Author + information
- Christopher R Thompson, MD, CM, FACC∗,* (, )
- Christopher E Buller, MD, FACC‡,
- Lynn A Sleeper, ScD∥,
- Tracy A Antonelli, MPH∥,
- John G Webb, MD, FACC∗,
- Wael A Jaber, MD§,
- James G Abel, MD†,
- Judith S Hochman, MD, FACC¶,
- for the SHOCK Investigators
- ↵*Reprint requests and correspondence: Dr. C.R. Thompson, Director, Cardiology Clinical Research, St. Paul’s Hospital, Room 5134-1081 Burrard Street, Vancouver, B.C., Canada V6Z 1Y6
Our objective was to define the outcomes of patients with cardiogenic shock (CS) due to severe mitral regurgitation (MR) complicating acute myocardial infarction (AMI).
Methods for early identification and optimal treatment of such patients have not been defined.
The SHOCK Trial Registry enrolled 1,190 patients with CS complicating AMI. We compared 1) the cohort with severe mitral regurgitation (MR, n = 98) to the cohort with predominant left ventricular failure (LVF, n = 879), and 2) the MR patients who underwent valve surgery (n = 43) to those who did not (n = 51).
Shock developed early after MI in both the MR (median 12.8 h) and LVF (median 6.2 h) cohorts. The MR patients were more often female (52% vs. 37%, p = 0.004) and less likely to have ST elevation at shock diagnosis (41% vs. 63%, p < 0.001). The MR index MI was more frequently inferior (55% vs. 44%, p = 0.039) or posterior (32% vs. 17%, p = 0.002) than that of LVF and much less frequently anterior (34% vs. 59%, p < 0.001). Despite having higher mean LVEF (0.37 vs. 0.30, p = 0.001) the MR cohort had similar in-hospital mortality (55% vs. 61%, p = 0.277). The majority of MR patients did not undergo mitral valve surgery. Those undergoing surgery exhibited higher mean LVEF than those not undergoing surgery; nevertheless, 39% died in hospital.
The data highlight opportunities for early identification and intervention to potentially decrease the devastating mortality and morbidity of severe post-myocardial infarction MR.
Severe mitral regurgitation (MR) complicating acute myocardial infarction (AMI) is an important cause of hemodynamic instability and cardiogenic shock (CS). Nonrandomized series that have reported favorable outcomes after early mitral valve surgery have led to recommendations that early surgery is appropriate in such patients (1–10). However, these series are subject to powerful selection and publication biases. In the absence of randomized trials, reports characterizing large unselected cohorts of hemodynamically unstable patients with severe MR complicating AMI are needed to provide broader context, assist clinical decision making, and highlight areas for prospective investigation.
A pre-trial SHould we use emergently revascularize Occluded Coronaries in cardiogenic shocK? (SHOCK) Trial Registry prospectively collected data on 251 patients with CS at 19 centers between January 1992 and April 1993 (11). In that preliminary registry 19 (7.6%) patients with CS had acute severe MR or rupture of the ventricular septum, accounting for shock. It is interesting that only 8 of 19 patients had cardiac catheterization and only 4 of 19 had cardiac surgery. Mortality was 100% in the surgical group and 80% in those who did not undergo surgery. Thus, despite the previous favorable reports promoting surgical treatment of mechanical CS, the SHOCK Trial Registry indicated that a significant proportion of patients with mechanical causes of CS did not undergo surgery and that surgical mortality was high. The SHOCK Trial Registry provides an opportunity to re-examine these findings in a much larger unselected population. There were two aims of the SHOCK Trial Registry analysis: 1) to describe the cohort in the Registry with acute severe MR and to compare it with the cohort with predominant left ventricular (LV) failure not accompanied by severe MR or other mechanical complications; and 2) to compare the characteristics and outcome of surgically and nonsurgically-treated severe MR patients.
Patients with suspected CS complicating AMI, whether meeting strict trial criteria for CS or not, were prospectively registered. Thirty-six enrolling centers were initiated in a staggered fashion, and the first patient was enrolled in April 1993. A local discharge diagnosis of AMI and CS (DRG’s 410 and 785.51) constituted criteria for being registered. Acute severe MR, ventricular septal rupture, isolated right ventricular failure, cardiac tamponade or rupture, prior severe valvular heart disease and iatrogenic shock constituted etiologies of shock other than predominant LV failure and were SHOCK Trial clinical exclusion criteria. Importantly, patients with acute severe MR without CS were not consistently registered, because a diagnosis of suspected CS was required.
