the presence or absence of the Thr54 FABP2 allele (12 Ala/Ala; 8 Thr/Ala, 2 Thr/Thr at codon 54 in FABP2), with Thr/Ala and Thr/Thr genotypes combined into a single group. Individuals with either Thr/Ala or Thr/Thr genotypes demonstrated a twofold greater postprandial glucose response (P<0.05) and a 70% greater postprandial insulin response (P=0.15) than Ala/Ala genotypes. The authors observed that these results were consistent with impaired insulin-mediated glucose uptake, due perhaps to fatty-acid-induced insulin resistance in the muscle. A significant genotype by meal interaction was observed for TAG fractional synthetic rate (P<0.05). This rate was not affected by the FA composition of the test meal for individuals in the Ala54 group, but individuals in the Thr54 group had a 20% higher rate after consumption of the TFA meal than after the cis MUFA meal (P<0.05).
The Ala54Thr polymorphism of FABP2 gene has been shown to increase affinity of intestinal fatty-acid-binding protein 2 for long-chain dietary FA, and a recent study has show accelerated incorporation of dietary unsaturated FAs, including TFA, into postprandial triglycerides in diabetic subjects carrying the FABP2 Thr54 allele. It was suggested this might increase the susceptibility of these subjects to adverse effects of specific FAs, but the study was not designed to address whether this dietary FA–gene interactions could increase risk of type 2 diabetes (Almeida et al., 2010).
Isolated pancreatic islets studies suggest that there is a differential effect of TFA compared with cis FA on the regulation of insulin secretion, with TFA potentiating GSIS more than cis isomers of identical chain length. There is also concordant data from cells and human studies, which suggest increased rates of oxidation of TFA compared with cis FA. There is limited data from rodent studies; the majority of existing studies have not shown any effect of TFA on markers of insulin resistance; the few that have reported a positive relationship have either used TFA levels of 20% of E or else are confounded by variations in other fat types.
Of the three large prospective cohort studies, two showed a positive association. In these studies, the impact of TFA on risk of diabetes were similar, with odds ratios of 1.31 for a range of TFA intakes from 1.3–2.9% of E (Salmerón et al., 2001), and 1.39 for TFA intakes from 0.7–2.0% of E (van Dam et al., 2002). However, in the latter study the association was no longer significant after adjustment for dietary factors and BMI. In the third study, an inverse association between TFA intakes and diabetes was reported, but these findings were limited by high likelihood of misclassification because of the collection of dietary data only at baseline, and potential misdiagnosis of diabetes in a large percentage of participants (Meyer et al., 2001). A recent study using data from the Nurses' Health Study has found a significant relationship between potato consumption and diabetes, even after adjusting for the standard factors including fiber, TFA and SFA consumption (Halton et al., 2006). This raises the possibility that variables around carbohydrate consumption may also need to be included in models for diabetes risk, and none of these studies included such adjustment.
One of the five population-based studies found an inverse relationship between TFA intake and diabetes (Hodge et al., 2007), but the level of TFA intake reported seem to be far outside of the normal range of intakes and unrealistically low, suggesting that the data may be incorrect or else that the population used in the study have significantly different dietary habit to the general population. The authors also made no adjustment for diet in their analysis. Another population study reported a strong correlation between hydrogenated fat consumption and diabetes (Esmaillzadeh and Azadbakht, 2008), but did not convert the levels of hydrogenated fat into TFA intake.
No effect on insulin sensitivity or glucose tolerance was found in five RCTs or meal studies of healthy individuals fed TFA diets at intakes between 2–7% of E. However, postprandial hyperinsulinemia was observed in obese subjects with type 2 diabetes fed diets containing 20% of E as TFA (Christiansen et al., 1997), and a study in 15 subjects with raised low-density lipoprotein-cholesterol found a diet containing 4% of E from TFA resulted in a higher HOMA value than diets containing MUFA (Vega-Lopez et al., 2006). In an acute meal study (Lefevre et al., 2005), a significant increase in insulin resistance was observed following meals high in TFA (10% of E). In the same study, other adverse effects of TFA (on triglyceride synthesis) were shown to be genotype dependent, with greater adverse response for individuals carrying the FABP2 Thr54 allele (∼28% of the population). A single study that compared diets containing TFA from ruminant and industrial sources at levels commonly found in the free-living population did not see any difference between diets (Tardy et al., 2009). Perhaps the most convincing evidence for a relationship between TFA intake and insulin resistance comes from the 6-year primate study comparing cis and trans isomers (Kavanagh et al., 2007). However, the diets contained 8% of E as TFA, which is between 4–8 × the levels currently consumed in most Western countries. Overall, the evidence at present is tentative, and further investigations at habitual TFA intakes are needed.
