Letters Incretins and pancreatic cancer

Cardiovascular benefits of incretins

BMJ 2013; 347 doi: https://doi.org/10.1136/bmj.f4382 (Published 10 July 2013) Cite this as: BMJ 2013;347:f4382
  1. Anthony H Barnett, emeritus professor of medicine and consultant physician1,
  2. Paul O’Hare, reader in medicine and honorary consultant physician2
  1. 1Heart of England NHS Foundation Trust and University of Birmingham, Diabetes Centre, Birmingham Heartlands Hospital, Birmingham B9 5SS, UK
  2. 2Warwick Medical School, University of Warwick, Coventry, UK
  1. anthony.barnett{at}heartofengland.nhs.uk

Cohen and colleagues make no mention of the evidence that treatment of type 2 diabetes with increasingly larger doses of sulfonylureas and insulin is not without serious risk from hypoglycaemia, weight gain, and possibly increased cardiovascular risk.1 A balanced account of this is needed in any review of incretins.

There have been concerns that such treatments, or the hypoglycaemia and weight gain they produce, explain failure to reduce cardiovascular disease despite tightening glycaemic control. Large meta-analyses of sulfonylureas suggest that they may increase cardiovascular risk.2 3 By contrast, a large meta-analysis of dipeptidyl peptidase 4 inhibitors suggests significant cardiovascular benefit.4 Although there is no definitive answer regarding cardiovascular outcomes, a balanced review would have considered the potential benefits of newer agents versus older established treatments.

Cohen also ignored the cohort study of 72 723 patients, which reported no increase in cardiovascular admissions or mortality and no increase in pancreatitis with sitagliptin.5 Instead, she referred only to the less well controlled cohort using patients only up to the age of 65, which reported an increase in hospital admissions from pancreatitis, but did not comment on cardiovascular events.6

Retrospective cohort studies are limited by confounding and bias. Cohen did not look at the definitive Cochrane meta-analysis of 28 randomised controlled trials from 170 articles. It analysed data from 10 910 patients treated with glucagon-like peptide-1 analogues and found only two cases of pancreatitis.7 Even allowing for some under-reporting, clinical pancreatitis appears rare and balanced by the benefits in glycaemic control without hypoglycaemia and with weight loss.

There are no data on increased risk of pancreatic cancer in humans. Experts and the regulators have reviewed the animal data; the science and relevance to humans are disputed and based on interpretations of dysplasia. The hypothesis of a “subclinical pancreatitis” in humans is not backed up by credible evidence and needs more balanced scrutiny.

Histological analysis was done on postmortem samples from eight patients only.8 The findings and interpretation need confirmation in greater numbers, with independent scrutiny, before they can be taken as evidence of potential harm. The serious concerns about data interpretation in this study are well reviewed.9

Once allegations of a potential for cancer are raised for any drug class, they can be refuted only by long term follow-up. Cardiovascular safety studies with these drugs may, when pooled, help do this. For now, if the hypothesis is that pancreatitis promotes carcinogenic changes, the clinical safety data reporting that pancreatitis is very rare is reassuring.

All treatments need to be kept under review to balance benefit against harm, but this must be done in a balanced way. We believe that the evidence presented by Cohen and colleagues lacks this balance. We wonder why, given that the safety issues are being investigated by the European Medicines Agency and US Food and Drug Administration, this is not acceptable to your journal.

Notes

Cite this as: BMJ 2013;346:f4382

Footnotes

  • Competing interests: Honorariums received for lectures and advisory work from MSD, Novartis, Boehringer-Ingelheim, BMS/Astra-Zeneca, Takeda, Roche, Sanofi-Aventis, Eli Lilly, and Novo Nordisk.

  • Full response at: www.bmj.com/content/346/bmj.f3680/rr/651338.

References

View Abstract

Sign in

Log in through your institution

Free trial

Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
Sign up for a free trial

Subscribe