Rigalleau et al. (1) provide interesting findings from their observational follow-up study of long-term sequelae of type 2 diabetes that suggest rapid reduction in A1C (>1.5% in less than 4 months) is associated with significant worsening in diabetic retinopathy (DR), cardiovascular events, and foot ulcers, with notably high hazard ratios (38.74, 3.34, and 7.03, respectively). After this early and considerable decrease of A1C (mean 3.1%), these events presumably occurred over the ensuing mean 54 months of follow-up of their study. The comparator group had a stable A1C (although “stable” was not defined) and was not associated with increased DR worsening or cardiovascular events. Foot ulcers did nevertheless occur with a lower hazard ratio (3.08). They emphasize that in addition to the deterioration that may be seen in DR with rapid reduction in A1C discussed in our commentary (2), deleterious outcomes may occur in macrovascular disease and perhaps foot ulcers as well (1).

There is considerable evidence that rapid A1C reduction in patients with poorly controlled diabetes may cause worsening of DR (1–3). It appears that it is not only the rapidity of A1C reduction that plays a role in worsening outcomes but also its magnitude and likely the duration and severity of preexisting disease. However, it is not clear what the threshold is, if there is one, for the degree of A1C reduction with rapid glucose control that causes worsening of chronic complications of diabetes. We proposed an A1C reduction of ≥2% based on a comparison of studies in the same population (3,4), but that threshold has not been rigorously defined.

Furthermore, if there is a relationship between the rate/magnitude of decline in A1C and various outcomes, the time delay in the appearance of these complications may not be the same. For DR, worsening appears to occur relatively early, but for cardiovascular outcomes this may be delayed. The timing is unclear from the brief description available of the study by Rigalleau et al. Data from clinical trials also are not helpful. For example, a post hoc study of the ACCORD (Action to Control Cardiovascular Risk in Diabetes) trial did not find an association between reduction in A1C in the first 4 months of that trial and later mortality (5), an outcome different from but related to that of cardiovascular events reported by Rigalleau et al. However, that study was not designed to test this question.

It will therefore take a larger study specifically designed to test the question of whether euglycemic progression (3) of diabetic microangiopathy after rapid and severe reduction of A1C is a more generalized vascular problem, as suggested by Rigalleau et al. (1).