Response to Comment on: Leeds et al. High Prevalence of Microvascular Complications in Adults With Type 1 Diabetes and Newly Diagnosed Celiac Disease. Diabetes Care 2011;34:2158–2163

We thank Bakker et al. (1) for their interest in our study (2) and are grateful for the opportunity to respond to their comments. Contention remains as to whether patients with type 1 diabetes should be routinely tested for celiac disease (CD), and the true effect of concomitant CD on microvascular complications has not been well studied (3). Our study set out to examine the prevalence of CD in type 1 diabetes and identified 12 new cases. This allowed us to compare the subjects’ baseline characteristics to age-, sex-, duration-, and weight-matched individuals with type 1 diabetes alone. HbA_1c was significantly higher in those with both type 1 diabetes and CD compared with those with type 1 diabetes alone, as was the prevalence of renal and retinal disease. HbA_1c is regarded as one of the most important predictors for the development of microvascular complications (4), but other risk factors have also been identified (5). As Bakker et al. point out, these cases were not matched to HbA_1c control subjects and suggest that this may be the underlying mechanism by which these microvascular complications occur. We agree that, ideally, to assess the effect of CD on microvascular complications HbA_1c should also be controlled for. Given that to identify 12 new cases of CD we had to screen 1,000 patients with type 1 diabetes, to perform the kind of study that Bakker et al. are describing would require a multicenter approach. We freely stated in the conclusion that the effect of CD on HbA_1c is likely to be the mechanism along with changes in other relevant risk factors such as HDL cholesterol. Our study, however, remains the first to follow up patients with newly identified CD in the context of type 1 diabetes and shows the effect of a gluten-free diet.

To finally answer the important question of the full effect of CD in type 1 diabetes, a multicenter study needs to be performed recruiting sufficiently large numbers to provide enough power to control for glycemic control as well as other risk factors.