Elevated Cardiac Troponin I and Relationship to Persistence of Electrocardiographic and Echocardiographic Abnormalities After Aneurysmal Subarachnoid Hemorrhage
Background and Purpose— Cardiac injury persistence after aneurysmal subarachnoid hemorrhage (aSAH) is not well described. We hypothesized that post-aSAH cardiac injury, detected by elevated cardiac troponin I (cTnI), is related to aSAH severity and associated with electrocardiographic and structural echocardiographic abnormalities that are persistent.
Methods— Prospective longitudinal study was conducted of patients with aSAH with Fisher grade ≥2 and/or Hunt/Hess grade ≥3. Serum cTnI was collected on Days 1 to 5; cohort dichotomized into peak cTnI ≥0.3 ng/mL (elevated) or cTnI <0.3 ng/mL. Relationships among cTnI and aSAH severity, 12-lead electrocardiography early (≤4 days) and late (≥7 days), Holter monitoring on Days 1 to 5, and transthoracic echocardiogram (left ventricular ejection fraction and regional wall motion abnormalities) early (Days 0 to 5) and late (Days 5 to 12) were evaluated.
Results— Of 204 subjects, 31% had cTnI ≥0.3 ng/mL. cTnI ≥0.3 ng/mL was incrementally related to aSAH severity by admission symptoms (Hunt/Hess P=0.001) and blood load (Fisher P=0.028). More patients with cTnI ≥0.3 ng/mL had prolonged QTc on early (63% versus 30%, P<0.0001) and late electrocardiography (24% versus 7%, P=0.024). On Holter monitoring, more patients with cTnI ≥0.3 ng/mL had ventricular tachycardia/fibrillation (22% versus 9%, P=0.018) but not atrial fibrillation/flutter (P=0.241). Cardiac troponin I ≥0.3 ng/mL was associated with both early ejection fraction <50% (44% versus 5%, P<0.0001) and regional wall motion abnormalities (44% versus 4%, P<0.0001). Regional wall motion abnormalities predominated in basal and midventricular segments and persisted to some degree in 73% of patients affected, whereas ejection fraction <50% persisted in 59% of patients affected.
Conclusions— Cardiac injury is incrementally worse with increasing aSAH severity and associated with persistent QTc prolongation and ventricular arrhythmias. Regional wall motion abnormalities and depressed ejection fraction persist to some degree in the majority of those affected.
- cardiac arrhythmia
- cardiac troponin I
- left ventricular ejection fraction
- neurocritical care
- subarachnoid hemorrhage
- wall motion abnormality
Patients with aneurysmal subarachnoid hemorrhage (aSAH) exhibit cardiac injury prevalence ranging from 17% to 40%.1,2⇓ A prevailing hypothesis is that a catecholamine surge3 at aneurysm rupture from cardiac sympathetic nerve endings causes subendocardial contraction band necrosis.4 Elevations in cardiac troponin I (cTnI) after aSAH are noted, but the relationship to electrocardiographic (ECG) or mechanical cardiac abnormalities and their persistence over time in a prospective sample has not been well described. We sought to determine the relationship between cTnI elevations (≥0.3 ng/mL) and ECG abnormalities over time (12-lead ECG at ≤4 days and ≥7 days, ongoing Holter monitoring on Days 1 to 5) and cardiac mechanical dysfunction (depressed ejection fraction and wall motion abnormalities by transthoracic echocardiogram early [Days 0 to 5] and later [Days 5 to 12]) during hospitalization.
Setting and Sample
The University of Pittsburgh Institutional Review Board approved this prospective, longitudinal study (National Institutes of Health R01HL074316). Informed consent was obtained from patients or proxy. From May 2004 to August 2008, we prospectively enrolled 239 consecutive patients admitted to the University of Pittsburgh Medical Center’s NeuroVascular Intensive Care Unit with aSAH documented by digital subtraction angiography and/or CT angiography. All were adults (ages 18 to 75 years) with a Fisher score of ≥2 and/or a Hunt/Hess grade ≥3 as assigned by the attending neurosurgeon. Exclusion criteria were pre-existing neurological disease, traumatic or mycotic aneurysms, or recent myocardial infarction. Enrollment rate was 84% of patients approached.
