Content uploaded by Maria Soledad Hershey
Author content
All content in this area was uploaded by Maria Soledad Hershey on Jun 09, 2021
Content may be subject to copyright.
Original article
The Mediterranean lifestyle (MEDLIFE) index and metabolic syndrome
in a non-Mediterranean working population
Maria S. Hershey
a
,
b
, Mercedes Sotos-Prieto
b
,
c
,
d
, Miguel Ruiz-Canela
a
,
e
,
Costas A. Christophi
b
,
f
, Steven Moffatt
g
, Miguel
Angel Martínez-Gonz
alez
a
,
e
,
h
,
Stefanos N. Kales
b
,
i
,
*
a
Department of Preventive Medicine and Public Health, Navarra Institute for Health Research (IdiSNA), University of Navarra, 31008, Pamplona, Spain
b
Department of Environmental Health, Harvard T.H. Chan School of Public Health, 677 Huntington Avenue, Boston, MA, 02115, USA
c
Department of Preventive Medicine and Public Health, School of Medicine, Universidad Aut
onoma de Madrid, 28049, Madrid, Spain
d
CIBER of Epidemiology and Public Health (CIBERESP), Carlos III Health Institute, 28029, Madrid, Spain
e
Biomedical Research Network Centre for Pathophysiology of Obesity and Nutrition (CIBEROBN), Carlos III Health Institute, 28029, Madrid, Spain
f
Cyprus University of Technology, School of Health Sciences, Cyprus International Institute for Environmental and Public Health in Association with Harvard
School of Public Health, Limassol, Cyprus
g
National Institute for Public Safety Health, Indianapolis, IN 324 E New York Street, Indianapolis, IN, 46204, USA
h
Department of Nutrition, Harvard T.H. Chan School of Public Health, 677 Huntington Avenue, Boston, MA, 02115, USA
i
Occupational Medicine, The Cambridge Health Alliance/Harvard Medical School, Boston, MA, 02139, USA
article info
Article history:
Received 18 November 2020
Accepted 17 March 2021
Keywords:
Metabolic syndrome
Lifestyle
Cardiovascular disease
Type 2 diabetes mellitus
Mediterranean diet
summary
Background &aims: The Mediterranean lifestyle (MEDLIFE), as an overall lifestyle pattern, may be
associated with a lower prevalence of metabolic syndrome. We assessed the association of a validated
MEDLIFE index with metabolic syndrome and its components in a non-Mediterranean working
population.
Methods: A cross-sectional analysis was conducted at baseline among 249 US career firefighters in
Feeding America's Bravest 2016e2018. The MEDLIFE index consisted of 26 items on food consumption,
dietary habits, physical activity, rest, and social interactions that scored 0 or 1 point. Thus, total scores
could range from 0 to 26 points. Multivariable logistic regression models were used to determine the
associations across tertiles of MEDLIFE adherence with metabolic syndrome and each of its individual
components. Multivariable linear models further assessed each component as a continuous outcome.
Results: The prevalence of metabolic syndrome was 17.7%. Participants with higher MEDLIFE adherence
(T3: 11e17 points) had 71% lower odds of having metabolic syndrome compared to those with lower
MEDLIFE adherence (T1: 2e7 points) (OR ¼0.29; 95%CI: 0.10 to 0.90, p for trend ¼0.04). Furthermore,
significant inverse associations were found for T3 versus T1 on abdominal obesity (OR ¼0.42; 95%CI: 0.18
to 0.99, p for trend ¼0.07) and hypertriglyceridemia (OR ¼0.24; 95%CI: 0.09 to 0.63, p for trend ¼0.002).
Significant inverse associations for continuous outcomes included total-cholesterol (total-c), low-density
lipoprotein (LDL) cholesterol, and total-c:high-density lipoprotein (HDL) cholesterol (p for trend <0.05).
Conclusion: Higher adherence to traditional Mediterranean lifestyle habits, as measured by a compre-
hensive MEDLIFE index, was associated with a lower prevalence of metabolic syndrome and a more
favorable cardiometabolic profile in a non-Mediterranean working population. Future studies employing
the MEDLIFE index in other populations are warranted to support this hypothesis.
©2021 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
1. Introduction
Metabolic syndrome is a condition of clustered cardio-metabolic
(cardiovascular disease (CVD) and Type 2 Diabetes Mellitus (T2DM)
risk factors, which requires at least three out of five criteria
including hyperglycemia, raised blood pressure, elevated
*Corresponding author. Department of Environmental Health, Harvard T.H. Chan
School of Public Health, 677 Huntington Avenue, Boston, MA, 02115, USA.
E-mail addresses: mhershey@alumni.unav.es (M.S. Hershey), msotosp@hsph.
harvard.edu,mercedes.sotos@uam.es (M. Sotos-Prieto), mcanela@unav.es
(M. Ruiz-Canela), costas.christophi@cut.ac.cy (C.A. Christophi), steven.moffatt@
ascension.org (S. Moffatt), mamartinez@unav.es (M.
A. Martínez-Gonz
alez),
skales@hsph.harvard.edu (S.N. Kales).
Contents lists available at ScienceDirect
Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu
https://doi.org/10.1016/j.clnu.2021.03.026
0261-5614/©2021 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Clinical Nutrition 40 (2021) 2494e2503
triglyceride levels, low HDL cholesterol levels, and abdominal
obesity, respectively [1]. According to a representative sample of
the United States (US) population from the National Health and
Nutrition Examination Survey (NHANES), the prevalence of meta-
bolic syndrome increased from 32.5% from 2011 to 2012 to 36.9% in
2015e2016, with significant increases in those aged 20e39 years,
women, Asians, and Hispanics [2]. Metabolic syndrome has been
largely attributed to an overconsumption of calories, in addition to
more sedentary lifestyles worldwide, which has led to increasing
obesity rates [3]. Such trends indicate a strong need for effective
lifestyle modifications targeted at high risk populations to lower
CVD and T2DM risk, as well as subsequent healthcare costs and
disability [1e3].
US career firefighters often experience several lifestyle risk
factors, including poor diet quality and suboptimal eating habits,
physiological stress from strenuous physical activity at work,
emotional stress, and environmental pollutants, in addition to low
fitness [4,5]. Such lifestyle factors have led to a high prevalence of
obesity and other chronic conditions, including hypertension, hy-
percholesterolemia, and high blood glucose, contributing to sig-
nificant morbidity and disability among this working population
[5e8]. Existing evidence on dietary workplace interventions for
health promotion support healthy dietary changes, such as the
current Dietary Guidelines for Americans, which includes the Med-
iterranean dietary pattern (with its characteristic frugality), for the
long-term prevention of diet-related chronic disease [9]. On the
other hand, evidence supports workplace physical activity in-
terventions to improve body composition [10]. However, evidence
of the joint effect of multicomponent lifestyle patterns, including
interrelated factors such as physical activity and sedentary behavior
within a general way of living, on metabolic syndrome is limited
[11,12].
MEDLIFE, as described and represented in a Mediterranean Diet
Pyramid by the Mediterranean Diet Foundation's International
Scientific committee, encompasses several distinctive habits
beyond the Mediterranean diet, such as resting patterns, social
structures, consumption of seasonal and diverse foods, and other
healthy culinary techniques [13,14]. In 2014, a MEDLIFE index was
developed and validated in a Spanish working population based on
these recommendations [15,16]. To our knowledge, the MEDLIFE
index has only been associated with lower CVD risk factors and
mortality in Mediterranean populations [16e18]. Therefore, the
aim of this cross-sectional study was to evaluate the association
between the MEDLIFE index and metabolic syndrome in a non-
Mediterranean working population at high CVD risk.