The SHOCK Trial Registry consisted of 1,190 patients. In order to compare well-categorized, distinct groups, five patients with shock due to predominant LV failure with moderate MR and 208 patients who had shock that was not caused by either MR or predominant LV failure, were excluded. The data set for this article therefore consists of 977 patients—98 patients who had CS with acute severe MR and a comparison group of 879 patients with predominant LV failure. The diagnosis of acute severe MR was made at the local SHOCK enrolling center.
Data were abstracted from the medical record by local SHOCK study coordinators who were centrally trained to complete standardized study report forms. Patient characteristics, MI characteristics, hemodynamics, utilization of medications and procedures, and vital status at hospital discharge were recorded. Cardiac catheterization and angiography reports were sent to the Clinical Coordinating Center for abstraction of information and completion of a standardized form. The following variables were recorded only on revised data collection forms and are therefore available from only two-thirds of the patient sample: LV ejection fraction, inotrope usage, the presence of ST segment elevation at shock, pulmonary edema, and the presence of rales.
Electrocardiogram (ECG) locations were defined according to the Global Utilization of Streptokinase and tPA for Occluded Coronary Arteries (GUSTO) 1 classification scheme; (i.e., V1 − V4 Anterior; II, III, AVF Inferior; V5 − V6 Apical; I, AVL Lateral; and V1 − V2 Posterior)(12).
Groups were compared using the Fisher exact test for categorical variables, the Wilcoxon rank-sum test for ordinal and non-normally distributed continuous variables and the Student t-test for normally distributed continuous variables. Covariate-adjusted in-hospital mortality by group was analyzed using logistic regression. In order to determine if group status was an independent predictor of mortality, a multivariate model was constructed by including all baseline patient characteristic variables with a univariate p value for group comparison of ≤0.20. All variables with a final p value of ≤0.05 were retained in the model. All analyses were conducted using the Statistical Analysis System (SAS, v. 6.12, SAS Institute, Inc., Cary, North Carolina).
The MR (n = 98) and LV failure (n = 879) groups had similar pre-existing cardiovascular conditions and major co-morbidities; however, a larger proportion of MR patients were female (52% vs. 37%, p = 0.004) and were admitted to the tertiary SHOCK Trial center via transfer (65% vs. 42%, p < 0.001) (Table 1). ⇓The ECG characteristics are described in Table 2. Both the presence of ST elevation at the time of shock diagnosis and the presence of ST elevation in at least two leads were less frequent in the MR cohort (41% vs. 63%, p < 0.001; 47% vs. 73%, p < 0.001). Among those with an identifiable index MI location by ECG, MR patients had a greater prevalence of inferior MI (55% vs. 44%, p = 0.039) and posterior MI (32% vs. 17%, p = 0.002) and a correspondingly lower prevalence of anterior MI (34% vs. 59%, p < 0.001). In those undergoing coronary angiography, the identity of the infarct artery was consistent with these observations.
Clinical and hemodynamic variables
Patients with MR had later shock (median 12.8 vs. 6.2 h post-MI, p < 0.001) (Table 2). Consistent with the known pathophysiology of severe MR, the MR cohort had higher median LV ejection fraction (0.37 [0.25, 0.48] n = 58 vs. 0.30 [0.20, 0.40] n = 335, p = 0.001) yet more often had clinical and radiographic evidence of pulmonary edema.
Patients with MR were significantly more likely to undergo all interventions except percutaneous transluminal coronary angioplasty (PTCA) (Tables 3 and 4). ⇓⇓Median time intervals from the onset of shock to right heart catheterization (3.7 h vs. 2.1 h, p = 0.030), left heart catheterization (5.8 h vs. 2.6 h, p = 0.009), and intra-aortic balloon pump (IABP) (5.0 h vs. 3.1 h, p = 0.036), while relatively short in both groups, were longer in the MR group. The median times from shock to bypass surgery were similar in the MR (16.6 h [5.1, 55.3], n = 36 and the LV failure groups (29.2 h [3.9, 115.0], n = 128), p = 0.397.