Aneurysms were secured in the operating room with surgical clips or the interventional radiology suite with endovascular embolization. Routine care included prophylactic nimodipine (generally 60 mg every 4 hours) and magnesium. Blood pressure was managed with antihypertensives (labetalol, nicardipine, nitroprusside) or vasopressor/inotropics (phenylephrine, norepinephrine, dopamine, dobutamine) to maintain systolic blood pressure <140 mm Hg before aneurysm-securing and >140 mm Hg afterward. Normovolemia was maintained with crystalloids and albumin guided by central venous pressure monitoring. Outcome was assessed on 3-month interview using the Glasgow Outcome Scale in 149 patients (55 not assessed due to attrition) and dichotomized as good (moderate to no disability) or poor (severe disability, vegetative state, death).
Cardiac Troponin I
Cardiac troponin I was measured with a fluorescent enzyme immunoassay (Bayer Health Care, Tarrytown, NY) for Days 0 to 5 after aSAH. Peak cTnI was used as a continuous variable and also dichotomized as abnormal for levels ≥0.3 ng/mL by local clinical criteria.
Cardiac Electric Abnormality
Twelve-lead ECGs were obtained ≤4 days after admission (n=204) and repeated ≥7 days later (n=91). A cardiologist (J.M.F.) analyzed ECGs by manually measuring RR, PR, QRS, and QT intervals and averaging 3 beats excluding U-waves from QT interval measurement and rejecting bundle branch block from QT determinations. Measurement was based on high-resolution, electronic ECGs using Cardio Calipers Version 3.3 (Iconico, Inc, New York, NY). QTc intervals were calculated using the Bazett correction. In the setting of atrial arrhythmias, intervals were averaged over 5 beats. Prolonged QTc was defined as ≥470 ms.
Continuous Holter monitoring was initiated on consent (recording goal ≥48 hours; available in 190 patients). A cardiologist blinded to cTnI levels and echocardiograms analyzed Holter output for cardiac arrhythmia or ectopy, which were categorized as: (1) atrial tachycardia; (2) atrial flutter; (3) atrial fibrillation; (4) nonsustained supraventricular tachycardia; (5) third-degree atrioventricular block; (6) premature ventricular contractions; (7) ventricular couplets; (8) ventricular bigeminy/trigeminy; (9) nonsustained ventricular tachycardia (≥3 beats); (10) ventricular tachycardia (≥10 seconds); (11) torsades de pointes; and (12) ventricular fibrillation. Patients could have >1 arrhythmia category. Items 2 and 3 were combined to categorize atrial fibrillation/flutter, and items 9 through 12 were combined to categorize ventricular tachycardia/fibrillation (VT/VF).
Two-dimensional transthoracic echocardiography (Vivid 7 GE-Vingmed, Horten, Norway) was performed early (Day 0 to 5 after subarachnoid hemorrhage) and repeated later (Day 5 to 12) in a patient subset (early n=125, late n=106; not obtained on total sample due to delay in beginning this procedure in the protocol, late study not always obtained due to patient discharge or death). Left ventricular ejection fraction (EF) was assessed with the biplane Simpson’s rule5 using manual tracing of apical 4- and 2-chamber views in a blinded manner.6 Interobserver and intraobserver variability for absolute difference of EF was 2.9%±2.1% and 2.0%±1.3%, respectively. EF determinations were not possible in a small subset of 8 studies (3.5%), 3 early and 5 late, in which apical views could not be obtained. EF ≥50% defined normal. Stroke volume and cardiac output were assessed using pulsed Doppler of the left ventricular outflow tract from apical 5-chamber or apical long axis views.7 All measurements were performed in triplicate and average values reported. Wall motion score was calculated using a standard 16-segment model. Segment scores were: 1=normal, 2=hypokinesis, 3=akinesis, and 4=dyskinesis. Wall motion score index was calculated as a sum of all segment scores divided by the number of segments visualized.8 Wall motion score index=1 was categorized as normal; wall motion score index >1 was categorized as a regional wall motion abnormality (RWMA). Wall motion scoring was performed by 2 experienced observers and values in agreement reported.
