Author + information
- Received November 1, 1999
- Revision received March 1, 2000
- Accepted April 12, 2000
- Published online September 1, 2000.
- David Hasdai, MD∗,
- Robert A. Harrington, MD, FACC†,
- Judith S. Hochman, MD, FACC‡,
- Robert M. Califf, MD, FACC†,
- Alexander Battler, MD, FACC∗,
- James W. Box, MS†,
- Maarten L. Simoons, MD, FACC§,
- Jaap Deckers, MD§,
- Eric J. Topol, MD, FACC∥ and
- David R. Holmes Jr., MD, FACC¶,* ()
- ↵*Reprint requests and correspondence: Dr. David R. Holmes, Jr., Cardiovascular Diseases and Internal Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN 55905
The study examined whether antiplatelet treatment with eptifibatide affected the frequency and outcome of shock among patients in the Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy (PURSUIT) trial who had acute coronary syndromes but not persistent ST-segment elevation.
Preliminary reports suggest a salutary effect of antiplatelet agents when shock complicates acute myocardial infarction.
We analyzed the impact of antiplatelet treatment with eptifibatide on the frequency and outcome of cardiogenic shock developing after enrollment. PURSUIT was a double-blind, randomized trial that examined the efficacy of eptifibatide (180 μg/kg bolus + continuous infusion of 2.0 μg/kg/min for ≤96 h) versus placebo among patients who had acute coronary syndromes but not persistent ST-segment elevation.
Shock developed in 2.5% of the 9,449 patients at a median (25th, 75th interquartiles) of 94.0 (38, 206) h. Death by 30 days occurred in 65.8% of shock patients. Patients who had acute myocardial infarction upon enrollment had a greater incidence of shock (2.9% vs. 2.1%, p = 0.01), developed shock earlier (40.2% <48 h vs. 20.9%, p = 0.001), and had higher 30-day mortality from shock (77.2% vs. 52.7%, p = 0.001). Randomization to eptifibatide did not affect the occurrence of shock (p = 0.71, adjusted odds ratio [OR] = 0.95, 95% confidence interval [CI] = 0.72–1.25). However, shock patients treated with eptifibatide had significantly reduced adjusted odds of 30-day death (p = 0.03, adjusted OR = 0.51, 95% CI = 0.28–0.94).
Patients with shock treated with eptifibatide had significantly reduced adjusted odds of death, suggesting a salutary effect of antiplatelet therapy on shock. This finding warrants verification in specifically designed studies.
The acute coronary syndromes encompass a wide range of clinical scenarios from unstable angina to acute myocardial infarction. Most patients present with either no electrocardiographic (ECG) abnormalities or changes in the ST-T segment other than persistent ST-segment elevation (1). Cardiogenic shock accounts for most deaths that occur in these patients (2).
Recent studies have demonstrated that platelet glycoprotein IIb/IIIa blockade improves the outcome of patients with acute coronary syndromes who do not have persistent ST-segment elevation (3–10). There have also been preliminary reports of a salutary effect exerted by platelet glycoprotein IIb/IIIa inhibitors in a small number of patients with shock in the setting of acute myocardial infarction (11,12). The aim of this retrospective analysis of the Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy (PURSUIT) trial (10) was to determine the impact of treatment with the platelet glycoprotein IIb/IIIa blocker eptifibatide (Integrilin™, COR Therapeutics, South San Francisco, California) on the frequency and outcome of shock among patients who had acute coronary syndromes but did not have persistent ST-segment elevation.
Details of the PURSUIT trial have been previously described in detail (7). Patients presenting with ischemic chest pain at rest lasting ≥10 min within the previous 24 h were eligible for randomization provided they had either transient ST-segment elevation >0.5 mm or transient or persistent ST-segment depression >0.5 mm or T-wave inversion >1 mm within 12 h of an episode of chest pain, or had a creatine kinase, MB fraction (CK-MB) above the upper limit of normal for that hospital. The exclusion criteria included persistent ST-segment elevation >1 mm. As in the original PURSUIT study, all analyses in the current study used the high-dose eptifibatide group (180 μg/kg bolus of eptifibatide followed by continuous infusion of 2.0 μg/kg/min) as the treatment group for comparison with the placebo arm. It was recommended, but not mandated, that the study drug be infused until hospital discharge or for up to 72 h. If a coronary intervention was performed near the end of the 72 h, then the infusion could be continued for an additional 24 h, up to 96 h. The use of assist devices and referral to cardiac catheterization with subsequent percutaneous or surgical revascularization interventions were at the discretion of the treating physician.
