Sei sulla pagina 1di 9

The Journal of Nutrition. First published ahead of print October 28, 2015 as doi: 10.3945/jn.115.220699.

The Journal of Nutrition


Nutritional Epidemiology

Total and Full-Fat, but Not Low-Fat, Dairy


Product Intakes are Inversely Associated with
Metabolic Syndrome in Adults1,2
Michele Drehmer,3,5* Mark A Pereira,6 Maria Ines Schmidt,4,5 Sheila Alvim,7 Paulo A Lotufo,8
Vivian C Luft,3,5 and Bruce B Duncan4,5
3
Department of Nutrition and Food and Nutrition Research Center, 4Department of Social Medicine, and 5Postgraduate Program in
Epidemiology, School of Medicine, Federal University of Rio Grande, Porto Alegre, Brazil; 6Division of Epidemiology and Community
Health, School of Public Health, University of Minnesota, Minneapolis, MN; 7Institute of Collective Health, Federal University of Bahia,
Salvador, Brazil; and 8Center for Clinical and Epidemiologic Research, University of Sao Paulo, Sao Paulo, Brazil

Abstract
Background: Growing evidence suggests that dairy products may have beneficial cardiometabolic effects. The current

Downloaded from jn.nutrition.org by guest on February 23, 2017


guidelines, however, limit the intake of full-fat dairy products.
Objective: We investigated the association of dairy consumption, types of dairy products, and dairy fat content with
metabolic syndrome (MetSyn).
Methods: We analyzed baseline data of the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil), a multicenter cohort
study of 15,105 adults aged 3574 y. We excluded participants with known diabetes, cardiovascular diseases, or other
chronic diseases, and those who had extreme values of energy intake, leaving 9835 for analysis. Dairy consumption was
assessed by a food-frequency questionnaire. We computed servings per day for total and subgroups of dairy intake. We
computed a metabolic risk score (MetScore) as the mean z score of waist circumference, systolic blood pressure, HDL
cholesterol (negative z score), fasting triglycerides, and fasting glucose. We performed multivariable linear regression to
test the association of servings per day of dairy products with MetScore.
Results: In analyses that adjusted for demographics, menopausal status, family history of diabetes, dietary intake,
nondietary lifestyle factors, and body mass index, we observed a graded inverse association for MetScore with total dairy
(20.044 6 0.01, P = 0.009 for each additional dairy servings per day) and full-fat dairy (20.126 6 0.03, P < 0.001) but not
with low-fat dairy intake. Associations were no longer present after additional adjustments for dairy-derived saturated
fatty acids.
Conclusions: Total and especially full-fat dairy food intakes are inversely and independently associated with metabolic
syndrome in middle-aged and older adults, associations that seem to be mediated by dairy saturated fatty acids. Dietary
recommendations to avoid full-fat dairy intake are not supported by our findings. J Nutr doi: 10.3945/jn.115.220699.

Keywords: metabolic syndrome, dairy consumption, saturated fatty acids, cohort study, diabetes

Introduction
Metabolic syndrome (MetSyn)9 is a clustering of interrelated
Noncommunicable diseases are the main cause of morbidity, risk factors for cardiovascular disease (CVD) and type 2 diabetes
disability, and health care costs, accounting for nearly two-thirds mellitus, which occur together more often than by chance alone,
of deaths worldwide. Cardiovascular disease is the leading cause including high blood pressure, low fasting HDL cholesterol, high
of death in the United States and globally (1, 2). fasting TGs, high fasting blood glucose, and abdominal obesity
(3). The consumption of dairy products and associated nutrients
1
Supported by the Brazilian Ministry of Health (Science and Technology has been suggested to improve MetSyn characteristics (4, 5).
Department) and the Brazilian Ministry of Science, Technology and Innovation
Evidence that supports a beneficial effect of dairy products on
(Funding Authority for Studies and Projects and the National Council for Scientific
and Technological Development) grants 01 06 0010.00 RS, 01 06 0212.00 BA, 01 CVD risk factors came initially from the Dietary Approaches to
06 0300.00 ES, 01 06 0278.00 MG, 01 06 0115.00 SP, 01 06 0071.00 RJ, and Stop Hypertension trial (6). A meta-analysis of randomized
postdoctoral grant 249320/2013.
2
Author disclosures: M Drehmer, MA Pereira, MI Schmidt, S Alvim, PA Lotufo,
9
VC Luft, and BB Duncan, no conflicts of interest. Abbreviations used: CVD, cardiovascular disease; ELSA-Brasil, Brazilian
* To whom correspondence should be addressed. E-mail: michele.drehmer@ Longitudinal Study of Adult Health; MetScore, metabolic risk score; MetSyn,
ufrgs.br. metabolic syndrome.

2016 American Society for Nutrition.


Manuscript received July 15, 2015. Initial review completed August 16, 2015. Revision accepted October 8, 2015. 1 of 9
doi: 10.3945/jn.115.220699.
Copyright (C) 2015 by the American Society for Nutrition
studies found small effects of moderate increases in dairy products for cheese and cream cheese, 80 g for ice cream, 50 g for dairy dessert
on cardiometabolic risk factors (7). A main limitation is that no (pudding and mousse), and 5 g for butter.
trials, to our knowledge, have directly compared the effect of low- The nondairy food groups were computed in the analysis by grams
and full-fat dairy intake on cardiometabolic risk (7). Dairy products per day. As covariates, we assessed the consumption of fruits, vegetables,
whole and refined grains, nondairy sweets and desserts, processed and
are complex and generally dense in nutrients, and the effects of
unprocessed red and white meats, eggs, beans, coffee, tea, juice, and
their saturated fat content on MetSyn needs to be elucidated (6, 8). sodas.
Our recent findings from the Brazilian Longitudinal Study of We used the University of Minnesota Nutrition Data System for
Adult Health (ELSA-Brasil) study suggested that the intake of Research database (14) to determine the nutritional composition of the
total dairy products, especially from fermented products, was food items from the FFQ based on the following equation:
inversely associated with measures of glycemia and insulinemia,
size of serving in standard portions 3 frequency of consumption
and the SFA myristic acid may have mediated this association 3 weight of a standard portion in grams
(9). Data from 2 other large cohort studies suggested that circu- 3 nutritional composition of the food serving
lating biomarkers of dairy fat were not significantly associated grams of nutrient per total grams of food item 1
with stroke and that SFAs, especially those that originate from In this equation, the values used for the frequency categories were as
dairy, do not increase CVD risk (5, 10). Considering the high follows: 3 for >3 times/d, 2.5 for 23 times/d, 1 for once per day, 0.8
prevalence of dairy consumption, even in low- and middle- for 56 times/wk, 0.4 for 24 times/wk, 0.1 for once per week, 0.07 for
income countries (11), and the recommendation to limit full- 13 times/mo, and 0 for never/almost never.
fat dairy intake (12), more studies are needed to evaluate the The following list of dairy nutrients and components were evaluated
association of dairy intake and its components with cardio- as possible confounders or mediators: animal protein (percentage energy
vascular risk. This study sought to investigate the association per day), lactose (g/d), sugar (g/d), total fat (percentage energy per day),
saturated fat from dairy products (percentage energy per day), choles-
between dairy consumption, types of dairy products, including
terol (mg/d), MUFA (percentage energy per day), PUFA (percentage
fat content, and MetSyn at the baseline of a large multicentric energy per day), linoleic FAs (percentage energy per day), vitamin A (IU),
cohort study.

