A patient reporting that his debilitating symptoms were completely relieved at six-month follow-up greatly surprised the researchers of the 1959 trial in which he was enrolled. (N Engl J Med 1959;260:1115.) Why? He was in the sham procedure arm of the trial.
It is with this in mind that we turn our attention to the recently published ORBITA trial examining the efficacy of percutaneous coronary intervention (PCI) for the symptomatic treatment of stable angina. (Lancet 2017; published online Nov 2; doi: 10.1016/S0140-673632714-9.)
The authors enrolled patients aged 18-85 with angina or equivalent symptoms and at least one angiographically significant lesion (>70%) in a single vessel that was clinically appropriate for PCI. Patients were randomized following medical optimization to undergo PCI or a sham procedure. Patients in the PCI group underwent angiography and stent placement using drug-eluding stents to treat all lesions that were deemed to be significant, with a mandate to achieve angiographic complete revascularization.
Patients in the sham procedure group were sedated for at least 15 minutes on the cath lab table and coronary catheters were removed without the patients undergoing any interventions. Patients underwent follow-up testing at six weeks to assess the burden of symptomatic disease.
The authors enrolled 200 patients between January 2014 and August 2017, and randomized 103 to the PCI group and 97 to the sham procedure arm. Baseline characteristics, pre- and post-randomization medical therapy, were similar between groups. The authors found no difference in their primary endpoint, the incremental improvement in exercise times between the patients who underwent PCI and those who had a sham procedure: 28.4 seconds vs 11.8 seconds with a difference of 16.6 seconds (95% CI -8.9 to 42.0) between the groups.
Nor was there a difference in the majority of secondary endpoints measured, including time to 1 mm of ST depression, peak oxygen uptake, physical limitations, angina frequency or stability, Duke Treadmill Score, or assessment of quality of life. The only difference the authors noted was an incremental change in peak wall stress of 0.09 (95% CI -0.15 to -0.04, p-value 0.0011).
These results are impressively negative. Some will argue that the trial was underpowered to demonstrate a difference in outcomes because of its sample size of 200 patients. In fact, there was a 16.6-second difference in incremental improvement in exercise time favoring the PCI arm that failed to reach statistical significance. Could this be a true benefit provided by PCI? Yes, strictly from a statistical standpoint, these authors are unable to differentiate a 16-second incremental improvement from random chance. But such a perspective fails to see the clinical forest for the statistical trees.
Statistically significant differences can be detected with fairly low sample sizes when continuous variables are measured. Differences in such cases can be identified that boast a p-value less than 0.05, but possess no clinical relevance. Such is the case with incremental improvement in exercise time. The authors powered the study to detect a 30-second difference in incremental improvement in exercise time, a value already below what is considered the minimal important difference in outcomes.
While an increased sample size may have provided a degree of statistical clarity, an incremental improvement in exercise tolerance of 16 seconds holds no clinically relevant value no matter how slight the p-value supporting its veracity, especially considering the well-documented harms associated with the procedure. (N Engl J Med 2007;356:1503.)
PCI has held a fairly esteemed status in the halls of medical interventions. Its popularity is seemingly resistant to mounting evidence discrediting its use for treating stable angina. We have known since the publication of the COURAGE trial that PCI does not reduce the rate of death or MI in patients with stable angina. (J Am Coll Cardiol 2012;60:2017.)
With the publication of ORBITA, even the claims of its ability to reduce the burden of symptoms have been called into question. (N Engl J Med 2008;359:677; N Engl J Med 1992;326:10.) If we continue to support its unrestricted use in patients with stable angina, one has to ask just who exactly we are treating. A powerful placebo effect indeed.Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.