You are here

Article Text

Article menu
Heart failure and cardiomyopathies
Original article
Early β-blocker use and in-hospital mortality in patients with Takotsubo cardiomyopathy
  1. Toshiaki Isogai1,2,
  2. Hiroki Matsui1,
  3. Hiroyuki Tanaka2,
  4. Kiyohide Fushimi3,
  5. Hideo Yasunaga1
  1. 1Department of Clinical Epidemiology and Health Economics, School of Public Health, The University of Tokyo, Tokyo, Japan
  2. 2Department of Cardiology, Tokyo Metropolitan Tama Medical Center, Tokyo, Japan
  3. 3Department of Health Policy and Informatics, Graduate School of Medicine, Tokyo Medical and Dental University, Tokyo, Japan
  1. Correspondence to Dr Toshiaki Isogai, Department of Clinical Epidemiology and Health Economics, School of Public Health, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan; toisogai-circ{at}umin.ac.jp

Abstract

Objective A catecholamine-mediated mechanism has been implicated in the pathogenesis of Takotsubo cardiomyopathy (TC). However, the impact of β-blockers in acute-phase management of TC remains uncertain. This study aimed to examine whether early β-blocker use in TC was associated with lower in-hospital mortality.

Methods This was a retrospective cohort study using the Diagnosis Procedure Combination nationwide inpatient database in Japan. Patients with TC aged ≥20 years who were admitted to acute-care hospitals between 2010 and 2014 were identified. Thirty-day in-hospital mortality was compared between patients who started β-blocker therapy on hospitalisation day 1 or 2 (early β-blocker group) and those who did not receive a β-blocker during hospitalisation (control group) using propensity score-matching and instrumental variable analyses.

Results Of 2672 eligible patients (female, 81.5%; 423 early β-blocker therapy, 2249 controls) from 615 hospitals, 1:4 propensity score-matching created a cohort of 2110 patients (422 early β-blocker therapy, 1688 controls). There was no significant difference in 30-day in-hospital mortality between the early β-blocker group and control group (2.4% vs 2.0%, p=0.703; risk difference, 0.4%; 95% CI, −1.2% to 2.0%). Logistic regression analysis did not show a significant association between early β-blocker use and 30-day in-hospital mortality (OR, 1.17; 95% CI 0.58 to 2.37). Instrumental variable analysis also found that early β-blocker use was not associated with lower 30-day in-hospital mortality (risk difference, 1.2%; 95% CI −3.1% to 5.5%).

Conclusions This study found no significant association between early β-blocker use and in-hospital mortality in patients with TC.

https://doi.org/10.1136/heartjnl-2015-308712

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Takotsubo cardiomyopathy (TC) is a cardiac syndrome with acute reversible ventricular dysfunction, generally triggered by extreme stress in the absence of culprit coronary artery disease.1–5 Recent observational studies revealed that the in-hospital prognosis of TC was not as favourable as previously expected, with in-hospital mortality ranging from 2.5% to 8.4%.6–11 However, if patients with TC overcome the acute critical phase, ventricular dysfunction generally recovers within days or weeks.1–5 It is important to establish optimal therapy for the acute phase of TC. However, no evidence-based therapy has been established as yet because of the lack of rigorous studies. Thus, TC is generally managed with supportive therapy, despite its severity in the acute phase.1 ,2 ,4 ,5 ,12

    β-Blockers are now established as one of the key drugs for patients with heart failure with reduced ejection fraction and for acute myocardial infarction (AMI) in current major guidelines.13–15 β-Blockers are reported to be useful in preventing acute life-threatening complications in AMI.16 ,17 The acute-phase complications of TC are similar to those of AMI.7–10 ,18 Several pathophysiological mechanisms of TC have been proposed. A catecholamine-mediated mechanism remains controversial; some early studies supported a role of catecholamines, but subsequent studies did not confirm this mechanism.19–22 A recent experimental study showed that β-blocker therapy was not effective but rather harmful in rats with induced TC.23 Furthermore, most recently, a large-scale multinational study showed that the use of β-blockers at hospital discharge was not significantly associated with improved 1-year survival in patients with TC.6 No study has shown that β-blocker use is effective in reducing mortality in patients with TC. The use of β-blockers in the acute phase of TC remains a matter of debate, because ventricular dysfunction in TC is transient and β-blockers may actually worsen haemodynamics through negative inotropic and chronotropic effects.4

    The purpose of the present study was to examine whether β-blocker use in the acute phase of TC was associated with lower short-term mortality, using a large-scale national inpatient database.

