Clinical Nutrition 30 (2011) 590e592
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Original article
Inverse association between serum phospholipid oleic acid and insulin resistance in subjects with primary dyslipidaemia A. Sala-Vila a, b, *, M. Cofán a, b, R. Mateo-Gallego c, A. Cenarro c, F. Civeira c, E. Ortega a, d, E. Ros a, b a
Endocrinology and Nutrition Department, Institut d’Investigacions Biomèdiques August Pi Sunyer (IDIBAPS), Hospital Clínic, Barcelona, Spain Ciber Fisiopatología de la Obesidad y Nutrición (CIBERobn), Instituto de Salud Carlos III (ISCIII), Spaine c Lipid Clinic and Molecular Investigation Laboratory, Hospital Universitario Miguel Servet, Instituto Aragonés de Ciencias de la Salud, Zaragoza, Spain d CIBER de Diabetes y Enfermedades Metabólicas Asociadas (CIBERdem), ISCIII, Spainf b
a r t i c l e i n f o
s u m m a r y
Article history: Received 25 October 2010 Accepted 26 February 2011
Background & aims: Data on intake of oleic acid (OA) and insulin resistance (IR) are inconsistent. We investigated whether OA in serum phosphatidylcholine relates to surrogate measures of IR in dyslipidaemic subjects from a Mediterranean population. Methods: Cross-sectional study of 361 non-diabetic subjects (205 men, 156 women; mean age 44 and 46 y, respectively; BMI 25.7 kg/m2). IR was diagnosed by BMI and HOMA values using published criteria validated against the euglycemic clamp. Alternatively, IR was defined by the 75th percentile of HOMA-IR of our study population. The fatty acid composition of serum phosphatidylcholine was determined by gas-chromatography. Results: The mean (SD) proportion of OA was 11.7 2.0%. Ninety-two subjects (25.5%) had IR. By adjusted logistic regression, including the proportions of other fatty acids known to relate to IR, the odds ratios (OR) (95% confidence intervals) for IR were 0.75 (0.62e0.92) for 1% increase in OA and 0.84 (0.71e0.99) for 1% increase in linoleic acid. Other fatty acids were unrelated to IR. When using the alternate definition of IR, OA remained a significant predictor (0.80 [0.65e0.99]). Conclusions: Higher phospholipid proportions of OA relate to less IR, suggesting an added benefit of increasing olive oil intake within the Mediterranean diet. Ó 2011 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Biomarker HOMA-IR Mediterranean diet Olive oil Phospholipid fatty acids
1. Introduction Data on the link between monounsaturated fatty acid (MUFA) consumption and insulin resistance (IR) are inconsistent.1,2 In the Mediterranean diet, MUFA and its principal species oleic acid (OA) are supplied to a great extent by olive oil. Replacement of food sources of saturated fatty acids (SFA) by olive oil has been related to reduced IR in a few studies from Southern Spain, as reviewed.2,3 Furthermore, recent reports from the PREDIMED study, a large nutrition-intervention trial, showed that a Mediterranean diet enriched with either olive oil or nuts (another source of MUFA) improved insulin sensitivity4 and reduced the incidence of diabetes5 in a high-risk population. Conversely, in non-Mediterranean Abbrevations: MUFA, monounsaturated fatty acids; IR, insulin resistance; OA, oleic acid; SFA, saturated fatty acids; PUFA, polyunsaturated fatty acids; HOMA, homeostasis model assessment; PC, phosphatidylcholine; LA, linoleic acid; AA, arachidonic acid; EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid. * Corresponding author. Clínica de Lípids, Edifici Helios Despatx 8, Hospital Clínic, Villarroel 170, 08036 Barcelona, Spain. Tel.: þ34932275400x2276; fax: þ34934537829. E-mail address:
