Participants, methods, and results

We conducted a population based, prospective, four year follow up study of men aged 54 or 60 in the Kuopio ischaemic heart disease risk factor study, a population study in eastern Finland.4 Of 633 eligible men, 555 (88%) participated in the four year follow up. Of these, 508 were not diabetic (fasting blood glucose concentration ≥6.7 mmol/l and no treatment for diabetes)at baseline. A participant was defined diabetic at the end of the follow up if he had a fasting blood glucose concentration≥6.7 mmol/l, a blood glucose concentration of ≥10.0 mmol/l two hours after a glucose load, or clinical diagnosis of diabetes requiring dietary, oral, or insulin treatment.

The G to A transition at nucleotide 845 of the HFE cDNA, resulting in a substitution of tyrosine for cysteine at codon 282, was assayed by a solid phase minisequencing technique.5 The other strongest predictors of diabetes—blood glucose and serum fatty acid concentration,4 maximal oxygen uptake, and body mass index—were assessed with standard methods. We estimated odds ratios for developing diabetes with adjustment for risk factors using multivariate logistic regression models. All tests of significance were two sided.

We found one homozygote and 34 heterozygotes for the HFE C282Y mutation, giving a carrier frequency of 6.9% (including the homozygote). Four (11%) of the carriers and 23 (5%) of the 473 non-carriers developed diabetes during the four year follow up. In a multivariate logistic model, significant predictors of diabetes, assessed at baseline, were high body mass index, high ratio of saturated to the sum of polyunsaturated and monounsaturated fatty acids in serum, and the occurrence of the C282Y mutation (table). The Tyr282 allele carriers had an odds ratio of over 3.5 (95% confidence interval 1.02 to 12.1, P=0.047) for developing diabetes, compared with non-carriers.

Comment

Our prospective cohort study suggests an association between the common HFE gene mutation and the incidence of type 2 diabetes. Although our study was relatively small, our data support the theory that abnormalities of iron homeostasis contribute to the development of type 2 diabetes. If the observed association were causal, the C282Y mutation would account for as much as 15% of the incidence of type 2 diabetes in northern European populations with a carrier frequency of 7%. As the incidence of diabetes in northern Europe is among the highest in the world and rising, screening for C282Y mutation by DNA analysis, monitoring of iron status, and iron depleting treatment could potentially constitute new important measures in the primary prevention of diabetes. The C282Y mutation also predisposes to coronary heart disease.5 It may therefore prove especially important to find out whether the C282Y mutation increasesthe risk of myocardial infarction in people with type 2 diabetes.