2. Methods and materials
2.1. Study population and design
Feeding America's Bravest is a cluster-randomized diet inter-
vention trial conducted among US career firefighters from the 44
fire stations of the Indianapolis Fire Department and the 6 fire
stations of the Fishers Fire Department (both in Indiana, USA). The
primary objective of this randomized controlled trial was to
compare a Mediterranean diet nutritional intervention with mul-
tiple behavior change strategies (diet/lifestyle education, dis-
counted access to key Mediterranean diet foods, electronic
education platforms and reminders) with a Midwestern-style diet
or “usual care”group, using a 2-year cross-over study design. In the
present cross-sectional study we analyzed baseline data to assess
the association between Mediterranean lifestyle factors and
metabolic syndrome or its individual components.
The study protocol was approved by the Harvard Institutional
Review Board (IRB16e10170) and is registered at Clinical Trials
(NCT029441757) [19]. Firefighters who met the eligibility criteria
at the time of enrollment were invited to participate. In accor-
dance with the Declaration of Helsinki, all potential participants
were informed of their right to refuse to participate or to with-
draw from the study at any time without retribution. Written
informed consent was obtained from all participants. More details
on this study's objective, design, and methods have been pub-
lished previously [20].
For the present analysis, we included participants who
completed a baseline lifestyle questionnaire between November
28, 2016 and April 16, 2018 (n ¼265) and excluded participants
with a missing FFQ or biochemical assessment (n ¼3) or whose
total daily energy intake exceeded predefined limits (men:
800e5000 kcal/d, women: 500e3500 kcal/d) [21](n¼13). After
exclusions, a total of 249 participants were left for evaluation.
2.2. The Mediterranean lifestyle (MEDLIFE) index
Dietary intake was collected at baseline using a validated 131-
item semi-quantitative 2007 grid Harvard food-frequency ques-
tionnaire (FFQ), also known as the Willett FFQ. This FFQ reflects the
previous year's habitual intake and has shown a mean correlation
coefficient of 0.59 with energy-adjusted nutrient intakes measured
by diet records [22]. Additional information on eating habits and
dietary behaviors, including a validated and previously employed
13-item modified Mediterranean diet score (mMDS), sleep behav-
iors, and physical activity were self-reported within a lifestyle
questionnaire administered at baseline for the participants to
complete and return online of their own accord [23e26].
The MEDLIFE index consists of 28 items divided into three
blocks describing food consumption; traditional dietary habits
(frugality; moderation; locally grown, biodiverse, seasonal, and
traditional products; culinary practices; conviviality during meals)
and physical activity, rest and social interactions [15]. Eleven
modifications and 2 exclusions were made to the original MEDLIFE
index items and criteria to best fit the available data from the
baseline questionnaires in the Feeding America's Bravest trial
(Table S1). Each item was weighted equally with 0 or 1 point,
creating a theoretical scoring range from 0 (worst) to 26 (best).
Final scores were then categorized into tertiles of MEDLIFE
adherence.
Abbreviations
MEDLIFE Mediterranean lifestyle
CVD Cardiovascular disease
T2DM Type 2 Diabetes Mellitus
HDL High-density lipoprotein
US United States
NHANES National Health and Nutrition Examination Survey
FFQ food frequency questionnaire
mMDS modified Mediterranean diet score
LDL low-density lipoprotein
BMI body mass index
total-c total cholesterol
PSM Public Safety Medical
SD Standard deviation
OR odds ratio
CI confidence intervals
SUN Seguimiento de Universidad de Navarra
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2495
2.3. Outcome assessment
The primary outcome of this analysis was metabolic syndrome.
The harmonized definition of metabolic syndrome established in
2009 requires meeting at least three of the following five criteria;
abdominal obesity (waist circumference 102 cm in me n, or
88 cm in women); elevated blood glucose (100 mg/dl or
treatment with antidiabetic drugs); high blood pressure (systolic
130 mm Hg or d iastolic 85 mm Hg, or receiving antihyper-
tensive drugs); triglycerides 150 mg/dl; serum HDL cholesterol
<40 mg/dl in men or <50 mg/dl in women [1]. Secondary out-
comes in our assessment included each of the five potential
metabolic syndrome components. Additionally, we evaluated
body mass index (BMI), waist circumference, percent body fat,
triglycerides, total cholesterol, high density lipoprotein (HDL)
cholesterol, low density lipoprotein (LDL) cholesterol, total
cholesterol (total-c):HDL cholesterol, and plasma glucose levels as
continuous outcomes.
Anthropometric measurements were collected by the study
team at the time of enrollment, which marked the participants'
baseline visit. Baseline lipid panels were collected separately
during participants’fire department medical examinations, which
were conducted at the contracted Public Safety Medical (PSM)
clinics independent of the research study. Blood samples were
collected after an overnight fast. Plasma and serum samples
were collected in 15-mL tubes as appropriate for each assay, ali-
quoted, frozen at 80
C, and then stored. Blood lipid profiles
were determined using standardized automated high-throughput
enzymatic analyses, which achieved coefficients of variation of
3% for cholesterol and 5% for triglycerides, using cholesterol
assay kit and reagents Ref:7D62e21 and triglyceride assay kit and
reagents Ref:7D74e21 by ARCHITECT c System, Abbott Labora-
tories, IL, USA. Baseline measures were gathered from the PSM
electronic medical record database within the last year from
enrollment in the study.
2.4. Covariate assessment
Information on sociodemographic characteristics, diet and
supplement intake, lifestyle habits, anthropometric measure-
ments, and medical history were collected at baseline through in-
person data collection, an online lifestyle questionnaire, and
medical records after informed consent was given. BMI was
calculated by dividing weight by height squared (kg/m
2
). Total
daily energy and micronutrient intakes were calculated using the
baseline FFQ. Additional alcohol intake was defined as servings of
beer and distilled beverages per week because wine consumption
was already included in the MEDLIFE index. Participants with
dyslipidemia, hypertension, or T2DM self-reported a previous
diagnosis of these conditions or treatment with lipid-lowering,
antihypertensive, or antidiabetic medications within the previ-
ous year prior to enrollment, respectively.
2.5. Statistical analysis
Variables with quantitative values were expressed as
means ±standard deviation (SD) and those characterized quali-
tatively as a percentage. The inverse probability weighting
method was used to present age-, sex-, and energy intake-
adjusted baseline characteristics of participants, as well as age-
and-sex adjusted MEDLIFE characteristics, according to tertiles
of MEDLIFE adherence. Statistical significance of between-group
comparisons for each characteristic was tested with a post-
estimation contrast of adjusted means across MEDLIFE tertiles.
To determine the contribution of each block to the between-
person variance of total MEDLIFE scores, a linear regression with
Shapley and Owen decomposition of R
2
analysis was conducted
[27]. The R
2
as a percentage identifies each block's contribution to
the total variability of MEDLIFE scores.
Multivariable logistic regression models were used to assess
the association between metabolic syndrome and adherence to
the MEDLIFE index. We also assessed the association between the
MEDLIFE index and each component of metabolic syndrome:
abdominal obesity, hyperglycemia, hypertension, hyper-
triglyceridemia, and low HDL cholesterol. Odds ratios (OR) were
reported with 95% confidence intervals (CI) and linear p-for-
trends calculated across tertile medians for each model. To control
for potential confounding, an initial multivariable adjusted model
included age (years), sex (M/F), BMI (kg/m
2
), total energy intake
(kcal/d), smoking status (never, current, or former), and education
level (technical school, some college, associate's degree/Bachelor's
degree or higher). A final multivariable model additionally
adjusted for potential confounders, including alcohol intake other
than wine (g/d), marital status (married/single), multivitamin use,
supplement use, sleep medication use (yes/no), prevalent T2DM,
hypertension, and dyslipidemia (yes/no). Possible confounders,
including BMI, prevalent T2DM, hypertension, and dyslipidemia,
were excluded from models with corresponding outcomes,
respectively.