[email protected] (A. Sala-Vila). e http://www.ciberobn.es f http://www.ciberdem.org
countries, where olive oil is used scarcely, prospective studies show a high correlation between MUFA and SFA intake. This might be due to shared sources of these fats in animal products (i.e., beef and dairy products), and might explain in part the lack of association of MUFA intake with the risk of type-2 diabetes in these populations.6 Patients with IR systematically disclose skeletal muscle and circulating lipids enriched with SFA in association with depletion of n-6 polyunsaturated fatty acids (PUFA).1 However, to the best of our knowledge, epidemiological studies failed to find any association between OA enrichment of circulating lipids and IR.1 We hypothesized that the proportion of OA in plasma phospholipids, as a marker of olive oil intake in a Mediterranean population,7 would be inversely associated with IR in primary dyslipidaemia subjects, a population with variable levels of cardiovascular risk, including IR. 2. Methods 2.1. Subjects In a cross-sectional study, 361 non-diabetic subjects (205 men and 156 women, mean age 44 and 46 years, respectively) were
0261-5614/$ e see front matter Ó 2011 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved. doi:10.1016/j.clnu.2011.02.008
A. Sala-Vila et al. / Clinical Nutrition 30 (2011) 590e592
evaluated from 1st June 2006 to end of September 2008 in two Lipid Clinics located at university hospitals in Northern Spain (Hospital Clínic of Barcelona [n ¼ 185] and Hospital Miguel Servet of Zaragoza [n ¼ 176]). These subjects were part of a larger group for which PC fatty acids in relation to preclinical atherosclerosis in the carotid arteries have been recently reported.8 All participants provided informed consent to a protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and approved by the ethical review boards of each institution. 2.2. Biochemical analyses All participants had a venipuncture performed to obtain fasting blood samples after 4 weeks without hypolipidemic drug treatment. Blood glucose was measured by the glucose-oxidase method. Insulin was measured by radioimmunoassay. Homeostasis model assessment (HOMA-IR) was estimated as fasting plasma insulin (mU/mL) glucose (mmol/L)/22.5. IR was diagnosed if any of the following conditions were met: BMI > 28.9 kg/m2; HOMAIR > 4.65; or BMI > 27.5 kg/m2 and HOMA-IR > 3.60, according to the criteria described by Stern et al.9 An alternative diagnostic criterion for IR (HOMA-IR > 75th percentile of the distribution in our patient group) was also used. Fatty acid methyl-esters from serum phosphatidylcholine (PC) were determined as described.8 Each fatty acid is expressed as percentage of total fatty acids. Further details on the clinical aspects and assessment of other biochemical variables can be found in supplementary electronic material. 2.3. Statistical analyses When appropriate, the unpaired t, ManneWhitney, and c2 tests were used to assess the effect of IR presence on cardiovascular risk factors and fatty acid composition of serum PC. We constructed logistic regression models to assess the risk of IR (OR and 95% CI) for an 1% increase in the proportion of OA in serum PC. The model was adjusted for potential confounders, including centre, gender, age, smoking, being physically active, fasting triglycerides, HDL cholesterol, and non-HDL cholesterol (increases by 1 SD, corresponding to 1.83, 0.40, and 1.56 mmol/L, respectively). To