New cases of diabetes were significantly more common among pancreatic cancer patients before their cancer diagnoses, compared with controls, according to data from 736 pancreatic cancer patients and 1,875 controls.
Although previous studies have shown a link between existing diabetes and pancreatic cancer, the temporal relationship between the two diseases—and whether this relationship might be used to predict cancer—is not well understood, wrote Dr. Suresh T. Chari and colleagues at the Mayo Clinic in Rochester, Minn.
To determine the prevalence of new-onset diabetes in pancreatic cancer patients and the temporal association between these conditions, the researchers reviewed records of pancreatic cancer patients and control patients seen at the Mayo Clinic between January 15, 1981, and July 9, 2004.
They assigned two matched controls to each cancer case. The mean age of both patients and controls was 69 years, and approximately 50% of the subjects in each group were men.
A subject was considered to have diabetes if he or she had a fasting blood glucose level greater than 126 mg/dL or was taking diabetes medication. The proportion of cases and controls with diabetes was compared in each 12-month interval, starting with 60 months prior to a cancer diagnosis for cancer patients (Gastroenterology 2008;134:95–101).
Overall, significantly more pancreatic cancer patients met the criteria for diabetes, compared with controls, any time during the 60-month period before pancreatic cancer diagnosis (40% vs. 19%). But the proportions of individuals with diabetes were not significantly different between the cancer group and the control group during the 12-month intervals from 60 months to 48 months and from 48 months to 36 months prior to cancer diagnosis.
In contrast, starting with 36 months before a cancer diagnosis, the prevalence of diabetes in the pancreatic cancer patients rose steadily for each 12-month interval, while the prevalence of diabetes in the controls remained relatively stable throughout the study period. New-onset diabetes was defined as diabetes with onset at 24 months or less prior to a cancer diagnosis.
Diabetes was more likely to be new onset in patients with pancreatic cancer than in controls (52.3% compared with 23.6%, respectively) among the subjects with diabetes for whom diabetes duration was known; this difference was highly significant.
“The very high prevalence of diabetes in pancreatic cancer and its close temporal association with the diagnosis of cancer provide strong epidemiologic evidence to support the notion that pancreatic cancer causes diabetes mellitus,” the researchers wrote.
The findings support data from small clinical studies in which the removal of tumors from pancreatic cancer patients with diabetes has improved their glucose tolerance and reversed their metabolic defects.
But prospective studies are needed to show the benefits of screening older adults with new-onset diabetes for pancreatic cancer, and such screening would be helpful only if a type of new-onset diabetes that is associated with pancreatic cancer could be distinguished from type 2 diabetes, perhaps with the use of a biomarker test, the researchers noted.
In an editorial that accompanied the article, Dr. Niels Teich, of the University of Leipzig in Germany wrote that there is a lack of practical criteria that could be used to rule out pancreatic cancer in new-onset diabetes patients (Gastroenterology 2008;134:344–5).
The study findings invite more research to determine whether new-onset diabetes in pancreatic cancer patients is different from new-onset type 2 diabetes mellitus in general, and whether new-onset diabetes could be an early sign of this cancer in otherwise asymptomatic persons, he noted.
“The data available today clearly suggest that diabetes mellitus can be both a long-standing cause of pancreatic cancer, and, as shown now, an early manifestation of the disease,” Dr. Teich said.
ELSEVIER GLOBAL MEDICAL NEWS