Involvement of dietary saturated fats, from all sources or of dairy origin only, in insulin resistance and type 2 diabetes

(1)

Nutrition Science$Policy

Involvement of dietary saturated fats, from all sources or of

dairy origin only, in insulin resistance and type 2 diabetes

Be´atrice Morio, Anthony Fardet, Philippe Legrand, and Jean-Michel Lecerf

Reducing the consumption of saturated fatty acids to a level as low as possible is a European public health recommendation to reduce the risk of cardiovascular dis-ease. The association between dietary intake of saturated fatty acids and develop-ment and managedevelop-ment of type 2 diabetes mellitus (T2DM), however, is a matter of debate. A literature search was performed to identify prospective studies and clini-cal trials in humans that explored the association between dietary intake of satu-rated fatty acids and risk of insulin resistance and T2DM. Furthermore, to assess whether specific foods, and not just the saturated fatty acid content of the food matrix, can have differential effects on human health, the relationship between consumption of full-fat dairy products, a main source of dietary saturated fatty acids, and risk of insulin resistance and T2DM was studied. There is no evidence that dietary saturated fatty acids from varied food sources affect the risk of insulin resistance or T2DM, nor is intake of full-fat dairy products associated with this risk. These findings strongly suggest that future studies on the effects of dietary satu-rated fatty acids should take into account the complexity of the food matrix. Furthermore, communication on saturated fats and their health effects should be prudent and well informed.

INTRODUCTION

Excess energy intake and positive energy balance are as-sociated with the development of obesity and insulin re-sistance, which are key features in the pathophysiology of type 2 diabetes (T2DM).1 Excess intake of dietary macronutrients may be an important risk factor in the development of obesity and insulin resistance, as high intakes of fat and sugar may contribute to excess energy intake.2Furthermore, the involvement of different types of dietary fat is a matter of scientific and clinical debate. In vitro experimental data have demonstrated that satu-rated fatty acids, especially the 16-carbon palmitate,

induce insulin resistance.3The recent report from Food and Agriculture Organization (FAO) of the United Nations concluded that saturated fat might be associ-ated with an increased risk of T2DM.4However, review of the current published literature indicates the need to re-evaluate this conclusion,5 particularly within the context of a diet in which saturated fat is replaced by other macronutrients. For example, several clinical tri-als and meta-analyses have demonstrated that replacing dietary fat with refined carbohydrates could decrease insulin sensitivity in healthy, overweight, or diabetic (T2DM) individuals.6–9 Furthermore, Forsythe et al.10 found that low-carbohydrate diets high in saturated Affiliation: B. Morio is with the CarMeN Laboratory (Inserm 1060, INRA 1397, INSA), University of Lyon, Faculty of Medicine Lyon – South, Oullins, France. B. Morio and A. Fardet are with the French National Institute for Agricultural Research (INRA), UMR1019 Human Nutrition, Center for Human Nutrition Research (CRNH) Auvergne, Clermont-Ferrand, France. P. Legrand is with the Biochemistry and Human Nutrition Laboratory, Agrocampus – French National Institute for Agricultural Research (INRA), Rennes, France. J-M. Lecerf is with the Nutrition Department, Pasteur Institute of Lille, Lille, France.

Correspondence: B. Morio, CarMeN Laboratory, Faculte´ de Me´decine Lyon Sud — BP 12, 165 Chemin du Grand Revoyet, 69921 Oullins Cedex, France. E-mail: beatrice.morio@lyon.inra.fr.

Key words: human, lipid nutrition, prevention of diabetes, public health, saturated fatty acids.

For Permissions, please e-mail: journals.permissions@oup.com.

doi: 10.1093/nutrit/nuv043

(2)

(86 g of saturated fat per day) or unsaturated (47 g of saturated fat per day) fats have equal impact on the con-centration of circulating saturated fatty acids and the se-rum insulin concentration in weight-stable men, thus raising questions about the effect of a high intake of sat-urated fat on insulin resistance and risk of T2DM.

Most of the evidence in the literature uses the term saturated fat for the classification of dietary fat, which makes it difficult to distinguish the effects of individual saturated fatty acids. Saturated fat refers mostly to the main dietary long-chain saturated fatty acids, C16 and C18. Still, certain foods such as dairy products are vec-tors of other saturated fatty acids, such as short- and medium-chain saturated fatty acids (C4–C10), whose biological characteristics and effect on health differ from those of long-chain fatty acids. A dose–response meta-analysis showed that a high intake of dairy prod-ucts was associated with a significant decrease in the risk of T2DM and found no significant association between the consumption of full-fat dairy products and risk of T2DM.11 This is further supported by the results from the prospective Multi-Ethic Study of Atherosclerosis,12which showed that trans-palmitoleate plasma concentrations, which are positively correlated with consumption of full-fat dairy products, were inde-pendently associated with a lower incidence of T2DM, including a 48% lower risk in quintile 5 compared with quintile 1 (hazard ratio [HR]: 0.52; 95% confidence in-terval [CI]: 0.32–0.85, P trend ¼ 0.02). Thus, some food sources rich in saturated fats have distinct impacts on health, depending on their composition of saturated fatty acids and other nutrients.

The objective of this review was to evaluate the evi-dence related to the association between dietary satu-rated fat and risk of insulin resistance and T2DM. The impact of reducing saturated fat intake in the context of replacement by carbohydrates was also examined. Finally, to assess whether a complex food matrix can impact human health beyond its content in saturated fat, the association between intake of full-fat dairy, a main food source of saturated fat with a relatively well-characterized matrix, and the development of insulin resistance and T2DM was studied.

METHODS Literature selection

An exhaustive literature search was performed through MEDLINE, the Food Science and Technology Abstracts database, and the Cochrane Library until August 31, 2014, to identify all studies that examined consumption of dietary fat, especially saturated fat, and the risk of

insulin resistance or T2DM. Search terms included “di-etary fat,” “saturated fat intake,” “fasting insulin,” “insu-lin resistance,” and “diabetes mellitus.”

A similar search was performed to identify studies that examined dairy consumption and dairy fat intake in relation to risk of insulin resistance or T2DM. Search terms included “dairy food/product/intake,” “dairy fat intake,” “fasting insulin,” “insulin resistance,” and “diabetes mellitus.” Finally, an additional query was performed using “pentadecanoic acid” and “trans-palmitoleic acid” as a circulating marker of dairy fat in-take, since circulating concentrations of pentadecanoic acid13 and trans-palmitoleic acid14 were shown to be valid biological markers of dairy fat intake.

The search was performed by the investigator (B.M.) together with 2 external experts (P.L. and J.M.L). Inclusion and exclusion criteria

Inclusion criteria were English-language articles and all prospective studies or clinical trials in adults (age 18 y). Animal experiments, ecological studies, commentaries, and general reviews were excluded. The aim was to identify randomized clinical trials and large prospective cohort studies, because their strengths and limitations are best suited for studying the risk of T2DM, a disease with a long latency.

When considering the literature on consumption of saturated fat, studies in which statistical models were not adjusted for total energy intake were excluded be-cause the metabolic pathways in which dietary saturated fatty acids are processed and the cellular consequences vary, depending on the daily energy balance. Furthermore, studies that used erythrocyte, serum, or plasma fatty acid profiles exclusively to examine the re-lationship between fat intake and insulin resistance or T2DM, in the absence of any dietary data, were also ex-cluded. Indeed, alterations in plasma saturated fatty acid content may be related to de novo liver lipogenesis, independent of changes in dietary fat intake.15,16Thus, elevated circulating levels of saturated fatty acids may result from changes in hepatic metabolism rather than from inadequate intake of dietary fats.

A different selection strategy was used for the liter-ature on dairy fat consumption. Since circulating con-centrations of pentadecanoic acid13 and trans-palmitoleic acid14 were shown to be valid biological markers of dairy fat intake, the few studies that used these measures were included.

Data extraction and synthesis

Data on study design, participant characteristics, mea-sures of fat intake or dairy product consumption,

(3)

statistical methods, and results were extracted. The in-formation has been summarized to provide a compara-tive analysis and is presented in tables separating prospective from intervention studies.

The purpose was to highlight the association be-tween the intake of saturated fats and the risk of insulin resistance or T2DM. In the studies that examined glu-cose homeostasis and insulin resistance specifically, the focus was on the association between dietary fat intakes and the following: (1) plasma concentrations of glucose, insulin, and glycated hemoglobin (HbA1c); (2) scores from the homeostasis model assessment of insulin re-sistance (HOMA-IR) or the revised quantitative insu-lin sensitivity check index; or (3) the results of oral glucose tolerance tests or intravenous glucose toler-ance tests.

