Review of the ORBITA trial

PCI versus placebo PCI in stable coronary artery disease

The Lancet Volume 391, Issue 10115, 6–12 January 2018, Pages 31-40

Background:  For decades, cardiologists commonly used percutaneous coronary intervention (PCI) for the relief of angina. It made sense because PCI resulted in near complete resolution of blood flow through a stenosed vessel. The problem facing evidence-based clinicians was that no previous trial had compared PCI to a placebo (sham) procedure. Instead, previous trials had compared PCI (a procedure) to tablets. In the absence of blinding, a procedure will exert a larger placebo effect than tablets.

The Objective Randomized Blinded Investigation with optimal medical Therapy of Angioplasty in stable angina (ORBITA) trial was designed to assess the effect of PCI versus placebo on exercise time in patients with stable ischemic symptoms.

ORBITA met ethical criteria because previous trials, primarily the COURAGE trial, had found that PCI in addition to medical therapy did not reduce hard outcomes, such as myocardial infarction or death due to cardiovascular causes, compared to medicine alone. In other words, PCI in patients with stable coronary artery disease was not a disease-modifying therapy; it was used to relieve symptoms.

Patients:  Patients had to have single-vessel coronary artery disease (≥ 70% stenosis) that was appropriate for PCI and angina or equivalent symptoms. The authors published in the appendix pictures of every patient enrolled in the trial. Exclusion criteria included acute coronary syndrome, previous bypass surgery, left main stenosis, chronic total occlusions, severe valvular disease or left ventricular dysfunction, moderate or severe pulmonary hypertension, or life-expectancy less than 2 years.  

Baseline Characteristics: The mean age of patients was 65 years. More than 79% were male. Almost 90% had normal left ventricular function. Canadian Cardiovascular Society class included about 60% with class 2 symptoms and nearly 40% with class 3 symptoms. Angina had been present for a mean of 9 months.

Trial Procedures: ORBITA had two phases. First was a 6-week medical optimization phase wherein patients were optimally treated with medical therapy. They had a questionnaires, dobutamine stress echo, and a cardiopulmonary exercise test.  They then had the blinded procedure with either PCI or placebo.

All PCI was done with drug-eluting stents. The procedure included measures to insure blinding, such as headphones during the procedure, sedation and a measure of hemodynamics such as fractional flow reserve. The second phase was a 6-week period of blanking in which patients underwent follow-up assessment. Testing procedures were similar to the pre-procedure protocol.

At all times, the staff were blinded to the procedural data. This included procedural details as well as post-procedural assessment. The recovery staff were well rehearsed in their role of maintenance of blinding. Patients and subsequent medical caregivers were also blinded to treatment allocation. The study physicians present during the procedure had no further contact with the patient during the study.

By the time of randomization, in the PCI group, 103 (98%) of 105 patients were taking aspirin, 103 (98%) were taking a second antiplatelet, and 99 (94%) were taking a statin, compared to 93 (98%), 94 (99%), and 91 (96%) of 95 patients, respectively, in the placebo group. At the same timepoint, in the whole study population, 156 (78%) of 200 patients were taking β blockers and 182 (91%) were taking calcium channel antagonists.

The mean number of antianginal medications in the PCI group was 0·90 (SD 0·8) at enrollment, 2·8 (1·2) at pre-randomization, and 2·9 (1·1) at follow-up, compared to the placebo group in which the mean number of medications was 1·0 (0·9; p=0·357), 3·1 (0·9; p=0·097), and 2·9 (1·1; p=0·891), respectively.

Endpoints: The primary endpoint of ORBITA was the difference between PCI and placebo groups in the change in treadmill exercise time. The power calculation relied on previous trials wherein PCI had resulted in a 48-55 second increase in exercise time over medicine. ORBITA authors designed the trial to detect a 30 second increase in exercise time.