further investigate whether the association between OA and IR was mediated through other fatty acids known to relate to IR (linoleic acid [LA], arachidonic acid [AA], sum of eicosapentaenoic [EPA] and docosahexaenoic [DHA] acids, and the palmitoleic acid/palmitic acid ratio), the proportions of these fatty acids in serum PC were also included in the model as confounders. A second model was constructed, with IR defined as HOMA-IR > 75th percentile as dependent variable with adjustment for the same variables plus BMI. For all tests statistical significance was defined as p < 0.05. Analyses were performed using SPSS software, release 16.0 (SPSS Inc., Chicago, IL). 3. Results Table 1 shows the clinical and biochemical characteristics of the study subjects by IR status. Predictably, subjects with IR were older and showed higher BMI, serum triglycerides, fasting glucose and insulin, and lower HDL cholesterol than those without. Fig 1 depicts statistically significant determinants of IR by the criteria of Stern et al.9 as assessed by multivariate logistic regression, which included age, fasting triglycerides, and the serum PC proportions of OA and LA. No significant associations were found for centre (0.69 [0.39e1.21]); being physically active (0.61 [0.35e1.06]); male gender (1.86 [0.99e3.51]); smoking (0.82 [0.44e1.51]); HDL cholesterol (1.09 [0.76e1.56]); non-HDL cholesterol (1.10 [0.85e1.42]), and serum PC
591
Table 1 Clinical characteristics, untreated lipid profiles, measures of glucose control, and main serum phosphatidylcholine fatty acids in 361 asymptomatic subjects with primary dyslipidaemia by presence or absence of insulin resistance (defined by Stern9). Variable
IR (n ¼ 92)
No IR (n ¼ 269)
Male, n (%) Age, years Current smoker, n (%) BMI, kg/m2 Waist circumference, cm Total cholesterol, mmol/L LDL cholesterol, mmol/L (n ¼ 334) HDL cholesterol, mmol/L Non-HDL cholesterol, mmol/L Triglycerides, mmol/L Fasting glucose, mmol/L Fasting insulin, mU/mL HOMA-IR Serum phosphatidylcholine fatty acid, % 16:0 18:0 S saturated fatty acids 16:1 n-7 18:1 n-9cis 18:1 n-9 trans S monounsaturated fatty acids 18:2 n-6 20:3 n-6 20:4 n-6 S n-6polyunsaturatedfattyacids 18:3 n-3 20:5 n-3 22:5 n-3 22:6 n-3 S n-3polyunsaturatedfattyacids
65 (70.7) 49 11 24 (26.1) 30.1 2.6 101.7 7.4 8.15 1.67 5.84 1.78
140 (52.0)* 44 13*** 72 (26.8) 24.2 2.8*** 86.4 10.7*** 8.00 1.46 5.90 1.54
1.24 0.36 6.87 1.79 2.01 [1.31e3.44] 5.36 0.61 16.5 [10.4e22.6] 4.0 [2.5e5.3]
1.38 0.41** 6.63 1.48 1.34 [0.88e2.04]*** 4.98 0.56*** 9.6 [7.8e12.4]*** 2.1 [1.7e2.7]***
28.29 [26.92e29.13] 14.50 [13.57e15.50] 42.87 [42.33e43.83] 0.55 [0.39e0.72] 11.39 [10.12e12.84] 1.42 [1.27e1.55] 13.75 [12.43e15.16] 21.40 [19.12e24.12] 3.56 [2.98e4.28] 10.10 [9.03e11.41] 35.77 [34.04e37.71] 0.12 [0.10e0.17] 0.98 [0.60e1.34] 0.75 [0.64e0.85] 4.73 [4.06e5.60] 6.56 [5.59e7.96]
27.77 [26.80e28.74] 14.18 [13.34e15.04]* 42.46 [41.62e43.22]** 0.49 [0.39e0.64] 11.59 [10.49e13.00] 1.50 [1.34e1.68]** 14.20 [12.73e15.65] 22.77 [20.59e24.68]* 3.17 [2.60e3.76]*** 9.84 [8.59e11.02] 36.05 [34.21e37.66] 0.14 [0.11e0.18]* 0.88 [0.59e1.46] 0.74 [0.64e0.84] 4.91 [4.11e5.72] 6.93 [5.58e8.11]
Data are means SD except for qualitative variables (expressed as n and %) and fasting triglycerides, fasting insulin, HOMA-IR, and fatty acids, expressed as medians [interquartile ranges]. HOMA-IR calculated as fasting plasma insulin [mU/mL] glucose [mmol/L]/22.5; IR, insulin resistance. *p < 0.05, **p < 0.01, ***p < 0.001 vs subjects with insulin resistance.