RESULTS

Association between dietary saturated fat intake and insulin resistance or type 2 diabetes mellitus

Figure 1shows the literature search examining the asso-ciation between dietary saturated fat and insulin

resistance or T2DM. Table 117–36 summarizes the in-cluded studies, showing characteristics of the partici-pants and main results related to dietary intake of saturated, monounsaturated, and polyunsaturated fats.

Table S1in the Supporting Information online presents additional characteristics of the included studies, along with the methods used to determine dietary fat intake, the statistical adjustment used, and other significant findings.

Results from prospective studies. Three prospective stud-ies found a positive association between saturated fat consumption and plasma glucose levels, assessed by a fasting glucose test or a 2-hour oral glucose tolerance test,17,18 or by plasma insulin in nondiabetic popula-tions.19Furthermore, Marshall et al.20found that satu-rated fat intake was higher in subjects with impaired glucose tolerance who converted to T2DM after 1 to 3 years than in those who reverted to normal glucose tol-erance. By contrast, van Dam et al.21found that the in-take of saturated fat was not associated with an increased risk of T2DM in men, even after adjustment for differences in body mass index. Similarly, using data adjusted for differences in body mass index, Meyer

Figure 1 Summary of the literature search examining the association between dietary saturated fat and risk of insulin resistance or type 2 diabetes.

(4)

Table 1 Results of studies reporting associations between the type of dietary fat intake and insulin resistance or type 2 diabetes

Reference Name and characteristics of study No. of subjects, sex Age in yearsa

Main findings related to fat intake and insulin resistance or T2DM

Mean intake of fat

Prospective studies Feskens et al. (1995)18

Seven Countries Study, Finland, The The Netherlands Follow-up: 20 y

338 M 75 (mean) SFA intake in past was as-sociated with 2-h plasma glucose levels (OGTT)

At study time:

Normal glucose tolerance: SFA 16.3, MUFA 12.5, PUFA 5.5 en% Impaired glucose tolerance: SFA 18.6,

MUFA 12.4, PUFA 5.1 en% New diabetes: SFA 19.1, MUFA 12.9,

PUFA 6.6 en% At baseline:

Normal glucose tolerance: SFA 20.1, MUFA 13.6, PUFA 4.2 en% Impaired glucose tolerance: SFA 20.3,

MUFA 13.6, PUFA 4.0 en% New diabetes: SFA 21.6, MUFA 14.5,

PUFA 4.3 en% Feskens & Kromhaut (1990)17 Zutphen Elderly Study, The Netherlands Follow-up: 10 y

394 M 50–70 SFA intake was

significantly associated with fasting glucose levels (r¼ 0.13) but not with the AUC of post-prandial glucose levels

SFA 49.7, MUFA 51.9, PUFA 20.2 g/d

Marshall et al. (1994)20

San Luis Valley Diabetes Study, USA

Follow-up: 11–40 mo

134 M/F 30–74 SFA intake was higher in subjects converting to T2DM compared with those reverting to nor-mal glucose tolerance

Reverted to normal: SFA 14.0, MUFA 14.9, PUFA 7.0 en%

Remained impaired glucose tolerance: SFA 14.6, MUFA 16.3, PUFA 6.6 en% Converted to T2DM: SFA 16.1, MUFA

17.1, PUFA 7.5 en% Mayer et al. (1993)19 Kaiser Permanente Women Twins Study, USA Follow-up: 11 y

544 F 51 (mean) Higher intake of SFA was positively related to higher fasting insulin levels (log r¼ 0.08).

SFA 12.7, MUFA (C18:1) 13.8, PUFA (C18:2) 7.6 en%

Meyer et al. 200122

Iowa Women’s Health Study, USA Follow-up: 11 y

35 988 F 55–69 SFA intake was not related to incident T2DM SFA: Q1¼ 19.3 and Q5 ¼ 31.8 g/d MUFA: Q1¼ 20.4 and Q5 ¼ 33.8 g/d PUFA: Q1¼ 8.9 and Q5 ¼ 16.6 g/d n-3 PUFA: Q1¼ 0.03 and Q5 ¼ 0.39 g/d trans-MUFA: Q1¼ 2.2 and Q5 ¼ 5.2 g/d Salmero´n et al. (2001)23 Nurses’ Health Study, USA Follow-up: 14 y

84 204 F 34–59 SFA intake was not related to incident T2DM

SFA: Q1¼ 15 and Q5 ¼ 15 en% MUFA: Q1¼ 14 and Q5 ¼ 17 en% PUFA: Q1¼ 2 and Q5 ¼ 7 en% trans-MUFA: Q1¼ 2 and Q5 ¼ 3 en% van Dam et al.

(2002)21

Health Professionals Follow-up Study, USA

Follow-up: 12 y

42 504 M 40–75 SFA intake was associated with increased risk of T2DM (RR 1.34; 95%CI 1.09–1.66; Ptrend¼ 0.01);

the association was no longer significant after adjustment for BMI

SFA: Q1¼ 7.6 and Q5 ¼ 14.0 en% Oleic acid: Q1¼ 8.0 and Q5 ¼ 14.0 en% Linoleic acid: Q1¼ 3.5 and Q5 ¼ 6.8 en% a-linolenic acid: Q1¼ 321 and

Q5¼ 671 mg/d

n-3 PUFA: Q1¼ 80 and Q5 ¼ 570 mg/d trans-MUFA: Q1¼ 0.7 and Q5 ¼ 2.0 en% Intervention studies Berglund et al. (2007)28 DELTA Program, USA Crossover design Follow-up: 7 wk

85 M/F 21–61 SFA diet had no effect on blood glucose or insulin levels compared with MUFA and low-fat diets

Control diet: fat 35.8 en% (SFA 15.5, MUFA 14.4, PUFA 5.8 en%)

MUFA diet: fat 35.7 en% (SFA 8.7, MUFA 20.8, PUFA 6.2 en%) Bos et al. (2010)30 The Netherlands Parallel design Follow-up: 8 wk 24 M 36 F

40–65 Replacing a high-SFA diet with a high-MUFA diet or a Mediterranean diet did not affect insulin sensitivity as assessed by clamp, by fasting serum glucose, insulin, or C-peptide levels, or by HOMA-IR score

High SFA: fat 37 en% (SFA 20, MUFA 11, PUFA 6 en%)

High MUFA: fat 40 en% (SFA 11, MUFA 21, PUFA 7 en%)

Med diet: fat 40 en% (SFA 11, MUFA 22, PUFA 7 en%)

(continued)

(5)

Table 1 Continued

Mean intake of fat

Due et al. (2008)33 Denmark Parallel design Follow-up: 6 mo 55 M 76 F

28.2 6 4.8 Control diet with SFAs did not affect fasting glu-cose level but main-tained higher fasting insulin level and HOMA-IR score compared with MUFA diet

Control diet: fat 32.1 en% (SFA > 15.1, MUFA 10.4, PUFA 4.0 en%)

MUFA diet: fat 38.4 en% (SFA 7.1, MUFA 20.2, PUFA 7.8 en%) Due et al. (2008)34 Subcohort of MUFObes Study, Denmark Parallel design Follow-up: 6 mo 20 M 26 F

28.0 6 0.7 MUFA diet reduced fasting glucose and insulin lev-els and HOMA-IR score. Standard diet with SFA increased these vari-ables, but glucose and insulin responses to OGTT did not differ be-tween diets

Control diet: fat 35 (30–40) en% (SFA > 15, MUFA 5–15, PUFA 0–10 en%)

MUFA diet: fat 40 (35–45) en% (SFA < 10, MUFA > 20, PUFA 5–10 en%) Haghighatdo-ost et al. (2012)31 Iran Crossover design Follow-up: 6 wk

17 20–50 No effect of SFA vs MUFA diet on fasting serum insulin level or HOMA-IR score

SFA diet: fat 34 en% (SFA 16, MUFA 8 en%)

MUFA diet: fat 34 en% (SFA 8, MUFA 16 en%) Jebb et al. (2010)26 RISCK trial Parallel design Follow-up: 24 wk 230 M 318 F 52 6 10 Isoenergetic replacement of SFAs by MUFAs did not affect insulin sensi-tivity index as assessed by ivGTT and R-QUICKI

SFA high-GI diet: fat 37.9 en% (SFA 16.5, MUFA 11.6, PUFA 5.8 en%), GI 63.5 MUFA high-GI diet: fat 35.6 en% (SFA

9.5, MUFA 16.2, PUFA 6.6 en%), GI 63.3 MUFA low-GI diet: fat 35.7 en% (SFA 9.6,

MUFA 16.3, PUFA 6.9 en%), GI 60.7 Low-fat, high-GI diet: fat 27.5 en% (SFA

9.2, MUFA 9.8, PUFA 5.2 en%), GI 69.1 Low-fat, low-GI diet: fat 26.1 en% (SFA

8.3, MUFA 9.7, PUFA 5.1 en%), GI 65.1 Lopez et al.