Downloaded from jn.nutrition.org by guest on February 23, 2017


vitamin D (mcg/d), calcium (mg/d), phosphorus (mg), magnesium (mg/d),
and sodium (mg/d). We evaluated butyric, caproic, caprylic, capric,
lauric, myristic, palmitic, stearic, and palmitoleic FAs as percentages of
Methods energy per day.
Study design Anthropometrics. Anthropometric measures (weight, height, and waist
ELSA-Brasil is a multicenter cohort study designed to address the circumference) were obtained from participants while they were
incidence of CVDs and diabetes and major associated risk factors. A standing, dressed in a light uniform standardized for the study, and
total of 15,105 civil servants from selected universities and research without shoes after an 815-h fast. We measured body weight to the
institutions between the ages of 35 and 74 y living in 6 cities (Salvador, nearest 0.1 kg with a calibrated Toledo 2096PP balance (Toledo do Brasil
Belo Horizonte, Rio de Janeiro, Sao Paulo, Vitoria, and Porto Alegre) Ltda, Brazil) and height with a vertical Seca-SE-216 stadiometer (Seca
were assessed between August 2008 and December 2010. All partici- Brasil, Brazil) to the nearest 0.1 cm. Waist circumference was measured
pants gave written consent to participate, and the research protocol was with a tape measure to the nearest 0.1 cm around the midpoint between
approved by the local ethics committee of each institution involved. the inferior border of the ribs and the iliac crest. BMI was calculated as
weight in kilograms divided by height in square meters. A BMI $30 kg/m2
Study participants was considered obese.
The exclusion criteria were known diabetes (diabetes status reported
during initial interviews), taking oral hypoglycemic medications or Blood pressure. Resting blood pressure was measured 3 times at 1-min
insulin (n = 1473), CVD (n = 1280), cancer (n = 695), and other chronic intervals while participants remained in the seated position after a 5-min
diseases (n = 2649). We also excluded both a fasting state below 12 h and rest and obtained using a 765CP oscillometric sphygmomanometer
above 15 h (n = 542) and participants with a reported energy intake of (Omron). The average of the second and third measurements was used
#1298 kcal/d (second percentile) or high $6372 kcal/d (98th percentile) in the analyses. Hypertension was defined as systolic blood pressure
(n = 629). Some participants had more than 1 exclusion criteria. Our $140 mm Hg or diastolic blood pressure $90 mm Hg or verified
final sample comprised 9835 participants. treatment with antihypertensive medication during the past 2 wk.

Data collection and key measurements Subclinical measures. Participants arrived at the clinic after an
overnight fast of 8 to 15 h and were instructed to avoid heavy physical
Dairy intake and other diet variables. A validated FFQ with 114 items activity during the previous day. Blood was drawn by venipuncture, and
was administered to evaluate the participants usual intake over the past a standard 2-h 75-g oral glucose tolerance test was administered for all
12 mo (13). For each food item ascertained, the ELSA-Brazil FFQ participants who did not report a diagnosis or previous treatment for
included measures of portions and frequency of consumption, the latter diabetes. Glucose was measured by an ADVIA 1200 chemistry hexoki-
of which had 8 response options: >3 times/d, 23 times/d, 1 time/d, 56 nase system (Siemens). Glycated hemoglobin was measured using an
times/wk, 24 times/wk, 1 time/wk, 13 times/mo, and never/almost HPLC assay (Bio-Rad D-10 Dual Program Laboratories), which is
never. certified by the National Glycohemoglobin Standardization Program.
The following items were included as dairy products: milk (skimmed We calculated the insulin sensitivity index composite with the
milk, low-fat milk, whole milk), yogurt (regular, low-fat), cheese (regu- formula 10; 000=Offasting glucosemg=dL3fasting insulinmU=mL3
lar, low-fat), cheese spread, cream cheese base (regular, low-fat), ice mean glucose3mean insuling and used fasting and 2-h values for the
cream, desserts made with milk (pudding, mousses), and butter. Mixed means (15).
dishes and products for which the contribution of dairy products was HDL cholesterol and TGs were estimated using enzymatic pro-
deemed negligible were not included in the derivation of our dairy intake cedures (ADVIA 1200). Low HDL cholesterol was defined by sex:
estimates. Total dairy intake was calculated as the sum of all dairy pro- men <1.03 mmol/L (40 mg/dL) and women <1.29 mmol/L (50 mg/dL).
ducts reported in the dietary questionnaire. We computed servings per Hypertriglyceridemia was considered when $1.69 mmol/L (150 mg/dL).
day for total dairy intake and the following dairy subgroups: full- and
low-fat dairy, milk, cheese, yogurt, cheese spread and cream cheese base, Metabolic syndrome definition. For diagnosing MetSyn, we used the
dairy-based desserts, butter, and fermented dairy. The amount of grams joint interim statement consensus criteria (3), which require the presence
per serving for each dairy food was 240 g for milk, 120 g for yogurt, 30 g of any 3 of the following 5 risk factors: elevated waist circumference

2 of 9 Drehmer et al.
($102 cm in men and $88 cm in women), elevated TGs or drug higher than low-fat consumption, corresponding to 189 and 155 g/d,
treatment ($150 mg/dL), reduced HDL cholesterol or drug treatment respectively. Mean dairy food subgroup intakes were 241 g/d for
(<40 mg/dL for men and <50 mg/dL for women), elevated blood milk, 38 g/d for cheese, 37.8 g/d for yogurt, 5.2 g/d for cream
pressure or drug treatment ($130 mm Hg and/or diastolic blood pressure cheese, 1.6 g/d for butter, 12 g/d for ice cream, and 8 g/d for dairy
$85 mm Hg), and elevated fasting glucose ($100 mg/dL), which included
desserts.
those classified as having diabetes as ascertained simply by laboratory
measures at baseline.
Table 1 presents baseline characteristics for the overall sample
In addition, a metabolic risk score (MetScore) was computed using and is stratified by categories of dairy consumption. Dairy intake
an approach similar to that previously reported (16) as the mean of was higher in whites than in other ethnic groups and in
z scores of the continuous metabolic risk factors (waist circumference, participants with a university degree than in those with less
systolic blood pressure, HDL cholesterol (negative z score), TGs, and formal education. Higher dairy consumption was also observed
fasting glucose as follows: among nonsmokers and those who drank less alcohol, did more
 moderate and vigorous physical activity, and ate fruits and
MetScore Zwaist 1 ZSBP 2 ZlnHDL-C 1 ZlnTriglycerides 1 ZFasting glucose 5 2
vegetables every day. Values of proteins, total fat, and saturated
HDL cholesterol and TGs were log-transformed to achieve normality fat intake, when expressed as a percentage of energy intake,
before computing z scores. The z scores for waist circumference and increased according to higher dairy product consumption.
HDL cholesterol were sex-specific. We also computed a second MetScore Means and frequencies of cardiometabolic risk factors for the
using log-transformed values of 2 h of postload glucose instead of fasting overall sample and by dairy servings per day are also presented in
glucose. Table 1. Plasma glucose and blood pressure values declined with a
greater consumption of dairy products. Thus, there were less cases
Covariates: demographics, socioeconomics, and behaviors. Inter-
views and examinations were conducted by trained health professionals
of diabetes (as ascertained solely by study-determined laboratory
with strict quality control. Standardized questionnaires provided infor- values) and hypertension in participants with higher dairy con-
mation on demographics (age, sex, race, educational level, family sumption. Although TG concentrations were lower in participants
with high dairy consumption, those of HDL cholesterol were

