Changes in Prandial Glucagon Levels After a 2-Year Treatment With Vildagliptin or Glimepiride in Patients With Type 2 Diabetes Inadequately Controlled With Metformin Monotherapy

The study was an extension to 2 years of a previously described study (4). Standard meal challenge was performed in selected centers at last treatment visit after administration of a morning dose of metformin plus 50 mg vildagliptin or up to 6 mg glimepiride (mean treatment period 1.8 years). Doses were selected to achieve comparable improvements in glycemia. After an overnight fast, a mixed meal was served (orange juice [180 ml], two slices [60 g] of white bread, 30 g jam, 15 g butter or margarine, 120 ml whole milk [3–4% fat] or equivalent amount of cheese plus 120 ml water, and, if desired, decaffeinated coffee or tea; 510 kcal, with 50% from carbohydrate, 38% from fat, and 12% from protein). Blood samples were taken before the meal and at 15, 30, 60, 90, 120, 180, and 240 min.

Plasma glucose concentration was determined by the glucose oxidase method (Beckman Glucose Analyzer II; Beckman Instruments, Fullerton, CA). Plasma insulin, C-peptide, and glucagon concentrations were determined by radioimmunoassay (Diagnostics Products, Los Angeles, CA). Plasma intact glucagon-like peptide (GLP)-1 concentration was measured by ELISA using an NH₂-terminal–specific antibody (Linco Research, St. Charles, MO) by Novartis.

Insulin secretory rate was determined as described previously (5). The absolute and incremental/decremental areas under the curve (AUCs) for time 0–2 h after the meal were calculated using the trapezoidal method. End point changes from baseline were assessed using ANCOVA.

The study was conducted in accordance with the Declaration of Helsinki. It was reviewed by an independent ethics committee or institutional review board for each center. Written informed consent was obtained from each subject.

Baseline characteristics were as follows: patient demographics of the subpopulation who participated in the meal challenge (n = 259) were essentially the same as reported previously (4). Mean baseline A1C was 7.3 ± 0.6%.

At the follow-up meal test (up to 2 years after start; mean 1.8 years), A1C was reduced by −0.1 ± 0.9% in the vildagliptin group (n = 137) versus −0.2 ± 0.8% in the glimepiride group (n = 121). Prandial glucose AUC_{0–2 h} was similarly reduced in both groups (−1.7 mmol · h⁻¹ · l⁻¹ for vildagliptin vs. −2.1 mmol · h⁻¹ · l⁻¹ for glimepiride), whereas prandial insulin 1 AUC_{0–2 h} increased to a greater extent in the glimepiride group (33 ± 18 pmol · h⁻¹ · l⁻¹ for vildagliptin vs. 91 ± 19 pmol · h⁻¹ · l⁻¹ for glimepiride) (P = 0.017).

Prandial glucagon AUC_{0–2 h} decreased from baseline with vildagliptin treatment but increased with glimepiride (−3.4 ± 1.6 pmol · h⁻¹ · l⁻¹ for vildagliptin vs. 3.8 ± 1.7 pmol · h⁻¹ · l⁻¹ for glimepiride; P < 0.001; Fig. 1,A). Prandial intact GLP-1 AUC_{0–2 h} increased in both groups but was much larger after vildagliptin (18.8 ± 1.8 pmol · h⁻¹ · l⁻¹ for vildagliptin vs. 1.6 ± 1.7 pmol · h⁻¹ · l⁻¹ for glimepiride; Fig. 1,B). Insulin secretion rate relative to glucose (0–2 h) increased in both groups (4.3 ± 0.9 pmol/min/m²/mmol/l for glimepiride vs. 1.6 ± 0.9 pmol/min/m²/mmol/l for vildagliptin; P = 0.022; Fig. 1 C). Prandial insulin-to-glucagon ratio (AUC_{0–2 h} for insulin/AUC_{0–2 h} for glucagon) changed by −1.1 ± 9.3 (baseline 7.4 ± 0.4) pmol insulin/pmol glucagon in the vildagliptin group versus −7.3 ± 9.8 (baseline 7.1 ± 0.4) pmol insulin/pmol glucagon in the glimepiride group (P = 0.62).

Fasting insulin (15.2 ± 1.6 pmol/l for glimepiride vs. 5.7 ± 1.6 pmol/l for vildagliptin; P < 0.001) and HOMA-IR (0.11 ± 0.10 for vildagliptin vs. 0.63 ± 0.10 for glimepiride; P < 0.001; Fig. 1 D) increased in both treatment groups with larger increases with glimepiride.

Vildagliptin reduces glucagon levels after a standard mixed meal in patients with type 2 diabetes who are treated with metformin, as was previously demonstrated in individuals with IGT (6), IFG (7), and type 1 diabetes (8). This effect appears to be the result of a GLP-1–induced improvement in glucose sensitivity of the α-cells (3,9). In contrast, glimepiride increases glucagon levels, which may be the result of attenuation of the glucose sensitivity of the α-cells with uncoupling of glucose dependency (10).

With respect to β-cell function, vildagliptin increases glucose sensitivity (11), whereas indirect data suggest that sulfonylurea uncouples the glucose dependency of the β-cells, which attenuates glucose sensitivity (10). In this study, this resulted in increased insulin secretion by both treatments, although vildagliptin did not increase the insulin secretion rate as much as glimepiride. The greater insulin secretion rate in the glimepiride group was balanced by less insulin resistance in the vildagliptin group, as reflected by HOMA-IR and lower postprandial glucagon levels resulting in similar glucose levels.

In summary, prandial glucagon is increased by glimepiride but reduced by vildagliptin. Hence, vildagliptin effectively targets glucagon secretion in the treatment of diabetes. This together with the lesser increase in insulin by vildagliptin than by glimepiride may provide a pathophysiological explanation for the observation that vildagliptin's effect of reducing A1C levels below 7% is associated with much less hypoglycemia than glimepiride while being associated with the same A1C reduction (4).

This study was funded by Novartis Pharmaceuticals. D.R.M. is a Senior Investigator of the National Institute of Health Research (NIHR), U.K., and acknowledges NIHR funding support. E.F. has received consulting fees from Novartis Pharmaceuticals. V.A.F. has received funding for research and consulting/speaking for Novartis as well as from other manufacturers of drugs in this class, including Merck, Novo-Nordisk, Eli Lilly, Takeda, and Glaxo. B.Z. has received research support and consulting fees from Novartis Pharmaceuticals. B.A. has received funding and consulting fees from Novartis Pharmaceuticals. D.R.M. has received consulting fees from Novartis as well as from other manufacturers of drugs in this class, including Novo-Nordisk and Merck. He has also received lecturing honoraria from Eli Lilly and Novo-Nordisk. J.F. and S. D. are employees of Novartis Pharmaceuticals. No other potential conflicts of interest relevant to this article were reported.

The authors gratefully acknowledge the investigators, those who participated in any part of the study (listed previously in Ferrannini et al. [4]), staff at the 402 participating sites, and the editorial assistance of Gali Venkateswara Prasad and Shivali Arora, PhD (both from Novartis, India).