Intended for healthcare professionals


Pressure mounts for companies to hand over data on antidiabetes drugs linked to pancreatic harm

BMJ 2013; 346 doi: (Published 17 June 2013) Cite this as: BMJ 2013;346:f3900
  1. Deborah Cohen
  1. 1BMJ

Pressure on manufacturers to release the data they hold on glucagon-like peptide-1 (GLP-1) based antidiabetes drugs is mounting after a joint BMJ and Channel 4 Dispatches investigation of unpublished data on the potential pancreatic harms in animal and human studies.1

The American Diabetes Association has asked all companies involved in the development or marketing of such drugs to make patient level data available for an independent review that could help settle the question of whether the drugs contribute to the development of pancreatitis or pancreatic cancer.

The association’s chief scientific and medical officer, Robert Ratner, said, “People who are taking these medications or who may consider taking them should have the benefit of all that is currently known about their risks and advantages in order to make the best possible decisions about their treatment and care in consultation with their healthcare providers.”

In the United Kingdom the charity Pancreatic Cancer UK issued a statement saying, “It is important to investigate any treatment where there are concerns about an increased risk of pancreatic cancer.” The charity added, “This case certainly highlights the need to make the full trial data publicly available.”

The charity also urged European regulators to prioritise their current investigation of the drugs, to determine the need for possible regulatory action. The US Food and Drug Administration and the European Medicines Agency are each reviewing data from a small study of pancreases from deceased organ donors.2 These researchers suggested that the drugs caused “marked” cell proliferation and damage, with potential for the cells’ eventual transformation into cancer.

In the course of their investigations the FDA and EMA told the BMJ that they had disproportionate numbers of reports of pancreatic cancer associated with the drugs, but they said that this did not mean the association was causal.

At a meeting at the US National Institutes of Health on 13 June the FDA said that data mining its database for adverse events would be unhelpful and that more data were needed to help resolve the controversy. The agency called for more “adequately powered, long term epidemiological studies.” Currently, the only ongoing clinical trials are cardiovascular outcome trials, and it is unlikely that these will be used to look for pancreatitis and pancreatic cancer.

At the meeting Novo Nordisk presented the unpublished results of an insurance claims database study showing no increased risk of pancreatitis for its GLP-1 based drug, liraglutide; meanwhile Merck criticised insurance claims studies as being uncontrolled when the meeting discussed a claims study showing a higher risk of pancreatitis from its drug, sitagliptin.

Writing in the BMJ, Victor Montori, professor of medicine at the Mayo Clinic, Rochester, Minnesota, said, “There is a compelling biological rationale linking GLP-1 based treatments with pancreatitis and pancreatic cancer.”3

Some of this rationale hinges on studies identifying the presence of the GLP-1 receptor on precancerous pancreatic tissue and cells lining pancreatic ducts. Researchers say that the drugs act on the underlying pathological lesions and have a proliferative effect—in the same way that oestrogen speeds up conversion to cancer in receptor positive breast lesions.

But this theory is now being fiercely contested. Speaking at the National Institutes of Health meeting, Alan Moses, senior vice president and global chief medical officer of Novo Nordisk, said that the company had developed a much more specific method for detecting the GLP-1 receptor. He presented new company data that he said showed that non-human primates did not express the GLP-1 receptor in pancreatic duct cells, and neither did pancreatic adenocarcinoma. The company also presented its negative data from animal studies.

But Edwin Gale, emeritus professor of endocrinology at Bristol University, is sceptical about Novo Nordisk’s findings. He said, “Their big problem is that the exocrine pancreas undoubtedly proliferates in response to GLP-1 in many situations, and it’s hard to imagine how that could happen in the absence of any receptors.”

As revealed in the BMJ, it is not just human data but safety data from monkey studies that had not been published. Gale said that calling for the publication of individual patient data alone was not enough.

“The remarkable thing about the ADA [American Diabetes Association] statement is that it casts doubt on the credibility of the data presentation and its interpretation by the companies. I don’t believe they [the association] have ever taken this step before. It represents a major erosion of public confidence.

“The statement actually asks for both clinical and pathological data at one point, although the emphasis is upon patient data. Since the most important aspect of any independent review is that it should be based upon the totality of the data, it’s absolutely essential that all the clinical and preclinical material should be made freely available,” he told the BMJ.

Indeed, the BMJ and Channel 4 Dispatches investigation found that the results of preclinical monkey studies on Amylin’s exenatide—the first GLP-1 agonist to gain market approval—are being contested in a lawsuit in California. Lawyers acting for the plaintiffs asked to see pancreas histology slides from the study, but the manufacturer refused.

However, Clive Taylor, a pathologist from the University of Southern California, was granted limited access to review them, and he saw changes in the treated monkeys that he did not see in the controls. “These changes are associated with pancreatitis and even, perhaps, with pancreatic [neoplasia], pancreatic tumours,” he said.

The BMJ found that these changes were not included in the pathology reports for market access—but the company disagrees with Taylor’s interpretation. The BMJ also found that the regulators have not independently scrutinised the pathology slides but have relied on reports submitted by the company.

And although the FDA has admitted that Taylor and the company pathologists were seeing the same histology slides but were attributing different interpretations, the agency has not decided whether an independent review would help.

Gale said that this was one of the most shocking facts thrown up by the BMJ investigation. “A respected pancreatic pathologist saw what he believed to be premalignant changes in monkey pancreas and wanted others to double check. The company has known this for two years and has apparently done nothing about it. Now we learn that the FDA also seems prepared to do nothing. [It’s] unbelievable,” he said.


Cite this as: BMJ 2013;346:f3900


View Abstract

Log in

Log in through your institution


* For online subscription