They estimated that a sample size of 100 patients per group had more than 80% power to detect a between-group difference in the increment of exercise duration of 30 seconds, at the 5% significance level, using the two-sample t test of the difference between groups. This calculation assumed a between-patient standard deviation of change in exercise time of 75 s. Since there had been no previous placebo-controlled trials of PCI, the authors initially allowed for a one-third dropout rate in the 6-week period of medical optimization between enrollment and randomization and therefore planned to enroll 300 patients. But the dropout rate was much lower, so only 230 patients had to be enrolled.

The primary endpoint was continuous, and it was calculated as a difference between groups. They also measured secondary endpoints, including measures of angina severity and quality of life.

Results: A total of 368 patients were screened for eligibility, and 200 were randomly assigned. Most were excluded from randomization because they declined to participate. There were 105 allocated to PCI (all but one had PCI) and 95 to placebo (4 patients had PCI due to a procedural complication).

Across all patients, the mean area stenosis by quantitative coronary angiography was 84·4% (SD 10·2), mean FFR was 0·69 (0·16), and mean iFR was 0·76 (0·22). 57 (29%) patients had FFR greater than 0·80 and 64 (32%) had iFR greater than 0·89.

The median length of stent implanted was 24 mm (IQR 18–33). After PCI, the mean FFR improved to 0·90 (SD 0·06; p<0·0001) and iFR to 0·95 (0·04; p<0·0001).

The mean increment of exercise time in the PCI was 28.4 seconds (95% CI 11·6 to 45·1) vs 11.8 seconds (95% CI −7·8 to 31·3) in the placebo arm. The difference was 16.6 seconds (95% CI −8·9 to 42·0). P-value = 0.20.

There was also no significant difference in the time to 1-mm of ST depression, peak oxygen uptake, SAQ-physical limitation, SAQ-angina-frequency, EQ-5D-5L QoL or Duke Treadmill score.

There was, however, an improvement in peak stress wall motion index score in the PCI arm over placebo. This difference was highly statistically significant for the PCI arm.  

During correspondence after the study, one group of writers suggested that ORBITA authors re-analyze their results using the more precise statistical method of analysis of covariance (ANCOVA). The authors accepted this request and found that using ANCOVA gave an effect size of 21·4 s (95% CI −3·4 to 41·1, p=0·09). Due to the novel nature of ORBITA, the authors responded to many letters of correspondence. Linked here.

No patient died during the trial, but there were notable adverse effects, including three periprocedural major bleeding events (two with PCI and one with placebo). In four patients in the placebo group, PCI was needed for a pressure-wire related complication. During the follow-up phase, in the placebo group, one patient developed an acute coronary syndrome and two patients had major bleeding on dual antiplatelet therapy.

Conclusion:

The ORBITA trial found that PCI when used in patients whose angina was medically optimized did not lead to a statistically significant increase in exercise time nor improvement in angina measures. This is a remarkable finding in light of the fact that PCI led to reversal of ischemia as measured by stress echocardiography.

Many of the criticisms of this trial fell short. Some said that 6 week assessment was too short. But ischemia was completely resolved the moment the PCI was done. And the stress echo confirmed improved wall motion. Others criticized the small numbers of patients. But ORBITA did not measure binary outcome events; instead, they measured a continuous variable of exercise time, and powered the study for a shorter difference in exercise time that had been shown in previous trials.

ORBITA also surprised some because of the experience of patients feeling better after PCI. The problem of course is that in clinical practice, neither doctor nor patient are blinded. And “feeling better” can stem from both physiology and placebo effect.

We were also struck by the angiographic severity of the lesions. We laud the authors’ choice to publish the angiograms in the supplement. These images help with external translation of the study, as these were lesions that doctors frequently perform PCI on.

For the first time, ORBITA authors have shown the strong placebo effect of PCI. It informed the treatment of patients with stable coronary artery disease, but it also showed the field of cardiology the feasibility and safety of performing placebo-controlled procedural trials.