LAUSR

We were interested by the recent article from Miyang Luo et al, who reviewed longitudinal studies of subjects with both types of diabetes and described long‐term trends in HbA1c. For type 1 diabetes, they identified stable, low to moderate (7.1% to 9% HbA1c) as the main pattern, whereas extremely low or high but stable, deteriorating or improving HbA1c were possible for a minority of subjects: 2% to 45% among the 11 reviewed studies. In Bordeaux university hospital, we follow a cohort of subjects with type 1 diabetes since 2009. One hundred and eighty‐five participants had HbA1c measurements both at inclusion and 7 years later (from 2009 to 2016). They were mainly men (N = 105, 56%), 51 ± 14 years old, with 22 ± 13 years of diabetes duration. HbA1c was 7.5 ± 0.9% at baseline, which did not differ from the values 6 months before (7.6 ± 0.8%) and 12 months before (7.6 ± 0.8%). Seven years later, the mean HbA1c remained at 7.5 ± 0.9% and related with baseline value (ß = +0.523, P < 0.001). We registered HbA1c at a limited number of time points, which is less accurate than clustering from serial HbA1c measurements, but on the mean our participants seemed to track in the main “stable, low to moderate” HbA1c pattern identified by Luo, at first view. However, these findings contrast with our daily practice, as we often see subjects with important HbA1c excursions, a common cause for seeking specialized diabetology care. We investigated these subjects, by categorizing our participants according to the changes of HbA1c from 2009 to 2016: HbA1c declinors were defined as < −0.7% decline, and HbA1c progressors as > +0.7% HbA1c progression. We chose a ±0.7% HbA1c change because this was the maximal amplitude of initial long‐term changes in the recent report from Nirantharakumar K et al on HbA1c tracking in type 1 diabetes. As