Author + information
- Received February 10, 1998
- Revision received May 5, 1998
- Accepted May 20, 1998
- Published online September 1, 1998.
- Matthias Pfisterer, MDa,
- Jafna L Cox, MD∗,
- Christopher B Granger, MD†,* (, )
- Sorin J Brener, MD‡,
- C.David Naylor, MD∗,
- Robert M Califf, MD†,
- Frans van de Werf, MD§,
- Amanda L Stebbins, MS†,
- Kerry L Lee, PhD†,
- Eric J Topol, MD‡,
- Paul W Armstrong, MD∥,
- for the GUSTO-I Investigators
- ↵*Address for correspondence: Dr. Christopher B. Granger, Duke Clinical Research Institute, 2024 West Main Street, Bay A-1, Durham, North Carolina 27705
Objectives. We assessed the use and effects of acute intravenous and later oral atenolol treatment in a prospectively planned post hoc analysis of the GUSTO-I dataset.
Background. Early intravenous beta blockade is generally recommended after myocardial infarction, especially for patients with tachycardia and/or hypertension and those without heart failure.
Methods. Besides one of four thrombolytic strategies, patients without hypotension, bradycardia or signs of heart failure were to receive atenolol 5 mg intravenously as soon as possible, another 5 mg intravenously 10 min later and 50 to 100 mg orally daily during hospitalization. We compared the 30-day mortality of patients given no atenolol (n = 10,073), any atenolol (n = 30,771), any intravenous atenolol (n = 18,200), only oral atenolol (n = 12,545) and both intravenous and oral drug (n = 16,406), after controlling for baseline differences and for early deaths (before oral atenolol could be given).
Results. Patients given any atenolol had a lower baseline risk than those not given atenolol. Adjusted 30-day mortality was significantly lower in atenolol-treated patients, but patients treated with intravenous and oral atenolol treatment vs. oral treatment alone were more likely to die (odds ratio, 1.3; 95% confidence interval, 1.0 to 1.5; p = 0.02). Subgroups had similar rates of stroke, intracranial hemorrhage and reinfarction, but intravenous atenolol use was associated with more heart failure, shock, recurrent ischemia and pacemaker use than oral atenolol use.
Conclusions. Although atenolol appears to improve outcomes after thrombolysis for myocardial infarction, early intravenous atenolol seems of limited value. The best approach for most patients may be to begin oral atenolol once stable.
Several therapies improve survival after acute myocardial infarction (MI), including aspirin (1), thrombolytics (2), angiotensin-converting enzyme inhibitors for patients with impaired left ventricular function (3)and beta-adrenergic blocking agents (4). An overview of 28 randomized trials of early beta blockade in acute MI found a significant, 14% reduction in mortality during the first week (5). In the largest study, the First International Study of Infarct Survival (ISIS-1) trial (conducted before the use of thrombolysis) (6), intravenous, then oral atenolol treatment reduced mortality by 15% compared with no atenolol (p < 0.04). Most of the reduction occurred on the first day, and patients randomized to atenolol had fewer suspected cardiac ruptures (7). In the Thrombolysis in Myocardial Infarction (TIMI) phase II trial (8), patients randomized to acute intravenous metoprolol had less early reinfarction and recurrent ischemia than patients treated with delayed oral metoprolol, but ventricular function and mortality did not differ, although the power to detect such differences was small. There was also a trend towards reduced intracranial hemorrhage with intravenous metoprolol (9). In a much smaller study, although early intravenous treatment with a beta-blocking or a specific bradycardic agent and thrombolysis was safe, this strategy appeared not to enhance myocardial salvage or preservation of left ventricular function (10).
Despite these inconclusive results, early intravenous beta blockade is generally recommended after MI, especially for patients with tachycardia or hypertension and those without heart failure (11). The aim of this study was to describe the use and effects of acute intravenous and later oral atenolol in patients also treated with thrombolytics. The analysis plan was to compare outcomes according to use of any atenolol, and use of intravenous atenolol, with and without adjustment for confounding factors. This was a prospectively planned, observational analysis of the Global Utilization of Streptokinase and TPA (alteplase) for Occluded Coronary Arteries (GUSTO-I) data (12).
