Takotsubo-Like Myocardial Dysfunction in Ischemic Stroke
A Hospital-Based Registry and Systematic Literature Review
Background and Purpose—We investigated clinical and radiological characteristics of ischemic stroke patients with Takotsubo-like myocardial dysfunction.
Methods—From multicenter stroke registry database, ischemic stroke patients who underwent transthoracic echocardiography were found. Among these, patients were classified if they had specific ventricular regional wall motion abnormalities discording with coronary artery distribution, such as apical (typical pattern) or nonapical ballooning (atypical pattern), considered as echocardiographic findings of Takotsubo cardiomyopathy. Patients with ischemic heart disease history, myocarditis, or pheochromocytoma were excluded. We compared patients with Takotsubo-like myocardial dysfunction with those without and further performed systematic literature review on those with Takotsubo cardiomyopathy.
Results—This study included 23 patients (0.42%). The mean age was 70.7±13.9 years, with predominance of women (73.9%) and typical pattern of Takotsubo-like myocardial dysfunction (91.3%). They were associated with short-term poor functional outcomes, including high mortality, neurological deterioration, and functional status at discharge, compared with those without (39.1% versus 2.4%, 47.8% versus 7.4%; and median [interquartile range], 5 [5–6] versus 3 [2–4]; all P<0.001). They had a higher inflammatory marker level and lower triglyceride level. Ischemic lesions were more commonly found in the right anterior circulation with specific dominant regions being the insula and peri-insular areas. In addition, a trend toward a remarkable mortality rate and higher prevalence of insular involvement was observed in the propensity-score matching, subgroup fulfilling the strict Takotsubo cardiomyopath criteria, and was as reported in literature review.
Conclusion—Stroke patients with Takotsubo-like myocardial dysfunction may differ from those without in clinical outcomes, laboratory findings, and radiological features.
Takotsubo cardiomyopathy (TTC) is a reversible condition characterized by midsegmental left ventricular (LV) dysfunction with or without involvement of the apex, abnormalities of cardiac biomarkers and ECG, and absence of obstructive coronary artery disease or angiographic evidence of acute plaque rupture.1 Several comorbidities resulting in excessive catecholamine production could trigger or induce TTC.2 Although an overall prognosis of TTC is widely known to be favorable because it might completely resolve, a small subset shows a higher risk of poor outcome based on comorbid illness.3 Therefore, in accordance with specific comorbidities, the identification and management of clinical factors that might predispose patients to TTC or affect subsequent clinical outcomes are an integral part of treatment.2
Of the comorbidities, ischemic stroke has gained attention from clinicians because of the potential links between ischemic stroke and TTC. Currently, however, the information obtained from several anecdotal case series4,5 and a wide-field approach5,6 not focusing on ischemic stroke and considering ischemic stroke as one of the neurological diseases is limited and insufficient to determine a relationship between them. This is mainly derived considering that ischemic stroke itself could be a serious condition where invasive coronary angiography is difficult to get performed as a major diagnostic tool even when laboratory findings such as ECG abnormalities, elevated cardiac enzymes, or wall motion abnormalities of transthoracic echocardiography (TTE) suggest acute coronary syndrome (ACS) or TTC. Thus, TTC could be under-recognized and underestimated for most of the patients with acute stroke. In addition, there is no universal diagnostic criterion of TTC.3 Considering these limitations, concept of ischemic stroke with Takotsubo-like myocardial dysfunction (TMD) from the Japanese Guidelines for diagnosis of TTC7 was likely to be broad, but more appropriate and comprehensive than that of TTC in stroke.
Therefore, we evaluated the clinical, neurological, and neuroimaging features of patients with ischemic stroke and TMD based on TTE by using the multicenter hospital-based stroke registry. Moreover, we performed an up-to-date systematic review of literature on TTC in ischemic stroke to compare with our results.