Because electrolyte imbalance may be a confounding cause of arrhythmia, daily serum samples were analyzed for potassium, magnesium, and calcium (uncorrected). Values at admission and 24 and 48 hours were analyzed as continuous variables.
Comparisons between subjects with and without cTnI ≥0.3 ng/mL were evaluated by χ2 analyses, Student t tests, or nonparametric tests where appropriate. Blocked sequential multivariate logistic regression analysis was performed to assess the likelihood of cTnI ≥0.3 ng/mL to predict poor Glasgow Outcome Scale controlling for age, race, gender, and Hunt/Hess grade.
In 204 patients with both cTnI and an admission ECG, 31% exhibited a peak cTnI ≥0.3 ng/mL (Table 1). Patients with cTnI ≥0.3 ng/mL were older (P=0.001), but both groups were relatively young. There was no difference in cTnI due to gender (P=0.697) or race (P=0.236). Few patients in either group had a coronary artery disease history nor were there differences in smoking, hypertension, or home medication history. More patients with cTnI <0.3 ng/mL had aneurysm clipping (45% versus 33%) and less embolization (55% versus 67%, P=0.016). Significantly more patients with cTnI ≥0.3 ng/mL had greater aSAH severity by clinical symptoms on admission (Hunt/Hess, P=0.001) and blood load on CT scan (Fisher grade, P=0.028), but there was no relationship with aneurysm site (P=0.295). There was a stepwise rise for mean values for cTnI at each Hunt/Hess grade for increasing severity (Figure 1). In clinical care, patients with cTnI ≥0.3 ng/mL were more likely to have vasopressor/inotrope infusions (41% versus 22%, P=0.006) and have the pressors started at an earlier time point (P=0.046). According to the 3-month Glasgow Outcome Scale, significantly more patients with cTnI ≥0.3 ng/mL had a poor outcome, and in regression analysis, cTnI ≥0.3 ng/mL remained a significant predictor of poor outcome after controlling for age, gender, race, and Hunt/Hess grade (OR, 2.61; 95% CI, 1.15 to 6.46; P=0.039).
Early 12-lead ECG (mean, 0.87±0.89 days) showed no between-group differences (Table 2) in PR intervals or QRS durations, but mean QTc was longer for patients with cTnI ≥0.3 ng/mL (485±48 ms versus 452±37 ms, P<0.0001). At a 470-ms cut point, significantly more patients with cTnI ≥0.3 ng/mL had prolonged QTc (63% versus 30%, P<0.0001). On late 12-lead ECG, the mean QTc interval decreased for both groups, but still remained longer for cTnI ≥0.3 ng/mL patients (443±34 ms versus 429±31 ms), although significance was not maintained (P=0.055). On late ECG, prolonged QTc ≥470 ms prevalence remained higher for patients with cTnI ≥0.3 ng/mL (24% versus 7%, P=0.024).
Holter monitoring data in 190 patients (Table 3) indicated overall atrial fibrillation/flutter prevalence of 7% (n=13) not present on admission but no between-group difference (P=0.241). There was an overall VT/VF prevalence, including nonsustained Vt of 13% (n=25). Significantly more patients with cTnI ≥0.3 ng/mL displayed VT/VF (22% versus 9%, P=0.018), but few of these had sustained VT/VF (n=2; one VT, one VF). Although patients with cTnI ≥0.3 ng/mL had a statistically lower uncorrected calcium level on admission (8.6±.9 versus 8.9±.7, P=0.043) and 24 hours (8.3±.8 versus 8.6±.7, P=0.006), the values were clinically similar. Otherwise, electrolyte values were similar between groups and within normal limits (data not shown).
