Neurocardiac Injury Assessed by Strain Imaging Is Associated With In-Hospital Mortality in Patients With Subarachnoid Hemorrhage
Original Research
Central Illustration
Abstract
Objectives
This study sought to test the hypothesis that speckle tracking strain echocardiography can quantify neurocardiac injuries in patients with aneurysmal subarachnoid hemorrhage (SAH), which is associated with worse clinical outcome.
Background
SAH may be a life-threatening disease associated with variable degrees of neurocardiac injury. Strain imaging has the potential to detect subtle myocardial dysfunction which is additive to conventional measurements.
Methods
A total of 255 consecutive patients were prospectively enrolled with acute SAH, who were admitted to the intensive care unit with echocardiography studies within 72 h. Left ventricular (LV) and right ventricular (RV) strains were acquired from standard apical views. Abnormal LV global longitudinal strain (GLS) and RV free-wall strain were pre-defined as <17% and <23% (absolute values), respectively.
Results
Performing LV GLS was feasible in 221 patients (89%) 53 ± 10 years of age, 71% female, after excluding those with previous cardiac disease. Abnormal LV GLS findings were observed in 53 patients (24%) and were associated with worse clinical severity, including a Hunt-Hess grade >3 (34% vs. 15%; p = 0.005) and biomarker evidence of neurocardiac injury and higher troponin values (1.50 [interquartile range (IQR): 0.01 to 3.87] vs. 0.01 [IQR: 0.01 to 0.22] ng/ml; p < 0.001). A reverse Takotsubo pattern of segmental strain was observed in 49% of patients (apical sparing and reduced basal strain). Importantly, LV GLS was more strongly associated with in-hospital mortality than left ventricular ejection fraction (LVEF), even after adjusting for clinical severity (odds ratio [OR]: 3.11; 95% confidence interval [CI]: 1.12 to 8.63; p = 0.029). RV strain was measured in 159 subjects (72%); abnormal RV strain was added to LV GLS for predicting in-hospital mortality (p = 0.007).
Conclusions
Neurocardiac injury can be detected by LV GLS and RV strain in patients with acute SAH. LV GLS was significantly associated with in-hospital mortality. RV strain, when available, added prognostic value to LV GLS. Abnormal myocardial strain is a marker for increased risk of in-hospital mortality in SAH and has clinical prognostic utility.
Introduction
Aneurysmal subarachnoid hemorrhage (SAH) remains a serious disease associated with considerable mortality and disability (1). Although SAH most often occurs in younger to middle-aged female patients without a history of heart disease, various degrees of associated neurocardiac injury may occur. Neurocardiac injury is typically detected clinically by abnormal electrocardiography (ECG) results, elevated serum troponin levels, or abnormal left ventricular (LV) function by routine echocardiography. Even modest degrees of neurocardiac injury have been associated with a worse physical and functional prognosis (2–5). Strain imaging by speckle tracking echocardiography may quantify LV function and is reported to be a more sensitive and more powerful prognostic marker than conventional parameters such as LV ejection fraction (EF) and wall motion score indexes (6–10). Accordingly, the objective of this study was to test the hypothesis that myocardial strain imaging in the acute phase of SAH may quantify neurocardiac injury by using LV global longitudinal strain (GLS), LV segmental strain patterns, and right ventricular (RV) strain to provide additional prognostic information to routine clinical measurements.
Methods
Study design
This was a prospective, single-center, longitudinal observational study conducted for the purpose of investigating the neurocardiac injury in patients with SAH, known as the SAHMII (SubArachnoid Hemorrhage and Myocardial Infarction and Ischemia) study. The protocol was approved by the Institutional Review Board, and all patients or their representatives gave written informed consent. A total of 255 patients with aneurysmal SAH in whom echocardiography was performed in the acute phase (<72 h) were prospectively enrolled. Inclusion criteria were >18 years of age; admission within 24 h of SAH symptoms; verification of SAH by computed tomography scan with Fisher grade ≥2 and corroborated by cerebral angiogram. Exclusion criteria were SAH from trauma, mycotic aneurysm or arterial venous malformation; history of myocardial infarction, cardiac surgery or trauma within 1 year; or serum creatinine concentration >2.0 mg/dl. Laboratory data included serum troponin I levels (collected at least daily) and daily 12-lead ECG and chest radiographs. ECG abnormalities were defined as a composite of T-wave inversion or ST-segment elevation or depression or atrial or ventricular arrhythmia episodes, or both. Severity of neurologic injury was scored using Hunt-Hess grade (11), and computed tomography findings were rated by Fisher grade (12). The primary endpoint was all-cause in-hospital mortality.