Cardiogenic shock was a predefined adverse event that was prospectively recorded. Criteria for shock included systolic blood pressure <90 mm Hg for ≥1 h, not responsive to fluid resuscitation alone, felt to be secondary to cardiac dysfunction, and associated with signs of hypoperfusion or a cardiac index of ≤2.2 liters/min/m2. If systolic blood pressure increased to >90 mm Hg as a result of positive inotropic agents alone in less than 1 h, the event was still classified as shock.
Myocardial infarction upon enrollment and during follow-up was diagnosed based on the presence of ischemic chest pain, ECG changes, and a rise in creatine kinase and/or its MB isoform (10). Suspected infarctions during follow-up were evaluated by a clinical-events committee blinded to the treatment received.
Continuous variables are presented as medians with 25th and 75th percentiles and discrete variables as frequencies and percentages. Logistic regression models were used to identify the baseline measures that had predictive value with regard to the outcomes. Because the precise timing of invasive procedures such as angioplasty, coronary bypass surgery, or intraaortic balloon counterpulsation were not available, we did not include these variables in our analyses. Predictors in the model were tested using the Wald chi-square test. Results are also presented as odds ratios (OR) with corresponding 95% confidence intervals (CI). Once the final model was developed, the individual patient’s risk was determined, and a logistic regression model was developed with shock as the outcome and log(risk) as the predictor. The quality of this final model as well as the original model is described with the use of the concordance index, which is a description of the discriminant power of the model to predict an outcome reliably (13).
On the basis of the coefficients in the model, a probability chart was developed for the occurrence of shock after enrollment. As we previously described (14,15), each variable in the model received a certain score based on the value of the variable. The total points were then transformed into predictive values.
For verification purposes, we applied the model developed in the PURSUIT cohort to the Global Use of Strategies to Open Occluded Coronary Arteries IIb (GUSTO-IIb) cohort of patients without persistent ST-segment elevation (1). In GUSTO-IIb, the outcome of patients randomized to receive either heparin or hirudin as adjunctive therapy was not significantly different.
Logistic regression modeling techniques were also used to evaluate the univariate relationship between demographic and clinical characteristics and the likelihood of dying within 30 days among patients with shock in PURSUIT. Multivariable logistic regression techniques were then used to develop a model to predict 30-day mortality among shock patients in PURSUIT.
Shock versus nonshock population
Of the 9,449 patients receiving either placebo or high-dose eptifibatide in PURSUIT, 237 (2.5%) developed shock after enrollment (Table 1). Of these 237 patients, 117 (49.4%) were randomized to placebo.
The patients who developed shock were older, shorter, and lighter in weight than the nonshock patients. In addition, shock patients more frequently had hypertension, peripheral vascular disease, and a prior history of either coronary artery disease or heart failure, whereas nonshock patients were more likely to be current smokers (Table 1). Shock patients had a faster heart rate at enrollment and more pulmonary rales (Table 1), and they more commonly had ST-segment depression in their initial ECG (Table 2).
Shock patients with and without myocardial infarction at enrollment
Of the 237 shock patients, 127 (53.6%) presented with myocardial infarction (Table 1). Of the patients who presented with myocardial infarction 2.9% developed shock, as compared with 2.1% of patients who did not have myocardial infarction at enrollment (p = 0.01). Significant differences existed between shock patients who presented with myocardial infarction and those who did not have myocardial infarction at enrollment (Tables 1 and 2). The former group had more diabetic patients, a faster heart rate and more pulmonary rales (Table 1). Their initial ECG more frequently depicted ST-segment depression (Table 2). Myocardial infarction or reinfarction within 30 days occurred in 54.0% of the shock patients (57.2% and 51.2% of the shock patients who presented without and with myocardial infarction, respectively).