Data Collection Based on Stroke Registry
Data were obtained from a prospectively collected multicenter hospital-based stroke registry (Korea University Stroke Registry [KUSR]) from January 2008 to February 2015. Details of the KUSR are described elsewhere.8 In brief, 3 university hospitals located in a separate urban area have been participating in the KUSR and sharing electronic documentation in KUSR format. Clinical, laboratory, and radiological data were consecutively collected in patients with acute ischemic stroke, within 7 days from symptom onset. Routine examination on patients diagnosed with stroke included cardiac evaluation consisting of ECG, TTE, and Holter monitoring. Lipid profiles were particularly examined with at least an 8-hour fasting at the first hospitalization day.
The study was approved by the institutional review boards of each hospital (KUGH14151, AN14170, and AS17044). Omission of written informed consent was permitted by each institutional review board because of the retrospective nature of this study. All investigators kept the participants’ anonymity by using a new identification code in a locked-area and managed in authorized computers.
Cardiac Evaluation and Patient Selection
The 3 participating centers had the same protocol for cardiac evaluation in patients with acute stroke. Every patient who was admitted because of ischemic stroke underwent ECG at the emergency department as the initial evaluation step. Abnormal ECG findings were T-wave inversion, ST-segment elevation, pathological Q wave,9 or QT prolongation. We used Bazett formula to calculate the corrected QT values, and an abnormal corrected QT in men and women was defined as corrected QT >450 ms and >470 ms, respectively.10 Additional cardiac enzymes such as serum troponin I or T, creatine kinase MB and N-terminal pro-brain natriuretic peptide were obtained in patients with cardiopulmonary symptoms such as chest pain or dyspnea or those suspected with ACS.
After admission, TTE was routinely performed as soon as possible, and the interpretation was performed by a certified and dedicated cardiologist at each center except for patients with conditions contraindicated for echocardiography. Particularly, coronary angiography was performed in accordance with the decision of the attending cardiologists. Cardiac enzymes and ECG were rechecked serially until these abnormalities were within normal range. Among the patients with abnormalities on initial TTE, repeat examination was performed in 2 to 3 weeks, if needed, when patients went into the recovery period. Normalization of LV wall motion abnormality was considered to occur when the moderately hypokinetic area disappeared and the LV ejection fraction (LVEF) improved to >0.6.11
We screened the patients with regional wall motion abnormality who were suspected of TTC from the initial echocardiographic report which staff cardiologists at each center with advanced training in echocardiography filled in. Subsequently, the echocardiographic findings were reviewed by a cardiologist (Y.H.K.) who was blinded to patients’ information, by using a viewer program (Centricity Enterprise Web; GE Medical Systems). The TMD based on the echocardiographic pattern of TTC was classified into typical and atypical patterns.12 Typical pattern was defined as LV dysfunction in mid- and apical segments, but normal contractility in basal segments. Atypical pattern implied involvement of basal- and midventricular segments and normal contractility of the apical segments.12 Those with a typical pattern of LV apical ballooning or atypical pattern of nonapical ballooning were enrolled. Patients with ischemic heart disease or myocarditis history were excluded, as well as those with pheochromocytoma.
Clinical Evaluation and Outcomes
Initial stroke severity and discharge outcome were assessed by using the National Institutes of Health Stroke Scale (NIHSS) and modified Rankin Scale, respectively. During hospitalization, neurological deterioration was defined as an increase in NIHSS score by at least ≥4 points within 10 days after admission.13
Lesion Laterality and Location
The topographical locations on acute infarction in patients with or without TMD were obtained from the KUSR database. Lesion laterality was divided into right anterior, left anterior, and posterior circulation. In addition, detailed distributions of infarction on initial diffusion-weighted magnetic resonance imaging were assessed by 2 investigators blinded to clinical information and echocardiographic findings (J.M.J. and J.K.K.).
Infarct lesions at the ganglionic level were manually outlined and overlapped on the basis of axial diffusion-weighted magnetic resonance imaging by using a normalized canonical image (ch2better.nii.gz) as provided in MRIcron (v6; http://www.mccauslandcenter.sc.edu/mricro/mricron/) by the third investigator (J.B.K.).
