Journal of Cardiac Failure Vol. 20 No. 1 2014
Consensus
Diagnostic Utility of Cardiac Biomarkers in Discriminating Takotsubo Cardiomyopathy From Acute Myocardial Infarction MANDEEP SINGH RANDHAWA, MD,1 ASHWAT SINGH DHILLON, MD,1 HARRIS C. TAYLOR, MD,2 ZHIYUAN SUN, MS,3 AND MILIND Y. DESAI, MD4 Cleveland, Ohio
ABSTRACT Background: Takotsubo cardiomyopathy (TC) mimics acute myocardial infarction (AMI). We postulated that ventricular dysfunction in TC in the absence of significant myocardial necrosis would produce higher B-type natriuretic peptide (BNP)/troponin T (TnT) and BNP/creatine kinase MB fraction (CKMB) ratios than in AMI. Methods and Results: We studied 58 consecutive TC (age 65.8 6 82.9) and 97 AMI patients (age 59.8 6 83.4). The ratios of BNP/TnT and BNP/CKMB were calculated with the use of first simultaneously drawn laboratory values. Receiver operating characteristic curves were used to distinguish TC from AMI with 95% specificity based on cardiac biomarker ratios. Median BNP/TnT and BNP/CKMB ratios were, respectively, 1,292 [interquartile range 443.4e2,657.9] and 28.44 [13.7e94.8] in the TC group and 226.9 [69.91e426.32] and 3.63 [1.07e10.02] in the AMI group (P ! .001). TC can be distinguished from AMI with 95% specificity with the use of BNP/TnT ratio $1,272 (sensitivity 52%) and BNP/CKMB ratio $29.9 (sensitivity 50%). Conclusions: The value of BNP is significantly higher in TC than in AMI. Early BNP/TnT and BNP/ CKMB ratios help to differentiate TC from AMI with greater accuracy than BNP alone. (J Cardiac Fail 2014;20:2e8) Key Words: Takotsubo cardiomyopathy, cardiac biomarkers, electrocardiogram and acute myocardial infarction.
Takotsubo cardiomyopathy (TC) is a disorder of transient ventricular systolic dysfunction thought to be precipitated by severe emotional or physical stress.1e4 Echocardiography and/or ventriculogram typically shows left ventricular apical ballooning.1,5 Patients usually present with chest
pain with electrocardiographic (ECG) changes mimicking an acute myocardial infarction (AMI). Approximately 1% e2% of patients presenting with acute coronary syndrome (ACS) are ultimately diagnosed with TC.6,7 An elevated plasma concentration of B-type natriuretic peptide (BNP) is an established marker of left ventricular dysfunction.8 Both troponin T (TnT) and creatinine kinase MB fraction (CKMB) are elevated in the setting of membrane leak caused by acute myocardial necrosis, whereas the production and release of BNP is related to ventricular distention with or without myocyte necrosis.8e10 It is known that BNP and TnT are increased in both TC and AMI.10,11 However, because TC is a disease primarily causing distention of the ventricles and characterized by reversible myocardial dysfunction without necrosis, a greater increase in plasma BNP compared with TnT or CKMB has been demonstrated, compared with AMI.10,12e15 TC has also been associated with significant QTc interval prolongation on ECG.16,17
From the 1Department of Internal Medicine, Cleveland Clinic, Cleveland, Ohio; 2Director Resident Research, Fairview Hospital - Cleveland Clinic, Clinical Professor of Medicine, Case Western Reserve University School of Medicine, Cleveland, Ohio; 3Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland, Ohio and 4Heart and Vascular Institute, Cleveland Clinic, Cleveland, Ohio. Manuscript received August 10, 2013; revised manuscript received December 8, 2013; revised manuscript accepted December 9, 2013. Reprint requests: Milind Y. Desai, MD, Tomsich Department of Cardiovascular Medicine, Heart and Vascular Institute, Cleveland Clinic, 9500 Euclid Ave, Desk J1-5, Cleveland, OH 44195. Tel: 216-445-5250; Fax: 216-445-6155. E-mail: [email protected] See page 8 for disclosure information. 1071-9164/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cardfail.2013.12.004
2
Biomarkers and Takotsubo Cardiomyopathy
Currently, TC is a diagnosis of exclusion according to the modified Mayo Clinic Criteria, which involve performing coronary angiography to confirm the absence of obstructive coronary artery disease (CAD).18 In the present study, we sought to determine if we could improve our noninvasive ability to distinguish between TC and AMI with the use of cardiac biomarker ratios obtained early during presentation. Methods This was a retrospective cohort analysis (matching for cases and respective control subjects) done at a single tertiary care center (Cleveland Clinic Foundation, Cleveland, Ohio), after appropriate Institutional Review Board approval. We studied 225 consecutive TC patients from the period of January 2006 to February 2012 based on the following criteria of diagnosis: acute presentation with ACS-like symptoms, with acute ST-segment and T-wave changes on the ECG and rise in the cardiac enzymes (CKMB and TnT), typical echocardiographic and angiographic findings (ballooning of the left ventricle), complete reversibility of ejection fraction (EF) based on repeated transthoracic echocardiography, and absence of significant (O50%) CAD and acute plaque rupture based on angiography. To be included in the study, a patient had to have had biomarkers determined at the time of admission. We excluded the following patients: !18 years old, known chronic kidney disease (CKD) stage IIIeV with estimated glomerular filtration rate !60 mL/min according to the Modification of Diet in Renal Disease equation,19 acute kidney injury (AKI) at time of admission according to the criteria proposed by the Acute Dialysis Quality Initiative group,20 history of congestive heart failure (CHF) with a history of left ventricular ejection fraction (LVEF) !45% on the last available echocardiogram, known nonischemic cardiomyopathy, microbial and toxic myocarditis, septic shock, acute pulmonary embolism, and severe aortic stenosis on the last available echocardiogram. A total of 225 consecutive patients with TC were studied. 118 patients were excluded because of unavailability of simultaneously drawn first available BNP, CKMB, and TnT. Of the others, 22 patients did not meet our diagnostic criteria for TC. Sixteen patients had preexisting CKD. Eight patients were excluded because of AKI. Two patients were excluded because of preexisting CHF. One patient was excluded because of concurrent chemotherapy that was potentially cardiotoxic. After applying the exclusion criteria, 58 patients (case subjects) were selected, and their baseline clinical characteristics were recorded by manually reviewing their electronic medical records. Subsequently, a control group of patients that met the criteria of AMI based on the Joint European Society of Cardiology/American College of Cardiology Committee for the Redefinition of Myocardial Infarction were assembled.21 Ninety-seven such AMI patients were selected to keep the study power O90% to show a significant difference in the 2 groups of TC and AMI. The QTc interval duration was recorded and initial ECG was analyzed for acute ST-segment and T-wave changes, and both, case and control subjects were divided into ST-segment elevation and noneST-segment elevation groups. QTc was deemed to be prolonged if it was O470 ms.22,23 These variables were followed for 72 hours after admission. All case and control subjects underwent coronary angiography during hospitalization in a standard
Randhawa et al
3
fashion. The degree of coronary obstruction was deemed to be significant in the setting of O50% stenosis in a major epicardial vessel. Although it is true that atypical variants of TC are less clinically confusing with coronary artery disease, our focus was to correlate the first available cardiac biomarkers regardless of echocardiographic findings. We did not study echocardiographic data for global versus segmental dysfunction. The study was done to retrospectively compare initial cardiac biomarker profiles in TC and AMI. Laboratory Measurements We recorded first simultaneously available levels of TnT, CKMB, and BNP after presentation. All measurements for TnT and CKMB were done by electrochemiluminescence immunoassay technique with the use of the Roche Cobas 6000e analyzer (Roche Diagnostics). All determinations of BNP were done by chemiluminescence immunoassay technique with the use of the Advia Centaur analyzer (Siemens Medical Solutions Diagnostics). As such, we standardized the laboratory analysis for all study subjects. Statistical Methods Continuous measures were described with the use of mean and standard deviation or median and interquartile range (IQR) and were compared between groups with the use of standard t testing or Wilcoxon rank sum test. Categoric measures were summarized as frequency and percentages and were compared between groups with the use of chi-square test. Receiver operating characteristic (ROC) analyses of absolute values of BNP, TnT, and CKMB, and 2 ratios (BNP/TnT and BNP/CKMB) were performed to assess their ability in differentiating TC case subjects from AMI control subjects. We wanted to assess whether the ratio of BNP/ TnT or BNP/CKMB would be better in distinguishing TC from AMI compared with individual cardiac biomarker values alone. Therefore, paired comparison was performed on the area under the ROC curve (AUC) of BNP, TnT, CKMB, and the ratios BNP/TnT and BNP/CKMB. In addition, analysis of ECG findings revealed that presence of QTc interval prolongation was more common in TC case subjects than in AMI control subjects. Logistic regression model was used to compare the distinguishing ability of combination of BNP/CKMB and presence of QTc interval prolongation compared with the BNP/CKMB ratio alone. All analyses were performed with the use of SAS 9.2 and R 2.12.1 software.