Multivariable linear regression models were used to determine
the extent to which each tertile of MEDLIFE adherence predicted
continuous outcomes, including BMI, waist circumference, body
fat percentage, triglycerides, total cholesterol, HDL cholesterol,
LDL cholesterol, total-c:HDL cholesterol, and plasma glucose
levels. Beta coefficients with 95%CI and p-for-trends were re-
ported across MEDLIFE tertiles for each model. An initial multi-
variable model adjusted for age, sex, BMI, total energy intake,
smoking status, and education level. A fully adjusted model
additionally adjusted for other sources of alcohol intake different
from wine, civil status, multivitamin use, supplement use, sleep
medication, prevalent hypertension, dyslipidemia, and T2DM.
Lastly, multivariable logistic regression models were con-
ducted to assess the effect of each item (1 pt vs 0 pts.), block, and
additional point of the MEDLIFE score (as continuous variable) on
metabolic syndrome, adjusting for age, sex, total daily energy
intake, other sources of alcohol intake different from wine,
smoking status, education level, civil status, multivitamin use,
supplement use, sleep medication, and the remaining items or
blocks, respectively.
Sensitivity analyses were conducted for metabolic syndrome
across MEDLIFE tertiles with additional exclusions of women
(n ¼13), participants reporting caloric intake beyond Willett's total
daily energy intake limits; 800e4000 kcal/d for men and
500e3500 kcal/d for women [21](n¼20), participants with
baseline clinical measurements earlier than 6 months prior to
enrollment (n ¼126) and those with prevalent hypertension,
T2DM, or dyslipidemia (n ¼50). Additional subgroup analyses on
metabolic syndrome were conducted for age (median cut-off
point ¼<47 years, 47 years), BMI (<30 kg/m
2
,30 kg/m
2
), total
daily energy intake (median cut-off point ¼<2204, 2204 kcal/d)
and smoking status (never/former or current), with corresponding
p-for-trend and p-for-interaction across T1 as the reference cate-
gory and T2 þT3 as a single category for higher MEDLIFE adher-
ence. Lastly, a substitution of Block 1: Mediterranean food
consumption for the mMDS was conducted to further test our
primary findings.
All analyses were conducted with Stata version 14.0 (StataCorp,
College Station, TX). All p-values are two-sided and were consid-
ered statistically significant at p <0.05.
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2496
3. Results
3.1. Study population characteristics
For included participants, MEDLIFE scores ranged from 2 to 17
points with a mean score of 8.8 ±3.0 points. The average age of the
total study population was 47 ±7.6 years, ranging from 30 to 62
years. Males represented 95% of the study population with an
average total daily energy intake of 2395 ±909 kcal per day,
whereas females had an average intake of 1886 ±662 kcal per day.
The prevalence of metabolic syndrome was 17.7%, abdominal
obesity 38.2%, hyperglycemia 37.0%, hypertension 7.2%, hyper-
triglyceridemia 26.5%, and low HDL cholesterol 24.5%.
Adjusted baseline characteristics of the participants (n ¼249)
are presented in Table 1 according to tertiles of MEDLIFE adherence.
Higher MEDLIFE adherence was significantly associated with
increased mMDS scores, whole grains, total fiber intake, and sup-
plement use, as well as decreased total daily energy intake, added
sugar, saturated fat, sodium intake, and sleep medication use.
MEDLIFE characteristics presented in Table S2 indicate that, as ex-
pected, higher MEDLIFE adherence was characterized by an
increased consumption of legumes, fish, nuts, fruits, vegetables,
olive oil, water, coffee, and tea, <2.3 g/d of sodium, whole grains,
local, seasonal, or organic products, heavy exercise, 3 naps per
week, 6e8 h of sleep per day, and 4 h of television per week.
Furthermore, MEDLIFE was significantly associated with decreased
Table 1
Adjusted baseline characteristics of participants according to MEDLIFE tertiles in Feeding America's Bravest.
Characteristic MEDLIFE adherence p-value
T1 (2e7 pts.) T2 (8e10 pts.) T3 (11e17 pts.)
N 90 99 60
MEDLIFE score (pts) 5.82 (1.15) 8.99 (0.83) 12.71 (1.58) <0.001
Female (%) 2.22 7.07 6.67 0.28
Age (yrs) 46.92 (6.98) 46.66 (7.57) 46.56 (8.08) 0.95
BMI (kg/m
2
) 29.82 (4.08) 30.20 (4.51) 28.91 (3.98) 0.19
mMDS
a
(pts) 19.22 (6.60) 24.75 (5.66) 28.83 (4.77) <0.001
Total daily energy intake (kcal/d) 2660.73 (927.98) 2250.27 (920.84) 2124.58 (717.10) <0.001
Protein intake (g/d) 101.25 (40.39) 106.61 (45.77) 100.73 (30.11) 0.63
Vegetable protein (g/d) 28.41 (11.69) 33.33 (16.43) 32.60 (11.77) 0.05
Animal protein (g/d) 72.84 (31.52) 73.28 (34.41) 68.13 (25.22) 0.52
Carbohydrate intake (g/d) 258.40 (102.73) 257.61 (107.78) 244.05 (96.28) 0.74
Whole grains (g/d) 31.88 (18.40) 40.81 (23.39) 37.25 (24.04) 0.02
Total fiber intake (g/d) 20.70 (8.65) 25.74 (11.03) 28.46 (10.03) <0.001
Added sugar (g/d) 70.28 (45.78) 56.00 (29.96) 43.19 (31.32) <0.001
Fat intake (g/d) 101.43 (38.70) 97.36 (45.30) 93.81 (41.80) 0.57
Saturated fat (g/d) 34.47 (14.44) 31.37 (14.78) 27.51 (11.63) 0.008
Polyunsaturated fat (g/d) 21.58 (8.36) 21.11 (11.02) 20.08 (8.73) 0.67
Monounsaturated fat (g/d) 37.53 (15.13) 37.00 (17.73) 38.75 (20.49) 0.89
Total micronutrient intake
Sodium (g/d) 2.88 (1.08) 2.85 (1.38) 2.34 (8.81) 0.006
Calcium (mg)
c
1103.2 (997.2) 1118.2 (57.0) 1119.2 (68.9) 0.98
Iodine (mcg) 14.67 (4.97) 16.05 (4.83) 16.20 (5.97) 0.97
Zinc (mg) 1.17 (0.06) 1.35 (0.9) 1.29 (0.09) 0.22
Iron (mg)
c
15.4 (0.8) 16.8 (0.8) 16.5 (1.1) 0.48
Vitamin B12 (mcg)
c
61.5 (21.7) 89.3 (23.5) 54.4 (24.3) 0.54
Vitamin D (IU)
c
379.7 (36.8) 528.3 (63.2) 658.7 (92.0) 0.006
Vitamin C (mg)
c
146.4 (15.7) 195.7 (22.9) 290.0 (48.6) 0.01
Nondrinkers (%) 8.83 8.80 9.09 0.99
Alcohol
b
(g/d) 10.27 (14.92) 12.28 (17.84) 20.14 (40.29) 0.37
Smoking status (%) 0.79
never 53.81 59.58 49.33
current 14.44 14.98 17.65
former 31.75 25.43 33.02
Education (%) 0.68
Technical school/some college/associates degree 61.17 61.04 67.93
Bachelor's degree or higher 38.83 38.96 32.07
Civil status (%) 0.55
married 78.38 82.80 75.24
single 21.62 17.20 24.76
Multivitamin use (%) 30.82 43.05 43.94 0.19
Supplement use (proteins, glutamine, amino acids, etc.) (%) 22.01 35.33 45.49 0.01
Sleep medication use (%) 20.83 16.91 4.42 0.03
Prevalent chronic disease
d
(%) 17.24 23.69 17.26 0.49
BMI: body mass index, d:day, g: gram, kg: kilograms, h: hour, kcal: kilocalories, m: meters, pts: points, yrs: years.