Echocardiography
Transthoracic echocardiography was performed (Vivid-7, Vivid-E9, or Vivid-E95, GE Healthcare, Horten, Norway) in the acute phase at the patient’s bedside (13). Images for strain measurements were acquired at 65 ± 20 fps and were analyzed offline by investigators blinded to all clinical and laboratory data, using commercially available software (EchoPAC version 201, GE Healthcare). LV GLS was calculated as the average strain of 18 segments from 3 standard apical views (13). Apical, mid, and basal strains were defined as the average of each 6 segments, respectively. RV strain was measured using the apical 4-chamber view as the average strain of the 3 free-wall segments. Longitudinal strain values were presented as absolute values in order for larger numbers to mean better cardiac function and smaller numbers to mean poorer cardiac function, which was more easily understood by our multidisciplinary team (14). We defined LV GLS <17% and RV strain <23% as abnormal LV and RV strain, respectively, according to previously reported normal ranges (13,15). A control group of 25 normal healthy volunteers, 35 ± 6 years of age, were studied by speckle tracking echocardiography to determine normal regional strain patterns. The apical-to-basal peak strain ratio was calculate by the following formula: [apical strain/basal wall strain] (16). Mean ± 2 SD in the control group was defined as the upper and lower limits of normal range for regional strain patterns.
Variability
Interobserver variability for the LV and RV strain analyses were evaluated independently in 20 randomly selected patients by 2 blinded observers. Intraobserver variability was evaluated by analysis of measurements in the 20 patients by the same observer at least 1 month after the first analysis.
Statistical analysis
Data are mean ± SD or median (interquartile range [IQR]) for continuous variables and a frequency (%) for categorical variables. Group differences were evaluated using Student’s t-tests, Mann-Whitney U tests, or one-way ANOVA tests for continuous variables and chi-square or Fisher’s exact tests for categorical variables. Repeated test results were compared using paired t-test. A p value for trend of mortality among 4 LV GLS groups was calculated by Cochran-Armitage test. To explore the independent association between LV GLS and in-hospital mortality, multivariate logistic regression models were constructed using the reported predictors of worse outcome after SAH (i.e., age, systolic blood pressure, heart rate, Hunt-Hess grade, Fisher grade, troponin level, abnormal ECG findings, LVEF, and wall motion score) (4,17–19). Consistent with the number of deaths as the primary outcome in this study, models were constructed including 2 independent variables in each. Because Hunt-Hess grade is established as a powerful prognostic marker in SAH (1,11,20), multivariate models were tested to assess the incremental value of strain measurements as follows: 1) only Hunt-Hess grade; 2) Hunt-Hess grade and LV GLS; and 3) Hunt-Hess grade and LV GLS and RV strain. Area under the curves (AUCs) of the receiver-operating characteristic (ROC) curves were calculated. Category-free net reclassification improvement was used to assess the incremental value of adding LV and RV strain (21). Interobserver and intraobserver variability of the strain analyses were assessed using intraclass correlation coefficients. All statistical analyses were performed with R software (R Foundation for Statistical Computing, Vienna, Austria). In all analyses, a 2-tailed p value <0.05 indicated statistical significance.
Results
Study cohort
The initial cohort consisted of 255 consecutive patients with available acute phase echocardiograms from the SAHMII study. Seven additional patients were excluded because they had a history of cardiac disease that was discovered after the initial screening. Analyses of LV GLS were able to be carried out in 221 patients (89%), and these data were included in the final analysis (Figure 1). Frame rates of the images were 65 ± 20 fps. Overall, the mean age was 52.5 ± 10.3 years, and 71% were female. Mean Hunt-Hess grade was 2.6 ± 1.1, with 20% assessed as Hunt-Hess grade >3, which indicates severe hemiparesis or coma. The prevalence of patients with abnormal ECG findings was 38%, and troponin I concentration ≥1.0 ng/ml was 22%. Additional patient characteristics are shown in Table 1.
LV Global Longitudinal Strain
Data are from a patient demonstrating LV strain acquisition from standard apical views. LV GLS was calculated as the average strain of the 18 segments. GLS = global longitudinal strain; LV = left ventricle.
