Beyond Blood Sugar: What Else May Link Obesity to Pancreatic Cancer?

Key Takeaways

A 2026 UK Biobank analysis of more than 462,000 adults found that glucose and HbA1c statistically mediated about 12% to 20% of the association between measures of obesity and pancreatic cancer risk, depending on the adiposity measure and glycemic marker examined.
These findings do not mean that blood sugar causes 12% to 20% of obesity-related pancreatic cancers, nor do they show that lowering glucose would eliminate that portion of risk.
Most of the statistical association was not mediated through glucose or HbA1c in these models. The study itself does not identify what accounts for that remainder.
Other candidate pathways include hyperinsulinemia and insulin/IGF signaling, chronic inflammation, adipose-derived signaling molecules, and fat accumulation in and around the pancreas. The strength of evidence differs considerably among these mechanisms.
The relationship between diabetes and pancreatic cancer is also complicated by reverse causation: in some adults, new-onset diabetes can be an early manifestation of an existing pancreatic cancer rather than simply a cause of future disease.
The most useful interpretation of the glucose finding is therefore a limited one: glycemic pathways appear to explain part, but not all, of the association between excess adiposity and pancreatic cancer.

Introduction

Obesity is associated with a higher risk of pancreatic cancer, but identifying a risk factor is not the same as understanding how that risk develops biologically. Excess adipose tissue changes glucose metabolism, insulin signaling, inflammatory activity, circulating hormones, lipid metabolism, and the tissue environment surrounding internal organs. Any attempt to explain the obesity–pancreatic cancer relationship with a single pathway is therefore likely to be incomplete.

A large UK Biobank study published in 2026 helped quantify one piece of that puzzle. Researchers investigated whether glucose and glycated hemoglobin (HbA1c) statistically mediate the association between several measures of body size and pancreatic cancer. Depending on the measure examined, these glycemic markers accounted for roughly 12% to 20% of the association (Amadou et al., 2026).

That finding is informative precisely because it is incomplete.

It suggests that altered glucose metabolism may be one pathway linking excess adiposity with pancreatic cancer, while most of the statistical association in these models was not mediated through glucose or HbA1c. What the study does not tell us is what biological mechanisms account for that remainder.

Answering that question requires looking beyond blood sugar.

What the UK Biobank Study Actually Found

Amadou and colleagues analyzed 462,300 adults aged 40 to 69 years enrolled in the UK Biobank. During a median follow-up of 10.9 years, 1,115 participants developed pancreatic cancer (Amadou et al., 2026).

The researchers examined several anthropometric measures, including body mass index (BMI), waist circumference, waist-to-hip ratio, and body-shape phenotypes created by combining multiple anthropometric traits. They then used a four-way mediation decomposition to investigate whether metabolic biomarkers helped explain the associations with pancreatic cancer.

Each one-standard-deviation increase in BMI was associated with a 20% higher incidence of pancreatic cancer (HR 1.20, 95% CI 1.12–1.28). Waist-to-hip ratio showed a similar association, with an HR of 1.24 (95% CI 1.14–1.36) per standard-deviation increase (Amadou et al., 2026).

For BMI, glucose accounted for an estimated 15.9% (95% CI 2.8–28.9%) of the association, while HbA1c accounted for 20.0% (95% CI 6.3–33.7%).

For a body-shape phenotype characterized by overall obesity, glucose accounted for 12.2% (95% CI 3.4–21.0%) and HbA1c for 15.0% (95% CI 5.7–24.2%) of the association (Amadou et al., 2026).

These are important estimates, but they require careful interpretation.

What Does “20% Mediated” Actually Mean?

A mediated proportion is a statistical estimate. It describes how much of an observed exposure–outcome association is compatible with an indirect pathway through a measured mediator under the assumptions of the mediation model.

It does not mean that 20% of pancreatic cancers among people with obesity are caused by high blood sugar.

It also does not establish that lowering HbA1c would remove 20% of pancreatic cancer risk associated with obesity.

Those conclusions would require stronger causal evidence, including evidence that deliberately modifying the mediator changes cancer incidence.

