AMSTERDAM – Incretin-based antidiabetic drugs do not appear to increase risk for pancreatic cancer, but the acute need for these drugs because of rapid worsening of type 2 diabetes may be a marker for early, occult pancreatic adenocarcinoma, investigators cautioned.
Results of a study of nearly 825,000 patients with type 2 diabetes in Belgium and northern Italy showed that patients who required a first-time prescription for an incretin-based antidiabetic drug had a 3.5-fold greater risk of being diagnosed with pancreatic cancer within 3 months, compared with patients who could be safely maintained on an oral noninsulin, nonincretin antidiabetic drug (NIAD), reported Alice Koechlin of the International Prevention Research Institute in Lyon, France, on behalf of coauthor Phillipe Autier, MD, also of the institute.
Alice Koechlin
“We think that, at the beginning, there is an asymptomatic pancreatic cancer, with no clinical findings, and its first health effects are to disturb glucose metabolism. Then patients are diagnosed with type 2 diabetes, they are prescribed antidiabetic drugs, and then, as the cancer progresses but is still asymptomatic, the diabetes is less well controlled, and the patients shift to incretins and insulin more rapidly. And when the symptoms [of cancer] appear, it is too late for treatment,” Ms. Koechlin said at an annual congress sponsored by the European Cancer Organisation.
Incretin hormones stimulate the release of insulin from the pancreas. Incretin-based therapies such as dipeptidyl peptidase-4 inhibitors (DPP4 inhibitors, or gliptins) and glucagon-like peptide-1 receptor agonists (GLP-1 RAs), are generally held in reserve for patients with type 2 diabetes who have poor or inadequate glycemic control on oral agents such as metformin or the sulfonylureas.
Data from laboratory studies have suggested that incretins in general, and GLP-1 RAs in particular may increase risk of pancreatic cancer because of their direct action on the gland, Ms. Koechlin said.
As a part of postmarketing studies of the GLP-1 RAs requested by the European Medicines Agency (EMA), the investigators drew data from the Belgian Cancer Register on 11 million people in Belgium and from a registry maintained by the University of Milano-Bicocca, which covers approximately 10 million people in the Lombardy region of Italy.
They identified patients with type 2 diabetes who received a first prescription of an incretin drug or NIAD from January 2008 through the end of 2013 in Belgium, and through the end of 2012 in Italy.
They found that on first blush, incretin use did indeed appear to be associated with risk of pancreatic cancer, compared with NIAD use. Hazard ratios for cancer with incretin were 2.12 in Belgium (95% confidence interval, 1.60-2.81) and 2.17 (95% CI, 1.50-3.13) in Lombardy. The combined HR was 2.14 (95% CI, 1.71-2.67).
“In Belgium, 25% of cases were diagnosed within 90 days, and in Lombardy it was 18%. After the first year, the proportion of diagnosed pancreatic cancers dropped dramatically,” Ms. Koechlin said.
When they looked at the risk of cancer from incretin use over time, however, they found that the risk was highest at 3.5-fold, compared with NIAD use, within 3 months of starting a first prescription, but diminished to 2.3-fold during months 3-6, 2-fold for months 6-12, and 1.7-fold after the first year.
“This is not compatible with a causal relationship, because if there was a causal relationship we would observe a small risk for shorter duration of use, and higher risk for higher duration of use,” she said.
To support their findings, they looked at the relationship between cancer risk over time and first prescriptions for NIADs, and saw a similar decrease in risk as the duration of therapy increased.
They then looked at the relationship between a first prescription for insulin during follow-up, and saw significant increases in cancer risk, compared with patients who did not require insulin, with an HR in Belgium of 6.61 (95% CI, 5.63-7.77), and in Lombardy of 7.46 (95% CI, 6.00-9.35).
The perceived association between incretin drugs and cancer risk, therefore, may be due to “protopathic” or “reverse causation” bias, Ms. Koechlin said.
“Should patients diagnosed with diabetes be screened for pancreatic cancer? For now, no such test exists, but we could imagine in the future something like this: If the patient is diagnosed with diabetes, and the diabetes degrades rapidly, there could be some tests to identify markers for pancreatic cancer, and the pancreatic cancer could then be diagnosed at early stage, where there is more chance of curing the disease,” she said.
The study results raise questions about how practitioners should discuss potential risks of pancreatic cancer in patients with type 2 diabetes, commented Ian Banks, MD, president of the European Men’s Health Forum and cochair of ECCO2017.
“What is the best way of getting this message out without causing unnecessary concern, while at the same time raising the level of suspicion to go and see their GP?” he asked Ms. Koechlin, who presented a summary of the data in a briefing.
“It’s quite a difficult question, actually, because as you said, it could cause unnecessary worries,” and there are “no easy answers,” she said.
The study was sponsored by Sanofi. The investigators and Dr. Banks reported no conflicts of interest.