Author + information
- Received November 4, 2004
- Revision received January 7, 2005
- Accepted January 11, 2005
- Published online July 19, 2005.
- Lynn A. Sleeper, ScD⁎,⁎ (, )
- Krishnan Ramanathan, MD†,
- Michael H. Picard, MD‡,
- Thierry H. LeJemtel, MD§,
- Harvey D. White, MD∥,
- Vladimir Dzavik, MD¶,
- Deborah Tormey, RN#,
- Nancy E. Avis, PhD⁎⁎,
- Judith S. Hochman, MD†,
- SHOCK Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Lynn A. Sleeper, New England Research Institutes, 9 Galen Street, Watertown, Massachusetts 02472.
Objectives Our goal was to describe the functional status of cardiogenic shock survivors, identify the correlates of cardiogenic shock, and compare global quality of life and functional status of patients randomly assigned to treatment with emergency revascularization (ERV) versus initial medical stabilization (IMS).
Background Historically, the hospital survival rate of patients with cardiogenic shock complicating acute myocardial infarction (MI) has been very low. Shock survivors are salvaged from a critically ill state, and their later functional status is not well documented. The SHould we emergently revascularize Occluded Coronaries for cardiogenic shocK (SHOCK) trial showed significantly improved one-year survival after ERV compared with IMS.
Methods The SHOCK trial survivors completed interviews at 2 weeks after discharge and at 6 and 12 months after MI. Functional status assessment included the Multidimensional Index of Life Quality and New York Heart Association (NYHA) congestive heart failure functional class.
Results Eighty-seven percent of one-year survivors of the SHOCK trial were in NYHA functional class I or II. Between two weeks after discharge and one year after MI, improvement was similar in the two treatment groups (18% overall), but fewer patients remained stable (44% vs. 71%), and more patients worsened or died (34% vs. 15%) in the IMS group compared with those assigned to ERV. Assignment to ERV was the only independent predictor of outcome at one year.
Conclusions Although one-year mortality after ERV is still high (54%), most survivors have good functional status. The ERV patients have a lower rate of deterioration than IMS patients. The level of recovery for shock patients undergoing ERV is similar to that of historical controls not in cardiogenic shock undergoing elective revascularization.
Since the mid-1990s, the in-hospital mortality rate for patients with acute myocardial infarction (MI) complicated by cardiogenic shock has ranged from 60% to 70% (1,2). The SHould we emergently revascularize Occluded Coronaries for cardiogenic shocK (SHOCK) trial showed that additional lives can be saved with emergency revascularization, reducing the mortality rate to 51% at six months (3). However, because of the extensive myocardial damage caused by the disease process before revascularization, shock survivors may have severely limited long-term functional status. Salvage of critically ill patients by revascularization improves survival, but it is not known whether survivors are restricted in their activities by symptoms of severe heart failure and whether they have a poor quality of life.
We report the effect of a strategy of emergency revascularization (ERV) for cardiogenic shock complicating acute MI on functional status and quality of life of patients enrolled in the SHOCK trial. The findings of patient interviews from 2 weeks after discharge and 6 and 12 months after MI are summarized. The association between functional status and long-term survival is explored, and the patterns of recovery for shock patients undergoing ERV and a historical sample of patients not in cardiogenic shock undergoing elective revascularization are compared.
Patients were enrolled from April 1993 to November 1998 at 30 international sites via computerized telephone randomization. All patients (or their surrogate) signed an informed consent form, approved by the institutional review board or ethics committee of the clinic site.
The SHOCK trial was a randomized, unblinded, multi-center trial comparing the effects of ERV and initial medical stabilization (IMS) strategies with respect to 30-day, 6-month, and 1-year survival (4). Patients with acute MI who developed cardiogenic shock caused by predominant left ventricular failure within 36 hours of MI onset were eligible for the trial if the electrocardiogram showed ST-segment elevations or Q waves, posterior infarction, or new or presumably new left bundle block. Randomization had to be accomplished within 12 h of shock diagnosis. Strict clinical and hemodynamic criteria for shock were required, including sustained hypotension (systolic blood pressure <90 mm Hg or need for vasopressors), tissue hypoperfusion, pulmonary capillary wedge pressure ≥15 mm Hg, and cardiac index ≤2.2 l/min/m2. Exclusion criteria included other etiologies of cardiogenic shock, including mechanical complications, prior cardiomyopathy or valvular heart disease, severe systemic illness, and inability to perform cardiac catheterization or known unsuitable coronary anatomy.
