Dietary Patterns and Metabolic Syndrome in Adult Subjects: A Systematic Review and Meta-Analysis

1Metabolic Syndrome (MetS) has become a relevant public health concern because of its increased prevalence partially explained by aging population and lifestyle factors, including diet.

2MetS is a pathophysiological state and a cluster of interrelated factors including abdominal obesity, insulin resistance, dysglycemia, hypertension, and dyslipidemia (triglycerides and HDL-C—high-density lipoprotein cholesterol). The diagnosis of MetS requires three or more of the following criteria: (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). A harmonization of the diagnostic criteria has been proposed, as the reference thresholds for abdominal obesity vary considerably among countries and international organizations. In particular, the recommended waist circumference cutoff points are lower for both men and women in Asia, Sub-Saharan Africa, and Central and South America.

3According to literature, the consumption of specific foods or nutrients is strongly related to the risk of developing MetS. Nutritional epidemiology currently applies dietary patterns to analyze the relation of diet with chronic diseases rather than focusing on individual foods and nutrients. Dietary patterns provide a closer representation of the overall dietary habits of the population in study. The statistical methods identifying dietary patterns are distinguished in a priori and a posteriori methods. A priori approaches assign dietary indices and scores (i.e., glycemic index, Mediterranean score) based on current nutritional knowledge of positive and negative effects of various nutrients or foods and identify an optimal pattern, the adherence to which could maximize health benefit. The a priori approach can prove more advantageous only if important dietary factors have been clearly defined to affect the outcome under study. Conversely, a posteriori methods identify dietary patterns (i.e., Western and Healthy patterns) based on available dietary data directly obtained from the studied population. Their major limit is that the identified dietary pattern may be sample specific and influenced by subjective decisions. The association of MetS outcomes with a priori patterns, such as the Mediterranean diet and inflammatory diet, have been analyzed. The Mediterranean diet reduced the risk of MetS, whereas the comparison of the most pro-inflammatory diet versus the most anti-inflammatory diet showed no significant association. A recent meta-analysis had evaluated the relationship between a posteriori dietary patterns and MetS and showed a risk reduction of 11% for prudent/healthy pattern and a risk increase of 16% for Western/unhealthy pattern. A previous meta-analysis, found that an inverse association of prudent/healthy pattern and a positive association of Western/unhealthy pattern with MetS in cross-sectional studies, but not in cohort studies. Since then several other studies have been published on this topic with contrasting results. Therefore, we conducted a meta-analysis for deriving a more precise estimation of this association.

4The aim of our systematic review and meta-analysis is to investigate and provide an estimate of the association between dietary patterns defined by a posteriori methods and MetS risk in adults.

5The present meta-analysis was conducted following the MOOSE (Meta-analysis of Observational Studies in Epidemiology) guidelines and PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) statement.

6We conducted a comprehensive literature search, without restrictions, up to 31 March 2019 through PubMed (http://www.ncbi.nlm.nih.gov/pubmed/), Web of Science (http://wokinfo.com/) and Scopus (https://www.scopus.com/) databases to identify all the original articles on the association between dietary patterns and MetS. The literature search included the following search medical subject headings (MeSH) and key words: (“Metabolic Syndrome” OR MetS) AND (“dietary pattern” OR “eating pattern” OR “food pattern” OR “dietary habit” OR “dietary score” OR “dietary index” OR "nutrient pattern” OR “diet diversity” OR “diet variety” OR “diet quality” OR “diet index” OR “diet score”) AND (“factor analysis” OR “principal component analysis” OR “cluster analysis” OR clustering OR “reduced rank regression” OR “data-driven approach” OR “a posteriori method”).

7We manually examined the reference lists of selected articles and recent relevant reviews to identify possible additional relevant publications.

8Articles were included if they met the following criteria: (i) evaluated the relationship between dietary patterns derived by a posteriori methods, such as principal component analysis (PCA), factor analysis (FA), and principal component factor analysis (PCFA), and by reduced rank regression (RRR, i.e., an integration of the a priori and the a posteriori approaches) and MetS in adults; (ii) used a case-control, prospective or cross-sectional study design; (iii) reported odds ratio (OR), relative risk (RR) or hazard ratio (HR) estimates with 95% confidence intervals (CIs). For each potentially included study, two investigators independently carried out the selection, data abstraction, and quality assessment. Disagreements were resolved by discussion or in consultation with the third author. Although useful to have background information, reviews and meta-analysis were excluded. No studies were excluded for weakness of design or data quality.

