Mediterranean diet and health outcomes: a systematic meta-review

Abstract

Background

The Mediterranean diet (MeDi) is considered a healthy dietary pattern, and greater adherence to this diet may improve health status. It also may reduce the social and economic costs of diet-related illnesses. This meta-review aims to summarize, synthesize and organize the effects of MeDi pattern on different health outcomes.

Methods

This meta-review was performed following the PRISMA guidelines. A systematic search was conducted in the electronic databases PubMed, Web of Science and Scopus. Two researchers screened all the records to eliminate any duplicate, and they selected the articles to be fully reviewed. A third researcher was consulted to resolve discrepancies and so reach a consensus agreement.

Results

Thirty-three articles were included, nine were systematic reviews and twenty-four were meta-analyses. Most of the diseases analysed are catalogued as non-communicable diseases (NCD), and the impact of these in populations may have major financial consequences for healthcare spending and national income. The results showed that the MeDi may improve health status, and it also may reduce total lifetime costs.

Conclusion

MeDi has been shown to be a healthy dietary pattern that may reduce risk related to NCD. The effect is larger if the pattern is combined with physical activity, and tobacco and excessive alcohol consumption are avoided. Promoting the MeDi as a healthy dietary pattern presents challenges which need the collaboration of all levels of society.

Introduction

Diet and nutrition are important factors in the promotion and maintenance of good health throughout the entire life course,¹ and should be promoted jointly with physical activity,² and the avoidance of other risk factors like tobacco and excessive alcohol consumption.³ The social and economic costs of diet-related illnesses are straining individuals, families and national healthcare budgets.⁴ The Mediterranean diet (MeDi) is considered a healthy dietary pattern, and greater adherence to this diet may improve health status.^5–8 A MeDi is characterized by the daily intake of cereals, vegetables, fruit, low fat dairy products, olive oil, nuts and wine in moderation, weekly intake of fish and legumes and limited meat consumption.⁹

Several reviews and meta-analyses examining the effect of MeDi on different health outcomes have been conducted. However, no review to date has systematically studied all outcomes in an aggregated manner. Therefore, the main aim of this meta-review is to summarize, synthesize and organize the effects of MeDi pattern on different health outcomes, by analyzing and grouping the evidence presented in previous research. Furthermore, the present work aims to provide information and scientific evidence for health professionals, researchers and policy-makers.

Methods

Search strategy

A meta-review is a review of systematic reviews and meta-analyses.¹⁰ This meta-review was performed following the PRISMA guidelines.¹¹ The search was conducted by one researcher (RM). Relevant studies published in English were identified from the appearance of the first report until the end of May 2017 by searching the following electronic databases: PubMed, Web of Science and Scopus. The Open Grey literature database was used to reduce the potential publication bias by searching for grey literature, but none was found. Table 1 shows the search strategy and the terms included. Gephi software¹² was used to develop the diagram of included articles and health outcomes (figure 1).

Table 1

Search strategy

‘Mediterranean diet’

AND

‘systematic review’ OR ‘meta-analysis’

Figure 1

Mediterranean diet: included articles and health outcomes

Study selection

Two researchers (IP, RM) screened all the records to eliminate any duplicates. Both researchers then independently reviewed the titles and abstracts of the 744 records obtained. They then selected the articles to be fully reviewed. A third senior researcher (FE) was consulted to resolve discrepancies and so reach a consensus agreement. Systematic reviews and meta-analyses meeting the following requirements were included: (i) they were published in English; (ii) they were a systematic review or meta-analysis; (iii) they focused on MeDi and health status. Moreover, papers were excluded if, following the AMSTAR checklist,¹³ the quality was not high or moderate. AMSTAR has been shown to be a good measurement tool to assess the methodological quality of systematic reviews of randomized control trials (RCT)¹⁴ and non-randomized studies.¹⁵ Furthermore, papers were excluded when an updated study with the same authorship was performed (see Supplementary Material).

