INDEPENDENT THINKING. WHOLE-FOOD LIVING.NO ANIMALS / NO OIL
THENANODIET
← Metabolism / Health Desk

Whole grains and diabetes risk: the refined-grain comparison matters

A dose-response meta-analysis found an association for whole grains, with substantial variation between cohorts.

Whole-food ingredients prepared without added oil

Illustrative food image, not a photograph of the study intervention.

A synthesis of future diagnoses

Aune and colleagues examined grain intake and subsequent type 2 diabetes in a systematic review and dose-response meta-analysis. Their search covered PubMed studies through June 5, 2013, and the analyses included 16 cohort studies. They used random-effects models to pool relative risks. The question was about developing diabetes over follow-up, not glucose changes after one meal or treatment of established diabetes. Cohort design also matters: the researchers combined observations of dietary habits rather than trials assigning people to grain menus. A pooled association can be informative without establishing that a grain substitution alone caused the difference in diabetes incidence.

Aune D, Norat T, Romundstad P, Vatten LJ. European journal of epidemiology, 2013. PMID 24158434 · DOI

Whole grains and refined grains differed

For whole grains, the pooled relative risk per three daily servings was 0.68, with a 95% confidence interval of 0.58–0.81, based on ten studies. For refined grains, the corresponding estimate was 0.95, with a confidence interval of 0.88–1.04, based on six studies. The refined-grain interval includes 1.0, whereas the whole-grain interval does not. These are relative comparisons, not percentage-point reductions in an individual's chance of diagnosis. Nor is three servings a tested treatment dose: it is the increment used to summarise dietary exposure in this observational dose-response analysis of future type 2 diabetes risk.

Aune D, Norat T, Romundstad P, Vatten LJ. European journal of epidemiology, 2013. PMID 24158434 · DOI

The average conceals uneven results

Variation between study results was substantial: the reported I-squared was 82% for whole grains and 53% for refined grains. The whole-grain association was nonlinear, with P for nonlinearity below 0.0001; the corresponding refined-grain value was 0.10. A nonlinear relationship means the single per-three-serving estimate should not be stretched into a claim that every additional serving brings an identical proportional benefit. The pooled number is a summary across studies and intakes, not a promise about an individual breakfast. These features are central to interpreting the synthesis, rather than minor qualifications to place after an otherwise universal dietary claim.

Aune D, Norat T, Romundstad P, Vatten LJ. European journal of epidemiology, 2013. PMID 24158434 · DOI

Do not flatten the grain categories

Inverse associations were reported for whole-grain bread, whole-grain cereals, wheat bran and brown rice, while white rice was associated with higher risk. The authors explicitly noted that several subtype results rested on few studies and warranted further investigation. It would therefore be misleading to transfer the broad refined-grain estimate to every particular refined grain, or to rank all whole grains confidently from these data. The authors supported replacing refined grains with whole grains. The careful takeaway is about the direction of the observed pattern, while retaining the observational design, sparse subtype evidence and considerable between-study variation in the same picture.

Aune D, Norat T, Romundstad P, Vatten LJ. European journal of epidemiology, 2013. PMID 24158434 · DOI