9For each selected study, we extracted the following information: first author’s last name, year of publication, country, study design, sample size (when possible, number of cases and controls; cohort size and incident cases), population characteristics (sex, age), duration of follow-up for cohort studies, MetS assessment method, dietary assessment and dietary pattern identification methods (FA, PCA and PCFA), characteristics of the dietary assessment method, name given to the dietary patterns and their characteristics, cutoff points of the different categories of adherence to the dietary pattern (dichotomy, tertile, quartile and quintile), risk estimates with 95% CIs for the different categories of adherence, p-value for trend, and confounding factors adjustment. When multiple estimates were reported in the article, we pulled out those adjusted for the most confounding factors.

10The estimated overall effect-size statistic was the average of the logarithm of the observed OR (approximated to RR, when necessary) associated with the highest versus the lowest level of adherence to the different dietary patterns. We used the random-effects model to calculate the summary OR and 95% CIs. We restricted the analysis to the dietary patterns defined a posteriori. Since the labeling of the patterns is arbitrary and the dietary patterns are population-specific, we considered only those patterns sharing most foods with similar factor loadings. For the inclusion in the meta-analysis, the two most common dietary patterns with similar factor loading of principle components were identified in 38 studies (out of 40). The first dietary pattern, named “Healthy”, was characterized by a high loading of vegetables and fruit, poultry, fish, and whole grains. The selected articles labeled this pattern as “Healthy”, “Healthy Japanese”, “Health-conscious”, “Prudent”, “Balanced”, “Fruit & vegetables”, “Vegetables, fruits, cereals, and tubers”, “Traditional Chinese”, “Minimally processed/processed”, “Mixed-traditional”, “Fruits, vegetables, nuts, and legumes”, “Refined Grains & Vegetables”, “Traditional”, “Traditional Lebanese”, “Balanced Korean”, “Fruit and dairy”, “Grains, vegetables, and fish”.

11The second dietary pattern, named “Meat/Western”, had a high loading of red meat, processed meat, animal fat, eggs and sweets. The included articles labeled this pattern as “Western”, “Traditional and protein”, “Unhealthy”, “Animal food”, “Common Brazilian meal”, “Ultra-processed”, “Westernized”, “Mixed-modern”, “High-protein/cholesterol”, “Meat”, “Refined and Processed”, “Animal protein”, “Organ Meat & Poultry”, “Fat, meat and alcohol”, “High-fat/Western”, “Animal food”, “Southern”, “High-Protein”, “Semi-Western”, “Alcohol and meat”, “Processed foods”, “High-protein/fat”, “Meat and French fries”, “High glycemic index and high-fat”.

12The chi-square-based Cochran’s Q statistic and the I2 statistic were used to evaluate heterogeneity in results across studies. The I2 statistic yields results ranged from 0% to 100% (I2 = 0%–25%, no heterogeneity; I2 = 25%–50%, moderate heterogeneity; I2 = 50%–75%, large heterogeneity; and I2 = 75%–100%, extreme heterogeneity). Results of the meta-analysis may be biased if the probability of publication is dependent on the study results. We used the methods of Begg and Mazumdar and Egger et al. to detect publication bias. Both methods tested for funnel plot asymmetry, the former being based on the rank correlation between the effect estimates and their sampling variances, and the latter on a linear regression of a standard normal deviate on its precision. If a potential bias was detected, we further conducted a sensitivity analysis to assess the robustness of combined effect estimates, and the possible influence of the bias, and to have the bias corrected. We also conducted a sensitivity analysis to investigate the influence of a single study on the overall risk estimate, by omitting one study in each turn. We considered the funnel plot to be asymmetrical, if the intercept of Egger’s regression line deviated from zero, with a p-value of <0.05. The analyses were performed using the ProMeta Version 3.0 statistical program (Internovi, Via Cervese, 47522, Cesena, Italy).