Crude (unadjusted) in-hospital mortality was similar for the two groups (MR vs. LV failure odds ratio [OR] 0.79; 55% for MR and 61% for LV failure, p = 0.277) and did not differ significantly after adjustment for patient outcome-related differences between the two groups—namely, transfer status, prior MI, and posterior MI (MR vs. LV failure OR 0.97, 95% confidence interval [CI] 0.60 to 1.56, p = 0.900). Pulmonary edema was not included as an adjustment factor, because it was considered to be a consequence of MR. Patients with MR had a longer median length of stay (10.7 [2.6, 20.6] days vs. 6.1 [1, 15.1] days following shock, p = 0.002). Among the survivors, 44 patients with MR were discharged after a median of 20.8 [12.3, 37.8] days, compared with 15.4 [10.1, 24.9] days for 343 LV failure patients p = 0.005.
Among the 98 patients with MR, data indicating whether or not valve surgery was performed were available for 94. Almost half (46%) underwent valve replacement (n = 37) or valve repair (n = 6). Six patients had mitral valve surgery without coronary artery bypass graft surgery (CABG), and 37 with CABG.
The characteristics of the patients with MR who underwent mitral valve surgery and those without mitral valve surgery were similar. Patients selected for surgery had lower median highest creatine kinase (932 [516, 1,875] vs. 1,659 [738, 3319], p = 0.030), and much higher in-hospital LV ejection fraction (40% [35, 52], n = 30 vs. 29% [24, 39], n = 28, p = 0.004) than those selected for nonsurgical care (Table 5). Intra-aortic balloon pump support was used in almost all the surgical patients but in less than half of the nonsurgical patients (Table 5). As expected, coincident revascularization with CABG was much more common in the surgical group; 4 of the 51 patients not undergoing valve surgery underwent CABG. Unadjusted mortality in those who underwent valve surgery was lower than in those who did not (40% vs. 71%; OR = 0.27, 95% CI = 0.12 to 0.64, p = 0.003). Of the few patient factors distinguishing the surgical and nonsurgical groups, only gender was even marginally related to mortality (better survival for men), and the adjusted odds ratio for death for surgical versus nonsurgical MR patients remained significant (OR 0.30, 95% CI 0.13 to 0.73, p = 0.008).
The primary reasons for not undertaking mitral valve surgery were 1) that the patient could not be stabilized or died awaiting surgery (half of patients) and 2) the presence of co-morbidities related to current illness or secondary to shock (one-third of patients).
The development of severe MR complicating AMI and leading to CS is widely recognized to be a medical catastrophe portending very poor prognosis. These data from the SHOCK Trial Registry will do little to alter that opinion. However, this observational study of a large and minimally selected cohort of patients with CS accompanied by acute severe MR provides insights not available in smaller or more selected series. The contemporaneous cohort of patients with predominant LV failure provides unique opportunities for comparison. Many baseline characteristics of the MR and LV failure groups were similar, reflecting a common risk profile for their underlying coronary artery disease and acute coronary syndrome. However, several potentially important differences emerged.
First, the distribution of electrocardiographic and angiographic infarct zones supports previous clinical and pathological work indicating that severe MR most often reflects necrosis of the posteromedial papillary muscle (13–15). Conversely however, anterior infarction was present in one-third of our population. Clearly, the presence of anterior infarction should not dissuade clinicians from considering acute MR when other clinical signs and symptoms suggest it. Furthermore, less than half the MR cohort displayed clinically recognized ST-segment elevation or new Q waves. It is a sobering observation that half or more of the instances of acute severe MR and shock develop in the absence of these markers of extensive necrosis. Nevertheless, despite the absence of ST elevation, complete vessel occlusion was likely present in many patients, posterior chest leads were not commonly used at the time of the SHOCK Trial Registry, and there were a substantial number of circumflex infarcts. Along with findings regarding the limited enzymatic elevations that characterized our MR cohort (median CPK elevation < 5 fold upper limit normal), these observations indicate that acute severe MR with shock is often a consequence of infarction or dysfunction of limited but exquisitely important myocardium.
The interesting observation of a higher prevalence of women with acute severe MR in our cohort, compared with patients having predominant LV failure, appears to confirm a similar observation by Tcheng et al. (5). There are further corresponding observations of an increased prevalence of women with acute severe MR causing shock, compared with other causes of shock, in the pre-SHOCK Trial Registry by Hochman et al. (11) and the main SHOCK Trial Registry (16). The Tcheng report predominantly included patients without shock. The effect, if true, therefore appears distinct from hemodynamic issues and implies gender-related factors specific to the mitral mechanism itself. Differences in patterns of vascular supply, collateralization, connective tissue, or in the clinical presentation and detection of MR are all possible explanations and warrant further study. The higher proportion of MR patients admitted via transfer likely reflects a belief among referring physicians that emergency surgery for acute severe MR is life-saving and indicated. Shock apparently due to a reversible mechanical cardiac defect such as acute MR seems intrinsically suited to emergent surgery. Until recently, it has been a less-than-intuitive concept that shock from LV failure, even when myocardial necrosis is well established, would benefit from emergency revascularization. This clinical predisposition to obtain emergency surgery for acute severe MR has been given additional credence from recommendations arising from nonrandomized case series.