Quality of the reviews

To assess the quality of the systematic reviews, two researchers (IP, RM) independently used a measurement tool for Assessment of Multiple Systematic Reviews (AMSTAR).¹³ The checklist works with 11 questions. Each component can score one point if the answer is ‘yes’ or zero if the answer is ‘no’, ‘can’t answer’ or ‘not applicable’. Scores between zero and three points were considered low quality; four to seven points moderate quality; and eight to eleven points high quality. Any differences in the assessment of a particular study were resolved by consensus. The mean AMSTAR score was 6.7 and the most common reasons for a score of zero were: conflict of interest (the AMSTAR team indicates that to score one, articles must indicate source of funding for the systematic review and for each of the included studies), lack of a priori design, status of publication as an inclusion criterion, and no list with included and excluded studies. The quality of individual studies included in each systematic review or meta-analysis was not assessed. Supplementary table S1 shows the AMSTAR scores for included studies. The AMSTAR scores for excluded studies are presented in the Supplementary table S1.

Data extraction

Two researchers (IP, RM) independently extracted the data from each study. The following information was extracted: (i) reference; (ii) study type; (iii) number of included studies; (iv) objective; (v) population; (vi) evaluated items; (vii) measure of effect in meta-analysis; (viii) results of pooled analysis (95% CI); (ix) heterogeneity (I²); (x) AMSTAR checklist scores.

Results

Study selection

The systematic database search identified 744 publications (figure 2). After eliminating duplicates, 473 remained. Only 33 met the inclusion and exclusion criteria. A list of included articles and references may be found in Supplementary Material. Of the 33 studies included, 9 were systematic reviews and 24 were meta-analyses. Excluded articles and references may be found in Supplementary Material.

Figure 2

Flow diagram

Study characteristics

The main characteristics of the included studies are listed in Supplementary table S1 (see Supplementary Material). One study analysed partial information of another six included studies, but the information Supplementary table S1 was not used for statistical analysis. Two studies were an updated systematic review and meta-analysis. The studies included analysed the relationship between MeDi and 14 issues: adherence, all-cause mortality, asthma, cancer, cognitive functioning, cardiovascular disease (CVD), economic evaluation, fractures, health related quality of life, hypertension, metabolic syndrome, obesity, body weight and body mass index (BMI), rheumatoid arthritis and type 2 diabetes. All diseases and topics with references are shown in figure 1. Some studies showed a combination of diseases or topic, and the line width, regular or bold, represented whether the results of the different studies were from a systematic review or meta-analysis (pooled analysis was conducted in the study), respectively (figure 1). When a study focused on a specific topic, splitting the results into subtopics was avoided. These were then were presented jointly in the corresponding epigraph (e.g. results of cardiovascular risk factors from the study by Nordmann et al.¹⁶).

Adherence to the MeDi

Two articles focused on adherence to the MeDi in children and young people, these being a meta-analysis and a systematic review, respectively. Results of the meta-analysis showed that adherence to the diet varied considerably, and 21% of the study sample presented low adherence to the MeDi versus 10% with high adherence (I² = 99.7% in both analyses). Furthermore, the results of the systematic review revealed that adherence varied across Mediterranean countries, and even within the same country. Differences might be due to social factors, but were mainly found to be a result of the different methodology applied.

All-cause mortality

One meta-analysis showed a significant association between two-point increased adherence to the diet and a reduction of 8% in risk of mortality from all causes, with little evidence of statistical heterogeneity across the studies (I² = 47%). Furthermore, a systematic review and meta-analysis showed that RCT revealed that greater adherence to MeDi had no effects on all-cause mortality.

Asthma

One meta-analysis showed a significant negative association in the overall effect between the highest tertile of adherence (compared with the lowest) of MeDi consumption with lower occurrence of ‘current wheeze’ (I² = 35.4%). However, the study showed a non-significant association in the overall effect of ‘current severe wheeze’. There was a lower prevalence of ‘asthma ever’ associated with the highest tertile of MeDi in the overall effect (I² = 19.4%). The study showed significant differences between Mediterranean and non-Mediterranean centres (‘current severe wheeze’ was only significant in the Mediterranean area), but differences might be explained by different diagnostic criteria in the included studies.

