Combining C-peptide with glucose could help better predict type 1 diabetes

Among autoantibody-positive people with elevated glucose on an oral glucose tolerance test (OGTT), 88% of those with low C-peptide levels developed clinical type 1 diabetes within five years, compared with 41% of those with high C-peptide levels, according to data highlighted in a perspective published April 30, 2026, in Diabetes Care.

Type 1 diabetes is an autoimmune disease where the immune system damages the insulin-producing beta cells of the pancreas. Since 2015, folks at risk have been grouped into three stages: stage 1, meaning two or more diabetes-related autoantibodies with normal blood sugar; stage 2, which adds abnormal blood sugar that falls below the diagnostic range and stage 3, being a clinical diagnosis, according to the piece led by Emily K. Sims, M.D., pediatric endocrinologist at Riley Hospital for Children and a physician scientist at the Center for Diabetes and Metabolic Diseases at Indiana University School of Medicine in Indianapolis.

Regulatory approval and a growing pipeline of therapies that modify the disease before symptoms have driven more autoantibody screening, Sims and her team added. The FDA approved Tzield (teplizumab) in November 2022 to delay stage 3 disease in people 8 and older with stage 2 type 1 diabetes. Better ways to measure progression would help with counseling, treatment timing and early readouts of therapy response, they added.

The authors argue that because staging relies on glucose, it can misread risk. They make the case for reading glucose alongside C-peptide, a marker of how much insulin the body is still making on its own. Sims reported consulting fees from Sanofi, which markets teplizumab, and is a coinventor on a patent application describing methods to measure the metabolic effects of immunotherapy. Many of the co-authors reported advisory or speaking ties to Sanofi.

Five findings from earlier research

Rather than report a single new trial, the authors highlighted five findings from earlier analyses, many of them their own. Most of the data came from the “Diabetes Prevention Trial–Type 1 (DPT-1)” and the “Type 1 Diabetes TrialNet Pathway to Prevention (TNPTP)” study, which follow autoantibody-positive relatives of people with type 1 diabetes. The researchers compared glucose alone with combined measures such as Index60, a score that blends glucose and C-peptide values. They also plotted glucose against C-peptide on a grid to track how each group's response shifted over time.

What the data showed

In the five-year analysis, which included 1,136 TNPTP participants with low C-peptide and 1,342 with high C-peptide, the gap held at every glucose level. Out of those with low C-peptide, risk rose steadily as glucose rose. Among those with high C-peptide, the pattern was irregular. A separate analysis found that abnormal blood sugar was unrelated, or even oppositely related, to known type 1 diabetes risk factors. Index60 tracked with the same risk factors as type 1 diabetes. In another analysis, 34 people who first tested in the diabetes range with a low Index60, and were later diagnosed had much higher C-peptide than two other groups that progressed, and that gap persisted at diagnosis.

The authors said this group may need a different treatment approach. Among 72 DPT-1 and 82 TNPTP participants who developed diabetes, C-peptide rose along with glucose until about 1.5 years before diagnosis. It then began to fall as glucose kept climbing, a shift the authors described as a loss of the body's ability to compensate.

Combined measures could also offer earlier answers in prevention trials. For example, the teplizumab prevention trial showed an effect on progression after several years of follow-up, but combined glucose and C-peptide endpoints detected a lasting drug effect within 3 months of treatment.

In a trial of abatacept in people with stage 1 disease, combined endpoints showed signs of a drug effect after 1 year, even though the trial was negative on its primary endpoint of time to abnormal blood sugar or diabetes. Without C-peptide, “there can be misinterpretations of glucose levels and missed opportunities to gain important information,” the authors wrote.

Limits and barriers to clinical use

The authors noted that findings were consistent across research groups. However, most of the data focus on relatives of people with type 1 diabetes. General population screening studies show similar progression once multiple autoantibodies appear, which the authors said suggests the findings apply broadly, though that was not tested directly here. Some subgroups were small, and many cited analyses come from the same team.

It’s important to note that hurdles still remain. C-peptide tests used in trials are not standardized and often are not run by local labs, so scores could vary by lab. A reference method exists, but manufacturers have not yet recalibrated their tests. OGTTs are also rarely done in routine care, though a one-hour OGTT may offer similar risk-score accuracy.

The authors stopped short of calling for a new staging system. Rebuilding staging around combined measures could improve risk prediction, but the recent push to educate clinicians on the current system would make that switch difficult, they said. Instead, the authors suggested using combined measures to refine risk within the existing stages. They also said regulatory requirements or business initiatives could speed test standardization and that removing barriers to C-peptide testing in clinical care for autoantibody-positive people “should be a high priority moving forward.”