Patients randomized to a strategy of attempted ERV were required to undergo either angioplasty or coronary artery bypass graft surgery, if feasible, within six hours of randomization. Patients randomized to a strategy of IMS could be considered for angioplasty or coronary artery bypass graft surgery 54 h after randomization. Intra-aortic balloon counterpulsation was recommended for all patients (used in 86%), and thrombolytic therapy was strongly recommended for the IMS group, unless an absolute contraindication was present.
The primary study end point of the SHOCK trial was 30-day mortality. Secondary end points of the trial presented in this article include six-month and one-year survival, functional status, and quality of life.
The SHOCK trial enrolled 302 patients. One hundred fifty-two patients were randomly assigned to ERV, and 150 were assigned to IMS. Figure 1summarizes the patient sample analyzed for this report. There were 158 in-hospital deaths and 2 patients who survived after heart transplantation during the shock hospitalization. The sample for this analysis is composed of the remaining 142 patients (77 assigned to ERV and 65 assigned to IMS) who were eligible for follow-up interviews. Sixteen of the hospital survivors had no interviews completed after discharge (11%). One patient in the IMS group was lost to follow-up at one year.
Data collection after hospital discharge
For patients discharged alive, up to three follow-up interviews to collect information on quality of life and physical function were conducted: 1) 2 weeks after discharge (window 10 days to 2 months, median 3 weeks), 2) 6 months after index MI (window 3 to 9 months, median 6.6 months), and 3) 1 year after index MI (window 9 to 18 months, median 12.6 months). Global quality of life was measured using three Likert scale questions, including the Andrews Ladder of Life (5) (Appendix). Functional status according to the New York Heart Association (NYHA) congestive heart failure functional classification system was measured using four standardized questions. For 35 patients who were randomized early in the study, NYHA functional classification was assessed in at least one interview using questions from the modified Sickness Impact Profile. Functional status was also measured using the Multidimensional Index of Life Quality (MILQ), a validated instrument designed specifically to assess quality of life in patients with cardiovascular disease (6). Each MILQ domain score was based on four questions answered using a seven-point Likert scale (Appendix). The MILQ and the life satisfaction questions are valid only if answered by the patient, and therefore were not administered to proxy respondents (40, or 14% of the 287 interviews obtained within window, with no significant difference between treatment groups).
There were 16 patients eligible for follow-up (11%) who did not complete any post-discharge interviews. Five of the 16 came from one clinic site, and 6 died in the year after MI. The non-interviewed cohort was similar to the group of 126 patients who had at least one follow-up interview with respect to treatment group assignment (50% vs. 55% in the ERV group, p = 0.79). Seventy percent (92 patients) of six-month survivors and 75% (89 patients) of one-year survivors were included in the six-month and one-year analyses, respectively. Seven of the interviews classified as unavailable were actually completed, but were outside of the time window specified for analysis. Inclusion of the out-of-window interviews (2 to 4 months after discharge; and 9 to 10 months rather than 6 months post-MI) did not significantly alter the study findings.
Survival end points were calculated from the time of randomization. Per protocol, follow-up data were not collected for two survivors who underwent heart transplantation during the shock hospitalization. The Wilcoxon rank sum test was used to compare the distribution of the global quality of life questions by treatment group at a given time point and by one-year vital status. The Mantel-Haenszel test for linear trend and the Fisher Exact test were used to compare the distributions of NYHA functional class (I to IV) by treatment group and by survival status at one year within treatment group. Logistic regression modeling (7) was used to identify independent correlates of one-year outcome (NYHA functional class I/II vs. III/IV/dead). The Wilcoxon rank sum test was used to compare physical function domain scores by treatment group. A model-based estimate of NYHA congestive heart failure functional classification was used for 35 patients for whom standardized NYHA functional classification responses were not available. The model was developed from patients with measured NYHA functional class, using logistic regression with questions from the modified Sickness Impact Profile as predictors. All statistical analyses were conducted with SAS software (Statistical Analysis System for Windows version 8.2, 2001, SAS Institute, Inc., Cary, North Carolina).