Boldface indicates statistical significance.
Characteristics are adjusted for age, sex, and total daily energy intake using the inverse probability weighting method, with the exception of age, sex, and total daily energy
intake.
Continuous variables are expressed as mean (SD) and categorical variables as a percentage.
a
Modified Mediterranean diet score previously defined for this study population [28].
b
Alcohol includes beer and distilled alcoholic beverages.
c
Includes dietary supplements.
d
Previous diagnosis or treatment for hypertension (n ¼15), dyslipidemia (n ¼34), or T2DM (n ¼4).
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2497
intakes of sweets, red meat, processed meat, potatoes, dairy
products, cereals, snacks, and sugary beverages.
3.2. MEDLIFE scores
Frequency of points awarded for each MEDLIFE item (Table S3)
showed that sleep (85.1%), vegetables (65.5%), and whole grain
products (65.5%) were the most frequently awarded MEDLIFE
items, whereas processed meat (1.6%), wine (10.4%), nuts (12.1%),
physical activity (12.5%), and olive oil (12.9%) were the least scored
items. The contribution of each block to the total between-person
variability (%R
2
) of MEDLIFE scores was the following: Block 1:
Mediterranean food consumption ¼48.0%, Block 2: Mediterranean
dietary habits ¼36.3%, and Block 3: Physical activity, rest, social
habits, and conviviality ¼15.7%, which explained 98% of the total
MEDLIFE variance.
3.3. The MEDLIFE index on metabolic syndrome and its components
Figure 1 shows OR (95%CI) for metabolic syndrome and its five
components when comparing higher MEDLIFE adherence (T3),
scores 11 to 17 points, with lower adherence (T1), scores 2 to 7
points. Metabolic syndrome was 71% less likely among participants
with higher MEDLIFE adherence compared to those with lower
MEDLIFE adherence (OR ¼0.29; 95%CI: 0.10 to 0.90, p for
trend ¼0.04). Sensitivity analysis for the exclusion of women,
Willett's predefined energy intake limits, baseline clinical mea-
surements earlier than 6 months prior to enrollment, and prevalent
chronic disease did not show any significant deviations from our
primary findings. We did not find any significant interactions with
age, BMI, total daily energy intake, and smoking status (p for
interaction >0.05). The substitution of Block 1 with the mMDS
further supported the robustness of our findings for the MEDLIFE
index (T3 vs. T1: OR ¼0.37; 95%CI: 0.14 to 0.98). Furthermore,
higher MEDLIFE scores were significantly associated with lower
levels of abdominal obesity and hypertriglyceridemia. All models
and tertile estimates are provided in Table S4.
3.4. The MEDLIFE index on continuous outcomes
Table 2 shows the results of multivariable linear regression
models by MEDLIFE tertiles. Participants in T3 showed significantly
lower averages of total cholesterol, LDL cholesterol and total-c:HDL
cholesterol ratio compared to participants in T1. Although statisti-
cal significance was not observed in fully adjusted models, higher
MEDLIFE adherence was inversely associated with waist circum-
ference, triglycerides, and HDL cholesterol in the partially adjusted
models. No associations were found in neither the crude nor the
multivariable models for BMI, body fat, and plasma glucose levels.
3.5. The MEDLIFE items and blocks on metabolic syndrome
The individual MEDLIFE items and their association with
metabolic syndrome are shown in Fig. 2, which show the OR (95%
CI) for each item, block, and MEDLIFE index for each additional
point. Inverse associations on metabolic syndrome were observed
for preference for whole grain products and watching television
4 h per week, Block 2: Mediterranean dietary habits, and for each
Fig. 1. Odds ratios (OR) and 95% confidence intervals (CI) of the MEDLIFE index (T3 vs T1) on metabolic syndrome and its components. Adjusted for age, sex, BMI, total daily energy
intake, alcohol intake (excluding wine), smoking status, education level, civil status, multivitamin use, supplement use, sleep medication, prevalent hypertension, dyslipidemia, and
T2DM. *adjusted for all covariables in the model with the exclusion of BMI, T2DM, hypertension, or dyslipidemia, respectively. CI: confidence intervals, HDL: high density lipo-
protein, MetSyn: metabolic syndrome, OR: odds ratio. Table S4 shows the OR, 95%CI and p for trends across MEDLIFE tertiles for crude and multivariable adjusted models.
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2498
additional point of the MEDLIFE index. No association was found
between Block 1: Mediterranean food consumption nor Block 3:
Physical activity, rest, social habits, and conviviality and metabolic
syndrome.
4. Discussion
4.1. Principal findings
In a cross-sectional study with US career firefighters, those with
better adherence to MEDLIFE exhibited a 71% lower odds of
metabolic syndrome compared to participants with poorer adher-
ence. MEDLIFE was inversely associated with abdominal obesity
and hypertriglyceridemia, as well as total cholesterol, LDL choles-
terol, and total-c:HDL cholesterol ratio, suggesting a more favorable
cardiometabolic profile. Additional sensitivity analyses further
supported the robustness of our findings.
4.2. Existing evidence and significance of the MEDLIFE index
To the best of our knowledge, the validated MEDLIFE index has
been previously employed in a cross-sectional study in a Croatian
working population, a Spanish cohort of university graduates, and a
representative cohort of the adult Spanish population [17,18,29].
Among 366 Croatian oil and gas company workers, positive asso-
ciations were found with protective CVD factors when comparing
Q4 (16e19 points) vs Q1-Q3 (8e15 points), including body fat
percentage within acceptable range (OR ¼1.3; 95% CI ¼1.1 to 1.6),
lower total cholesterol (OR ¼1.2; 95% CI ¼1.0 to 1.4), and HDL
cholesterol higher than recommended (OR ¼1.9; 95% CI ¼1.0 to
3.6) [17]. In the Seguimiento Universidad de Navarra (SUN) cohort
with 20,494 participants, high adherence (>14 points) to MEDLIFE
was associated with a 41% decreased risk of all-cause mortality
(hazard ratio (HR) ¼0.59; 95% CI: 0.42e0.83, p<0.001 for trend)
and a 65% decreased risk for CVD death (HR ¼0.35; 95% CI:
0.14e0.85, p for trend ¼0.006) compared to low adherence (3e10
points) [18]. In addition, this same study population demonstrated
MEDLIFE was inversely associated with primary CVD events
(HR ¼0.50; 95%CI: 0.31e0.81) when comparing highest scores
(14e23 points) to lowest scores (0e9 points) [30]. More recently,
MEDLIFE was inversely associated with metabolic syndrome,
abdominal obesity, low HDL-cholesterol, HOMA-IR, and C-reactive
protein, as well as total mortality (HR
Q4vs Q1
¼0.58; 95%CI:
0.37e0.90) and CVD mortality (HR
Q4vs Q1
¼0.33; 95%CI: 0.11e1.02)
Table 2
Multivariable linear regression coefficients (
b
) and 95% confidence intervals (CI) according to MEDLIFE tertiles.