    Methods

    Study design and data source

    This study was a retrospective cohort study using the Diagnosis Procedure Combination (DPC) nationwide inpatient database, which has been described in detail elsewhere.7 The database includes the following information: unique hospital identifier; admission and discharge dates; patient age and sex; diagnosis for admission, comorbidities already present at admission and postadmission complications, described by the International Classification of Diseases, 10th Revision (ICD-10) codes and Japanese text; consciousness level at admission measured with the Japan Coma Scale7; drugs and devices during hospitalisation; surgical and non-surgical procedures; and discharge status (dead or alive). The database includes dates of application of drugs, devices and procedures used during hospitalisation. In the DPC database, all data are de-identified and anonymised and cannot be linked with original medical charts. Patient follow-up can be performed in the period from the date of admission to the date of hospital discharge because the DPC database only includes in-hospital data.

    The present study was approved by the Institutional Review Board of The University of Tokyo. The requirement for informed consent was waived because of the anonymous nature of the data.

    Selection of participants

    TC was identified from both the Japanese text-based diagnosis of TC and the relevant ICD-10 code (see online supplementary table S1), as in our previous study.7 In this study, ‘early’ β-blocker use was defined as starting intravenous or oral β-blocker therapy on the day of admission (day 1) or the following day (day 2).

    We identified patients who were admitted to acute-care hospitals with TC and underwent coronary angiography during index hospitalisation between 1 July 2010 and 31 March 2014. Thus, the present study did not include patients who developed TC after admission. For the main analyses in the present study, we set inclusion and exclusion criteria to exclude haemodynamically unstable patients who were unlikely to receive an early β-blocker. Inclusion criteria were as follows: (1) ≥20 years of age; (2) underwent coronary angiography on day 1 or 2; and (3) did not undergo coronary revascularisation (percutaneous coronary intervention or coronary artery bypass grafting) during the index hospitalisation. Exclusion criteria were as follows: (1) pheochromocytoma or acute myocarditis; (2) status asthmaticus; (3) a history of myocardial infarction; (4) mitral stenosis, aortic stenosis or aortic regurgitation; (5) received intra-aortic balloon pumping (IABP), extracorporeal membrane oxygenation (ECMO) or temporary pacing on day 1 or 2; (6) received inotropes or vasopressors (dopamine, dobutamine, noradrenaline, adrenaline, phosphodiesterase-III inhibitor, isoprenaline, vasopressin, etilefrine or phenylephrine) on day 1 or 2; (7) underwent surgery under general anaesthesia on day 1 or 2; (8) discharged on day 1 or 2; and (9) started treatment with a β-blocker on day 3 or later. Eligible patients were divided into two groups: patients who started receiving a β-blocker on day 1 or 2 (early β-blocker group) and those who did not receive any β-blocker during hospitalisation (control group).

    Baseline variables

    For baseline variables, we extracted the following data on patient characteristics and clinical treatment initiated on day 1 or 2: patient age and sex; ambulance use; type of hospital (academic or non-academic); Japan Coma Scale at admission;7 comorbidities already present at admission (see online supplementary table S1), in reference to our previous study7; invasive arterial pressure monitoring; invasive pulmonary artery pressure monitoring; drugs given; and mechanical support (mechanical ventilation and renal replacement therapy). Drug therapy included nitrates, nicorandil, carperitide (α-human atrial natriuretic peptide), class III antiarrhythmics (amiodarone, nifekalant or sotalol), digitalis, loop diuretics, aldosterone antagonists, ACE inhibitors/angiotensin II receptor blockers, calcium antagonists, anticoagulants, aspirin/thienopyridines and statins. Levosimendan, a calcium-channel sensitiser, is not approved for use in Japan. The Japan Coma Scale at admission was categorised into four groups: 0 (alert), 1–3 (drowsy), 10–30 (somnolent) and 100–300 (comatose).