The UK Biobank analysis was observational, and the authors noted limitations that include the possibility of residual confounding and the challenges inherent in drawing causal conclusions from mediation analyses of observational data (Amadou et al., 2026).

The safest interpretation is narrower: glucose and HbA1c appear to statistically mediate a meaningful but minority portion of the association between excess adiposity and pancreatic cancer in this population.

Infographic showing how glucose and HbA1c may explain only part of the association between obesity and pancreatic cancer, with insulin, inflammation, adipokines, pancreatic fat, and the microbiome among other possible pathways.

Glucose and HbA1c statistically mediate only part of the association between obesity and pancreatic cancer. Researchers are investigating additional pathways involving insulin signaling, inflammation, adipose-derived factors, pancreatic fat, and the microbiome.

The Obesity–Pancreatic Cancer Link Predates This Study

The 2026 analysis did not establish obesity as a pancreatic cancer risk factor for the first time. Instead, it attempted to clarify mechanisms within an association already supported by epidemiologic and genetic evidence.

Mendelian randomization research has provided additional support for a potentially causal relationship. In an analysis using genetic data from 7,110 pancreatic cancer cases and 7,264 controls, genetically predicted higher BMI was associated with pancreatic cancer risk (Carreras-Torres et al., 2017).

That study is particularly relevant to the question raised by the newer mediation analysis because it also identified fasting insulin as a potential metabolic contributor.

The important scientific question has therefore shifted from simply asking whether excess adiposity is associated with pancreatic cancer toward understanding which biological consequences of excess adiposity contribute to that risk.

If Glucose Explains Only Part of the Association, What Could Explain the Rest?

There is no established percentage that can currently be assigned to each remaining pathway.

The residual association in the UK Biobank models should not be treated as a pie chart waiting to be divided neatly among insulin, inflammation, adipokines, pancreatic fat, and other mechanisms. These pathways interact, may share upstream causes, and may partly overlap with glucose metabolism itself.

Still, several mechanisms have stronger biological and epidemiologic support than others.

Hyperinsulinemia May Matter Independently of Glucose

One of the strongest candidates is insulin.

Obesity commonly contributes to insulin resistance. As tissues become less responsive to insulin, the pancreas may compensate by producing more of it, allowing glucose concentrations to remain relatively controlled for a period even while circulating insulin levels are elevated.

That distinction matters because glucose and insulin are related but are not interchangeable biological exposures.

Insulin and insulin-like growth factor pathways can activate intracellular signaling involved in cell proliferation and survival. These include PI3K/Akt and MAPK-related pathways that are relevant to carcinogenesis (Ruze et al., 2023).

Human genetic evidence also supports closer attention to insulin. Carreras-Torres and colleagues found that genetically predicted higher fasting insulin was associated with greater pancreatic cancer risk, with an odds ratio of 1.66 (95% CI 1.05–2.63) per standard-deviation increase in fasting insulin (Carreras-Torres et al., 2017).

That does not prove that hyperinsulinemia accounts for the portion left unexplained by glucose in the UK Biobank study. It does, however, make insulin signaling one of the more credible pathways to investigate separately.

Chronic Inflammation Provides Another Plausible Route

Adipose tissue is biologically active. As adipose tissue expands and becomes dysfunctional, its immune and metabolic behavior changes.

Obesity is associated with chronic low-grade inflammation and altered production of inflammatory mediators. Cytokines and other inflammatory signals can affect cellular proliferation, oxidative stress, immune responses, and the tissue microenvironment—all processes relevant to cancer development (Ruze et al., 2023).

Inflammation is also intertwined with insulin resistance and altered lipid metabolism, which makes it difficult to assign a completely independent fraction of pancreatic cancer risk to inflammatory pathways.

That overlap is one reason single-mediator models have limitations. Biological pathways do not necessarily behave as isolated statistical variables.

Adipokines May Carry Signals From Fat Tissue to the Pancreas

Adipose tissue also releases signaling molecules known as adipokines.