Categories of adherence to MEDLIFE p for trend
T1 (2e7 pts.) T2 (8e10 pts.) T3 (11e17 pts.)
BMI (kg/m
2
):
Crude model (95% CI) 1 Ref. 0.13 (1.36 to 1.11) 1.37 (2.78 to 0.05) 0.02
Multivariable adjusted model (95% CI)
a,c
1 Ref. 0.44 (0.78 to 1.66) 0.80 (2.20 to 0.60) 0.07
Multivariable adjusted model (95% CI)
b,c
1 Ref. 0.29 (0.94 to 1.53) 0.98 (2.45 to 0.50) 0.08
Waist circumference (cm):
Crude model (95% CI) 1 Ref. 1.48 (4.82 to 1.87) 5.82 (9.65 to 2.00) 0.004
Multivariable adjusted model (95% CI)
a
1 Ref. 0.65 (2.16 to 0.85) 2.30 (4.04 to 0.57) 0.01
Multivariable adjusted model (95% CI)
b
1 Ref. 0.48 (2.02 to 1.06) 1.77 (3.61 to 0.08) 0.07
Body fat (%):
Crude model (95% CI) 1 Ref. 0.62 (1.34 to 2.57) 1.54 (3.78 to 0.70) 0.24
Multivariable adjusted model (95% CI)
a
1 Ref. 0.21 (0.78 to 1.20) 0.39 (1.53 to 0.75) 0.57
Multivariable adjusted model (95% CI)
b
1 Ref. 0.53 (0.46 to 1.52) 0.05 (1.13 to 1.23) 0.82
Triglycerides (mg/dL):
Crude model (95% CI) 1 Ref. 21.05 (40.43 to 1.66) 32.87 (55.05 to 10.68) 0.003
Multivariable adjusted model (95% CI)
a
1 Ref. 16.35 (35.21 to 2.51) 23.99 (45.73 to 2.26) 0.03
Multivariable adjusted model (95% CI)
b
1 Ref. 15.84 (35.30 to 3.63) 21.89 (45.17 to 1.39) 0.05
Total cholesterol (mg/dL):
Crude model (95% CI) 1 Ref. 6.17 (16.42 to 4.09) 13.30 (25.04 to 1.56) 0.03
Multivariable adjusted model (95% CI)
a
1 Ref. 6.45 (16.98 to 4.08) 14.20 (26.33 to 2.06) 0.02
Multivariable adjusted model (95% CI)
b
1 Ref. 4.85 (15.36 to 5.66) 14.31 (26.88 to 1.74) 0.03
HDL cholesterol (mg/dL):
Crude model (95% CI) 1 Ref. 1.06 (1.98 to 4.11) 5.88 (2.39e9.37) 0.002
Multivariable adjusted model (95% CI)
a
1 Ref. 0.23 (2.71 to 3.16) 4.51 (1.13e7.89) 0.02
Multivariable adjusted model (95% CI)
b
1 Ref. 0.35 (2.51 to 3.21) 3.18 (0.25 to 6.60) 0.09
LDL cholesterol (mg/dL):
Crude model (95% CI) 1 Ref. 2.70 (11.74 to 6.33) 11.28 (21.61 to 0.94) 0.04
Multivariable adjusted model (95% CI)
a
1 Ref. 2.70 (12.02 to 6.60) 12.23 (22.96 to 1.50) 0.03
Multivariable adjusted model (95% CI)
b
1 Ref. 1.49 (10.93 to 7.94) 11.80 (23.09 to 0.52) 0.05
Total-c:HDL cholesterol:
Crude model (95% CI) 1 Ref. 0.27 (0.55 to 0.01) 0.75 (1.08 to 0.43) <0.001
Multivariable adjusted model (95% CI)
a
1 Ref. 0.20 (0.48 to 0.75) 0.65 (0.97 to 0.33) <0.001
Multivariable adjusted model (95% CI)
b
1 Ref. 0.20 (0.48 to 0.09) 0.60 (0.94 to 0.25) 0.001
Glucose (mg/dL):
Crude model (95% CI) 1 Ref. 0.86 (7.04 to 5.32) 2.23 (4.84 to 9.30) 0.59
Multivariable adjusted model (95% CI)
a
1 Ref. 0.77 (6.97 to 5.43) 2.78 (4.37 to 9.92) 0.50
Multivariable adjusted model (95% CI)
b
1 Ref. 0.75 (7.12 to 5.62) 2.88 (4.74 to 10.50) 0.51
BMI: body mass index, CI: confidence interval, mMDS: modified Mediterranean diet score, total-c: total cholesterol.
Boldface indicates statistical significance (p <0.05).
a
Adjusted for age, sex, BMI, total daily energy intake, smoking status, and education level.
b
Adjusted for age, sex, BMI, total daily energy intake, alcohol intake (excluding wine), smoking status, education level, civil status, multivitamin use, supplement use, sleep
medication, prevalent hypertension, dyslipidemia, and T2DM.
c
Adjusted for all variables in the model except BMI.
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2499
among an adult Spanish population [31]. These findings provide
evidence for MEDLIFE on CVD risk and mortality among Mediter-
ranean populations.
Lifestyle modifications for the prevention and management of
metabolic syndrome most frequently consider the joint effect of
energy restricted diets and physical activity for weight loss pro-
motion and subsequent improvement of metabolic syndrome
components [32e36]. Limited evidence is available on more
comprehensive lifestyle patterns beyond these two factors [37e41].
The SUN cohort employed a healthy lifestyle score (HLS) comprised
of nine habits. After a minimum of six years of follow-up among
participants initially free of metabolic syndrome, the highest cate-
gory (7e9 points) was associated with a reduced risk of developing
metabolic syndrome (OR ¼0.66; 95% CI: 0.47e0.93) compared to
the lowest category (0e3 points) [39]. This index, among others,
addresses a general healthy lifestyle, whereas the MEDLIFE index
specifically measures the traditional Mediterranean lifestyle.
4.3. Significance of our findings
Our study employed the unique MEDLIFE index, which con-
siders the combined effect of numerous lifestyle factors associated
with CVD risk in addition to diet and leisure time activity [16]. We
observed a significant inverse association of MEDLIFE with meta-
bolic syndrome when considered as the sum of all 26 items. Overall
lifestyle patterns may capture both direct and indirect underlying
effects of numerous lifestyle factors on metabolic syndrome
[42e46]. The possible synergism produced by the combination of
several components may create an effect greater than the sum of
the individual effects [47]. This overall effect may be attributed to
biological mechanisms supporting healthy physiological pathways
and molecular mechanisms that combat chronic stress and
inflammation. The healthy functioning of systems impede disrup-
tions to the autonomic nervous system, hypothal-
amicepituitaryeadrenal axis, cardiovascular, metabolic, and
immune systems, which contribute to the biochemical changes
characteristic of metabolic syndrome [43].
4.4. Strengths and limitations
Strengths of this study include the aim of the MEDLIFE index to
holistically capture the multifactorial etiology of chronic lifestyle
diseases, which have risen with the cultural divergence from
traditional ways of living [13,14]. Low adherence to the Mediter-
ranean diet was attributed to specific items, including processed
meat consumption (1.6%), wine (10.4%), and olive oil (12.9%), which
observed lower frequencies of adherence compared to previous
MEDLIFE studies [15e18,31]. Although cultural relevance of the
Mediterranean diet in non-Mediterranean populations may be
debated, our findings are robust and the MEDLIFE items support the
proposed shifts to improve food, beverage, physical activity and
other lifestyle factors in the US for general health promotion [48].
