Adherence to the Mediterranean diet is inversely associated with metabolic syndrome occurrence: a meta-analysis of observational studies.

Abstract Diet plays a role in the onset and progression of metabolic disorders, including metabolic syndrome (MetS). We aimed to systematically review and conduct a quantitative meta-analysis of results from observational cross-sectional and prospective cohort studies on adherence to the Mediterranean dietary pattern and risk of MetS. Literature databases including PubMed, SCOPUS and EMBASE were searched from the beginning to May 2016. Eight cross-sectional and four prospective studies were included in this meta-analysis, accounting for a total of 33,847 individuals and 6342 cases of MetS. High adherence to the Mediterranean diet was associated with a risk of MetS (RR: 0.81, 95%CI: 0.71, 0.92). Regarding individual components of the MetS, the inverse associations were significant for waist circumference, blood pressure and low HDL-C levels. In conclusion, adoption of a Mediterranean dietary pattern was associated with lower risk of the MetS and it can be proposed for the primary prevention of the MetS.


Introduction
The metabolic syndrome (MetS) represents a clinical condition characterized by a number of metabolic risk factors that predispose to increased risk of developing diabetes mellitus and cardiovascular disease (CVD) (Grundy et al. 2004;Alberti et al. 2009). A chronic state of subclinical inflammation seems to be a key mechanism underlying the pathophysiology of MetS. Metabolic alterations characterizing the MetS involve blood glucose, blood lipids [triglycerides and highdensity lipoprotein cholesterol (HDL-C)], blood pressure and abdominal obesity. Over the past decades, several criteria for the definition of MetS have been proposed. According the National Cholesterol Education Program-Adult Treatment Panel-III (NCEP-ATP III), diagnosis of MetS requires three or more of the following: (i) waist circumference >102 cm in men and >88 cm in women; (ii) HDL-C <40 mg/dL (<1.04 mmol/L) in men and <50 mg/dL (<1.29 mmol/L) in women; (iii) triglycerides 150 mg/dL (1.7 mmol/L); (iv) blood pressure 130/85 mmHg and (v) fasting glucose 110 mg/dL (6.1 mmol/L) (Grundy et al. 2004). More recently, the Joint Interim Societies (including the International Diabetes Federation Task Force on Epidemiology and Prevention, the National Heart, Lung, and Blood Institute, the American Heart Association, the World Heart Federation, the International Atherosclerosis Society, and the International Association for the Study of Obesity) proposed a harmonization of the criteria maintaining the same thresholds for blood pressure, triglycerides and HDL-C, but introduced population-and country-specific cutoff points for waist circumference and modified those for fasting glucose levels [100 mg/dL (5.6 mmol/L)] (Alberti et al. 2009). Despite MetS is characterized by a variety of metabolic alteration with no univocal pathogenetic mechanism, there are genetic and environmental factors that most likely are associated to this condition (Anagnostis 2012;Di Renzo et al. 2014b). Besides the resulting hormonal abnormalities, engagement in unhealthy food patterns and lifestyles, such as overnutrition and sedentary lifestyle, may worsen the clinical condition, with body fat accumulation and potential disregulation of the aforementioned parameters characterizing the MetS.
A dietary pattern following the principles of the traditional eating habits of individuals living in the Mediterranean areas during 1960s has been the focus of attention over the last decades for its potential ability to prevent and ameliorate metabolic disorders, including those associated with MetS (Grosso et al. 2014c). The Mediterranean diet is characterized by a number of key features: use of olive oil as the main or exclusive culinary fat, high intake of olive oil, fruits and nuts, vegetables, non-refined cereals and legumes as main sources of fiber and plant-derived antioxidants, such as vitamins and polyphenols (Zamora-Ros et al. 2012;Zamora-Ros et al. 2013;Grosso et al. 2014d); frequent consumption of fish as the main source of proteins and poly-unsaturated fatty acids (PUFA) (Marventano et al. 2015); a high intake of monounsaturated fatty acids (MUFA) derived from olive oil, and a high MUFA to saturated fat intake (Rondanelli et al. 2015); moderate consumption of wine (mainly red), which provides limited amount of alcohol and peculiar polyphenol compounds (such as stilbenes) (Giacosa et al. 2014); low consumption of meat and sweets, as sources of unhealthy fats, such as cholesterol and trans-fatty acids (Di Daniele et al. 2014) and low consumption of butter and cream. There are several peculiar foods characterizing diet of populations living in the Mediterranean region, such as oranges (Grosso et al. 2013b), prickly pears (Silveira et al. 2015), pomegranates (Gonzalez-Trujano et al. 2015) and artichoke (Rondanelli et al. 2013). The synergic effect of all the components of the diet, rather than any individual food or nutrient, has been considered the key for the success of this dietary pattern in improving human health and prolonging lifespan (Giacosa et al. 2013;Sofi et al. 2014). The peculiar profile of the MUFA:SFA ratio and the high intake of PUFA characterizing the Mediterranean diet, as well as the high content in fiber, antioxidants and polyphenols with anti-inflammatory properties, have been associated with lower risk of obesity and an overall better metabolic status (Abenavoli et al. 2014;Shin et al. 2015). Despite the evidence on the beneficial effect of the MedDiet on individual components of the MetS is convincing (Kastorini et al. 2011), no summary quantitative analyses have been conducted on population studies to explore whether this dietary pattern may exert preventive effects toward MetS itself. Thus, the aim of this study was to systematically review and perform meta-analysis of existing observational studies exploring the relation between adoption of a Mediterranean dietary pattern and the prevalence or the risk of MetS.

