Review of the STREAM trial
Early Fibrinolysis vs Primary PCI in STEMI
N Engl J Med 2013;368:1379-1387
Background In 2013, it had been established that primary PCI for STEMI was the preferred strategy. Yet many patients did not have prompt access to primary-PCI capable hospitals and transfer delays could impact outcomes. The vast majority of patients with STEMI who present to non-PCI facilities do not subsequently get primary PCI within recommended times.
Delays led to the development of prehospital care, such as ECGs in the ambulance, and pre-hospital delivery of fibrinolysis. The Strategic Reperfusion Early after Myocardial Infarction (STREAM) study evaluated whether a fibrinolytic-therapy approach consisting of prehospital or early fibrinolysis with contemporary antiplatelet and anticoagulant therapy, coupled with timely coronary angiography, provides a clinical outcome similar to that with primary PCI in patients with STEMI who present early after symptom onset.
Patients Eligible patients had a) STEMI within three hours, b) could not have primary PCI within one hour of first medical contact. No formal exclusion criteria were listed in the main manuscript.
Baseline Characteristics A total of 1892 patients underwent randomization in 1:1 fashion. The mean age of patients was 59 years. Less than 15% of both groups were older than 75 years. Females were 20%. More than 90% of patients were Killip class 1. Less than 10% of enrolled patients had had prior CHF, MI, or PCI.
Procedures Patients were randomized in a 1:1 ratio to fibrinolysis followed by timely coronary angiography or primary PCI. All patients were transferred to a PCI-capable hospital; for all non-PCI community hospitals participating in the study, a well-developed hub-and-spoke relationship with a PCI-capable site was required.
The fibrinolytic strategy included early use of concomitant antiplatelet and anticoagulant medications, as well as additional discretionary glycoprotein IIb/IIIa antagonists. Tenecteplase was administered in a weight-based dose and was combined with low-molecular-weight enoxaparin, weight and age adjusted.
Antiplatelet therapy consisted of clopidogrel in a 300-mg loading dose (omitted for patients ≥75 years of age) followed by 75 mg daily and aspirin (150 to 325 mg) immediately followed by 75 to 325 mg daily.
Urgent coronary angiography in the fibrinolysis group was permitted at any time in the presence of hemodynamic or electrical instability, worsening ischemia, or progressive or sustained ST-segment elevation requiring immediate coronary intervention, according to the investigator's judgment.
Endpoints The primary end point of the trial was a 30-day composite of death from any cause, shock, congestive heart failure, or reinfarction. Single efficacy end points as well as safety end points consisting of ischemic stroke, intracranial hemorrhage, nonintracranial bleeding, and other serious clinical events were recorded.
The statistical analysis plan was complicated. A sample size of 1000 patients per study group was planned, and the rate of the primary end point in the primary PCI group was projected to be 15.0%. After one-fifth of patients had been enrolled, trialists amended the protocol to reduce the dose of tenecteplase by 50% in patients older than 75 years because of excess ICH. ECG criteria for inferior MI was also changed to require at least 3 mm (up from 2) of ST elevation in two contiguous leads.
This trial was designed as a proof-of-concept study. All statistical tests were of an exploratory nature.
Results The median time delay from the onset of symptoms to first medical contact and randomization was similar in the two groups ( 61-62 minutes). The median times between symptom onset and start of reperfusion therapy (bolus tenecteplase or arterial sheath insertion) were 100 minutes and 178 minutes, respectively (P<0.001).
As expected, the median time from randomization to angiography was longer in the fibrinolysis group than in the primary PCI group, with a delay of 2.2 hours for the 36% of patients who required rescue or urgent intervention and 17 hours for the remaining 64% of patients.
The primary end point (death from any cause, shock, congestive heart failure, or reinfarction up to 30 days) occurred in 116 of 939 patients (12.4%) in the fibrinolysis group and 135 of 943 patients (14.3%) in the primary PCI group (relative risk in the fibrinolysis group, 0.86; 95% confidence interval [CI], 0.68 to 1.09; P=0.21). None of the components of primary endpoint differed statistically.
The incidence of the primary end point in the prespecified subgroups was generally similar to the overall result. No significant treatment interactions were found.
Rates of stroke were low in the two study groups, but both intracranial hemorrhagic (1.0 vs 0.2) and primary ischemic strokes (0.6 vs 0.3) were more frequent in the fibrinolysis group than in the primary PCI group. After the dose reduction of tenecteplase in patients 75 years of age or older, there were no cases of intracranial hemorrhage (0 of 97 patients), as compared with 3 of 37 patients (8.1%) in this age group before the amendment. There was no statistically significant difference in major non-intracranial bleeding (6.5% vs 4.8%).
Conclusions Patients with STEMI who presented early after symptom onset with an ST-segment elevation of at least 2 mm in two contiguous leads had similar rates of the primary composite end point of death, shock, congestive heart failure, or reinfarction at 30 days, regardless of whether they underwent prehospital fibrinolysis or primary PCI.
Important caveats include the fact that patients who could have had early PCI (within one hour of first medical contact) were excluded and these results should not be applied to these patients. Similarly, these results should not be applied to patients who present with STEMI of longer than 3 hours after symptom onset.
In addition, the sample size was moderate and there was no formal hypothesis testing. The 95% confidence intervals allow for a 32% lower rate of the primary endpoint as well as 9% worse outcome relative to primary PCI. The authors explain that the upper bound of a 9% relative risk increase would likely have fell below typical non-inferiority margins—though this was not prespecified as a noninferiority trial.
Finally, the lack of clear superiority of fibrinolysis using this approach, in addition to the increase risk of bleeding associated with its use, limited the adoption of this study results.