Author + information
- Timothy P. Murphy, MD∗ ()
- ↵∗Department of Diagnostic Imaging, Rhode Island Hospital, 593 Eddy Street, Providence, Rhode Island 02903
I want to thank Drs. Savas and Kalay for their letter. They state that we “failed” to demonstrate a significant difference in outcomes on the basis of our examined variables. When we study a sample we draw inferences on a population, and assume that our results probably reflect those in the larger population, so the word failed would not be applicable unless there was evidence that the results are wrong.
Regarding the question of the endpoints, the CORAL (Cardiovascular Outcomes in Renal Atherosclerotic Lesions) study used endpoints such as heart attack, stroke, and heart failure because they are hard, clinically important events (1). Prior studies using blood pressure as a surrogate endpoint did not show any benefit of renal artery intervention (2,3). Although blood pressure is a widely studied and validated endpoint for studies of essential hypertension and correlates strongly with events such as heart attack, stroke, and heart failure in that population, the validity of hypertension as a surrogate endpoint in a population of renal artery hypertension patients has never been validated and there are data that suggest that even with blood pressure control, those events are experienced often in patients with renal artery hypertension (4). The CORAL study was powered to show a 25% difference in hard clinical endpoints at 2 years. Of course those events may be attributable to comorbid conditions and not attributable to the intervention in some cases, but an invasive, expensive procedure has to have some clinical effect if its performance is justifiable, and calculation of the sample size considers the minimum attributable treatment effect that would justify the procedure. In other words, if revascularization of the renal artery (which is expensive and entails some risk) improves blood pressure (which is asymptomatic), but does not reduce hard clinical events, its performance clinically would not be justifiable. In fact, in the CORAL study there was a statistically significant difference of 2.3 mm Hg in systolic blood pressure favoring renal artery stenting, but no clinically significant difference in the incidence of endpoint events. Furthermore, the baseline risk that is not attributable to renal artery stenosis should be controlled by randomization.
Regarding the importance of improving renal function, progressive renal insufficiency and renal replacement therapy were components of the primary endpoint in the CORAL study. The ASTRAL (Angioplasty and Stenting for Renal Artery Lesions) and CORAL studies both showed no treatment group renal function outcome differences (1,5). However, the CORAL study was mostly a study of patients with preserved renal function, with a mean baseline estimated glomerular filtration rate of 58 ml/min/1.73m2, and it would be erroneous to draw inferences on patients with stage IV or V renal failure from the CORAL study. But, to clarify, in the CORAL study there was no statistically significant benefit in patients with creatinine ≤1.6 mg/dl, although there was a trend for interaction at that cutpoint (p = 0.09) (1).
Please note: Dr. Murphy is an employee of Rhode Island Medical Imaging, Inc.; has ownership interest in Sentient Bioscience, Summa Therapeutics, Saphena Medical, and Anaxiom, Inc.; has received research funding from Cordis/Johnson & Johnson, Abbott Vascular, the National Institutes of Health, Novate Medical, The Medicines Company, and the National Heart, Lung, and Blood Institute; and has received honoraria from Merit Medical.
- American College of Cardiology Foundation