    Outcome measures

    The primary outcome was 30-day in-hospital mortality, which was defined as in-hospital deaths within 30 days after admission. The secondary outcomes were 15-day in-hospital mortality and overall in-hospital mortality. Cardiovascular complications (ventricular tachycardia or fibrillation, second-degree or third-degree atrioventricular block and cerebral infarction) that occurred after admission were also identified.

    Statistical analyses

    Categorical variables are presented as numbers and percentages. Continuous variables are presented as mean and SD or median and IQR. The balance of the baseline variables between the groups was examined using the standardised difference.24 An absolute standardised difference of >10 was regarded as a significant imbalance in a baseline variable.

    We performed 1:4 propensity score-matching between the early β-blocker group and the control group.24 A propensity score was estimated for early β-blocker use by using a logistic regression model including all 36 baseline variables as covariates. The C-statistic was calculated to evaluate the goodness of fit. Each patient in the early β-blocker group was matched with four patients in the control group, with the closest estimated propensity score within a calliper (≤0.20 of the pooled SD of estimated logits) by using the nearest neighbour method with replacement. The outcomes were then compared between the two groups using Fisher's exact test. Risk ratios and risk differences as well as their 95% CIs were calculated using the numbers of patients and deaths in each treatment group.

    To assess the association between early β-blocker use and mortality, logistic regression analysis fitted with generalised estimating equations was also performed to take into account the matched nature of the data in the propensity score-matched cohort.24 The dichotomous variable of early β-blocker use or no use was the independent variable in the logistic regression analyses in the propensity score-matched groups, and ORs and 95% CIs were determined. Because there were no missing data in the measured variables in this study, all patients were included in the multivariable analyses.

    For confirmatory analysis of our propensity score-matching analysis, we also performed instrumental variable analysis.25 ,26 We used the frequency of early β-blocker therapy per hospital as an instrumental variable. This analysis included only patients admitted to hospitals in which at least four eligible patients with TC were admitted during the study period. We calculated the frequency of early β-blocker therapy per hospital by dividing the total number of patients receiving early β-blocker therapy by the total number of patients with TC at each hospital. The F-statistic was calculated to confirm that the instrumental variable was not a weak instrument. An F-statistic >20 indicated a strong instrumental variable. Risk differences and 95% CIs were estimated using two-stage, least-squares regression that also adjusted for propensity scores.

    Analysis of 30-day in-hospital mortality was also performed in subgroups defined according to age (<75 years or ≥75 years), sex, ambulance use for admission, type of hospital and presence or absence of comorbidity at admission. The significance of interaction was assessed using the Mantel–Haenszel test.

    For the sensitivity analyses, we created another cohort (termed the ‘extended cohort’), which included haemodynamically unstable patients (defined as requiring an inotrope/vasopressor or IABP/ECMO on day 1 or 2), and performed the propensity score-matching and instrumental analyses again.

    All hypothesis tests had a two-sided significance level of 0.05. The propensity score-matching and the instrumental variable analyses were performed using the psmatch2 and ivreg2 procedures in Stata/SE V.13.0 (StataCorp, College Station, Texas, USA). Other statistical analyses were performed using IBM SPSS, V.22 (IBM, Armonk, New York, USA).

    Results

    Study population

    Figure 1 presents a flow chart of the patient selection procedure, which led to identification of 2672 eligible patients in 615 acute-care hospitals. In this overall cohort, the mean age was 73.0 (SD, 11.2) years and the proportion of females was 81.5%. The median length of stay was 11 days (IQR, 7–16 days). In the early β-blocker group (n=423), β-blockers were started on day 1 in 227 (53.7%) patients and on day 2 in 196 (46.3%) patients. Nineteen (4.5%) patients received initial intravenous and then oral β-blockers, and 27 (6.4%) received only intravenous β-blockers (propranolol, landiolol and esmolol are available in Japan), while the remaining 377 (89.1%) received oral β-blockers only. Of 396 patients receiving oral β-blockers, 294 received carvedilol, 78 received bisoprolol and 24 received other β-blockers. The median duration of β-blocker therapy in the early β-blocker group was 9 days (IQR, 6–15 days) and the median proportion of β-blocker therapy days during the whole hospitalisation days was 92.3%.

    Figure 1
    Figure 1

    Patient selection. AR, aortic regurgitation; AS, aortic stenosis; ECMO, extracorporeal membrane oxygenation; IABP, intra-aortic balloon pumping; MS, mitral stenosis; TC, Takotsubo cardiomyopathy.