Search strategy and study selection
Literature databases including PubMed, SCOPUS and EMBASE were searched from the beginning through May 2016. Relevant keywords associated with Mediterranean diet ("mediterranean diet") were searched in combination with keywords associated with MetS ("metabolic syndrome" or "metabolic impairment" or "x syndrome" or "metabolic disease") in combination with keywords relevant to the study methods ("incidence" or "cohort" or "follow-up" or "association"). Reference lists of retrieved articles were manually searched for missing citations. The literature search was limited to English. If more than one article was published using the same cohort, the most recent article with the longest follow-up period was considered. Studies included in this meta-analysis met the following inclusion criteria: (i) evaluated the risk or association between Mediterranean diet adherence and occurrence or risk of MetS with a prospective or casecontrol/cross-sectional design; (ii) used an a priori method to evaluate adherence to the diet; (iii) clearly defined the category of exposure (high versus low adherence) and provided risk measures. Articles were excluded if they did not report sufficient statistics. Two authors (JG and GZ) independently assessed the articles for compliance with the inclusion and exclusion criteria and solved disagreements through consensus.

Data extraction
The following information was extracted from each study: (i) name of the first author; (ii) year of publication; (iii) study cohort or name; (iv) country; (v) number of participants and cases; (vi) gender and age of the study population at baseline; (vii) follow-up period; (viii) endpoints and cases; (ix) diet adherence score; (x) MetS criteria; (xi) odds ratios (ORs) or hazard ratios (HRs) with 95% confidence intervals (CIs) of MetS and its components for the highest adherence versus the lowest category of exposure and (xii) covariates used in adjustments.
The quality of observational studies was assessed according to the Newcastle-Ottawa quality assessment scale (Wells et al. 1999), consisting of three parameters of quality: selection (four points), comparability (two points) and outcome (three points), with a score of seven or more points reflecting high quality.

