• diabetes

In our recently published analysis of Newcastle Heart Project data we concluded that in South Asians diabetes or impaired glucose tolerance was associated with an increased likelihood of reporting sibling but not parental diabetes.¹ By contrast, in Europeans diabetes or impaired glucose tolerance were associated with both sibling and parental diabetes. In response to feedback, we re-analysed our data to assess whether awareness of a diagnosis of diabetes influenced reporting and to re-consider the possible influence of our direct standardisation procedure.

In brief, the Newcastle Heart Project studied an age and sex stratified population sample of 684 South Asian (259 Indian, 305 Pakistani, 120 Bangladeshi) and 825 European men and women aged 25–74 years in Newcastle upon Tyne, UK. Diabetes was defined as a plasma glucose sample ≥11.1 mmol/l two hours after a 75 g oral glucose load and impaired glucose tolerance >7.8 and <11.1 mmol/l, following World Health Organisation (1985) criteria. Respondents prescribed oral hypoglycaemic agents or insulin were classified as having diabetes. Full details of the study have been published.²

Because of the small numbers clear cut patterns did not emerge from age specific data. We were therefore reliant on age standardised analysis. Table 1 shows typical results. In Europeans the age/sex adjusted prevalence of parental diabetes rose across the categories of normal glucose tolerance, impaired glucose tolerance and previously undiagnosed diabetes (p=0.017 for linear trend across these categories). Parental diabetes was more frequent in those already known to have diabetes than those with newly diagnosed diabetes. In South Asians the equivalent associations were less clear cut. There was no appreciable difference in reporting between those with newly diagnosed or previously known diabetes.

In Europeans the prevalence of sibling diabetes showed a strong positive relation with respondent glucose tolerance (p<0.001 for linear trend) There was little difference in the prevalence of sibling diabetes between those with newly diagnosed and known diabetes. Once again, in South Asians the equivalent associations were less clear cut. Those with known diabetes were more likely than those with newly diagnosed diabetes to report diabetes in their siblings.

Even after excluding respondents with known diabetes, we found that among Europeans there was a clear relation between impaired glucose tolerance and diabetes in respondents and reported diabetes in parents and siblings. The relation with parental and sibling diabetes in South Asians was weaker. Exclusion of those with known diabetes also weakened the previously reported association with diabetes in South Asian siblings, raising the possibility that awareness of a diagnosis of diabetes may influence reporting of family histories of diabetes both in Asians and in Europeans.

This study shares with the majority of previous reports the weakness that it is dependent on respondent reports rather than biochemical testing for family diagnoses of diabetes. It is also possible that there are ethnic differences in family reporting. The relatively large number of missing responses (which may reflect uncertainty about family history) also urge caution in interpreting these results.

With these caveats, we conclude that familial clustering of diabetes is more clear cut in Europeans than in South Asians. Larger studies with biochemically verified information about parental and sibling diabetes may help to elucidate the causal basis of ethnic variations in diabetes.