Downloaded from jn.nutrition.org by guest on February 23, 2017


income, occupation status, study center), family history of diabetes,
smoking (current and previous), and menopausal status. Alcohol intake higher only in women who consumed more dairy products.
was estimated as the sum of ethanol (g/wk) of all beverages consumed. MetScore was inversely related to categories of dairy intake.
Leisure-time physical activity was defined using the long form of the An examination of the cardiometabolic risk factors as
International Physical Activity Questionnaire according to its guidelines categorical variables revealed that the frequency of elevated
for data processing and analysis. Median values and interquartile ranges waist circumference was 32.2% in the overall sample. Partici-
were computed for walking, moderate intensity activities, vigorous-
pants who had elevated TGs or who were taking fibrates or
intensity activities, and a combined total leisure-time physical activity
score (17). All continuous scores were expressed in metabolic equivalent
nicotinic acid represented 29.9% of the sample, and those with
minutes per week. low HDL cholesterol or taking those medications represented
17.2%. Those with blood pressure $130 mm Hg for systolic
Statistical analysis and/or $85 mm Hg for diastolic or medication for hypertension
Multivariable linear regression analysis was performed to test the within the last 2 wk made up 38.9%. Those with fasting glucose
association of servings per day of dairy products with the MetSyn >100 mg/dL or using antidiabetic medication corresponded to
components: waist circumference, TGs, HDL cholesterol, systolic 68.6% of participants.
blood pressure, and fasting glucose. We adjusted for covariates Table 2 presents adjusted means of cardiometabolic risk
(potential confounders) in 2 models, the first of which, model 1, factors across levels of total dairy intake. Significant inverse dose-
included demographic covariates, menopausal status, and family
response associations were found between total dairy intake and
history of diabetes; model 2 included all variables from model 1 plus
nondietary (smoking status, physical activity) and dietary factors
adjusted mean values of systolic and diastolic blood pressure, 2-h
(alcohol intake, energy intake, and nondairy food groups). For the postload glucose, and TGs. We found a positive graded associ-
nondairy food groups, we retained in the models those with a P value ation between total dairy intake and HDL cholesterol levels in
<0.20. This approach, while decreasing the number of covariates in any women only. We did not find a significant linear trend between
given model, resulted in slightly different sets of covariates in models total dairy consumption and fasting glucose in model 2 after full
that evaluated different outcomes. We used the same modeling adjustment. Visual inspection of the trends across the categories
approach to evaluate total dairy and dairy subgroups associations presented in Table 2 showed no important nonlinearity in the
with MetScore. Multiple logistic regression was used to evaluate the relevant range of dairy servings per day, a finding corroborated by
association between dairy consumption and MetSyn when expressed no or minimal significance for quadratic in terms of servings per
categorically, and we included in model 1 demographics, menopausal
day when added in additional models (data not shown).
status, family history of diabetes, and behaviors; for model 2 we added
nondairy food groups. A linear trend was tested by modeling categor-
Table 3 summarizes the adjusted means of the MetScore by
ical dairy servings per day (<1, 12, >24, and >4) as a continuous total dairy intake categories and across subgroups of dairy (full-
variable in the multivariable regression models. and low-fat) intake. We observed a graded inverse association
We evaluated the potential mediation of the dairy associations by between total and full-fat dairy with MetScore, but we did not
adding dairy nutrient components to the final models. A correlation find the same association for low-fat dairy products in models
matrix was used to explore multicollinearity, and variance inflation adjusted for demographics, behavioral risk factors, caloric
factor values >10 indicated multicollinearity. All analyses were intake, and nondairy food groups. In a fully adjusted model fit
performed with the statistical package SPSS version 18 (IBM), and 5% with both full- and low-fat dairy intake, thus each being adjusted
was considered significant. for the other, we found similar bs 6 SEs and P values (20.070 6
0.02, P = 0.001 for a 1 serving per day difference in full fat and
20.022 6 0.02, P = 0.280 for a 1 serving per day difference in
Results
low-fat dairy intake). When we computed the MetScore in
The mean 6 SD age at baseline was 50.7 6 8.7 y, and 54.8% (n = similarly adjusted additional analyses using log-transformed 2-h
5390) were women. On average, participants consumed 344 g/d postload glucose instead of fasting glucose, the graded inverse
total dairy products, and the intake of full-fat products was association between total dairy with MetScore became stronger
Dairy intake and metabolic syndrome in ELSA-Brasil 3 of 9
TABLE 1 Characteristics and cardiometabolic risk factors of Brazilian participants without previously diagnosed diabetes by categories
of dairy servings per day: ELSA-Brasil 200820101

Categories of dairy consumption


Final sample (n = 9835) ,1 sv/d (n = 1038) 12 sv/d (n = 1956) .24 sv/d (n = 3763) .4 sv/d (n = 3078)

Age, y 50.7 6 8.7 50.6 6 8.0 50.5 6 8.5 50.6 6 8.7 51.0 6 8.9
Sex, n (%)
Men 4445 (45.2) 634 (14.3) 951 (21.4) 1581 (35.6) 1279 (28.8)
Women 5390 (54.8) 404 (7.5) 1005 (18.6) 2182 (40.5) 1799 (33.4)
Self-identified skin color/race category, n (%)
White 5220 (51.7) 394 (7.8) 865 (17.1) 2004 (39.6) 1796 (35.5)
Black 1532 (15.7) 226 (14.8) 370 (24.2) 544 (35.5) 392 (25.6)
Brown (pardo or mixed) 2828 (29.0) 358 (12.7) 634 (22.4) 1048 (37.1) 788 (27.9)
Asian 218 (2.2) 29 (13.3) 46 (21.1) 95 (43.6) 48 (22.0)
Indigenous 102 (1.0) 18 (17.6) 23 (22.5) 37 (36.3) 24 (23.5)
Educational level, n (%)
Complete secondary school 3619 (35.8) 478 (13.6) 813 (23.1) 1312 (37.3) 910 (25.9)
University degree 5414 (53.5) 347 (6.6) 857 (16.2) 2123 (40.2) 1955 (37.0)
Family history of diabetes, n (%) 3577 (36.4) 415 (11.6) 721 (20.2) 1340 (37.5) 1101 (30.8)
Current smoker, n (%) 1243 (12.4) 201 (16.2) 292 (23.5) 416 (33.5) 334 (26.9)
Alcohol intake, g ethanol/d 67 6 130 107 6 202 69.7 6 136 60.9 6 113 59.4 6 112
Physical activity (moderate and vigorous), min/wk 426 6 926 304 6 783 339 6 779 448 6 946 496 6 1019