Continuous variables were presented as mean±SD, median (interquartile range or range) based on their distribution. Categorical variables were expressed as fractions (%). For between-group comparisons, the independent 2-sample t test or Mann–Whitney U test and the χ2 test or Fisher exact test were used for continuous and categorical variables, respectively. In addition, we performed propensity-score (PS) matching to adjust the baseline differences between the 2 groups (with TMD versus without TMD). The PS in each patient was obtained by using a logistic regression model, and matching was performed by using Griddy methods.14 We matched a randomly selected 1 TMD patient with 5 patients without TMD who had the nearest estimated logit value. Covariates for multivariable logistic regression model for calculating PS included age, sex, time-to-emergency room, thrombolysis, hypertension, diabetes mellitus, smoking, white blood cell count, hemoglobin, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglycerides, systolic blood pressure, diastolic blood pressure, hs-CRP (high-sensitivity C-reactive protein), atrial fibrillation, initial NIHSS, and Trial of Org 10172 in Acute Stroke Treatment classification. After PS matching, we further performed the same analyses for comparison of the 2 groups. SPSS 20.0 for Windows (IBM Corporation, Armonk, NY) was used. PS matching was performed with SAS version 9.3 (SAS Institute, Inc., Cary, NC).
Systematic Review of Literature
We searched the literature by using a comprehensive database, including the MEDLINE, Embase, Kmbase, and KoreaMED, from 1990 to August 2016. Keywords were as follows: stroke, Takotsubo cardiomyopathy, stress-induced cardiomyopathy, apical ballooning, stunned myocardium, case report, and case series. Only articles with brain imaging data or description of the involved brain lesion were included. English abstracts were also included if patient information was well described. No limitation on language was set. Two investigators (J.M.J. and J.Y.C.) independently identified the included cases and extracted the clinical and radiological information of those patients. Discrepancy was resolved through the consensus of the investigators. We followed the PRISMA guideline (Preferred Reporting Items for Systematic Reviews and Meta-Analyses).
Baseline Characteristics of Included Patients
During the study period, 6278 patients composed of 5817 stroke and 461 transient ischemic attack patients were registered in KUSR. Among 5817 patients with stroke, 5098 were selected for analysis. Of these, 23 (women, 73.9%; mean age, 70.7±13.9 years) met the inclusion criteria for TMD and were finally selected (Figure 1). Hypertension and atrial fibrillation were the most frequent cardiovascular risk factors (Table 1). Of these stroke subtypes, cardioembolic stroke was the most prevalent (52.2%). The median time from symptom onset to arrival at the emergency room was 3.4 hours (interquartile range, 1.07–14.5 hours). No previous emotional or psychological stress was addressed by patients or by caregivers. The presumable medical condition leading to TMD was postrenal acute renal failure in 1 patient who developed stroke after TMD.
Cardiac manifestations and evaluations are presented in Table I in the online-only Data Supplement. Cardiopulmonary symptoms, such as chest pain or dyspnea, were found in 3 patients, whereas 12 patients were unable to complain of these symptoms because of neurological deficits. T-wave inversion was the most common electrocardiographic finding (65.2%). The median (range) time from arrival at the emergency room to performing TTE was 2 days (0–9 days). Representative cases of TTE are depicted in Figure I in the online-only Data Supplement. Repeat TTEs were performed in 8 patients (34.8%) with a median of 21 days (3–240 days) from the initial echocardiography. Follow-up TTE showed full and partial recovery of LV dysfunction in 5 and 2 patients, respectively. No interval change was found in 1 patient, which might be because of early follow-up duration (3 days). The mean LV ejection fraction at follow-up TTE was 55.8±7.5%. Typical and atypical echocardiographic patterns were found in 21 (91.3%) and 2 (8.7%) patients, respectively. Coronary arteries were nonstenotic for all 5 patients who underwent coronary angiography.