Early echocardiograms were available in 125 patients (Table 2). Early mean EF was significantly lower in patients with cTnI ≥0.3 ng/mL (52%±12% versus 63%±7%, P<0.0001). When dichotomizing depressed EF as <50%, significantly more patients with cTnI ≥0.3 ng/mL had early EF <50% (44% versus 5%, P<0.0001). Both early stroke volume and cardiac output were significantly less for patients with cTnI ≥0.3 ng/mL (P=0.007 and P=0.012, respectively). The mean duration between the early and late echocardiograms was 4.5 days (range, 2 to 9 days). Late echocardiograms were available in 106 patients, and the mean EFs for each cTnI group remained almost identical to their early value, maintaining significance (P<0.001). On late echocardiogram, the percentage of patients with cTnI ≥0.3 ng/mL with EF <50% had decreased somewhat but was still persistent in 33%. Stroke volume on the late echocardiogram remained significantly less for patients with cTnI ≥0.3 ng/mL (P=0.039). The overall RWMA prevalence was 15% on the early echocardiogram (19 of 125), and patients with cTnI ≥0.3 ng/mL were significantly more likely to exhibit a RWMA (44% versus 4%, P<0.0001). On the late echocardiogram, the RWMA prevalence and between-group difference had changed very little from the early study (41% versus 3%, P<0.0001).
The 19 patients with early RWMA were evaluated on their characteristics and course (Table 4). Their mean time from injury for the early echocardiogram was 2.7 days (range, 0 to 5 days) and 6.7 days (range, 5 to 12 days) for the late echocardiogram. cTnI was ≥0.3 ng/mL in 15 of the 19 patients (mean cTnI for patients with RWMA 6.5±10.3 ng/mL versus 0.3±0.7 ng/mL for 106 patients without RWMA, P=0.017). The peak cTnI exceeded 1.0 ng/mL in 13 of 19 patients with RWMA (68%), which represents 30% of all patients in the sample with cTnI ≥1.0 ng/mL. The RWMA persisted in 14 of the 19 (74%), and although improvement was seen in 68% (13 of 19), only 26% (5 of 19) had returned to normal at the late echocardiogram, whereas 21% (4 of 19) has worsened. Only 2 patients with early RWMA had an early EF >50%, and patients with the highest wall motion score index indices had the lowest EFs as expected. Nevertheless, EF depression was modest; at baseline, 14 had EF of 40% to 50%, whereas only 3 had EF 20% to 30%. Of the 17 patients with early EF <50%, the finding persisted at the late study in 10 (59%). Early QTc was prolonged in the majority of patients with RWMA (14 of 19), and half required vasopressor/inotrope infusions (10 of 19). Of 13 patients with 3-month outcome data, only 3 had good functional recovery, whereas 2 had severe disability, 6 had moderate disability, and 2 died.
The prevalence of dysfunction within each of the 16 regional segments for the 19 patients with RWMA is illustrated in Figure 2. In the early evaluation, prevalence predominated in the basal and midventricular areas of the anteroseptal and anterolateral regions with the apex affected in only approximately one fourth. At the late study, the basal area appeared more likely to recover followed by the midventricular area, whereas wall motion was less improved at the apex.
This article validates that cardiac injury is common after aSAH and was observed in 31% of our young patient population with minimal cardiac disease history. Although our elevated cTnI prevalence is lower than reported by some others (Naidech et al, 42%3; Rampappa et al, 37%9), our data demonstrate cardiac injury prevalence across our entire prospectively evaluated aSAH population (study enrollment rate of 84% and cTnI sampled on nearly all enrolled). Because we did not restrict cTnI evaluation to patients with clinical suspicion of cardiac injury, this study is unique in assessing for evidence of myocardial damage in a large, consecutive heterogeneous group of patients with aSAH.
