(© Peakstock - stock.adobe.com)
In A Nutshell
- An Early But Unreliable Signal: In Mayo Clinic records, average fasting blood sugar in people later diagnosed with pancreatic cancer began pulling ahead of matched people without cancer about 30 to 36 months before diagnosis, but those studies tracked group averages and did not produce a test.
- Most New Diabetes Is Not Cancer: In a large prospective study of adults 50 and older with newly detected diabetes, an estimated 0.62 percent, or about 1 in 160, were diagnosed with pancreatic cancer within three years, and a separate analysis found HbA1c alone too weak for widespread screening.
- Patterns Matter, Screening Stays Selective: High or rapidly rising blood sugar, weight loss, older age, pancreatitis, gallstones and family history were linked to higher risk, yet the U.S. Preventive Services Task Force still advises against screening adults without symptoms whose risk comes from factors such as new diabetes or older age.
The pancreas helps keep blood sugar in check. So it may not be surprising that when a tumor grows there, blood sugar can start to shift, sometimes long before anyone suspects cancer.
For a disease that usually causes symptoms only after it has spread, according to the National Cancer Institute, that possibility has drawn years of research. The findings point to a real signal, but a messy one. The change can begin early, yet the institute notes that most people who develop diabetes will never get pancreatic cancer. The first question is how early that signal appears.
When Can Blood Sugar Begin To Change?
About three years before diagnosis, on average, in Mayo Clinic records. That does not mean a single lab result can reveal a hidden cancer.
A 2018 case-control study in Gastroenterology pulled five years of fasting blood sugar tests for 219 people in Olmsted County, Minnesota, who were diagnosed with pancreatic ductal adenocarcinoma, a form of pancreatic cancer. The Mayo Clinic researchers compared them with people of the same age and sex who did not have cancer. Average fasting glucose in the cancer group first rose significantly above the comparison group 30 to 36 months before diagnosis and kept climbing. It crossed the diabetes threshold of 126 milligrams per deciliter roughly six to 12 months before diagnosis. In another group of patients in the same paper, high blood sugar went away after the tumor was surgically removed. That suggests the tumor itself may have been driving the rise.
An earlier Mayo Clinic records study, published in 2009 in The American Journal of Gastroenterology, compared 736 patients with pancreatic cancer and 1,875 people without it. From five to three years before diagnosis, average fasting glucose did not differ meaningfully between the groups. In the cancer group it then rose step by step over the final three years.
Both studies started with people already known to have cancer and looked backward. The 2018 team described the idea that cancer can be found when blood sugar first starts rising as “an unproven but hopeful premise.” Long-standing diabetes is a separate matter. In earlier work, the same Mayo group noted that long-standing diabetes is thought to raise pancreatic cancer risk, while new diabetes may instead be a symptom of the cancer.
Does New-Onset Diabetes Usually Mean Pancreatic Cancer?
Usually not, and the best numbers show how rare the link is.
The clearest estimate of the risk comes from a prospective study in Gastroenterology, published online in 2025 and in the journal’s January 2026 issue. Prospective means the researchers identified people first and then followed them forward. Using real-time monitoring of electronic health records, the team flagged 18,838 adults age 50 or older whose lab results newly met the definition of diabetes. Over a median of 2.3 years, 82 were diagnosed with pancreatic cancer.
After adjusting for the racial makeup of Americans with new diabetes, the estimated three-year risk was 0.62 percent, or about 1 in 160. Risk varied by group, from 0.84 percent among non-Hispanic white patients to 0.22 percent among Asian and Pacific Islander patients. Among people who did develop cancer, diabetes showed up an average of eight months before the diagnosis.
A different blood test tells a similar story. A 2020 analysis in JAMA Network Open examined 851,402 Kaiser Permanente Southern California members ages 50 to 84 who had HbA1c tests, a lab measure of blood sugar. Pancreatic cancer risk rose with higher HbA1c cutoffs and stricter definitions of newly high blood sugar, but the pattern was not consistent across racial and ethnic groups. The authors concluded that screening based on HbA1c alone was unlikely to work. Depending on the cutoff and definition used, doctors would need to evaluate 206 to 600 patients to find one cancer.
Which Patterns Raise More Concern?
A cluster of signs, rather than one lab result.
A 2022 systematic review and meta-analysis in Cancers pooled 22 observational studies covering 576,210 people with new-onset diabetes, 3,560 of whom had pancreatic ductal adenocarcinoma. A meta-analysis combines results from many studies to get a more precise estimate. Across four studies, high or rapidly rising blood sugar was tied to more than twice the risk (a pooled estimate of 2.33 that combined risk and odds ratios). Weight loss, older age, pancreatitis, gallstones or gallbladder inflammation, and a family history of pancreatic cancer were also tied to higher risk.
The authors cautioned that most of the studies looked backward, definitions varied from study to study, and results still differed considerably among them.