Based on a chronological sequence, ischemic stroke preceded TMD in 11 patients, and vice versa in 2 patients. In contrast, TMD occurrence seemed unclear in 10 patients because abnormal ECG findings or elevated cardiac enzyme levels had been already found at arrival at the emergency room, or onset of cardiopulmonary symptoms was not obtained because of neurological deficits.
About short-term outcomes during hospitalization, 11 patients developed neurological deterioration, which was attributed to primary brain dysfunction, including stroke progression, new stroke, secondary brain dysfunction due caused by infectious etiology, and comorbidity aggravation in 8, 1, 1, and 1 patients, respectively. Based on discharge modified Rankin Scale, functional dependence (modified Rankin Scale score ≥ 4) was found in 22 patients, but not in 1 (95.7%). Of those, 9 patients (39.1%) died. The cause of death was cerebral, cerebellar, pulmonary edema, and sepsis in 6, 1, 1, and 1 patients, respectively.
Comparisons Between Those With TMD and Without TMD
Female predominance, low LVEF, increased levels of inflammatory markers, including white blood cell count and hs-CRP, and decreased triglyceride level were observed in patients with TMD compared with those without (Table 2). Those with TMD had older age and severe stroke with a higher range of initial NIHSS. Short-term outcomes, including neurological deterioration, functional dependence, and in-hospital mortality, were poorer in that group. Ischemic lesions were mainly located in the right anterior circulation compared with those from the KUSR database, which seemed evenly distributed (P=0.001; Figure 2). For specific anatomic substrate, involvement of the insular cortex and peri-insular areas, including the frontal cortex or subcortex, was more prevalent than that of other regions (Figure 3A). Particularly, the right insular cortex was the most frequently overlapped in the ganglionic level (Figure 3B).
Missing values were present only in those without TMD: white blood cell count (n=3), hemoglobin (n=3), total cholesterol level (n=80), high-density lipoprotein (n=113), low-density lipoprotein (n=97), triglyceride (n=105), and hs-CRP (n=117). We performed PS matching using data set without missing values because the proportion (all <2.1%) with missing values in group without TMD was minimal and main variables such as demographics and clinical outcomes had no missing value. Because 1 patient in the TMD group was not matched with one of group without TMD, 22 and 110 patients were finally selected in those with and without TMD, respectively. Interestingly, higher probability of poor outcomes was still sustained in the TMD group although the baseline characteristics between the 2 groups were comparable, except LVEF. Ischemic lesion predominance of the right anterior circulation was also found with a nonsignificant trend (P=0.251; Figure II in the online-only Data Supplement).
Subgroup Analysis for Those Fulfilling the Diagnostic Criteria of TTC
We selected only 8 patients (patient numbers 4, 7, 11, 17, 18, 19, 20, and 21) based on the strict TTC criteria.15 They fulfilled the condition that coronary angiography was normal or follow-up TTE was normalized. Poor short-term outcomes were observed (37.5%, 25%, and 100% for neurological deterioration, in-hospital mortality, and functional dependency at discharge, respectively). Ischemic lesion predominance of the right anterior circulation (7 patients; 87.5%) and insular involvement (5 patients; 62.5%) was similar to that of those with TDM.
This hospital-based registry study investigated the comprehensive relationship between ischemic stroke and TMD. Although this study focused on TMD rather than on TTC, a small number of patients with acute ischemic stroke actually had Takotsubo-like ventricular abnormality. However, those with TMD had distinguishing characteristics from those without. These findings were verified to be robust through results from PS matching. In addition, subgroup analysis adopted the strict TTC criteria, and systematic literature review demonstrated that patients with TMD had similar trends to those with TTC although this study had innate limitations for a possibility of mis- or underdiagnosis of ACS. Here, we discussed about the clinical features, laboratory findings, topographical brain lesion patterns, and time sequence in patients with TMD and ischemic stroke.