Results A total of 155 patients were studied from January 2006 to February 2012. Of these, 58 were diagnosed with TC (case subjects) and 97 had angiographically proven AMI (control subjects). Of the case subjects, 22/58 (38%) were found to have ST-segment elevation (STE-TC) and 36/58 (62%) had no ST-segment elevation (NSTE-TC) in their initial ECGs. In the control group, 38/97 (39%) had ST-segment-elevation MI (STEMI) and 59/97 (61%) had noneST-segment-elevation MI (NSTEMI). Within the NSTEMI subgroup of the control subjects, most patients had significant multivessel disease. There were 7 patients with significant left main disease, 26 with significant
4 Journal of Cardiac Failure Vol. 20 No. 1 January 2014 left anterior descending artery disease, 33 with significant disease in the left circumflex, 27 with significant disease in the right coronary artery, and 10 with a history of coronary bypass surgery. In the STEMI subgroup of the control subjects, 1 patient had significant left main disease, 19 had significant left anterior descending disease, 13 had significant circumflex disease, 24 had significant right coronary disease, 1 had cocaine-induced AMI, and 1 had bypass graft occlusion. No significant CAD was found on coronary angiography of TC case subjects. The vast majority of case subjects (55/58, 95%) presented with symptoms, and 3/58 (5%) case subjects developed symptoms after hospitalization. Among the control subjects, 95/97 (98%) presented with symptoms, and 2/97 (2%) developed symptoms after hospitalization. Baseline characteristics and cardiac biomarker values of the cases and controls are summarized in Tables 1 and 2. In the TC group, BNP levels ranged from 16 to 2,511 pg/mL (median 456.5 pg/mL), whereas the median BNP level in the AMI group was 97 pg/mL (P ! .0001). The median [IQR] BNP/TnT and BNP/CKMB ratios were, respectively, 1,292 [443.4e2657.9] and 28.44 [13.7e94.8] in the TC group and 226.9 [69.91e426.32] and 3.63 [1.07e10.02] in the AMI group (P ! .001). Figure 1 depicts a scatter plot of BNP and CKMB values in the study population. We found that 20/58 TC case subjects had BNP $600 pg/mL, and in 52/58 case subjects the value of CKMB was #25 ng/mL. This was in contrast to the AMI population in our study, all of whom had a first BNP level #600 pg/mL. A presumed precipitating factor was identified in 37/58 (64%) of the TC cases (Supplemental Table 1). Discrimination of Takotsubo Cardiomyopathy From Acute Myocardial Infarction
The data on incremental ability of cardiac biomarkers and their ratios to distinguish TC from AMI are presented in Table 3. TC case subjects could be distinguished from AMI control subjects with the use of the earliest BNP/ TnT ratio $1,272 with a specificity of 95% and sensitivity of 52% (95% confidence interval [CI] 0.751e0.894, AUC 0.822; P ! .001). As such, of 57 TC cases for whom CKMB data was available, the earliest BNP/CKMB ratio $29.9 distinguished TC from AMI with a specificity of
95% and sensitivity of 50% (95% CI 0.802e0.922, AUC 0.862; P ! .001). These results also show that the BNP/ CKMB ratio has greater AUC than the BNP/TnT ratio (Fig. 2). On exploring further, we found that 53% of TC patients had a BNP/CKMB ratio $25 and conversely, 89/97 (92%) of AMI patients had a BNP/CKMB ratio !25 (Fig. 3). A slightly higher BNP/CKMB ratio of $38 distinguishes TC from AMI with 99% specificity and 46% sensitivity. We also found that the ratios of BNP/CKMB as well as BNP/TnT were both significantly (P ! .001) more accurate in distinguishing TC from AMI than the value of BNP alone (Table 4). Subset Analysis of STE-TC and STEMI (Table 3)
In the ST-segment elevation group, a BNP/CKMB ratio $13.1 distinguished TC from AMI with 95% specificity and 68% sensitivity (AUC 0.888, 95% CI 0.800e0.975; P ! .001) and a BNP/TnT ratio $622 distinguished TC from AMI with 95% specificity and 52% sensitivity (AUC 0.863, 95% CI 0.764e0.963; P ! .001). Subset Analysis of NSTE-TC and NSTEMI
In the noneST-segment elevation group, a BNP/CKMB ratio $34.7 distinguished TC from AMI with 95% specificity and 56% sensitivity (AUC 0.868, 95% CI 0.795e0.942; P ! .001) and a BNP/TnT ratio $2,094 distinguished TC from AMI with 95% specificity and 54% sensitivity (AUC 0.830, 95% CI 0.735e0.925; P ! .001). On comparing the AUC of BNP/CKMB and BNP alone in both subsets (ST-segment elevation and noneSTsegment elevation) the ratio of BNP/CKMB distinguished TC from AMI with significantly higher accuracy. The same was not true for the ratio of BNP/TnT (Table 4). Discussion To the best of our knowledge, this analysis is one of the largest cohorts of TC and AMI patients studied to compare the ability of cardiac biomarker ratios to provide diagnostic yield to distinguish between these 2 disease processes with very different management strategies and outcomes.10,15,24 We exclusively assessed first simultaneously available values of cardiac biomarkers to distinguish TC from
Table 1. Baseline Clinical Characteristics Variable Age (y) BMI (kg/m2) Female, n (%) HTN, n (%) DM, n (%) HLD, n (%) T-wave inversion, n (%) QTc prolongation, n (%)
Acute Myocardial Infarction Control Subjects (n 5 97)
Takotsubo Case Subjects (n 5 58)
P Value
59.8 6 13 29.45 6 5.75 77 (79) 71 (73) 23 (24) 56 (58) 52 (57) 16 (16)
65.8 6 12.9 26.4 6 4.9 57 (98) 39 (68) 10 (18) 33 (59) 42 (79) 35 (60)
.0044 .0062 .0009 .53 .37 .89 .006 !.0001
BMI, body mass index; HLD, hyperlipidemia; HTN, hypertension; DM, diabetes mellitus.
Biomarkers and Takotsubo Cardiomyopathy
Randhawa et al
5
Table 2. Cardiac Biomarkers and Their Ratios, Median [Interquartile Range] Variable CKMB (ng/mL) TnT (ng/mL) BNP (pg/mL) BNP/TnT BNP/CKMB
Acute Myocardial Infarction Control Subjects (n 5 97)
Takotsubo Case Subjects (n 5 58)
P Value
25 [13.75e62.7] 0.52 [0.19e1.45] 97 [45.5e248.5] 226.9 [69.91e426.32] 3.63 [1.07e10.02]
10.5 [5.8e14.93] 0.38 [0.16e0.65] 456.5 [120.25e734.5] 1,292.1 [443.4e2,657.9] 28.44 [13.7e94.8]
!.0001 .0092 !.0001 !.001 !.001
CKMB, creatinine kinase MB fraction; TnT, troponin T; BNP, B-type natriuretic peptide.