Our findings corroborate that elevated cTnI prevalence increases as aSAH severity increases.9,10⇓ We also demonstrate that the cTnI rises in proportion to the degree of aSAH severity in a stepwise fashion (ie, the higher the Hunt/Hess grade, the higher the mean cTnI value). If, as hypothesized, a catecholamine surge at rupture causes neurocardiac injury,11 then our finding suggests that the degree of surge corresponds to aSAH severity. Nevertheless, not all cases fit this hypothesis. The cTnI <0.3 ng/mL group included 16% with Hunt/Hess Grades 4 to 5 or Fisher Grade 4, whereas conversely, 20% of patients with elevated cTnI ≥0.3 ng/mL had low Hunt/Hess or Fisher grades. Thus, variation in catecholamine sensitivity or other factors might explain nonuniformity of cTnI elevations in response to aSAH severity for all cases, warranting further exploration.
We also corroborate frequently prolonged QTc on admission ECG, occurring more commonly in patients with elevated cTnI.12,13⇓ Although our overall 40% prevalence of QTc prolongation is higher than noted by some others,12–14⇓⇓ our use of individual visualization, averaging, and rejection of bundle branch block for QTc determination rather than reliance on ECG software may reflect more accurate prevalence assessment. We demonstrated that not only is prolonged QTc significantly more prevalent in patients with cTnI elevation (60%), but it persists in 40% of patients at the late ECG. We further demonstrate that cTnI elevation is associated with dynamic cardiac arrhythmias. Although Frontera15 reported a lower overall prevalence of clinically significant arrhythmia (4%), our methodology of Holter monitoring may have more effectively captured arrhythmia. Kawahara16 noted increased ventricular ectopy, but did not specifically report on VT/VF. We importantly noted a relatively high overall Vt/VF prevalence in patients with elevated cTnI, which, although only infrequently sustained, might contribute to transient, subtle brain perfusion abnormalities.
We previously reported RWMA prevalence after aSAH of 17% with the basal anterior and anterospetal segments most affected.8 Our current study indicates that not only is elevated cTnI associated with both depressed EF and RWMA, but that, although for the most part mild, both persist to some degree in nearly half of those who develop. Our late echocardiography noted that although 68% of patients with RWMA showed some improvement, only 26% had normalized. Kovethal17 reported that 70% of patients with RWMA had improved by 5 days postaSAH but did not report if return was to normal. Khush18 indicated that 75% to 90% of patients with RWMA improved at 5 days, again not specifying if improvement indicated normalization. Sugimoto19 reported on 11 patients with RWMA, all normalizing by Day 10. Our data demonstrated that at late echocardiogram (mean, 7 days; range, 5 to 12 days), most RWMA had improved but failed to normalize. This finding is consistent with myocardial stunning behavior wherein the time needed for full recovery ranges from 14 to 90 days, perhaps making it prudent to follow patients with RWMA with a repeat echocardiogram at 6 to 12 weeks to determine if dysfunction has normalized as expected with stunning.20 Our finding that the segmental abnormality distribution at the early study is dominant at the base and midventricular areas is interesting and not consistent with apical ballooning attributed to an intense catecholamine surge with severe stress known as tako-tsubo cardiomyopathy.11 Nevertheless, our findings of greater RWMA predominance at the base and midventricular areas in early evaluation is similar to aSAH population findings by Sugimoto19 and Zaroff,21 who hypothesized that lesser predominance of sympathetic nerve endings at the apex might be explanatory. However, we noted that the apex, although less affected early, demonstrated a slower return to normalcy as compared with other segmental areas, and to our knowledge, we are the first to report on later evaluation of regional segments. We do not know if this trend would persist at an even later follow-up period.