Clinical Features and Outcomes
Given that our study did not use the strict diagnostic TTC criteria, cumulative prevalence of TMD in ischemic stroke was much lower than the previously known incidence of TTC ranging from 1 to 12%.2 Female predominance was observed, in accordance with the previous report.16 However, emotional stress was not evident as a predisposing factor in cases of stroke as an important physical stressor. Initial presenting symptoms were not informative to be suspected or diagnosed. Most patients (82.6%) complained of no cardiopulmonary symptoms as initial presenting symptoms. Patients mainly presented with severe stroke attributed to atrial fibrillation and underwent acute thrombolytic treatment more frequently. Poor short-term outcomes including early neurological deterioration, in-hospital mortality, and unfavorable functional status at discharge were more frequently observed in stroke patients with TMD of unmatched and matched datasets. TMD may be a plausible prognostic factor or severity marker in ischemic stroke patients in that our study showed the relationship between those with TMD and higher NIHSS, atrial fibrillation, low LVEF, and higher inflammatory marker level considered as predictors for poor outcomes.17–19 In addition, a systematic review of published cases or case series showed a considerable mortality rate although their findings slightly vary with ours, which could imply an underestimation because of publication bias. Therefore, patients with ischemic stroke and TMD should be differentiated from idiopathic cases of TTC.20
Electrocardiographic changes and cardiac enzyme increases were confused with ACS although they might provide a clue to the presence of TMD. However, several laboratory findings would differ from those of stroke patients without TMD. Levels of inflammatory markers, such as white blood cell count and hs-CRP, increased in those with TMD. This could reflect the presence of an acute inflammatory change in the myocardium with regard to brain natriuretic peptide/N-terminal-pro-brain natriuretic peptide release21 or poststroke inflammation predicting prognosis or outcome.21,22 In addition, patients with TMD had significantly lower triglycerides level. The total and low-density lipoprotein cholesterol levels seemed to be decreased. These findings could suggest that pathophysiology of TMD was different from that of ACS.18 High prevalence of cardioembolic stroke or atrial fibrillation among the patients with TMD could be another possible explanation because triglycerides level is usually lower in the cardioembolic stroke than in the other subtypes among patients with acute stroke.23 Further studies are needed to evaluate whether these laboratory findings were associating factors, risk factors, or epiphenomenon for predicting TMD in ischemic stroke.
Topographical Brain Lesion Patterns
A well-designed study demonstrated that cerebral infarction on a specific anatomic brain lesion could predispose to TTC.4 Specific brain regions such as the insula (particularly the right) and medulla had been reported to play an important role in cardiovascular autonomic function on the basis of experimental24,25 and clinical26,27 evidence. In our study, ischemic lesions in the right anterior circulation were associated with TMD. When ischemic lesions were divided into involved regions, the insula cortices and peri-insular areas (especially the right) were commonly detected. However, the involvement of the brain stem including the medulla was less frequently observed. These brain lesion involvements corresponded to TMD cases before stroke occurrence. In addition, this systematic review showed similar lesion preponderance of the insular cortices. These results suggested another evidence supporting brain–heart interaction.
Time Sequence Between TMD and Stroke
The association between TTC and ischemic stroke could be divided into 3 categories: (1) TTC caused by central autonomic network dysfunction associated with cerebral infarction, (2) cardioembolic stroke caused by LV thrombus associated with TTC, and (3) unknown cause and effect relation.28 Ischemic stroke preceded TMD in 11 patients in our study. Two patients had TMD, which led to ischemic stroke. One had LV thrombus in TTE, suggesting TMD possibly preceded by the index stroke and the other patient developed ischemic stroke after TMD diagnosis. Other 10 patients had an unclear sequence of relationship between TMD and ischemic stroke. All had abnormal electrocardiographic findings and elevated cardiac marker levels at arrival of the emergency room. They had no preceding event of psychological or physical stress as a precipitating factor except stroke. In addition, TTC developed several hours after ischemic stroke.6 Combined with these findings, most patients with an unclear time sequence relationship might belong to the first group.
