Dietary Glycemic Index and Load and the Risk of Type 2 Diabetes: Assessment of Causal Relations

While dietary factors are important modifiable risk factors for type 2 diabetes (T2D), the causal role of carbohydrate quality in nutrition remains controversial. Dietary glycemic index (GI) and glycemic load (GL) have been examined in relation to the risk of T2D in multiple prospective cohort studies. Previous meta-analyses indicate significant relations but consideration of causality has been minimal. Here, the results of our recent meta-analyses of prospective cohort studies of 4 to 26-y follow-up are interpreted in the context of the nine Bradford-Hill criteria for causality, that is: (1) Strength of Association, (2) Consistency, (3) Specificity, (4) Temporality, (5) Biological Gradient, (6) Plausibility, (7) Experimental evidence, (8) Analogy, and (9) Coherence. These criteria necessitated referral to a body of literature wider than prospective cohort studies alone, especially in criteria 6 to 9. In this analysis, all nine of the Hill’s criteria were met for GI and GL indicating that we can be confident of a role for GI and GL as causal factors contributing to incident T2D. In addition, neither dietary fiber nor cereal fiber nor wholegrain were found to be reliable or effective surrogate measures of GI or GL. Finally, our cost–benefit analysis suggests food and nutrition advice favors lower GI or GL and would produce significant potential cost savings in national healthcare budgets. The high confidence in causal associations for incident T2D is sufficient to consider inclusion of GI and GL in food and nutrient-based recommendations.

1. Table S1. Observations in single-sex studies by duration (time, years) of exposure. 2. Normalisation of fasting blood glucose when lowering the dietary glycemic index in recent medium to long term intervention studies Figure 1. Differences in fasting blood glucose in participants consuming diets of higher and lower glycemic index. Meta-regression curves (grey lines) are 95% confidence intervals. Shown also are the first author names associated with the studies extracted. Simply throwing these into a forest plot suggest no significant effects but meta-regression shows a significant relation to baseline fasting glucose, considered together with earlier observations [6], there is evidence of lower GI carbohydrate diets reducing both hypoglycaemia and hyperglycaemia among the otherwise random effects.
Following observations in a meta-analyses of human intervention studies published by us elsewhere in 2008 [6] indicating a transition point on the fasting blood glucose concentration for the direction of effect, which was at approx. 5 to 6 mmol/L, the present meta-regression analysis of subsequent findings from randomised control studies in humans (n = 615) from 0.25 to 1.5 years treatment duration [7][8][9][10][11][12][13] are confirmatory of the existence of a transition ( Figure S1). The slope is the meta-regression relationship based on studies having equal weights with a median effect size SE value of 3.9 mmol/L. The slope was -0.23 mmol/L per mmol/L basal fasting glucose, was significant (P > t = 0.03), and intersected the Y-axis (at y = 0) at a fasting value of 5.4 mmol/L (the transition point) at which the direction effect on fasting glucose changes with baseline fasting blood glucose, and where no effect is expected to be observed.
A weakness of the present meta-regression analysis is the small number of study effects (n = 12), the use of findings from 4 studies at more than one duration of treatment ( [12,13], [7,13]), the use of equal weights to alleviate excessive leverage from two studies and heterogeneities in disease states related to blood glucose control and treatment modalities for body weight maintenance and body weight reduction. A strength of the analysis is the agreement with the prior meta-analysis on 50 similar observations [6] over a shorter duration and which showed also a continued greater fall in fasting blood glucose with increasing severity of T2D when lower GI diets were eaten.
Together these analyses indicate both the effect size and direction of effect can be dependent on the effectiveness of blood glucose control in the study groups, and with a transition at about 5.4 mmol/L fasting blood glucose, which otherwise explains heterogeneity in the treatment effect of lower versus higher glycemic index diets. In the present analysis heterogeneity (I 2 ) was zero after taking account of the status of baseline glucose control. Together with a lack of effect on fasting insulin concentrations by lower versus higher GI diets at below 100 pmol/L [1] these analyses imply a similar implication for the assessment of insulin sensitivity, which is often approximated in models (HOMA IR) by the product of fasting glucose and insulin concentrations and which need robust analytical data [14,15].
A further implication of the meta-analysis ( Figure S1 above and in [6]) is that low glycemic diets help to normalise the fasting blood glucose whether it is below a value close to normal or too high. Too few observations are available to verify this at fasting glucose values below 4.5 mmol/L and above 13 mmol/L). Delay in the short term gluco-regulatory mechanisms such as the glucose-fatty-acid cycle [16] and the Staub-Traugott effect likely explain the blood glucose raising potential of lower GI diets at below the transition point.