The MEDLIFE index may be particularly useful among high risk
populations, including US career firefighters, who have shown a
surprisingly high prevalence of metabolic syndrome [49,50].
Limitations of this study include the cross-sectional nature,
which hinders the ability to infer causality, since the temporal
sequence is not well defined. It is possible that part of the effect we
Fig. 2. Odds ratio (OR) and 95% confidence intervals (CI) for each item (1 pt vs 0 pts), block, and for each additional point in MEDLIFE scores on metabolic syndrome. All models were
adjusted for age, sex, total daily energy intake, alcohol intake (excluding wine), smoking status, education level, civil status, multivitamin use, supplement use, sleep medication, and
the remaining items or blocks, respectively. MEDLIFE item 3 for processed meat 1 serving/wk is not presented due to the negligible number of participants who met the criteria
(n ¼4) the regression could not be conducted. Mediterranean food consumption is comprised of block 1 items, Mediterranean dietary habits is comprised of block 2 items, physical
activity, rest, social habits, and conviviality is comprised of block 3 items of MEDLIFE.
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2500
have observed is a consequence of metabolic syndrome itself.
However, given previous knowledge, less healthy lifestyle behav-
iors seem to be more likely causes than consequences of metabolic
syndrome. Another caveat might be a possible misclassification
bias, due to the self-reported data. Nonetheless, our measure of
physical activity showed a moderate correlation with maximal
oxygen uptake (V02max) (r ¼0.41) in this study population
[26,51,52]. Moreover, the MEDLIFE index showed good to moderate
concordance for nearly 60% of items (kappa ¼0.41e1) in another
population [16]. In addition, completion of the lifestyle question-
naire by all enrolled participants would have increased sample size
and provided greater statistical power. The observed prevalence of
metabolic syndrome was almost 18%, therefore the ORs may have
overestimated the relative risk. In any case, this study may be used
to establish hypotheses for larger longitudinal cohorts or inter-
vention trials, while limiting the possibility of reverse causality
[53]. Despite the multivariable adjustments, residual confounding
cannot be completely eliminated [54].
Our findings were specific to Midwestern US career firefighters,
81% of which identified as Caucasian, which are not representative
of the general US population. Moreover, due to the predominately
male prevalence of the firefighter profession, our results must be
extrapolated to women with precaution. The association between
lifestyle and metabolic syndrome may vary according to sex and
cultural differences, particularly through social components when
comparing collectivistic and individualistic cultures [42,43]. Future
studies with greater sample size should include greater socio-
demographic (i.e. geographical, ethnic, gender, etc.) and cultural
diversity, including the roles of extended and immediate family on
lifestyle behaviors. Even though our participants were primarily
healthy middle-aged men, MEDLIFE criteria may not adequately
consider lifestyle behaviors recommended to individuals with
health conditions [16].
Furthermore, MEDLIFE items, block classification, and scoring
criteria may be debatable, nonetheless, these relied on the inde-
pendent recommendations from the Mediterranean Diet Founda-
tion for intake and behavioral cut off points. A more comprehensive
data collection may have allowed for greater reproducibility of the
original MEDLIFE index (i.e. herbs and spices, socializing with
friends, collective sports). Nonetheless, the overall index in this
study has holistically reflected the original MEDLIFE. Although a
significant increase in supplementation use was observed among
higher MEDLIFE adherers and evidence suggests nutraceutical
supplementation with a calorie-restricted Mediterranean diet and
lifestyle modifications improves cardiometabolic risk factors,
MEDLIFE does not include dietary supplements within the tradi-
tional MEDLIFE recommendations targeted at a healthy adult
population [55].
5. Conclusion
Adherence to MEDLIFE in a population of US career firefighters
was significantly associated with a decreased prevalence of meta-
bolic syndrome. Future findings from prospective studies on overall
healthy lifestyle patterns, beyond diet and physical activity, could
positively contribute towards the implementation of effective
public health strategies for the primordial prevention of metabolic
syndrome, its components, and other chronic diseases in non-
Mediterranean populations, particularly among those with high
CVD and T2DM risk.
Funding statement
This research was funded by the U.S. Department of Homeland
Security, grant number EMW-2014-FP-00612.
Author contribution
Conceptualization, M.R.-C and M.S.-P; methodology, M.R.-C;
formal analysis, M.S.H and M.R.-C; resources and data curation,
M.S.-P, S.M, and C.C.; writingdoriginal draft preparation, M.S.H;
writingdreview and editing, M.S.H, M.S.-P, M.R.-C, M.A.M.-G., C.C,
and S.N.K; visualization, M.R.-C, M.S.-P, and C.C.; supervision and
project administration, M.R.-C, M.S.-P, S.M and S.N.K; funding
acquisition, S.N.K. All authors had full access to all of the data in this
study and take complete responsibility for the integrity of the data,
the accuracy of the data analysis, and have read and agreed to the
published version of the manuscript.
Conflicts of interest
Commercial sponsors of this study are: Kroger Company (cou-
pons and customer loyalty discounts); Barilla America (Barilla Plus
Products), Arianna Trading Company, Innoliva and Molino de Zafra,
Spain (extra virgin olive oil samples and discounts) and the Almond
Board of California (free samples of roasted unsalted almonds). The
sponsors had no role in the overall study design; in the collection,
analyses, or interpretation of data; in the writing of the manuscript,
or in the decision to publish the results.
Maria Soledad Hershey, Miguel Ruiz-Canela, Mercedes Sotos-
Prieto, Miguel
Angel Martínez-Gonz
alez, Costas A Christophi and
Steven Moffatt declare no conflict of interest. Dr. Kales reports
grants from US Dept. of Homeland Security, non-financial support
from Barilla America, non-financial support from California Almond
Board, non-financial support from Arianna Trading Company, non-
financial support from Innoliva/Molina de Zafra, during the conduct
of the study; personal fees from Medicolegal Consulting, personal
fees from Mediterranean Diet Roundtable, outside the submitted
work.
Acknowledgment
The authors acknowledge the Advisory Board and the data
advisory monitoring board (DAMB), Indianapolis Fire Department
(IFD), Fishers Fire Department, Indianapolis Local 416 support,
and the National Fire Organizations; International Association of
Fire Fighters, National Volunteer Fire Council, National Fallen
Firefighters’Foundation, The Fire Protection Research Foundation,
and International Association of Fire Chiefs that support the
research, as well as the firefighters and their families for their
participation.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.clnu.2021.03.026.
References
[1] Alberti KGMM, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, et al.
Harmonizing the metabolic syndrome. Circulation 2009;120:1640e5. https://
doi.org/10.1161/CIRCULATIONAHA.109.192644.
[2] Hirode G, Wong RJ. Trends in the prevalence of metabolic syndrome in the
United States, 2011-2016. J Am Med Assoc 2020;323:2526e8. https://doi.org/
10.1001/jama.2020.4501.
[3] Saklayen MG. The global epidemic of the metabolic syndrome. Curr Hypertens
Rep 2018;20. https://doi.org/10.1007/s11906-018-0812-z.
[4] Smith DL, Graham E, Stewart D, Mathias KC. Cardiovascular disease risk factor
changes over 5 Years among male and female US firefighters. J Occup Environ
Med 2020;62:398e402. https://doi.org/10.1097/JOM.0000000000001846.
[5] Kales SN, Smith DL. Firefighting and the heart. Circulation 2017;135:1296e9.
https://doi.org/10.1161/CIRCULATIONAHA.117.027018.
