Rosiglitazone and the "Holy Grail"

.
On Wednesday (July 14, 2010), the FDA advisory council looking at whether to ban rosiglitazone (Avandia ®), a drug for the treatment of diabetes that has been found to significantly increase the rate of heart attacks (MIs) decided to not ban it, but to place significant restrictions on its use. On the NPR story Thursday (July 15, 2010), a panel member who is a consumer said (and I paraphrase, but this is close)
“We were very concerned about it but it is effective for lowering blood sugar, which for we diabetics is the Holy Grail.”

What concerns me is that a very well-informed (at least by virtue of being on this panel) consumer (and, from the comment, presumably a person with diabetes) would refer to lowering blood sugar as “The Holy Grail”. As I have noted in several recent posts (most recently Statins and scientific integrity, July 6, 2010), the only patient-important outcomes are premature death and quality of life (morbidity); the only reason any intermediate, or “surrogate”, variable (e.g., blood sugar or Hemoglobin A1c [HbA1c – a longer-term measure of blood sugar], blood pressure, cholesterol level, etc.) is the degree to which it is linked to one of the two patient-important outcomes (mortality or morbidity); none is in itself a patient important outcome, not to mention a “Holy Grail”. Who cares what one’s blood sugar (or cholesterol or blood pressure) is unless we know that it is likely to lead to a bad outcome? It is the outcome that is important. (again, see GY Gandhi, et. al., “Patient-important outcomes in registered diabetes trials”, JAMA. 2008 Jun 4;299(21):2543-9).

In the case of diabetes and blood sugar, there is good evidence that lower blood sugar significantly reduces the rate of cataracts in the eyes, which is certainly a morbidity. There is some evidence that it may be linked to lower rates of microvascular disease (such as kidney and retinal conditions) which definitely cause morbidity. Much weaker evidence relates it to death. Indeed, in a recent publication from the ACCORD trial[1], “tight” control of diabetes (getting the HbA1c below 6, rather than 7) did not decrease microvascular morbidity and led to increased mortality.

We also know that the major cause of death for people with diabetes is macrovascular disease, primarily heart attack. Thus, a drug which increases the rate of heart attacks (definitely increasing morbidity and probably mortality) even if it decreases blood sugar (a surrogate variable) is of great concern.

The actual decision (if the FDA itself chooses to follow the panel’s recommendation and not ban rosiglitazone) to balance the increased risk of heart attack with the benefit of reduced blood sugar, will rest with the patient and the physician. However, responsible physicians – and certainly the leading diabetes experts who are on this panel – need to move away from emphasizing surrogate variables, like blood sugar and HbA1c, to such a degree that patients lose sight of the real outcomes they want to try to achieve.
.
[1] Ismail-Beigi F et al. Effect of intensive treatment of hyperglycaemia on microvascular outcomes in type 2 diabetes: An analysis of the ACCORD randomised trial. Lancet 2010 Jun 29, 1-12
.