| Overall (N = 221) | LV GLS ≥17% (n = 168) | LV GLS <17% (n = 53) | p Value | |
|---|---|---|---|---|
| Age, yrs | 52.5 ± 10.3 | 52.4 ± 10.1 | 52.9 ± 11.1 | 0.73 |
| Females | 157 (71.0) | 116 (69.0) | 41 (77.4) | 0.30 |
| Systolic BP, mm Hg | 150 ± 30 | 149 ± 28 | 150 ± 35 | 0.85 |
| Diastolic BP, mm Hg | 83 ± 18 | 83 ± 17 | 83 ± 19 | 0.94 |
| Heart rate, beats/min | 78 ± 16 | 77 ± 15 | 83 ± 18 | 0.007 |
| History of hypertension | 82 (37.6) | 61 (37.0) | 21 (39.6) | 0.75 |
| Troponin I, ng/ml | 0.01 (0.01–0.69) | 0.01 (0.01–0.22) | 1.50 (0.01–3.87) | <0.001 |
| Troponin I >1.0 ng/ml | 48 (21.8) | 21 (12.6) | 28 (52.8) | <0.001 |
| Abnormal ECG | 85 (38.5) | 50 (29.8) | 35 (66.0) | <0.001 |
| Pulmonary edema | 93 (42.3) | 61 (36.5) | 32 (60.4) | 0.003 |
| Hunt-Hess grade | 2.6 ± 1.1 | 2.5 ± 1.0 | 3.0 ± 1.1 | 0.002 |
| Hunt-Hess grade >3 | 43 (19.5) | 25 (14.9) | 18 (34.0) | 0.005 |
| Fisher grade | 0.21 | |||
| 2 | 97 (43.9) | 78 (46.4) | 19 (35.8) | |
| 3 | 86 (38.9) | 65 (38.7) | 21 (39.6) | |
| 4 | 38 (17.2) | 25 (14.9) | 13 (24.5) | |
| Echocardiography | ||||
| LA diameter, cm | 3.37 ± 0.50 | 3.36 ± 0.48 | 3.39 ± 0.55 | 0.72 |
| LV diastolic diameter, cm | 4.61 ± 0.57 | 4.55 ± 0.55 | 4.81 ± 0.60 | 0.004 |
| LV systolic diameter, cm | 3.03 ± 0.66 | 2.93 ± 0.56 | 3.38 ± 0.82 | <0.001 |
| LV ejection fraction, % | 61.7 ± 9.1 | 64.2 ± 5.6 | 53.4 ± 12.4 | <0.001 |
| Cardiac index, l/min/m2 | 2.82 ± 0.79 | 2.88 ± 0.83 | 2.62 ± 0.60 | 0.036 |
| LV GLS (absolute), % | 19.4 ± 4.0 | 21.1 ± 2.7 | 14.0 ± 2.6 | <0.001 |
| RV strain (absolute), % | 25.3 ± 5.4 | 26.4 ± 5.0 | 21.3 ± 4.8 | <0.001 |
Routine left ventricular functional measurements and strain
Patients underwent echocardiography at a mean time of 2.3 days from admission. Mean LV GLS was 19.4 ± 4.0% (absolute values), and there were 53 patients (24.0%) with abnormal LV GLS results (<17%). The scatterplot in Figure 2 shows the relationship between LVEF and LV GLS. Overall, there was a significant correlation between LVEF and LV GLS (r = 0.59; p < 0.001). However, LV GLS was weakly correlated in patients with preserved to hyperdynamic LVEF (LVEF ≥50%), which was most of the patients (r = 0.29; p < 0.001). In patients with reduced LVEF, that is <50%, the correlation was closer (r = 0.83; p < 0.001). Reproducibility of LV GLS was excellent, with an interobserver intraclass coefficient of 0.98 and an intraobserver intraclass coefficient of 0.96.
Scatterplot of GLS Versus EF
The relationship between LV EF and LV GLS. In the patients with preserved LVEF ≥50%, LV GLS had a wide range, and the correlation between LVEF and LV GLS was weaker (r = 0.29; p < 0.001), whereas in patients with reduced LVEF <50%, the correlation was closer (r = 0.83; p < 0.001). EF = ejection fraction; other abbreviations as in Figure 1.
Reverse takotsubo pattern of LV regional strain
The abnormality in regional strain was more pronounced in basal segments, with relative sparing in apical segments. Using the normal control group as a reference, abnormal apical sparing strain above the upper limit of normal was defined as >1.43. This apical sparing has been previously referred to in SAH as the “reverse Takotsubo pattern” (22,23). A reverse Takotsubo pattern of apical sparing was observed in 49% of the present patient cohort overall and was observed in all LV GLS subgroups, regardless of LV GLS values (p < 0.005 vs. controls in every group) (Figure 3). There were only 3 patients (1.4%) with apical akinesis and no patients had apical ballooning. There were no significant differences in clinical characteristics and acute clinical findings between those with and those without the reverse Takotsubo pattern, including respiratory parameters, blood pressure, troponin level, and level of consciousness (Supplemental Table 1). In 125 patients in whom follow-up strain analyses were able to be carried out, median time between the initial echocardiography and the follow-up echocardiography was 4 (range: 3 to 5) days. Within this short period of time to follow-up, the apical/basal strain ratio significantly improved (1.51 to 1.41; p = 0.006), and the prevalence of the reverse Takotsubo pattern was significantly decreased (59% to 40%; p < 0.001).
