Editorials
Pharmacology
Volume 195 - Issue 6

Safety of incretin-based therapies for type 2 diabetes

Author:  Timothy M E Davis

Med J Aust 2011; 195 (6): 312-313. || doi: 10.5694/mja11.10648
Published online: 19 September 2011
Download PDF (103 KB)
Share: Share on Blue Sky Share on Facebook Share on LinkedIn Share by email

Australian database linkages could be used for postmarketing surveillance of antidiabetic therapy side effects

Several incretin hormone-based therapies for type 2 diabetes are marketed in Australia. These are exenatide (the glucagon-like peptide-1 [GLP-1] analogue) and sitagliptin, vildagliptin and saxagliptin (inhibitors of the enzyme dipeptidyl-peptidase-4 [DPP-4]). These drugs are attractive because they improve blood glucose control without weight gain and with less hypoglycaemia than insulin and its secretagogues. Their distinct mode of action means that they improve glycaemic control when added to more established blood glucose-lowering therapies. This underlies their current Pharmaceutical Benefits Scheme listing as part of dual or, in the case of exenatide, triple therapy with metformin and/or sulfonylurea treatment.

With increasing use of these drugs, more data relating to their adverse effects have emerged. A case linking pancreatitis with exenatide therapy was described 5 years ago,1 and further case reports have meant that specific warnings are now included in the product information for all GLP-1 analogues and DPP-4 inhibitors. Animal studies with the GLP-1 analogue liraglutide (registered but not currently available in Australia) raised the possibility of medullary (C-cell) thyroid cancer. Although an increased risk of this tumour was not evident in humans in preregistration trials, liraglutide is contraindicated in patients with a personal or family history of medullary thyroid cancer and in patients with multiple endocrine neoplasia syndrome type 2 (http://www.novo-pi.com/victoza.pdf). Inhibition of DPP-4 may predispose to hypersensitivity reactions through prolonged action of neuropeptides such as substance P,2 and alter immune function, with a possible increased risk of infections.3 However, the relative infrequency of adverse effects such as pancreatitis, C-cell cancer and severe allergy means that large case–control databases are needed to provide accurate estimates of their incidence and predictors.

In an attempt to provide better information on incretin therapy-associated adverse effects, United States researchers led by Peter Butler recently used the US Food and Drug Administration (FDA) publicly available Adverse Event Reporting System (AERS). An initial accepted report containing the results of these analyses appeared in February 2011 on the website of the journal Gastroentrology. The main findings were of a significantly increased risk of pancreatitis and pancreatic cancer with exenatide and sitagliptin (the two most-used incretin-based therapies in the US), as well as an increased risk of all thyroid cancer (not just medullary) with exenatide. After the validity of these findings was challenged by the manufacturers,4 the paper was withdrawn but a modified version subsequently reappeared online and in print.5

The use of AERS to identify adverse drug effects is controversial and its substantial limitations are acknowledged by the FDA itself. Reporting is uncontrolled, voluntary, from multiple sources and often incomplete. Overreporting of events for new drugs, especially during the first 2 years and for serious outcomes, is well recognised.6 The choice of comparator treatments, from which odds ratios for the occurrence of index events are generated, has a bearing on the results, as does knowledge of other therapies or patient factors that might also increase the risk of an event. For pancreatitis, this includes the association between diabetes and obesity, as well as drugs that have relatively high use in type 2 diabetes, such as angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, statins and fibrates. Although mining of data from the AERS has successfully found previously unrecognised adverse effects, a study of 21 unrelated therapies suggests that only about 50% of the signals thus identified subsequently appear in the product information or lead to regulatory actions, while an equivalent number of clinically important signals are missed.7 In the case of incretin-based therapies, the FDA’s automated algorithm Empirica, which is designed to signal AERS safety alerts, did not do so. In addition, one unrelated retrospective analysis of a large US medical and pharmacy claims database, involving 786 656 patients, did not show either exenatide or sitagliptin to be associated with an increased pancreatitis risk.8

The Butler group state that their analyses do not establish that pancreatitis, pancreatic cancer and thyroid cancer are caused by incretin-based therapy but suggest that appropriate studies are required to rule out these associations.5 Accurate incidence of pancreatitis will be difficult to obtain without adequately powered long-term prospective studies, but at least the product information for these drugs warns of this infrequent potential adverse effect. For the possible association between cancer and incretin-based therapy (and since other diabetes treatments including glargine insulin9 and pioglitazone10 might be cancer-promoting), it seems sensible to urgently develop coordinated links between diabetes prescription databases and cancer registries. Available Australian databases and links mean that government agencies could take a lead in this and provide independent, postmarketing observational data for the two incretin-based classes and individual drugs that are more robust than those generated by systems such as AERS.

The history of diabetes treatment, from phenformin to rosiglitazone and even newer compounds, illustrates the continuing requirement for comprehensive preclinical and clinical safety and efficacy data before registration. Robust multifaceted postmarketing surveillance and data for important adverse effects, even if they are uncommon, are also necessary. Analyses such as those by the Butler group4 should not direct clinical decision making, but are part of the pharmacovigilance process that aims to provide regulatory authorities, clinicians and patients with the best evidence of the risks and benefits of individual blood glucose-lowering therapies.

Author

  • Timothy M E Davis MB BS, DPhil, FRACP1
  • 1. University of Western Australia, Fremantle, WA.

Correspondence: tdavis@cyllene.uwa.edu.au

Competing interests

I have served on advisory boards for Merck Sharp and Dohme (MSD), Novartis and Bristol-Myers Squibb (BMS), companies which market DPP-4 inhibitors in Australia, and Novo Nordisk, which manufactures liraglutide. I have had attendance costs for sacientific meetings covered by MSD and Novartis, and have received lecture fees from MSD, Novartis, BMS and Novo Nordisk.

Acknowledgements

I am supported by a Practitioner Fellowship from the National Health and Medical Research Council.

References

View this article on Wiley Online Library