13The primary literature search through PubMed (n = 90), Web of Science (n = 227) and Scopus (n = 143) databases identified a total of 460 articles. Duplicates (n = 158) were removed and the remaining 302 records were identified for title and abstract revision (Figure 1).

14Among these, 236 articles were excluded as not investigating the association between dietary patterns and the outcome of interest. Sixty-five articles were subjected to full-text revision. Hand searching of reference lists of both selected articles and recent relevant reviews led to the identification of seven additional items. Subsequently, 32 papers were excluded because they did not meet the inclusion criteria as follows: 9 studies considered a different dietary pattern as the comparison reference; 6 studies were carried out on adolescents; 5 studies reported the MetS risk combined with genotype; 4 studies derived the dietary patterns considering nutrients instead of food items; 3 studied reported the correlation instead of risk estimate; one study used a control group (no MetS) as reference; one study was carried out on transplant recipients; and one study was carried out on type 2 diabetes. Therefore, at the end of the selection process, 40 studies were enclosed for the identification of the different dietary patterns in the systematic review and meta-analysis.

15General characteristics of the 40 studies evaluating the association between adherence to a posteriori dietary patterns with MetS risk are shown in Table 1.

16These studies were published between 2007 and 2019. Eight studies were conducted in Korea, eight in Europe; six in Iran; four in the USA; three in China; two in Japan, Brazil, Samoan Islands and Lebanon; and one each in Thailand, Australia and Mexico. Four were cohort studies, one was a case-control study and all others were cross-sectional studies. Six studies were conducted on women and men separately, three were on women only and all others estimated the MetS risk on women and men together. One study did not report the parameters used to identify the MetS.

17Thirty-one studies used a food frequency questionnaire (FFQ; 43 to 168 items) while six studies used a 24-h dietary recall to collect dietary information. In addition, three studies used a diet history questionnaire, 3-day food diary and 4 weeks face-to-face dietary history interview, respectively. One study derived dietary patterns through RRR, another study used a “partial least squares regression” method, while all the other studies derived dietary patterns through a posteriori methods (PCA, PCFA, and FA). Nine studies reported the association of MetS risk with two different dietary patterns, 24 studies considered three different dietary patterns, six studies considered four different dietary patterns and one study considered five different dietary patterns.

18We identified two common dietary patterns with similar factor loading of principle components: “Healthy” and “Meat/Western” patterns. Thirty-eight out of 40 articles included in the systematic review were used for the overall risk estimation. Two studies were excluded because they reported dietary patterns that could not be clearly assumed in “Healthy” nor in “Meat/Western” patterns. In the studies by Agodi et al. and by Wang et al., the “Healthy” dietary pattern was the only pattern identified, whereas in the study by Cattafesta et al. the “Meat/Western” was the only pattern selected. The meta-analyses on the MetS risk in association with “Healthy” and “Meat/Western” dietary patterns (studies comparing the highest intake to the lowest intake) are shown in Figure 2A,B, respectively.

19The overall analysis showed that the MetS risk significantly decreased in association with the adherence to the “Healthy” pattern (OR = 0.85; 95% CI: 0.79–0.91) and significantly increased in association with the adherence to the “Meat/Western” pattern (OR = 1.19; 95% CI: 1.09–1.29). These results did not essentially change when the studies not comparing the highest vs. the lowest dietary pattern adherence values were excluded (Table 2).

20In the “Healthy” pattern meta-analysis, the stratification by study design showed a significant reduced MetS risk in the cross-sectional studies only (Table 2). Stratifying the analysis by geographic area, MetS risk decreased significantly in Eastern countries (OR = 0.78; 95% CI: 0.71–0.86), particularly in Asia (OR = 0.77; 95% CI: 0.70–0.85). The preventive effect of the “Healthy” pattern resulted statistically significant in both sexes (Table 2).

21In the “Meat/Western” pattern meta-analysis, the stratification by study design showed a significantly higher MetS risk in both cohort and cross-sectional studies (Table 2). Similarly, when stratifying the analysis by the geographic area the MetS risk significantly increased in Asia, America and Europe, and in Eastern and Western countries (Table 2). No significant association was found when stratifying by sex (Table 2).