(2011)36

Spain

Crossover design Follow-up: 8 h

14 M 33 6 7 SFA high-fat meal de-creased postprandial in-dexes of insulin sensitivity calculated from blood glucose and insulin responses com-pared with MUFA high-fat meal and control meal

Control meal: fat 0 en%

High-SFA, high-fat meal: fat 72 en% (SFA 47, MUFA 22, PUFA 3 en%)

High-MUFA, high-fat meal: fat 72 en% (SFA 11, MUFA 58, PUFA 3 en%)

Masson & Mensink (2011)35 The Netherlands Crossover design Follow-up: 8 h 13 M 18–70 High-SFA meal vs n-6 PUFA meal did not af-fect postprandial blood glucose and insulin responses

High-SFA meal: fat 51 en% High n-6 PUFA meal: fat 53 en%

Ortega et al. (2013)32 Spain Parallel design Follow-up: 2 wk 10 M 4 F

24.8 6 1.8 SFA diet did not alter fast-ing blood glucose and insulin levels, HOMA-IR score, or insulin sensitiv-ity index from OGTT

Basal diet: fat 34.3 en% (SFA 11.9 en%) SFA diet: fat 36.2 en% (SFA 21.7 en%)

Paniagua et al. (2007)29 Spain Crossover design Follow-up: 4 wk 4 M 7 F

35–75 Lower blood HbA1c and glucose levels with MUFA diet than with SFA diet. Plasma insulin was not affected. HOMA-IR score was de-creased only with MUFA diet when compared with SFA diet

Low-fat diet: fat 20 en% (SFA 6, MUFA 8, PUFA 6 en%)

MUFA diet: fat 38 en% (SFA 9, MUFA 23, PUFA 6 en%)

SFA diet: fat 38 en% (SFA 23, MUFA 9, PUFA 6 en%)

(continued)

(6)

et al.22and Salmero´n et al.23also concluded that dietary saturated fat was not associated with risk of T2DM in women of the Iowa Women’s Health Study or the Nurses’ Health Study.

Results from intervention studies. Mixed results were ob-tained from the 13 studies selected. Varying conclusions were reached by the 4 larger studies, i.e., the Women’s Health Initiative trial and the KANWU,25 RISCK,26 and LIPGENE27 intervention studies, all of which used similar design and methodological approaches.

In the large Women’s Health Initiative trial,24 19 541 (of 48 835) postmenopausal women were as-signed to an 8-year diet in which both total fat and satu-rated fat were reduced (28.6% and 9.5% of total energy intake, respectively) and replaced largely with carbohy-drates. The group that followed the fat, low-saturated-fat dietary pattern showed no differences in fasting glucose or insulin levels, HOMA-IR score, or in-cident T2DM compared with the control group, which consisted of of 29 294 women whose diet contained an average of 36.9% of total energy intake from fat and 12.4% from saturated fat.

In 162 weight-stable healthy subjects of the KANWU study,25insulin sensitivity measured by an in-travenous glucose tolerance test was decreased by 10% in the group of individuals who received a diet high in saturated fat (37% of energy as fat, >17% of energy as

saturated fat) for 3 months, but it remained unchanged in the group who received a diet high in monounsatu-rated fat (37% of energy as fat, >21% of energy as monounsaturated fat). Interestingly, when the subjects were divided into those consuming above and those consuming below the median total fat intake (i.e., 37% of energy as fat), the favorable effect of substituting monounsaturated fat for saturated fat on insulin sensi-tivity was seen only in individuals consuming less than 37% of energy as fat. All individuals with a total fat intake higher than 37% of energy showed de-creased insulin sensitivity in response to dietary intervention, regardless of the quality of fat consumed.

By contrast, the RISCK trial26 found, in 548 weight-stable individuals at risk for metabolic syn-drome, that a diet high in saturated fat (39% of energy as fat, 17% of energy as saturated fat) over 6 months had no impact on the revised quantitative insulin-sensi-tivity check index or on insulin sensiinsulin-sensi-tivity as measured using the intravenous glucose tolerance test, in compar-ison with a diet high in monounsaturated fat (36% of energy as fat, 16% of energy as monounsaturated fat). Similarly, in the LIPGENE cohort,27which consisted of 417 weight-stable subjects with metabolic syndrome, a diet high in saturated fat (40% of energy as fat, >17% of energy as saturated fat) over 3 months had no impact on plasma glucose or insulin concentrations, on the HOMA-IR score, or on insulin sensitivity as measured

Table 1 Continued

Mean intake of fat

Tierney et al. (2011)27 LIPGENE study, EU Parallel design Follow-up: 12 wk 183 M 234 F 54–55 (mean) Isoenergetic replacement of SFAs by MUFAs had no effect on fasting blood glucose and insu-lin levels, or on insuinsu-lin sensitivity as assessed by ivGTT

High-fat SFA diet: fat 36 en% (SFA 16, MUFA 12, PUFA 6 en%)

High-fat MUFA diet: fat 36 en% (SFA 8, MUFA 20, PUFA 6 en%)

Low-fat MUFA diet: fat 28 en% (SFA 8, MUFA 11, PUFA 6 en%þ MUFA 1 g/d) Low-fat n-3 PUFA diet: fat 26 en% (SFA

8, MUFA 11, PUFA 6 en%þ 1.24 g/d n-3 PUFA) Tinker et al. (2008)24 Women’s Health Initiative trial Parallel design Follow-up: 8.1 y

48 835 F 50–79 SFA vs low-fat diet showed no evidence of altering risk of T2DM (RR 0.95, 95%CI 0.90– 1.03)

Control diet: fat 36.9 en% (SFA 12.4, MUFA 17, PUFA 7.5 en%)

Intervention diet: fat 28.6 en% (SFA 9.5, MUFA 13.1, PUFA 6 en%)

Vessby et al. (2001)25 KANWU study, EU Parallel design Follow-up: 90 d 86 M 76 F

30–65 SFA diet decreased insulin sensitivity as assessed by ivGTT but did not af-fect insulin secretion

Fat 37 en%

SFA diet: SFA 17, MUFA 14, PUFA 6 en% MUFA diet: SFA 8, MUFA 23, PUFA 6

en%

Abbreviations: AUC, area under the curve; BMI, body mass index; CI, confidence interval; en%, percent of energy intake; EU, European Union; F, female; GI, glycemic index; HOMA-IR, homeostasis model assessment–estimated insulin resistance; ivGTT, intravenous glucose tolerance test; M, male; MUFA, monounsaturated fatty acid; OGTT, oral glucose tolerance test; PUFA, polyunsaturated fatty acid; Q, quintile; R-QUICKI, Revised Quantitative Insulin Sensitivity Check Index; RR, relative risk; SFA, saturated fatty acid; T2DM, type 2 diabe-tes mellitus.

a

Age at baseline for prospective studies.

(7)

using the intravenous glucose tolerance test, in compar-ison with a diet high in monounsaturated fat (39% of energy as fat, >18% of energy as monounsaturated fat).

Five other studies were performed in weight-stable overweight subjects with28,29 or without30–32 increased metabolic risk factors. Diets similar to those used in the above-mentioned trials were used, but over shorter du-rations (2–8 weeks). The studies used either parallel30,32 or crossover28,29,31designs. Conclusions were mostly in favor of an absence of effect of high saturated fat intake on glucose homeostasis assessed by fasting plasma glu-cose and insulin and by plasma gluglu-cose and insulin re-sponses to an oral glucose tolerance test.28,30,32 Only Paniagua et al.29 showed, in 11 insulin-resistant off-spring of obese and T2DM patients, that replacing satu-rated fat with monounsatusatu-rated fat for 28 days lowers the blood glucose and HbA1c levels and, thus, reduces the HOMA-IR score in comparison with a diet high in saturated fat.

In a population of obese individuals who lost more than 8% of their body weight,33,34a diet administered for 6 months and containing 32% of energy as fat with approximately 15% as saturated fat enhanced the fasting plasma insulin level and thus reduced the HOMA-IR score compared with basal values. However, this had no significant impact on the glucose and insulin responses to an oral glucose tolerance test, thus suggesting no sig-nificant impact on insulin sensitivity.33

Finally, 2 crossover studies explored the short-term effect of meals containing high amounts of saturated or unsaturated fat on blood glucose and insulin responses. In healthy adults, Masson and Mensink35 observed no significant alterations in postprandial glucose and insu-lin responses, whatever the diet. In contrast, in men newly diagnosed with type IIb or IV hyperlipoproteine-mia, Lopez et al.36found a higher postprandial insulin peak after a high-fat, high-saturated-fat meal (72% of energy as fat, 47% of energy as saturated fat) than after a high-fat, high-monounsaturated-fat meal (72% of en-ergy as fat, 11% of enen-ergy as saturated fat).

