In response to the recent article by Kilpatrick et al. (1) and the accompanying editorial (2), a number of letters to the editor have been published in Diabetes Care. Given the relevance of the topic, this is not surprising and actually is a good opportunity for additional comments to the article.
Monnier et al. (3), the Diabetes Research in Children Network (DirecNet) Study Group (4), and Irsch and Brownlee (5) all correctly point out that the seven-point blood glucose profile from which the calculations of Kilpatrick et al. (1) are based may not accurately reflect glucose variability. Therefore, they claim that the conclusion of Kilpatrick et al. (1) that glucose variability is not a risk factor for long-term microvascular complications is not valid (3–5).
No one can argue that the mean amplitude of glycemic excursion index is a more sophisticated assessment than the SD of the seven-point blood glucose profile and that, at least in theory, continuous glucose monitoring would be superior to single points of blood glucose measurements. However, it is important to remember that mean amplitude of glycemic excursions require continuous blood glucose monitoring. Moreover, there currently are limitations to measuring glucose continuously in the interstitial space. Factors such as poor knowledge of glucose dynamics in the interstitial versus vascular space and the unpredictable performance of the glucose sensor(s) and its limited ability to reliably transmit data of glucose reading make it unrealistic to use this approach to generate an extraordinarily large and solid database like the one used by Kilpatrick et al. (1) for their analysis, at least at the present time. Similarly, whereas the idea of additional studies in the area with more accurate assessment of glucose variability is acceptable in principle, it is unrealistic to think that a second Diabetes Control and Complications Trial study will be planned to assess the role of glucose variability on the development of microangiopathy.
Their critical comments (3–5) may also discount the simple finding by Kilpatrick et al. (1) that glucose variability did not correlate with the appearance of microangiopathy. Moreover, the limited value of the SD of glucose in the seven-point blood glucose profile is counter-balanced by the fact that: 1) the database of blood glucose readings in hundreds of subjects with type 1 diabetes is extraordinarily large, which makes it unlikely that important spikes of hyperglycemia were missed; 2) many subjects were followed-up for many years in their real life conditions; 3) real blood measurements, not interstitial glucose approximations, were done; and 4) blood glucose measurements correlated with “hard” end points, i.e., microangiopathy (1), and not surrogate markers such as oxidative stress (3). The limitations of the observation (1), including the fact that it is confined to type 1 diabetes and microangiopathy and does not allow extrapolation to type 2 diabetes and/or to macroangiopathy, have already been made clear (2).
The study of Kilpatrick et al. (1) clearly does not diminish the importance of insulin substitution and glycemic control in type 1 diabetes targeting an A1C <7.0%. However, the average risk for severe hypoglycemia in the intensively treated group of the Diabetes Control and Complications Trial was high, and at present, it is possible to use more precise modes of insulin delivery including either continuous subcutaneous insulin infusion (6) or multiple daily injections (7) based on insulin analogs to achieve an A1C <7.0% and at the same time limit hypoglycemia. These approaches also reduce large glycemic excursions. However, the Kilpatrick et al. article (1) suggests that these excursions alone do not translate into risk for microangiopathy.
