Increase Font Size Decrease Font Size View as PDF Print

What is the relationship between adherence to dietary guidelines/recommendations or specific dietary patterns, assessed using reduced rank regression analysis, and risk of type 2 diabetes?

Conclusion

There is insufficient evidence, due to a small number of studies, to examine the relationship between dietary patterns derived using reduced rank regression and risk of type 2 diabetes. The differences in the methods used and populations studied made it difficult to compare results, and therefore no conclusions were drawn.
 

Grade

IV – Not Assignable

 

Key Findings

The three positive quality prospective cohort studies included in this review used reduced rank regression (see appendix A) analysis to examine the relationship between dietary patterns and the risk of type 2 diabetes (T2D). Comparison across studies was limited by the small number of studies, differences in methodology, and in the populations studied. Therefore, no conclusions were drawn.

Evidence Summary Overview

Description of the Evidence
Three prospective cohort studies that used RRR to examine the relationship between dietary patterns and T2D were included in this systematic review (Liese, 2009; McNaughton, 2008; Imamura, 2009). All of the studies were rated positive quality. Two of the studies were conducted in the USA and one in the United Kingdom. The sample sizes were 880 for Liese (2009), 2,879 for Imamura (2009), and 6,699 for McNaughton (2008). All three studies were conducted in adults, included females and males, used validated food frequency questionnaires to assess dietary intake, and included incidence of T2D as the primary study outcome.
 
The independent variables in these studies were dietary pattern scores and biomarkers used as response variables in two of the studies. The response variables used and the respective dietary patterns extracted for each study are described in more detail below.
 

Evidence Summary Paragraphs

Liese 2009 (positive quality) used plasminogen activator Inhibitor-I (PAI-1) and fibrinogen as response variables. One dietary pattern was extracted that was characterized by high intake of red meats, low-fiber bread and cereal, dried beans, fried potatoes, tomato vegetables, eggs, cheese, and cottage cheese and low intake of wine. Red meat and low-fiber bread/cereal explained 19.3 percent and 18.1 percent, respectively, of the variation in the pattern score. Taken together, all nine food groups within the pattern explained 72.8 percent of food pattern score variation.
 
McNaughton 2008 (positive quality) used Homeostasis Model assessment of Insulin resistance index (HOMA-IR index) as the response variable. One dietary pattern was extracted that was characterized by high consumption of low-calorie/diet soft drinks, onions, sugar-sweetened beverages, burgers and sausages, crisps and other snacks, and white bread; and low consumption of medium-/high-fiber breakfast cereals, jam, French dressing/vinaigrette, and whole meal bread. The extracted dietary pattern explained 5.7 percent of the variation in HOMA-IR, and the 10 food items with factor loadings >0.2 explained 66.5 percent of the variation in the dietary pattern score.
 
Imamura 2009 (positive quality) conducted confirmatory and exploratory analyses to compare internally and externally derived dietary patterns on the incidence of T2D using data from the Nurses’ Health Study (NHS), European Prospective Investigation into Cancer and Nutrition Potsdam Study (EPIC), Whitehall II Study (WS). Response variables were: NHS: Inflammatory cytokines; EPIC: HDL, glycated hemoglobin, c-reactive protein, adiponectin; WS: HOMA-IR. For the exploratory analyses, three RRR analyses were done within the FOS cohort using BMI, fasting glucose, triglycerides (TG), HDL cholesterol, and hypertension as response variables and each food grouping of the NHS, EPIC, and WS cohorts was applied. All three exploratory scores had similar positive contributions (increased risk; meat, processed meat, eggs, margarine, fried products, refined grains, and caloric/noncaloric soft drinks), but the negative contributors (decreased risk) differed (with the exceptions of tea and whole grains).

Click for Table 4-C-III-1 Summary of Findings

 

Assessment of the Body of Evidence

This review included three positive-quality prospective cohort studies. There was a positive association between dietary patterns that included meat intake and incident T2D in the two studies (Liese, 2009; McNaughton, 2008) that used biomarkers as response variables, though the definitions of meat differed. However, because there were so few studies available, variability in the methodology used and different populations considered, there was insufficient information from which to assess consistency or draw conclusions about the relationship between dietary patterns derived using RRR and risk of T2D.
 

Limitations of the Evidence

Methodological Differences:
  • All of the studies used different types of biomarkers as response variables, such as PAI-1 and fibrinogen; HOMA-IR index; and BMI, fasting glucose, TG, HDL, and hypertension, making it difficult to make comparisons across these studies.
  • The dietary patterns described in each of these studies were directly linked to the response variables selected; therefore, the variation in the response variables used suggest that the resulting dietary patterns may not be comparable.
  • There were variations in dietary assessment methods used to assess dietary intake, as well as the food groupings used in the analyses across the studies. For example, Liese (2009) used a 114-item validated semi-quantitative FFQ, created 33 food groups on the basis of similarities in food and nutrient composition, and queried alcoholic beverages separately. McNaughton (2008) used a 127-item validated FFQ and the food and beverage items were aggregated into 71 groups on the basis of nutrient content, cooking, and preparation methods. Imamura 2009 used a 126-item validated semi-quantitative FFQ (FOS) and used food groupings from previous studies to each RRR-derived dietary pattern and applied to the FOS data to create three different sets of food groups used in their analyses. These methodological differences make it difficult to compare the resulting dietary patterns across studies and to determine how these differences may have contributed to differences in relationships between the patterns and type 2 diabetes risks.
  • The studies were not consistent in their use of confounders in the analyses. For example, as compared to McNaughton (2008), alcohol intake was not included as a confounder in the analyses by Liese (2009), and alcohol, BMI, and smoking status were not included as confounders by Imamura (2009).
 Population Differences:
Two of the studies were conducted in the United States and one in the United Kingdom and represented populations in different regions of the world, which limited the ability to compare and interpret the results due to potential differences in dietary patterns between these regions.
 

Research Recommendations

More research using reduced rank regression analyses should be conducted to investigate the relationship between dietary patterns and type 2 diabetes, particularly among U.S.-based populations, and including both intermediate outcomes (glucose intolerance, insulin resistance), as well as incidence of disease. Additionally, standardization in methodology, such as response variables and food groupings used, are also needed.
 

References

  1. Liese AD, Weis KE, Schulz M, Tooze JA.  Food intake patterns associated with incident type 2 diabetes: the Insulin Resistance Atherosclerosis Study. Diabetes Care. 2009 Feb; 32(2):263-8. doi: 10.2337/dc08-1325. Epub 2008 Nov 25. PMID: 19033409
  2. McNaughton SA, Mishra GD, Brunner EJ. Dietary patterns, insulin resistance, and incidence of type 2 diabetes in the Whitehall II Study. Diabetes Care. 2008 Jul;31(7):1343-8. doi: 10.2337/dc07-1946. Epub 2008 Apr 4.  PMID: 18390803
  3. Imamura F, Lichtenstein AH, Dallal GE, Meigs JB, Jacques PF. Generalizability of dietary patterns associated with incidence of type 2 diabetes mellitus. Am J. Clin. Nutr. 2009 Oct; 90(4):1075-83. Epub 2009 Aug 26. PubMed PMID: 19710193; PubMed Central PMCID: PMC2744626.

 


Research Design and Implementation
For a summary of the Research Design and Implementation results, click here.