Reports describing cases presenting with TTC or TMD (coronary angiography not performed like a part of our study patients) in stroke were included. Fifty reports describing a total of 67 cases were finally included (Figure III and Tables II and III in the online-only Data Supplement). The mean age was 69.5±13.8 years. Literature showed more women (87.9%). Diagnostic coronary angiography or follow-up 2-dimensional echocardiography was neither performed nor described in the reports (43.3% and 23.9%, respectively). For the chronological sequence, TTC preceded ischemic stroke in 25 patients, and vice versa in 29. Unclear time sequence was also observed in 13 patients, most of whom presented with drowsy mental status or aphasia. Initial neurological symptoms ranged from mild to severe. Outcome was relatively good, and follow-up echocardiography was normalized or improved in most patients who underwent echocardiography. However, 5 patients (7.5%) died at discharge or 1 month after. Seven patients were discharged with severe neurological deficits or transferred to the rehabilitation department.
Brain images were presented in 29 reports, including only 39 cases. Brain lesions were predominantly confined to the insular cortex (20 [51.3%] of 39 scans; right, 6; left, 12; both 2). Brain stem lesions were observed in 4 patients (10.5%). Insular cortex involvement (14 [51.9%] of 27 scans; right, 5; left, 7; both, 2) was slightly increased in 27 patients with stroke before TTC or unclear chronological sequence of stroke and TTC. Otherwise, brain images of patients with TTC before stroke revealed ischemic lesions in the insula (5 [41.7%] of 12 scans; left, 4; right, 1).
This study has several limitations. First, this study was based on retrospective chart review although prospectively collected. Second, this study could not strictly discriminate TMD from TTC or ACS because of the low performance rate of coronary angiography (14.3%). This is presumed to be because of poor neurological or general conditions of patients with stroke. This situation corresponded well with the previous case reports and with the real clinical setting, wherein some are reluctant to perform invasive studies, such as coronary angiography,29 or further imaging studies, such as cardiac magnetic resonance imaging or positron emission tomography to patients with severe stroke. In addition, most of them (52.2%) did not undergo repeated echocardiography because of the poor general conditions arising from severe stroke and higher mortality rate although follow-up echocardiography showed LV dysfunction improvement or normalization in others. Furthermore, short-term follow-up echocardiography would be insufficient to demonstrate initial LV dysfunction improvement or normalization because its normalization occurs usually several weeks to months later.15 It is possible that Takotsubo-like ventricular abnormality was also derived in part from coronary artery disease in our patients although coronary artery disease would be a bystander, not a causative factor.30 However, this situation was rare.15 Third, we used TMD mixed with TTC in although TMD was different from TTC. However, subgroup analysis and systematic literature review demonstrated that patients with TMD had similar characteristics as those of patients with TTC.
Considering the poor clinical outcomes and topographical brain lesion features related with TMD, clinical relevance of such findings in ischemic stroke is important although TMD prevalence is relatively low, and under- or misdiagnosis of ACS is possible. Further studies investigating proper prevention, identifying patients at risk, and optimal managements for such conditions are needed.
We thank Eun Ju Lee for much endeavor, manager from the Medical Library, Korea University, and for all her efforts in searching for related abstracts and articles.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.116.014304/-/DC1.
- Received June 7, 2016.
- Revision received August 16, 2016.
- Accepted August 29, 2016.
- © 2016 American Heart Association, Inc.
- Thygesen K,
- Alpert JS,
- White HD,
- Jaffe AS,
- Apple FS,
- Galvani M,
- et al
- Kaarisalo MM,
- Immonen-Räihä P,
- Marttila RJ,
- Salomaa V,
- Kaarsalo E,
- Salmi K,
- et al
- Simats A,
- García-Berrocoso T,
- Montaner J