AMI. Also, the first simultaneous cardiac biomarkers were drawn within 6 hours of admission in 83% of case subjects and 87% of control subjects. Moreover, we used strict inclusion and exclusion criteria for cases to reduce false positives.10,15,25,26 We further maximized consistency, choosing only those patients for whom the same laboratory methods were used for measurement of cardiac biomarkers. Our study population is a select population owing to the inherent difficulty in finding TC patients whose BNP levels are related only to their new-onset cardiac dysfunction. Levels of BNP can be affected by preexisting or coexisting clinical conditions other than acute heart failure, such as acute and chronic kidney disease, chronic heart failure, septic shock, and acute pulmonary embolism. Because accurate measurement of BNP as affected by acute heart failure secondary to TC was the key to our study, we had to exclude patients who had preexisting elevation of BNP. Therefore, the results may be of use in those patients who were known to have normal cardiac function. It must be noted that although TnT and CKMB are usually drawn simultaneously as part of a ‘‘cardiac enzyme panel,’’ BNP is often not ordered along with them. Our study was
designed to specifically assess the first available, simultaneously drawn BNP/TnT and BNP/CKMB ratios. We demonstrated that by using the simultaneous admission cardiac biomarker values, a BNP/TnT ratio O1,272 and a BNP/CKMB ratio O29.9 were both 95% specific for differentiating TC from AMI, albeit with low sensitivity (52% and 50%, respectively). Furthermore, a BNP/CKMB ratio $38 can distinguish TC from AMI with a specificity of 99% although with a slightly lower sensitivity of 46% (Fig. 3). Additionally, using the paired comparison, we demonstrate that the ratio of BNP/CKMB was significantly (P ! .001) more accurate in differentiating TC from AMI compared with the value of BNP alone (Table 4). While analyzing the ECG of TC patients, we noted that 63% manifested QTc interval prolongation within the first 72 hours after admission. TC is known to be associated with significant prolongation of the QTc interval on ECG.16,17 One earlier study found that the QTc interval was significantly prolonged in the subacute phase of TC compared with the acute and chronic phases.16 We found that QT interval prolongation was present in a higher number of TC patients compared with the AMI group. Logistic
Fig. 1. Distribution of B-type natriuretic peptide (BNP) and creatine kinase MB fraction (CKMB) in study population.
6 Journal of Cardiac Failure Vol. 20 No. 1 January 2014 Table 3. Distinguishing TC From AMI Using the Earliest Simultaneously Drawn Cardiac Biomarker Ratios and AUC Values for ROC Analysis Cutoff Specificity 95% Differentiate NSTE-TC from NSTEMI BNP BNP/TnT BNP/CKMB Differentiate STE-TC from STEMI BNP BNP/TnT BNP/CKMB Differentiate all TC from AMI BNP BNP/TnT BNP/CKMB QT prolongation þ BNP/CKMB
Sensitivity*
AUC
95% CI
P Value
442 2094 34.7
56% 54% 56%
0.770 0.830 0.868
0.667e0.874 0.735e0.925 0.795e0.942
!.001 !.001 !.001
522 622 13.1
41% 52% 68%
0.794 0.863 0.888
0.672e0.914 0.764e0.963 0.800e0.975
!.001 !.001 !.001
479 1272 29.9 0.473
50% 52% 50% 73%
0.776 0.822 0.862 0.890
0.698e0.854 0.751e0.894 0.802e0.922 0.832e0.949
!.001 !.001 !.001 !.001
AMI, acute myocardial infarction; AUC, area under the receiver operating characteristic curve; BNP, B-type natriuretic peptide (pg/mL); CKMB, creatinine kinase MB fraction (ng/mL); ROC, receiver operator curve; NSTE, noneST-segment elevation; STE, ST-segment elevation; TC, takotsubo cardiomyopathy; TnT, troponin T (ng/mL). *All sensitivities based on a specificity of 95%.
regression also revealed that the AUC for diagnosing TC was higher if both BNP/CKMB ratio and QTc interval prolongation were combined. However, subsequent paired comparison of BNP/CKMB versus the combination of BNP/CMKB and QTc prolongation did not show a significant difference (Table 4). . In practice, TC is mainly a diagnosis of exclusion, and coronary angiography is necessary to rule out obstructive coronary artery disease or acute plaque rupture. Currently available data suggest that 1%e2% of patients with symptoms of ACS have been found to have TC.18 TC has only recently begun to be recognized as a distinct disease entity. As a result, there are comparatively little data available on cardiac biomarkers in TC. Our study provides data that may
help to further characterize TC based on initial cardiac biomarker levels. . The significance of BNP in heart failure is well known.27 The N-terminal precursor of BNP (NT-proBNP) in the normal heart is mainly produced in the atrium. However, in the failing heart, there is increased ventricular production of NT-proBNP.9 TC is primarily a disorder of ventricular systolic dysfunction and not surprisingly has been found to be associated with higher circulating levels of BNP and NT-proBNP.2,25 It is also established that a rise in BNP concentration is caused by circumstances that increase ventricular wall stress, whereas troponin and CKMB are elevated in the setting of myocardial necrosis.12 Recent studies have shown that TC presents with an elevation of natriuretic peptides out of proportion to troponin levels.15,18 Frohlich et al found that the use of the NT-proBNP/peak TnT ratio appeared to be most accurate to distinguish ACS from TC.15 Following these findings, we decided to assess only the first available simultaneously drawn laboratory values of BNP, TnT, and CKMB in patients presenting with symptoms suggestive of ACS and to determine whether TC could be reliably distinguished from AMI early in the disease course. Finally, despite the difficulty in finding simultaneously available biomarkers and in using strict criteria to limit the number of patients with falsely elevated BNP values, we were able to include 58 TC patients in our study. Though not obviating the need for cardiac catheterization, we think that these results are sufficiently encouraging to justify an adequately powered prospective study to determine more definitively whether biomarkers and a positive discriminant, which combines biomarkers and QTc interval, can reliably distinguish these 2 disorders noninvasively at the time of presentation.
Fig. 2. Receiver operator characteristic curve analysis for the study population to help distinguish between acute myocardial infarction and Takotsubo cardiomyopathy. See Tables 3 and 4 for further details.
Study Limitations
A number of limitations in the present study need to be acknowledged. In addition to our study being retrospective
Biomarkers and Takotsubo Cardiomyopathy
Randhawa et al
7
Fig. 3. Distribution of BNP/CKMB ratio in the study population. Abbreviations as in Figure 1.
with the attendant limitations, we had to exclude as many as 118 of 225 TC case subjects because of lack of simultaneously drawn biomarkers. Second, despite our attempts to match case and control subjects according to sex, 98% of TC case subjects were female whereas 79% of AMI control subjects were female. Nonetheless, it should be recognized that TC occurs predominantly in postmenopausal women.1,28 Similarly, it was not possible to eliminate a significant difference in age between the case and control subjects. Third, the study is not applicable to patients who have
Table 4. Paired Comparison of AUC in Study Population Estimated Difference [95% CI] NoneST-segment elevation BNP vs BNP/TnT BNP vs BNP/CKMB BNP/TnT vs BNP/CKMB ST-segment elevation BNP vs BNP/TnT BNP vs BNP/CKMB BNP/TnT vs BNP/CKMB Overall BNP vs BNP/TnT BNP vs BNP/CKMB BNP/TnT vs BNP/CKMB BNP/CKMB vs QT prolongation þ BNP/CKMB Abbreviations as in Table 3.