Interestingly, 5 of those intervention studies showed that replacing the dietary saturated fat with car-bohydrates does not reduce the risk factors insulin re-sistance.26,27,29,33,34 In the RISCK26 and LIPGENE27 trials, isoenergetic replacement of fat (from 36%–38% to 28% of energy intake) with carbohydrates (from 43% to 52% of energy intake) had no effect on fast-ing blood glucose and insulin levels or on the insulin sensitivity index.26,27In studies by Due et al.,33,34 lower-ing the fat intake from 32% to 24% of energy intake (on average) and increasing the carbohydrate intake from 50% to 58% of energy intake (on average) had a similar impact on plasma glucose and insulin concentrations as the saturated fat diet. Paniagua et al.29also found that

the HOMA-IR score was similarly reduced in response to carbohydrate (65% of energy intake) or high-saturated fat (23% of energy intake) diets compared with a diet high in monounsaturated fat (23% of energy intake).

Association between dairy fat intake and insulin resistance or type 2 diabetes mellitus

The literature search examining the association between dairy fat intake and insulin resistance or T2DM is com-piled in Figure 2. Table 237–55summarizes the charac-teristics of the included studies, the main results, and the characteristics of the dairy products.Table S2in the Supporting Information online presents additional characteristics of the included studies, the statistical ad-justment used, and other significant findings. Table S3

shows the methods used to determine dairy product and dairy fat intake.

Results from prospective studies. Of the 16 prospective studies that examined the total consumption of dairy products, 5 showed an inverse relation between total dairy consumption and T2DM incidence,37–41 and 5 found no association.42–47In addition, Kirii et al.47 con-cluded that dairy intake was inversely associated with T2DM incidence in women, but not in men. In con-trast, only 1 study identified a positive relation between dairy intake at the age of 14 years and incidence of high HbA1c levels in adulthood.48

Of the studies that examined the consumption of full-fat dairy products, 4 found no association39,41–42,46 and 1 found an inverse relationship38 with the risk or incidence of T2DM. More specifically, an inverse corre-lation was found between the proportions of pentadeca-noic and heptadecanoic acids in erythrocyte membranes and the incidence of T2DM in 2 stud-ies,49,50including the recent EPIC-InterAct case–cohort study performed in a large cohort of European adults.50 Furthermore, the case–control studies from the Va¨sterbotten Intervention Program and the Monitoring of Trends and Determinants in Cardiovascular Disease,51along with the Northern Sweden Health and Disease Study,52 established an inverse association be-tween the proportions of pentadecanoic and heptadeca-noic acids in plasma phospholipids and fasting insulin51 and glucose.52

Additional contrasting conclusions were noted when subgroups of dairy products were examined. In particular, consumption of whole milk was positively associated with T2DM incidence in the Health Professionals Follow-Up Study cohort,42but consump-tion of high-fat cheese was not46 or was inversely43,45 correlated with T2DM risk.

(8)

Results from intervention studies. Three intervention studies were identified, and their combined results indi-cate that increasing the intake of dairy products may help prevent insulin resistance and T2DM. Stancliffe et al.53investigated 40 overweight and obese adults with metabolic syndrome who were randomly assigned to ei-ther a low-dairy (<0.5 dairy serving per day and <600 mg calcium per day) or an adequate-dairy (>3.5 dairy servings per day, 1200 mg calcium per day) weight-maintenance diet for 3 months. They concluded that increasing total dairy intake to adequate levels does not affect blood glucose and reduces plasma insulin, HOMA-IR score, and circulating biomarkers of inflammation (tumor necrosis factor a, monochemoat-tractant protein 1, interleukin 6) and oxidative stress (malondialdehyde, oxidized low-density lipoprotein). Wennersberg et al.54 also found that 6 months of increased dairy consumption lowered fasting blood in-sulin in 121 middle-aged overweight subjects with traits of metabolic syndrome when compared with the control group. Finally, Benatar et al.55 recently ran-domly assigned 180 healthy volunteers to increase (þ12.5 6 15.7 g/d), reduce (10.4 6 101 g/d), or not

change (3.4 6 7.9 g/d) their dairy fat intake for 1 month. The dietary manipulations had no significant ef-fect on plasma glucose or insulin.

DISCUSSION

Association between dietary saturated fat intake and insulin resistance or type 2 diabetes mellitus

Several reports have examined the effect of dietary satu-rated fat on the risk of T2DM but could not reach a de-finitive conclusion.4,56,57 Whereas the FAO concluded there is a possible positive relationship between satu-rated fatty acid intake and increased risk of T2DM,4the French Agency for Food, Environmental and Occupational Health & Safety found that only high in-takes of saturated fat (i.e., >20% of total energy intake) might be associated with an increased risk of T2DM.57 In the present analysis, data compiled from the litera-ture, including prospective and intervention studies, point to the lack of conclusive evidence or consensus on the role of dietary saturated fat in T2DM.

Figure 2 Summary of the literature search examining the association between dairy fat intake and risk of insulin resistance or type 2 diabetes.

(9)

Table 2 Results of studies reporting associations between dairy consumption and insulin resistance or type 2 diabetes

Main findings related to dairy and insulin resistance or T2DM

Fat content specified

Prospective studies Choi et al. (2005)42

Health Professionals Follow-up Study, USA Follow-up: 12 y

41 254 M 40–75 Dairy intake associated with a modestly lower risk of T2DM (RR Q5 0.77, 95%CI 0.62–0.95; Ptrend¼ 0.003). High-fat dairy

intake not associated with T2DM incidence.

Whole milk positively associated with T2DM (RR Q4 1.19; 95%CI 1.00–1.43; Ptrend¼ 0.07)

Low-fat dairy: skim/ low-fat milk, sherbet, yogurt, cottage/ricotta cheese

High-fat dairy: whole milk, cream, sour cream, cream cheese and other cheeses

Forouhi et al. (2014)50 EPIC-InterAct case– cohort study Follow-up: 16 y

27 779 M/F 52.3 6 9.2 Plasma PL content in pentadeca-noic and heptadecapentadeca-noic acids inversely associated with incident T2DM (HR [95%CI] per 1 SD difference 0.79 [0.73–0.85] and 0.67 [0.63–0.741], respec-tively, Ptrend< 0.0001)

Butter, cheese, yogurt and thick milk, milk, dairy products

Fumeron et al. (2011)37

DESIR study, France Follow-up: 9 y

3435 M/F 30–64 Intake of dairy products inversely associated with T2DM and fast-ing glycemia (OR¼ 0.83; 95%CI 0.75–0.92, Ptrend¼ 0.001; OR¼ 0.85, 95%CI 0.76–0.94, Ptrend¼ 0.001) No Kirii et al. (2009)47

Japan Public Health Center-based Prospective Study Follow-up: 5 y

59 796 M/F 40–59 No association between dairy products, milk, cheese, or yogurt intake and T2DM in men. In women, dairy product intake

(300 g/d), but not milk, cheese, or yogurt intake, was inversely associated with T2DM (OR¼ 0.71; 95%CI 0.51–0.98; Ptrend¼ 0.054) No Krachler et al. (2008)49 Incident case-referent study from the Va¨sterbotten Invervention Program survey population, Sweden

Follow-up: 5.4 y for dia-betes cases, 8.8 y for referents Diabetes cases: 159 Referents: 291 M/F

51.7 6 7.7 Milk fat biomarkers in erythrocyte PL inversely associated with development of T2DM (OR¼ 0.54; 95%CI 0.35–0.83; Ptrend¼ 0.005) Whole milk: 3% Skim milk: 0.5% Malik et al. (2011)38

Nurses’ Health Study II, USA

Follow-up: 58 y

37 038 F 24–42 Current dairy intake:

High dairy intake inversely associ-ated with T2DM (RR 0.75, 95%CI 0.55–1.02; Ptrend¼ 0.03).