The methods, results and clinical definitions for GUSTO-I have been published (12). Briefly, 41,021 patients within 6 h of symptom onset with ST-segment elevation were randomly assigned to receive streptokinase (1.5 million U) with subcutaneous heparin (12,500 U) twice daily beginning 4 h after the start of thrombolysis, streptokinase (1.5 million U) with intravenous heparin, accelerated alteplase with intravenous heparin or combined alteplase (1.0 mg/kg, up to 90 mg) and streptokinase (1 million U) with intravenous heparin. All patients received 160 to 325 mg of aspirin daily. The primary end point of the study was all-cause mortality at 30 days.
Use of atenolol
The protocol recommended that patients without hypotension, bradycardia or signs of heart failure be given atenolol (Tenormin®, ICI Pharma, Wilmington, DE) 5 mg intravenously over 5 min as soon as possible after enrollment, followed 10 min later by another 5 mg intravenously over 5 min. Oral atenolol (50 mg given 10 min after the last intravenous dose, followed by 50 to 100 mg daily) was to be given if no contraindications existed. All other medications and invasive procedures were prescribed at the discretion of the physician.
Five main subgroups were established for this analysis: patients receiving no atenolol, patients receiving any intravenous atenolol, patients receiving only oral atenolol and patients receiving both intravenous and oral atenolol. Groups were further defined by inclusion of only patients “eligible” for atenolol, inclusion of only patients who received oral atenolol the day of or the day after enrollment (days 0 or 1) and inclusion of only patients who survived for at least 24 or 48 h after entry. Eligible patients were those who had a heart rate >55 beats/min, systolic blood pressure >90 mm Hg and Killip class I or II at entry, and who did not develop congestive heart failure or cardiogenic shock during the first 24 h after enrollment. Atenolol initiation after hospital discharge was not collected.
This was a prospectively planned, observational analysis. Baseline characteristics were summarized by frequencies and percentages for categorical variables and by the median, 25th and 75th percentiles for continuous variables.
When predicting 30-day mortality, we used a validated model to adjust for baseline differences (13). To control for the confounding issue of early deaths (occurring before oral atenolol could be given), patients dying within 24 and 48 h of entry were separately removed and the analyses repeated. Likewise, an event such as cardiogenic shock, which independently predicts mortality, might be a reason for, rather than a result of, withholding atenolol. When predicting 30-day mortality or the nonuse of atenolol, we adjusted for these two types of confounding variables by repeating the analyses after excluding patients who developed cardiogenic shock within 24 h of enrollment. To account for the greater opportunity for atenolol use among survivors, another adjustment was performed in which we included only patients who received atenolol during the first 2 days.
For the prediction of intracranial hemorrhage, adjustments for differences in baseline predictors were made with another validated predictive model (14).
To allow comparison with the TIMI-II population, we used their criteria to identify a low-risk cohort (8). Patients in this group had no previous infarction, location of infarction other than anterior, systolic blood pressure >100 mm Hg, heart rate under 100 beats/min and no atrial fibrillation, presented in Killip class I or II and were under 70 years old.
This observational study represents an ancillary analysis of the GUSTO-I database, which has undergone multiple comparisons. To minimize the risk of applying “significance” to differences possibly due to chance alone, p values in the 0.01 to 0.05 range should be interpreted as only suggestive of statistical significance.
Of the 41,021 patients enrolled in GUSTO-I, 24.8% received no atenolol and 75.2% received any atenolol: 44.5% received any intravenous atenolol and 30.7% received it orally only. Patients who received both intravenous and oral atenolol had their first oral dose earlier than did patients given oral atenolol alone. Of patients who received both intravenous and oral atenolol, 92% received the first oral dose by the end of the day after enrollment. In contrast, only 68% of patients managed solely with oral atenolol had received the drug within this interval.
In general, patients receiving atenolol had lower-risk features at baseline, including younger age, higher systolic blood pressure and lower Killip class (Table 1). With regard to MI location, patients receiving atenolol were at higher risk, owing to a greater incidence of anterior MI compared with patients not receiving atenolol.
Unadjusted clinical outcomes, including mortality, generally were better for patients treated with atenolol than for those who were not (Table 2). A higher proportion of patients who received no atenolol died, and their deaths occurred earlier: half of the deaths among such patients occurred within 21 h of entry, and the median time to death was 23 h earlier for patients who received intravenous and oral atenolol than for those given only oral drug.