P Value
0.060 [0.137e0.017] 0.098 [0.160 to 0.036] 0.038 [0.101e0.025]
.13 .002 .23
0.069 [0.173e0.035] 0.093 [0.187e0] 0.024 [0.074e0.026]
.19 .050 .34
0.046 0.086 0.040 0.028
[0.107e0.014] .14 [0.136 to 0.035] !.001 [0.081e0.002] .059 [0.065e0.008] .13
underlying renal dysfunction or heart failure (exclusion criteria). Although it is true that levels of BNP are age and sex dependent, we do not consider this to be a limitation because it applies to both TC and AMI.29 Fourth, although QTc interval prolongation was present on the first available ECG in 43% of cases, we followed serial ECGs and recorded QTc interval prolongation if present at any time in the first 72 hours after admission. Therefore, the QTc interval used to calculate the AUC in our study may not have been concurrent with the original BNP, CKMB, or TnT. Finally, although we did not include patients with known baseline QTc interval prolongation, subsequent drug-induced QTc interval prolongation in both case and control subjects can not be completely ruled out. However, the main emphasis of this study was on the cardiac biomarker profiles in TC and AMI, rather than the QTc interval. Conclusion The ratios of BNP/TnT and BNP/CKMB, obtained early during admission, can be used as independent indices to help differentiate TC from actual AMI with a high specificity in the appropriate clinical setting. The ratio of BNP/CKMB is more accurate in distinguishing TC from AMI than the value of BNP alone. An adequately powered prospective study is warranted to confirm the validity of this approach.
8 Journal of Cardiac Failure Vol. 20 No. 1 January 2014 Disclosures None.
Supplementary Data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.cardfail.2013.12.004.
References 1. Wittstein IS, Thiemann DR, Lima JA, Baughman KL, Schulman SP, Gerstenblith G, et al. Neurohumoral features of myocardial stunning due to sudden emotional stress. N Engl J Med 2005;352:539e48. 2. Neil CJ, Nguyen TH, Sverdlov AL, Chirkov YY, Chong CR, Stansborough J, et al. Can we make sense of takotsubo cardiomyopathy? An update on pathogenesis, diagnosis and natural history. Expert Rev Cardiovasc Ther 2012;10:215e21. 3. Sharma AK, Singh JP, Heist EK. Stress cardiomyopathy: diagnosis, pathophysiology, management, and prognosis. Crit Pathw Cardiol 2011;10:142e7. 4. Virani SS, Khan AN, Mendoza CE, Ferreira AC, de Marchena E. Takotsubo cardiomyopathy, or broken-heart syndrome. Tex Heart Inst J 2007;34:76e9. 5. Hurst RT, Prasad A, Askew JW 3rd, Sengupta PP, Tajik AJ. Takotsubo cardiomyopathy: a unique cardiomyopathy with variable ventricular morphology. JACC Cardiovasc Imaging 2010;3:641e9. 6. Wedekind H, Moller K, Scholz KH. [Takotsubo cardiomyopathy. Incidence in patients with acute coronary syndrome]. Herz 2006;31: 339e46. 7. Azzarelli S, Galassi AR, Amico F, Giacoppo M, Argentino V, Tomasello SD, et al. Clinical features of transient left ventricular apical ballooning. Am J Cardiol 2006;98:1273e6. 8. Mukoyama M, Nakao K, Saito Y, Ogawa Y, Hosoda K, Suga S, et al. Increased human brain natriuretic peptide in congestive heart failure. N Engl J Med 1990;323:757e8. 9. Hall C. NT-proBNP: the mechanism behind the marker. J Card Fail 2005;11:S81e3. 10. Madhavan M, Borlaug BA, Lerman A, Rihal CS, Prasad A. Stress hormone and circulating biomarker profile of apical ballooning syndrome (takotsubo cardiomyopathy): insights into the clinical significance of B-type natriuretic peptide and troponin levels. Heart 2009;95: 1436e41. 11. Akashi YJ, Nakazawa K, Sakakibara M, Miyake F, Koike H, Sasaka K. The clinical features of takotsubo cardiomyopathy. QJM 2003;96:563e73. 12. Bonaca MP, Morrow DA. Defining a role for novel biomarkers in acute coronary syndromes. Clin Chem 2008;54:1424e31. 13. Akashi YJ, Nef HM, Mollmann H, Ueyama T. Stress cardiomyopathy. Annu Rev Med 2010;61:271e86.
14. Bielecka-Dabrowa A, Mikhailidis DP, Hannam S, Rysz J, Michalska M, Akashi YJ, et al. Takotsubo cardiomyopathydthe current state of knowledge. Int J Cardiol 2010;142:120e5. 15. Frohlich GM, Schoch B, Schmid F, Keller P, Sudano I, Luscher TF, et al. Takotsubo cardiomyopathy has a unique cardiac biomarker profile: NT-proBNP/myoglobin and NT-proBNP/troponin T ratios for the differential diagnosis of acute coronary syndromes and stress induced cardiomyopathy. Int J Cardiol 2012;154:328e32. 16. Matsuoka K, Okubo S, Fujii E, Uchida F, Kasai A, Aoki T, et al. Evaluation of the arrhythmogenecity of stress-induced ‘‘takotsubo cardiomyopathy’’ from the time course of the 12-lead surface electrocardiogram. Am J Cardiol 2003;92:230e3. 17. Behr ER, Mahida S. Takotsubo cardiomyopathy and the long-QT syndrome: an insult to repolarization reserve. Europace 2009;11: 697e700. 18. Prasad A, Lerman A, Rihal CS. Apical ballooning syndrome (tako-tsubo or stress cardiomyopathy): a mimic of acute myocardial infarction. Am Heart J 2008;155:408e17. 19. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D, Modification of Diet in Renal Disease Study Group. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med 1999;130:461e70. 20. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P. Acute Dialysis Quality Initiative w. Acute renal failureddefinition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8:R204e12. 21. Alpert JS, Thygesen K, Antman E, Bassand JP. Myocardial infarction redefinedda consensus document of The Joint European Society of Cardiology/American College of Cardiology Committee for the Redefinition of Myocardial Infarction. J Am Coll Cardiol 2000;36: 959e69. 22. Reardon M, Malik M. QT interval change with age in an overtly healthy older population. Clin Cardiol 1996;19:949e52. 23. Lepeschkin E, Surawicz B. The measurement of the Q-T interval of the electrocardiogram. Circulation 1952;6:378e88. 24. Ahmed KA, Madhavan M, Prasad A. Brain natriuretic peptide in apical ballooning syndrome (takotsubo/stress cardiomyopathy): comparison with acute myocardial infarction. Coron Artery Dis 2012;23: 259e64. 25. Nguyen TH, Neil CJ, Sverdlov AL, Mahadavan G, Chirkov YY, Kucia AM, et al. N-Terminal proebrain natriuretic protein levels in takotsubo cardiomyopathy. Am J Cardiol 2011;108:1316e21. 26. Srisawasdi P, Vanavanan S, Charoenpanichkit C, Kroll MH. The effect of renal dysfunction on BNP, NT-proBNP, and their ratio. Am J Clin Pathol 2010;133:14e23. 27. Hall C. Essential biochemistry and physiology of (NT-pro)BNP. Eur J Heart Fail 2004;6:257e60. 28. Akashi YJ, Goldstein DS, Barbaro G, Ueyama T. Takotsubo cardiomyopathy: a new form of acute, reversible heart failure. Circulation 2008; 118:2754e62. 29. Redfield MM, Rodeheffer RJ, Jacobsen SJ, Mahoney DW, Bailey KR, Burnett JC Jr. Plasma brain natriuretic peptide concentration: impact of age and gender. J Am Coll Cardiol 2002;40:976e82.