22The high heterogeneity in the pooled analysis of both “Healthy” and “Meat/Western” patterns was slightly reduced in the stratification by geographic area.

23Sensitivity analyses suggested that the estimates were not substantially modified by any single study. Small changes were found in the risk estimates after removal of the outlier studies by Naja et al. (OR = 0.84; 95% CI: 0.78–0.91) and by Nasreddine et al. (OR = 0.85; 95% CI: 0.79–0.92) in the “Healthy” pattern analysis, and by Shokrzadeh et al. (OR = 1.20; 95% CI: 1.09–1.32) and by Gadgil et al. (OR = 1.23; 95% CI: 1.11–1.35) in the “Meat/Western” pattern analysis.

24In the meta-analysis on the “Healthy” pattern, a significant publication bias was detected by the Egger’s test in the overall analysis (p = 0.005) and in cross-sectional studies (p = 0.016), but not by the Begg’s method (Table 2). In the analysis performed excluding the studies by Bell et al., by Arisawa et al. and by Panagiotakos et al., the publication bias, although reduced, remained significant (p = 0.011) (Table 2). In the meta-analysis on “Meat/Western” pattern, a significant publication bias was detected by Egger’s method in the Eastern countries (p = 0.021) and by the Begg’s test in men (p = 0.042) (Table 2).

25The funnel plots of the meta-analyses on the “Healthy” pattern and on the “Meat/Western” pattern are shown in Figure 3A,B, respectively.

26Our systematic review and meta-analysis investigated the effect of dietary patterns extracted via a posteriori methods on MetS risk. According to literature, several different health outcomes are associated with unhealthy and healthy dietary patterns. In particular, the Western/unhealthy pattern increases the risk of cancer in different sites and the risk of low bone mineral density and osteoporotic fracture. Moreover, the prudent/healthy pattern is associated with lower risk of cardiovascular disease and coronary heart disease, diabetes mellitus, and cognitive decline and dementia.

27Considering the 40 included articles, we identified two prevalent dietary patterns: “Healthy” and “Meat/Western”. The “Healthy” pattern was associated with a lower MetS risk and significantly decreased the risk in both sexes and in Eastern countries, particularly in Asia. Adherence to the “Meat/Western” pattern was positively associated with MetS risk and this association persisted in the stratified analysis by geographic area and study design. Similarly, the recent meta-analyses by Shab–Bidar et al. and Rodríguez–Monforte et al. showed that a Western/unhealthy pattern significantly increased MetS risk, whereas a prudent/healthy pattern significantly lowered MetS risk. In our study, MetS risk through unhealthy dietary patterns increased by 19%, while it increased by 22% in the study by Shab–Bidar et al. and by 28% in the study by Rodríguez–Monforte et al.. Healthy dietary patterns significantly decreased MetS risk by 15% in our analysis, by 11% in the meta-analysis by Shab–Bidar et al. and by 17% in the meta-analysis by Rodríguez–Monforte et al.. It should be noted that the meta-analysis of Shab–Bidar et al. was performed on cross-sectional studies only and that Rodríguez–Monforte et al. selected 31 studies including those which identified the dietary patterns via cluster analysis (a priori method).