Body weight and BMI

Two meta-analyses and two systematic reviews focused on MeDi and weight loss. In general, the results of the four studies showed there was an inverse association between higher adherence to the MeDi and body weight and BMI. The results of one meta-analysis showed that MeDi may be a useful tool to reduce body weight, especially when the MeDi is energy-restricted, associated with physical activity, and >6 months in length. BMI loss was significantly higher compared with the control diet group (I² = 91.45%). Weight loss was greater compared with the control diet group (I² = 94.93%). The mean weight loss was 1.75 kg (3.88 kg with energy restriction; 4.01 kg increased physical activity). The other meta-analysis showed similar results: the MeDi was more effective in decreasing BMI compared with control diets (mean difference -0.29 kg m^-2; I² = 0%) and body weight (-0.29 kg; I² = 0%), but not in waist circumference. In addition, one systematic review showed that the MeDi resulted in greater weight loss than the low-fat diet after 12 months (range of mean values: -4.1 to -10.1 kg), but produced similar weight loss to other comparator diets. The other systematic review showed that the prevalence of overweight, obesity and central obesity was inversely associated with the MeDi score.

Cancer

Three meta-analyses presented statistical results on the association between MeDi and cancer. All these studies showed risk reduction in different cancers. One study showed that the group in the highest quantile of adherence to a MeDi had a 14% reduction in total cancer mortality (I² = 77%), 4% reduction in total cancer incidence (I² = 0%), 9% reduction in colorectal cancer incidence (I² = 63%) and 4% reduction in breast cancer incidence (I² = 53%). Another study showed a 4% reduction of neoplastic disease incidence. The third meta-analysis showed that the highest MeDi adherence score was significantly associated with a 13% risk reduction of overall cancer mortality (I² = 84%), and among cancer survivors the association between the highest MeDi adherence score and risk of cancer mortality and cancer recurrence was not statistically significant. In addition, the study showed a risk reduction of colorectal cancer (17%; I² = 56%), breast cancer (7%; I² = 15%), gastric cancer (27%; I² = 66%), prostate cancer (4%; I² = 0%), liver cancer (42%; I² = 0%) and head and neck cancer (60%; I² = 90%) in people with the highest adherence scores. No significant association was detected for oesophageal, ovarian, endometrial and bladder cancer.

Cognitive functioning

Four systematic reviews and four meta-analyses studied the relation between MeDi and cognitive functions. Six studies found that high MeDi adherence was associated with the risk reduction of different cognitive diseases. One study, despite methodological heterogeneity and limited statistical power in some of the included studies, showed a pattern of associations: higher MeDi adherence was associated with better cognitive function, lower rates of cognitive decline, and reduced risk of Alzheimer disease in 75% of the studies. Results for mild cognitive impairment, however, were inconsistent. A further study showed that higher MeDi adherence was associated with significantly decreased risks for all-cause dementia, Alzheimer’s disease, Alzheimer’s disease mortality and infarcts detected by magnetic resonance imaging. Another study found that the MeDi was associated with improved cognitive function, a decreased risk of cognitive impairment or decreased risk of dementia or Alzheimer’s disease. One meta-analysis showed that high MeDi adherence was consistently associated with reduced risk of stroke (29%; I² = 69.1%), depression (32%; I² = 53.4%) and cognitive impairment (40%; I² = 76.4%). Moderate adherence was also associated with reduced risk of depression and cognitive impairment, but marginal risk of stroke. Another meta-analysis showed protective function in the highest Mediterranean score for the developing of cognitive disorders (21%; I² = 22%). Subgroup analysis showed a reduction of the incident risk for mild cognitive impairment (17%; I² = 0%) and Alzheimer’s disease (40%; I² = 0%), but the association was not detected in the incident risk of dementia. The median adherence score was not associated with cognitive disorders. Another meta-analysis showed that the highest MeDi adherence versus the lowest adherence was associated with reduced risk of cognitive impairment (33%; I² = 0%). Furthermore, the study showed that higher adherence to the diet was associated with a reduced risk of mild cognitive impairment (27%; I² = 0%), and Alzheimer’s disease (36%; I² = 0%) among cognitively normal subjects. One systematic review showed reductions in the risk for mild cognitive impairment (46%) and dementia (64%) compared with the control group in one RCT, but also showed no association in other included studies. However, one study showed conflicting results and conclusions regarding the efficacy of the diet as a therapeutic approach for age-related cognitive function.

