The investigation of dietary patterns has emerged as a complementary approach to single food investigations by taking into account human nutrition as a combination of food items. Dietary pattern (DP) approaches can be divided in a priori approach using hypotheses about the benefit or detriment of specific food items on a certain health outcome or a posteriori approach applying exploratory methods on data at hand without previous knowledge. Furthermore, reduced rank regression (RRR), as a mixed approach, uses hypothesis-driven biomarkers or nutrients that are involved in the development of type 2 diabetes and identifies exploratory derived dietary patterns which explain the maximum variance in these biomarkers. However, due to differences in methodological approaches and compositions of dietary patterns, conclusions on healthful patterns with regard to diabetes were rather limited. Attempts to summarize evidence from existing studies were either limited to healthful dietary patterns omitting information from dietary patterns that were not associated with type 2 diabetes or summarized risk estimates derived from different approaches without considering the constraint comparability.

Associations between dietary patterns and type 2 diabetes

Therefore, we summarized evidence of prospective studies which examined associations of dietary patterns with type 2 diabetes by considering different methodological pattern approaches. The search in MEDLINE and Web of Science identified prospective studies (cohorts or trials) associating dietary patterns with diabetes incidence in nondiabetic and apparently healthy participants and resulted in 48 articles from 16 cohorts. The risk to develop type 2 diabetes for selected commonly applied diet quality indices was reduced:

• Mediterranean diet characterized by a high consumption of fruit, vegetables, legumes, nuts, fish and seafood and a moderate consumption of alcohol (RR* for comparing extreme quantiles: 0.87; 95% CI: 0.82, 0.93)
• Dietary Approaches to Stop Hypertension (DASH) based on plants and rich in fruit, vegetables and nuts with low-fat dairy and meats (RR*: 0.81; 95% CI: 0.72, 0.92)
• Alternative Healthy Eating Index (AHEI) including whole grains, vegetables, fruit (RR*: 0.79; 95% CI: 0.69, 0.90)
  Besides, studies reported a variety of heterogeneous indices (n = 15) in terms of components and the way of constructing.  For eight of these, a significant risk reduction (14% – 68%) was observed, but not meta-analyzed due to completely different approaches. For exploratory dietary patterns, we identified 32 patterns derived in 14 studies. Because meta-analyses would not be meaningful for patterns of different composition, we evaluated similarities between individual component food groups and identified two overall groups of patterns:
  • “mainly unhealthy” group of patterns characterized by red and processed meat, refined grains, high-fat dairy, eggs, and fried products was associated with an increased diabetes risk by 44% (RR: 1.44; 95% CI: 1.27, 1.62)
  • “mainly healthy” group of patterns characterized by vegetables, legumes, fruit, poultry, and fish was associated with a reduced diabetes risk by 16% (RR: 0.84; 95% CI: 0.77, 0.91) Due to the population-specificity and heterogeneity of pattern structure, comparability between exploratory dietary patterns still remains limited and requires careful consideration in metaanalyses.

The search identified six RRR-patterns which were related significantly to diabetes risk, with the strongest association observed using thrombosis markers as responses. Meta-analyses of studies on RRR-patterns were only reasonable for originally derived RRR-patterns and their replications in other study populations. Across these identified RRR-patterns, irrespective of the biomarker sets and overall pattern composition, those characterized by high intakes of refined grains, sugar-sweetened soft drinks, and processed meat, were all significantly associated with increased diabetes risk.

Diets rich in F&V could prevent diabetes

Diets according to the Mediterranean Diet, DASH and AHEI have each a strong potential for the prevention of diabetes, irrespective of different particular components. The two identified groups of exploratory DP characterized by concordant food groups were significantly associated with diabetes risk, suggesting that the combination of several food groups allows the identification of diabetes-associated DPs, despite limited evidence for an association of single food groups. Still, exploratory DPs quite frequently remain population-specific observations. For three RRR-patterns consistent positive associations with diabetes were observed across independent populations, which might favor the RRR approach over purely exploratory approaches.

* Relative Risk

Based on: F. Jannasch, J. Kröger, and MB. Schulze. Dietary Patterns and Type 2 Diabetes: A Systematic Literature Review and Meta-Analysis of Prospective Studies.
The Journal of Nutrition 2017; 147(6): 1174-1182.