    The propensity score-matching generated a cohort of 2110 patients, including 422 in the early β-blocker group and 1688 in the control group. The C-statistic was 0.74.

    Baseline characteristics

    In the unmatched groups, compared with the control group, the early β-blocker group was older, had a lower Japan Coma Scale score at admission, a lower proportion of malignancy and chronic pulmonary disease, and a higher proportion of invasive pulmonary artery pressure monitoring from day 1 or 2 (table 1). Patients in the early β-blocker group were significantly more likely to receive several drugs on day 1 or 2. After propensity score-matching, all baseline characteristics were well balanced between the two groups.

    View this table:
    Table 1

    Baseline characteristics in unmatched and propensity score-matched groups

    Comparison of mortalities between groups

    There were no significant differences in in-hospital mortality rates between the early β-blocker and control groups in the unmatched analyses or in the propensity score-matched analyses (table 2). There were also no significant differences in postadmission cardiovascular complications (see online supplementary table S2).

    View this table:
    Table 2

    Comparisons of mortalities between early β-blocker group and control group

    Association between early β-blocker use and mortality

    In the logistic regression analysis of the propensity score-matched patients, no significant associations were observed between early β-blocker use and in-hospital mortality rates (table 3). In the instrumental variable analysis (n=2075), the frequency of early β-blocker therapy per hospital was a sufficiently strong instrumental variable for predicting early β-blocker therapy (F-statistic=503). This analysis also showed that early β-blocker therapy was not significantly associated with lower in-hospital mortality rates (table 3).

    View this table:
    Table 3

    Association between early β-blocker use and mortalities

    Subgroup and sensitivity analyses

    Figure 2 shows the results of subgroup analyses of 30-day in-hospital mortality in the propensity score-matched patients. There was a significant interaction between early β-blocker therapy and sex on 30-day in-hospital mortality. In the sensitivity analyses using the extended cohort (n=3378), we obtained results consistent with those in the main analyses (see online supplementary figures 1, 2 and tables S3–S5).

    Figure 2
    Figure 2

    Subgroup analyses of 30-day in-hospital mortality in propensity score-matched patients. *p Values were estimated by Fisher's exact test. †Comorbidity at admission was defined as having at least one chronic or acute comorbidity at admission.

    Discussion

    This retrospective nationwide cohort study, using propensity score-matching and instrumental variable analyses, examined whether early β-blocker use for TC was associated with lower in-hospital mortality. Our results found no significant association between early β-blocker use for TC and in-hospital mortality.

    Although the pathogenesis of TC remains to be fully defined, one hypothesis is that it involves a catecholamine-mediated mechanism.19–21 Given the putative harmful effects of catecholamines in patients with TC, β-blockers may potentially decrease the intensity of the detrimental cardiac effects in TC by blocking these effects. Some review articles mentioned that empirical β-blocker use should be considered in the acute phase, if tolerated,2 ,3 while others mentioned that β-blockers could carry the risk of adverse effects in TC.4 ,5 This controversial issue of β-blockers in TC has been examined in two small studies. A multicentre study (n=36) showed that 30-day treatment with an oral β-blocker initiated at the time of diagnosis was not significantly associated with more rapid improvement of ejection fraction than treatment without a β-blocker.27 An American single-centre study (n=64) showed that pretreatment with β-blockers before admission was not significantly associated with the severity of myocardial dysfunction or damage in patients with TC.28 However, these observational studies were subject to significant bias because their small sample sizes preclude performance of multivariable adjustment. In the present study, a large number of patients with TC were included and propensity score-matching and instrumental variable analyses were performed to reduce potential bias.

    One potential reason for our results is that the deleterious effects of catecholamines on the myocardium may not have continued after development of TC with marked myocardial dysfunction. Therefore, administration of a β-blocker for TC may have been too late to prevent or recover the loss of cardiac function. Although the present study did not investigate the underlying pathophysiology of TC, our results regarding early β-blocker therapy may suggest that other pathophysiological pathways such as an inflammatory29 or energetic–metabolic pathway30 need to be explored to identify a suitable therapy for TC. Importantly, however, this was not a randomised controlled study but an observational study. Therefore, a randomised controlled study is necessary to further investigate the issues surrounding β-blocker use in TC.