At randomization, the mean age of the 126 patients with at least one post-discharge interview was 63.4 ± 11.1 years. The two treatment groups were similar with respect to age and other baseline characteristics (Table 1):69% were male, 28% had prior MI, 30% had diabetes mellitus, and 42% had a history of hypertension. Ninety-three percent of the ERV and 40% of the IMS patients in the interview cohort underwent a revascularization attempt during the shock hospitalization. The post-discharge revascularization rates at one year were similar (4%) in the ERV and IMS groups: three ERV patients underwent percutaneous coronary intervention or coronary artery bypass graft after discharge, and two IMS patients underwent coronary artery bypass graft (one of whom underwent percutaneous coronary intervention during the hospitalization for shock). Of the 27 patients with at least one proxy interview, 3 were noted to have had a stroke. At one year, the 9 proxy respondents (9% of 98 interviews at one year) indicated 1 patient unwilling to be interviewed, 4 unable to answer questions or mostly bedridden, and 4 in the hospital (1 non-cardiac, 3 cardiac).
NYHA functional classification
At two weeks after discharge, the distributions of NYHA functional class by treatment assignment were similar (p = 0.131): 62.5% of 48 patients assigned to IMS and 75.9% of 58 patients assigned to ERV were class I or II (Table 2).Figure 2shows that the proportion of patients in NYHA functional class I or II remained fairly constant over time. However, the proportion of NYHA functional class III and IV patients decreased over time as deaths occurred, with proportionately more deaths in the IMS group (treatment group comparison of class I or II vs. III or IV vs. deceased: p = 0.035 at six months post-MI and p = 0.014 at one year post-MI). The finding of more one-year survivors in the ERV group compared with the IMS group without a proportionate increase in the number of NYHA functional class III or IV patients who underwent ERV indicates a marked improvement in survival without disability. Table 3shows the transition rates from functional class at two weeks after discharge to one-year post-MI. Similar proportions of patients improved in both treatment groups (22% in IMS and 15% in ERV group), but fewer patients remained stable (44% vs. 71%) and more patients worsened (34% vs. 15%) in the IMS group. Among those alive at each follow-up, however, the distributions of NYHA functional class for the two treatment groups were similar (Table 2).
Among one-year survivors, the MILQ physical function domain scores of the two treatment groups (Table 4)were similar at one year (19.1 vs. 19.3). Table 4also shows mean MILQ scores separately for angioplasty and coronary artery bypass graft surgery patients in the ERV group. An increase in physical functioning over one year was observed for coronary artery bypass graft patients. Physical function scores for angioplasty patients were highest at six months.
At two weeks after discharge, patients assigned to the ERV group had better quality of life scores in three of the nine MILQ domains: mental health (21.7 ± 4.8 vs. 19.3 ± 5.8, p = 0.048), physical health, measuring energy and pain levels (19.6 ± 5.1 vs. 15.9 ± 5.8, p = 0.003), and financial stability (20.2 ± 6.9 vs. 16.3 ± 7.4, p = 0.019). The mean MILQ physical function domain score (measuring ability to perform physical tasks) of patients assigned to the ERV group was similar to that of their IMS counterparts at two weeks (16.4 ± 5.6 vs. 14.4 ± 7.4, p = 0. 198). At later time points, the MILQ domain scores of the ERV and IMS surviving cohorts were similar.
Overall life satisfaction
The distributions of responses regarding global quality of life from one-year survivors were similar for the two treatment assignments at each time point. The Andrews Ladder of Life (1 = worst possible life, 10 = best possible life) had median scores of 5.0 for IMS patients (n = 31) and 7.0 for ERV patients (n = 49) at two weeks after discharge (p = 0.256). The median score for both groups was 7.0 at both six months and one year post-MI. The summary health question (“In general, how would you say your health is right now?”), which ranged from 1 denoting excellent health to 5 denoting poor health, had two-week median scores of 3.5 and 3.0 (good health) for IMS and ERV patients, respectively. The median score continued to be 3.0 for both groups in later interviews. The overall life satisfaction question ranged from 1 denoting complete dissatisfaction to 7 denoting complete satisfaction. Median scores for this question were 5.0 and 6.0 for IMS and ERV patients, respectively, at two weeks after discharge (p = 0.142) and six months post-MI, and 5.0 for both groups at one year.