CVD

Two systematic reviews and seven meta-analyses studied the relation between MeDi and CVD. All studies showed a positive effect of MeDi on CVD. One systematic review revealed strong evidence on the association between adherence to a MeDi pattern and the incidence of CVD. The other systematic review showed that persons with diabetes, in summary, may improve cardiovascular risk factors, while another study showed risk reduction for a new myocardial infarction (68%) and cardiovascular death (68%) in a RCT, but also showed no association in other included studies. One meta-analysis showed that MeDi might protect against vascular disease, but the available evidence is limited. The combined effect for all studies showed protection against major vascular events (37%), coronary events (35%), stroke (35%) and heart failure (70%), but not for all-cause mortality or cardiovascular mortality. The heterogeneity was not assessed due to the small number of included studies. Another meta-analysis revealed that a Mediterranean dietary pattern is associated with lower risks of CVD incidence and mortality. Subjects in the highest quantile of adherence to the diet had lower CVD incidence (27%; I² = 36%) and mortality (24%; I² = 75%). The study also showed a significant reduction of risk for coronary heart disease (28%; I² = 0%), myocardial infarction (33%; I² = 0%), and stroke (24%; I² = 52%). It revealed that the protective effects of the diet appeared to be most attributable to olive oil, fruits, vegetables and legumes. A further meta-analysis found that MeDi appear to be more effective than low-fat diets in inducing clinically relevant long-term changes in cardiovascular risk factors and inflammatory markers. Body weight, BMI, waist circumference, blood pressure (systolic and diastolic), total cholesterol, high-sensitive C-reactive protein and plasma glucose decreased more in subjects randomized to MeDi than in subjects randomized to low-fat diets. Another meta-analysis provided evidence that a MeDi decreases inflammation and improves endothelial function. MeDi regimes resulted in a significantly more pronounced increase in flow mediated dilatation (1.86%; I² = 43%), and adiponectin (1.69 µg ml^-1; I² = 78%). Additionally, there was a reduction in high-sensitive C reactive protein (-0.98 mg l^-1; I² = 91%), interleukin-6 (-0.42 pg ml^-1; I² = 81%) and intracellular adhesion molecule-1 (-23.73 ng ml^-1; I² = 34%). A final meta-analysis showed that a 2-point increase in adherence score to the MeDi was reported to determine a 10% reduced risk of CVD (I² = 38%).

Economic evaluation

One systematic review studied the cost-effectiveness of adherence to the MeDi as a prevention strategy against mortality and morbidity, comparing it to a non-MeDi and investigating the impact of adopting a MeDi based on dietary cost. In summary, the study showed that the MeDi had the potential to modify undesirable disease outcomes and the costs of managing them.

Fractures

Two meta-analyses studied the association between adherence to a MeDi and the risk of fractures. One also studied bone mineral density. The results were similar for hip fractures (18–21% of risk reduction), but one of the meta-analyses showed no association with the risk of any or total fractures (pooled analysis of only two studies). With reference to bone mineral density, the highest MeDi scores were positively associated with measures for lumbar spine, femoral neck and total hip.

Health-related quality of life

One meta-analysis reviewed the results from one article reporting health-related quality of life in two cohort studies. Neither study found an association between MeDi indices and mental scores on the Short Form-36 Health Survey. Furthermore, results of the physical Short Form-36 Health Survey score were mixed.

Hypertension

One meta-analysis studied the effect of interventions of at least 1-year duration on blood pressure, comparing MeDi with a low-fat diet. The results showed reduced systolic (-1.44 mm Hg; I² = 87%) and diastolic (-0.70 mm Hg; I² = 63%) blood pressure.

Metabolic syndrome

Four meta-analyses and one systematic review studied MeDi, and metabolic syndrome as well as its components. In summary, all studies reported adopting MeDi pattern resulted in lower risk of metabolic syndrome. One meta-analysis studied metabolic risk factors, with the results showing that MeDi was significantly beneficial on five of the six metabolic factors (waist circumference, triglycerides, blood glucose, systolic and diastolic blood pressure) when the intervention was longer in duration, was conducted in Europe, used a behavioural technique, and was conducted using small groups. Another meta-analysis showed that MeDi was associated with lower risk of metabolic syndrome (19%; I² = 74%). Furthermore, the study revealed inverse associations for waist circumference, blood pressure and low HDL cholesterol levels. The systematic review showed that most of the reviewed studies provided strong evidence on the inverse association between high MeDi adherence and the prevalence of metabolic syndrome. Other meta-analysis showed that higher MeDi adherence was associated with reduced risk of metabolic syndrome (50%; I² = NA). The study also showed significant effect in waist circumference, HDL cholesterol, triglycerides, glucose and HOMA-IR, but not in blood pressure. Finally, another meta-analysis showed that participants with metabolic syndrome allocated to a MeDi, as compared with those following a control diet, had a 49% (I² = 71%) increased probability of remission from metabolic syndrome during a 2–5-year follow-up.