Reverse Takotsubo Pattern in Subarachnoid Hemorrhage
Polar plots of regional strain in normal control subjects (upper left) and all subarachnoid patients (upper right) with apical and basal ratios shown. (Bottom row) Polar plots are grouped by 4 equal ranges of GLS ranges. Apical/basal strains were significantly higher in all LV GLS subgroups than in normal controls (all p < 0.005). Abbreviations as in Figure 1.
Neurocardiac injury among LV GLS subgroups
There were no significant differences between basic characteristics of the patients with abnormal (<17%) and those with normal (≥17%) LV GLS. Patients with abnormal LV GLS had significantly lower cardiac index, higher troponin levels, abnormal ECG findings, and higher clinical severity of SAH by Hunt-Hess grade than the patients with normal LV GLS. Furthermore, in the 4 groups with equal ranges of LV GLS, both the troponin level and the prevalence of poor Hunt-Hess grade (>3) were highest in the group with the most abnormal LV GLS patterns and decreased linearly as LV GLS improved (Figure 4). These findings strongly suggested that more severe neurocardiac injury in SAH is associated with more reduced LV GLS.
Neurocardiac Injury in Patients Grouped by Left Ventricular Global Longitudinal Strain Values
Bar graphs of patient data grouped by 4 equal ranges of LV GLS ranges. Clinical severity, using Hunt-Hess grades (left) and biomarker evidence of neurocardiac injury by using troponin I (right) were significantly associated with LV GLS values using ANOVA.
LV GLS and in-hospital mortality
Overall, there were 18 in-hospital deaths (8.1%). The cause of death in the majority of patients (78% [n = 14]) was attributed to multiorgan system failure, 3 patients (17%) had rerupture of cerebral aneurysms, and 1 patient (6%) died of refractory cardiac failure. There were 15 patients (83%) with hypotension that required inotropic support, which was associated with cardiac dysfunction in previous studies (24). In-hospital mortality was significantly higher in the patients with abnormal LV GLS than in the group with normal LV GLS (17.0% vs. 5.4%, respectively; p = 0.017). Importantly, when dividing patients into 4 groups by equal ranges of LV GLS, mortality was significantly associated with LV GLS and linearly increased as absolute values of LV GLS became more abnormal (p = 0.0057) (Figure 5). Multivariate analysis demonstrated that, even after adjusting for previously reported factors associated with mortality in SAH, including LVEF and Hunt-Hess grade, abnormal LV GLS remained a significant independent predictor of in-hospital mortality (Table 2). Specifically, when LV GLS was added to Hunt-Hess grade, the AUC of the model for predicting in-hospital mortality increased and significant category-free net reclassification improvement (0.57; 95% confidence interval [CI]: 0.09 to 1.04; p = 0.020) was observed (Supplemental Figure 1).
LV Global Longitudinal Strain and In-Hospital Mortality
Bar graphs of in-hospital mortality rates among patients grouped by 4 equal ranges of LV GLS ranges. All-cause mortality was significantly associated with LV GLS and linearly increased as LV GLS decreased in absolute values (p = 0.0057). Abbreviations as in Figure 1.
| OR | 95% CI | p Value | |
|---|---|---|---|
| Model 1 | |||
| Age | 1.04 | 0.99–1.09 | 0.13 |
| Abnormal LV GLS | 3.56 | 1.32–9.58 | 0.012 |
| Model 2 | |||
| Systolic BP | 1.00 | 0.99–1.02 | 0.77 |
| Abnormal LV GLS | 3.61 | 1.35–9.67 | 0.011 |
| Model 3 | |||
| Heart rate | 0.99 | 0.96–1.02 | 0.55 |
| Abnormal LV GLS | 3.85 | 1.41–10.5 | 0.009 |
| Model 4 | |||
| Hunt-Hess grade ≥3 | 16.4 | 2.13–126 | 0.007 |
| Abnormal LV GLS | 3.11 | 1.12–8.63 | 0.029 |
| Model 5 | |||
| Troponin I | 0.98 | 0.87–1.11 | 0.77 |
| Abnormal LV GLS | 3.79 | 1.34–10.7 | 0.012 |
| Model 6 | |||
| Abnormal ECG | 1.48 | 0.52–4.23 | 0.46 |
| Abnormal LV GLS | 3.14 | 1.10–8.96 | 0.032 |
| Model 7 | |||
| Reduced LVEF <50% | 5.61 | 0.64–48.6 | 0.12 |
| Abnormal LV GLS | 5.42 | 1.92–15.3 | 0.001 |
| Model 8 | |||
| Wall motion score index | 0.55 | 0.05–6.33 | 0.63 |
| Abnormal LV GLS | 4.05 | 1.40–11.7 | 0.010 |
Right ventricular strain and clinical outcomes