Consensus
Diagnostic Utility of Cardiac Biomarkers in Discriminating Takotsubo Cardiomyopathy From Acute Myocardial Infarction MANDEEP SINGH RANDHAWA, MD,1 ASHWAT SINGH DHILLON, MD,1 HARRIS C. TAYLOR, MD,2 ZHIYUAN SUN, MS,3 AND MILIND Y. DESAI, MD4 Cleveland, Ohio
ABSTRACT Background: Takotsubo cardiomyopathy (TC) mimics acute myocardial infarction (AMI). We postulated that ventricular dysfunction in TC in the absence of significant myocardial necrosis would produce higher B-type natriuretic peptide (BNP)/troponin T (TnT) and BNP/creatine kinase MB fraction (CKMB) ratios than in AMI. Methods and Results: We studied 58 consecutive TC (age 65.8 6 82.9) and 97 AMI patients (age 59.8 6 83.4). The ratios of BNP/TnT and BNP/CKMB were calculated with the use of first simultaneously drawn laboratory values. Receiver operating characteristic curves were used to distinguish TC from AMI with 95% specificity based on cardiac biomarker ratios. Median BNP/TnT and BNP/CKMB ratios were, respectively, 1,292 [interquartile range 443.4e2,657.9] and 28.44 [13.7e94.8] in the TC group and 226.9 [69.91e426.32] and 3.63 [1.07e10.02] in the AMI group (P ! .001). TC can be distinguished from AMI with 95% specificity with the use of BNP/TnT ratio $1,272 (sensitivity 52%) and BNP/CKMB ratio $29.9 (sensitivity 50%). Conclusions: The value of BNP is significantly higher in TC than in AMI. Early BNP/TnT and BNP/ CKMB ratios help to differentiate TC from AMI with greater accuracy than BNP alone. (J Cardiac Fail 2014;20:2e8) Key Words: Takotsubo cardiomyopathy, cardiac biomarkers, electrocardiogram and acute myocardial infarction.
Takotsubo cardiomyopathy (TC) is a disorder of transient ventricular systolic dysfunction thought to be precipitated by severe emotional or physical stress.1e4 Echocardiography and/or ventriculogram typically shows left ventricular apical ballooning.1,5 Patients usually present with chest
pain with electrocardiographic (ECG) changes mimicking an acute myocardial infarction (AMI). Approximately 1% e2% of patients presenting with acute coronary syndrome (ACS) are ultimately diagnosed with TC.6,7 An elevated plasma concentration of B-type natriuretic peptide (BNP) is an established marker of left ventricular dysfunction.8 Both troponin T (TnT) and creatinine kinase MB fraction (CKMB) are elevated in the setting of membrane leak caused by acute myocardial necrosis, whereas the production and release of BNP is related to ventricular distention with or without myocyte necrosis.8e10 It is known that BNP and TnT are increased in both TC and AMI.10,11 However, because TC is a disease primarily causing distention of the ventricles and characterized by reversible myocardial dysfunction without necrosis, a greater increase in plasma BNP compared with TnT or CKMB has been demonstrated, compared with AMI.10,12e15 TC has also been associated with significant QTc interval prolongation on ECG.16,17
From the 1Department of Internal Medicine, Cleveland Clinic, Cleveland, Ohio; 2Director Resident Research, Fairview Hospital - Cleveland Clinic, Clinical Professor of Medicine, Case Western Reserve University School of Medicine, Cleveland, Ohio; 3Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland, Ohio and 4Heart and Vascular Institute, Cleveland Clinic, Cleveland, Ohio. Manuscript received August 10, 2013; revised manuscript received December 8, 2013; revised manuscript accepted December 9, 2013. Reprint requests: Milind Y. Desai, MD, Tomsich Department of Cardiovascular Medicine, Heart and Vascular Institute, Cleveland Clinic, 9500 Euclid Ave, Desk J1-5, Cleveland, OH 44195. Tel: 216-445-5250; Fax: 216-445-6155. E-mail: [email protected] See page 8 for disclosure information. 1071-9164/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cardfail.2013.12.004
2
Biomarkers and Takotsubo Cardiomyopathy
Currently, TC is a diagnosis of exclusion according to the modified Mayo Clinic Criteria, which involve performing coronary angiography to confirm the absence of obstructive coronary artery disease (CAD).18 In the present study, we sought to determine if we could improve our noninvasive ability to distinguish between TC and AMI with the use of cardiac biomarker ratios obtained early during presentation. Methods This was a retrospective cohort analysis (matching for cases and respective control subjects) done at a single tertiary care center (Cleveland Clinic Foundation, Cleveland, Ohio), after appropriate Institutional Review Board approval. We studied 225 consecutive TC patients from the period of January 2006 to February 2012 based on the following criteria of diagnosis: acute presentation with ACS-like symptoms, with acute ST-segment and T-wave changes on the ECG and rise in the cardiac enzymes (CKMB and TnT), typical echocardiographic and angiographic findings (ballooning of the left ventricle), complete reversibility of ejection fraction (EF) based on repeated transthoracic echocardiography, and absence of significant (O50%) CAD and acute plaque rupture based on angiography. To be included in the study, a patient had to have had biomarkers determined at the time of admission. We excluded the following patients: !18 years old, known chronic kidney disease (CKD) stage IIIeV with estimated glomerular filtration rate !60 mL/min according to the Modification of Diet in Renal Disease equation,19 acute kidney injury (AKI) at time of admission according to the criteria proposed by the Acute Dialysis Quality Initiative group,20 history of congestive heart failure (CHF) with a history of left ventricular ejection fraction (LVEF) !45% on the last available echocardiogram, known nonischemic cardiomyopathy, microbial and toxic myocarditis, septic shock, acute pulmonary embolism, and severe aortic stenosis on the last available echocardiogram. A total of 225 consecutive patients with TC were studied. 118 patients were excluded because of unavailability of simultaneously drawn first available BNP, CKMB, and TnT. Of the others, 22 patients did not meet our diagnostic criteria for TC. Sixteen patients had preexisting CKD. Eight patients were excluded because of AKI. Two patients were excluded because of preexisting CHF. One patient was excluded because of concurrent chemotherapy that was potentially cardiotoxic. After applying the exclusion criteria, 58 patients (case subjects) were selected, and their baseline clinical characteristics were recorded by manually reviewing their electronic medical records. Subsequently, a control group of patients that met the criteria of AMI based on the Joint European Society of Cardiology/American College of Cardiology Committee for the Redefinition of Myocardial Infarction were assembled.21 Ninety-seven such AMI patients were selected to keep the study power O90% to show a significant difference in the 2 groups of TC and AMI. The QTc interval duration was recorded and initial ECG was analyzed for acute ST-segment and T-wave changes, and both, case and control subjects were divided into ST-segment elevation and noneST-segment elevation groups. QTc was deemed to be prolonged if it was O470 ms.22,23 These variables were followed for 72 hours after admission. All case and control subjects underwent coronary angiography during hospitalization in a standard
Randhawa et al
3
fashion. The degree of coronary obstruction was deemed to be significant in the setting of O50% stenosis in a major epicardial vessel. Although it is true that atypical variants of TC are less clinically confusing with coronary artery disease, our focus was to correlate the first available cardiac biomarkers regardless of echocardiographic findings. We did not study echocardiographic data for global versus segmental dysfunction. The study was done to retrospectively compare initial cardiac biomarker profiles in TC and AMI. Laboratory Measurements We recorded first simultaneously available levels of TnT, CKMB, and BNP after presentation. All measurements for TnT and CKMB were done by electrochemiluminescence immunoassay technique with the use of the Roche Cobas 6000e analyzer (Roche Diagnostics). All determinations of BNP were done by chemiluminescence immunoassay technique with the use of the Advia Centaur analyzer (Siemens Medical Solutions Diagnostics). As such, we standardized the laboratory analysis for all study subjects. Statistical Methods Continuous measures were described with the use of mean and standard deviation or median and interquartile range (IQR) and were compared between groups with the use of standard t testing or Wilcoxon rank sum test. Categoric measures were summarized as frequency and percentages and were compared between groups with the use of chi-square test. Receiver operating characteristic (ROC) analyses of absolute values of BNP, TnT, and CKMB, and 2 ratios (BNP/TnT and BNP/CKMB) were performed to assess their ability in differentiating TC case subjects from AMI control subjects. We wanted to assess whether the ratio of BNP/ TnT or BNP/CKMB would be better in distinguishing TC from AMI compared with individual cardiac biomarker values alone. Therefore, paired comparison was performed on the area under the ROC curve (AUC) of BNP, TnT, CKMB, and the ratios BNP/TnT and BNP/CKMB. In addition, analysis of ECG findings revealed that presence of QTc interval prolongation was more common in TC case subjects than in AMI control subjects. Logistic regression model was used to compare the distinguishing ability of combination of BNP/CKMB and presence of QTc interval prolongation compared with the BNP/CKMB ratio alone. All analyses were performed with the use of SAS 9.2 and R 2.12.1 software.
