Although tako-tsubo cardiomyopathy (TTC) is regarded as a reversible condition with favorable outcome, a malignant clinical course evolves in some subjects. In this single-institution experience, we describe the clinical profile of patients with adverse TTC outcome. A cohort of 249 consecutive patients with TTC was interrogated for those with acute unstable presentation during the first 24 hours. Forty-seven patients (19%) experienced early complicated clinical course with cardiac arrest in 9 (ventricular fibrillation, n = 4, pulseless electrical activity, n = 3, and asystole, n = 2) or marked hypotension in 38 (systolic blood pressure ≤90 mm Hg requiring vasopressors and/or balloon pump). Of the 47 patients, Killip class III to IV heart failure was present in 30 (64%). Despite treatment, 8 patients (3%; all women) died inhospital due to respiratory failure, cardiogenic shock, or anoxic brain injury. All 8 inhospital deaths occurred among the 47 patients with unstable presentation, including 2 after cardiac arrest and 6 with marked hypotension. Post-TTC event mortality for a period of 4.7 ± 3.4 years significantly exceeded that in a matched general US population (standardized mortality ratio 1.4; 95% confidence interval 1.1 to 1.9; p = 0.005) largely due to noncardiac co-morbidities. In conclusion, contrary to widespread perception, TTC is not an entirely benign and reversible condition. Among this large cohort, a high-risk subgroup was identified with cardiac arrest or hemodynamic instability, accounting for all hospital deaths. Hospital nonsurvivors had a variety of irreversible co-morbid conditions with the potential to compromise clinical status and adversely affect short-term survival. Long-term survival after hospital discharge was also reduced compared with the general population because of noncardiac co-morbidities.
Tako-tsubo (stress) cardiomyopathy (TTC) is a distinctive cardiac condition with unique left ventricular (LV) contraction profile, often triggered by stressful events, which has been considered reversible and associated with favorable outcome. However, our experience in a large, single-institution cohort has suggested that this benign characterization of TTC may not be as consistent as previously regarded. Therefore, we report here a high-risk subset of TTC patients with malignant clinical presentation. Such observations provide a more robust profile of the TTC clinical spectrum, useful in guiding effective management strategies for patients with this complex but incompletely understood condition.
Methods
From August 2001 to March 2012, 249 consecutive patients presented with a first TTC event to the Minneapolis Heart Institute at the Abbott Northwestern Hospital (Minneapolis, Minnesota). As previously described, patients with TTC shared the following features: (1) acute presentation typically with chest pain/discomfort or dyspnea, (2) systolic dysfunction with marked LV contraction abnormality, extending beyond the geographic territory of a single coronary artery, assessed with contrast LV angiography, cardiovascular magnetic resonance imaging (CMR), or 2-dimensional echocardiography, (3) absence of obstructive coronary stenosis (i.e., ≤50% luminal narrowing of the major coronary arteries by angiography) or evidence of acute plaque rupture. The left anterior descending coronary artery was carefully examined since myocardial ischemia in this distribution can cause an LV contraction abnormality mimicking TTC , (4) absence of myocarditis or ischemic transmural late gadolinium enhancement on CMR. The magnitude of acute heart failure was quantitated using the Killip classification. Selected data from 136 study patients have been published previously.
On admission, the ejection fraction (EF) and LV contraction pattern were assessed by LV angiography (n = 149), 2-dimensional echocardiography (n = 95), CMR (n = 2), computed tomography angiography (n = 2), or nuclear imaging (n = 1). LV contraction patterns were categorized: “apical ballooning” (abnormal contraction of mid and apical LV segments), “midventricular ballooning” (abnormal contraction limited to the mid-LV segments), and “basal ballooning” (abnormal contraction limited to the basal LV segments). CMR imaging was performed after admission (n = 154) at the discretion of the attending cardiologist. Details of CMR imaging methods have been recently published.
Two-dimensional echocardiography included analysis of LV and right ventricular (RV) wall motion, degree of mitral regurgitation, and mitral valve systolic anterior motion (SAM). Left ventricular outflow tract gradients were estimated with continuous wave Doppler. Post-hospital EF was measured with echocardiography or CMR.