proportions of arachidonic acid (0.92 [0.74e1.15]), sum of EPA and DHA (0.83 [0.67e1.03]), and the palmitoleic acid/palmitic acid ratio (0.94 [0.68e1.30]). When including the sum of SFA or the proportion of palmitic acid as confounders instead of the palmitoleic acid/
Age, per 10 y
Fasting TG, per 1.83 mmol/L
Oleic acid, by 1% increase
Linoleic acid, by 1% increase 0.5
1
2
4
Odds Ratio Fig. 1. Significant determinants of insulin resistance (according to the criteria described by Stern et al.9) in 361 subjects with primary dyslipidaemia by multivariate logistic regression. Data are presented as point estimates with 95% CI. Values relating to risk of insulin resistance are 1.44 (1.14e1.81) for age; 1.52 (1.14e2.04) for fasting triglycerides (TG); 0.75 (0.62e0.92) for oleic acid; and 0.84 (0.71e0.99) for linoleic acid. Other variables included in the model (gender, centre, smoking, being physically active, fasting HDL cholesterol and non-HDL cholesterol, and serum phosphatidylcholine proportions of arachidonic acid, sum of eicosapentaenoic and docosahexaenoic acids, and the palmitoleic acid/palmitic acid ratio) were not significant predictors of insulin resistance (ORs and 95% CI in text).
592
A. Sala-Vila et al. / Clinical Nutrition 30 (2011) 590e592
palmitic acid ratio, the direction and strength of the associations was not materially changed (data not shown). In a similar model with IR diagnosed as HOMA-IR > 75th percentile, OA remained a significant determinant of IR (0.80 [0.65e0.99]) together with age (1.19 [0.91e1.54]) and fasting triglycerides (1.84 [1.30e2.61]). BMI also became a significant predictor in this model (1.22 [1.11e1.33]). 4. Discussion The principal and novel finding of our study is that, in asymptomatic dyslipidaemic men and women from a Mediterranean population, higher proportions of OA in serum PC (as a biomarker of inner membranes) are associated with less IR as defined by two different criteria. This result supports circumstantial evidence relating olive oil-derived MUFA intake within the Mediterranean diet to beneficial effects on glycaemic control and IR.2,3 Our results regarding less IR with increasing LA in serum PC, but not with higher proportions of EPA þ DHA, concur with prior epidemiological evidence.1 Some methodological issues should be considered. First, although the strength of the association between OA intake and its enrichment in circulating lipids is weaker than those found for fatty acids not synthesized endogenously such as trans or odd-chain fatty acids, LA, a-linolenic acid, and fish-derived EPA and DHA, populations with high olive oil consumption consistently show higher OA proportions in circulating phospholipids.7 Our values of OA in serum PC concur with those reported for different Spanish population groups in a recent survey.7 Second, despite the fact that adiposity and fat distribution are important determinants of IR, we did not include BMI or waist circumference as confounders in the multivariate analysis, since BMI was used for the diagnosis of IR.9 However, when an alternative definition of IR according to the distribution of HOMA-IR was used, BMI was included in the model as a confounder and expectedly became an independent determinant of IR. Third, the palmitoleic acid/palmitic acid ratio (a proxy of the activity of stearoyl-coA desaturase) was included in the model due to its direct relationship with IR and validity as an indirect marker of SFA intake.1 There are limitations to our study. First, we used neither the goldstandard test for IR, the euglycemicehyperinsulinaemic clamp, nor the second-to-best minimal-model analysis frequently sampled intravenous glucose tolerance test. Both tests are invasive, laborintensive, and expensive, and their use is not feasible for IR diagnosis in epidemiological studies. However, the criteria proposed by Stern et al.9 that we used have been developed within a large multicentric collection of euglycemicehyperinsulinaemic clamp data and thus permit using easier-to-obtain clinical measurements. Second, the criteria of Stern et al.9 might not be valid for a group of Mediterranean subjects with primary dyslipidaemia. For this reason, we used a second criterion for the diagnosis of IR more specific to our population. The fact that OA was an independent determinant in both models reinforces the validity of the observed association. Finally, fatty acids in serum PC do not reflect long-term