Dietary Patterns and Non-Communicable Disease Biomarkers: A Network Meta-Analysis and Nutritional Geometry Approach

Abstract

Quantitative rankings of multiple dietary patterns for their effects on non-communicable disease (NCD) biomarkers is lacking and would inform primary prevention strategies. Accordingly, a network meta-analysis (NMA) was conducted to compare and rank the effects of different dietary patterns on NCD biomarkers, and associations of dietary patterns’ underlying macronutrient composition with NCD biomarkers were determined by a nutritional geometry approach. Randomised controlled trials (RCTs) were eligible for inclusion if they enrolled healthy participants, employed food-based dietary pattern interventions without energy restriction, and reported NCD biomarker outcomes. NCD biomarkers were included as an outcome if ≥10 trials were available. A systematic search of five electronic databases identified 4008 records. Sixty-eight articles from 59 RCTs reporting lipids, glycemic, and inflammatory biomarkers were included for quantitative syntheses. Risk-of-bias was predominantly categorized as low or having some concerns, and confidence-of-evidence low. Relative to western habitual diet, the Mediterranean, Dietary Approaches to Stop Hypertension (DASH), dietary guidelines-based, plant-based, and low-fat diets reduced low-density lipoprotein cholesterol (mean difference range: −0.29 to −0.17 mmol/L), total cholesterol (−0.36 to −0.24 mmol/L), and apolipoprotein B (−0.11 to −0.07 g/L) (all p < 0.05); the Paleo, plant-based and dietary guidelines-based diets reduced homeostasis model assessment of insulin resistance (−0.95 to −0.35, all p < 0.05). No dietary pattern ranked consistently highest. The Paleo diet received the highest all-outcomes-combined average Surface Under the Cumulative Ranking Curve value (67%), followed by DASH (62%) and Mediterranean diets (57%), whereas western habitual diet was lowest (36%). Our findings were independent of macronutrient composition, highlighting the significance of dietary pattern-level analysis.