The FDA and diabetes medications

On 8th November 2012, the Federal Drugs Administration (FDA) in the United States received guidance from its Endocrinologic and Metabolic Drugs Advisory Committee on the application for a license for degludec insulin and it’s combination preparation degludec/aspart. Degludec has a protracted delivery following subcutaneous injection allowing for flexibility in the timing of its once daily dosing and low risk of night-time hypoglycaemia. Although the committee had voted in favour of a license, this was only by an 8-to-4 split and there had been a unanimous recommendation (12:0) for a study of the cardiovascular (CV) safety of degludec. This is in marked contrast to the opinion of the European Medicines Agency (EMA), which recommended the granting of a marketing authorisation for degludec on 18th October 2012 with no apparent concerns over CV safety. The Japanese authorities were also satisfied with the CV analyses. So, why should there be this divergence of opinion and who has got it right?

The story goes back to May 2007 when the New England Journal of Medicine published a meta-analysis which suggested that the oral hypoglycaemic agent, rosiglitazone, increased the risk of myocardial infarction. Clearly, for a condition like type 2 diabetes (T2DM) where cardiovascular disease is the major cause of death, this was a significant problem and led to much discussion. Although, in my view at least, the debate was never settled, the end result was the 2010 withdrawal of the license for rosiglitazone in the EU and severe restrictions applied in the US by the FDA.

During this period, the FDA published guidance for industry, which advised on how pharmaceutical companies should evaluate the cardiovascular risk of new therapies for T2DM. This document mandated that ‘major adverse cardiovascular events’ (termed MACE) seen in phase 2 and phase 3 clinical studies should be aggregated and compared with the event rates seen in the studies’ comparator arms. Given that these studies are typically small and of short (24-week) duration, it was acknowledged that event rates would be low and it was unlikely that any statistically significant differences in MACE would be seen. However, where the upper confidence interval could not exclude a 30% increase in MACE associated with a new drug, the FDA stated that it would require a CV safety study to be performed after the drug had been licensed. For those drugs where an 80% increase in MACE could not be excluded, a license would not be granted until the results of a CV safety study were available.

The FDA has continued with this recommendation and as a result there are currently at least six massive CV studies recruiting in the UK, typically placebo-controlled and lasting for between 2 and 7 years (this difference depending on the level of CV risk in the patients under study). Only one drug, alogliptin, has been required to complete a CV safety study prior to licensing although it has been given the go-ahead in Japan (the country which hosts its manufacturer).

The FDA has also adopted a much more aggressive attitude to other safety signals, especially in relation to the risk of malignancy. Liraglutide, which is currently the leader in the GLP-1 receptor agonist market, had its’ US license held up for over a year due to the FDA’s concerns over C-cell thyroid malignancies seen in rodent studies. Likewise dapagliflozin, the first of the new class of sodium-glucose transporter 2 inhibitors was recently granted a license for use in Europe but is currently on-hold in the US due to concerns of increased rates of breast and bladder cancers.

So, was the degludec decision entirely what one might have expected for a new insulin molecule, especially since the MACE event rates (80/8,918) had an upper confidence interval just below 80%. Well, not quite. The FDA had previously signaled that it would not require CV studies to be performed on insulins, presumably reflecting the almost 90 years experience with this class of agent. This change of heart may impact on other insulin developments, not only at the upper end of the market but also including the development of cheaper so-called ‘bio-similar’ versions of the current analogue insulins.

On this background dominated so much by safety, the FDA’s view of the new anti-obesity agent, Qsymia, is perhaps surprising. Anti-obesity therapies have a poor track record with the withdrawal of several drugs in the past (for example, dexfenfluramine, rimonobant and most recently sibutramine). Qsymia is a combination product which includes phentermine, a stimulant that suppresses the appetite, and has long been used for short-term weight loss. Recognised side-effects include tachycardia and blood pressure elevation, which immediately raises the possibility of CV safety. The FDA approved Qsymia in July 2012 for use in adults with a BMI >30Kg/M2 or for those with a BMI >27Kg/M2 and a co-morbidity such as hypertension or type 2 diabetes. At least they have requested a CV safety study!

Professor Steve Bain