Downloaded from jn.nutrition.org by guest on February 23, 2017


Intake of vegetable every day, n (%) 5046 (51.3) 437 (8.7) 903 (17.9) 2010 (39.8) 1696 (33.6)
Intake of fruit every day, n (%) 5472 (55.6) 428 (7.8) 961 (17.6) 2211 (40.4) 1872 (34.2)
Nutrients, % energy
Carbohydrate 52.9 6 8.2 55.8 6 8.8 54.4 6 8.6 52.9 6 7.7 50.9 6 7.9
Protein 18.8 6 3.8 17.5 6 4.0 18.3 6 3.9 18.9 6 3.7 19.2 6 3.7
Total fat 28.5 6 5.2 25.9 6 5.3 27.2 6 5.0 28.4 6 4.8 30.4 6 5.1
Total SFAs 9.8 6 2.6 7.3 6 1.9 8.5 6 1.9 9.7 6 2.1 11.5 6 2.6
SFAs from dairy products 3.4 6 2.3 0.7 6 0.5 1.8 6 0.9 3.3 6 1.4 5.6 6 2.4
Cardiometabolic risk factors
Fasting glucose, mg/dL 106 6 17.1 109 6 19.9 107 6 19.5 106 6 15.9 105 6 15.6
2-h postload glucose, mg/dL 131 6 44.4 135 6 49.1 133 6 47.4 131 6 44.4 128 6 40.3
ISI composite 7.9 6 7.8 7.8 6 7.7 7.9 6 7.9 7.9 6 7.7 7.9 6 7.9
TGs, mg/dL 135 6 95.6 159 6 131 139 6 99.4 132 6 87.2 128 6 87.1
HDL cholesterol, mg/dLmen (n = 4442) 51.2 6 12.3 51.5 6 13.1 51.5 6 12.9 51.1 6 11.8 50.8 6 11.9
HDL cholesterol, mg/dLwomen (n = 5389) 62.1 6 14.4 59.9 6 13.4 60.8 6 13.7 62.4 6 14.4 62.9 6 14.7
Systolic blood pressure, mm Hg 120 6 16.6 125 6 19.1 122 6 17.1 119 6 16.2 118 6 15.5
Diastolic blood pressure, mm Hg 76.1 6 10.8 78.7 6 11.8 77.2 6 10.9 75.5 6 10.6 75.2 6 10.3
BMI, kg/m2men (n = 4443) 26.7 6 4.2 26.4 6 4.1 26.6 6 4.1 26.7 6 4.3 26.8 6 4.3
BMI, kg/m2women (n = 5389) 26.6 6 4.8 26.9 6 5.0 27.0 6 4.9 26.5 6 4.7 26.5 6 4.8
Waist, cmmen (n = 4444) 94.2 6 11.5 93.0 6 10.9 93.7 6 11.6 94.4 6 11.6 94.8 6 11.6
Waist, cmwomen (n = 5390) 86.4 6 11.9 87.3 6 12.0 87.4 6 12.4 86.0 6 11.7 86.2 6 11.9
Obesity (BMI $ 30), n (%) 20.1 199 (19.2) 416 (21.3) 745 (19.8) 621 (20.2)
Abdominal obesity,2 n (%)men (n = 4444) 21.9 118 (18.6) 195 (20.5) 348 (22.0) 313 (24.5)
Abdominal obesity,2 n (%)women (n = 5390) 39.8 165 (40.8) 442 (44.0) 831 (38.1) 707 (39.3)
Diabetes,3 n (%) 10.6 140 (13.5) 227 (11.6) 410 (10.9) 267 (8.7)
Hypertension,4 n (%) 38.9 491 (47.3) 840 (42.9) 1406 (37.4) 1088 (35.4)
Hypertriglyceridemia,5 n (%) 29.9 407 (39.2) 621 (31.8) 1086 (28.9) 824 (27.8)
Low HDL cholesterol,6 n (%) 17.2 183 (17.6) 348 (17.8) 650 (17.3) 506 (16.4)
MetScore 0.00 6 3.03 0.69 6 3.20 0. 26 6 3.05 20.12 6 2.99 20.23 6 2.99
MetSyn,7 n (%) 30.7 384 (37.0) 646 (33.0) 1104 (29.4) 885 (28.8)
1
Values are means 6 SDs unless otherwise indicated. ELSA-Brasil, Brazilian Longitudinal Study of Adult Health; ISI, insulin sensitivity index; MetScore, metabolic risk score;
MetSyn, metabolic syndrome; sv, serving.
2
Men: waist .102 cm; women: waist .88 cm.
3
Diabetes diagnosed by fasting or 2-h postload glucose or glycated hemoglobin.
4
Systolic blood pressure $130 mm Hg, diastolic blood pressure $85 mm Hg, or hypertensive medication used in last 2 wk.
5
We excluded cases with fasting times less than 12 h or greater than 15 h for values $150 mg/dL.
6
HDL cholesterol: men, ,40 mg/dL; women, ,50 mg/dL.
7
MetScore: mean z scores of continuous metabolic syndrome risk factors (waist circumference, systolic blood pressure, negative HDL cholesterol, TGs, and fasting glucose),
n = 9825, we excluded cases with fasting times less than 12 h or greater than 15 h.

4 of 9 Drehmer et al.
TABLE 2 Estimated mean systolic and diastolic blood pressure, fasting glucose, 2-h postload glucose, TGs, HDL cholesterol, and
waist circumference according to dairy servings per day of Brazilian adults: ELSA-Brasil 200820101

Adjusted2 mean 6 SE Linear estimate Linear


,1 sv/d (n = 1074) 12 sv/d (n = 2023) .24 sv/d (n = 3864) .4 sv/d (n = 3153) for 1 sv/d3 trend P value4

SBP, mm Hg
Model 1 128 6 1.22 127 6 1.19 126 6 1.17 125 6 1.17 20.91 6 0.16 ,0.001
Model 2 127 6 1.24 126 6 1.21 124 6 1.19 123 6 1.21 21.18 6 0.17 ,0.001
DBP, mm Hg
Model 1 80.6 6 0.84 80.2 6 0.81 79.3 6 0.80 79.0 6 0.80 20.58 6 0.11 ,0.001
Model 2 80.3 6 0.83 79.7 6 0.81 78.8 6 0.79 78.2 6 0.80 20.73 6 0.11 ,0.001
Fasting glucose, mg/dL
Model 1 110 6 1.36 110 6 1.32 109 6 1.30 109 6 1.30 20.38 6 0.18 0.032
Model 2 111 6 1.38 111 6 1.35 110 6 1.33 110 6 1.35 20.21 6 0.20 0.29
2-h postload glucose, mg/dL
Model 1 134 6 3.58 135 6 3.49 134 6 3.43 131 6 3.44 21.19 6 0.47 0.012
Model 2 135 6 3.59 135 6 3.49 135 6 3.45 132 6 3.49 21.15 6 0.50 0.021
TGs, mg/dL (n = 8644)
Model 1 165 6 7.65 153 6 7.44 149 6 7.32 145 6 7.34 25.62 6 1.01 ,0.001
Model 2 178 6 7.70 167 6 7.50 165 6 7.41 160 6 7.51 25.18 6 1.11 ,0.001
HDL cholesterol, mg/dLmen (n = 4382)
Model 1 56.1 6 1.46 54.9 6 1.44 55.2 6 1.42 54.8 6 1.40 20.30 6 0.18 0.10