High-fat dairy intake inversely as-sociated with T2DM (RR 0.72, 95%CI 0.53–0.99; Ptrend¼ 0.03)

High-fat dairy: whole milk, whole chocolate milk, cream cheese, other cheeses, ice cream, milkshakes, and butter Margolis et al. (2011)39 Women’s Health Initiative observa-tional study, USA Follow-up: 3 y

82 076 F 50–79 High yogurt consumption (2 servings/wk) associated with significant decrease in T2DM risk (RR 0.52; 95%CI 0.42–0.64; Ptrend¼ 0.0001). No relationship

between high-fat dairy product and T2DM risk

Low-fat dairy: milk (nonfat milk, skim milk, 1%), cheese (low-fat cottage, part-skim, or reduced-fat cheese), nonfat yogurt, low-fat or nonfat frozen desserts Pereira et al. (2002)40 CARDIA study Follow-up: 10 y

3157 M/F 18–30 Incidence of abnormal glucose ho-meostasis significantly reduced in overweight individuals who consumed35 servings of dairy per week (OR¼ 0.35; 95%CI 0.16–0.66; Ptrend< 0.001)

Milk and milk drinks: whole, reduced fat Cheese and sour cream:

whole, reduced fat

(continued)

(10)

Table 2 Continued

Sluijs et al. (2012)43 EPIC-InterAct study, EU Follow-up: 11–13 y 16 154 M: 38% F: 62%

52 6 10 Total dairy not associated with T2DM (HR¼ 1.01; 95% CI: 0.83–1.34).

Cheese intake tended to have an inverse association with T2DM (HR¼ 0.88; 95%CI 0.76–1.02; Ptrend¼ 0.01) No Snijder et al. (2008)44

Hoorn Study, The Netherlands Follow-up: 6.4 y

1124 M/F 50–75 Dairy consumption not associated with changes in fasting or post-load glycemia (OR¼ 0.86; 95%CI 0.52–1.42)

Milk and yogurt: low-fat, skim, whole Soedamah-Muthu et al. (2012)46 Whitehall II cohort, UK Follow-up: 9.8 6 1.9 y 4526 M: 72% F: 28%

56 6 6 Total dairy not associated with T2DM (HR¼ 1.13; 95% CI 0.84– 1.51).

Intakes of high- and low-fat dairy, total milk, low-fat milk, fer-mented dairy products, yogurt, and cheese not associated with incident T2DM. Additional ad-justment for changes in BMI did not alter results

Low-fat dairy: cottage cheese, semi-skim, skim milk, and milk-based hot drinks

High-fat dairy: full-fat cheese, yogurt, milk puddings, whole and Channel Islands milk

Struijk et al. (2013)45 Inter99 study Denmark Follow-up: 5 y 5953 M: 47.5% F: 52.5%

45.8 6 7.8 Total dairy or any of the dairy sub-groups were not related to T2DM incidence, HOMA-IR, or HOMA-B.

Cheese intake (20 g/d) inversely associated with 2-h OGTT plasma glucose (b¼ 0.048; 95%CI0.095 to 0.001). Fermented dairy intake (150 g/d) inversely associated with fasting plasma glucose and HbA1c (b¼ 0.028; 95%CI 0.048 to 0.008)

Low-fat dairy, <2 g of fat per100 g of product; low-fat cheese, <20 g of low-fat per 100 g of product Full-fat dairy,2 g of fat

per 100 g of product; full-fat cheese,20 g of fat per 100 g of product

te Velde et al. (2011)48

Amsterdam Growth and Health Longitudinal Study, The Netherlands Follow-up: 23 y

374 M/F 13–36 Total dairy intake at age 14 y was higher in participants with high HbA1c levels at age 36 y com-pared with other participants. This appeared to be mainly a re-sult of differences in high-fat dairy intake at age 14 y, be-cause no differences in low-fat dairy intake were found be-tween the 2 groups of participants

Low-fat dairy:2% fat High-fat dairy: >2% fat

van Dam et al. (2006)41

Black Women’s Health Study, USA Follow-up: 8 y

41 186 F 21–69 (mean, 39)

High-fat dairy intake not related to T2DM incidence. Total dairy intake (2/d) inversely related to T2DM in the age-adjusted model (HR Q4¼ 0.75, 95%CI 0.61–0.93; Ptrend¼ 0.001), but

not in the multi-adjusted model

Milk: whole, 2%, 1%, or skim

Low-fat dairy: 2% fat High-fat dairy: all other

items

Warensjo¨ et al. (2004)51

Case-control study from the Va¨sterbotten Intervention Program and the Monitoring of Trends & Determinants in Cardiovascular Disease survey popu-lations, Sweden

Cases: 78 Controls:

156 M/F

Inverse correlation between pen-tadecanoic and heppen-tadecanoic acid content in serum PLs and fasting insulin, specific insulin, and proinsulin (r¼ 0.27 and 0.21, respectively)

(continued)

(11)

The majority of intervention studies compared a normal-fat (35%–40% of total energy intake, which is the adequate intake according to the French guide-lines57), high-saturated-fat (>15% of total energy in-take) diet with a normal-fat, low-saturated-fat (<10% of total energy intake), high-unsaturated-fat diet.25–34 Despite these comparable approaches, the results were heterogeneous. Four25,29,33,34of the 1025–34clinical trials demonstrated a deleterious effect of an intake of satu-rated fat higher than 15% of total energy intake on glu-cose-insulin homeostasis. This is corroborated by the results of 4 of 7 prospective studies17–20but not by the 3 larger prospective studies21–23or the pooled analysis of prospective cohort studies of Micha and Mozaffarian,58 who concluded that consumption of saturated fat is not associated with the onset of T2DM (pooled relative risk ¼ 0.98 [95%CI: 0.87–1.10]). Furthermore, 6 inter-vention studies showed that, in overweight individuals with26–28 or without30–32 features of metabolic syn-drome, decreasing saturated fat to approximately 10% in comparison with approximately 15% of energy intake does not improve glucose-insulin homeostasis, as as-sessed by an oral glucose tolerance test. Thus, most of

the prospective and intervention studies demonstrate that dietary saturated fat at levels ranging from 13% to more than 20% of energy intake is not associated with an increased risk of T2DM.

Contradictory results in prospective studies could be due in part to the lack of adjustment for differences in the quantity and the quality of carbohydrate intake. This is supported by findings from dietary intervention stud-ies, which showed that saturated fat and carbohydrates have similar effects on key risk factors of T2DM. Indeed, 7 intervention studies24,26–29,33,34compared a normal-fat, high-saturated-fat diet with a low-fat (25%–29% of total energy intake), high-carbohydrate (50%–65% of total en-ergy intake) diet. All showed that high-saturated-fat and high-carbohydrate diets have similar effects on insulin sensitivity,26,27on markers of insulin sensitivity (plasma glucose and insulin, HOMA-IR score),26–29,33,34and on risk of T2DM.24Furthermore, recent meta-analyses have shown that a high glycemic index and a high glycemic load were associated with an increased risk of chronic diseases, such as T2DM, independent of dietary fat in-takes.59,60The mechanism potentially involves postpran-dial hyperglycemia60and is still under investigation.

Table 2 Continued

Warensjo¨ et al. (2010)52

Case-control study from the Northern Sweden Health & Disease Study populations, Sweden MI cases: 444 Controls: 556 M/F

49–63 Inverse correlation between pen-tadecanoic and heppen-tadecanoic acid content in serum PLs and fasting glucose (r¼ 0.085)

Cheese: 17% and 28% fat Milk: skim (0.5%), low-fat (1.5%), and full-fat (3%)

Intervention studies Benatar et al. (2014)55

Auckland, New Zealand Healthy normal-weight (BMI 24.5 6 4.0) and normotensive (BP 110/70 6 10/8 mmHg) subjects Follow-up: 1 mo 180 M: 30% F: 70% 38–55 (mean, 47) No effect of increasing (þ12.5 6 15.7 g/d), reducing (10.4 6 10.1 g/d), or maintaing (3.4 6 7.9 g/d) dairy fat in-take on fasting glucose or insulin

Fat content calculated from manufacturers’ labels Stancliffe et al. (2011)53 USA

Overweight and obese subjects with3 MetS risk factors Follow-up: 12 wk

20 persons per group Obese: 50% M/F

37.0 6 9.9 Adequate dairy intake (>3.5 serv-ings/d) did not affect blood glu-cose in overweight, obese, or all subjects compared with low dairy intake (<0.5 serving/d)

No Wennersb-erg et al. (2009)54 Finland, Norway, Sweden Overweight subjects with3 MetS risk factors Follow-up: 6 mo 54–57 persons per group M/F

30–65 High dairy intake did not affect blood glucose levels. It lowered blood insulin levels compared with levels in controls, but not compared with baseline levels

Milk: 0.5%–3% fat Yogurt or sour milk:

1%–5.4% fat Cream or cre`me fraiche:

12%–40% fat Cheese: 15%–30% fat Butter: 40%–80% fat Cottage cheese: 2%–8% fat

a

Age at baseline for prospective studies.