Among those studied in relation to obesity and pancreatic carcinogenesis are leptin, adiponectin, and fatty-acid-binding proteins such as FABP4. Obesity can alter the balance of these molecules, potentially influencing inflammation, insulin sensitivity, cell proliferation, and the tumor microenvironment (Ruze et al., 2023).

Experimental studies have identified mechanisms through which leptin and other adipose-derived signals could promote pancreatic tumor biology, while observational studies have examined associations involving circulating adipokines.

This is a biologically plausible field, but the evidence should not be presented as though researchers already know how much of obesity-associated pancreatic cancer risk is attributable to individual adipokines.

They do not.

At present, adipokines are better understood as components of a broader metabolic and inflammatory network than as quantified explanations for the residual association.

Fat Inside the Pancreas Is an Emerging Area of Interest

Another possibility is that where fat accumulates matters in addition to how much total body fat a person carries.

Intrapancreatic fat deposition, sometimes referred to as fatty pancreas, has attracted increasing attention because adipose tissue located within or around the pancreas could influence the organ’s local metabolic and inflammatory environment.

A 2023 systematic review and meta-analysis examined 2,956 participants and found that fatty pancreas was substantially more common among people with pancreatic cancer than among controls. However, the included evidence was primarily case-control and cross-sectional, making the direction of the relationship difficult to establish (Lipp et al., 2023).

This distinction is critical.

An association between pancreatic fat and pancreatic cancer does not prove that pancreatic fat caused the cancer. Cancer itself, metabolic changes preceding diagnosis, or shared risk factors could contribute to the observed relationship.

Prospective studies measuring pancreatic fat years before cancer develops are needed to determine whether local fat deposition is an independent causal contributor.

The Microbiome Is Interesting—but the Evidence Is Earlier

The gut and pancreatic microbiomes have become increasingly active areas of cancer research. Obesity can alter microbial communities and intestinal barrier function, potentially affecting systemic inflammation, metabolites, and immune signaling.

These mechanisms could theoretically influence pancreatic carcinogenesis, and microbiome-related pathways are discussed in contemporary reviews of obesity- and diabetes-associated pancreatic cancer biology (Ruze et al., 2023).

However, this area should be distinguished from better-established epidemiologic evidence involving adiposity and insulin.

Microbiome findings can vary substantially across populations, laboratory methods, diet, medication exposure, and disease stage. It is also difficult to determine whether microbial differences precede cancer or arise because of the disease and its metabolic consequences.

For now, the microbiome belongs in the category of promising mechanistic research, not a quantified explanation for the unexplained portion of obesity-associated pancreatic cancer risk.

Diabetes Complicates the Story Because the Relationship Runs Both Ways

Any discussion of glucose and pancreatic cancer also has to address reverse causation.

Long-standing metabolic dysfunction may contribute to pancreatic cancer risk, but pancreatic cancer itself can disrupt glucose regulation. New-onset diabetes can therefore sometimes be an early manifestation of an already developing pancreatic tumor.

In a study of adults older than 50 with new-onset diabetes, approximately 1% were diagnosed with pancreatic cancer within three years. Researchers developed a model incorporating age, changes in blood glucose, and changes in weight to identify individuals at higher risk (Sharma et al., 2018).

This does not mean that one in every 100 people with diabetes has pancreatic cancer, nor does new-onset diabetes by itself justify assuming that cancer is present. It illustrates why pancreatic cancer epidemiology involving glucose and diabetes is unusually difficult to interpret: metabolic abnormalities can potentially operate as risk factors, consequences of occult disease, or both.

That bidirectional relationship is another reason simplistic statements such as “high blood sugar causes pancreatic cancer” are inadequate.

Why the Unexplained Percentage Should Not Be Taken Literally

It is tempting to subtract 20% from 100% and conclude that science now needs to identify the biological cause of the remaining 80%.

That is not quite how mediation analysis works.

The residual portion can contain several things at once: genuine biological pathways not included in the model, imperfect measurement of the mediator, interactions among pathways, residual confounding, measurement error, and limitations imposed by the statistical model.

Glucose and HbA1c themselves are also related. Insulin resistance influences glucose, inflammation can influence insulin sensitivity, and adipose-derived signals can affect both.

