Is there a better way for the Food and Drug Administration (FDA) to go about approving new medications than the process that's in place right now?
Approval of new drugs to treat chronic disorders is currently based on the results of randomized clinical trials that show improvements in surrogate markers. Examples of surrogate markers include blood cholesterol levels and blood pressure measurements (risk factors for cardiovascular disease); blood glucose or hemoglobin A1c levels (predictors of control and complications of diabetes); and DEXA scan scores (a measure of bone mineral density used to detect osteoporosis).
But what patients and doctors really need to know is whether a new drug actually improves patient outcomes; that is, does a drug that lowers cholesterol or blood pressure also lower the risk of heart attack or stroke?
Does a drug that lowers blood glucose levels also lower the risk of becoming blind or dying of a heart attack? Does slowing the loss of bone mineral density diminish fractures of the hip or spine?
In other words, our real concern here is what actually happens to the patient, not what changes may occur in the surrogate markers we've happened to choose.
This issue of surrogate measures versus actual patient outcomes was highlighted recently when advisory committees to the FDA met to make recommendations on what to do about rosiglitazone, a drug for diabetes.
Using the surrogate marker HbA1c, scientists found that rosiglitazone helped people with diabetes lower their glucose levels over time. But then later, longer-term analyses of the drug's performance showed that rosiglitazone also raised the incidence of heart attacks and strokes in these patients.
The FDA's advisory committees evaluating these new data recommended that both rosiglitazone (Avandia®) and a similar drug, pioglitazone (Actos®), should not be removed from the market but should carry a "black box" warning.
The FDA's drug-approval approach has drawn sharp criticism from many, including Clifford Rosen, M.D., a respected diabetes researcher who wrote an article for a recent issue of the New England Journal of Medicine. Rosen argued that new criteria should be adopted for drug approvals for the treatment of type 2 diabetes, criteria that would hinge on clinical outcomes, rather than on a drug's effect on surrogate markers.
In his article, Rosen asks whether, just because clinical trials showed that rosiglitazone lowered a surrogate marker (HbAlc levels), that drug should have been put on the market without first conducting long-term studies of outcomes — studies that have now shown that the drug raises the risk of heart attacks?
Presumably, Rosen might require similar outcome results for the approval of any new medication to improve cholesterol values or blood pressure, or to slow deterioration of the DEXA scan.
I agree that better patient outcomes will always trump an improvement in a surrogate measurement. I find it difficult, however, to accept this heightened criterion that lengthy and large outcome studies must also be conducted before any new drugs to treat chronic disorders can be approved.
The requirement for outcome results would make the drug-approval process so slow and costly that pharmaceutical companies might shy away from developing new drugs altogether.
For example, it took a number of years after statins were approved, based on their cholesterol-lowering effects, before we had the later outcome studies that eventually proved conclusively that these drugs did in fact reduce heart attacks and strokes. Likewise, similar outcome evidence has shown that lowering blood pressure does prevent heart attacks and strokes, and that slowing the loss of bone mineral density does prevent fractures.
But delaying approval of statins for the prevention of cardiovascular disease and of bisphosphonates for osteoporosis until the results of long-term outcome studies become available would have denied millions of people the benefits of these drugs for many years.
On the other hand, the discovery of the heart attack dangers of Avandia and the increased heart failure risk of Actos, as well as the recent finding of a rise in deaths in patients taking an experimental drug that dramatically raised HDL ("good") cholesterol, illustrates the potential dangers of basing drug approvals solely on surrogate measurements and limited safety data.
I believe the common-sense solution here is for the FDA to continue to approve new medications based on surrogate markers, but for the agency to then require post-marketing studies that would identify serious side effects as quickly as possible.
