Review of the OAT Trial

Coronary Intervention for Persistent Occlusion after Myocardial Infarction

N Engl J Med 2006;355:2395-407

Am Heart J 2011;161:611-21

Background: Registry data suggests that 10-20% of patients with a STEMI present more than 12 hours after the onset of symptoms. The optimal treatment for such patients is unknown. In some cases, the inciting event may have occurred weeks prior and been mistaken for indigestion or another non-life threatening condition. Such patients may present to the hospital with a new diagnosis of congestive heart failure or atrial fibrillation. Echocardiography often reveals a a large wall motion abnormality, perfusion testing demonstrates an infarct with peri-infarct ischemia and an occluded vessel is seen on angiography. Should we try to open it?

On the one hand, the damage has been done. Attempting to open an occluded vessel is associated with higher procedural risks and the patient’s themselves are more often than not sub-optimal candidates for intervention; often having some combination of heart failure, LV dysfunction, older age, multimorbidity and hemodynamic instability.

But on the other hand, revascularization restores blood flow and that has to count for something, right?

The Occluded Artery Trial (OAT) tested the hypothesis that a strategy of routine PCI for total occlusion of the infarct-related artery 3 to 28 days after AMI would improve cardiac outcomes compared to medical therapy alone.

Patients: Patients were eligible if coronary angiography, performed 3 to 28 days after MI, showed a total occlusion of the infarct-related artery with poor antegrade flow and either an EF less than 50% or the occlusion was in the proximal portion of a major coronary vessel with a large risk region, or both. The qualifying period of 3 to 28 days was based on calendar days with day 1 being the onset of symptoms and thus, the minimal time from the AMI to angiography was just over 24 hours. [This is important, readers should not take the inclusion criteria of 3 to 28 days to mean that patients were not eligible if angiography was performed <72 hours from symptom onset]

Patients were excluded with NYHA III or IV heart failure, shock, serum creatinine >2.5 mg/dl, left main or 3 vessel disease, angina at rest, and severe ischemia on stress testing (stress testing was required if the infarct zone was not akinetic or dyskinetic).

Baseline characteristics: The trial included 2,166 patients - 1,082 randomized to PCI and 1,084 to medical therapy. The average age of patients was 59 years and 78% were men. Over 80% were white. The median time between AMI and randomization was 8 days. Patients had normal kidney function with an average GFR of 81 ml/min. The mean EF was 48% with 20% of patients having an EF <40%. Most patients were NYHA class 1 (83%) and the remainder were class II. Nearly 90% of patients had either STE on ECG (67%), a new Q wave (67%) or R-wave loss or some combination of the 3. Less than 30% underwent stress testing but among those who did, approximately 40% had mild or moderate ischemia. [This is pertinent to contemporary practice where the finding of peri-infarct ischemia on perfusion testing is often used to support the decision to pursue revascularization]

Procedures: Patients were randomized to PCI or medical therapy in a 1:1 ratio. All patients received optimal medical therapy with aspirin, anticoagulation if indicated, ACEi, beta blocker and lipid lowering therapy unless contraindicated. Thienopyridines were recommended in both study groups for 1 year after MI.

Patients assigned to PCI were to undergo the procedure within 24 hours after randomization. Stenting was recommended for the occluded vessel as well as for other high-grade stenoses in major coronary segments, whenever technically feasible in the PCI group. PCI of other non-infarct related stenosis was also permitted in the medical therapy; however, such non-infarct related stenting was not commonly performed (7% in PCI group and 6% in medical therapy group).

A subgroup of 124 patients underwent baseline viability scanning to assess myocardial viability before the patient received the assigned treatment. Another subgroup of 381 patients underwent repeat cardiac catheterization at 1 year.

Endpoints: The primary endpoint was a composite of death from any cause, reinfarction, or hospitalization for NYHA class IV heart failure. Certain stipulations were imposed to minimize post-PCI MI’s including exclusion of troponin level for diagnostic use within 10 days of index AMI. An MI in this case would have required required presence of symptoms for 30 minutes or more and electrocardiographic changes consistent with ischemia.

It was originally estimated that 3,200 patients would be required for the study to have 90% power to detect a 25% reduction in the primary endpoint in the PCI group, assuming a 3-year event rate of 25% in the medical therapy group. The sample size was subsequently reduced to 2,400 because of recruitment challenges and a less than expected cross-over rate. The final enrollment of 2,166 patients afforded 94% power to detect a 25% reduction in the primary endpoint.

Results: A total of 1,082 patients were randomized to PCI and 1,084 to medical therapy. PCI of the qualifying occlusion was successful in 87% of patients assigned to PCI. PCI of the non-infarct related artery was performed in 7% of patients in the PCI group and 6% in the medical therapy group. Crossover to PCI was 9% in the medical therapy group (3% within 30 days after randomization) and 4 patients (0.4%) in each of the 2 groups underwent CABG within 30 days. Patients were followed for an average of 2.9 years.

In the main paper, event rates are reported as the estimated 4-year cumulative event rate and not the raw percentage of events per group. For this reason, there may appear to be discrepancies between the event rates and HR. Compared to patients assigned to medical therapy, PCI did not reduce the primary endpoint (17.2% vs 15.1%; HR 1.16; 95% CI 0.92-1.45) or any of the individual secondary endpoints. In an as-treated analysis comparing 937 patients in the PCI group who had successful PCI with 1057 patients in the medical therapy group who did not cross over to PCI within 30 days, the HR was 1.15.

In the subgroup of patients who underwent repeated angiography, the infarct-related artery was patent at 1 year in 83% of patients in the PCI group and 25% in the medical therapy group.

No major subgroup interactions were identified.

The results of viability subgroup study were reported in a separate publication. At baseline, mean infarct size was 26% of the LV, mean infarct zone viability was 43% of peak uptake, and most patients (70%) had at least moderately retained infarct zone viability. In multivariable models, increasing baseline viability independently predicted improvement in EF but there was no interaction between infarct zone viability and treatment assignment for any measure of LV remodeling. The authors concluded that, “in the contemporary era of MED, PCI of the infarct-related artery compared with MED alone does not impact LV remodeling irrespective of viability.”

Conclusions: In patients with a STEMI and an occluded infarct-related artery who are more than 24 hours out from the onset of symptoms, PCI does not improve outcomes compared to medical therapy alone. This is true whether there is evidence of viability or not.

In the OAT trial, there were more events in the PCI group but the difference was not statistically significant; however, in the real world, it is very reasonable to assume that PCI would cause harm to patients meeting the general eligibility requirements for OAT. This is because patients in OAT were young (average age <60) and hemodynamically stable with normal kidney function and mild heart failure symptoms at most. In clinical practice, the typical late-presenting patient is older and sicker. Unfortunately, the authors did not provide information on screening so we cannot determine how truly unique this population was. Regardless, OAT tested the open-artery hypothesis more than 24 hours after STEMI under best-case circumstances and PCI failed.

Based on data from the OAT trial, we strongly advise against PCI in STEMI patients with occluded infarct arteries who are more than 24 hours from the index event. But furthermore, we think this trial is pertinent for all patients with occluded vessels in whom PCI is being considered. If a benefit could not be found in this trial for opening the occluded artery, under what circumstance could it possibly be beneficial?