Exposure and outcome measures
Adherence to a Mediterranean was defined through scores that estimated the conformity of the dietary pattern of the studied population with the traditional Mediterranean dietary pattern (Davis et al. 2015). Overall, people more adherent to the Mediterranean diet were considered those included in the highest quantile of the score used in each study.
MetS was defined according the criteria set up by the study researchers in each included study. When the study provided also risk estimates for individual criteria included in the definition of MetS, we also performed separate meta-analyses for each of these outcomes to test whether the association with adherence to the Mediterranean diet was relying on a specific clinical feature rather than on the overall condition.

Statistical analysis
In this meta-analysis, ORs and HRs were deemed equivalent to relative risks (RRs) (Greenland 1987). ORs or HRs with 95% CIs for all categories of exposure were extracted for the analysis and random-effects models were used to calculate pooled RRs with 95% CIs for the highest compared with the lowest category of exposure. Heterogeneity was assessed by using the Q test and I 2 statistic. The I 2 statistic represents the amount of total variation that could be attributed to heterogeneity. I 2 values <25%, <50% and <75% indicated little, moderate and significant heterogeneity, respectively. A sensitivity analysis was conducted by excluding one study at a time and by grouping studies by design, gender, sample size, Mediterranean diet assessment tool, MetS criteria and adjustment for dietrelated (i.e. total energy intake) and health-related (i.e. history of CVD) variables. Publication bias was assessed by visual observation of funnel plot. All analyses were performed with Review Manager (RevMan) version 5.2 (Copenhagen: The Nordic Cochrane Center, The Cochrane Collaboration).

Study characteristics
The full process of identification and selection of studies is shown in Figure 1. The relevance of studies was assessed with a hierarchical approach on the basis of title, abstract and the full manuscript. Of the 21 studies considered relevant, 9 were excluded for the following reasons: 1 study had different design; 1 study evaluated different outcomes; 5 studies defined the Mediterranean dietary pattern through principal component analysis; 2 studies reported insufficient statistics. Overall, eight cross-sectional (Panagiotakos et al. 2004;Alvarez Leon et al. 2006;Babio et al. 2009;Gouveri et al. 2011;Grosso et al. 2014a;Yang et al. 2014;Grosso et al. 2015b;Veissi et al. 2016) and four prospective studies (Kesse-Guyot et al. 2013;Steffen et al. 2014;Mirmiran et al. 2015;Pimenta et al. 2015) were included in this meta-analysis, accounting for a total of 33,847 individuals and 6342 cases of MetS.

Association of Mediterranean diet and MetS
Fourteen datasets from all 12 studies were pooled together to estimate the association between adherence to the Mediterranean diet and MetS. High adherence was inversely associated with decreased risk of MetS (RR: 0.81, 95%CI: 0.71, 0.92) compared to lowest adherence ( Figure 2), with significant evidence of heterogeneity (I 2 ¼ 74%, p < .001) but no asymmetry of funnel plot (Supplementary Figure 1). Sensitivity analyses conducted by removing one study at a time showed that heterogeneity was driven mainly by one study (Babio et al. 2009), after exclusion of which results were substantially identical but without residual heterogeneity. Reason for such heterogeneity may depend on the fact that the study of Babio et al. (2009) was conducted on high CVD risk individuals, which may have a different sensibility to the potential effects of adherence to the Mediterranean diet. When considering studies by design, the association was found in both cross-sectional and prospective studies, despite with evidence of significant heterogeneity (Figure 2). Among the former, the contrasting results were due to the aforementioned study by Babio et al. (2009), while none of the prospective studies was primarily contributing to the heterogeneity. Stratification of analyses by variables of interest revealed that results were stable and significant when including studies with larger samples conducted on general population (rather than high CVD-risk patients) and using the Joint Interim Societies criteria for MetS (Table 2). Despite there was no difference in effect size either using the Trichopoulou or the Panagiotakos Mediterranean diet score, findings of studies adopting the latter definition resulted less heterogeneous ( Table 2).