Prediction of long-term outcome
It was of interest to identify factors from the time of hospitalization to discharge that were correlated with a poor outcome (defined as NYHA functional class III or IV or death) at one year post-MI. Table 5compares 74 patients in NYHA functional class I or II at one year post-MI with 37 patients with poor outcome at one year post-MI. The only statistically significant factors were assignment to the emergency revascularization group (62.2% of NYHA functional class I or II vs. 40.5% of class III or IV or deceased patients, p = 0.043) and NYHA functional class at two weeks after discharge (p = 0.045; 22 patients with missing two-week data). Other factors with a suggestive association were a history of hypertension (p = 0.067), chest pain at presentation (p = 0.084), index MI with new left bundle branch block (p = 0.077), and systolic blood pressure at the time of shock (p = 0.075). A multivariate logistic regression model (N = 106) examining these factors showed that assignment to emergency revascularization (odds ratio [OR] 0.31, 95% confidence interval 0.13 to 0.76, p = 0.011), systolic blood pressure from the time of shock onset (OR per 10-mm Hg increase 0.78, 95% confidence interval 0.60 to 1.02, p = 0.066) and chest pain at presentation (OR 3.96, 95% confidence interval 0.77 to 20.30, p = 0.099) were the only independent predictors at the 0.10 significance level of patient status one year after MI.
Patients who initially survive cardiogenic shock, typically a lethal complication of an acute MI, have improved long-term survival with good functional status. A strategy of early revascularization is an independent strong predictor of survival with good functional status at one year.
The SHOCK trial showed that a strategy of early revascularization confers a one-year survival advantage over initial medical stabilization with delayed revascularization for patients with cardiogenic shock complicating acute MI (3). The results of this trial confirmed previous non-randomized study findings suggesting improved survival in cardiogenic shock with early revascularization. The functional capacity and life satisfaction of patients with established severe heart failure is poor (8) but is not known for survivors of cardiogenic shock. Previous studies lacked the necessary study design to assess long-term functional status and life satisfaction. They were non-randomized, retrospective, and confounded by selection bias with referral of less ill patients for revascularization. Small series of cardiogenic shock patients supported the use of ventricular assist devices using surrogates such as “return to work” (9). The SHOCK trial included functional status and quality of life as secondary end points to be evaluated in all hospital survivors.
Early revascularization in cardiogenic shock is thought to relieve ischemia and thereby interrupt the cycle of events that results in further myocyte necrosis and depressed myocardial contractility. Furthermore, early revascularization may influence the remodeling process that follows MI. However, shock classically occurs in patients with severe left ventricular damage (10). The substantial survival advantage of early revascularization could theoretically result in the salvage of critically ill patients with severe left ventricular dysfunction and poor functional status. Reassuringly, this was not the case. Eighty-three percent of one-year survivors (85% in the ERV group and 80% in the IMS group) reported minimal functional limitation (NYHA functional class I or II). This good outcome may be in part attributable to the fact that these patients have not experienced the years of physical deconditioning that are characteristic of those with chronic heart failure, and shows the dissociation between systolic ventricular performance and functional status.
The SHOCK trial patient physical function domain scores from the MILQ scale are useful for comparison with other cardiac patient samples. The authors of the MILQ scale studied a cohort of 22 patients undergoing scheduled percutaneous revascularization and 66 undergoing coronary artery bypass surgery (5). This cohort was 61 ± 10 years old and 81% male. Over one-half (53%) had a previous MI, 23% had diabetes, and 46% had hypertension. The level and timing of recovery of SHOCK trial patients assigned to emergency angioplasty was similar to that of the historical cohort with scheduled angioplasty (Table 4). The SHOCK trial emergency bypass surgery patients also had recovery of function similar to that of patients with scheduled procedures, but with a six-month delay; that is, the one-year scores of SHOCK trial patients are similar to the six-month scores of the scheduled bypass surgery patients.
