Review of the ISIS-1 Trial
Randomised Trial of Intravenous Atenolol Among 16,027 Cases of Suspected Acute Myocardial Infarction: ISIS-1
Lancet 1986:57-65
Background The International Study of Infarct Survival (ISIS) was an international collaboration to evaluate the effects on survival of treatments with broad applicability during the early phase of myocardial infarction. This will be the first of many trials presented by this group. Here the investigators sought to test the hypothesis that early beta-blockade with atenolol would reduce vascular mortality in patients with MI. The average start date of propranolol following MI in the BHAT trial was day 14, here it will be immediate.
Patients Patients with suspected myocardial infarction, who the responsible physician thought were within 12 hours of symptom onset and not already on beta-blockers or verapamil were eligible. Contraindications included: heart rate persistently <50 beats per minute, systolic blood pressure persistently below 100 mmHg, second or third degree heart block, “severe” heart failure or bronchospasm. Criteria for defining severe heart failure are not described in the main paper.
The authors designed the trial to be pragmatic—in order to facilitate patient enrollment. Once patients were randomized they could not be regarded as “ineligible”, even if MI was later refuted as the diagnosis. There were no special procedures dictating patient follow-up after their hospitalization ended.
Baseline characteristics The average age was 59 and the majority were male (77%). The mean SBP and heart rate at entry were 145 mmHg and 79 beats per minute, respectively. Less than 20% had a previous MI and only 6% had diabetes. Approximately half of the participants had a definite MI, which by the definitions used in this trial meant an ST-segment elevation MI meeting ST-segment elevation cutpoints. Presumably many more participants had STEMI’s with less pronounced ST elevations and another significant percentage would have had NSTEMI’s.
Procedures Patients were immediately randomized to atenolol or control. Those allocated to active treatment received the following sequence of interventions:
An immediate intravenous injection of 5 mg of atenolol, given over 5 minutes and stopped if the heart rate fell below 40 beats per minute or if any other contraindication developed.
If, after 10 minutes, the heart rate was >60 beats per minute, up to 5 mg more of atenolol was injected.
If, after another 10 minutes, the heart rate was >40 beats per minute, 50 mg of oral atenolol was given.
12 hours later another 50 mg of oral atenolol was given.
100 mg daily or 50 mg twice daily of oral atenolol was given thereafter for another 6 days or until discharge if it occurred earlier. *If thought necessary, oral treatment could be reduced to 50 mg daily or stopped altogether. Rules defining this are not provided.
In the control group, NO placebo was given but beta blockers were to be avoided in hospital unless thought to be clearly indicated (i.e., chest pain unresponsive to nitrates or calcium antagonists, control of hypertension).
Endpoints Vascular mortality was the primary endpoint. Follow-up after discharge involved only mortality through government records wherever possible.
Results 16,027 participants were randomized from 245 hospitals. 94% of beta blocker-allocated participants received at least the first 5 mg IV dose of atenolol, 2% had the first IV dose discontinued and the other 4% developed a contraindication before a dose could be given. Oral atenolol was not received by 7% of the atenolol-allocated patients, and overall, 27% had later discontinuation or a significant dose reduction. 2% of control patients received some dose of IV beta blocker in the hospital and 7% received some oral beta blocker. Medication use in hospital and on discharge was different between groups.
During the treatment period (days 0-7) there were 313 (3.9%) vascular deaths in the atenolol group compared with 365 (4.6%) in the control group, which was statistically significant (p<0.04). The difference in all-cause death was also significant as there were only 6 additional non-vascular causes of death. All of the apparent benefit on vascular mortality was observed in days 0-1 with no further difference thereafter.
At 1 year, the difference between the atenolol and control groups for vascular and all-cause death were 10.3% vs 11.6% and 10.8% vs 11.9%, respectively. Even over the course of 1 year, non-vascular death contributed very little toward mortality, which confers information about the starting health of the patients who were enrolled.
Inspection of various subgroups yields interesting additional information. First, men accounted for 77% of all participants and 65% of vascular deaths from days 0-7 but did not benefit from beta blockade in terms of vascular death reduction (3.5% vs 3.7%). In contrast, women, who accounted for only 23% of participants and 35% of vascular deaths, derived a benefit over twice as large as the trial average (5.2% vs 7.5%). This subgroup difference was not addressed by the authors in the main manuscript.
Other interesting subgroups included patients with signs of hemodynamic instability based on HR >90 bpm or SBP <120 mmHg. In both groups, overall vascular death rates were higher than average; however, there was no benefit from beta blockade. In fact, vascular death was higher in the beta blocker groups. And participants with both of these features, who made up only a small percentage of the overall trial population (2.2%) experienced a much greater increase in vascular death (17.6% vs 15.1%). This information is critical for clinical translation and will be seen again in a subsequent large trial of beta blockers in acute MI.
Conclusions IV atenolol followed by oral administration over 7 days reduced vascular and all-cause mortality in patients presenting with suspected MI with an NNT of approximately 100 over the course of 7 days and 1 year. About half of the patients had STEMI and the other half had NSTEMI’s.
Patients presenting with heart failure, hemodynamic instability (SBP <100 mmHg) or bradycardia (HR <50 bpm) were excluded.
All of the benefit appears to have been derived between days 0-1. Almost all deaths over the first 7 days and the overwhelming majority of deaths over 1 year of follow-up were vascular. Notably the rate of death in this trial was over twice as high as in BHAT, which had much stricter inclusion criteria and did not initiate treatment until after patients were stabilized (the average start day was day 14); whereas patients were immediately initiated on treatment in ISIS-1. Despite the higher mortality rate observed in this trial, the treatment effect was less impressive compared to BHAT.
Treatment effect heterogeneity was identified by the trial authors for patients with soft hemodynamics (SBP <120 mmHg and HR >90 bpm and the combination of both was additive).
The BHAT and ISIS-1 trials provide the bulk of evidence from seminal trials supporting beta blocker use in hemodynamically stable patients with acute heart attacks. They also provide direct and indirect support for the notion that beta blockade is harmful in patients with hemodynamic instability. Neither trial supports beta blocker use in MI patients with acute heart failure.
Subscribe to listen to our podcast coming later today discussing the BHAT and ISIS-1 trials. Next week, we will stay in the Acute Coronary Syndrome category with reviews and a podcast featuring the GISSI-1 and ISIS-2 trials.
Learned lots of details here which I never knew. IV atenolol has never had much buy-in in my 21 years at a large Canadian community hospital (and I don’t recall much uptake while I was a trainee either). Even IV metoprolol is now falling out of favour, although early oral beta blockade remains ensconced in our pattern of practice. Look forward to the final summary of evidence for BB in the post MI non HF space.
This is very useful for a cardiology resident, especially since it comes in bits, it is not overwhelming, and you have the time to catch up and reflect. The audio feature is great also.
I hope that this series will advance my knowledge on the trials that got us where we are now in terms of clinical practice. It is information that can be easily skipped throughout training, but still very important. I am very grateful for this!