Time interval from enrollment to shock
Shock developed in the majority of patients >48 h after enrollment (Fig. 1), with a median time (25th, 75th interquartiles) of 94.0 (38, 206) h. Patients presenting with myocardial infarction developed shock earlier (40.2% vs. 20.9% within 48 h, p = 0.001). The median time from enrollment to shock for patients who presented with myocardial infarction was 64.8 (23, 168) h and 70.0 (33, 174) h for patients randomized to placebo and eptifibatide, respectively. For patients without myocardial infarction upon presentation, the corresponding times were 116.1 (52, 240) h and 148.9 (67, 247) h.
Shock developed after enrollment but before drug infusion in only two patients. Shock developed while on treatment in 37 (15.7%) patients, 29 presenting with myocardial infarction (15 randomized to placebo and 14 to eptifibatide) and 8 presenting without myocardial infarction (7 randomized to placebo and 1 to eptifibatide) (p = 0.001 for comparison of myocardial infarction vs. non-myocardial infarction). The remaining 198 patients developed shock after cessation of drug infusion.
Predictors of shock development
In the multivariable analysis, the age of the patient, the presence of ST-segment depression in the initial ECG, and the physical findings at the time of enrollment accounted for a great part of the information that predicted the subsequent development of shock (Table 3). The treatment modality did not impact on the occurrence of shock (p = 0.71, adjusted OR = 0.95, 95% CI = 0.72–1.25). This model was transformed into a scoring algorithm that can be used by the clinician to evaluate an individual patient’s risk of developing shock (Table 4). The concordance index of the original logistic model was 0.710, and of the validated model it was 0.670. When we verified the validity of this model derived from the PURSUIT cohort in GUSTO-IIb patients who did not present with persistent ST-segment elevation, the concordance index was 0.682, indicating that the model had similar reliability in predicting the development of shock in both cohorts.
Coronary angiography, revascularization, and assist devices
Coronary angiography, performed in 68% of the patient cohort, was more frequently performed among nonshock patients (Table 5). A similar trend was evident in terms of percutaneous coronary interventions. In contrast, shock patients were more commonly referred for coronary bypass surgery or underwent intraaortic balloon counterpulsation or pulmonary artery catheterization. Among the shock patients, coronary angiography and coronary bypass surgery were more commonly performed among those who did not present with myocardial infarction, as was the use of intraaortic balloon counterpulsation.
Compared with the 2.0% incidence of death within 30 days among nonshock patients, 65.8% of shock patients died within 30 days (p = 0.001). Among shock patients, those who presented with myocardial infarction had a higher incidence of 30-day death (77.2% vs. 52.7%, p = 0.001). In the multivariable analysis (Table 6), shock patients with myocardial infarction upon presentation, ST-segment depression in the initial ECG, or diabetes mellitus were at increased risk of dying.
The 30-day incidence of death was 58% and 48% for shock patients who did not have myocardial infarction at enrollment who received placebo or eptifibatide, respectively. The corresponding incidence of death for shock patients with myocardial infarction at enrollment was 85% and 69%. In the multivariable model (Table 6), treatment with eptifibatide significantly reduced the adjusted odds of 30-day death from shock (p = 0.03, adjusted OR = 0.51, 95% CI = 0.28–0.94).
We also examined whether there was an interaction between treatment and shock status on outcome (30-day death). Among patients who did not develop shock, the incidence of 30-day death was 2.0% for patients treated with either placebo or eptifibatide. Among patients who did develop shock, the incidence of 30-day death was 73.5% for placebo and 58.5% for eptifibatide. The p value for an interaction is 0.032, indicating that the benefit of eptifibatide in reducing 30-day death is restricted to shock patients.
The major finding in our retrospective analysis of the PURSUIT trial was that patients who developed shock and had been treated with eptifibatide had significantly improved outcome, although eptifibatide did not attenuate the development of shock. The adjusted odds of dying within 30 days among shock patients who had received eptifibatide was reduced by 49%. Overall, cardiogenic shock developed in 2.5% of patients presenting with acute coronary syndromes without persistent ST-segment elevation. Most cases of shock occurred >48 h after enrollment. Patients who had acute myocardial infarction upon enrollment had a greater incidence of shock than did patients who did not present with myocardial infarction (2.9% vs. 2.1%), developed shock earlier (40.2% before 48 h vs. 20.9%), and subsequently died more frequently from shock (death within 30 days in 77.2% vs. 52.7% of cases, respectively). The major predictors of shock were the patient’s age, the presence of ST-segment depression in the initial ECG, and physical findings upon enrollment.