The clinical status of the 249 study patients after the TTC event was assessed as of September 1, 2014, by telephone interview, clinic visit, review of electronic medical records, or interrogation of the US Social Security Death Index. Follow-up was achieved for 233 patients (94%) for a period of 4.7 ± 3.4 years (median 4.2; range 0.1 to 12.8 years). For patients with TTC discharged from the hospital after the initial event, the fraction surviving at each follow-up time was estimated using the Kaplan-Meier method. The expected fraction surviving at each time was computed by assigning to each patient the probability of surviving after presentation, appropriate to patient age at diagnosis and gender, and based on the US Census data. Actual and expected surviving fractions were compared using the 1-sample log-rank test, which also provides an estimate and confidence interval (CI) for the standardized mortality ratio and 95% CI. All computations used the “survival” package (version 2.34-1) of the R Software System, version 2.7.2 R (Development Core Team 2008, R Foundation for Statistical Computing, Vienna, Austria).
Data are displayed as mean and SD for continuous variables with number and percentage for categorical variables. Categorical variables were analyzed using the Pearson’s chi-square or Fisher’s exact tests. Continuous variables were assessed using the Student’s t test. A value of p <0.05 was considered significant, p values are 2 sided where appropriate.
Variables with p <0.05 for univariable associations (e.g., gender, heart rate, EF, peak troponin, smoker, and physical TTC trigger) were entered into stepwise multivariable logistic regression models. In this model, variables with p <0.05 were retained; only patients with complete data on all covariates were included in the multivariable analysis. Statistical calculations and plots were performed with Stata 11.2 (StataCorp. 2009. Stata Statistical Software: Release 11 . College Station, TX: StataCorp LP). Institutional review board approval was obtained for data collection, follow-up, and data analysis.
Results
Forty-seven of the 249 study patients (19%) experienced an early complicated course with cardiac arrest in 9 patients (ventricular fibrillation, n = 4, pulseless electrical activity, n = 3, and asystole, n = 2) or marked hypotension in 38 patients (systolic blood pressure ≤90 mm Hg requiring intravenous vasopressors and/or intra-aortic balloon pump [IABP]; Figure 1 , Table 1 ). Of these 47 patients, Killip class III to IV heart failure was present in 30 (64%), including 6 of those with cardiac arrest.
| Variable | With (n = 47) | Without (n = 202) | p-value |
|---|---|---|---|
| Age (years) | 67 ± 14 | 68 ± 13 | 0.85 |
| Male | 8 (17%) | 7 (3%) | < 0.001 |
| Heart rate (beats/minute) | 103 ± 28 | 92 ± 22 | 0.004 |
| Ejection fraction (%) | 25 ± 9 | 33 ± 10 | < 0.001 |
| Peak troponin (ng/ml) | 0.84 ± 0.98 | 0.49 ± 0.47 | 0.016 |
| Beta-blocker (pre-admission) | 13 (28%) | 45 (22%) | 0.39 |
| Diabetes mellitus | 4 (9%) | 29 (14%) | 0.31 |
| History of systemic hypertension | 20 (43%) | 115 (57%) | 0.13 |
| Current smoker | 14 (30%) | 33 (16%) | 0.022 |
| ST-segment elevation (admission ECG) | 23 (49%) | 74 (37%) | 0.12 |
| Physical stressor | 31 (66%) | 82 (41%) | 0.002 |
| LV apical ballooning | 27 (57%) | 116 (57%) | 0.47 |
| LV mid-ventricular ballooning | 18 (38%) | 83 (41%) | |
| LV basal ballooning | 2 (4%) | 3 (2%) | |
| Length of hospital stay (days) | 13 ± 12 | 5 ± 5 | < 0.001 |
| Hospital mortality | 8 (17%) | 0 | < 0.001 |
When the 47 patients presenting with hemodynamic instability or cardiac arrest were compared with the other 202 patients with TTC, more markers of clinical severity were evident, including higher heart rates, lower EF, and higher peak troponin and also with a greater proportion of male gender and presence of physical stressor ( Table 1 ). In this subset, a physical stressor was present in 7 of 8 men (88%) versus 24 of 39 women (62%; p = 0.16).