Association of mediterranean diet and MetS components
A separate analysis was performed to evaluate the association between adherence to the Mediterranean diet and individual components of the MetS were provided in four studies (Alvarez Leon et al. 2006;Babio et al. 2009;Grosso et al. 2015b;Mirmiran et al. 2015) (six datasets) (Figure 3). The resulting associations were significant for waist circumference (RR ¼0.82, 95%CI: 0.70, 0.96; I 2 ¼ 22%, p for heterogeneity ¼ .27), blood pressure (RR ¼0.87, 95%CI: 0.77, 0.97; I 2 ¼ 0%, p for heterogeneity ¼ .77) and HDL-C levels (RR ¼0.87, 95%CI: 0.77, 1.00; I 2 ¼ 0%, p for heterogeneity ¼ .50) with no significant evidence of heterogeneity and asymmetry of funnel plot in any of the analysis (Supplemental Figure 1). The analysis on serum triglycerides showed an inverse non-significant association with high adherence to the Mediterranean diet (RR ¼0.84, 95%CI: 0.70, 1.01; I 2 ¼ 44%, p for heterogeneity ¼ .11) with little evidence of heterogeneity due to the results of one study (Babio et al. 2009), after exclusion of which the association was not significant. Null results were found for the association between Mediterranean diet and blood glucose criterion for MetS (Figure 3).