Despite the addition of new evidence since the SACN review (SACN, 2007), the data available to assess evidence for an association between TFA intakes and incidence of diabetes remains limited. At high TFA intakes there is some evidence for a weak association, but the relevance of these observations to western populations where the intake is much lower is debatable.
Overall, it seems unlikely that the levels of TFA consumed as part of a standard western diet will have a significant contribution to risk of diabetes or insulin resistance. The possibility of increased risk for individuals with particular genotypes (such as the FABP2 Thr54 allele) is of interest, but this has only been considered in a single study to date. Any further studies should be aware of the need for careful design and adjustment for the appropriate diet and lifestyle factors, and the importance of using tissue TFA levels or recent, well-validated questionnaires to determine TFA intake.
Almeida JC, Gross JL, Canani LH, Zelmanovitz T, Perassolo MS, Azevedo MJ (2010). The Ala54Thr polymorphism of the FABP2 gene influences the postprandial fatty acids in patients with type 2 diabetes. J Clin Endocrinol Metab 95: 3909–3917.
Alstrup KK, Brock B, Hermansen K (2004). Long-term exposure of INS-1 cells to cis and trans fatty acids influences insulin release and fatty acid oxidation differentially. Metabolism 53, 1158–1165.
Alstrup KK, Gregersen S, Jensen HM, Thomsen JL, Kjeld H (1999). Differential effects of cis and trans fatty acids on insulin release from isolated mouse islets. Metabolism 48, 22–29.
Bernal CA, Rovira J, Colandré ME, Cussó R, Cadefau JA (2006). Effects of dietary cis and trans unsaturated and saturated fatty acids on the glucose metabolites and enzymes of rats. Br J Nutr 95, 947–954.
Christiansen E, Schnider S, Palmvig B, Tauber-Lassen E, Pedersen O (1997). Intake of a diet high in trans monounsaturated fatty acids or saturated fatty acids. Effects on postprandial insulinemia and glycemia in obese patients with NIDDM. Diabetes Care 20, 881–887.
Colditz GA, Manson JE, Stampfer MJ, Rosner B, Willett WC, Speizer FE (1992). Diet and risk of clinical diabetes in women. Am J Clin Nutr 55, 1018–1023.
Cromer KD, Jenkins TC, Thies EJ (1995). Replacing cis octadecenoic acid with trans isomers in media containing rat adipocytes stimulates lipolysis and inhibits glucose utilization. J Nutr 125, 2394–2399.
Dorfman SE, Laurent D, Gounarides JS, Li X, Mullarkey TL, Rocheford EC et al. (2009). Metabolic implications of dietary trans-fatty acids. Obesity.
Esmaillzadeh A, Azadbakht L (2008). Consumption of hydrogenated versus nonhydrogenated vegetable oils and risk of insulin resistance and the metabolic syndrome among Iranian adult women. Diabetes Care 31, 223–226.
European Food Safety Authority (2004). Opinion of the scientific panel on the dietetic products, nutrition and allergies on a request from the commission related to the presence of trans fatty acids in foods and the effect on human health of the consumption of trans fatty acids. EFSA J 81, 1–49.
Food Standards Australia New Zealand (2009). Intakes of trans fatty acids in New Zealand and Australia. Internet: http://www.foodstandards.gov.au/scienceandeducation/publications/transfattyacidsrepor4560.cfm.
Halton TL, Willett WC, Liu S, Manson JE, Stampfer MJ, Hu FB (2006). Potato and french fry consumption and risk of type 2 diabetes in women. Am J Clin Nutr 83, 284–290.
Health Canada (2006). Transforming the food supply, report of the trans fat task force. internet: http://www.hc-sc.gc.ca/fn-an/nutrition/gras-trans-fats/tf-ge/tf-gt_rep-rap-eng.php.
Hodge AM, English DR, O′Dea K, Sinclair AJ, Makrides M, Gibson RA et al. (2007). Plasma phospholipid and dietary fatty acids as predictors of type 2 diabetes: interpreting the role of linoleic acid. Am J Clin Nutr 86, 189–197.