There were 159 patients (72%) in whom RV strain analyses were feasible (Figure 6). There were 54 patients (34%) with abnormal RV strains, pre-defined as <23%. These patients with abnormal RV strain also had significantly greater LV size (mean LV diastolic diameter 4.79 ± 0.59 cm vs. 4.51 ± 0.54 cm, respectively; p = 0.004) and decreased LV systolic function (LVEF 57.3 ± 10.8% vs. 64.0 ± 7.2%, respectively; p < 0.001; and LV GLS 17.7 ± 4.3% vs. 20.8 ± 3.5%, respectively; p < 0.001) than those with normal RV strain. Also, the prevalence of a poor Hunt-Hess grade (>3; 29.6% vs. 14.3%, respectively; p = 0.03) and troponin level (0.13 [IQR: 0.01 to 1.58] ng/ml vs. 0.01 [IQR: 0.01 to 0.28 ng/ml]; p = 0.02) were significantly greater in patients with abnormal RV strain than those with normal RV strain. In-hospital mortality was also significantly higher in the patients with abnormal RV strain (13.0 vs. 3.8%, respectively; p = 0.046). To assess biventricular strain as markers of degree of neurocardiac injury, patients were divided into 3 groups, as follows, group A were patients in whom both LV GLS and RV strain were abnormal, group B consisted of patients in whom either the LV GLS or RV strain was abnormal, and group C, patients in whom neither the LV GLS nor the RV strain was abnormal. Mortality was significantly increased in group A (22.7%) with abnormal biventricular strain compared with group B (4.7%) and group C (4.3%) (p = 0.007). Furthermore, adding RV strain to the logistic regression model consisting of Hunt-Hess grade and LV GLS showed statistically significant net reclassification index improvement, supporting the fact that RV strain was superior to the model with Hunt-Hess grade and LV GLS alone (0.64; 95% CI: 0.05 to 1.23; p = 0.034). (Supplemental Figure 1). However, the change in AUC of ROC curves from 0.72 to 0.73 was not significant. The reproducibility of RV free-wall strain was similar with an interobserver and interclass coefficient of 0.88 and an intraobserver and intraclass coefficient of 0.88. Interobserver and intraobserver agreement of LV GLS was 90% and 85%, respectively, and both of those for RV strain were 75%. Mean differences in interobserver and intraobserver variability analyses were 0.2 ± 1.2% and 0.8 ± 1.1% for LV GLS and 0.8 ± 2.8% and 0.5 ± 2.7% for RV strain. Based on variability analysis, it was determined that a clinically acceptable margin of error for LV GLS was 2% (in strain units) and that for RV strain was 3% (in strain units).
RV Strain Measurement
An example from a patient demonstrating RV acquisition from the apical 4-chamber view. Segmental free-wall and average free-wall stain from 3 segments are shown in absolute values. RV = right ventricle.
Discussion
This study demonstrates that cardiac strain imaging is an effective means to quantify neurocardiac injury associated with clinical outcomes in a large prospective series of patients with aneurysmal SAH. GLS was significantly associated with in-hospital mortality and was added to routine clinical markers of prognosis. Specifically, it was observed that abnormal LV GLS and RV strains in acute SAH patients commonly occurred even in the setting of preserved or hyperdynamic LVEF. Regional strain abnormalities in SAH appeared to be a reverse Takotsubo pattern, which occurred in all ranges of LV GLS and provided mechanistic insight. Abnormal LV GLS and abnormal RV strain were significantly associated with evidence of neurocardiac injury. Furthermore, abnormal LV GLS remained significantly associated with in-hospital mortality even after adjusting for known factors associated with outcome after SAH, including Hunt-Hess grade. Also, this is the first study showing RV dysfunction in SAH by RV strain, which could be added to LV GLS as an indicator of biventricular neurocardiac injury in SAH and furthered the association with in-hospital mortality. The important potential clinical implication of assessing LV and RV strain in the acute setting of SAH is that these novel markers of neurocardiac injury were prognostic markers of clinical outcomes that were superior to routine means, such as troponin values and LVEF.