Abbreviations: BMI, body mass index; BP, blood pressure; CI, confidence interval; HOMA-B, homeostasis model assessment–b-cell func-tion; HOMA-IR, homeostasis model assessment–estimated insulin resistance; HR, hazard ratio; MetS, metabolic syndrome; MI, myocar-dial infarction; OGTT, oral glucose tolerance test; OR, odds ratio; L, phospholipids; Q, quintile; RR, relative risk; SD, standard deviation; T2DM, type 2 diabetes mellitus.

(12)

The lack of consensus may be due in part to the dif-ferent populations that were examined, which ranged from healthy subjects to individuals with metabolic syndrome or high cardiometabolic risks, in whom sen-sitivity to saturated fat could differ, as suggested by Micha and Mozaffarian.58 Some genetic determinants of obesity could also affect the individual response to di-etary lipids, especially the response to saturated fats. Using data from the LIPGENE-SU.VI.MAX study, Phillips et al.61,62demonstrated that some FTO, ApoB, and ApoA1 genotypes exacerbate the metabolic re-sponse to dietary lipids in predisposed subjects, thus in-creasing the risk for obesity, insulin resistance, and T2DM.

Cardiovascular studies have shown that saturated fatty acids are differentially associated with risks of met-abolic disorders, depending on the number of carbon atoms,58,63,64 thus indicating that saturated fatty acids should not all be considered metabolically equivalent. As far as can be determined, the clinical trials that ex-amined the association between individual saturated fatty acids and risk of insulin resistance or T2DM did not consider dietary fat intakes but explored the fatty acid composition of serum phospholipids, erythrocytes, or skeletal muscles.65–69

Interestingly, in a recent study of 3004 individuals, Ma et al.70 showed that circulating concentrations of the main saturated fatty acids, palmitic acid and stearic acid, were associated with increased risk of T2DM. Compared with findings in the lowest quintile, the mul-tivariate-adjusted HR (95%CI) in the highest quintile was 1.89 (1.27–2.83; P trend ¼ 0.001) for palmitic acid and 1.62 (1.09–2.41; P trend ¼ 0.006) for stearic acid. However, after additional adjustment for circulating markers of metabolic risk (high-density lipoprotein cholesterol ratio, plasma triglycerides, C-reactive pro-tein, fibrinogen, and HOMA-IR), the association of pal-mitic acid with T2DM was only slightly attenuated, whereas the association of stearic acid with T2DM was attenuated and no longer significant.70As described in the Methods section, altered concentrations of circulat-ing saturated fatty acids may result from changes in he-patic metabolism rather than from inadequate dietary fat intakes.14,15 Supporting this hypothesis, Ma et al.70 showed that circulating saturated fatty acids were weakly correlated with the estimated dietary intake of saturated fatty acids. Furthermore, they found that, in contrast to circulating concentrations of saturated fatty acids, dietary intakes of saturated fatty acids were not significantly associated with incident diabetes in a co-hort of 4221 individuals. The HR (95%CI) in the highest quintile was 1.14 (0.72–1.79; P trend ¼ 0.55) for pal-mitic acid and 1.00 (0.62–1.62; P trend ¼ 0.93) for stea-ric acid.70

Association between dairy fat intake and insulin resistance or T2DM

A considerable number of prospective studies allow ex-ploration of the association between dietary dairy intake and insulin resistance or T2DM, but very few interven-tion studies are available to support the conclusions.

If the prospective studies that specifically examined intake of dairy fat are considered, half of the prospective and case–control studies found an inverse association be-tween dairy fat intake and T2DM incidence,38,49–52and the other half found no association.39,41,42,46,47These con-clusions are consistent with those of the prospective stud-ies that examined dairy intake, whatever the fat content. Indeed, of the 15 prospective studies analyzed, only 1 identified a positive relation between dairy intake at the age of 14 years and incidence of high HbA1c levels in adulthood.48Of the other studies, half found an inverse relation between total dairy consumption and T2DM in-cidence,37,38,40–42and half found no association.43–47

Nevertheless, evidence from randomized controlled trials is limited, since only 3 intervention studies were available.53–55Furthermore, overall, those studies provide only modest support that enhancing dairy consumption may help prevent features of T2DM. The lack of evidence may be attributable to the short duration of these inter-ventions (1,553,53and 654 months), the low number of subjects (20–60 persons per group), and the differences in dairy matrices and nutritional designs (seeTable S3in the Supporting Information online).

The majority of the studies included in this review demonstrate either no association or an inverse associa-tion between dairy intake and risk of T2DM, with only 1 study reporting a positive association.48 Therefore, taken together, combined data from the present review suggest that the constituents of dairy products do not alter – and may even reduce – the risk of T2DM. The recent conclusions from the large EPIC-InterAct case– cohort study showed an inverse association between dairy fat consumption, assessed from the proportions of pentadecanoic and heptadecanoic acids in plasma phos-pholipids, and the incidence of T2DM (HR [95%CI] per 1 standard deviation difference: 0.79 [0.73–0.85] for pentadecanoic acid and 0.67 [0.63–0.71] for heptadeca-noic acid, P trend < 0.0001).50Furthermore, other com-ponents of the dairy matrix (in addition to short- and medium-chain fatty acids and stearic acid), such as cal-cium, magnesium and other minerals, bioactive pep-tides and amino acids, vitamins B and D, and probiotics, may also contribute to the inverse relation often described between total dairy intake and features of T2DM.71 The health benefits of these nutrients and probiotics have been the topic of meta-analyses72–75and reviews.76,77

(13)

Most studies showed that dairy fat intake does not alter blood glucose. In contrast, the effect of dairy prod-uct intake on insulin secretion could be a matter of de-bate. Indeed, milk products deviate from other carbohydrate-containing foods in that they produce high insulin responses, despite their low glycemic index. Recent short-term intervention trials have demon-strated that the insulinotropic effect of milk is related to its protein composition rather than to its lipid con-tent.78,79 Furthermore, prospective studies show that the insulinotropic effect is not observed over the long term.45

CONCLUSION

The present review provides additional support to ear-lier studies that found no evidence that saturated fat in moderate amounts affects the risk of insulin resistance or T2DM. There is also increasing evidence to support the conclusion that full-fat dairy products neither alter nor decrease the risks of T2DM. Nevertheless, it is pru-dent to discourage saturated fat intake that exceeds 15% of energy intake since it might alter the glucose–insulin homeostasis in comparison with diets that are high in mono- and polyunsaturated fats.

The effect of specific foods on the risk of insulin resis-tance and T2DM cannot be predicted solely from their saturated fat content because individual saturated fatty acids have different health effects, and some main food sources of saturated fat contain other constituents that also could influence metabolic risk factors. Indeed, foods are complex matrices in which different components have different effects on human health, depending on whether they are consumed alongside other nutrients or in isola-tion.80,81This statement of common sense calls for caution in drawing conclusions about a compound or a specific nutrient outside the context of the diet as a whole. Acknowledgments

Author contributions. B.M. performed the literature search, extracted and analyzed the data, wrote the man-uscript, and was the guarantor of the manuscript; A.F. contributed to the draft of the manuscript and gave fi-nal approval of the manuscript; and P.L. and J.M.L. ana-lyzed the literature, contributed to the draft of the manuscript, and gave final approval of the manuscript. Financial disclosures. The present review was supported by the authors’ institutions, INRA and Pasteur Institute of Lille. No external funding was received.

Declaration of interest. The authors have no relevant in-terests to declare.

SUPPORTING INFORMATION

The following Supporting Information is available through the online version of this article at the publish-er’s website:

Table S1 Description of the included papers examining the relationship between dietary fat quality and insulin resistance or type 2 diabetes (T2DM).

Table S2 Description of the included papers examining the relationship between dairy consumption and insulin resistance or type 2 diabetes (T2DM).

Table S3 Methods used for assessing dairy fat intakes in the included studies.

REFERENCES

1. Uauy R, Dı´az E. Consequences of food energy excess and positive energy balance. Public Health Nutr. 2005;8:1077–1099.

2. World Health Organization. Global strategy on diet, physical activity and health. http://www.who.int/dietphysicalactivity/strategy/eb11344/strategy_english_web. pdf. Published 2004. Accessed January 2014.

3. Boden G. Role of fatty acids in the pathogenesis of insulin resistance and NIDDM. Diabetes. 1997;46:3–10.

4. Food and Agricultural Organization of the United Nations. Fats and fatty acids in hu-man nutrition. Report of an expert consultation. http://foris.fao.org/preview/25553-0ece4cb94ac52f9a25af77ca5cfba7a8c.pdf. Published 2010. Accessed January 2014. 5. Astrup A, Dyerberg J, Elwood P, et al. The role of reducing intakes of saturated fat

in the prevention of cardiovascular disease: where does the evidence stand in 2010? Am J Clin Nutr. 2011;93:684–688.