Shock developing among patients without persistent ST-segment elevation
Most recent publications regarding the frequency and outcome of cardiogenic shock have focused on patients who presented with persistent ST-segment elevation, many of whom were eligible to receive fibrinolytic therapy. There are few current data regarding shock developing in other acute coronary syndromes.
Holmes et al. (16) recently analyzed the occurrence of shock in GUSTO-IIb, a double-blind study designed to examine the efficacy of hirudin versus heparin as adjunctive therapy for patients with a wide array of acute coronary syndromes, including patients with and without persistent ST-segment elevation upon enrollment (1). The researchers found that shock occurred in 2.5% of patients who did not have persistent ST-segment elevation upon enrollment, as opposed to 4.2% of those with persistent ST-segment elevation. In addition, the mortality rates were 73% and 63%, respectively. In that study, the qualifying ischemic episode was acute myocardial infarction in approximately 90% of the patients with ST-segment elevation, compared with approximately 50% of the non-ST-segment-elevation population. Our study extends the findings of Holmes et al. (16), focusing on patients without persistent ST-segment elevation and analyzing the impact of myocardial infarction at enrollment on the frequency and outcome of shock. The incidence of shock in our cohort was identical to that of the non-ST-segment-elevation population in GUSTO-IIb (16), and the mortality rates were similarly high. Together, these two studies indicate that once cardiogenic shock occurs in the setting of acute coronary syndromes, the prognosis is dismal regardless of the initial clinical and ECG features.
Shock most commonly developed >48 h after enrollment in our cohort (median of 94 h), with a longer lag period for patients who did not have myocardial infarction upon enrollment. (This is in striking contrast to the rapid development of shock among patients with persistent ST-segment elevation. Indeed, as we recently reported , the median time to shock after enrollment in the Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries study [GUSTO-I] was 11 h). In GUSTO-IIb (16), shock also developed significantly later among patients without ST-segment elevation (median time of 76.2 h vs. 9.6 h among patients with ST-segment elevation). Thus, the window of opportunity to identify the patients at risk of developing shock and to implement measures to possibly avert its development is significantly wider among patients who do not have persistent ST-segment elevation upon presentation.
Predictors of shock
The Braunwald classification is often used to stratify the risk of death or nonfatal reinfarction among patients presenting with unstable angina (17). Based on the Braunwald classification, the presence of profound ischemia or heart failure is associated with a high risk of subsequent adverse events. Our findings complement the Braunwald classification by stratifying the risk of developing shock among patients presenting with unstable angina or non-ST-elevation myocardial infarction. We developed a simple scoring system that was relatively accurate in predicting the risk of shock in the PURSUIT cohort. Moreover, we validated this algorithm in the GUSTO-IIb patients without persistent ST-segment elevation. Based on this algorithm, cardiogenic shock was predicted primarily by the patient’s age, the presence of ST depression in the initial ECG, and physical findings. We recently examined the predictors of shock developing after thrombolytic therapy for acute myocardial infarction with persistent ST-segment elevation and found that the patient’s age and physical findings accounted for >85% of the information needed to predict the development of shock (14). Thus, despite the many differences between patients with and without persistent ST-segment elevation who develop shock (16), the baseline demographic and clinical variables associated with the development of shock are similar.
Eptifibatide and outcome of shock
Two recent observational, nonrandomized reports have shown that treatment with the platelet glycoprotein IIb/IIIa inhibitor abciximab during angioplasty resulted in improved outcome of patients with shock complicating acute myocardial infarction (11,12). In our analysis, platelet glycoprotein IIb/IIIa blockade with eptifibatide did not attenuate the development of shock. However, treatment with eptifibatide was associated with 49% lower adjusted odds of dying once shock developed, regardless of whether patients presented with or without acute myocardial infarction. Eptifibatide exerted a beneficial effect despite the fact that shock developed during drug administration in only a small minority of patients, indicating that platelet glycoprotein IIb/IIIa blockade may have a sustained beneficial effect in this particular group of patients.