Results A total of 155 patients were studied from January 2006 to February 2012. Of these, 58 were diagnosed with TC (case subjects) and 97 had angiographically proven AMI (control subjects). Of the case subjects, 22/58 (38%) were found to have ST-segment elevation (STE-TC) and 36/58 (62%) had no ST-segment elevation (NSTE-TC) in their initial ECGs. In the control group, 38/97 (39%) had ST-segment-elevation MI (STEMI) and 59/97 (61%) had noneST-segment-elevation MI (NSTEMI). Within the NSTEMI subgroup of the control subjects, most patients had significant multivessel disease. There were 7 patients with significant left main disease, 26 with significant
4 Journal of Cardiac Failure Vol. 20 No. 1 January 2014 left anterior descending artery disease, 33 with significant disease in the left circumflex, 27 with significant disease in the right coronary artery, and 10 with a history of coronary bypass surgery. In the STEMI subgroup of the control subjects, 1 patient had significant left main disease, 19 had significant left anterior descending disease, 13 had significant circumflex disease, 24 had significant right coronary disease, 1 had cocaine-induced AMI, and 1 had bypass graft occlusion. No significant CAD was found on coronary angiography of TC case subjects. The vast majority of case subjects (55/58, 95%) presented with symptoms, and 3/58 (5%) case subjects developed symptoms after hospitalization. Among the control subjects, 95/97 (98%) presented with symptoms, and 2/97 (2%) developed symptoms after hospitalization. Baseline characteristics and cardiac biomarker values of the cases and controls are summarized in Tables 1 and 2. In the TC group, BNP levels ranged from 16 to 2,511 pg/mL (median 456.5 pg/mL), whereas the median BNP level in the AMI group was 97 pg/mL (P ! .0001). The median [IQR] BNP/TnT and BNP/CKMB ratios were, respectively, 1,292 [443.4e2657.9] and 28.44 [13.7e94.8] in the TC group and 226.9 [69.91e426.32] and 3.63 [1.07e10.02] in the AMI group (P ! .001). Figure 1 depicts a scatter plot of BNP and CKMB values in the study population. We found that 20/58 TC case subjects had BNP $600 pg/mL, and in 52/58 case subjects the value of CKMB was #25 ng/mL. This was in contrast to the AMI population in our study, all of whom had a first BNP level #600 pg/mL. A presumed precipitating factor was identified in 37/58 (64%) of the TC cases (Supplemental Table 1). Discrimination of Takotsubo Cardiomyopathy From Acute Myocardial Infarction
The data on incremental ability of cardiac biomarkers and their ratios to distinguish TC from AMI are presented in Table 3. TC case subjects could be distinguished from AMI control subjects with the use of the earliest BNP/ TnT ratio $1,272 with a specificity of 95% and sensitivity of 52% (95% confidence interval [CI] 0.751e0.894, AUC 0.822; P ! .001). As such, of 57 TC cases for whom CKMB data was available, the earliest BNP/CKMB ratio $29.9 distinguished TC from AMI with a specificity of
95% and sensitivity of 50% (95% CI 0.802e0.922, AUC 0.862; P ! .001). These results also show that the BNP/ CKMB ratio has greater AUC than the BNP/TnT ratio (Fig. 2). On exploring further, we found that 53% of TC patients had a BNP/CKMB ratio $25 and conversely, 89/97 (92%) of AMI patients had a BNP/CKMB ratio !25 (Fig. 3). A slightly higher BNP/CKMB ratio of $38 distinguishes TC from AMI with 99% specificity and 46% sensitivity. We also found that the ratios of BNP/CKMB as well as BNP/TnT were both significantly (P ! .001) more accurate in distinguishing TC from AMI than the value of BNP alone (Table 4). Subset Analysis of STE-TC and STEMI (Table 3)
In the ST-segment elevation group, a BNP/CKMB ratio $13.1 distinguished TC from AMI with 95% specificity and 68% sensitivity (AUC 0.888, 95% CI 0.800e0.975; P ! .001) and a BNP/TnT ratio $622 distinguished TC from AMI with 95% specificity and 52% sensitivity (AUC 0.863, 95% CI 0.764e0.963; P ! .001). Subset Analysis of NSTE-TC and NSTEMI
In the noneST-segment elevation group, a BNP/CKMB ratio $34.7 distinguished TC from AMI with 95% specificity and 56% sensitivity (AUC 0.868, 95% CI 0.795e0.942; P ! .001) and a BNP/TnT ratio $2,094 distinguished TC from AMI with 95% specificity and 54% sensitivity (AUC 0.830, 95% CI 0.735e0.925; P ! .001). On comparing the AUC of BNP/CKMB and BNP alone in both subsets (ST-segment elevation and noneSTsegment elevation) the ratio of BNP/CKMB distinguished TC from AMI with significantly higher accuracy. The same was not true for the ratio of BNP/TnT (Table 4). Discussion To the best of our knowledge, this analysis is one of the largest cohorts of TC and AMI patients studied to compare the ability of cardiac biomarker ratios to provide diagnostic yield to distinguish between these 2 disease processes with very different management strategies and outcomes.10,15,24 We exclusively assessed first simultaneously available values of cardiac biomarkers to distinguish TC from
Table 1. Baseline Clinical Characteristics Variable Age (y) BMI (kg/m2) Female, n (%) HTN, n (%) DM, n (%) HLD, n (%) T-wave inversion, n (%) QTc prolongation, n (%)
Acute Myocardial Infarction Control Subjects (n 5 97)
Takotsubo Case Subjects (n 5 58)
P Value
59.8 6 13 29.45 6 5.75 77 (79) 71 (73) 23 (24) 56 (58) 52 (57) 16 (16)
65.8 6 12.9 26.4 6 4.9 57 (98) 39 (68) 10 (18) 33 (59) 42 (79) 35 (60)
.0044 .0062 .0009 .53 .37 .89 .006 !.0001
BMI, body mass index; HLD, hyperlipidemia; HTN, hypertension; DM, diabetes mellitus.
Biomarkers and Takotsubo Cardiomyopathy
Randhawa et al
5
Table 2. Cardiac Biomarkers and Their Ratios, Median [Interquartile Range] Variable CKMB (ng/mL) TnT (ng/mL) BNP (pg/mL) BNP/TnT BNP/CKMB
Acute Myocardial Infarction Control Subjects (n 5 97)
Takotsubo Case Subjects (n 5 58)
P Value
25 [13.75e62.7] 0.52 [0.19e1.45] 97 [45.5e248.5] 226.9 [69.91e426.32] 3.63 [1.07e10.02]
10.5 [5.8e14.93] 0.38 [0.16e0.65] 456.5 [120.25e734.5] 1,292.1 [443.4e2,657.9] 28.44 [13.7e94.8]
!.0001 .0092 !.0001 !.001 !.001
CKMB, creatinine kinase MB fraction; TnT, troponin T; BNP, B-type natriuretic peptide.
