We thank Mikhail and Cope (1) for their comments on our study (2). Mikhail and Cope are correct in noting that the decrease in 2-h postprandial glucose was much larger for the inhaled human insulin (INH) group compared with metformin. As the authors suggest, this could have contributed to the lower HbA1c (A1C) level in the INH arm. A1C is the sum of both fasting and postprandial glucose excursions. From work by Monnier et al. (3), it is known that there is a variable relationship between fasting and postprandial glucose based on current A1C levels, such that the influence of fasting and postprandial glucose is equivalent when the A1C is ∼8.4%. However, the lower the A1C, the greater the contribution of postprandial glucose and the higher the A1C, the greater the contribution of the fasting glucose. Thus, all components (postprandial glucose, fasting glucose, and A1C) must be treated for optimal glycemic control (3).
Mikhail and Cope draw comparisons between our trial and a three-arm study by Rosenstock et al. (4), in which INH was substituted for or added to oral combination therapy, and suggest that premeal INH added to one or more oral agents may not be sufficient to achieve glycemic control. While the majority of patients in these trials did not achieve current recommended A1C targets (5), it should be noted that the American Diabetes Association A1C action level at the time of these studies was <8% (6). Blood glucose titration targets were therefore not planned to achieve an A1C of <7%.
We do not believe insulin-induced hypoglycemia was the limiting factor that prevented investigators from increasing the insulin dose in our study. While hypoglycemia was the most commonly reported adverse event, there were no discontinuations due to hypoglycemia. Although not presented in our article, detailed analyses of hypoglycemic events including time of onset were conducted. We can report that crude event rates increased gradually over time and were highest at 8 weeks after randomization, after which there was a gradual decline in event rates toward the end of the study. The distribution of time of onset of hypoglycemic events revealed that the INH group had most occurrences during the period of 11:00 a.m.–11:00 p.m.
In our trial, patients with very high A1C (≥9.5%) were more effectively treated with the addition of INH to oral agent monotherapy from the standpoint of glycemic control. Head-to-head comparisons of INH versus a long-acting insulin analog treating to current glycemic targets are ongoing.
References
A.H.B. has received honoraria and research grants from Eli Lilly, Novo Nordisk, and Roche. M.D. has received honoraria from sanofi-aventis, Novo Nordisk, Eli Lilly, GlaxoSmithKline, and AstraZeneca. P.L. has served on an advisory panel and received consulting fees from sanofi-aventis and Pfizer.