Neurocardiac injury in SAH
Neurocardiac injury represents an important phenomenon which remains incompletely understood as a secondary complication of SAH (3,4,18,25). The prevailing hypothesis is that acute brain injury from SAH is associated with abrupt increases in intracranial pressure which elicits an intense autonomic trigger. This results in copious release of catecholamines locally in the brain as well as systemically, including at the level of the adrenergic receptors upon the myocytes, which can lead to cell injury and death (Central Illustration). In addition, catecholamine toxicity may result in cellular functional changes including conduction or contractile impairment or death. Interestingly, the pattern of the cellular injury due to intense sympathetic stimulation follows the distribution of the sympathetic nerve terminal endings, rather than following the distribution of the coronary arteries as seen in embolic and ischemic injury causes. It has been noted that the myocardial cellular damage associated with neurocardiac injury has been previously observed as cardiac troponin release in approximately 20 to 35% of patients, with most studies following cardiac troponin (2,5,26). In contrast, the number of patients with wall motion abnormalities detected by routine conventional echocardiography was much less (5,24,27,28). More recently, strain imaging has been reported to be more sensitive to and more reproducible for detecting subtle cardiac dysfunction in other diseases (29). In present study, a large number of patients were found with abnormal LV GLS despite having preserved LVEF. Importantly, LV GLS was significantly and linearly associated with other markers of neurocardiac injury, including serum troponin levels, abnormal ECG results, and pulmonary edema (4,5,30–32).
Neurocardiac Injury Assessed by Strain Imaging in Subarachnoid Hemorrhage
Subarachnoid hemorrhage abruptly increases intracranial pressure, which results in an extreme catecholamine surge. Speckle tracking strain imaging was able to quantify LV and RV myocardial injury associated with subarachnoid hemorrhage. Strain values were significantly associated with evidence of neurocardiac injury and in-hospital mortality. LV = left ventricle; RV = right ventricle.
Strain and mortality after SAH
One of the most important findings in this study was the significant associations between strain evaluation at an early time point during hospitalization and eventual in-hospital mortality. The exact mechanistic basis for this finding remains uncertain, because the extent of neurocardiac injury was associated with a degree of SAH clinical severity. Previous studies have shown associations between neurocardiac injury assessed by troponin level and abnormal ECG and poor outcomes (2,4,28). ECG changes are frequently observed in patients with SAH and are reported to occur more commonly than elevated troponin I (TnI) or LV wall motion abnormalities (2,5). ECG changes are not considered a sensitive marker for neurocardiac injury because they may result from hypothalamic stimulation without myocardial damage. Furthermore, the prognostic significance of an abnormal ECG is weaker than TnI or wall motion abnormalities (2,5). In other words, the high prevalence of ECG abnormalities in SAH does not necessarily mean that ECG is more sensitive than GLS or TnI for detection of myocardial damage. The prevalence of elevations of TnI in the present cohort of SAH patients is consistent with that in previous studies. Cinotti et al. (33) reported that elevated troponin T occurred in a subset of SAH patients with relatively preserved GLS >−16%. In the present study, elevated troponin levels and abnormal ECGs were observed in SAH patients with relatively normal GLS >17%. An abnormal GLS <17% was associated with mortality independently of LVEF, abnormal ECG, or serum troponin level. These results indicate that GLS may more accurately stratify the “significant” neurocardiac injury, in terms of the association with clinical outcome, than these conventional markers. Furthermore, acute phase measurements of LV GLS were linearly related to clinical severity and in-hospital mortality, even in the patients with LV GLS values of 17% to 23%, which is considered normal in other settings. These findings suggest that, in such stressed conditions, cardiac function is boosted and normal range of GLS should be higher. Other clinical scenarios are observed with severe mitral or aortic regurgitation, where slight reduction of GLS, which is usually considered within normal range (19% to 22%), is significantly associated with increased mortality in patients with preserved LVEF (8,34). Similarly, GLS values of 17 to 23% may represent subtle myocardial damage in the acute phase of SAH. Mild degrees of myocardial dysfunction detected by strain imaging may be a marker for subtle systemic and cerebral perfusion impairment in SAH as well (24). Recent studies have reported RV function has an important prognostic role in other diseases (35,36), and ours is the first study extending RV functional assessment by RV strain to patients with SAH. The present study showed that reduced RV strain was associated with the evidence of severe biventricular neurocardiac injury and added incremental prognostic value to LV GLS.
Reverse takotsubo pattern in SAH
Takotsubo cardiomyopathy or apical ballooning is another form of stress-related cardiomyopathy (22,23). Interestingly, the most of the patients in the present study had a reverse Takotsubo strain pattern with apical sparing, and only 3 (1.2%) had apical wall motion abnormalities, similar to those reported by other studies in SAH, ranging from 1.2% to 2.2% (37,38). Apical sparing of wall motion in SAH was previously observed by visual inspection (22,23,28) and has now been confirmed by using quantitative strain imaging. The apical sparing pattern is not associated with a coronary artery distribution. It has been reported that beta-2 receptors are more frequently expressed in apical rather than basal segments, in contrast to that of sympathetic nerve terminals and norepinephrine receptors, which are expressed more highly at basal than apical segments (9,39). Although the role of adrenoreceptor activation in these diseases is not fully understood, the differences among receptor activation patterns suggest that systemic release of epinephrine is associated with apical dysfunction in Takotsubo cardiomyopathy whereas local release of norepinephrine is more likely associated with basal dysfunction in SAH.