Hu FB (2001). Diet and risk of type II diabetes: the role of types of fat and carbohydrate. Diabetologia 44, 805–817.
Huang Z, Wang B, Pace RD, Yoon S (2009). Trans fat intake lowers total cholesterol and high-density lipoprotein cholesterol levels without changing insulin sensitivity index in Wistar rats. Nutr Res 29, 206–212.
Ibrahim A, Natarajan S, Ghafoorunissa R (2005). Dietary trans-fatty acids alter adipocyte plasma membrane fatty acid composition and insulin sensitivity in rats. Metabolism 54, 240–246.
Kavanagh K, Jones KL, Sawyer J, Kelley K, Carr JJ, Wagner JD et al. (2007). Trans fat diet induces abdominal obesity and changes in insulin sensitivity in monkeys. Obesity 15, 1675–1684.
Koppe SWP, Elias M, Moseley RH, Green RM (2009). Trans fat feeding results in higher serum alanine aminotransferase and increased insulin resistance compared with a standard murine high-fat diet. Am J Physiol Gastrointest Liver Physiol 297, G378–G384.
Larque E, Gil-Campos M, Ramirez-Tortosa MC, Linde J, Canete R, Gil A (2006). Postprandial response of trans fatty acids in prepubertal obese children. Int J Obes 30, 1488–1493.
Lefevre M, Lovejoy JC, Smith SR, DeLany JP, Champagne C, Most MM et al. (2005). Comparison of the acute response to meals enriched with cis- or trans-fatty acids on glucose and lipids in overweight individuals with differing FABP2 genotypes. Metabolism 54, 1652–1658.
Lemaitre RN, King IB, Patterson RE, Psaty BM, Kestin M, Heckbert SR (1998). Assessment of trans-fatty acid intake with a food frequency questionnaire and validation with adipose tissue levels of trans-fatty acids. Am J Epidemiol 148, 1085–1093.
Lichtenstein A, Erkkilä A, Lamarche B, Schwab U, Jalbert S, Ausman L (2003). Influence of hydrogenated fat and butter on CVD risk factors: remnant-like particles, glucose and insulin, blood pressure and C-reactive protein. Atherosclerosis 171, 97–107.
Louheranta AM, Turpeinen AK, Vidgren HM, Schwab US, Uusitupa MIJ (1999). A high-trans fatty acid diet and insulin sensitivity in young healthy women. Metabolism 48, 870–875.
Lovejoy JC, Champagne CM, Smith SR, DeLany JP, Bray GA, Lefevre M et al. (2001). Relationship of dietary fat and serum cholesterol ester and phospholipid fatty acids to markers of insulin resistance in men and women with a range of glucose tolerance. Metabolism 50, 86–92.
Lovejoy JC, Smith SR, Champagne CM, Most MM, Lefevre M, DeLany JP et al (2002). Effects of diets enriched in saturated (palmitic), monounsaturated (oleic), or trans (elaidic) fatty acids on insulin sensitivity and substrate oxidation in healthy adults. Diabetes Care 25, 1283–1288.
Meyer KA, Kushi LH, Jacobs Jr DR, Folsom AR (2001). Dietary fat and incidence of type 2 diabetes in older Iowa women. Diabetes Care 24, 1528–1535.
Mozaffarian D, Clarke R (2009). Quantitative effects on cardiovascular risk factors and coronary heart disease risk of replacing partially hydrogenated vegetable oils with other fats and oils. Eur J Clin Nutr 63, S22–S33.
SACN (2002). A framework for evaluation of evidence that relates food and nutrients to health. Internet: http://www.sacn.gov.uk/pdfs/sacn_02_02a.pdf.
SACN (2007). Update on Trans Fatty Acids and Health. Position Statement by the Scientific Advisory Committee on Nutrition (SACN). HMSO: London.
Salmerón J, Hu FB, Manson JE, Stampfer MJ, Colditz GA, Rimm EB et al. (2001). Dietary fat intake and risk of type 2 diabetes in women. Am J Clin Nutr 73, 1019–1026.
Saravanan N, Haseeb A, Ehtesham NZ, Ghafoorunissa R (2005a). Differential effects of dietary saturated and trans-fatty acids on expression of genes associated with insulin sensitivity in rat adipose tissue. Eur J Endocrinol 153, 159–165.
