The safety of incretin based drugsBMJ 2014; 348 doi: http://dx.doi.org/10.1136/bmj.g2779 (Published 24 April 2014) Cite this as: BMJ 2014;348:g2779
- Victor M Montori, professor of medicine
- 1Division of Endocrinology, Diabetes, Metabolism and Nutrition, Mayo Clinic, Rochester, Minnesota 55905, USA
- Correspondence to:
Important concerns have been raised about the impact of incretin based drugs on the risk of acute pancreatitis in patients with type 2 diabetes. Two linked papers by Li and colleagues (doi:10.1136/bmj.g2366) and Faillie and colleagues (doi:10.1136/bmj.g2780) could help us decide whether patients and clinicians should consider these concerns when choosing an antihyperglycemic drug.1 2
Like all other antihyperglycemic drugs, incretin based drugs—glucagon-like peptide-1 (GLP-1) receptor agonists and dipeptidyl peptidase-4 (DPP-4) inhibitors or gliptins—despite their elegant mechanism, have no special ability to reduce the risk of complications of diabetes. Patients might prefer to use these drugs though as they have almost no risk of hypoglycemia, are weight neutral (gliptins) or lead to weight loss (GLP-1 agonists), and have convenient oral (gliptins) or injectable (GLP-1 agonists) dosing. Because of these favorable features, experts recommend them as second line drugs after metformin to help patients achieve glycemic control.3 Some patients might also need to consider cost, which is about 70 times higher than metformin or sulfonylureas in the United States. Should the possibility of acute pancreatitis with these drugs also affect their decision making?
A safety signal suggesting that incretin based drugs could increase the risk of acute pancreatitis, emerged as soon as these treatments entered the market but reached broad notoriety in 2013.4 Since then, the US Food and Drug Administration and the European Medicines Agency have reviewed and conducted toxicological studies in animals and reviewed existing trials and observational studies, finding reassuring but insufficient evidence to reach a conclusion about a causal link.5 Both agencies have decided to keep these drugs available with the potential for pancreatitis described in their label.
The first linked paper is a rigorous and comprehensive systematic review and meta-analysis of randomized and observational studies.1 Most of the summarized evidence comes from 55 randomized trials funded by industry that were at low to moderate risk of bias; none was specifically designed to ascertain and adjudicate episodes of acute pancreatitis. The extent to which patients with known risk factors for acute pancreatitis (such as old age, obesity, duration of diabetes, prior episode of acute pancreatitis, gallstones, alcohol abuse) participated in these trials is unclear. Exclusion of high risk participants and the brevity of the trials can produce an unbiased yet narrowly applicable estimate of effect pertinent to low risk patients taking these drugs for less than six months. The pooled result suggests no significant increase in the risk of acute pancreatitis with incretin based drugs.
The systematic review also included five observational studies that, as expected, included more episodes of acute pancreatitis than the collection of randomized trials and a larger and perhaps more representative population of patients, maybe enriched with patients of higher socioeconomic status given the out of pocket cost of these drugs in some markets. These studies were judged to be at moderate to high risk of bias, had limited ascertainment of exposure and outcomes, and had limited adjustment for potential confounders. Also, the risk of publication bias can be substantial in observational studies, although I suspect that here it could bias against the drugs. Their results were inconsistent, with most suggesting no significant increase in the risk of acute pancreatitis.
In the other linked paper, Faillie and colleagues report on a large population based cohort drawn from the UK Clinical Practice Research Datalink practice database and the Hospital Episodes Statistics database; its rigor warrants moderate confidence in its results.2 This study usefully reports on the risk of acute pancreatitis in patients with type 2 diabetes not treated with incretin based drugs: 15 in 10 000 patients per year. This study also reports no significant association between these drugs and acute pancreatitis (hazard ratio of 1.00, 95% confidence interval 0.59 to 1.70). This result, which is consistent with the result of the meta-analysis of randomized trials (odds ratio 1.11, 95% confidence interval 0.57 to 2.17), was robust across several analytical assumptions when incretin based drugs were compared with sulfonylureas, metformin, or insulin, and after adjustment for previous sulfonylurea exposure. Neither here nor in previous studies is there evidence of a relation between the risk of acute pancreatitis and dose or duration of exposure to incretin based drugs or of differential effect of GLP-1 agonists or DPP4 inhibitors on acute pancreatitis.
The available evidence suggests, with low to moderate confidence, that there is no significant increase in the risk of acute pancreatitis with these drugs. There is no question that acute pancreatitis can be severe and even fatal, but it is rare, affecting about 15 out of 10 000 patients with type 2 diabetes each year. If these drugs were truly linked to an increase in this risk, the best guess based on this evidence is that the risk would increase to 16 in 10 000 (95% confidence interval 7 to 23 in 10 000) per year. And theoretically these magnitudes would be lower if incretin based drugs were used only in low risk patients. Ongoing cardiovascular safety trials and new “big data” observational analyses might soon provide additional evidence.
This evidence should be shared with patients in the context of a shared decision making process,6 with selection of drug treatment reflecting their informed preference rather than a prespecified “next step” after metformin. In essence, patients will be asked to consider the desirability of other drugs against the possibility that incretin based drugs could increase the risk of acute pancreatitis. Patients and their clinicians will uncover the final word on the safety of incretin based drugs for each patient after a careful evidence based discussion, in part supported by the research The BMJ publishes today. Given the state of this evidence, however, they would be wise to schedule that decision for review.
Cite this as: BMJ 2014;348:g2779
Competing interests: I have read and understood the BMJ Group policy on declaration of interests and declare the following interests: None.
Provenance and peer review: Commissioned, not externally peer reviewed.