Downloaded from jn.nutrition.org by guest on February 23, 2017


Model 2 52.5 6 1.41 53.1 6 1.39 52.7 6 1.38 52.6 6 1.40 20.05 6 0.20 0.81
HDL cholesterol, mg/dLwomen (n = 5324)
Model 1 62.0 6 1.70 62.8 6 1.61 64.1 6 1.58 64.4 6 1.58 0.75 6 0.22 , 0.001
Model 2 59.8 6 1.66 60.4 6 1.60 61.3 6 1.57 61.7 6 1.58 0.68 6 0.23 0.003
Waist, cmmen (n = 4515)
Model 1 95.9 6 1.37 96.5 6 1.35 96.6 6 1.33 96.8 6 1.33 0.25 6 0.18 0.15
Model 2 95.3 6 1.35 95.9 6 1.34 96.2 6 1.32 96.2 6 1.34 20.28 6 0.19 0.38
Waist, cmwomen (n = 5325)
Model 1 87.9 6 1.37 88.4 6 1.31 87.6 6 1.29 87.6 6 1.29 20.23 6 0.18 0.19
Model 2 87.9 6 1.38 88.5 6 1.32 88.60 6 1.30 87.0 6 1.32 20.48 6 0.19 0.012
1
ELSA-Brasil, Brazilian Longitudinal Study of Adult Health; sv, serving.
2
Adjusted mean determined by ANCOVA for each of the following variables: model 1: demographics, including age (y), sex, race, occupational status, education, family income,
study center, menopausal status, and family history of diabetes; model 2: model 1 + dietary [calorie intake (kcal/d) and nondairy food groups (g/d)], nondietary lifestyle factors
[current and previous smoking status, alcohol intake (grams of ethanol per day), physical activity (metabolic equivalent min/wk), and BMI (except for the waist model)] (nondairy
food groups included in the models varied across outcomes because only variables with P , 0.20 in the final model were maintained); systolic blood pressure: fruits, vegetables,
whole grains, coffee, sodas, and processed red meat; diastolic blood pressure: fruits, whole grains, coffee, sodas, processed red meat, non-dairy sweets and desserts; fasting
glucose: fruits, vegetables, refined grains, coffee, sodas, processed red meat, eggs, and nondairy sweets and desserts; 2-h postload glucose: fruits, whole grains, coffee, eggs,
and nondairy desserts; TGs: fruits, vegetables, refined grains, coffee, tea, sodas, and nondairy sweets and desserts; HDL cholesterol (men): fruits, refined grains, sodas, and juice;
HDL cholesterol (women): fruits, refined grains, coffee, tea, sodas, juice, unprocessed red meat, and beans; waist (men): fruits, vegetables, whole grains, sodas, juice,
unprocessed red meat, processed red and white meat, beans, eggs, and nondairy sweets and desserts; and waist (women): vegetables, refined grains, coffee, tea, sodas,
unprocessed red and white meat, processed red and white meat, eggs, and nondairy sweets and desserts.
3
Values are b 6 SE.
4
Linear trend was tested by modeling dairy servings per day (,1, 12, .24, and .4 sv/d) as a continuous variable in the multivariable regression models.

(20.132 6 0.04, P = 0.001), and we found a similar association We observed a strong inverse association between full-fat
between full-fat dairy and MetScore (20.123 6 0.03, P < dairy and MetScore (difference of 20.062 6 0.02 servings per
0.001). The findings were consistent across categories of sex and day, P = 0.002) when analyzing subgroups of dairy consump-
race in a stratified analysis. tion. This association was also present for butter, yogurt, and
Despite a strong association between continuous MetScore fermented dairy. For dairy-based desserts, we found a positive
and dairy products, when using the current MetSyn criteria of association with MetScore (0.169 6 0.07, P = 0.023) (Table 5).
the joint interim statement consensus (3), we found statistically Several nutrients were evaluated as possible mediators of
significant associations only in the model adjusted by demo- dairy consumption and MetScore. These nutrients included
graphics, caloric intake, smoking, alcohol, and physical activity animal protein, total FAs, and SFAs from dairy products
(consumption of >4 dairy servings per day represented a 27% (including specific SFAs) and butyric, caproic, caprylic, capric,
lower odds of MetSyn; 95% CI: 0.62, 0.87). In the fully adjusted lauric, myristic, palmitic, stearic, palmitoleic/monounsaturated,
model, we did not observe a statistically significant association and polyunsaturated acids when expressed as a percentage of
(OR: 0.90; 95% CI: 0.75, 1.08). However, when we analyzed energy intake. We also considered as possible mediators choles-
the relation between total dairy intake and a high MetScore terol, lactose, sucrose, vitamins A and D, calcium, phosphorus,
(above the 75th percentile), participants with a reported intake magnesium, sodium, and potassium. After adjusting addition-
of >4 dairy servings per day had 25% lower odds (OR: 0.75; ally for saturated fat from dairy products (Table 6), associations
95% CI: 0.61, 0.92) of MetSyn in the fully adjusted model of total dairy (20.023 6 0.03, P = 0.37) and full-fat dairy
(Table 4). (20.025 6 0.02, P = 0.33) intake with MetScore were no longer
Dairy intake and metabolic syndrome in ELSA-Brasil 5 of 9
TABLE 3 Estimated mean MetScore according to type and intake of dairy servings per day of Brazilian adults: ELSA-Brasil
200820101

Adjusted2 mean 6 SE Linear estimate Linear


Dairy Type ,1 sv/d (n = 1074) 12 sv/d, n = 2023 .24 sv/d (n = 3864) .4 sv/d (n = 3153) for dairy intake,3 sv/d trend P value4

Total dairy
MetScore5
Model 1 0.99 6 0.24 0.88 6 0.23 0.69 6 0.23 0.58 6 0.23 20.053 6 0.01 ,0.001
Model 2 1.03 6 0.24 0.94 6 0.23 0.82 6 0.23 0.69 6 0.23 20.044 6 0.01 0.009
Full-fat dairy
MetScore
Model 1 0.95 6 0.23 0.83 6 0.23 0.69 6 0.23 0.43 6 0.24 20.155 6 0.03 ,0.001
Model 2 0.98 6 0.23 0.86 6 0.23 0.76 6 0.23 0.57 6 0.24 20.126 6 0.03 ,0.001
Low-fat dairy
MetScore
Model 1 0.83 6 0.23 0.86 6 0.23 0.71 6 0.24 0.59 6 0.25 20.071 6 0.03 0.025
Model 2 0.87 6 0.23 0.94 6 0.23 0.82 6 0.24 0.79 6 0.26 20.022 6 0.03 0.49
1
ELSA-Brasil, Brazilian Longitudinal Study of Adult Health; MetScore, metabolic risk score; sv, serving.
2
Adjusted through ANCOVA for each of the following variables: model 1: demographics [age (y), sex, race, occupational status, education, family income [amount], study center,
menopausal status, and family history of diabetes]; model 2: model 1 plus behaviors [current and previous smoking status, alcohol intake (grams of ethanol per day), physical
activity (metabolic equivalent min/wk)], and nondairy variables] (nondairy dietary variables included in models vary across dairy type because only variables with P , 0.20 in the final
model were maintained); total dairy: calorie intake (kcal/d) and nondairy food groups [fruits, vegetables, whole and refined grains, tea, sodas, juice, unprocessed red and white

Downloaded from jn.nutrition.org by guest on February 23, 2017


meat, processed red meat, eggs, and nondairy sweets and desserts (g/d)]; full-fat dairy: calorie intake (kcal/d) and nondairy food groups [fruits, vegetables, whole and refined
grains, tea, sodas, juice, unprocessed and processed red meat, eggs, and nondairy sweets and desserts (g/d)]; and low-fat dairy: calorie intake (kcal/d) and nondairy food groups
[fruits, vegetables, whole and refined grains, tea, sodas, juice, unprocessed red and white meat, processed red meat, eggs, and nondairy sweets and desserts (g/d)].
3
Values are b 6 SE.
4
Linear trend was tested by modeling dairy servings per day (,1, 12, .24, and .4 sv/d) as a continuous variable in the multivariable regression models.
5
MetScore: metabolic risk score computed as the sum of z scores over all continuous metabolic risk factors (waist circumference, systolic blood pressure, HDL cholesterol, TGs,
and fasting glucose).