Overall stroke and intracranial hemorrhage rates were lower among patients given atenolol; however, the rates of intracranial hemorrhage did not differ between patients who received any intravenous atenolol (0.64%) and those who received oral drug alone or no atenolol (0.67%). Even among patients who presented with marked hypertension (entry systolic blood pressure >160 mm Hg), intravenous atenolol use did not confer a lower risk of intracranial hemorrhage (1.3% for any intravenous atenolol vs. 1.4% for patients not given intravenous atenolol). After adjustment for baseline differences, the use of intravenous atenolol was not associated with any difference in the rate of intracranial hemorrhage (p = 0.81).
The reinfarction rate was similar between patients given intravenous atenolol (3.9%) vs. no intravenous atenolol (4.1%, p = 0.61) and between patients given intravenous and oral (3.8%) vs. oral atenolol alone (4.2%, p = 0.07). Among patients given atenolol beginning the day of or after enrollment, 3.8% suffered a reinfarction compared with 3.7% of patients not given early treatment (odds ratio 1.02, 95% confidence interval 0.90 to 1.15). Recurrent ischemia occurred more often among patients given atenolol, although the timing of ischemia relative to the start of atenolol use was not recorded.
The highest incidence of ventricular tachycardia, ventricular fibrillation and asystole occurred in patients receiving no atenolol. Intravenous and oral atenolol use was not associated with a lower incidence of ventricular arrhythmias than oral atenolol alone, but was associated with a higher incidence of asystole.
Suspected cardiac tamponade occurred less often among patients given intravenous atenolol (0.6%) than among patients given no intravenous atenolol (0.9%, p < 0.0001), but the incidence was similar compared with patients given only oral atenolol (0.5%).
When early intravenous atenolol was compared with early (within the first 24 h) oral atenolol use, there were increased rates of congestive heart failure (14.3% vs. 10.7%, p < 0.01), cardiogenic shock (3.3% vs. 2.2%, p < 0.01), recurrent ischemia (22.1% vs. 18.9%, p < 0.01) and the need for pacing (6.0% vs. 4.0%, p < 0.01).
The use of any atenolol conferred a mortality risk fivefold lower than if no atenolol had been given (Table 3). A significantly lower mortality persisted after adjustment for baseline differences, after exclusion of early deaths and when only “atenolol-eligible” patients were considered. This pattern persisted for the use of any intravenous atenolol and for only oral atenolol.
There was a higher unadjusted mortality for patients who received intravenous and oral atenolol than for those given early oral atenolol alone. The increased risk of death persisted after adjustment for baseline characteristics and after exclusion of patients dying within 24 hours of enrollment. After the removal of patients dying during the first 48 hours, the higher mortality among patients receiving intravenous atenolol was of decreased, borderline significance (p = 0.02). After adjusting for baseline differences and including only patients given oral atenolol begun the day of or after enrollment, the likelihood of dying by 30 days did not differ significantly between patients given intravenous and oral atenolol and those given only oral atenolol (p = 0.15).
Early intravenous vs. oral atenolol treatment in clinical subgroups
The point estimate suggesting slight harm with the use of acute intravenous atenolol vs. oral atenolol alone was similar for the overall group, for patients with systolic blood pressure ≥140 mm Hg (or pulse ≥90) and Killip class I at entry and for patients who underwent thrombolysis within 2 h of symptom onset (Fig. 1). “Low-risk” patients were of such low risk (only 7 deaths of 462 patients) that modeling was not performed. Baseline characteristics associated with early death after intravenous atenolol were similar to baseline characteristics associated with early death overall (Table 4).
The use of atenolol in 75% of patients in GUSTO-I, and of intravenous atenolol in 44%, is similar to that of previous trials in which beta blockade was protocol-mandated for all eligible patients (8). This represents a much higher rate of use than in recent United States registry data, which show that only 36% of patients with acute MI receive oral beta-blockers and only 17% are given intravenous drug after thrombolytic therapy (15).