Several possible mechanisms exist by which eptifibatide exerts a salutary effect during shock. Platelet glycoprotein IIb/IIIa blockade has improved outcome after coronary angioplasty in both high- and low-risk subgroups (8–10). Because angioplasty was performed in only approximately 25% of shock patients in our cohort, it is unlikely that the beneficial effect of eptifibatide was only through the improvement of outcome after angioplasty. Topol et al. (18) demonstrated that the long-term benefit of acute platelet glycoprotein IIb/IIIa blockade during angioplasty extends to 36 months, perhaps by causing sustained “passivation” of the coronary artery (19). Thus, treatment with eptifibatide in our cohort possibly resulted in coronary “passivation” that rendered the milieu amenable to treatment once shock developed. This may be especially true for the coronary microcirculation (20); platelet glycoprotein IIb/IIIa blockade may relieve the microvascular obstruction in the shock patients, resulting in improved prognosis. This hypothesis is best supported by the exquisite sensitivity to platelet glycoprotein IIb/IIIa blockade of patients with unstable angina and elevated serum troponin T levels (21). These patients with elevated troponin T levels are presumed to have had microvascular obstruction. Finally, a direct protective effect of eptifibatide on the myocardium is also possible.
Few interventions, pharmacological or mechanical, have significantly and consistently altered the outcome of shock patients in recent years. Because this was a retrospective subgroup analysis, our results should be taken cautiously until proven by specifically designed studies. Nonetheless, the very poor prognosis of shock patients even in recent studies (22) underscores the possible significance of our findings.
We also found a statistically significant interaction (p = 0.032) between treatment and shock status on 30-day death, suggesting that eptifibatide was beneficial in reducing the incidence of death only among shock patients. However, because shock status is not a baseline variable but rather an outcome in itself, any conclusion with respect to the differential benefit of eptifibatide should be made with caution.
The current study demonstrates that shock occurred in 2.5% of patients with acute coronary syndromes without persistent ST-segment elevation, with a slightly greater incidence among patients who had myocardial infarction upon enrollment. Cardiogenic shock was predicted primarily by the patient’s age, the presence of ST-depression in the initial ECG, and physical findings upon presentation. Outcome was very poor once shock developed, with greater mortality among those presenting with myocardial infarction. Patients treated with eptifibatide seemed to have reduced adjusted odds of death from shock, suggesting a possible salutary effect of platelet glycoprotein IIb/IIIa blockade during shock. This finding derived from a post hoc analysis should be verified in specifically designed studies.
☆ Supported by COR Therapeutics (South San Francisco, California) and Schering Plough Research Institute (Kenilworth, New Jersey).
- GUSTO I
- Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries
- GUSTO IIb
- Global Use of Strategies to Open Occluded Coronary Arteries IIb
- creatine kinase-myocardial band
- Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy
- Received November 1, 1999.
- Revision received March 1, 2000.
- Accepted April 12, 2000.
- American College of Cardiology
- PARAGON Investigators
- Platelet Receptor Inhibition in Ischemic Syndrome Management in Patients Limited by Unstable Angina and Symptoms (PRISM-PLUS) Study Investigators
- Giri S.,
- Mitchel J.F.,
- Kiernan F.J.,
- et al.
- Hasdai D.,
- Califf R.M.,
- Thompson T.D.,
- et al.
- Holmes D.R. Jr.,
- Berger P.B.,
- Hochman J.S.,
- et al.
- ↵Braunwald E, Mark DB, Jones RH, et al. Unstable angina: diagnosis and management. Clinical guidelines number 10. AHCPR Publication No. 94-0602. Rockville, MD: Agency for Health Care Policy and Research and the National Heart, Lung, and Blood Institute, Public Health Service, U.S. Department of Health and Human Services. March 1994.
- Topol E.J.,
- Ferguson J.J.,
- Weisman H.F.,
- et al.