Discussion
In the present meta-analysis, a significant and consistent inverse association between adherence to the Mediterranean diet and the risk of MetS was found all over the population and cohort studies investigated. Previous meta-analyses on Mediterranean diet and MetS have been conducted only on a limited number of studies or only on each specific criteria but not on the overall syndrome (Kastorini et al. 2011), while this is the first time that a quantitative meta-analysis using the overall MetS as the outcome has been performed involving a large number of individuals. Among the existing clinical trials evaluating MetS as an outcome (Esposito et al. 2013), only the PREDIMED study provided data on the potential efficacy of the Mediterranean diet in preventing this condition (Salas-Salvado et al. 2008). However, the researchers showed that the administration of a Mediterranean diet supplemented with extra-virgin olive oil or nuts led to null results on the risk of MetS (despite their finding of a significantly higher reversion with MedDiet of the condition in patients affected at baseline) (Babio et al. 2014). Current evidence is thus contrasting. Despite not conclusive, our findings are suggestive of an inverse association, though they are affected by small evidence of heterogeneity. Besides, evidence from a large randomized trial supported that the Mediterranean diet is able to ameliorate the MetS traits in subjects who have already developed it (i.e. a significantly higher reversion rate was apparent). However, further clinical trials conducted on high-risk individuals are needed to confirm that this dietary pattern could be considered as a first option for the primary prevention of MetS. Among the components of the MetS, blood pressure, blood lipids and central obesity criteria resulted significantly inversely associated adherence to the Mediterranean diet. Overall, results on individual metabolic alterations are in line with previous summary analyses showing that Mediterranean diet is associated with decreased risk of hypertension (Nissensohn et al. 2016) as well as an improvements in blood lipid levels and body weight (Huo et al. 2015). However, we did not find any significant association between higher adherence to the Mediterranean diet and the blood glucose criterion. This finding is substantially in disagreement with existing meta-analyses of observational studies reporting that high adherence to this dietary pattern was associated with a reduced risk of developing type-2 diabetes (Koloverou et al. 2014). It is noteworthy that among the studies included in the aforementioned meta-analysis, those conducted in non-Mediterranean countries or involving multiethnic populations mainly reported null results. There is no clear reason for such a lack of association between a Mediterranean-type diet and diabetes risk in non-Mediterranean countries. Notably, we observed a similar lack of association also in the studies conducted in non-Mediterranean countries included in the present meta-analysis. The geographical localization of the population may condition whether or not the operational definition of the Mediterranean diet used in a study may truly capture or not the traditional Mediterranean dietary pattern. Despite a progressive shifting away from traditional dietary patterns have been documented in Mediterranean countries (De Lorenzo et al. 2001;Bonaccio et al. 2014;Grosso et al. 2014b), factors potentially related with quality of individual components of the Mediterranean diet (i.e. consumption of local organic products versus mass production) (De Lorenzo et al. 2010;Di Daniele et al. 2013;Grosso et al. 2013c), culinary practices, lifestyles linked to cultural and demographic backgrounds (Bonaccio et al. 2013;Grosso et al. 2013d;Buscemi et al. 2014), or genetic profiles (Di Daniele et al. 2014) may differ between Mediterranean and non-Mediterranean countries.
There are several Mechanisms relating the Mediterranean diet and its potential effects toward metabolic disorders. Known biological effects of the compounds characterizing this dietary pattern provide plausibility for its potential protective role against MetS. The synergic role of each of the Mediterranean diet components seems to provide protection against the chronic state of subclinical inflammation characterizing the early stage of MetS Casas et al. 2014). Vitamins and flavonoids contained in fruit and vegetable have been extensively studied for their antioxidant and inflammatory properties (Grosso et al. 2013a;Gregorio et al. 2016). Vitamin C, vitamin E and carotenoids are free-radical scavengers, which mainly benefit toward MetS depends on their antioxidant activity (Dakshinamurti 2015). Flavonoids inhibit lipid peroxidation, promote vascular relaxation and ameliorate endothelial function via promoting antiatherogenic, antithrombotic and anti-inflammatory effects (Amiot et al. 2016). Intake of red wine naturally enriched with resveratrol has been demonstrated to lead to the expression of inflammation and oxidative stress-related genes and reduce post-prandial oxidation of blood lipids (Di Renzo et al. 2014a;Di Renzo et al. 2015). Whole grains have been demonstrated to protect against metabolic disorders due to their content of fiber, which are rich in several bioactive compounds (including minerals, trace elements, vitamins, carotenoids, polyphenols and alkylresorcinols) and act themselves in ameliorating carbohydrates metabolism (Mellen et al. 2008). Whole-grain wheat is also a source of methyl donors and lipotropes (methionine, betaine, choline, inositol and folates), which are involved in lipid metabolism and cardiovascular and hepatic protection (Borneo & Leon 2012). Finally, the high unsaturated:saturated fatty acid ratio has been suggested to exert beneficial effects toward insulin sensitivity, inflammation, vascular function and thrombosis (Siri-Tarino et al. 2015). Overall, all such compounds may counteract the release of proinflammatory cytokines occurring when the excess of adipose tissue trigger the inflammation and disregulate immune function as well as insulin sensitivity, blood pressure homeostasis and lipid metabolism (Welty et al. 2016).
Results of the present study should be considered in light of some limitations. First, most of the studies included in the meta-analysis had an observational design. Thus, long-term large-scale clinical trials are needed to confirm findings from this meta-analysis. Second, we found evidence of heterogeneity across the studies, which was not entirely explained. Third, the use of different dietary scores to evaluate the adherence to the Mediterranean diet may lead to some bias due to the non-uniformity of the indices, for instance, regarding food group classification (i.e. fish and nut consumption not considered separately) and quantification of food intake (such as frequency of consumption versus portion size). Forth, as mentioned above, investigations conducted in different geographical areas may lead to substantially different dietary intakes across individuals labeled as "highly adherent" to the Mediterranean diet, as those conducted in Mediterranean countries have reported higher intakes of fish and legumes than those in non-Mediterranean countries (Grosso et al. 2015a).
In conclusion, the adoption of a Mediterranean dietary pattern may be considered for the primary prevention of the MetS, but the available results are not entirely consistent. Most components of the MetS seem to be reduced by the Mediterranean dietary pattern, with the exception that conflicting results on the blood glucose criterion have been found. Background, cultural and geographical barriers may play a role with this regard, and further efforts should be made to better identify such factors and improve the efficacy of a Mediterranean-type dietary pattern also in non-Mediterranean countries.