Consequently, future studies cannot simply test each candidate mediator separately and expect the percentages to add neatly to 100%.

The biology is interconnected.

What Research Could Clarify the Remaining Pathways?

The next generation of studies will need to move beyond single measurements and isolated mediators.

Multi-mediator analyses could examine glucose, insulin, inflammatory markers, lipids, and adipokines within the same framework. Repeated blood measurements would help researchers distinguish long-term metabolic exposure from a single snapshot.

Genetic approaches can provide another layer of causal inference. The fasting-insulin findings from Mendelian randomization demonstrate how genetically informed analyses can help distinguish correlated metabolic traits, although Mendelian randomization has assumptions and limitations of its own (Carreras-Torres et al., 2017).

Imaging studies that quantify visceral and intrapancreatic fat before cancer develops could help determine whether local fat deposition predicts future disease independently of BMI and other metabolic factors (Lipp et al., 2023).

Ultimately, intervention research would provide stronger evidence. If sustained changes in weight or specific metabolic pathways were shown to reduce pancreatic cancer incidence, that would move the field beyond statistical mediation toward direct evidence of prevention.

Because pancreatic cancer is relatively uncommon, however, such studies require large populations and long follow-up.

What Does This Mean for Patients?

The study should not be interpreted as evidence that people can eliminate obesity-associated pancreatic cancer risk simply by keeping glucose or HbA1c within a particular range.

Nor does it mean that glucose control is unimportant.

Instead, the findings reinforce a broader view of metabolic health. Excess adiposity affects numerous interconnected systems, and pancreatic cancer risk is unlikely to depend on any one biomarker.

Maintaining a healthy weight, engaging in regular physical activity, avoiding tobacco, and managing metabolic conditions remain sensible components of overall health and cancer-risk reduction. The 2026 mediation study adds biological detail to that picture; it does not create a glucose-specific pancreatic cancer prevention prescription (Amadou et al., 2026).

People with diabetes should continue to manage their condition according to individualized medical advice rather than altering treatment because of pancreatic cancer research.

What About New-Onset Diabetes?

New-onset diabetes deserves a separate note because of its relationship with pancreatic cancer.

Most people who develop diabetes do not have pancreatic cancer. Even among adults over 50 with new-onset diabetes, pancreatic cancer remains uncommon (Sharma et al., 2018).

However, clinicians may pay closer attention when new metabolic abnormalities occur alongside features such as unexplained weight loss or other concerning symptoms, particularly in older adults.

The appropriate response is clinical assessment based on the individual’s overall risk and presentation—not self-screening or assuming that a new diabetes diagnosis indicates cancer.

Final Thoughts

The most useful finding from the UK Biobank analysis may be what it didn’t explain.

Glucose and HbA1c statistically mediated roughly 12% to 20% of the association between selected measures of obesity and pancreatic cancer. That supports a role for glycemic metabolism, but most of the association remained outside those pathways in the models (Amadou et al., 2026).

Current evidence points toward a network rather than a single missing mechanism. Hyperinsulinemia has meaningful epidemiologic and genetic support. Inflammation and adipose signaling have extensive mechanistic support. Intrapancreatic fat is increasingly interesting but still requires stronger prospective evidence. Microbiome pathways remain comparatively early in their development (Carreras-Torres et al., 2017; Lipp et al., 2023; Ruze et al., 2023).

None can currently be assigned the unexplained percentage left behind by glucose.

That distinction is what makes the research useful rather than disappointing. Science has begun to quantify one pathway. The next task is not to invent a single explanation for the remainder, but to determine how several interacting metabolic and tissue-level pathways contribute—and which of them can actually be modified to reduce risk.

Frequently Asked Questions

Understanding the Study

What does it mean that glucose or HbA1c “mediated” part of the obesity–pancreatic cancer association?

Mediation analysis asks whether part of the statistical relationship between an exposure and an outcome may operate through an intermediate factor. In this study, obesity measures were the exposures, pancreatic cancer was the outcome, and glucose or HbA1c was evaluated as a potential mediator.