present. In each of these models, greater SFA intake from dairy Men and women who consumed >4 servings of dairy products
products was associated with a lower MetScore (P < 0.001). These per day had lower levels of systolic and diastolic blood pressure,
findings suggest potential mediation through SFAs consumed in 2-h postload glucose, and TGs. There was a positive monotonic
dairy products. No other potential mediators were found. trend between the intake of dairy products and adjusted mean
HDL cholesterol in women only. We also found evidence that
SFAs from dairy products may mediate these associations.
Discussion
Our results are consistent with previous prospective and
After adjusting for demographics, menopausal status, family cross-sectional research (4, 1821). The protective effect of dairy
history of diabetes, dietary intake, nondietary lifestyle factors, intake in relation to MetSyn has been reported previously in
and BMI in this large cohort study of Brazilian adults, we multicenter prospective cohort studies. In the Atherosclerosis
observed a strong inverse association between total dairy intake Risk in Communities study, a cohort of 9514 participants from
and cardiometabolic risk with a linear dose-response pattern. 4 US communities, individuals in the highest quintile of dairy
Interestingly, this association was observed for full-fat dairy consumption had a 13% (CI 95%: 0.77, 0.98) lower risk of
products, butter, and yogurt but not for low-fat dairy products. developing MetSyn, and this finding was consistent across sex

TABLE 4 Adjusted associations of MetSyn with total dairy intake of Brazilian adults: ELSA-Brasil
200820101

OR (95% CI) Linear


MetSyn criteria ,1 sv/d (n = 1074) 12 sv/d (n = 2023) .24 sv/d (n = 3864) .4 sv/d (n = 3153) trend P value

JIS
Model 1 1.00 0.92 (0.78, 1.08) 0.82 (0.70, 0.96) 0.73 (0.62, 0.87) ,0.001
Model 2 1.00 0.95 (0.80, 1.13) 0.91 (0.77, 1.07) 0.90 (0.75, 1.08) 0.26
High MetScore2
Model 1 1.00 0.91 (0.75, 1.10) 0.76 (0.64, 0.91) 0.65 (0.53, 0.78) ,0.001
Model 2 1.00 0.93 (0.77, 1.12) 0.83 (0.69, 0.99) 0.75 (0.61, 0.92) 0.002
1
Adjusted through multiple logistic regression for the following variables: model 1: demographic characteristics, including age (y), sex, race,
occupational status, education, family income (amount), study center, menopausal status, family history of diabetes, current and previous
smoking status, alcohol intake (grams of ethanol per day), physical activity (metabolic equivalent min/wk), and calorie intake (kcal/d); model
2: model 1 + nondairy food groups [fruit, vegetables, refined grains, sodas, and processed and unprocessed red and white meat (g/d)].
ELSA-Brasil, Brazilian Longitudinal Study of Adult Health; JIS, joint interim statement definition (3); MetScore, metabolic risk score; MetSyn,
metabolic syndrome; sv, serving.
2
High MetScore = mean z score $75th percentile.

6 of 9 Drehmer et al.
TABLE 5 Adjusted difference in MetScore for a 1-unit differ- between dairy products and MetSyn and diabetes risk (25).
ence in servings per day in the intake of each dairy subgroup of Similar to our study, the Epidemiological Study on the Insulin
Brazilian adults: ELSA-Brasil 200820101 Resistance Syndrome also found no association between cheese
intake and MetSyn or diabetes. By contrast, some studies have
Dairy subgroup Difference mean 6 SE P value
suggested positive or null associations between the intake of
Full-fat dairy 20.062 6 0.02 0.002 dairy products and MetSyn or CVD risk (2628).
Low-fat dairy 0.002 6 0.02 0.92 Because of the somewhat inconsistent evidence, meta-
Milk 0.00 6 0.03 0.76 analyses can be helpful. The relative risk of MetSyn when
Cheese 20.02 6 0.02 0.29 combining results of the 4 observational studies in subjects with
Yogurt 20.15 6 0.06 0.024 high milk or dairy consumption was 0.74 (95% CI: 0.64, 0.84)
Cream cheese 20.070 6 0.08 0.41 relative to the risk in those with low consumption (29). Data
Desserts 0.169 6 0.07 0.023 from a dose-response meta-analysis that included 13,518
Butter 20.129 6 0.04 0.003 participants and 2283 CVD (fatal and nonfatal) cases were
Fermented dairy 20.05 6 0.02 0.047 analyzed in 4 other prospective cohort studies, with milk as the
main exposure. An inverse association was found between milk
1
Adjusted through multivariable linear regression for demographics variables [includ-
intake and risk of overall CVD (RR = 0.94 per 200 mL/d; 95%
ing age (y), sex, race, occupational status, education, family income (amount), and
study center], menopausal status, family history of diabetes, current and previous
CI: 0.89, 0.99 per 200 mL/d) (30). A different meta-analysis of
smoking status, alcohol intake (grams of ethanol per day), physical activity (metabolic observational studies evaluated the association between high-fat
equivalent min/wk), calorie intake (kcal/d), and nondairy food groups [fruits, vegeta- dairy products and obesity, cardiovascular disease, and MetSyn
bles, whole and refined grains, coffee, tea, sodas, juice, unprocessed and processed and found high-fat dairy intake to be inversely associated with
red and white meats, beans, eggs, and nondairy sweets and desserts (g/d)]. ELSA-
the outcomes in 11 of 16 studies (31). However, a meta-analysis
Brasil, Brazilian Longitudinal Study of Adult Health; MetScore, metabolic risk score.
of 20 mostly short-term randomized controlled studies with
1677 healthy adults found no significant effects of increased