Study limitations
It is a limitation that strain imaging was only possible in 89% of patients for LV GLS and 72% for RV strain, using echocardiography in ill patients in the neurologic intensive care unit. Although myocardial strain imaging has demonstrated additive clinical prognostic utility in SAH, it remains unknown if myocardial dysfunction from neurocardiac injury is contributory to higher mortality or a marker for the extent of neurological injury. Certainly, there are limited data for strain imaging in patients with SAH, accordingly the cutoff values used for GLS and RV strain may be improved with future study. The number of events was relatively small, with 18 deaths, which limited the statistical methods. However, the linear relationship between GLS and mortality and multivariate models of GLS were consistently prognostically superiority to troponin values, LVEF, and Hunt-Hess grade. Finally, this is the very first prospective study showing prognostic benefit of strain imaging in patients with SAH, and further studies are warranted for external validation.
Conclusions
Neurocardiac injury may be detected by strain imaging soon after hospitalization in patients with SAH. Abnormal LV GLS was significantly associated with greater clinical severity of SAH and higher troponin levels. Furthermore, LV GLS and RV strain provided additional prognostic utility for predicting in-hospital mortality. Strain imaging assessment has the ability to provide novel future insight into the mechanisms of neurocardiac injury in SAH and has promising impact on risk assessment in patients with acute SAH.
Perspectives
COMPETENCY IN MEDICAL KNOWLEDGE: This paper represents an opportunity to build on competency in medical knowledge. In patients with subarachnoid hemorrhage, reduced cardiac strain was significantly associated with other markers of neurocardiac injury and more strongly associated with in-hospital mortality than left ventricular ejection fraction. Left ventricular global longitudinal strain was associated with in-hospital mortality and right ventricular strain was of additive prognostic value.
TRANSLATIONAL OUTLOOK: The present study results suggest that subtle myocardial dysfunction can be appreciated by strain echocardiography and is associated with poor clinical outcome. Although a mechanistic conclusion cannot be drawn, it appears that neurocardiac injury detected by strain imaging is a marker for the extent of neurological insult and is a new prognostic marker in patients with subarachnoid hemorrhage.
Abbreviations and Acronyms
| EF | ejection fraction |
| GLS | global longitudinal strain |
| IQR | interquartile range |
| LV | left ventricular |
| RV | right ventricle |
| SAH | subarachnoid hemorrhage |
References
1. "Subarachnoid hemorrhage". N Engl J Med 2017;377:257-266.
2. "Cardiac Troponin elevation and outcome after subarachnoid hemorrhage: a systematic review and meta-analysis". J Stroke Cerebrovasc Dis 2015;24:2375-2384.
3. "Impact of cardiac complications on outcome after aneurysmal subarachnoid hemorrhage: a meta-analysis". Neurology 2009;72:635-642.
4. "Cardiac troponin elevation, cardiovascular morbidity, and outcome after subarachnoid hemorrhage". Circulation 2005;112:2851-2856.
5. "Elevated cardiac troponin I and relationship to persistence of electrocardiographic and echocardiographic abnormalities after aneurysmal subarachnoid hemorrhage". Stroke 2009;40:3478-3484.
6. "Prognosis of light chain amyloidosis with preserved LVEF: added value of 2D speckle-tracking echocardiography to the current prognostic staging system". J Am Coll Cardiol Img 2017;10:398-407.
7. "Impaired systolic function by strain imaging in heart failure with preserved ejection fraction". J Am Coll Cardiol 2014;63:447-456.
8. "Strain echocardiography and functional capacity in asymptomatic primary mitral regurgitation with preserved ejection fraction". J Am Coll Cardiol 2016;68:1974-1986.
9. "Global longitudinal strain is a superior predictor of all-cause mortality in heart failure with reduced ejection fraction". J Am Coll Cardiol Img 2015;8:1351-1359.
10. "Incremental prognostic utility of left ventricular global longitudinal strain in hypertrophic obstructive cardiomyopathy patients and preserved left ventricular ejection fraction". J Am Heart Assoc 2017;6:
11. "Surgical risk as related to time of intervention in the repair of intracranial aneurysms". J Neurosurg 1968;28:14-20.
12. "Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning". Neurosurgery 1980;6:1-9.
13. "Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging". J Am Soc Echocardiogr 2015;28:1-39.