For example, the estimate that HbA1c mediated 20% of the BMI association means that about one-fifth of that statistical association was compatible with an indirect pathway involving HbA1c under the assumptions of the model. It does not mean that HbA1c caused 20% of pancreatic cancer cases (Amadou et al., 2026).

What is the difference between glucose and HbA1c?

Blood glucose measures the concentration of glucose in the blood at a particular point in time. HbA1c reflects the proportion of hemoglobin with glucose attached to it and provides information about average blood glucose exposure over approximately the previous two to three months.

They are related measures, but they do not capture exactly the same aspect of glucose metabolism. This helps explain why their estimated mediation percentages were similar but not identical.

Why did the researchers examine BMI, waist circumference, and waist-to-hip ratio instead of BMI alone?

BMI estimates overall body size relative to height but does not describe where fat is distributed. Waist circumference and waist-to-hip ratio provide additional information about central adiposity, while the body-shape measures used in the study combined multiple anthropometric characteristics.

Examining several measures allows researchers to investigate whether different patterns of adiposity show similar relationships with pancreatic cancer rather than relying on BMI as the only indicator (Amadou et al., 2026).

What does a hazard ratio of 1.20 mean?

A hazard ratio of 1.20 means that, within the study and after accounting for variables included in the model, each one-standard-deviation increase in BMI was associated with an approximately 20% higher rate of developing pancreatic cancer during follow-up (Amadou et al., 2026).

It does not mean that an individual’s absolute risk rises by 20 percentage points. Relative risk and absolute risk are different concepts, and pancreatic cancer remains much less common than many other diseases.

What the Percentages Do and Do Not Mean

If HbA1c explains 20%, does that mean researchers have to explain the other 80%?

Not quite. It is tempting to treat the result like a pie chart, with HbA1c occupying 20% and other biological mechanisms filling the remaining 80%. Mediation estimates do not work that neatly.

The remaining statistical association may reflect multiple biological pathways, interactions among them, measurement limitations, residual confounding, and factors not included in the model. Candidate mechanisms such as insulin signaling and inflammation can also influence one another.

The remaining percentage therefore cannot simply be divided among insulin, inflammation, pancreatic fat, adipokines, and the microbiome.

Does the study prove that high blood sugar causes pancreatic cancer?

No. The study was observational. Its mediation analysis provides evidence consistent with glucose metabolism being involved in part of the association, but it cannot by itself establish that elevated glucose directly causes pancreatic cancer (Amadou et al., 2026).

Establishing causality requires evidence from multiple approaches, potentially including genetic studies, prospective biomarker research, mechanistic experiments, and interventions.

Does the study mean that lowering HbA1c would reduce obesity-related pancreatic cancer risk by 20%?

No. The 20% figure is a statistical mediation estimate, not an estimate of the effect of treatment.

To conclude that lowering HbA1c reduces pancreatic cancer risk by a particular percentage, researchers would need evidence showing that deliberately changing HbA1c produces that reduction. The UK Biobank analysis did not test such an intervention.

Why are the confidence intervals around the mediation estimates important?

A point estimate such as 20% is not perfectly precise. The confidence interval represents the range of values compatible with the data and statistical model at the specified confidence level.

For example, the estimated HbA1c-mediated proportion for BMI was 20.0%, with a 95% confidence interval of 6.3% to 33.7% (Amadou et al., 2026). That relatively broad interval is an important reminder that readers should not treat 20.0% as an exact biological constant.

Blood Sugar, Insulin, and Metabolism

If glucose explains only part of the association, could insulin explain more?

Possibly, but the current evidence does not establish what percentage insulin explains.

Obesity can produce insulin resistance and compensatory hyperinsulinemia even before glucose becomes substantially elevated. Insulin also has biological effects beyond regulating blood glucose, including effects on cellular growth signaling.

Genetic evidence has associated higher fasting insulin with pancreatic cancer risk, making insulin an important candidate for further study (Carreras-Torres et al., 2017). However, that does not demonstrate that insulin accounts for the portion left unexplained by glucose in the UK Biobank analysis.