AMI. Also, the first simultaneous cardiac biomarkers were drawn within 6 hours of admission in 83% of case subjects and 87% of control subjects. Moreover, we used strict inclusion and exclusion criteria for cases to reduce false positives.10,15,25,26 We further maximized consistency, choosing only those patients for whom the same laboratory methods were used for measurement of cardiac biomarkers. Our study population is a select population owing to the inherent difficulty in finding TC patients whose BNP levels are related only to their new-onset cardiac dysfunction. Levels of BNP can be affected by preexisting or coexisting clinical conditions other than acute heart failure, such as acute and chronic kidney disease, chronic heart failure, septic shock, and acute pulmonary embolism. Because accurate measurement of BNP as affected by acute heart failure secondary to TC was the key to our study, we had to exclude patients who had preexisting elevation of BNP. Therefore, the results may be of use in those patients who were known to have normal cardiac function. It must be noted that although TnT and CKMB are usually drawn simultaneously as part of a ‘‘cardiac enzyme panel,’’ BNP is often not ordered along with them. Our study was
designed to specifically assess the first available, simultaneously drawn BNP/TnT and BNP/CKMB ratios. We demonstrated that by using the simultaneous admission cardiac biomarker values, a BNP/TnT ratio O1,272 and a BNP/CKMB ratio O29.9 were both 95% specific for differentiating TC from AMI, albeit with low sensitivity (52% and 50%, respectively). Furthermore, a BNP/CKMB ratio $38 can distinguish TC from AMI with a specificity of 99% although with a slightly lower sensitivity of 46% (Fig. 3). Additionally, using the paired comparison, we demonstrate that the ratio of BNP/CKMB was significantly (P ! .001) more accurate in differentiating TC from AMI compared with the value of BNP alone (Table 4). While analyzing the ECG of TC patients, we noted that 63% manifested QTc interval prolongation within the first 72 hours after admission. TC is known to be associated with significant prolongation of the QTc interval on ECG.16,17 One earlier study found that the QTc interval was significantly prolonged in the subacute phase of TC compared with the acute and chronic phases.16 We found that QT interval prolongation was present in a higher number of TC patients compared with the AMI group. Logistic
Fig. 1. Distribution of B-type natriuretic peptide (BNP) and creatine kinase MB fraction (CKMB) in study population.
6 Journal of Cardiac Failure Vol. 20 No. 1 January 2014 Table 3. Distinguishing TC From AMI Using the Earliest Simultaneously Drawn Cardiac Biomarker Ratios and AUC Values for ROC Analysis Cutoff Specificity 95% Differentiate NSTE-TC from NSTEMI BNP BNP/TnT BNP/CKMB Differentiate STE-TC from STEMI BNP BNP/TnT BNP/CKMB Differentiate all TC from AMI BNP BNP/TnT BNP/CKMB QT prolongation þ BNP/CKMB
Sensitivity*
AUC
95% CI
P Value
442 2094 34.7
56% 54% 56%
0.770 0.830 0.868
0.667e0.874 0.735e0.925 0.795e0.942
!.001 !.001 !.001
522 622 13.1
41% 52% 68%
0.794 0.863 0.888
0.672e0.914 0.764e0.963 0.800e0.975
!.001 !.001 !.001
479 1272 29.9 0.473
50% 52% 50% 73%
0.776 0.822 0.862 0.890
0.698e0.854 0.751e0.894 0.802e0.922 0.832e0.949
!.001 !.001 !.001 !.001
AMI, acute myocardial infarction; AUC, area under the receiver operating characteristic curve; BNP, B-type natriuretic peptide (pg/mL); CKMB, creatinine kinase MB fraction (ng/mL); ROC, receiver operator curve; NSTE, noneST-segment elevation; STE, ST-segment elevation; TC, takotsubo cardiomyopathy; TnT, troponin T (ng/mL). *All sensitivities based on a specificity of 95%.
regression also revealed that the AUC for diagnosing TC was higher if both BNP/CKMB ratio and QTc interval prolongation were combined. However, subsequent paired comparison of BNP/CKMB versus the combination of BNP/CMKB and QTc prolongation did not show a significant difference (Table 4). . In practice, TC is mainly a diagnosis of exclusion, and coronary angiography is necessary to rule out obstructive coronary artery disease or acute plaque rupture. Currently available data suggest that 1%e2% of patients with symptoms of ACS have been found to have TC.18 TC has only recently begun to be recognized as a distinct disease entity. As a result, there are comparatively little data available on cardiac biomarkers in TC. Our study provides data that may
help to further characterize TC based on initial cardiac biomarker levels. . The significance of BNP in heart failure is well known.27 The N-terminal precursor of BNP (NT-proBNP) in the normal heart is mainly produced in the atrium. However, in the failing heart, there is increased ventricular production of NT-proBNP.9 TC is primarily a disorder of ventricular systolic dysfunction and not surprisingly has been found to be associated with higher circulating levels of BNP and NT-proBNP.2,25 It is also established that a rise in BNP concentration is caused by circumstances that increase ventricular wall stress, whereas troponin and CKMB are elevated in the setting of myocardial necrosis.12 Recent studies have shown that TC presents with an elevation of natriuretic peptides out of proportion to troponin levels.15,18 Frohlich et al found that the use of the NT-proBNP/peak TnT ratio appeared to be most accurate to distinguish ACS from TC.15 Following these findings, we decided to assess only the first available simultaneously drawn laboratory values of BNP, TnT, and CKMB in patients presenting with symptoms suggestive of ACS and to determine whether TC could be reliably distinguished from AMI early in the disease course. Finally, despite the difficulty in finding simultaneously available biomarkers and in using strict criteria to limit the number of patients with falsely elevated BNP values, we were able to include 58 TC patients in our study. Though not obviating the need for cardiac catheterization, we think that these results are sufficiently encouraging to justify an adequately powered prospective study to determine more definitively whether biomarkers and a positive discriminant, which combines biomarkers and QTc interval, can reliably distinguish these 2 disorders noninvasively at the time of presentation.
Fig. 2. Receiver operator characteristic curve analysis for the study population to help distinguish between acute myocardial infarction and Takotsubo cardiomyopathy. See Tables 3 and 4 for further details.
Study Limitations
A number of limitations in the present study need to be acknowledged. In addition to our study being retrospective
Biomarkers and Takotsubo Cardiomyopathy
Randhawa et al
7
Fig. 3. Distribution of BNP/CKMB ratio in the study population. Abbreviations as in Figure 1.
with the attendant limitations, we had to exclude as many as 118 of 225 TC case subjects because of lack of simultaneously drawn biomarkers. Second, despite our attempts to match case and control subjects according to sex, 98% of TC case subjects were female whereas 79% of AMI control subjects were female. Nonetheless, it should be recognized that TC occurs predominantly in postmenopausal women.1,28 Similarly, it was not possible to eliminate a significant difference in age between the case and control subjects. Third, the study is not applicable to patients who have
Table 4. Paired Comparison of AUC in Study Population Estimated Difference [95% CI] NoneST-segment elevation BNP vs BNP/TnT BNP vs BNP/CKMB BNP/TnT vs BNP/CKMB ST-segment elevation BNP vs BNP/TnT BNP vs BNP/CKMB BNP/TnT vs BNP/CKMB Overall BNP vs BNP/TnT BNP vs BNP/CKMB BNP/TnT vs BNP/CKMB BNP/CKMB vs QT prolongation þ BNP/CKMB Abbreviations as in Table 3.