6. Nordmann AJ, Nordmann A, Briel M, et al. Effects of low-carbohydrate vs low-fat diets on weight loss and cardiovascular risk factors: a meta-analysis of randomized controlled trials. Arch Intern Med. 2006;166:285–293.

7. Kirk JK, Graves DE, Craven TE, et al. Restricted-carbohydrate diets in patients with type 2 diabetes: a meta-analysis. J Am Diet Assoc. 2008;108:91–100.

8. Kodama S, Saito K, Tanaka S, et al. Influence of fat and carbohydrate proportions on the metabolic profile in patients with type 2 diabetes: a meta-analysis. Diabetes Care. 2009;32:959–965.

9. Hu FB. Are refined carbohydrates worse than saturated fat? Am J Clin Nutr. 2010;91:1541–1542.

10. Forsythe CE, Phinney SD, Feinman RD, et al. Limited effect of dietary saturated fat on plasma saturated fat in the context of a low carbohydrate diet. Lipids. 2010;45:947–962.

11. Aune D, Norat T, Romundstad P, et al. Dairy products and the risk of type 2 diabe-tes: a systematic review and dose-response meta-analysis of cohort studies. Am J Clin Nutr. 2013;98:1066–1083.

12. Mozaffarian D, de Oliveira Otto MC, Lemaitre RN, et al. trans-Palmitoleic acid, other dairy fat biomarkers, and incident diabetes: the Multi-Ethnic Study of Atherosclerosis (MESA). Am J Clin Nutr. 2013;97:854–861.

13. Wolk A, Furuheim M, Vessby B. Fatty acid composition of adipose tissue and se-rum lipids are valid biological markers of dairy fat intake in men. J Nutr. 2001;131:828–833.

14. Micha R, King IB, Lemaitre R, et al. Food sources of individual plasma phospholipid trans fatty acid isomers: the Cardiovascular Health Study. Am J Clin Nutr. 2010;91:883–893.

15. Chong MF, Hodson L, Bickerton AS, et al. Parallel activation of de novo lipogenesis and stearoyl-CoA desaturase activity after 3 d of high-carbohydrate feeding. Am J Clin Nutr. 2008;87:817–823.

16. Wilke MS, French MA, Goh YK, et al. Synthesis of specific fatty acids contributes to VLDL-triacylglycerol composition in humans with and without type 2 diabetes. Diabetologia. 2009;52:1628–1637.

17. Feskens EJ, Kromhout D. Habitual dietary intake and glucose tolerance in eugly-caemic men: the Zutphen Study. Int J Epidemiol. 1990;19:953–959.

18. Feskens EJ, Virtanen SM, Ra¨sa¨nen L, et al. Dietary factors determining diabetes and impaired glucose tolerance. A 20-year follow-up of the Finnish and Dutch co-horts of the Seven Countries Study. Diabetes Care. 1995;18:1104–1112. 19. Mayer EJ, Newman B, Quesenberry CP Jr, et al. Usual dietary fat intake and

insulin concentrations in healthy women twins. Diabetes Care. 1993;16: 1459–1469.

20. Marshall JA, Hoag S, Shetterly S, et al. Dietary fat predicts conversion from im-paired glucose tolerance to NIDDM. The San Luis Valley Diabetes Study. Diabetes Care. 1994;17:50–56.

(14)

21. van Dam RM, Willett WC, Rimm EB, et al. Dietary fat and meat intake in relation to risk of type 2 diabetes in men. Diabetes Care. 2002;25:417–424.

22. Meyer KA, Kushi LH, Jacobs DR Jr, et al. Dietary fat and incidence of type 2 diabe-tes in older Iowa women. Diabediabe-tes Care. 2001;24:1528–1535.

23. Salmero´n J, Hu FB, Manson JE, et al. Dietary fat intake and risk of type 2 diabetes in women. Am J Clin Nutr. 2001;73:1019–1026.

24. Tinker LF, Bonds DE, Margolis KL, et al. Women’s Health Initiative. Low-fat dietary pattern and risk of treated diabetes mellitus in postmenopausal women: the Women’s Health Initiative randomized controlled dietary modification trial. Arch Intern Med. 2008;168:1500–1511.

25. Vessby B, Uusitupa M, Hermansen K, et al. Substituting dietary saturated for monounsaturated fat impairs insulin sensitivity in healthy men and women: the KANWU Study. Diabetologia. 2001;44:312–319.

26. Jebb SA, Lovegrove JA, Griffin BA, et al. Effect of changing the amount and type of fat and carbohydrate on insulin sensitivity and cardiovascular risk: the RISCK (Reading, Imperial, Surrey, Cambridge, and Kings) trial. Am J Clin Nutr. 2010;92:748–758.

27. Tierney AC, McMonagle J, Shaw DI, et al. Effects of dietary fat modification on insulin sensitivity and on other risk factors of the metabolic syndrome – LIPGENE: a European randomized dietary intervention study. Int J Obes. 2011;35:800–809.

28. Berglund L, Lefevre M, Ginsberg HN, et al. Comparison of monounsaturated fat with carbohydrates as a replacement for saturated fat in subjects with a high met-abolic risk profile: studies in the fasting and postprandial states. Am J Clin Nutr. 2007;86:1611–1620.

29. Paniagua JA, de la Sacristana AG, Sa´nchez E, et al. A MUFA-rich diet improves pos-prandial glucose, lipid and GLP-1 responses in insulin-resistant subjects. J Am Coll Nutr. 2007;26:434–444.

30. Bos MB, de Vries JH, Feskens EJ, et al. Effect of a high monounsaturated fatty acids diet and a Mediterranean diet on serum lipids and insulin sensitivity in adults with mild abdominal obesity. Nutr Metab Cardiovasc Dis. 2010;20:591–598. 31. Haghighatdoost F, Hosseinzadeh-Attar MJ, Kabiri A, et al. Effect of substituting

sat-urated with monounsatsat-urated fatty acids on serum visfatin levels and insulin resis-tance in overweight women: a randomized cross-over clinical trial. Int J Food Sci Nutr. 2012;63:772–781.

32. Ortega JF, Ferna´ndez-Elı´as VE, Hamouti N, et al. Increased blood cholesterol after a high saturated fat diet is prevented by aerobic exercise training. Appl Physiol Nutr Metab. 2013;38:42–48.

33. Due A, Larsen TM, Hermansen K, et al. Comparison of the effects on insulin resis-tance and glucose tolerance of 6-mo high-monounsaturated-fat, low-fat, and con-trol diets. Am J Clin Nutr. 2008;87:855–862.

34. Due A, Larsen TM, Mu H, et al. Comparison of 3 ad libitum diets for weight-loss maintenance, risk of cardiovascular disease, and diabetes: a 6-mo randomized, controlled trial. Am J Clin Nutr. 2008;88:1232–1241.

35. Masson CJ, Mensink RP. Exchanging saturated fatty acids for (n-6) polyunsaturated fatty acids in a mixed meal may decrease postprandial lipemia and markers of inflammation and endothelial activity in overweight men. J Nutr. 2011;141:816–821. 36. Lopez S, Bermudez B, Ortega A, et al. Effects of meals rich in either

monounsatu-rated or satumonounsatu-rated fat on lipid concentrations and on insulin secretion and action in subjects with high fasting triglyceride concentrations. Am J Clin Nutr. 2011;93:494–499.

37. Fumeron F, Lamri A, Emery N, et al. Dairy products and the metabolic syndrome in a prospective study, DESIR. J Am Coll Nutr. 2011;30(5 suppl 1):454S–463S. 38. Malik VS, Sun Q, van Dam RM, et al. Adolescent dairy product consumption and

risk of type 2 diabetes in middle-aged women. Am J Clin Nutr. 2011;94:854–861. 39. Margolis KL, Wei F, de Boer IH, et al. A diet high in low-fat dairy products lowers

diabetes risk in postmenopausal women. J Nutr. 2011;141:1969–1974. 40. Pereira MA, Jacobs DR Jr, Van Horn L, et al. Dairy consumption, obesity, and the

insulin resistance syndrome in young adults: the CARDIA Study. JAMA. 2002;287:2081–2089.

41. van Dam RM, Hu FB, Rosenberg L, et al. Dietary calcium and magnesium, major food sources, and risk of type 2 diabetes in U.S. black women. Diabetes Care. 2006;29:2238–2243.

42. Choi HK, Willett WC, Stampfer MJ, et al. Dairy consumption and risk of type 2 dia-betes mellitus in men: a prospective study. Arch Intern Med. 2005;165:997–1003. 43. Sluijs I, Forouhi NG, Beulens JW, et al. The amount and type of dairy product

in-take and incident type 2 diabetes: results from the EPIC-InterAct Study. Am J Clin Nutr. 2012;96:382–390.