At multivariable analysis, variables associated with unstable presentation included male gender (p = 0.003), physical trigger (p = 0.007), and magnitude of troponin elevation (p = 0.003). The remaining 202 patients all survived hospitalization and were stable at discharge.
Thirty-eight of the 249 patients (15%) demonstrated an unstable clinical presentation within 24 hours of admission with acute hypotension, systolic blood pressure ≤90 mm Hg (range 59 to 90 mm Hg) requiring acute intervention with vasopressor drugs, and/or IABP. Clinical presentation for these 38 patients included dyspnea (n = 16), chest pain (n = 15), altered consciousness or focal neurologic symptoms (n = 4), syncope (n = 2), or hypotension during general anesthesia (n = 1).
Of the 38 patients, 24 were in cardiogenic shock or with pulmonary edema (Killip class III or IV heart failure) and 14 were Killip class I or II heart failure ( Figure 1 ). Lowest systolic blood pressure in these 38 patients was 59 to 79 mm Hg (n = 26) or 80 to 90 mm Hg (n = 12), and mechanical ventilation was necessary in 24 patients.
Acute treatment with intravenous pharmacologic agents and/or IABP was judged obligatory in the 38 patients with unstable hemodynamic status, including administration of catecholamine drugs: dopamine (n = 23), phenylephrine (n = 19), norepinephrine (n = 16), dobutamine (n = 11), epinephrine (n = 1), or other agents: vasopressin (n = 7) and milrinone (n = 1). A single drug was sufficient to support blood pressure in 15, 2 drugs were required in 13, and ≥3 drugs in 10; an IABP was used in 8 patients. Ten of the 38 patients (26%) had SAM, and each was receiving a catecholamine drug (2 also with IABP), including 7 with mitral-septal contact and estimated gradients by Doppler echocardiography ≥30 mm Hg under conditions at rest, range 30 to 115 mm Hg (average 64 ± 39 mm Hg). Of these 10 patients, 5 also had basal ventricular septal hypertrophy (18 to 21 mm thickness) without uncontrolled systemic hypertension, consistent with hypertrophic cardiomyopathy. The remaining 5 patients had normal LV wall thickness and SAM resolution after TTC treatment.
Nine patients presented in cardiac arrest due to ventricular fibrillation (n = 4), pulseless electrical activity (n = 3), and asystole (n = 2; Figures 1 and 2 , Table 2 ). In 8 patients, these events occurred immediately before admission and in the other patient in the cardiac catheterization laboratory before coronary angiography. After resuscitation and stabilization, 6 of these patients were judged to be in Killip class III or IV heart failure and the other 3 patients were Killip class I or II. Cardiac arrest patients were younger than the 38 patients with unstable hypotension (58 ± 10 vs 69 ± 14 years; p = 0.03). Intravenous epinephrine was administered during resuscitation in 6 patients.
| Patient | Age (yrs)/ sex | Rhythm | Initial ECG | Vasopressor IABP | Stressor | Therapeutic Hypothermia | Irreversible Anoxia | Survival | Ejection Fraction % (Admit,Discharge) | LOS (days) | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| STE | QTc (ms) | ||||||||||
| 1 | 45/F | PEA | + | 500 | + | Drug overdose | + | 0 | + | 30, 55 | 10 |
| 2 | 48/F | VF | 0 | 441 | + | Migraine headache | + | 0 | + (ICD) | 25, 58 | 7 |
| 3 | 50/M | VF | 0 ∗ | 454 | + | Pulmonary embolism | + | 0 | + | 15, 65 | 21 |
| 4 | 51/F | VF | + | 420 | + | Sibling death | 0 | 0 | + (ICD) | 15, 60 | 34 |
| 5 | 58/F | PEA | + | 440 | + | COPD | + | + | 0 | 35, deceased | 3 |
| 6 † | 59/M | VF | + | 523 | 0 | Panic attack | 0 | 0 | + (ICD) | 30, 55 | 8 |
| 7 †† | 64/F | Asystole | + | 408 | + | COPD | 0 | 0 | + | 20, 70 | 12 |
| 8 | 69/F | PEA | + | 459 | + | Subarachnoid hemorrhage | + | + | 0 | 35, deceased | 2 |
| 9 | 76/F | Asystole | 0 | 477 | 0 | COPD | 0 | 0 | + | 30, 60 | 27 |
† Occurred in the emergency department.