Rheumatoid arthritis

One meta-analysis reviewed the results from a cohort study. Results from the cohort study showed similar rates of rheumatoid arthritis comparing subjects with the highest and lowest MeDi scores.

Type 2 diabetes

Six meta-analyses and two systematic reviews studied the effect of MeDi on type 2 diabetes. In summary, a MeDi pattern may help to prevent and manage type 2 diabetes. One meta-analysis studied the effect of different dietary strategies on type 2 diabetes. The MeDi reduced HbA1c level significantly compared with usual care (-0.31), but not compared with the Palaeolithic diet. Furthermore, none of the interventions was significantly better than the others in lowering glucose parameters. Another meta-analysis showed that, compared with control diets in type 2 diabetes patients, MeDi led to greater reductions in HbA1c (-0.30; I² = 67.2%), fasting plasma glucose levels (-0.72 mmol l^-1; I² = 66.1%), fasting insulin levels (-0.55 µU ml^-1; I² = 0%), total cholesterol (-0.14 mmol l^-1; I² = 0%), triglycerides (-0.29 mmol l^-1; I² = 58%), while high-density lipoprotein was increased (0.06 mmol l^-1; I² = 53.6%). Furthermore, systolic and diastolic blood pressure were reduced (1.45 mm Hg, 1.41 mm Hg; I² = 0%). Another meta-analysis showed that higher MeDi adherence was associated with reduced risk of developing type 2 diabetes (23%; I² = 58%). The main limitations were due to variations in the tools used for assessing MeDi adherence, confounder adjustment, duration of follow-up and number of events with diabetes. A further meta-analysis showed that for highest MeDi adherence (versus lowest), the pooled effect was a 19% reduction in risk of diabetes (I² = 55%). Another meta-analysis showed a reduction of HbA1c (mean difference = -0.47; I² = 3.5%). One systematic review showed that the evidence from the included studies suggested that MeDi might help prevent type 2 diabetes, and the other systematic review showed that three of the four included studies found that a higher MeDi score was predictive of a lower incidence of type 2 diabetes. Finally, one meta-analysis showed that, despite limitations, a MeDi with no restriction on fat intake may reduce the incidence of type 2 diabetes.

Discussion

Summary of evidence

This meta-review aimed to combine the available evidence examining the effect of MeDi on health outcomes in 9 systematic reviews and 24 meta-analyses incorporating results from 636 studies. Our results are useful for public health policy and related decision-making processes, because, in an aggregated and structured manner, and by means of a meta-review of high and moderate quality systematic reviews and meta-analyses, they provide all the available content associated with MeDi, health outcomes and topics of interest. The systematic reviews and meta-analyses included focus on effects of MeDi on fourteen topics: adherence, all mortality causes, asthma, cancer, cognitive functioning, CVD, economic evaluation, fractures, health related quality of life, hypertension, metabolic syndrome, obesity, body weight and BMI, rheumatoid arthritis and type 2 diabetes.

Most of the diseases analysed are catalogued as non-communicable diseases (NCD),¹⁷ and the impact of these in populations may have major financial consequences for healthcare spending and national income.¹⁸ More than 36 million people die annually from NCDs (63% of global deaths), including >14 million people who die too young between the ages of 30 and 70.^19,20 However, most of these pre-mature deaths from NCDs could be prevented by enabling health systems to respond more effectively and equitably to the health-care needs of people with NCDs, and by influencing public policies in sectors outside health that tackle shared risk factors (i.e. tobacco use, unhealthy diet, harmful alcohol consumption or physical inactivity).²⁰