Can someone have high insulin even if their blood sugar is normal?

Yes. During earlier stages of insulin resistance, the pancreas may compensate by producing more insulin. This additional insulin can help maintain glucose within or near the normal range for a period.

Consequently, glucose concentration alone does not capture every aspect of metabolic dysfunction. This is one reason researchers are interested in studying insulin separately from glucose.

Is HbA1c the same thing as insulin resistance?

No. HbA1c reflects average glycemic exposure over the preceding months. Insulin resistance describes reduced responsiveness of tissues to insulin.

A person can develop insulin resistance and compensatory elevations in insulin before glucose or HbA1c crosses a diagnostic threshold for diabetes.

Obesity and Pancreatic Cancer Risk

Does obesity cause pancreatic cancer?

It is more accurate to say that greater adiposity is an established risk factor for pancreatic cancer and that several lines of evidence support a potentially causal relationship. Genetic research using Mendelian randomization has also supported an association between genetically predicted higher BMI and pancreatic cancer risk (Carreras-Torres et al., 2017).

However, pancreatic cancer is multifactorial. Obesity does not mean that a person will develop pancreatic cancer, and pancreatic cancer occurs in people without obesity as well.

Does losing weight eliminate pancreatic cancer risk associated with obesity?

Current evidence does not allow researchers to assign a specific pancreatic cancer risk reduction to a given amount of intentional weight loss.

Weight management can benefit many aspects of metabolic and cardiovascular health, but the mediation study did not test whether losing weight prevents pancreatic cancer. Its results should therefore not be converted into a specific weight-loss prescription for cancer prevention.

Is abdominal fat more important than BMI?

Central and visceral adiposity may provide metabolic information that BMI alone does not capture. This is one reason studies investigate waist circumference and waist-to-hip ratio alongside BMI.

The UK Biobank analysis found associations across several measures of adiposity rather than identifying BMI as the only relevant measure (Amadou et al., 2026).

Diabetes and Pancreatic Cancer

Does diabetes increase the risk of pancreatic cancer?

Diabetes and pancreatic cancer have a complicated, bidirectional relationship. Long-standing metabolic dysfunction is associated with increased pancreatic cancer risk, but pancreatic cancer itself can also interfere with glucose regulation.

This means diabetes may sometimes be a risk marker and, in other circumstances, a consequence of an already developing pancreatic tumor.

Does new-onset diabetes mean someone has pancreatic cancer?

Usually not.

Pancreatic cancer remains uncommon among people newly diagnosed with diabetes. In research involving adults older than 50 with new-onset diabetes, approximately 1% developed pancreatic cancer within three years (Sharma et al., 2018).

New-onset diabetes therefore should not be interpreted as evidence that pancreatic cancer is present. Clinicians consider it alongside age, weight changes, symptoms, medical history, and other risk factors.

Should everyone with new-onset diabetes be screened for pancreatic cancer?

No universal pancreatic cancer screening program is currently recommended simply because someone develops diabetes. Researchers are investigating ways to identify the relatively small subgroup of people with new-onset diabetes whose risk is sufficiently elevated to justify further pancreatic evaluation.

Anyone concerned about unexplained symptoms or changes accompanying a new diabetes diagnosis should discuss them with a healthcare professional rather than attempting to interpret cancer risk from glucose results alone.

Other Possible Biological Pathways

How could inflammation contribute to pancreatic cancer?

Excess and dysfunctional adipose tissue can contribute to chronic low-grade inflammation. Inflammatory signaling can influence oxidative stress, immune function, cellular proliferation, insulin sensitivity, and the tissue environment in ways that may favor carcinogenesis (Ruze et al., 2023).

However, inflammation overlaps substantially with other metabolic pathways, making it difficult to isolate its independent contribution to pancreatic cancer risk.

What are adipokines?

Adipokines are signaling molecules produced or influenced by adipose tissue. Examples include leptin and adiponectin.