    In our subgroup analysis, there was a significant interaction between early β-blocker therapy and sex on 30-day in-hospital mortality. This result should be interpreted cautiously because the number of male patients receiving early β-blocker therapy was small (n=75). Nevertheless, this result may suggest that the effectiveness of early β-blocker therapy was different between male and female patients with TC. Previous literature demonstrated that higher positive inotropic responses mediated by β-adrenoceptors were observed in males than in age-matched females.21 In addition, cardioprotective effects of oestrogen have been reported.4 ,5 ,21 Effects of β-blocker therapy may be affected by the sex differences in β-adrenoceptor-mediated responses and effects of oestrogen.

    The current guidelines for AMI recommend that β-blockers should be initiated in the first 24 h in patients with AMI with no contraindication to β-blockers (class I).14 ,15 Although TC and AMI often present with similar clinical manifestations, there are fundamental differences in their pathophysiology and epidemiology. Our results may support that the evidence for β-blocker use in AMI should not be directly applied to patients with TC.

    Limitations

    Several limitations should be acknowledged in the present study. First, the present study was a retrospective observational study. Although we performed propensity score-matching and instrumental variable analyses, there could be residual bias in this study. In addition, recorded diagnoses are commonly less well validated in retrospective studies than in prospective studies. Second, the database does not include data about vital signs and laboratory findings. Third, various subclasses, doses and durations of β-blocker therapy were used in the early β-blocker group in this study. This heterogeneity of β-blocker therapy in the early β-blocker group may have affected the outcomes. Fourth, the effectiveness of intravenous β-blockers was not clarified in this study because intravenous β-blockers were used in only a small number of patients. The effectiveness on outcomes in patients with TC may be different between intravenous and oral β-blockers, because intravenous β-blockers have a shorter half-life and can show adrenergic blocking effects more rapidly and safely than oral β-blockers. Fifth, the effectiveness of early β-blocker use on recovery of cardiac function remains unknown. Sixth, it was impossible to examine the effectiveness of β-blockers on postdischarge prognosis in this study. Finally, our results may not be generalisable to patients with TC in other countries.

    Conclusions

    In this retrospective observational study, propensity score-matching and instrumental variable analyses found neither a beneficial nor harmful association between early, short-term β-blocker use and in-hospital mortality in patients with TC. A randomised controlled study is needed to confirm the findings of this study.

    Key messages

    What is already known on this subject?

    • Takotsubo cardiomyopathy (TC) is generally managed with supportive therapy because no evidence-based therapy has yet been established. Although the pathogenesis of TC remains to be fully elucidated, one hypothesis is that a catecholamine-mediated mechanism is involved. The effectiveness of β-blocker use in the acute phase of TC on in-hospital mortality has not yet been examined.

    What might this study add?

    • In this retrospective nationwide study in Japan, propensity score-matching and instrumental variable analyses showed no significant association between early β-blocker use (within the first 2 days after admission) and in-hospital mortality in patients with TC.

    How might this impact on clinical practice?

    • There is no evidence to support early β-blocker use as routine practice in acute-phase management of TC.

