Diabetes CareType 2 DiabetesRCTDecember 23, 2025

Differential Longitudinal Effects of Glucose-Lowering Medications on Glucagon and C-peptide Responses in the Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE).

Kahn SE, Tripputi M, Lachin JM, Balasubramanyam A, Banerji MA, Barzilay J, Cohen RM, Garvey WT, Gramzinski MR, Rasouli N, Rhee M, Seegmiller JC, Singh V, Sivitz WI, Steffes MW, Utzschneider K, DeFronzo RA, GRADE Research Group*

View on PubMed DOI: pii: dc252186. 10.2337/dc25-2186

Key Finding

In 724 people with type 2 diabetes, beta-cell function (but not alpha-cell function) predicted blood sugar control, with every standard deviation improvement in C-peptide index reducing the risk of poor glycemic control by 17%.

What This Study Found

Think of your pancreas as having two types of workers: alpha-cells that make glucagon (which raises blood sugar when it's too low) and beta-cells that make insulin (which lowers blood sugar when it's too high). This study followed 724 people with type 2 diabetes for years to see which worker mattered more for keeping blood sugar controlled. The researchers tested four different diabetes medications added to metformin - like having four different management strategies for these pancreatic workers. What they discovered was surprising: how well the alpha-cell workers were doing (measured by glucagon responses) didn't predict whether someone's diabetes would get worse over time. However, the beta-cell workers' performance (measured by C-peptide, which is made alongside insulin) was the real predictor of success. It's like discovering that in a two-person assembly line, one worker's efficiency doesn't matter much for the final product, but the other worker's performance is crucial. For every standard deviation improvement in beta-cell function (C-peptide index), people had a 17% lower chance of their blood sugar control deteriorating to concerning levels.

Statistics Decoded

P = 0.55 for alpha-cell measures means there's a 55% chance this result was just random luck - not convincing evidence. P = 0.04 and P < 0.001 for beta-cell measures means these results were very unlikely to be coincidence - like rolling snake eyes multiple times in a row. The 17% risk reduction means if 100 people typically develop poor blood sugar control, improving beta-cell function by one standard deviation would reduce that to about 83 people. The 2 df refers to degrees of freedom in the statistical test - essentially how many variables they were comparing at once.

Why This Matters

This study suggests doctors should focus on choosing diabetes medications based on how they support beta-cell (insulin-making) function rather than worrying about alpha-cell (glucagon-making) effects, potentially simplifying treatment decisions and improving long-term outcomes for millions of people with type 2 diabetes.

Original Abstract

To determine the longitudinal effects on α-cell function of four classes of glucose-lowering agents added to metformin in adults with type 2 diabetes. In a multicenter, randomized study, 5,047 participants ≥30 years of age with type 2 diabetes taking 1,000-2,000 mg metformin daily, HbA1c 6.8%-8.5%, were randomized to receive insulin glargine U100, glimepiride, liraglutide, or sitagliptin. In a subset of 724 participants, baseline and longitudinal measures of α-cell function were assessed as the fasting glucagon concentration and change in glucagon from fasting to 30 min following oral glucose (glucagon index [GGI]). Whether these measures and those of β-cell function (fasting C-peptide and C-peptide index [CPI]) were related to the primary metabolic outcome (HbA1c ≥7.0%) was examined. Baseline fasting glucagon and GGI were not associated with the primary metabolic outcome (2 df; P = 0.55), whereas the β-cell measures were (2 df; P = 0.04). Treatment-associated changes in fasting glucagon, GGI, and fasting C-peptide were not associated with the primary metabolic outcome, while changes in CPI were, in models unadjusted (P < 0.05) or adjusted (P < 0.001) for α-cell function. A 1-SD increase in CPI was associated with a 17% reduction in the risk of the primary metabolic outcome. There were no clear differential effects of the medications on glucagon responses. There were different patterns of effects of the four glucose-lowering medications on the α-cell, with no relationship of α-cell function with worsening glycemia. Thus, in treating type 2 diabetes, for choice of medication(s) the focus should primarily be on their impact on β-cell function.