Obesity can alter adipokine concentrations and signaling, potentially affecting inflammation, insulin sensitivity, cellular growth, and the tumor microenvironment. These mechanisms are biologically plausible, but researchers cannot currently say what percentage of obesity-associated pancreatic cancer risk they explain (Ruze et al., 2023).

What is fatty pancreas?

Fatty pancreas generally refers to excess fat accumulation within the pancreas. Researchers are investigating whether this local fat deposition influences inflammation, metabolic signaling, or the pancreatic tissue environment.

A systematic review and meta-analysis found fatty pancreas to be more common among people with pancreatic cancer, but much of the available evidence was observational and cannot establish that pancreatic fat causes the disease (Lipp et al., 2023).

Does the gut microbiome explain the remaining association?

There is not enough evidence to say that.

Microbiome research is investigating whether changes in microorganisms, intestinal permeability, microbial metabolites, and immune signaling contribute to pancreatic carcinogenesis. These mechanisms are scientifically interesting but currently much less quantified than the established epidemiologic relationship between adiposity and pancreatic cancer (Ruze et al., 2023).

What Readers Should Take From the Research

Should people focus on blood sugar to prevent pancreatic cancer?

Blood glucose is only one component of metabolic health. The study does not support targeting glucose as a stand-alone pancreatic cancer prevention strategy.

Maintaining a healthy weight, staying physically active, avoiding tobacco, and appropriately managing diabetes and other metabolic conditions remain more defensible general health strategies than focusing on a single laboratory value.

Should someone change diabetes medication because of this study?

No. This research does not establish that changing a particular diabetes medication prevents pancreatic cancer.

Diabetes medications should be selected based on an individual’s medical circumstances and discussed with their treating clinician. Patients should not stop, start, or change medication because of the mediation percentages reported in this study.

What is the most important unanswered question?

The major question is no longer simply whether obesity is associated with pancreatic cancer. Researchers increasingly need to determine which consequences of excess adiposity are causally responsible, how those pathways interact, and whether modifying them actually reduces cancer risk.

The 2026 study helps by quantifying part of one pathway. It does not provide a complete explanation—and that limitation is an important part of the finding itself.

References

Amadou, A., Freisling, H., Mercoeur, B., Bohmann, P., Stein, M. J., Noh, H., Gebremariam, A., Sedlmeier, A. M., Peruchet-Noray, L., Gan, Q., Leitzmann, M. F., Baurecht, H., & Fervers, B. (2026). Anthropometric traits, metabolic biomarkers, and pancreatic cancer risk: A causal mediation analysis in UK Biobank. British Journal of Cancer. https://doi.org/10.1038/s41416-026-03524-9

Carreras-Torres, R., Johansson, M., Gaborieau, V., Haycock, P. C., Wade, K. H., Relton, C. L., Martin, R. M., Davey Smith, G., Brennan, P., & Gunter, M. J. (2017). The role of obesity, type 2 diabetes, and metabolic factors in pancreatic cancer: A Mendelian randomization study. Journal of the National Cancer Institute, 109(9), djx012. https://doi.org/10.1093/jnci/djx012

Lipp, M., Tarján, D., Lee, J., Zolcsák, Á., Szalai, E., Teutsch, B., Faluhelyi, N., Erőss, B., Hegyi, P., & Mikó, A. (2023). Fatty pancreas is a risk factor for pancreatic cancer: A systematic review and meta-analysis of 2956 patients. Cancers, 15(19), 4876. https://doi.org/10.3390/cancers15194876

Ruze, R., Song, J., Yin, X., Chen, Y., Xu, R., Wang, C., & Zhao, Y. (2023). Mechanisms of obesity- and diabetes mellitus-related pancreatic carcinogenesis: A comprehensive and systematic review. Signal Transduction and Targeted Therapy, 8, 139. https://doi.org/10.1038/s41392-023-01376-w

Sharma, A., Kandlakunta, H., Nagpal, S. J. S., Feng, Z., Hoos, W., Petersen, G. M., & Chari, S. T. (2018). Model to determine risk of pancreatic cancer in patients with new-onset diabetes. Gastroenterology, 155(3), 730–739.e3. https://doi.org/10.1053/j.gastro.2018.05.023