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2501
[6] Soteriades ES, Smith DL, Tsismenakis AJ, Baur DM, Kales SN. Cardiovascular
disease in US firefighters: a systematic review. Cardiol Rev 2011;19:202e15.
https://doi.org/10.1097/CRD.0b013e318215c105.
[7] Yang J, Teehan D, Farioli A, Baur DM, Smith D, Kales SN. Sudden cardiac death
among firefighters 45 years of age in the United States. Am J Cardiol
2013;112:1962. https://doi.org/10.1016/j.amjcard.2013.08.029.e7.
[8] Smith DL, Haller JM, Korre M, Fehling PC, Sampani K, Grossi Porto LG, et al.
Pathoanatomic findings associated with duty-related cardiac death in US
firefighters: a caseecontrol study. J Am Heart Assoc 2018;7. https://doi.org/
10.1161/JAHA.118.009446.
[9] Schliemann D, Woodside JV. The effectiveness of dietary workplace in-
terventions: a systematic review of systematic reviews. Publ Health Nutr
2019;22:942e55. https://doi.org/10.1017/S1368980018003750.
[10] Mulchandani R, Chandrasekaran AM, Shivashankar R, Kondal D, Agrawal A,
Panniyammakal J, et al. Effect of workplace physical activity interventions on
the cardio-metabolic health of working adults: systematic review and meta-
analysis. Int J Behav Nutr Phys Activ 2019;16:134. https://doi.org/10.1186/
s12966-019-0896-0.
[11] van Namen M, Prendergast L, Peiris C. Supervised lifestyle intervention for
people with metabolic syndrome improves outcomes and reduces individual
risk factors of metabolic syndrome: a systematic review and meta-analysis.
Metabolism 2019;101:153988. https://doi.org/10.1016/
j.metabol.2019.153988.
[12] Ekelund U, Steene-Johannessen J, Brown WJ, Fagerland MW, Owen N,
Powell KE, et al. Does physical activity attenuate, or even eliminate, the
detrimental association of sitting time with mortality? A harmonised meta-
analysis of data from more than 1 million men and women. Lancet
2016;388:1302e10. https://doi.org/10.1016/S0140-6736(16)30370-1.
[13] Bach-faig A, Berry EM, Lairon D, Reguant J, Trichopoulou A, Dernini S, et al.
Mediterranean diet pyramid today. Science and cultural updates. Publ Health
Nutr 2011;14. https://doi.org/10.1017/S1368980011002515.
[14] Diolintzi A, Panagiotakos DB, Sidossis LS. From Mediterranean diet to Medi-
terranean lifestyle: a narrative review. Publ Health Nutr 2019;22:2703e13.
https://doi.org/10.1017/S1368980019000612.
[15] Sotos-Prieto M, Moreno-Franco B, Ordov
as JM, Le
on M, Casasnovas JA,
Pe~
nalvo JL. Design and development of an instrument to measure overall
lifestyle habits for epidemiological research: the Mediterranean Lifestyle
(MEDLIFE) index. Publ Health Nutr 2015;18:959e67. https://doi.org/10.1017/
S1368980014001360.
[16] Sotos-Prieto M, Santos-Beneit G, Bodega P, Pocock S, Mattei J, Luis Pe ~
nalvo J.
Validation of a questionnaire to measure overall Mediterranean lifestyle
habits for research application: the MEDiterranean LIFEstyle index (MEDLIFE).
Nutr Hosp 2015;32:1153e63. https://doi.org/10.3305/nh.2015.32.3.9387.
[17] Pavi
ci
c
Ze
zelj S, KenCel Jovanovi
c G, Draga
s Zubalj N, Mi
covi
c V, Sesar
Z. As-
sociations between adherence to the Mediterranean diet and lifestyle
assessed with the MEDLIFE index among the working population. Int J Envi-
ron Res Publ Health 2018;15:2126. https://doi.org/10.3390/ijerph15102126.
[18] Hershey MS, Fernandez-Montero A, Sotos-Prieto M, Kales S, Gea A, Ruiz-
Estigarribia L, et al. The association between the Mediterranean lifestyle index
and all-cause mortality in the Seguimiento Universidad de Navarra cohort.
Am J Prev Med 2020;59:e239e48. https://doi.org/10.1016/
j.amepre.2020.06.014.
[19] NCT02941757. Feeding America's bravest: mediterranean diet-based in-
terventions to change firefighters' eating habits. https://ClinicaltrialsGov/
Show/NCT02941757; 2016.
[20] Sotos-Prieto M, Cash SB, Christophi C, Folta S, Moffatt S, Muegge C, et al.
Rationale and design of feeding America's bravest: Mediterranean diet-based
intervention to change firefighters' eating habits and improve cardiovascular
risk profiles. Contemp Clin Trials 2017;61:101e7. https://doi.org/10.1016/
j.cct.2017.07.010.
[21] Willett W. Nutritional epidemiology. 3rd ed. New York, NY: Oxford University
Press; 2012. https://doi.org/10.1093/acprof:oso/9780199754038.001.0001.
[22] Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett WC.
Reproducibility and validity of an expanded self-administered semi-
quantitative food frequency questionnaire among male health professionals.
Am J Epidemiol 1992;135:1114e26. https://doi.org/10.1093/
oxfordjournals.aje.a116211.
[23] Yang J, Farioli A, Korre M, Kales SN. Modified Mediterranean diet score and
cardiovascular risk in a north American working population. PLoS One
2014;9:e87539. https://doi.org/10.1371/journal.pone.0087539.
[24] Sotos-Prieto M, Christophi C, Black A, Furtado JD, Song Y, Magiatis P, et al.
Assessing validity of self-reported dietary intake within a Mediterranean diet
cluster randomized controlled trial among US firefighters. Nutrients 2019;11:
2250. https://doi.org/10.3390/nu11092250.
[25] Romanidou M, Tripsianis G, Hershey MS, Sotos-Prieto M, Christophi C,
Moffatt S, et al. Association of the modified Mediterranean diet score (mMDS)
with anthropometric and biochemical indices in US career firefighters. Nu-
trients 2020;12:3693. https://doi.org/10.3390/nu12123693.
[26] Jackson AS, Blair SN, Mahar MT, Wier LT, Ross RM, Stuteville JE. Prediction of
functional aerobic capacity without exercise testing. Med Sci Sports Exerc
1990;22:863e70. https://doi.org/10.1249/00005768-199012000-00021.
[27] Huettner F, Sunder M. Axiomatic arguments for decomposing goodness of fit
according to Shapley and Owen values. Electron J Stat 2012;6:1239e50.
https://doi.org/10.1214/12-EJS710.
[28] Hershey MS, Sotos-Prieto M, Ruiz-Canela M, Martinez-Gonzalez MA,
Cassidy A, Moffatt S, et al. Anthocyanin intake and physical activity: associ-
ations with the lipid profile of a US working population. Molecules 2020;25:
4398. https://doi.org/10.3390/molecules25194398.
[29] Zaragoza-Martí A, Caba~
nero-Martínez MJ, Hurtado-S
anchez JA, Laguna-
P
erez A, Ferrer-Cascales R. Evaluation of Mediterranean diet adherence
scores: a systematic review. BMJ Open 2018;8:e019033. https://doi.org/
10.1136/bmjopen-2017-019033.
[30] Mata-Fern
andez A, Hershey MS, Pastrana-Delgado JC, Sotos-Prieto M, Ruiz-
Canela M, Kales SN, et al. A Mediterranean lifestyle reduces the risk of car-
diovascular disease in the “Seguimiento Universidad de Navarra”(SUN)
cohort. Nutr Metabol Cardiovasc Dis 2021. https://doi.org/10.1016/
j.numecd.2021.02.022.