    References

    1. ↵
      1. Ghadri JR,
      2. Ruschitzka F,
      3. Lüscher TF, et al
      . Takotsubo cardiomyopathy: still much more to learn. Heart 2014;100:1804–12. doi:10.1136/heartjnl-2013-304691
      OpenUrlFREE Full Text
    2. ↵
      1. Prasad A,
      2. Lerman A,
      3. Rihal CS
      . Apical ballooning syndrome (Tako-Tsubo or stress cardiomyopathy): a mimic of acute myocardial infarction. Am Heart J 2008;155:408–17. doi:10.1016/j.ahj.2007.11.008
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Kurisu S,
      2. Kihara Y
      . Clinical management of Takotsubo cardiomyopathy. Circ J 2014;78:1559–66. doi:10.1253/circj.CJ-14-0382
      OpenUrlCrossRefPubMed
    4. ↵
      1. Akashi YJ,
      2. Goldstein DS,
      3. Barbaro G, et al
      . Takotsubo cardiomyopathy: a new form of acute, reversible heart failure. Circulation 2008;118:2754–62. doi:10.1161/CIRCULATIONAHA.108.767012
      OpenUrlFREE Full Text
    5. ↵
      1. Yoshikawa T
      . Takotsubo cardiomyopathy, a new concept of cardiomyopathy: clinical features and pathophysiology. Int J Cardiol 2014;182C:297–303.
      OpenUrlPubMed
    6. ↵
      1. Templin C,
      2. Ghadri JR,
      3. Diekmann J, et al
      . Clinical features and outcomes of Takotsubo (stress) cardiomyopathy. N Engl J Med 2015;373:929–38. doi:10.1056/NEJMoa1406761
      OpenUrlCrossRefPubMed
    7. ↵
      1. Isogai T,
      2. Yasunaga H,
      3. Matsui H, et al
      . Out-of-hospital versus in-hospital Takotsubo cardiomyopathy: analysis of 3719 patients in the Diagnosis Procedure Combination database in Japan. Int J Cardiol 2014;176:413–7. doi:10.1016/j.ijcard.2014.07.110
      OpenUrlCrossRefPubMed
    8. ↵
      1. Brinjikji W,
      2. El-Sayed AM,
      3. Salka S
      . In-hospital mortality among patients with takotsubo cardiomyopathy: a study of the National Inpatient Sample 2008 to 2009. Am Heart J 2012;164:215–21. doi:10.1016/j.ahj.2012.04.010
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Murakami T,
      2. Yoshikawa T,
      3. Maekawa Y, et al
      . Characterization of predictors of in–hospital cardiac complications of takotsubo cardiomyopathy: multi-center registry from Tokyo CCU Network. J Cardiol 2014;63:269–73. doi:10.1016/j.jjcc.2013.09.003
      OpenUrlCrossRefPubMed
    10. ↵
      1. Schneider B,
      2. Athanasiadis A,
      3. Schwab J, et al
      . Complications in the clinical course of tako-tsubo cardiomyopathy. Int J Cardiol 2014;176:199–205. doi:10.1016/j.ijcard.2014.07.002
      OpenUrlCrossRefPubMed
    11. ↵
      1. Redfors B,
      2. Vedad R,
      3. Angerås O, et al
      . Mortality in takotsubo syndrome is similar to mortality in myocardial infarction—A report from the SWEDEHEART registry. Int J Cardiol 2015;185:282–9. doi:10.1016/j.ijcard.2015.03.162
      OpenUrlCrossRefPubMed
    12. ↵
      1. Elesber AA,
      2. Prasad A,
      3. Lennon RJ, et al
      . Four-year recurrence rate and prognosis of the apical ballooning syndrome. J Am Coll Cardiol 2007;50:448–52. doi:10.1016/j.jacc.2007.03.050
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. McMurray JJ,
      2. Adamopoulos S,
      3. Anker SD, et al
      . ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2012: The Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association (HFA) of the ESC. Eur J Heart Fail 2012;14:803–69.
      OpenUrlCrossRefPubMed
    14. ↵
      1. Steg PG,
      2. James SK,
      3. Atar D, et al
      . ESC Guidelines for the management of acute myocardial infarction in patients presenting with ST-segment elevation. Eur Heart J 2012;33:2569–619. doi:10.1093/eurheartj/ehs215
      OpenUrlFREE Full Text
    15. ↵
      1. Amsterdam EA,
      2. Wenger NK,
      3. Brindis RG, et al
      . 2014 AHA/ACC guideline for the management of patients with non-ST-elevation acute coronary syndromes: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 2014;130:e344–426. doi:10.1161/CIR.0000000000000134
      OpenUrlFREE Full Text
    16. ↵
      1. Chen ZM,
      2. Pan HC,
      3. Chen YP, et al
      . Early intravenous then oral metoprolol in 45,852 patients with acute myocardial infarction: randomised placebo-controlled trial. Lancet 2005;366:1622–32. doi:10.1016/S0140-6736(05)67661-1
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Mehta RH,
      2. Starr AZ,
      3. Lopes RD, et al