Saravanan N, Ibrahim A, Ghafoorunissa R (2005b). Dietary trans fatty acids alter diaphragm phospholipid fatty acid composition, tryacylglycerol content and glucose transport in rats. Br J Nutr 93, 829–833.
Stein DT, Stevenson BE, Chester MW, Basit M, Daniels MB, Turley SD et al. (1997). The insulinotropic potency of fatty acids is influenced profoundly by their chain length and degree of saturation. J Clin Invest 100, 398–403.
Steyn NP, Mann J, Bennett PH, Temple N, Zimmet P, Tuomilehto J et al. (2004). Diet, nutrition and the prevention of type 2 diabetes. Public Health Nutr 7, 147–165.
Sundram K, Karupaiah T, Hayes KC (2007). Stearic acid-rich interesterified fat and trans-rich fat raise the LDL/HDL ratio and plasma glucose relative to palm olein in humans. Nutr Metabol 4, 3.
Tardy A-L, Giraudet C, Rousset P, Rigaudiere J-P, Laillet B, Chalancon S et al. (2008). Effects of trans MUFA from dairy and industrial sources on muscle mitochondrial function and insulin sensitivity. J Lipid Res 49, 1445–1455.
Tardy A-L, Lambert-Porcheron S, Malpuech-Brugere C, Giraudet C, Rigaudiere J-P, Laillet B et al. (2009). Dairy and industrial sources of trans fat do not impair peripheral insulin sensitivity in overweight women. Am J Clin Nutr 90, 88–94.
Tholstrup T, Raff M, Basu S, Nonboe P, Sejrsen K, Straarup EM (2006). Effects of butter high in ruminant trans and monounsaturated fatty acids on lipoproteins, incorporation of fatty acids into lipid classes, plasma C-reactive protein, oxidative stress, hemostatic variables, and insulin in healthy young men. Am J Clin Nutr 83, 237–243.
van Dam RM, Willett WC, Rimm EB, Stampfer MJ, Hu FB (2002). Dietary fat and meat intake in relation to risk of type 2 diabetes in men. Diabetes Care 25, 417–424.
van de Vijver LPL, Kardinaal AFM, Couet C, Aro A, Kafatos A, Steingrimsdottir L et al. (2000). Association between trans fatty acid intake and cardiovascular risk factors in Europe: the TRANSFAIR study. Eur J Clin Nutr 54, 126–135.
Vega-Lopez S, Ausman LM, Jalbert SM, Erkkila AT, Lichtenstein AH (2006). Palm and partially hydrogenated soybean oils adversely alter lipoprotein profiles compared with soybean and canola oils in moderately hyperlipidemic subjects. Am J Clin Nutr 84, 54–62.
Wang Y, Jacome-Sosa MM, Ruth MR, Goruk SD, Reaney MJ, Glimm DR et al (2009). Trans-11 vaccenic acid reduces hepatic lipogenesis and chylomicron secretion in JCR:LA-cp rats. J Nutr 139, 2049–2054.
Wang Y, Lu J, Ruth MR, Goruk SD, Reaney MJ, Glimm DR et al. (2008). Trans-11 vaccenic acid dietary supplementation induces hypolipidemic effects in JCR:LA-cp rats. J Nutr 138, 2117–2122.
Wild S, Roglic G, Green A, Sicree R, King H (2004). Global prevalence of diabetes. Diabetes Care 27, 1047–1053.
Xu J, Eilat-Adar S, Loria CM, Howard BV, Fabsitz RR, Begum M et al. (2007). Macronutrient intake and glycemic control in a population-based sample of American Indians with diabetes: the Strong Heart Study. Am J Clin Nutr 86, 480–487.
We thank the Food Standards Agency (UK) for providing funding. Funding for Update on trans fatty acids and health provided by Food Standards Agency (UK).
The authors declare no conflict of interest.
About this article
Cite this article
Thompson, A., Minihane, AM. & Williams, C. Trans fatty acids, insulin resistance and diabetes. Eur J Clin Nutr 65, 553–564 (2011). https://doi.org/10.1038/ejcn.2010.240
- insulin resistance
- insulin sensitivity trans fatty acids
- trans FA
Lipids in Health and Disease (2020)
Saturated and monounsaturated fatty acid status is associated with bone strength estimated by calcaneal ultrasonography in Inuit women from Nunavik (Canada): A cross-sectional study
The journal of nutrition, health & aging (2014)
European Journal of Clinical Nutrition (2012)