The National Cancer Institute lists other signs to bring to a doctor: jaundice, which can mean yellow skin or eyes, dark urine, light-colored stools and itching; pain in the upper or middle abdomen and back; weight loss with no known cause; and loss of appetite. It adds that less serious conditions can cause these symptoms too.
What Does The Evidence Actually Support?
Not routine screening for adults at average risk who have no symptoms.
In 2019 the U.S. Preventive Services Task Force reaffirmed its recommendation against screening adults without symptoms for pancreatic cancer. That covers people whose extra risk comes from factors such as new-onset or existing diabetes, older age, smoking, obesity or chronic pancreatitis. It does not apply to people with certain inherited syndromes or a history of familial pancreatic cancer. The task force found no evidence that screening improves outcomes and judged the harms to be at least moderate. Common precursor growths in the pancreas often never progress, so finding them can lead to overtreatment, and pancreatic surgery carries real risks of complications and death.
The best-known blood marker doesn’t solve the problem either. According to Mayo Clinic, a CA 19-9 result can’t confirm pancreatic cancer on its own, and the test is not offered as general-population screening.
Checking out symptoms is different from screening everyone. The National Cancer Institute advises people who develop diabetes without known risk factors for it, such as family history, obesity or heart disease, to ask a doctor about checking for other signs of pancreatic cancer. For concerning symptoms, it says doctors typically begin with an abdominal ultrasound or CT scan. The blood sugar clue is real, but on its own it is a reason to ask a doctor, not a diagnosis.
Disclaimer: This article summarizes observational research, a pooled analysis of observational studies, and guidance from the U.S. Preventive Services Task Force and the National Cancer Institute. The two Mayo Clinic studies looked backward at people already diagnosed with pancreatic cancer and describe group averages, not individual blood sugar paths, so they cannot show that cancer is detectable when blood sugar first rises. The prospective study measured how often pancreatic cancer followed newly detected diabetes but did not test whether screening those patients saves lives. None of the studies evaluated a blood sugar threshold as a diagnostic test. The key uncertainty is whether acting on early blood sugar changes would find cancers sooner without causing more harm than benefit. This article is for general information and is not medical advice; anyone with new diabetes, unexplained weight loss or other symptoms should talk with a clinician.
Paper Notes
Limitations
Sharma et al., Gastroenterology, 2018. The authors describe the work as a retrospective study based on medical record review. Its glycemic profiles describe the cohort as a whole, and the authors note that individual profiles vary. They also state that the hyperglycemia-based estimate of how long cancer is present before diagnosis applies only to tumors that cause high blood sugar. The main temporal cohort came from one Minnesota county, and the paper calls early detection at the onset of high blood sugar an unproven premise.
Pannala et al., The American Journal of Gastroenterology, 2009. This was a backward-looking, nested case-control analysis of Mayo Clinic records for cases diagnosed from 1981 through 2004, limited to people who had at least one outpatient fasting glucose test in the five years before diagnosis. Glucose values were grouped into 12-month windows, which shows average trends rather than individual trajectories. The full limitations section could not be opened (the PubMed Central page was blocked), so design-based limits are listed here rather than the authors’ own wording.
Chari et al., Gastroenterology, 2026 (online 2025). The paper is an interim analysis with a median follow-up of 2.3 years and 82 cancers; the authors note that longer-term risk needs further study. Three-year incidence came from Kaplan-Meier estimates, and some racial and ethnic subgroups had wide confidence intervals. New-onset diabetes was defined from lab values flagged in electronic health records. The full text could not be opened, so the authors’ complete limitations section was not reviewed.
Wu et al., JAMA Network Open, 2020. The authors list these limits: only HbA1c was analyzed, not other glucose measures; antidiabetic medication use was not examined; selection bias is possible because clinicians chose whom to test; cases identified from death files lacked tissue confirmation, so some non-ductal cancers may be included; and a precise number needed to screen could not be estimated because the benefit of early detection strategies is unknown. The cohort came from one California health system.
Mellenthin et al., Cancers, 2022. The authors note that most included studies were retrospective, registry-based studies risk misclassification, definitions of risk factors varied and influenced results, and heterogeneity remained considerable for several factors (for example, I² of 85.5 percent for pancreatitis and 87.0 percent for gallstones or cholecystitis; glycemia was 6.7 percent). The paper’s weight-loss estimate differs between the abstract (2.49) and the results section (4), so the article does not cite a weight-loss number.
U.S. Preventive Services Task Force, 2019. This is a reaffirmation of a 2004 recommendation based on a targeted evidence review. It found no evidence that screening improves outcomes; much of the screening-harms evidence came from high-risk populations. It does not apply to people with certain inherited syndromes or familial pancreatic cancer.
Funding and Disclosures
Sharma et al., 2018: Supported by the NIH U01 consortium on chronic pancreatitis, diabetes and pancreatic cancer (CPDPC), the Pancreatic Cancer Action Network, the Kenner Family Research Fund and the Prokopanko Gift to Mayo Foundation. The authors disclosed no conflicts. (Checked in the article PDF.)