      . Incidence of and outcomes associated with ventricular tachycardia or fibrillation in patients undergoing primary percutaneous coronary intervention. JAMA 2009;301:1779–89. doi:10.1001/jama.2009.600
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Kumar S,
      2. Kaushik S,
      3. Nautiyal A, et al
      . Cardiac rupture in takotsubo cardiomyopathy: a systematic review. Clin Cardiol 2011;34:672–6. doi:10.1002/clc.20957
      OpenUrlCrossRefPubMed
    19. ↵
      1. Redfors B,
      2. Shao Y,
      3. Ali A, et al
      . Current hypotheses regarding the pathophysiology behind the takotsubo syndrome. Int J Cardiol 2014;177:771–9. doi:10.1016/j.ijcard.2014.10.156
      OpenUrlCrossRefPubMed
    20. ↵
      1. Wittstein IS,
      2. Thiemann DR,
      3. Lima JA, et al
      . Neurohumoral features of myocardial stunning due to sudden emotional stress. N Engl J Med 2005;352:539–48. doi:10.1056/NEJMoa043046
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Akashi YJ,
      2. Nef HM,
      3. Lyon AR
      . Epidemiology and pathophysiology of Takotsubo syndrome. Nat Rev Cardiol 2015;12:387–97. doi:10.1038/nrcardio.2015.39
      OpenUrlCrossRefPubMed
    22. ↵
      1. Madhavan M,
      2. Borlaug BA,
      3. Lerman A, et al
      . 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;95:1436–41. doi:10.1136/hrt.2009.170399
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Paur H,
      2. Wright PT,
      3. Sikkel MB, et al
      . High levels of circulating epinephrine trigger apical cardiodepression in a beta2-adrenergic receptor/Gi-dependent manner: a new model of Takotsubo cardiomyopathy. Circulation 2012;126:697–706. doi:10.1161/CIRCULATIONAHA.112.111591
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Austin PC
      . A critical appraisal of propensity-score matching in the medical literature between 1996 and 2003. Stat Med 2008;27:2037–49. doi:10.1002/sim.3150
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Bateman BT,
      2. Bykov K,
      3. Choudhry NK, et al
      . Type of stress ulcer prophylaxis and risk of nosocomial pneumonia in cardiac surgical patients: cohort study. BMJ 2013;347:f5416 doi:10.1136/bmj.f5416
      OpenUrlAbstract/FREE Full Text
    26. ↵
      1. Brookhart MA,
      2. Rassen JA,
      3. Schneeweiss S
      . Instrumental variable methods in comparative safety and effectiveness research. Pharmacoepidemiol Drug Saf 2010;19:537–54. doi:10.1002/pds.1908
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Fazio G,
      2. Pizzuto C,
      3. Barbaro G, et al
      . Chronic pharmacological treatment in takotsubo cardiomyopathy. Int J Cardiol 2008;127:121–3. doi:10.1016/j.ijcard.2007.04.013
      OpenUrlCrossRefPubMed
    28. ↵
      1. Palla AR,
      2. Dande AS,
      3. Petrini J, et al
      . Pretreatment with low-dose beta-adrenergic antagonist therapy does not affect severity of Takotsubo cardiomyopathy. Clin Cardiol 2012;35:478–81. doi:10.1002/clc.21983
      OpenUrlCrossRefPubMed
    29. ↵
      1. Neil C,
      2. Nguyen TH,
      3. Kucia A, et al
      . Slowly resolving global myocardial inflammation/oedema in Tako-Tsubo cardiomyopathy: evidence from T2-weighted cardiac MRI. Heart 2012;98:1278–84. doi:10.1136/heartjnl-2011-301481
      OpenUrlAbstract/FREE Full Text
    30. ↵
      1. Dawson DK,
      2. Neil CJ,
      3. Henning A, et al
      . Tako-Tsubo cardiomyopathy: a heart stressed out of energy? JACC Cardiovasc Imaging 2015;8:985–7. doi:10.1016/j.jcmg.2014.10.004
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors TI and HY were involved in study design and drafting of the manuscript. HY and KF obtained the data. TI, HY, HM and HT analysed and interpreted the data. TI, HY and HM performed the statistical analyses. All authors have read and approved the final version of the manuscript.

    • Funding This work was supported by grants for Research on Policy Planning and Evaluation from the Ministry of Health, Labour and Welfare, Japan (grant numbers: H27-Policy-Designated-009 and H27-Policy-Strategy-011). The funders had no role in the execution of this study or interpretation of the results.

    • Competing interests None declared.

    • Ethics approval The Institutional Review Board of The University of Tokyo.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    Linked Articles