Consistent with our findings, the neuroscience literature indicates that neurocardiac injury contributes to poorer outcomes independent from subarachnoid hemorrhage severity alone,2,14,22–24⇓⇓⇓⇓ but mechanisms remain unclear. Our data demonstrate that the relationship between elevated cTnI and arrhythmia is present but limited, and depression of EF and stroke volume is, although present and persistent, for the most part modest. Nevertheless, even modest and transient cardiac functional abnormalities may impinge on perfusion to the area of brain injury. Another possible explanation is that other coexisting processes such as inflammation may contribute to damage.25,26⇓ For example, elevated cTnI has been noted in other critically ill populations including sepsis27 and acute respiratory distress syndrome,28 diseases wherein sustained stress and inflammation are implicated. However, an inflammatory hypothesis would not explain the RWMA distribution.
Our study limitations include inability to obtain echocardiograms on every patient or late studies for all with early ECGs; variability of duration between time from injury and the early echocardiogram; and variability in the time window between early and late echocardiograms. It also would have been interesting (although not a part of our protocol) to have longer term follow-up by echocardiogram to identify the trajectory of recovery over a longer time period.
Cardiac injury occurs commonly after aSAH, and the persistence of electrocardiographic and echocardiographic manifestations may be longer than previously thought. Vigilance to cardiac injury consequences after aSAH should extend throughout hospitalization and possibly beyond in those patients affected.
Sources of Funding
Supported by the National Institutes of Health National Heart, Lung and Blood Institute R01HL074316 (M.H., E.A.C., J.G., Y.C., M.B.H.) and the American Heart Association AHA0725482U (J.M.F.).
- Received April 30, 2009.
- Revision received June 5, 2009.
- Accepted June 17, 2009.
- ↵Naidech AN, Kreiter KT, Janjua N, Ostapkovich ND, Parra A, Commichau C, Fitzsimmons BM, Connolly S, Mayer S. Cardiac troponin elevation, cardiovascular morbidity, and outcome after subarachnoid hemorrhage. Circulation. 2005; 112: 2851–2856.
- ↵Zaroff JG, Pawlikowska L, Miss JC, Yarlagadda S, Ha C, Achrol A, Kwok P, McCulloch CE, Lawton MT, Ko N, Smith W, Young WL. Adrenoreceptor polymorphisms and the risk of cardiac injury after subarachnoid hemorrhage. Stroke. 2006; 37: 1680–1685.
- ↵Lang RM, Beirig M, Devereux RB, Flachskampf FA, Foster E, Pelikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ; Chamber Quantification Writing Group; American Society of Echocardiography’s Guidelines and Standards Committee; European Association of Echocardiography. Recommendations for chamber quantification: a report from the Am Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group. J Am Soc Echocardiogr. 2005; 18: 1440–1463.
- ↵Cannesson M, Tanabe M, Suffoletto MS, McNamara DM, Madan S, Lacomis JM, Gorcsan J. A novel two-dimensional echocardiographic image analysis system using artificial intelligence-learned pattern recognition for rapid automated ejection fraction. J Am Coll Cardiol. 2007; 49: 217–326.
- ↵Tung P, Kopelnik A, Banki N, Ong K, Ko N, Lawton MT, Gress D, Drew B, Foster E, Parmley W, Zaroff J. Predictors of neurocardiac injury after subarachnoid hemorrhage. Stroke. 2004; 35: 548–553.
- ↵Schuiling WJ, Dennesen PJ, Tans JT, Algra A, Rinkel GJ. Troponin I in predicting cardiac or pulmonary complications and outcome in subarachnoid hemorrhage. J Neurol Neurosurg Psychiatry. 2005; 76: 1565–1569.
- ↵Mayer SA, Lin J, Homma S, Solomon RA, Lennihan L, Sherman D, Fink ME, Beckford A, Klebanoff LM. Myocardial injury and left ventricular performance after subarachnoid hemorrhage. Stroke. 1999; 30: 780–786.
- ↵Mashaly HA, Provencio JJ. Inflammation as a link between brain injury and heart damage: the model of subarachnoid hemorrhage. Cleveland Clinic J Med. 2008; 75 (suppl 2): S26–S30.