Incretin therapy: should adverse consequences have been anticipated?

In a linked investigation (doi:10.1136/bmj.f3680), Cohen shines a spotlight on the safety of the incretins—drugs for the treatment of diabetes that mimic or enhance the biological effects of glucagon-like peptide-1 (GLP-1).1 Her investigation raises important questions as to the nature of the dialogue between drug companies and the regulators, and the extent to which potentially harmful effects of these drugs have been hidden from prescribers and patients.

The drugs that have transformed the modern treatment of chronic illness—β blockers, angiotensin converting enzyme inhibitors, and statins—are typically inhibitors of defined molecular pathways, with consequences that are relatively predictable. By contrast, the incretins are agonists that act on multiple targets with multiple effects. They are “magic shotguns” rather than “magic bullets.”

The thiazolidinediones—nuclear receptor agonists that modulate the activity of numerous genes—are good examples of this. They were introduced with great fanfare, but unwanted effects such as weight gain, fluid retention, and osteopenia have limited their use in the management of diabetes. Troglitazone, the first in class, was withdrawn because of hepatic injury; rosiglitazone was withdrawn from most countries in 2011 because of cardiovascular problems; and pioglitazone has been implicated in bladder cancer.

In each case the potential problem was spotted early in development but the regulatory response was disconcertingly slow. Troglitazone eventually came off the market because of the “termites”—Food and Drug Administration officials who alerted members of Congress to the problem (and were disciplined for doing so)—and because of the work of a Pulitzer prize winning journalist.2 Rosiglitazone was brought down by a cardiologist who published an independent analysis of clinical trials that the drug’s manufacturer was fortuitously obliged to make public in the wake of an unrelated misdemeanour.3 If the clinical trials had not been made public, would we still be using this drug?

GLP-1 is a short acting gut peptide that interacts with receptors in tissues including the brain, cardiovascular system, renal tubules, thyroid, and pancreatic exocrine and islet cells. GLP-1 deficiency does not seem to be an intrinsic feature of type 2 diabetes,4 and the therapeutic actions of the incretins are achieved at pharmacological doses that are much higher and more prolonged than in the physiological situation. The long term consequences of this exposure are unknown. GLP-1 has both neurocrine and endocrine effects and is also an enterogastrone (an agent that affects stomach motility). Another important effect is proliferation of cell growth.

The first GLP-1 agonist to reach the market was exenatide, first identified as exendin-4 in the venom of the Gila monster, Heloderma suspectum, a North American species of poisonous lizard. Few have paused to wonder why this predator should produce a non-toxic peptide in its saliva. The answer is that H suspectum is a desert lizard that goes for weeks or months between meals and conserves energy during the intervals by involution of its digestive apparatus, including its intestinal epithelium and exocrine pancreas. Production of exendin-4, a human GLP-1 agonist, causes rapid proliferation of intestinal tissue and a 50% increase in the size of the pancreas when it feeds.5

The growth stimulating effects of the incretins have long been known, and it was initially hoped that they would stimulate regeneration of pancreatic β cells and reverse the progression of diabetes. However, duct cells in the mammalian exocrine pancreas also carry the GLP-1 receptor and proliferate in response to GLP-1 receptor stimulation. Pancreatic enlargement has been noted in several species, and ductal hyperplasia offers a plausible mechanism for the occurrence of acute pancreatitis, an increasingly undeniable class effect of the incretins.6 7

Postmortem studies in people who have been taking incretins confirm that the use of exenatide or sitagliptin is associated with pancreatic enlargement and morphological change.8 Regulatory documents unearthed by Cohen’s investigation provide further evidence that the incretins produce greater subclinical fluctuations in pancreatic enzymes than other treatments for diabetes. Increased concentrations of pancreatic enzymes do not prove subclinical pancreatic inflammation but are consistent with it, and silent pancreatic inflammation is associated with the development of pancreatic cancer.9 Signals that the use of the incretins may be associated with pancreatic and thyroid tumours are now clearly present in regulatory databases.10

Marked α cell and β cell hyperplasia, as well as exocrine expansion, has been seen in pancreatic tissue obtained at autopsy from people taking incretins. Hyperplasia of α cells was associated with microadenomas in three of eight pancreases examined, one of which also contained a neuroendocrine tumour. Remarkably similar findings have emerged from animal models of glucagon inhibition11 and may limit the use of this approach to treat diabetes.12

It has taken eight years from the introduction of exenatide for all this to come to light. What went wrong? It is always easy to blame the regulators—they can’t answer back. I was an adviser to the European Medicines Agency for several years and can testify to the high quality, motivation, and training of the people who work there. Nor is big pharma the evil empire. The problem lies in a system that subordinates the public interest to commercial secrecy and allows the perceived need for such secrecy to define the legal, administrative, and cultural limits of the interaction between the mirror bureaucracies of the regulators and companies. Regulatory documents released to the BMJ under the Freedom of Information Act make it abundantly clear that the European Medicines Agency raised the right questions at an early stage, but the agency took each one off the agenda when plausible responses were supplied by the applicants. Each concern was treated in isolation rather than as another clue to an emerging pattern of biological effect.

The fate of the incretins has yet to be determined, but it has once again been shown that current regulatory procedures are inadequate to deal with the challenges presented by drugs that act on many targets. Similar scenarios will play out again while secrecy rules and the companies control access to the data. Safety requires more than tidy paperwork or the performance of yet more clinical studies. Cohen’s investigation has shown us—if further proof is needed—how much we need transparency.