LAUSR

We are grateful to Dr. Srinivas for his thorough review of our data. He points out potential differences in the pharmacodynamic effects of insulin preparations between healthy people and people with diabetes and discusses several reasons for these differences, including insulin dose, effect of endogenous insulin, anti-insulin antibodies, and local injection site issues in people with diabetes [1]. Indeed, there have been some discussions about the validity of pharmacodynamic results with the euglycemic glucose clamp technique in healthy people, mostly related, however, to the assessment of duration of action and late metabolic activity of basal insulins [2, 3]. Because of the high metabolic activity of prandial insulins, endogenous insulin in healthy people is largely suppressed under glucose clamp conditions and should not contribute significantly to the observed pharmacodynamic effects. This is confirmed by previously published data with regular human insulin and insulin aspart injected at different injection sites in healthy people where the observed differences between the two study insulins were consistent with and without adjustment for endogenous insulin using C-peptide concentrations [4]. While we focused on pharmacokinetic results in our study investigating the impact of injection site on the properties of faster aspart [5] to avoid any discussions on the validity of pharmacodynamic data in healthy people, the results of a slower and slightly blunted insulin absorption with injection in the thigh compared with abdominal injections is consistent with published data of other insulins in both healthy people and patients with type 1 diabetes. Dr. Srinivas cited our study investigating the pharmacokinetic/pharmacodynamic effects of insulin degludec when injected into different regions [6] and the study by Süsstrunk et al. [7] comparing the absorption of regular human insulin after injection into abdomen or thigh. While the differences between injection sites did not reach statistical significance in the study by Süsstrunk et al. (which is hardly surprising considering a sample size of n = 4), the tendency to a delayed and lower absorption was consistent across the two studies. In our paper we listed the results of other investigations with short-acting insulin analogues done in healthy people all showing slower and impaired insulin absorption from the thigh compared with abdominal injections [4, 8]. Similar results were obtained in people with type 1 diabetes, e.g. in the study by Koivisto et al. [9] that Dr. Srinivas referred to, but also in other investigations [10, 11]. Moreover, a slower and impaired absorption from the thigh has been reported for other injectable drugs such as growth hormone [12] or exenatide [13]. The latter study was done in people with type 2 diabetes, which clearly demonstrates that these results are consistent across populations. The reason for the slower absorption has been investigated by Vora et al., who could demonstrate a positive relationship between the absorption of soluble insulin and subcutaneous blood flow [14] for all injection sites. The so-called ‘‘lag phase’’ (i.e., the initial phase of slow insulin absorption) was shortest with abdominal insulin injections, so that a higher subcutaneous blood flow in the abdomen versus the thigh is the most This reply refers to the article available at doi:10.1007/s40261-0170499-y.