The use of atenolol in GUSTO-I was associated with decreased mortality, stroke, shock and arrhythmias, but increased recurrent ischemia and reinfarction. The interpretation of these data is confounded in that patients treated with atenolol were generally less ill at presentation and may have had greater preservation of left ventricular function. Because there was a delay in beginning atenolol (especially oral atenolol) after enrollment and a very high proportion of deaths (19% of the eventual 30-day mortality) occurred within 4 h after randomization (16), atenolol was used more often in survivors unrelated to any treatment effect. The reduction in adverse events seen with atenolol use, then, may have been due to lower baseline risk, longer survival time, the beneficial effects of atenolol or to a combination of these factors.
To address the issue of differences in baseline risk, validated risk-adjustment models were applied. Another way to adjust for baseline differences was to include only patients who were “atenolol-eligible.” To adjust for the confounding influence of early death (reducing the opportunity for use of atenolol), we excluded all patients who died early and included only patients who began atenolol treatment within 2 days of enrollment.
Studies conducted before thrombolytic therapy came into use have shown a 15% to 25% reduction in mortality with the use of a variety of beta-blockers after acute myocardial infarction (4–6). Although nonrandomized, the current observational study reinforces these findings and extends them into the thrombolytic era. The markedly lower mortality seen with atenolol use was highly significant even after adjustment for confounding factors.
The mortality reductions presented here are larger than those previously reported. The treatment effects may have been exaggerated because the untreated cohort likely consisted of three subpopulations that post hoc adjustments could not adequately tease out: low-risk individuals for whom treatment was deliberately withheld (biasing against an effect of atenolol), otherwise eligible patients whose physicians were uncertain about the efficacy of early atenolol with thrombolysis (the group with whom a reasonably unbiased comparison is possible) and very high-risk patients who were deemed too sick to receive atenolol (biasing towards an exaggerated treatment effect). This caveat notwithstanding, the results are directionally consistent with previous descriptions and are consistent with the short-term mortality reduction from an overview of almost 30,000 patients across 28 trials (4).
Given the acceptance of the general benefit of beta blockade after MI, the most important unresolved clinical issue may be whether patients benefit from acute intravenous beta blockade (beyond oral therapy). The TIMI-IIb trial compared early intravenous, then oral, metoprolol vs. a strategy of significantly delayed oral metoprolol (at 6 days). Although acute intravenous and oral metoprolol conferred a lower risk of reinfarction and recurrent ischemia in the TIMI-II study (8), the lower event rate occurred only during the first week. Because oral metoprolol use did not begin until day 6 in the comparison group, however, their higher ischemic event rate may reflect the delay until oral metoprolol use rather than the lack of immediate intravenous metoprolol. An incremental benefit of immediate intravenous atenolol in addition to oral drug, vs. slightly later oral drug alone, was not seen in GUSTO-I patients after adjusting for differences in baseline characteristics and excluding early cardiogenic shock. In TIMI-II, prospectively defined subgroup analysis suggested that patients presenting within 2 h of symptom onset had greater benefit from intravenous metoprolol; we could not replicate this finding. TIMI-II also identified a “low-risk” group that appeared to derive greater benefit. In GUSTO-I, this subgroup (see Methods) had such low mortality that the effect of early intravenous atenolol could not be meaningfully measured. Although the American College of Cardiology/American Heart Association Guidelines recommend the use of intravenous beta-blockers for patients with hypertension and/or tachycardia in the absence of congestive heart failure (11), we could not show that this subgroup benefited.
Although the TIMI-II study found a trend (p = 0.11) toward reduced intracranial hemorrhage among patients assigned to intravenous metoprolol (9), we were unable to show such an association after adjusting for baseline characteristics. Even for patients presenting with hypertension, the rates of intracranial hemorrhage were the same regardless of use of early intravenous atenolol.
Recurrent ischemia and reinfarction
We found no conclusive evidence of protection against in-hospital reinfarction with the use of any atenolol, although the wide confidence intervals do not rule out the potential for a clinically important reduction in event rates. Recurrent ischemia was more common among patients who received atenolol, but the timing of recurrent ischemia was not recorded. Although impossible to prove, this association may simply reflect the use of atenolol to manage the recurrent ischemia.
