Harrison et al. (1) analyzed pancreatic β-cell function in drug-naïve patients with type 2 diabetes before and 42 months after treatment with insulin plus metformin or a combination of metformin, glyburide, and pioglitazone. The pretreatment evaluation was not exactly “before” the treatment. It was carried out after 3 months of the run-in period with insulin plus metformin therapy for all participants. The study is well designed, the analysis is appropriate, and the data are straightforward except for two potential drawbacks: absence of a wash-out period prior to the final meal test and failure to pay attention to alteration in insulin sensitivity caused by the treatments. The authors withheld hypoglycemic agents only for 24 h before the final meal test. This is inappropriate to evaluate the disease-modifying effect of the two treatment modalities. Nevertheless, the title says “preservation of β-cell function,” i.e., presence of a disease-modifying effect. Glyburide is especially problematic: the insulinotropic effect of glyburide does not fade away within 24 h (2). Therefore, it is impossible to distinguish a direct insulinotropic effect of the drug and the disease-modifying effect of it with this protocol. It was absolutely needed to stop glyburide at least for a few days, ideally for a week, before the “post-treatment” meal test in order to identify the β-cell preserving effect, if any. Of course, the final meal test should have been carried out after the same wash-out period in the insulin- and metformin-treated group for the fair comparison of the two treatment modalities. The second point is also profoundly important. Pioglitazone’s insulin-sensitizing effect may last longer than 24 h (3). In addition, glucose lowering by any means restores insulin sensitivity of the body to some extent (4). Therefore, β-cell function should have been evaluated with alteration of insulin sensitivity taken into consideration, especially after an appropriate wash-out period for pioglitazone. The authors found no change in plasma C-peptide or C-peptide–to–glucose ratio by the two treatments and took the finding as evidence for “preserved β-cell function” (1). However, it may imply improved or worsened β-cell function if insulin sensitivity had gone up or down, respectively. The most popular approach to obviate this problem is calculating oral disposition index (5). At least, the authors should have commented on the above-mentioned two issues as limitations of the study. The title might be misleading in that it gives an impression that the disease-modifying effect, which is of paramount importance, was successfully achieved by the conventional pharmacological therapies.