Although the LV ejection fraction was higher in the MR group than in the LV failure group, it is important to recognize that the median LV ejection fraction of 37% represents marked impairment of LV systolic performance in the presence of MR. The higher ejection fraction reflects both a smaller infarct size and the reduced impedance to LV ejection contributed by ejection into the left atrium. The higher prevalence of pulmonary edema in the MR patients might be expected, considering the sudden regurgitant volume into the left atrium and pulmonary veins seen with acute severe MR.
Comparison of surgical versus nonsurgical treatment of severe MR
Because reports of nonrandomized studies suggest that surgery is desirable when shock results from mechanical disruption of the mitral apparatus (1–5), valve surgery is the treatment of choice in many centers. Our data, however, reveal the degree to which selection bias may have influenced the outcomes in such series, including our own. In this multicenter database involving numerous cardiologists and cardiovascular surgeons, we observed systematic surgical selection of patients with better LV function and smaller index infarctions. The effects of such selection on outcome was no doubt amplified 1) by the deferment or death of patients considered for surgery who were deemed too ill to operate on immediately and 2) by the exclusion of a cohort of gravely ill MR patients who may have died during or prior to transport, and were therefore never registered. Finally, the relative contribution of revascularization versus repair of the mitral valve is unclear.
Even among registered patients, fewer than half underwent surgery. A potential criticism of the low surgical treatment rate is that clinicians caring for these patients were unduly conservative when selecting patients for mitral valve surgery. This criticism should be tempered by the 40% hospital mortality rate of those who received mitral valve surgery (a mortality rate comparable to that reported in other series), as well as by the multicenter nature of the Registry. Perhaps, however, surgery should have been performed more promptly in the patients who were considered too ill or who died while waiting for surgery, who comprised over one-third of the MR cohort. In conjunction with approximately 10% of MR patients in the SHOCK Trial Registry who were treated medically because of co-morbidity arising secondary to CS, these observations highlight the need for very early recognition, support and decision making in any future prospective evaluation of emergency surgery for this condition. Some patients in the registry received PTCA rather than mitral valve surgery with CABG. Although a favorable response of acute severe MR to percutaneous transluminal coronary angioplasty (PTCA) has previously been reported (17–20), Tcheng observed that acute reperfusion with thrombolysis or angioplasty did not usually reverse MR in a group of 50 patients with moderately severe or severe MR receiving treatment for AMI (5). In that series, the early and late mortality was higher in the PTCA group than in those treated medically or with surgery. Nine patients with severe MR in the present registry were treated with PTCA alone, six of whom subsequently died. An additional five patients underwent CABG within 24 h of PTCA, and three of these patients died.
Physicians should consider clinically undetected MR in CS—particularly in women and those with non-ST elevation MI, inferoposterior MI and pulmonary edema. Despite the selection of less than half of severe MR patients for surgery, in-hospital surgical mortality was extremely high at 40%. Clearly, efforts are needed to enhance earlier recognition of severe MR complicating AMI, because earlier surgery (before shock develops) may lead to improved prognosis.
☆ Supported by Grants #HL50020-018Z and HL-49970 (1994–1999) from the National Heart, Lung, and Blood Institute, Bethesda, Maryland.
- acute myocardial infarction
- cardiogenic shock
- coronary artery bypass graft surgery
- electrocardiogram, electrocardiographic
- intra-aortic balloon pump
- left ventricular failure
- severe mitral regurgitation
- percutaneous transluminal coronary angioplasty
- SHould we emergently revascularize Occluded Coronaries in cardiogenic shocK?
- American College of Cardiology
- Hochman J.S,
- Boland J,
- Sleeper L.A,
- et al.
- Calvo F.E,
- Figueras J,
- Cortadellas J,
- Soler-Soler J
- ↵Hochman JS, Buller CE, Sleeper LA, et al., for the SHOCK Investigators. Cardiogenic shock complicating acute myocardial infarction—etiologies, management and outcome: a report from the SHOCK Trial Registry. J Am Coll Cardiol 2000;36:1063–70.
- Shawl F.A,
- Forman M.B,
- Punja S,
- Goldbaum T.S