†† Occurred in catheterization laboratory before catheter insertion; all other cases are out-of-hospital.
Of the 9 patients with cardiac arrest, 5 had significant anoxic brain injury (reversible in 3), each of whom was treated with therapeutic hypothermia. Initial electrocardiogram demonstrated ST-segment elevation in 5 patients and QTc prolongation ≥500 ms in 2 patients; torsades de pointes ventricular tachycardia was absent in all.
Important co-morbid medical conditions were present in 6 of the 9 cardiac arrest patients including acute respiratory failure from advanced chronic obstructive lung disease (n = 3), subarachnoid hemorrhage (n = 1), submassive pulmonary embolism (n = 1), and drug overdose (opiates, alcohol, and citalopram) (n = 1). Circulatory support for hypotension or acute heart failure was required in 7 of these 9 patients including intravenous vasopressors (n = 6) and IABP (n = 2).
After cardiac arrest, 7 of the 9 patients survived to discharge without permanent neurologic or cardiac disability. Initial EF improved from 24 ± 7% to 60 ± 5% at follow-up (p <0.001). In 3 patients (each with ventricular fibrillation), a secondary prevention implantable cardioverter-defibrillator was placed before discharge, and none have experienced an implantable cardioverter-defibrillator intervention. After hospital discharge, 3 additional cardiac arrest patients died: at 1 month (chronic obstructive pulmonary disease [COPD]), at 3 months (suicide), and at 4 years (COPD). The 4 surviving cardiac arrest patients have had no recurrent cardiac events at follow-up for a period of 2 to 10 years.
Of the 249 study patients, 241 survived their acute event and were discharged from the hospital, with EF improved from 32 ± 10% on admission to 57 ± 10% at follow-up (p <0.001). The 8 inhospital deaths occurred in the 47 patients with hemodynamically unstable TTC presentation, including 2 with cardiac arrest (pulseless electrical activity) and 6 with marked hypotension. Mechanism of death included cardiogenic shock (n = 4), respiratory failure from pneumonia (n = 2), and anoxic encephalopathy (n = 2). Of the 39 survivors with unstable TTC presentation, EF improved from 24 ± 8% on admission to 59 ± 6% at follow-up (41 ± 81 days; p <0.001).
Nonsurvivors and survivors differed significantly with respect to the presence of major co-morbid conditions (8 [100%] vs 12 [31%]; p <0.001) and body weight (57 ± 17 vs 72 ± 18 kg; p = 0.04). However, no significant differences were present with respect to: age (75 ± 13 vs 66 ± 14 years; p = 0.09), female gender (8 [100%] vs 31 [79%]; p = 0.16), initial heart rate (114 ± 21 vs 101 ± 29 beats/min; p = 0.27), EF (30 ± 11 vs 24 ± 8%; p = 0.09), presence of physical stressor (7 [88%] vs 24 [62%]; p = 0.16), apical ballooning pattern (6 [75%] vs 21 [54%]; p = 0.27), RV segmental wall motion abnormality (4 [50%] vs 8 [21%]; p = 0.08), Killip class III to IV heart failure (6 [75%] vs 24 [62%]; p = 0.47), LV end-diastolic pressure (21 ± 5 vs 23 ± 7 mm Hg; p = 0.5), moderately severe or severe mitral regurgitation (2 [25%] vs 6 [15%]; p = 0.51), vasopressor drug use (8 [100%] vs 36 [92%]; p = 0.51), or IABP use (1 [13%] vs 9 [23%]; p = 0.45).
Average age at inhospital death was 75 ± 13 years (range 58 to 92), all were women (7 of 8 with a physical stressor), and admission EF 30 ± 11%. Notably, each of the 8 patients who died had an irreversible co-morbid condition judged to be an important contributor to demise, including cardiac arrest with anoxic encephalopathy (n = 2), subarachnoid hemorrhage without cardiac arrest (n = 1), particularly advanced age >90 years (n = 1), dementia (n = 1), lung cancer (n = 1), Crohn’s disease with short bowel syndrome (n = 1), and meningioma with cerebral edema (n = 1; Table 3 ). Four of these patients also had pneumonia.