28According to our findings, the “Meat/Western” pattern significantly increased MetS risk of 20% in Asia, 15% in Europe and 33% in America. In dietary patterns derived a posteriori, the factor loadings indicate the most commonly consumed foods, reflecting the cultural influence on food consumption. It is noteworthy that the usual diet of European populations, especially in Mediterranean countries, tend to include the consumption of healthy foods, such as seafood, vegetables, and fruit, whereas American populations mostly adhere to Westernized dietary patterns, containing high pro-inflammatory foods. As reported in the study by Calton et al., other pre-defined representative dietary patterns exist worldwide, such as the Dietary Approaches to Stop Hypertension (DASH) diet, which is characterized by high intake of fruit, vegetables, whole grains and dairy, and the Northern Europe dietary pattern, which is characterized by high intake of fruit, vegetables, legumes, low-fat dairy, fatty fish, oats, barley and almonds. These patterns can affect MetS risk and should be evaluated when investigating the effect of the dietary patterns on developing MetS, as culture and society influence adherence to healthy or unhealthy dietary pattern. Our study combined dietary patterns derived a posteriori from world countries with very different eating habits, in particular, traditional dietary patterns from Eastern Asian countries (Japan, China, Korea), from Western Asian countries (Iran), from the Mediterranean area (Greece, Lebanon), from Northern Europe (Sweden), from Middle Europe (Germany, Czech Republic and Poland), from North America (USA), from South America (Brazil), and from Australia. Indeed, the traditional dietary pattern in Asian countries is characterized by high intake of rice and/or kimchi, fish and sea food, soybean and soybean products, mushrooms, vegetables, and fruit, in Poland by red meat, fish, potatoes, soup, refined grains and sugars, and high-fat milk, and in Iran by refined grains, nuts, eggs, vegetables and legumes, potatoes, and hydrogenated fats.

29Despite the influence of sex-related factors on MetS, we observed no sex-related difference on the association of dietary pattern with MetS, but, notably, the “Healthy” pattern showed a stronger protective effect in women.

30The “Meat/Western” pattern, characterized by high intake or red and processed meat, eggs, refined grains, and sweets, resulted associated with an increased (+19%) MetS risk. These foods plausibly represent the main cause of the observed effect on MetS risk, particularly meat, since refined carbohydrates, red and processed meats, and fried foods have pro-inflammatory properties and can increase inflammatory cytokines. Indeed, although the meta-analysis by Namazi et al. found no significant association between the most pro-inflammatory diet and MetS, inflammatory factors are involved in insulin resistance and lipid disorders.

31Our results showed the association of the “Healthy” pattern with a lower (−15%) MetS risk. The healthy patterns are characterized by the consumption of foods with high content of vitamins, minerals, antioxidants, fiber, MUFA and n-3 fatty acids, which could contribute to explain the protective effect of the “Healthy” pattern on MetS. Indeed, higher adherence to healthy dietary patterns is associated with a lower risk of glucose intolerance, weight gain, inflammation, insulin resistance and a higher level of HDL cholesterol.

32The main limitation of our study is that the risk of developing MetS could be associated with dietary patterns other than the two (“Healthy” and “Meat/Western”) discussed in this meta-analysis. Differences in the populations in study and in the referral values for MetS diagnosis represent another study limitation and result in heterogeneity. Indeed, the high heterogeneity may be related to the wide variability in dietary data collection and analysis, in the various and not uniformly adjusted confounding factors, and in the identification of the dietary patterns. Heterogeneity is more evident in the analysis on “Meat/Western” pattern, as a possible consequence of the difficulty in characterizing this pattern across the selected studies. Another limitation is that pooled data were directly driven by the included studies, presenting their own weaknesses in study design. Moreover, the cross-sectional nature of many included studies precludes causal inference and the dietary pattern may represent a post hoc event. Only the OR of the highest and the lowest quantile of healthy or unhealthy dietary patterns were included in our analysis, limiting the evaluation of the presence of any trend. Finally, some studies reported risk estimates for quintiles, others for quartiles, and others for tertiles. As dietary intakes are influenced by sex, race/ethnicity, and societal factors, our findings should be considered in the different geographic contexts. Thus, these aspects may have affected the reproducibility of the association between dietary patterns and MetS.

33To further advance this field of research, future studies are needed to examine the association between dietary patterns in geographic context not yet described and MetS, and to evaluate the impact of dietary patterns on the determinants of MetS.

34A protective effect on MetS is attributed to adherence to the “Healthy” pattern, which is characterized by high consumption of fruit, vegetables, whole grains, poultry, fish, nuts, legumes, and low-fat dairy products, whereas the “Meat/Western” pattern is positively associated with MetS. Nutrition is one of the most important modifiable factors affecting health. Public health efforts should aim to adopt healthy dietary patterns and to reduce the burden of MetS, providing guidance for nutritional intervention. For further advance in research, more prospective studies are needed to investigate the association between dietary patterns and MetS in each gender and in different geographic context.