Our review shows that, according to different studies,^4,21–24 a MeDi pattern may improve health status. Furthermore, the implementation of a MeDi pattern may reduce total lifetime costs.²⁵ Nonetheless, food habits may be influenced by different factors,²⁶ and our study also includes an article showing that adherence to MeDi varied across countries, and the percentage of population (children and young people) with low MeDi adherence is higher than the percentage with high adherence.²⁷ Worldwide trends in MeDi adherence show that the majority of Mediterranean countries has suffered a significant decrease in adherence, with this decrease being higher in the Mediterranean European group.²⁸ The traditional MeDi is now progressively eroding due to the widespread dissemination of the Western-type economy, urban and technology-driven culture, as well as the globalization of food production and consumption, related to the homogenization of food behaviours in the modern era.⁹ This finding is also supported by another study,⁴ revealing that the abandonment of traditional habits and the emergence of new lifestyles associated with socio-economic changes pose important threats to the preservation and transmission of the MeDi to future generations. A recent study also shows that psychosocial factors may influence levels of MeDi adherence, and need to be taken into account when developing prevention and health promotion initiatives.²⁹ Moreover, the MeDi model has been proven to have a lower environmental impact, mainly due to the consumption of more plant-derived products and fewer animal products, compared with other current dietary patterns.³⁰ In other words, the MeDi is not only a healthy diet. The MeDi is a sustainable dietary pattern that can achieve long-term sustainable benefits as improving population’s health and nutrition, getting a lower environmental impact, the enhancement of richness in biodiversity, as well as generating positive local economic returns.^30,31

Hence, improving adherence to a MeDi pattern is a highly important goal,³² across all age and social groups, especially in individuals of a lower educational level,³³ not only due to the reported health outcomes in the majority of included studies, but also to preserve the cultural heritage.³⁴ Adopting a sustainable dietary pattern can contribute to food security and healthy life for present and future generations.³⁵

Limitations

The results of this meta-review should be interpreted with caution. There are several limitations that must be highlighted. The quality across studies varies. The heterogeneity of statistical analysis in the meta-analyses was high in many studies, but was explained in each study by using different methods: subgroup analysis, meta-regression, etc. We used the validated AMSTAR checklist to assess the quality of included studies. The quality of studies was evaluated by two researchers resolving conflicts by consensus (see the authors’ contributions). Furthermore, by summarizing reviews and meta-analyses that have already synthesized other articles, information may be lost, and hasty conclusions may be drawn from results, especially in topics for which only one article was included (i.e. economic evaluation, rheumatoid arthritis, etc.). We attempted to avoid duplicate results from the same article included in different studies, but when a meta-analysis was performed, we could not extract the duplicate article from the analysis. However, we considered this when formulating results and conclusions. Moreover, a meta-review can only include reviews and meta-analyses. This means that some primary research may be excluded, but we used original articles in the discussion to verify information. Finally, we attempted to minimize any bias in the review process by involving three researchers in the study selection and data extraction (see Author’s Contributions).

Conclusions

MeDi has been shown to be a healthy dietary pattern that may reduce risk related to NCD. It also provides health, economic and environmental benefits. Furthermore, the effect is larger if the pattern is combined with physical activity, and tobacco and excessive alcohol consumption are avoided. Worldwide variations in food habits have influenced adherence to this dietary pattern. It should also be noted that, despite the MeDi being a recommended dietary pattern, it is easier to follow it in Mediterranean countries than in other geographical areas. Promoting the MeDi as a healthy dietary pattern presents challenges which need the collaboration of all levels of society. These results may be used in clinical and healthy eating recommendations.

Funding

The authors received no funding for this work.

Conflict of interest: None declared.

Key points

• Twenty-four meta-analyses and nine systematic reviews and 636 studies were included; this research included results from 636 studies.

• The included studies analysed the relationship between Mediterranean diet and fourteen issues: adherence, all-cause mortality, asthma, cancer, cognitive functioning, cardiovascular disease (CVD), economic evaluation, fractures, health related quality of life, hypertension, metabolic syndrome, obesity, body weight and body mass index (BMI), rheumatoid arthritis and type 2 diabetes; a diagram was presented relating health outcomes and included studies.

• Most of the diseases analysed are catalogued as non-communicable diseases (NCD), and the impact of these in populations may have major financial consequences for healthcare spending and national income.

• Mediterranean diet pattern may improve health status and it may reduce total lifetime cost. Improving adherence to the Mediterranean Diet pattern is a highly important goal, across all age and social groups, especially in individuals of a lower educational level; it may help to preserve the diet cultural heritage.