Pannala et al., 2009: Not verified. The PubMed and PubMed Central pages could not be opened from this workspace; the paper is an NIH author manuscript, which indicates NIH support, but specific grants and disclosures were not confirmed.
Chari et al., 2026: Not verified. The PubMed Central full text could not be opened. The paper is listed by the NCI Early Detection Research Network and cited in the ClinicalTrials.gov record for the related Early Detection Initiative (NCT04662879), whose officials include investigators from MD Anderson and Kaiser Permanente, but the paper’s own funding and disclosure statements were not read directly.
Wu et al., 2020: Funded by National Cancer Institute grant 5U01CA200468-04; the funder had no role in the study. No conflicts of interest were reported. (Checked in the PubMed Central full text.)
Mellenthin et al., 2022: Supported by HFR Research Grant-2108; the funders had no role in the study. The authors declared no conflicts of interest. (Checked on the MDPI article page.)
USPSTF, 2019: Task force members reported no disclosures related to the statement, according to MDedge coverage of the JAMA publication. The statement itself was not checked for a funding line.
Publication Details
Sharma A, Smyrk TC, Levy MJ, Topazian MA, Chari ST. “Fasting Blood Glucose Levels Provide Estimate of Duration and Progression of Pancreatic Cancer Before Diagnosis.” Gastroenterology. 2018;155(2):490–500.e2. Published online April 30, 2018. DOI: 10.1053/j.gastro.2018.04.025. Affiliations: Division of Gastroenterology and Hepatology (Sharma, Levy, Topazian, Chari) and Department of Anatomic Pathology (Smyrk), Mayo Clinic, Rochester, Minnesota.
Pannala R, Leibson CL, Rabe KG, Timmons LJ, Ransom J, de Andrade M, Petersen GM, Chari ST. “Temporal Association of Changes in Fasting Blood Glucose and Body Mass Index With Diagnosis of Pancreatic Cancer.” The American Journal of Gastroenterology. 2009;104(9):2318–2325. Published online June 9, 2009. DOI: 10.1038/ajg.2009.253. Affiliations: Mayo Clinic, Rochester, Minnesota (department-level affiliations not verified).
Chari ST, Wu B, Lopez C, Lustigova E, Chen Q, Van Den Eeden SK, Leimpeter AD, Fisher W, Wood A, Alexander A, and colleagues (23 authors). “Risk of Pancreatic Cancer in Glycemically Defined New-Onset Diabetes: A Prospective Cohort Study.” Gastroenterology. 2026;170(1):106–117. Published online July 7, 2025. DOI: 10.1053/j.gastro.2025.06.025. Institutions include The University of Texas MD Anderson Cancer Center, Kaiser Permanente Southern California, Kaiser Permanente Northern California, Fred Hutchinson Cancer Center, Mayo Clinic, Baylor College of Medicine, Ohio State University Wexner Medical Center, Kelsey Research Foundation and the National Cancer Institute (per Nature Index listing).
Wu BU, Butler RK, Lustigova E, Lawrence JM, Chen W. “Association of Glycated Hemoglobin Levels With Risk of Pancreatic Cancer.” JAMA Network Open. 2020;3(6):e204945. Published June 12, 2020. DOI: 10.1001/jamanetworkopen.2020.4945. Affiliations: Center for Pancreatic Care, Kaiser Permanente Los Angeles Medical Center; Department of Research and Evaluation, Kaiser Permanente Southern California, Pasadena.
Mellenthin C, Balaban VD, Dugic A, Cullati S. “Risk Factors for Pancreatic Cancer in Patients With New-Onset Diabetes: A Systematic Review and Meta-Analysis.” Cancers. 2022;14(19):4684. Published September 26, 2022. DOI: 10.3390/cancers14194684. Affiliations include Hospital Fribourg and the University of Fribourg, Switzerland; Carol Davila University of Medicine and Pharmacy, Bucharest, Romania; and the University of Geneva, Switzerland (full list not verified).
Chari ST, Leibson CL, Rabe KG, Timmons LJ, Ransom J, de Andrade M, Petersen GM. “Pancreatic Cancer-Associated Diabetes Mellitus: Prevalence and Temporal Association With Diagnosis of Cancer.” Gastroenterology. 2008;134(1):95–101. DOI: 10.1053/j.gastro.2007.10.040. Cited for background on long-standing vs. new-onset diabetes.
US Preventive Services Task Force (Owens DK, et al.). “Screening for Pancreatic Cancer: US Preventive Services Task Force Reaffirmation Recommendation Statement.” JAMA. 2019;322(5):438–444. Published August 6, 2019. DOI: 10.1001/jama.2019.10232.
Non-academic sources: National Cancer Institute, “Pancreatic Cancer Symptoms” (cancer.gov); Mayo Clinic, “Pancreatic cancer: Diagnosis and treatment” (mayoclinic.org); MD Anderson Cancer Center research newsroom, “New-onset diabetes may serve as an early marker of pancreatic cancer,” July 2025 (mdanderson.org).
