Confirming the Link Between New-Onset Diabetes and Pancreatic Cancer

Recent research results validate the link between a new diagnosis of diabetes and pancreatic cancer for a small percentage of patients.

A paper published in January 2026 in the journal Gastroenterology showed that about 1 in 160 people were diagnosed with pancreatic cancer within three years of their new-onset diabetes diagnosis. In that group, diabetes was identified an average of eight months before a clinical pancreatic cancer diagnosis was made. These are interim results, because full data analysis is not yet complete.

Although the study results were not necessarily a surprise, “the data was very striking,” says lead investigator Suresh Chari, M.D., professor in the department of gastroenterology, hepatology and nutrition at the University of Texas MD Anderson Cancer Center, Houston, Texas.

Initial Research on the Diabetes–Pancreatic Cancer Link

Back in 2005, Chari and colleagues published a paper showing that for three years after their diagnosis with diabetes, patients have eight times the risk of developing pancreatic cancer.

They studied 2,122 patients from Rochester, Minnesota, age 50 and older, who were diagnosed with diabetes between 1950 and 1995. Although pancreatic cancer is not a common disease, 18 of the participants were diagnosed with this cancer within three years of their diabetes diagnosis. The researchers then compared this rate with that expected for people of similar age and gender who do not have diabetes. The group with newly diagnosed diabetes had eight times the expected rate.

How do I find clinical trials?

Find a trial

“This was our first published paper on the topic and we needed to conduct a prospective study to prove our findings,” Chari explains. “That took about 20 years, but the results are very important. Pancreatic cancer is so difficult to detect early. Patients are generally diagnosed when the disease is advanced and we need to change that.”

What Is the REGARD Study?

The REGARD (Recent Elevation of Glucose and Recent Diabetes) study is what is called a prospective study, where the researchers look at participants over time. For this study, the team used electronic health records to follow a large group of people who received a new diabetes diagnosis. Chari and colleagues wanted to validate the association between new-onset diabetes and pancreatic cancer.

One of the issues, Chari explains, is that the definition of “new-onset diabetes” in earlier studies often varied. Glycemic onset of diabetes is defined as an A1c level of 6.5 percent and above and a fasting plasma glucose of 126 mg/dL or higher with confirmed lack of prior evidence of diabetes. In the earlier study there was often a gap of sometimes months or years between the onset of hyperglycemia (when the blood sugar levels consistently become higher) and a physician diagnosing diabetes in the patient. Chari states, “What we needed to do was identify diabetes at the very first glycemic onset, which is based on A1c levels and fasting plasma glucose test results.”

“These glycemic parameters actually served as predictors of pancreatic cancer risk,’’ he adds. “In these patients, pancreatic cancer is causing the A1c elevation. It is not simply an association with diabetes.”

The study participants included adults aged older than 50 years with confirmed lack of prior diabetes and who had been receiving care at Kaiser Permanente Southern California, Kaiser Permanente Northern California, Baylor College of Medicine/Kelsey-Seybold Clinic, and Mayo Clinic Minnesota. Based on a sophisticated algorithm and active real-time surveillance of electronic health records, the researchers identified 18,838 adults who had new-onset diabetes based on glycemic criteria. The race and ethnicity breakdown included 6,518 non-Hispanic White participants, 5,984 Hispanic participants, 3,360 Asian/Pacific Islander participants, and 2,192 African-American participants.

The median follow-up for the interim analysis reached 2.3 years, and during this period 82 participants were diagnosed with pancreatic cancers. The mean age of diagnosis was 71, and 60 percent of those diagnosed were men. Stage IV disease accounted for 65 percent of the diagnoses.

There were racial disparities. The study’s interim results showed overall race-adjusted three-year pancreatic cancer incidence of 0.62 percent, with higher absolute three-year incidence among non-Hispanic White people than Hispanic, African-American, or Asian/Pacific Islander individuals.

The Diabetes–Pancreatic Cancer Relationship

“What was striking is that new-onset diabetes occurred on average about eight months before an individual received a clinical pancreatic cancer diagnosis,” Chari says. “And about 40 percent of cancers were diagnosed a year or even longer after new-onset diabetes. These findings show that we have a window of opportunity to potentially diagnose pancreatic cancer earlier. And that gives people a better chance at a better outcome. This disease is so aggressive and so difficult to detect that we really need to use all the information and tools we have available to detect pancreatic cancer earlier.”

Chari emphasizes the difference between those with known established diabetes, which is the vast majority of patients with diabetes, and those with new-onset diabetes. The risk of developing pancreatic cancer in those with established diabetes is very small. “New-onset diabetes is not very common in a physician’s everyday practice,” says Chari. “About one in 200 patients over age 50 would have it. Paying attention to these patients may help battle one of the deadliest cancers.

How do I find clinical trials?

Find a trial

“Although we are certainly making progress in treating the disease, we aren’t where we need to be, especially when the disease is advanced. Diagnosing an aggressive cancer like pancreatic cancer eight months earlier by doing something as simple as a CT scan if that patient has new-onset diabetes could potentially make a big difference in someone’s life.”

What’s Next for Diabetes­–Pancreatic Cancer Research?

The researchers are planning a follow-up study to establish the full five-year risk of developing pancreatic cancer after glycemic onset of diabetes. They suspect the risk will level off after three years in this group of patients and will be about 1.5- to 2-fold higher than that of the general population.

The researchers are also working on a prospective study designed to evaluate the validity of the ENDPAC (Enriching New-Onset Diabetes for Pancreatic Cancer) score. This score classifies people with new-onset diabetes into low, intermediate, or high risk for pancreatic cancer, based on body weight, age, and glucose/HbA1c levels. The REGARD investigators went back to collect prior weight data to study the ENDPAC score. These results will be validated in another prospective study, the Early Detection Initiative, specifically designed to study the ENDPAC score; this study is sponsored by the Pancreatic Cancer Action Network (PanCAN).

Earlier detection remains an important focus of Chari’s work. He is involved in a study led by Ajit Goenka, M.D., and researchers at the Mayo Clinic (Rochester, Minnesota) showing that AI-assisted CT scans can pick up pancreatic cancer on an average 475 days before clinical diagnosis. Chari also notes he is working with Goenka on a prospective study of AI-assisted CT scans to diagnose pancreatic cancers earlier in those with new-onset diabetes and high ENDPAC scores.