Downloaded from jn.nutrition.org by guest on February 23, 2017


and BMI categories (22). In the Coronary Artery Risk Devel- dairy food on the cardiometabolic risk factors evaluated herein.
opment in Young Adults study that followed ;3,000 par- The authors cautioned that most of the trials were small and of
ticipants, the highest category of dairy consumption ($5/d) modest quality (7).
compared with the lowest was associated with a 70% reduction We found a graded inverse association between total and
(95% CI: 0.14, 0.70) in a 10-y cumulative incidence of the full-fat dairy consumption and MetSyn in this study but no
insulin resistance syndrome (defined differently than MetSyn) in association for low-fat dairy consumption after fully adjusting
over- but not normal-weight participants (23). In the Multi- for covariates. When we added dairy-related nutrients to the
Ethnic Study of Atherosclerosis, 316 incident CVD cases were fully adjusted model, the results suggested that SFAs found in
observed in 5209 participants between the ages of 45 and 84 y. dairy products may be mediating the inverse association between
Each 5-unit increment in the percentage of energy from dairy fat dairy intake and MetSyn. The literature on SFAs and cardiovas-
was associated with a 38% lower risk of CVD (95% CI: 0.47, cular disease risk has been inconsistent and controversial (3234).
0.82) (24). The Epidemiological Study on the Insulin Resistance Our results that suggest that it may in fact be the dairy fat con-
Syndrome, a French cohort, also found inverse associations tent that explains the potential protective effect of dairy intake
highlight an important issue in this area of research.
In our previous analysis of this same population, ELSA-
Brasil, we found inverse associations between dairy consump-
TABLE 6 Change in the adjusted difference in MetScore for a tion and levels of glycemia and insulinemia that were independent
1 sv/d change in total dairy and full-fat dairy consumption with an
of obesity. This association was possibly mediated by myristic
additional adjustment for dairy SFA intake of Brazilian adults:
ELSA-Brasil 200820101
acid (9). Myristic acid has been associated with increased
LDL cholesterol and cardiovascular risk in epidemiological
Difference 6 SE P value studies (3537). However, the role of myristic acid in raising
plasma lipids and its impact on health outcomes are still
Total dairy controversial. Myristic and palmitic acids (both present in
Model without potential mediator dairy products) increase LDL cholesterol, but myristic acid
Total dairy 20.044 6 0.01 0.009 also increases HDL cholesterol (38). In an experimental
Model with potential mediator study, moderate quantities of myristic acid (1.2% and 1.8%
Total dairy 20.023 6 0.03 0.37 of total energy) were found to be associated with decreases in
SFAs from dairy, % energy 20.074 6 0.02 0.001 total cholesterol, LDL cholesterol, TGs, and the ratio of total
Full-fat dairy to HDL cholesterol (39). In another study of healthy young
Model without potential mediator subjects, myristic acid, present in dairy products, was chosen
Full-fat dairy 20.062 6 0.02 0.002 as a marker for the intake of milk fat and was inversely related
Model with potential mediator to the LDL cholesterol level (40).
Full-fat dairy 20.025 6 0.02 0.33 Dairy products are dense in nutrients, and the impact of
SFAs from dairy, % energy 20.062 6 0.02 0.001 dietary SFAs on cardiovascular risk may be influenced by the
1
Model without potential mediator adjusted by demographics [including age (y), sex, food and nutrient matrix involved with dairy intake. The effects
race, occupational status, education, family income (amount), study center], meno- of other components of dairy intake should also be considered.
pausal status, family history of diabetes, current and previous smoking status, alcohol For example, milk proteins have an angiotensin-converting
intake (grams of ethanol per day), physical activity (metabolic min/wk), calorie intake enzyme-inhibitory effect (41). The inhibition of the renin
(kcal/d), and nondairy food groups (g/d); model with potential mediator additionally
adjusted for SFAs from dairy products (with difference expressed for a 1% change in
angiotensin system in adipocytes can potentially reduce obesity
total energy intake of these saturated fats). ELSA-Brasil, Brazilian Longitudinal Study of and hypertension. As a rich source of calcium, dairy products
Adult Health; sv, serving. have been reported to reduce blood pressure and to be inversely
Dairy intake and metabolic syndrome in ELSA-Brasil 7 of 9
associated with adiposity (18). In our analyses, however, the 7. Benatar JR. Effects of low fat dairy food on cardio-metabolic risk factors:
only nutrient that materially changed the associations of dairy a meta-analysis of randomized studies. PLoS One 2013;8:e76480.
intake with the MetSyn was SFAs from dairy products. Because 8. Lamarche B. It is time to revisit current dietary recommendations for
saturated fat. Appl Physiol Nutr Metab 2014;39:140911.
of collinearity and limited ranges of intakes for some FAs, we
9. Drehmer M, Pereira MA, Schmidt MI, Molina MCB, Alvim S, Lotufo
could not reach any clear conclusions on specific SFAs. Consid-
PA, Duncan BB. Associations of dairy intake with glycemia and
ering that this mediation analysis is exploratory in nature, we insulinemia, independent of obesity, in Brazilian adults: the Brazilian
cannot draw definitive conclusions. Longitudinal Study of Adult Health (ELSA-Brasil). Am J Clin Nutr
This study has some limitations. We measured dairy intake and 2015;101:77582.
nutrients based on what participants reported in the FFQ, a typical 10. Yakoob MY, Shi P, Hu FB, Campos H, Rexrode KM, Orav EJ, Willett
choice in large epidemiologic studies, but this method is subject to WC, Mozaffarian D. Circulating biomarkers of dairy fat and risk of
random and systematic errors. We used cross-sectional data to incident stroke in U.S. men and women in 2 large prospective cohorts.
Am J Clin Nutr 2014;100:143747.
identify the association of dairy consumption with the MetSyn,
11. Souza AM, Pereira RA, Yokoo EM, Levy RB, Sichieri R. Most
and because of the observational nature of our study, it is possible consumed food in Brazil: National Dietary Survey 20082009. Rev
that our findings resulted from residual confounding despite the Saude Publica 2013;47 (Suppl 1):190S9S.
fact that we extensively adjusted for other dietary variables and 12. Ministry of Health of Brazil. Dietary guidelines for the Brazilian
risk factors. Strengths of this study include the direct measure- population; 2014. Braslia: Ministry of Health of Brazil [cited 2015 Jun
ments of the cardiometabolic risk factors and exclusion of subjects 3]. Available from: http://www.foodpolitics.com/wp-content/uploads/
with known CVD, diabetes, stroke, or other chronic diseases. Brazilian-Dietary-Guidelines-2014.pdf.
The findings of this study, in which we adjusted for demo- 13. Molina MC, Bensenor IM, Cardoso Lde O, Velasquez-Melendez G,
Drehmer M, Pereira TS, Faria CP, Melere C, Manato L, Gomes AL, et al.
graphics, menopausal status, family history of diabetes, dietary
Reproducibility and relative validity of the food frequency questionnaire
intake, nondietary lifestyle factors, and BMI, suggest that used in the ELSA-Brasil. Cad Saude Publica 2013;29:37989.
greater dairy intake, especially of full-fat dairy products, may 14. Nutrition Coordinating Center. Nutrition data system for research
decrease the risk of MetSyn in middle-aged and older adults.

