DEPARTMENTS: Cochrane Nursing Care Corner
Is the provision of dietary advice to follow a “Mediterranean” diet, or the supply of foods appropriate to such a diet, effective in the primary prevention of cardiovascular disease (CVD) when compared with no or minimal intervention?
Relevance to Nursing
Cardiovascular disease is a major cause of mortality and morbidity worldwide, but there are geographical variations in the pattern of illness. Observational studies suggest that one factor in these variations could be differences in dietary patterns. For instance, there are higher levels of heart disease in Northern and Western areas of Europe, where diets high in saturated fat are prevalent, contrasting with Southern and Eastern Europe, where the typical Mediterranean diet consists of greater consumption of fruit, vegetables, and legumes, with the use of olive oil as the major source of fat in the diet and low to moderate intake of red wine and dairy products and, additionally, a high ratio of fish to meat consumption, with meals freshly prepared from local, seasonal produce.
Randomized controlled trials of interventions promoting Mediterranean dietary patterns corroborate the findings of these observational studies in the secondary prevention of CVD, but less evidence exists of its effectiveness in primary prevention. This review investigates the role of Mediterranean dietary patterns in the primary prevention of CVD. It is relevant to nurses involved in formal or informal primary prevention because it is important that any advice given is credible and evidence based if it is to effect appropriate behavior change.
This is a summary1 based on a Cochrane review with narrative synthesis and, where appropriate, meta-analyses of 11 randomized controlled trials involving 52,044 participants; a single trial accounted for 48,835 of the participants.
All participants were adults (18 years or older) and recruited from the general population or were identified as high risk by study authors; studies in which more than 25% of the participants had CVD or diabetes at baseline were excluded.
All the studies that met the inclusion criteria were of interventions that provided dietary advice to individuals; most interventions were delivered by dietitians or nutritionists, although details of who delivered the intervention were not given for every trial. The interventions were diverse in intensity, duration, and method of delivery. The minimum follow-up for inclusion in this review was 3 months, and for the purposes of this review, a Mediterranean diet was defined as 2 or more components from the list:
- High monounsaturated-saturated fat ratio
- Low to moderate red wine consumption
- High consumption of legumes
- High consumption of fruits and vegetables
- Low meat consumption with increased consumption of fish
- Moderate consumption of dairy products
The comparator was no or minimal intervention, specified as no person-to-person intervention or reinforcement but allowing for the provision of leaflets.
Primary outcomes were cardiac/all-cause mortality and nonfatal clinical events such as myocardial infarction; secondary outcomes were blood lipid levels, type 2 diabetes, health-related quality of life, adverse effects, and costs.
Most of the included trials were small and at high or unclear risk for bias. Four of 11 trials were judged to be at low risk for bias for random sequence generation, with 3 trials at low risk because of the methods used for allocation concealment. There was a lack of clarity in relation to blinding of participants and personnel in all 11 trials; however, this is acknowledged as a product of the difficulties of blinding individuals to behavior change interventions. Four trials were found to be at low risk for bias related to assessment of outcomes because these were performed blinded. Four trials were judged to be at low risk for bias relating to completeness or otherwise of outcome data. Seven studies were at low risk for bias because of selective reporting.
The results were as follows:
- Primary outcomes of all-cause/cardiovascular mortality and nonfatal clinical events, such as myocardial infarction, were reported in only 1 trial; no statistically significant effects were identified.
- Small but statistically significant intervention effects were found on total cholesterol (8 trials, 4151 participants) of −0.16 mmol/L (95% confidence interval, −0.26 to −0.06; P = 0.003) and low-density lipoprotein (6 trials, 3227 participants) of −0.7 mmol/L (95% confidence interval, −0.13 to −0.01; P = 0.016).
- Subgroup analysis of interventions explicitly described by the study authors as Mediterranean or including core components of a Mediterranean diet as defined by the review authors (increased fruit and vegetable intake and replacement of saturated fats with monounsaturated fats) had a statistically significantly greater effect size than those that met the review inclusion criteria but were not described as Mediterranean or did not include both core components.
- Of 7 trials reporting outcome data for high-density lipoprotein cholesterol, only 1 trial reported a statistically significant intervention effect.
- Seven of 9 trials reported no statistically significant effect of the intervention on triglyceride levels; of the remaining 2 trials, 1 trial reported a statistically significant decrease in levels, whereas the other identified a statistically significant converse effect.
- Three of 5 trials reporting blood pressure identified a statistically significant intervention effect on, both systolic and diastolic, blood pressure.
- One trial reported incidence of diabetes; no statistically significant effect was found.
- None of the trials reported on costs, health-related quality of life, or adverse events.
Best Practice Recommendations
There is some limited evidence from this systematic review to support the use of interventions promoting a Mediterranean-style diet in the primary prevention of CVD, specifically in relation to reducing blood pressure and cholesterol levels; however, the included studies were of low quality, and most of the participants come from 1 large trial of postmenopausal women, potentially affecting the generalizability of the results.