Medical Nutrition Therapy (MNT) Evidence Update: Comparative Effectiveness of Dietary Programs for Reducing Mortality and Cardiovascular Events in Adults with Increased Cardiovascular Disease Risk

Health services question

In adults with established cardiovascular disease (CVD) risk factors, as compared with minimal intervention, what is the most effective dietary program intervention, with or without pharmacological management, physical activity, and behavioral support cointerventions, for reducing risk of early mortality and major cardiovascular events based on the best available systematic review and network meta-analyses of randomized clinical trials (RCTs)?

Bottomline

Based on 40 RCTs evaluating 7 dietary programs, moderate certainty evidence suggests that Mediterranean dietary programs (for example, high in vegetables, fruits, extra virgin olive oil, nuts, legumes, and fish), accompanied by varying cointerventions including pharmacological management (for example, statins), physical activity and behavioral support (for example, nutrition education, smoking cessation, and stress management), were superior to minimal interventions for reducing risk of all cause [1.7% absolute risk reduction (ARR)], cardiovascular mortality (1.3% ARR), stroke (0.7% ARR), and myocardial infarction (1.7% ARR) in patients with established CVD risk factors (for example, obesity, hypertension, dyslipidemia, or a previous cardiovascular event) over a 5-y period. Results from randomized trials with food provisions (for example, extra virgin olive oil, mixed nuts, primarily walnuts) among those living in Mediterranean regions had the largest treatment effects. Similarly, moderate certainty evidence demonstrated that low-fat dietary programs (for example, 20–30% total fat, <10% saturated fat, and high in fish, vegetables, and fruits together with varying cointerventions) were superior to minimal intervention for reducing all-cause mortality (0.9% ARR) and myocardial infarction (0.7% ARR) based on trials conducted in Mediterranean, North American, and Northern European regions. Network metaregression did not detect statistically significant differences in estimates when controlling for the presence of pharmacological management, physical activity, and behavioral support.