P Value
0.060 [0.137e0.017] 0.098 [0.160 to 0.036] 0.038 [0.101e0.025]
.13 .002 .23
0.069 [0.173e0.035] 0.093 [0.187e0] 0.024 [0.074e0.026]
.19 .050 .34
0.046 0.086 0.040 0.028
[0.107e0.014] .14 [0.136 to 0.035] !.001 [0.081e0.002] .059 [0.065e0.008] .13
underlying renal dysfunction or heart failure (exclusion criteria). Although it is true that levels of BNP are age and sex dependent, we do not consider this to be a limitation because it applies to both TC and AMI.29 Fourth, although QTc interval prolongation was present on the first available ECG in 43% of cases, we followed serial ECGs and recorded QTc interval prolongation if present at any time in the first 72 hours after admission. Therefore, the QTc interval used to calculate the AUC in our study may not have been concurrent with the original BNP, CKMB, or TnT. Finally, although we did not include patients with known baseline QTc interval prolongation, subsequent drug-induced QTc interval prolongation in both case and control subjects can not be completely ruled out. However, the main emphasis of this study was on the cardiac biomarker profiles in TC and AMI, rather than the QTc interval. Conclusion The ratios of BNP/TnT and BNP/CKMB, obtained early during admission, can be used as independent indices to help differentiate TC from actual AMI with a high specificity in the appropriate clinical setting. The ratio of BNP/CKMB is more accurate in distinguishing TC from AMI than the value of BNP alone. An adequately powered prospective study is warranted to confirm the validity of this approach.
8 Journal of Cardiac Failure Vol. 20 No. 1 January 2014 Disclosures None.
Supplementary Data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.cardfail.2013.12.004.
References 1. Wittstein IS, Thiemann DR, Lima JA, Baughman KL, Schulman SP, Gerstenblith G, et al. Neurohumoral features of myocardial stunning due to sudden emotional stress. N Engl J Med 2005;352:539e48. 2. Neil CJ, Nguyen TH, Sverdlov AL, Chirkov YY, Chong CR, Stansborough J, et al. Can we make sense of takotsubo cardiomyopathy? An update on pathogenesis, diagnosis and natural history. Expert Rev Cardiovasc Ther 2012;10:215e21. 3. Sharma AK, Singh JP, Heist EK. Stress cardiomyopathy: diagnosis, pathophysiology, management, and prognosis. Crit Pathw Cardiol 2011;10:142e7. 4. Virani SS, Khan AN, Mendoza CE, Ferreira AC, de Marchena E. Takotsubo cardiomyopathy, or broken-heart syndrome. Tex Heart Inst J 2007;34:76e9. 5. Hurst RT, Prasad A, Askew JW 3rd, Sengupta PP, Tajik AJ. Takotsubo cardiomyopathy: a unique cardiomyopathy with variable ventricular morphology. JACC Cardiovasc Imaging 2010;3:641e9. 6. Wedekind H, Moller K, Scholz KH. [Takotsubo cardiomyopathy. Incidence in patients with acute coronary syndrome]. Herz 2006;31: 339e46. 7. Azzarelli S, Galassi AR, Amico F, Giacoppo M, Argentino V, Tomasello SD, et al. Clinical features of transient left ventricular apical ballooning. Am J Cardiol 2006;98:1273e6. 8. Mukoyama M, Nakao K, Saito Y, Ogawa Y, Hosoda K, Suga S, et al. Increased human brain natriuretic peptide in congestive heart failure. N Engl J Med 1990;323:757e8. 9. Hall C. NT-proBNP: the mechanism behind the marker. J Card Fail 2005;11:S81e3. 10. Madhavan M, Borlaug BA, Lerman A, Rihal CS, Prasad A. Stress hormone and circulating biomarker profile of apical ballooning syndrome (takotsubo cardiomyopathy): insights into the clinical significance of B-type natriuretic peptide and troponin levels. Heart 2009;95: 1436e41. 11. Akashi YJ, Nakazawa K, Sakakibara M, Miyake F, Koike H, Sasaka K. The clinical features of takotsubo cardiomyopathy. QJM 2003;96:563e73. 12. Bonaca MP, Morrow DA. Defining a role for novel biomarkers in acute coronary syndromes. Clin Chem 2008;54:1424e31. 13. Akashi YJ, Nef HM, Mollmann H, Ueyama T. Stress cardiomyopathy. Annu Rev Med 2010;61:271e86.
14. Bielecka-Dabrowa A, Mikhailidis DP, Hannam S, Rysz J, Michalska M, Akashi YJ, et al. Takotsubo cardiomyopathydthe current state of knowledge. Int J Cardiol 2010;142:120e5. 15. Frohlich GM, Schoch B, Schmid F, Keller P, Sudano I, Luscher TF, et al. Takotsubo cardiomyopathy has a unique cardiac biomarker profile: NT-proBNP/myoglobin and NT-proBNP/troponin T ratios for the differential diagnosis of acute coronary syndromes and stress induced cardiomyopathy. Int J Cardiol 2012;154:328e32. 16. Matsuoka K, Okubo S, Fujii E, Uchida F, Kasai A, Aoki T, et al. Evaluation of the arrhythmogenecity of stress-induced ‘‘takotsubo cardiomyopathy’’ from the time course of the 12-lead surface electrocardiogram. Am J Cardiol 2003;92:230e3. 17. Behr ER, Mahida S. Takotsubo cardiomyopathy and the long-QT syndrome: an insult to repolarization reserve. Europace 2009;11: 697e700. 18. Prasad A, Lerman A, Rihal CS. Apical ballooning syndrome (tako-tsubo or stress cardiomyopathy): a mimic of acute myocardial infarction. Am Heart J 2008;155:408e17. 19. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D, Modification of Diet in Renal Disease Study Group. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med 1999;130:461e70. 20. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P. Acute Dialysis Quality Initiative w. Acute renal failureddefinition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8:R204e12. 21. Alpert JS, Thygesen K, Antman E, Bassand JP. Myocardial infarction redefinedda consensus document of The Joint European Society of Cardiology/American College of Cardiology Committee for the Redefinition of Myocardial Infarction. J Am Coll Cardiol 2000;36: 959e69. 22. Reardon M, Malik M. QT interval change with age in an overtly healthy older population. Clin Cardiol 1996;19:949e52. 23. Lepeschkin E, Surawicz B. The measurement of the Q-T interval of the electrocardiogram. Circulation 1952;6:378e88. 24. Ahmed KA, Madhavan M, Prasad A. Brain natriuretic peptide in apical ballooning syndrome (takotsubo/stress cardiomyopathy): comparison with acute myocardial infarction. Coron Artery Dis 2012;23: 259e64. 25. Nguyen TH, Neil CJ, Sverdlov AL, Mahadavan G, Chirkov YY, Kucia AM, et al. N-Terminal proebrain natriuretic protein levels in takotsubo cardiomyopathy. Am J Cardiol 2011;108:1316e21. 26. Srisawasdi P, Vanavanan S, Charoenpanichkit C, Kroll MH. The effect of renal dysfunction on BNP, NT-proBNP, and their ratio. Am J Clin Pathol 2010;133:14e23. 27. Hall C. Essential biochemistry and physiology of (NT-pro)BNP. Eur J Heart Fail 2004;6:257e60. 28. Akashi YJ, Goldstein DS, Barbaro G, Ueyama T. Takotsubo cardiomyopathy: a new form of acute, reversible heart failure. Circulation 2008; 118:2754e62. 29. Redfield MM, Rodeheffer RJ, Jacobsen SJ, Mahoney DW, Bailey KR, Burnett JC Jr. Plasma brain natriuretic peptide concentration: impact of age and gender. J Am Coll Cardiol 2002;40:976e82.