44. Snijder MB, van Dam RM, Stehouwer CD, et al. A prospective study of dairy con-sumption in relation to changes in metabolic risk factors: the Hoorn Study. Obesity. 2008;16:706–709.

45. Struijk EA, Heraclides A, Witte DR, et al. Dairy product intake in relation to glucose regulation indices and risk of type 2 diabetes. Nutr Metab Cardiovasc Dis. 2013;23:822–828.

46. Soedamah-Muthu SS, Masset G, Verberne L, et al. Consumption of dairy products and associations with incident diabetes, CHD and mortality in the Whitehall II study. Br J Nutr. 2012;7:1–9.

47. Kirii K, Mizoue T, Iso H, et al. Calcium, vitamin D and dairy intake in relation to type 2 diabetes risk in a Japanese cohort. Diabetologia. 2009;52:2542–2550.

48. te Velde SJ, Snijder MB, van Dijk AE, et al. Dairy intake from adolescence into adulthood is not associated with being overweight and metabolic syndrome in adulthood: the Amsterdam Growth and Health Longitudinal Study. J Hum Nutr Diet. 2011;24:233–244.

49. Krachler B, Norberg M, Eriksson JW, et al. Fatty acid profile of the erythrocyte membrane preceding development of type 2 diabetes mellitus. Nutr Metab Cardiovasc Dis. 2008;18:503–510.

50. Forouhi NG, Koulman A, Sharp SJ, et al. Differences in the prospective association between individual plasma phospholipid saturated fatty acids and incident type 2 diabetes: the EPIC-InterAct case-cohort study. Lancet Diabetes Endocrinol. 2014;2:810–818.

51. Warensjo¨ E, Jansson JH, Berglund L, et al. Estimated intake of milk fat is negatively associated with cardiovascular risk factors and does not increase the risk of a first acute myocardial infarction. A prospective case-control study. Br J Nutr. 2004;91:635–642.

52. Warensjo¨ E, Jansson JH, Cederholm T, et al. Biomarkers of milk fat and the risk of myocardial infarction in men and women: a prospective, matched case-control study. Am J Clin Nutr. 2010;92:194–202.

53. Stancliffe RA, Thorpe T, Zemel MB. Dairy attentuates oxidative and inflammatory stress in metabolic syndrome. Am J Clin Nutr. 2011;94:422–430.

54. Wennersberg MH, Smedman A, Turpeinen AM, et al. Dairy products and meta-bolic effects in overweight men and women: results from a 6-mo intervention study. Am J Clin Nutr. 2009;90:960–968.

55. Benatar JR, Jones E, White H, et al. A randomized trial evaluating the effects of change in dairy food consumption on cardio-metabolic risk factors. Eur J Prev Cardiol. 2014;21:1376–1386.

56. EFSA Panel on Dietetic Products, Nutrition, and Allergies (NDA). Scientific Opinion on Dietary Reference Values for fats, including saturated fatty acids, polyunsatu-rated fatty acids, monounsatupolyunsatu-rated fatty acids, trans fatty acids, and cholesterol. EFSA J. 2010;8:1461. doi:10.2903/j.efsa.2010.1461.

57. Agence Nationale de Se´curite´ Sanitaire Alimentation, Environnement, Travail. Actualisation des apports nutritionnels conseille´s pour les acides gras. Rapport d’expertise collective. Maisons-Alfort, France: ANSES; Mai 2011. Saisine no. 2006-SA-0359.

58. Micha R, Mozaffarian D. Saturated fat and cardiometabolic risk factors, coronary heart disease, stroke, and diabetes: a fresh look at the evidence. Lipids. 2010;45:893–905.

59. Alhazmi A, Stojanovski E, McEvoy M, et al. Macronutrient intakes and develop-ment of type 2 diabetes: a systematic review and meta-analysis of cohort studies. J Am Coll Nutr. 2012;31:243–258.

60. Barclay AW, Petocz P, McMillan-Price J, et al. Glycemic index, glycemic load, and chronic disease risk – a meta-analysis of observational studies. Am J Clin Nutr. 2008;87:627–637.

61. Phillips CM, Goumidi L, Bertrais S, et al. Gene-nutrient interactions and gender may modulate the association between ApoA1 and ApoB gene polymorphisms and metabolic syndrome risk. Atherosclerosis. 2011;214:408–414.

62. Phillips CM, Kesse-Guyot E, McManus R, et al. High dietary saturated fat intake ac-centuates obesity risk associated with the fat mass and obesity-associated gene in adults. J Nutr. 2012;142:824–831.

63. Kris-Etherton P, Yu S. Individual fatty acids on plasma lipids and lipoproteins: hu-man studies. Am J Clin Nutr. 1997;65(5 suppl):1628S–1644S.

64. Temme E, Mensink RP, Hornstra G. Comparison of the effects of diets enriched in lauric, palmitic, or oleic acids on serum lipids and lipoproteins in healthy women and men. Am J Clin Nutr. 1996;63:897–903.

65. Kusunoki M, Tsutsumi K, Nakayama M, et al. Relationship between serum concen-trations of saturated fatty acids and unsaturated fatty acids and the homeostasis model insulin resistance index in Japanese patients with type 2 diabetes mellitus. J Med Invest. 2007;54:243–247.

66. Lovejoy JC, Champagne CM, Smith SR, et al. Relationship of dietary fat and se-rum cholesterol ester and phospholipid fatty acids to markers of insulin resistance in men and women with a range of glucose tolerance. Metabolism. 2001;50:86–92. 67. Pelika´nova´ T, Kazdova´ L, Chvojkova´ S, et al. Serum phospholipid fatty acid composition and insulin action in type 2 diabetic patients. Metabolism. 2001;50:1472–1478.

68. Vessby B, Tengblad S, Lithell H. Insulin sensitivity is related to the fatty acid com-position of serum lipids and skeletal muscle phospholipids in 70-year-old men. Diabetologia. 1994;37:1044–1050.

69. Borkman M, Storlien LH, Pan DA, et al. The relation between insulin sensitivity and the fatty-acid composition of skeletal-muscle phospholipids. N Engl J Med. 1993;328:238–244.

70. Ma W, Wu JH, Wang Q, et al. Prospective association of fatty acids in the de novo lipogenesis pathway with risk of type 2 diabetes: the Cardiovascular Health Study. Am J Clin Nutr. 2015;101:153–163.

71. Fardet A, Boirie Y. Associations between diet-related diseases and impaired physi-ological mechanisms: a holistic approach based on meta-analyses to identify tar-gets for preventive nutrition. Nutr Rev. 2013;71:643–656.

72. Soedamah-Muthu SS, Verberne LD, Ding EL, et al. Dairy consumption and inci-dence of hypertension: a dose-response meta-analysis of prospective cohort stud-ies. Hypertension. 2012;60:1131–1137.

(15)

73. Chen M, Pan A, Malik VS, et al. Effects of dairy intake on body weight and fat: a meta-analysis of randomized controlled trials. Am J Clin Nutr. 2012;96:735–747.

74. Pittas AG, Lau J, Hu FB, et al. The role of vitamin D and calcium in type 2 diabetes. A systematic review and meta-analysis. J Clin Endocrinol Metab. 2007;92:2017–2029.

75. Agerholm-Larsen L, Bell ML, Grunwald GK, et al. The effect of a probiotic milk product on plasma cholesterol: a meta-analysis of short-term intervention studies. Eur J Clin Nutr. 2000;54:856–860.

76. Tremblay A, Gilbert JA. Milk products, insulin resistance syndrome and type 2 dia-betes. J Am Coll Nutr. 2009;28(suppl 1):91S–102S.

77. Pfeuffer M, Schrezenmeir J. Milk and the metabolic syndrome. Obes Rev. 2007;8:109–118.

78. Arnberg K, Mølgaard C, Michaelsen KF, et al. Skim milk, whey, and casein increase body weight and whey and casein increase the plasma C-peptide concentration in overweight adolescents. J Nutr. 2012;142:2083–2090.

79. Gunnerud U, Holst JJ, O¨stman E, et al. The glycemic, insulinemic and plasma amino acid responses to equi-carbohydrate milk meals, a pilot-study of bovine and human milk. Nutr J. 2012;11:83. doi:10.1186/1475-2891-11-83.

80. Burton-Freeman BM, Sesso HD. Whole food versus supplement: comparing the clinical evidence of tomato intake and lycopene supplementation on cardiovascu-lar risk factors. Adv Nutr. 2014;5:457–485.

81. Fardet A, Rock E. Toward a new philosophy of preventive nutrition: from a reduc-tionist to a holistic paradigm to improve nutritional recommendations. Adv Nutr. 2014;5:430–446.