14. "Global longitudinal strain to predict mortality in patients with acute heart failure". J Am Coll Cardiol 2018;71:1947-1957.
15. "Sex- and method-specific reference values for right ventricular strain by 2-dimensional speckle-tracking echocardiography". Circ Cardiovasc Imaging 2016;9:003866.
16. "Effect of combined systolic and diastolic functional parameter assessment for differentiation of cardiac amyloidosis from other causes of concentric left ventricular hypertrophy". Circ Cardiovasc Imaging 2013;6:1066-1072.
17. "Clinical prediction models for aneurysmal subarachnoid hemorrhage: a systematic review". Neurocrit Care 2013;18:143-153.
18. "Cardiac dysfunction after aneurysmal subarachnoid hemorrhage: relationship with outcome". Neurology 2014;82:351-358.
19. "Cardiovascular predictors of in-patient mortality after subarachnoid hemorrhage". Neurocrit Care 2006;5:102-107.
20. "Definition of initial grading, specific events, and overall outcome in patients with aneurysmal subarachnoid hemorrhage. A survey". Stroke 1994;25:1623-1627.
21. "Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers". Stat Med 2011;30:11-21.
22. "Differences between Takotsubo cardiomyopathy and reverse Takotsubo cardiomyopathy associated with subarachnoid hemorrhage". Int J Cardiol Heart Vasc 2016;11:99-103.
23. "Clinical and echocardiographic characteristics of acute cardiac dysfunction associated with acute brain hemorrhage- difference from Takotsubo cardiomyopathy". Circ J 2016;80:2026-2032.
24. "Relationship between cardiac dysfunction and cerebral perfusion in patients with aneurysmal subarachnoid hemorrhage". Neurocrit Care 2016;24:202-206.
25. "Neurocardiac injury after cerebral and subarachnoid hemorrhages". Cardiology in review 2017;25:89-95.
26. "Neurogenic stunned myocardium following acute subarachnoid hemorrhage: pathophysiology and practical considerations". J Intensive Care Med 2015;30:318-325.
27. "Acute neurocardiogenic injury after subarachnoid hemorrhage". Circulation 2005;112:3314-3319.
28. "Relation of elevation in cardiac troponin I to clinical severity, cardiac dysfunction, and pulmonary congestion in patients with subarachnoid hemorrhage". Am J Cardiol 2008;102:1545-1550.
29. "Reliability and feasibility of longitudinal AFI global and segmental strain compared with 2D left ventricular volumes and ejection fraction: intra- and inter-operator, test-retest, and inter-cycle reproducibility". Eur Heart J Cardiovasc Imaging 2015;16:642-652.
30. "Cardiac troponin I predicts myocardial dysfunction in aneurysmal subarachnoid hemorrhage". J Am Coll Cardiol 2000;36:1328-1335.
31. "Left ventricular dysfunction and cerebral infarction from vasospasm after subarachnoid hemorrhage". Neurocrit Care 2010;13:359-365.
32. "Contractile behavior of the left ventricle in diastolic heart failure: with emphasis on regional systolic function". Circulation 2006;113:296-304.
33. "Speckle tracking analysis allows sensitive detection of stress cardiomyopathy in severe aneurysmal subarachnoid hemorrhage patients". Intensive Care Med 2016;42:173-182.
34. "Incremental prognostic utility of left ventricular global longitudinal strain in asymptomatic patients with significant chronic aortic regurgitation and preserved left ventricular ejection fraction". J Am Coll Cardiol Img 2018;11:673-682.
35. "Prognostic value of right ventricular longitudinal peak systolic strain in patients with pulmonary hypertension". Circ Cardiovasc Imaging 2012;5:628-636.
36. "Impact of right ventricular involvement on the prognosis of Takotsubo cardiomyopathy". Eur Heart J Cardiovasc Imaging 2016;17:210-216.
37. "Tako-tsubo cardiomyopathy in aneurysmal subarachnoid hemorrhage: an underappreciated ventricular dysfunction". J Neurosurg 2006;105:264-270.
38. "Takotsubo cardiomyopathy in aneurysmal subarachnoid hemorrhage: Institutional experience and literature review". Clin Neurol Neurosurg 2016;141:65-70.
39. "Takotsubo cardiomyopathy and neurogenic stunned myocardium: similar albeit different". Eur Heart J 2016;37:2830-2832.
Footnotes
Supported by U.S. National Institutes of Health grant R01NR04221. Dr. Friedlander is the Chief Medical Officer and a stockholder of Neubase Therapeutics; and is on the Scientific Advisory Board and a stockholder of Diffusion, Inc. Dr. Gorcsan has received research grants from Medtronic, GE Healthcare, Biotronik, Hitachi, and EBR Systems. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
