Laugesen and Elliott’s epidemiological analysis (http://www.nzma.org.nz/journal/116-1168/295/) suggests a correlation between regional A1 β-casein cow milk consumption and ischaemic heart disease or Type 1 diabetes. Correlations were not significant for the A2 variant of β-casein. The result of their analysis with respect to Type 1 diabetes is similar to a previous report of a high correlation between milk protein consumption and diabetes incidence across various countries.1 Unfortunately, food disappearance data are notoriously unreliable, representing disappearance of foods on a country basis, and therefore only provide a rough approximation of food consumption by individuals. The data do not necessarily reflect either the type or amount of food consumed by individuals at risk for various chronic diseases. The only way to begin to associate any environmental exposure such as diet with development of chronic disease is to follow a cohort of individuals prospectively and monitor food intake (plus other exposures) and biomarkers of health, disease and/or disease outcome. Therefore, because randomised trials of A1 versus A2 β-casein-based diets in children have not been conducted, an assessment of a possible causal relationship between diet and disease must presently rely on feeding studies in animals.

We were involved in several such studies including a trial of A1 versus A2 β-casein initiated by Professor Robert Elliott and in part supported by Dr Jeremy Hill of the New Zealand Dairy Research Institute (now Fonterra Research Centre). Our experience from these studies can be summarised thus: cow milk proteins or whole milk preparations only modestly promote the development of immune-mediated insulin dependent Type 1-like diabetes in mice or rats genetically predisposed to develop such disease (NOD mice, BB rats). Partial protection from diabetes development was seen if hypoallergenic diets with amino acids from hydrolysed proteins were fed.2 For this reason, intervention studies in infants at risk are currently being conducted, comparing conventional and hypoallergenic infant formula (http://www.trigr.org/).

Diets entirely devoid of milk but containing other protein were also found to promote diabetes development in animals, and the most diabetes-promoting diet contained no milk products and was mainly wheat based (37%).2,3 The comparison of A1 versus A2 β-casein-based diets in NOD mice did not demonstrate different risks for the two diets. The latter study is the only published trial with blinding of the investigators with regard to the type of diets fed to the different groups.2 This study did not confirm earlier findings of a higher diabetes risk for A1 versus A2 β-casein diets in NOD mice.4 The blinded animal trial also included BB rats. Here, one of two comparisons showed a significantly higher diabetes incidence (p <0.05) for the A1 β-casein-containing diet. The other comparison showed a trend towards the opposite effect.

Hence, the experience from two well accepted animal models of Type 1 diabetes fails to demonstrate a consistently higher diabetes-promoting potential of A1 as compared to A2 β-casein milk. As judged from feeding studies in animals, avoidance of A1 β-casein milk consumption may not decrease the risk of Type 1 diabetes.5 In fact, A2 β-casein milk led to a very high diabetes rate in one experiment mentioned above. Furthermore, several studies reported on substantial diabetes rates after feeding soybean or wheat protein to NOD mice and/or BB rats.6 In conclusion, many protein-based diets promote the development of Type 1 diabetes-like disease in genetically predisposed animals, and these include A1 β-casein deficient milk.