Consistent with the lower incidence of suspected tamponade associated with the use of early intravenous atenolol in ISIS-1 (6), a lower incidence of tamponade was seen with intravenous atenolol use in GUSTO-I, although the absolute difference was small (0.6% vs. 0.9% for no intravenous drug, p < 0.0001). A theoretical appeal of early intravenous beta blockade is the potential to reduce the incidence of cardiac rupture, which contributes to the increased risk of early death associated with thrombolysis (2). The prompt administration of a more potent thrombolytic regimen may itself be associated with less “early hazard,” however, and this may lessen any relative beneficial effect of very early beta blockade on mortality (16).
Risk of intravenous atenolol
After adjustment for baseline differences and control for timing of death, the point estimate still suggested slight harm with acute intravenous atenolol use. Thus, although the limits of the analysis are substantial, the results suggest that oral beta-blockers should be the standard of care and that early intravenous atenolol seems of limited value.
This study is limited by its observational nature. Interpretation of the effect of atenolol on outcomes is confounded by the nonrandomized assignment of such therapy. Enrollment of patients into GUSTO-I was based on their eligibility for thrombolytic therapy, and baseline characteristics were those at the time of randomization. The decision to give or withhold atenolol was made some time after randomization, when the status of the patient may have changed. Thus, the risk adjustments herein pertain to the time of randomization to reperfusion regimen and not necessarily to the time of the decision about atenolol use.
Even more uncertain is the influence of early deaths on the results. A clinician observing a patient after randomization may have delayed the decision to begin atenolol pending clinical findings or for logistical reasons; patients dying during this period are counted in the “no atenolol” group. Although eliminating early deaths using various criteria can give a perspective on the effect of this issue, there is no completely satisfactory approach. Additionally, although more than 41,000 patients were enrolled in GUSTO-I, once subgroups are examined the power to detect meaningful differences in outcomes is limited. The limited power is compounded by even fewer events in the analyses from which early deaths are removed. Moreover, a major portion of the benefit of intravenous atenolol may be in the reduction of deaths during the first 24 h. Therefore, removing the patients with early death from the analysis may remove the ability to detect an early beneficial effect.
Given these limitations, although this report can provide a detailed account of the clinical outcomes of patients selected for atenolol therapy, it cannot provide definitive evidence about whether and which type of beta-blocker therapy should be used. Rather, observations about differences in outcome with and without atenolol should be interpreted in light of other sources of evidence and used only to generate hypotheses.
These observational GUSTO-I data do not corroborate reports that early beta blockade reduces the risk of intracranial hemorrhage, but are entirely consistent with available randomized evidence regarding reduced risk of cardiac rupture and death with this therapy. Use of beta-blockers is thus unequivocally beneficial for patients who have received thrombolytic therapy. However, a large, randomized trial still is required to determine the optimal timing and route of such therapy. Until then, data from this analysis of more than 41,000 patients with MI show that the addition of routine intravenous atenolol adds only limited value to a strategy of early oral atenolol given once the patient is stable.
- Global Utilization of Streptokinase and TPA (alteplase) for Occluded Coronary Arteries (trial)
- First International Study of Infarct Survival
- myocardial infarction
- Thrombolysis In Myocardial Infarction (trial)
- Received February 10, 1998.
- Revision received May 5, 1998.
- Accepted May 20, 1998.
- American College of Cardiology
- Fibrinolytic Therapy Trialists’ (FTT) Collaborative Group
- on behalf of the SAVE Investigators,
- Pfeffer M.A,
- Braunwald E,
- Moye L.A,
- et al.
- the TIMI Investigators,
- Gore J.M,
- Sloan M,
- Price T.R,
- et al.
- Van de Werf F,
- Janssens L,
- Brzostek T,
- et al.
- ↵(1990) Guidelines for the early management of patients with acute myocardial infarction: A report of the American College of Cardiology/American Heart Association task force on assessment of diagnostic and therapeutic cardiovascular procedures (subcommittee to develop guidelines for the early management of patients with acute myocardial infarction). J Am Coll Cardiol 16:249–292.
- for the GUSTO Investigators,
- Lee K.L,
- Woodlief L.H,
- Topol E.J,
- et al.
- for the GUSTO Investigators,
- Gore J.M,
- Granger C.B,
- Simoons M.L,
- et al.
- Rogers W.J,
- Bowlby L.J,
- Chandra N.C,
- et al.
- for the GUSTO Investigators,
- Kleiman N.S,
- White H.D,
- Ohman E.M,
- et al.