Downloaded from jn.nutrition.org by guest on February 23, 2017


software. Minneapolis (MN): University of Minnesota; 2010.
Typical dietary guidelines, including the 2015 US guidelines now 15. DeFronzo RA, Matsuda M. Reduced time points to calculate the
under review and the new Brazilian dietary guidelines published composite index. Diabetes Care 2010;33:e93.
in 2014, recommend consuming low-fat instead of full-fat dairy 16. Brage S, Wedderkopp N, Ekelund U, Franks PW, Wareham NJ,
products because of the concern that dairy fat could increase the Andersen LB, Froberg K. Features of the metabolic syndrome are
risk for obesity and dyslipidemia (7, 12, 42). However, our associated with objectively measured physical activity and fitness in
Danish children: the European Youth Heart Study (EYHS). Diabetes
findings and the literature on the association of dairy pro-
Care 2004;27:21418.
ducts with obesity, cardiovascular disease, and diabetes do not
17. Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth
support dietary guidelines that recommend only a low-fat dairy BE, Pratt M, Ekelund U, Yngve A, Sallis JF, et al. International physical
intake. activity questionnaire: 12-country reliability and validity. Med Sci Sports
Exerc 2003;35:138195.
18. Azadbakht L, Mirmiran P, Esmaillzadeh A, Azizi F. Dairy consumption
Acknowledgments
is inversely associated with the prevalence of the metabolic syndrome in
MD, MIS, SA, PAL, VCL, and BBD designed and conducted the Tehranian adults. Am J Clin Nutr 2005;82:52330.
research and provided essential materials; MD and MAP 19. Ruidavets JB, Bongard V, Dallongeville J, Arveiler D, Ducimetie`re P,
analyzed the data and wrote the paper; MD, MAP, MIS, SA, Perret B, Simon C, Amouyel P, Ferrie`res J. High consumptions of grain,
PAL, VCL, and BBD had primary responsibility for final con- fish, dairy products and combinations of these are associated with a low
tent. All authors read and approved the final manuscript. prevalence of metabolic syndrome. J Epidemiol Community Health
2007;61:8107.
20. Shin H, Yoon YS, Lee Y, Kim CI, Oh SW. Dairy product intake is
inversely associated with metabolic syndrome in Korean adults:
References Anseong and Ansan cohort of the Korean genome and epidemiology
study. J Korean Med Sci 2013;28:14828.
1. WHO. Global status report on noncommunicable diseases. Geneva 21. Kim J. Dairy food consumption is inversely associated with the risk of
(Switzerland): WHO; 2014 [cited 2015 Jun 3]. Available from: http://
the metabolic syndrome in Korean adults. J Hum Nutr Diet 2013;26
apps.who.int/iris/bitstream/10665/148114/1/9789241564854_eng.pdf?ua=1.
(Suppl. 1):1719.
2. Mendis S, Davis S, Norrving B. Organizational update: the World
22. Lutsey PL, Steffen LM, Stevens J. Dietary intake and the development of
Health Organization global status report on noncommunicable diseases
the metabolic syndrome: the atherosclerosis risk in communities study.
2014; one more landmark step in the combat against stroke and
Circulation 2008;117:75461.
vascular disease. Stroke 2015;46:e1212.
23. Pereira MA, Jacobs DR, Jr., Van Horn L, Slattery ML, Kartashov AI,
3. Alberti KGMM, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato
Ludwig DS. Dairy consumption, obesity, and the insulin resistance
KA, Fruchart JC, James WPT, Loria CM, Smith SC. Harmonizing the
metabolic syndrome: a joint interim statement of the International Diabetes syndrome in young adults: the CARDIA Study. JAMA 2002;287:2081.
Federation Task Force on Epidemiology and Prevention; National Heart, 24. de Oliveira O, Mozaffarian D, Kromhout D, Bertoni AG, Sibley CT,
Lung, and Blood Institute; American Heart Association; World Heart Jacobs DR Jr, Nettleton JA. Mozaffarian D, Kromhout D, Bertoni AG,
Federation; International Atherosclerosis Society; and International Associ- Sibley CT, Jacobs DR Jr, Nettleton JA. Dietary intake of saturated fat by
ation for the Study of Obesity. Circulation 2009;120:16405. food source and incident cardiovascular disease: the Multi-Ethnic Study
4. Beydoun MA, Gary TL, Caballero BH, Lawrence RS, Cheskin LJ, Wang of Atherosclerosis. Am J Clin Nutr 2012;96:397404.
Y. Ethnic differences in dairy and related nutrient consumption among 25. Fumeron F, Lamri A, Abi Khalil C, Jaziri R, Porchay-Balderelli I,
US adults and their association with obesity, central obesity, and the Lantieri O, Vol S, Balkau B, Marre M. Dairy Consumption and the
metabolic syndrome. Am J Clin Nutr 2008;87:191425. Incidence of Hyperglycemia and the Metabolic Syndrome: results from
5. Astrup A. A changing view on saturated fatty acids and dairy: from a french prospective study, Data from the Epidemiological Study on the
enemy to friend. Am J Clin Nutr 2014;100:14078. Insulin Resistance Syndrome(DESIR). Diabetes Care 2011;34:8137.
6. Sacks FM, Svetkey LP, Vollmer WM, Appel LJ, Bray GA, Harsha D, 26. Lawlor DA, Ebrahim S, Timpson N, Davey Smith G. Avoiding milk is
Obarzanek E, Conlin PR, Miller, III ER, Simons-Morton DG, et al.. Effects associated with a reduced risk of insulin resistance and the metabolic
on blood pressure of reduced dietary sodium and the Dietary Approaches to syndrome: findings from the British Womens Heart and Health Study.
Stop Hypertension (DASH) diet. N Engl J Med 2001;344:310. Diabet Med 2005;22:80811.

8 of 9 Drehmer et al.
27. Snijder MB, van der Heijden AA, van Dam RM, Stehouwer CD, 35. Hu FB, Stampfer MJ, Manson JE, Ascherio A, Colditz GA, Speizer FE,
Hiddink GJ, Nijpels G, Heine RJ, Bouter LM, Dekker JM. Is higher Hennekens CH, Willett WC. Dietary saturated fats and their food
dairy consumption associated with lower body weight and fewer sources in relation to the risk of coronary heart disease in women. Am
metabolic disturbances? The Hoorn Study. Am J Clin Nutr 2007;85: J Clin Nutr 1999;70:10018.
98995. 36. Mensink RP, Zock PL, Kester AD, Katan MB. Effects of dietary fatty
28. Soedamah-Muthu SS, Verberne LD, Ding EL, Engberink MF, Geleijnse acids and carbohydrates on the ratio of serum total to HDL cholesterol
JM. Dairy consumption and incidence of hypertension. A dose-response and on serum lipids and apolipoproteins: a meta-analysis of 60
meta-analysis of prospective cohort studies. Hypertension 2012; controlled trials. Am J Clin Nutr 2003;77:114655.
60:11317.
37. Mayneris-Perxachs J, Guerendiain M, Castellote AI, Estruch R, Covas
29. Elwood PC, Givens DI, Beswick AD, Fehily AM, Pickering JE, MI, Fito M, Salas-Salvado J, Martnez-Gonzalez
MA, Aros F, Lamuela-
Gallacher J. The survival advantage of milk and dairy consumption: Raventos RM, et al. Plasma fatty acid composition, estimated desaturase
an overview of evidence from cohort studies of vascular diseases, activities, and their relation with the metabolic syndrome in a population
diabetes and cancer. J Am Coll Nutr 2008;27:723S34S.
at high risk of cardiovascular disease. Clin Nutr 2014;33:907.
30. Soedamah-Muthu SS, Ding EL, Al-Delaimy WK, Hu FB, Engberink MF,
38. Zock PL, de Vries JH, Katan MB. Impact of myristic acid versus
Willett WC, Geleijnse JM. Milk and dairy consumption and incidence
palmitic acid on serum lipid and lipoprotein levels in healthy women
of cardiovascular diseases and all-cause mortality: dose-response meta-
and men. Arterioscler Thromb 1994;14:56775.
analysis of prospective cohort studies. Am J Clin Nutr 2011;93:15871.
39. Dabadie H, Motta C, Peuchant E, LeRuyet P, Mendy F. Variations in
31. Kratz M, Baars T, Guyenet S. The relationship between high-fat dairy
consumption and obesity, cardiovascular, and metabolic disease. Eur J daily intakes of myristic and a-linolenic acids in sn-2 position modify
Nutr 2013;52:124. lipid profile and red blood cell membrane fluidity. Br J Nutr 2006;
96:2839.
32. Jakobsen MU, OReilly EJ, Heitmann BL, Pereira MA, Balter K, Fraser
GE, Goldbourt U, Hallmans G, Knekt P, Liu S, et al. Major types of 40. Samuelson G, Bratteby LE, Mohsen R, Vessby B. Dietary fat intake in
dietary fat and risk of coronary heart disease: a pooled analysis of 11 healthy adolescents: inverse relationships between the estimated intake
cohort studies. Am J Clin Nutr 2009;89:142532. of saturated fatty acids and serum cholesterol. Br J Nutr 2001;85:333
33. Siri-Tarino PW, Sun Q, Hu FB, Krauss RM. Meta-analysis of prospec- 41.
tive cohort studies evaluating the association of saturated fat with 41. Pihlanto-Leppala A, Koskinen P, Piilola K, Tupasela T, Korhonen H.

Downloaded from jn.nutrition.org by guest on February 23, 2017


cardiovascular disease. Am J Clin Nutr 2010;91:53546. Angiotensin I-converting enzyme inhibitory properties of whey protein
34. Astrup A, Dyerberg J, Elwood P, Hermansen K, Hu FB, Jakobsen MU, digests: concentration and characterization of active peptides. J Dairy
Kok FJ, Krauss RM, Lecerf JM, LeGrand P, et al. The role of reducing Res 2000;67:5364.
intakes of saturated fat in the prevention of cardiovascular disease: 42. Weaver CM. How sound is the science behind the dietary recommen-
where does the evidence stand in 2010? Am J Clin Nutr 2011;93:6848. dations for dairy? Am J Clin Nutr 2014;99:1217S22S.

Dairy intake and metabolic syndrome in ELSA-Brasil 9 of 9

Potrebbero piacerti anche