intake as accurately as adipose tissue or red blood cells do. There are also strengths to our study, such as the comprehensive characterization of study subjects for cardiovascular risk factors and the use of PC fatty acids as objective biomarkers of intake that are not subject to the possibilities of bias or inaccurate reporting. In conclusion, higher phospholipid proportions of OA, as a marker of high olive oil intake, are associated with a lower HOMAIR in a Mediterranean population of primary dyslipidaemia subjects. However, the extent to which this association is only attributable to OA deserves further research, given that dietary polyphenols (antioxidant compounds naturally present in virgin olive oil) are
suggested to modulate insulin sensitivity.10 Our results add to current knowledge on the relationships among MUFA, insulin sensitivity, and diabetes risk.2e5 They also strengthen the evidence of MUFA derived from olive oil as a healthy fat within the Mediterranean diet. Eventually the large-scale PREDIMED trial, which uses Mediterranean diets supplemented with virgin olive oil or nuts against a low-fat diet for primary cardiovascular prevention in persons at high cardiovascular risk,4,5 might provide stronger evidence on the beneficial effects of OA in IR and diabetes risk. Conflict of interest None of the authors had a personal or financial conflict of interest. Author contribution The author’s responsibilities were as follows: ER and AS-V designed the study; AS-V and MC set up the laboratory method of analysis of fatty acids and performed these analyses; AS-V, EO and ER drafted the manuscript; all other authors substantially contributed to the analysis and/or interpretation of data and revised the manuscript critically for important intellectual content. Acknowledgements Emili Corbella provided expert assistance with statistical analyses. CIBERobn and CIBERdem are initiatives of ISCIII, Spain. Work supported by grants FIS PS09/00673, PS09/01292 and RTIC C06/01 (RECAVA) from the Spanish Health Ministry and Fundació Privada Catalana de Nutrició i Lípids, Barcelona, Spain. AS-V is supported by post-doctoral contract FIS CD07/0083. Appendix. Supplementary material Supplementary data related to this article can be found online at doi:10.1016/j.clnu.2011.02.008. References 1. Risérus U, Willett WC, Hu FB. Dietary fats and prevention of type 2 diabetes. Prog Lipid Res 2009;48:44e51. 2. Tierney AC, Roche HM. The potential role of olive oil-derived MUFA in insulin sensitivity. Mol Nutr Food Res 2007;251:1235e48. 3. López-Miranda J, Pérez-Jiménez F, Ros E, De Caterina R, Badimón L, Covas MI, et al. Olive oil and health: summary of the II international conference on olive oil and health consensus report, Jaén and Córdoba (Spain) 2008. Nutr Metab Cardiovasc Dis 2010;(20):284e94. 4. Estruch R, Martínez-González MA, Corella D, Salas-Salvadó J, Ruiz-Gutiérrez V, Covas MI, et al, PREDIMED Study Investigators. Effects of a Mediterranean-style diet on cardiovascular risk factors: a randomized trial. Ann Intern Med 2006;145:1e11. 5. Salas-Salvadó J, Bulló M, Babio N, Martínez-González MA, Ibarrola-Jurado N, Basora J, et al, For the PREDIMED Study investigators. Reduction in the incidence of type 2-diabetes with the Mediterranean diet: results of the PREDIMED-Reus nutrition intervention randomized trial. Diabetes Care; 2010 Oct 13 [Epub ahead of print] PMID: 20929998. 6. Hu FB, van Dam RM, Liu S. Diet and risk of type II diabetes: the role of types of fat and carbohydrate. Diabetologia 2001;44:805e17. 7. Saadatian-Elahi M, Slimani N, Chajès V, Jenab M, Goudable J, Biessy C, et al. Plasma phospholipid fatty acid profiles and their association with food intakes: results from a cross-sectional study within the European prospective investigation into cancer and nutrition. Am J Clin Nutr 2009;89:331e46. 8. Sala-Vila A, Cofán M, Pérez-Heras A, Núñez I, Gilabert R, Junyent M, et al. Fatty acids in serum phospholipids and carotid intima-media thickness in Spanish subjects with primary dyslipidemia. Am J Clin Nutr 2010;92:186e93. 9. Stern SE, Williams K, Ferrannini E, DeFronzo RA, Bogardus C, Stern MP. Identification of individuals with insulin resistance using routine clinical measurements. Diabetes 2005;54:333e9. 10. Hanhineva K, Törrönen R, Bondia-Pons I, Pekkinen J, Kolehmainen M, Mykkänen H, et al. Impact of dietary polyphenols on carbohydrate metabolism. Int J Mol Sci 2010;11:1365e402.