Patients assigned to emergency revascularization who were in NYHA functional class I or II at two weeks after discharge remained remarkably stable over one year. In contrast, patients assigned to initial medical stabilization who were in NYHA functional class I or II at two weeks had a much higher rate of deterioration. The death rate in shock patients with low two-week functional status assigned to IMS was similar to that of patients with stage D heart failure (11).
At two weeks after discharge, mental health status was significantly higher in patients who were assigned to emergency revascularization. The association is likely even stronger than estimated because incapacitated patients or those living in a nursing home did not complete the MILQ questionnaire. These excluded patients tended to more often be from the IMS group, presumably had lower mental health status, and had a high post-discharge death rate of 36%. The relationship between life satisfaction and risk of death is consistent with previous findings showing the independent prognostic value of life satisfaction over traditional markers including left ventricular ejection fraction (12,13).
A strategy of early revascularization, higher systolic blood pressure on support measures after shock onset, and the absence of chest pain at presentation were found to be independent predictors of a good one-year outcome. The systolic blood pressure at shock onset reflects residual left ventricular systolic reserve and the compensatory vasoconstrictive response of the systemic vasculature. Inflammatory mediators and inducible nitric oxide synthase expression affect the latter, and these mediators may also play a role in the outcome (14). Early revascularization in the acute setting, despite its acute risks, is a superior strategy with respect to one-year survival and functional status. These findings should help dispel the bias against early revascularization in critically ill patients with moderate to severe left ventricular dysfunction. Patients who remain severely limited in their functional capacity after revascularization should be monitored closely with early referral to additional therapies such as wearable left ventricular assist devices, to cardiac resynchronization therapy, and to participation in protocols evaluating newer therapies (gene and cell therapies).
This analysis of functional class in the SHOCK trial has several limitations. Six-month and one-year interviews within the specified time window were not available for approximately 25% of patients. However, age is a potent predictor of physical functioning, and in this study the age distribution of patients with and without an interview was similar. In addition, 16% of the two-week interviews were obtained by proxy when the patient was incapacitated or unwilling to be interviewed, and therefore it was not appropriate to collect quality-of-life data using the MILQ questionnaire. Therefore, the quality-of-life scores reported here may be biased upward.
At one year, 83% of patients discharged alive after an initial hospitalization for cardiogenic shock were in NYHA functional class I or II. In particular, patients who underwent early revascularization had better mental health immediately after discharge and were protected against deterioration in their functional status over time. Moreover, survivors of early revascularization after shock can have recovery of physical function similar to that of patients with elective revascularization. The intense use of resources in the salvage of critically ill patients is justified.
We acknowledge the dedication of the SHOCK trial coordinators and investigators for their work in maintaining contact with the survivors of the SHOCK trial, and Kevin Smith, MS, for providing the historical data in Table 4.
Appendix Supplementary data
For the Andrews’ Ladder of Life, global quality of life measurements, and other pertinent information, as well as a list of the Committee Members, Principal Investigators, and Study Coordinators of the SHOCK trial, please see the July 19, 2005, issue of JACCat www.onlinejacc.org.
This work was completed with support from research grants R01 HL49970 and R01 HL50020 from the National Heart, Lung, and Blood Institute of the National Institutes of Health, Bethesda, Maryland.
- Abbreviations and Acronyms
- emergency revascularization
- initial medical stabilization
- myocardial infarction
- Multidimensional Index of Life Quality
- New York Heart Association
- SHould we emergently revascularize Occluded Coronaries for cardiogenic shocK
- Received November 4, 2004.
- Revision received January 7, 2005.
- Accepted January 11, 2005.
- American College of Cardiology Foundation
- Hochman J.S.,
- Sleeper L.A.,
- Godfrey E.,
- et al.
- Andrews F.M.,
- Withey S.B.
- Rosner B.
- Ruzevich S.A.,
- Kanter K.R.,
- Pennington G.,
- et al.
- Alonso D.R.,
- Scheidt S.,
- Post M.,
- et al.
- CONSENSUS Trial Study Group
- Spertus J.A.,
- Jones P.,
- McDonell M.,
- et al.