[31] Sotos-Prieto M, Ortol
a R, Ruiz-Canela M, Garcia-Esquinas E, Martínez-
G
omez D, Lopez-Garcia E, et al. Association between the Mediterranean life-
style, metabolic syndrome and mortality: a whole-country cohort in Spain.
Cardiovasc Diabetol 2021;20:5. https://doi.org/10.1186/s12933-020-01195-1.
[32] Magkos F, Yannakoulia M, Chan JL, Mantzoros CS. Management of the meta-
bolic syndrome and Type 2 Diabetes through lifestyle modification. Annu Rev
Nutr 2009;29:223e56. https://doi.org/10.1146/annurev-nutr-080508-141200.
[33] P
erez-Martínez P, Mikhailidis DP, Athyros VG, Bullo M, Couture P, Covas MI,
et al. Lifestyle recommendations for the prevention and management of
metabolic syndrome: an international panel recommendation. Nutr Rev
2017;75:307e26. https://doi.org/10.1093/nutrit/nux014.
[34] Salas-Salvad
o J, Díaz-L
opez A, Ruiz-Canela M, Basora J, Fit
o M, Corella D, et al.
Effect of a lifestyle intervention program with energy-restricted Mediterra-
nean diet and exercise on weight loss and cardiovascular risk factors: one-
year results of the PREDIMED-Plus trial. Diabetes Care 2019;42:777e88.
https://doi.org/10.2337/dc18-0836.
[35] Pavi
c E, Had
ziabdi
c MO, Mucalo I, Martinis I,
Z Romi
c, Bo
zikov V, et al. Effect of
the Mediterranean diet in combination with exercise on metabolic syndrome
parameters: 1-year randomized controlled trial. Int J Vitam Nutr Res 2019;89:
132e43. https://doi.org/10.1024/0300-9831/a000462.
[36] Gomez-Huelgas R, Jansen-Chaparro S, Baca-Osorio AJ, Mancera-Romero J,
Tinahones FJ, Bernal-L
opez MR. Effects of a long-term lifestyle intervention
program with Mediterranean diet and exercise for the management of pa-
tients with metabolic syndrome in a primary care setting. Eur J Intern Med
2015;26:317e23. https://doi.org/10.1016/j.ejim.2015.04.007.
[37] Yamaoka K, Tango T. Effects of lifestyle modification on metabolic syndrome:
a systematic review and meta-analysis. BMC Med 2012;10:1e10. https://
doi.org/10.1186/1741-7015-10-138.
[38] Lee JA, Cha YH, Kim SH, Park HS. Impact of combined lifestyle factors on
metabolic syndrome in Korean men. J Public Health 2017;39:82e9. https://
doi.org/10.1093/pubmed/fdv208.
[39] Garralda-Del-Villar M, Carlos-Chiller
on S, Diaz-Gutierrez J, Ruiz-Canela M,
Gea A, Martínez-Gonz
alez M, et al. Healthy lifestyle and incidence of meta-
bolic syndrome in the SUN cohort. Nutrients 2018;11:65. https://doi.org/
10.3390/nu11010065.
[40] Sotos-Prieto M, Bhupathiraju SN, Falc
on LM, Gao X, Tucker KL, Mattei J.
A healthy lifestyle score is associated with cardiometabolic and neuroendo-
crine risk factors among Puerto Rican adults. J Nutr 2015;145:1531e40.
https://doi.org/10.3945/jn.114.206391.
[41] Chang S-H, Chen M-C, Chien N-H, Wu L-Y. Examining the links between
lifestyle factors and metabolic syndrome. Am J Nurs 2016;116:26e36. https://
doi.org/10.1097/01.NAJ.0000508662.88220.7a.
[42] Kim K, Jung SJ, Baek JM, Yim HW, Jeong H, Kim DJ, et al. Associations between
social network properties and metabolic syndrome and the mediating effect
of physical activity: findings from the Cardiovascular and Metabolic Diseases
Etiology Research Center (CMERC) Cohort. BMJ Open Diabetes Res Care
2020;8:1272. https://doi.org/10.1136/bmjdrc-2020-001272.
[43] Pedersen JM, Lund R, Andersen I, Clark AJ, Prescott E, Rod NH. Psychosocial
risk factors for the metabolic syndrome: a prospective cohort study. Int J
Cardiol 2016;215:41e6. https://doi.org/10.1016/j.ijcard.2016.04.076.
[44] Borel A-L. Sleep apnea and sleep habits: relationships with metabolic syn-
drome. Nutrients 2019;11:2628. https://doi.org/10.3390/nu11112628.
[45] Qian Y, Xu H, Wang Y, Yi H, Guan J, Yin S. Obstructive sleep apnea predicts risk
of metabolic syndrome independently of obesity: a meta-analysis. Arch Med
Sci 2016;12:1077e87. https://doi.org/10.5114/aoms.2016.61914.
[46] Zhang D, Liu X, Liu Y, Sun X, Wang B, Ren Y, et al. Leisure-time physical ac-
tivity and incident metabolic syndrome: a systematic review and dose-
response meta-analysis of cohort studies. Metabolism 2017;75:36e44.
https://doi.org/10.1016/j.metabol.2017.08.001.
[47] Van Der Weele TJ, Knol MJ. A tutorial on interaction. Epidemiol Methods
2014;3:33e72. https://doi.org/10.1515/em-2013-0005.
[48] 2015 dietary Guidelines advisory committee. 2015-2020 dietary Guidelines
for Americans. https://health.gov/dietaryguidelines/2015/guidelines/; 2015.
[49] Choi B, Ko S, Kojaku S. Resting heart rate, heart rate reserve, and metabolic
syndrome in professional firefighters: a cross-sectional study. Am J Ind Med
2017;60:900e10. https://doi.org/10.1002/ajim.22752.
[50] Baur DM, Christophi CA, Kales SN. Metabolic syndrome is inversely related to
cardiorespiratory fitness in male career firefighters. J Strength Condit Res
2012;26:2331e7. https://doi.org/10.1519/JSC.0b013e31823e9b19.
[51] Jurca R, Jackson AS, LaMonte MJ, Morrow JR, Blair SN, Wareham NJ, et al.
Assessing cardiorespiratory fitness without performing exercise testing. Am J
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2502
Prev Med 2005;29:185e93. https://doi.org/10.1016/j.amepre.2005.06.004.
[52] Mailey EL, White SM, W
ojcicki TR, Szabo AN, Kramer AF, Mcauley E. Construct
validation of a non-exercise measure of cardiorespiratory fitness in older
adults. 2010.
[53] Davies HTO, Crombie IK, Tavakoli M. When can odds ratios mislead? BMJ
1998;316:989e91. https://doi.org/10.1136/bmj.316.7136.989.
[54] Rothman KJ. Modern epidemiology. 1st ed. Boston: Little, Brown; 1986.
[55] Tripp ML, Dahlberg CJ, Eliason S, Lamb JJ, Ou JJ, Gao W, et al. A low-glycemic,
Mediterranean diet and lifestyle modification program with targeted nutra-
ceuticals reduces body weight, improves cardiometabolic variables and
longevity biomarkers in overweight subjects: a 13-week observational trial.
J Med Food 2019;22:479e89. https://doi.org/10.1089/jmf.2018.0063.
M.S. Hershey, M. Sotos-Prieto, M. Ruiz-Canela et al. Clinical Nutrition 40 (2021) 2494e2503
2503
