Secondary forms of Takotsubo cardiomyopathy: A whole different prognosis

Abstract

Introduction

Takotsubo syndrome (TKS)¹ is a condition that presents features mimicking an acute coronary syndrome.^2–10 On top of this it has been deemed to be related to stressful situations and the prognosis usually seems to be good.^2–6,8,9 Nevertheless, it is well known that in the acute setting it is not lacking of complications, mainly heart failure.^2–13 In recent years, probably due to the dissemination of knowledge about it, its numbers and reports published are growing quickly but without a clear nomenclature or classification.

In different case series the syndrome has been linked to stressful, both psychic and physical, triggers.^2–15 However, other cases have been presented spontaneously without known apparent cause or precipitating factor. Thus, since the clinical profile and the trigger are so different, it is possible that the disease may not be exactly the same (different disease, same syndrome), and therefore, the prognosis would differ between patients.

Our aim was to study the TKS regarding its triggering circumstances and to compare baseline features, clinical presentation, natural history and long-term prognosis.

Material and methods

Study population

We performed a post-hoc analysis including patients with TKS between January 2003–December 2013 from our own prospective local database² together with the National Registry (RETAKO) for this disease,¹² with a protocol approved by our institution’s human research committee. Inclusion (Mayo-modified) criteria were:^2,8 (a) transient hypokinesis, akinesis or dyskinesis of left ventricular segments; (b) electrocardiographic new abnormalities and elevated cardiac troponin; (c) absence of recent significant head trauma and of obstructive coronary artery disease, angiographic evidence of acute plaque rupture or thrombosis (luminal narrowing >50%), as published elsewhere.^2,15

Medical history, laboratory findings and in-hospital course were thoroughly recorded for each patient. All patients underwent at least one complete echocardiographic study before and after discharge (excepting in hospital deaths). Early cardiac catheterisation was performed in all patients, ruling out coronary stenosis potentially responsible. All, excepting some dead patients, displayed a complete left ventricular motion recovery.

Patients were divided in two groups regarding their potential triggers: (a) none or an important psychic stress as ‘primary forms’ and (b) physical factors (i.e. asthma, surgery, trauma, etc.) as ‘secondary forms’.

Diagnosis and management were performed according to the attending physician’s criteria. Every patient completed all the procedures of this study and provided informed consent. After discharge, all patients were followed in the outpatient clinic or by telephone interview.

During follow-up, as main endpoints were recorded: recurrence, considered as development of a new TKS episode or pretty similar symptoms, mainly chest discomfort; readmission in cardiology because cardiovascular causes and any cause death. A combined event variable (MACE), considering all the former complications was also analysed. For event-free survival analysis, the first event during follow-up was considered, censoring the case.

Statistical analysis

SPSS 15.0 software for Windows (SPSS, Chicago, Illinois, USA) was used for analysis. Unless otherwise indicated, data are shown as mean value±standard deviation. Comparison between groups was performed using the Student’s t-test for continuous variables and the Pearson’s χ² test for categorical variables, when feasible. Long-term follow-up survival and events curves were prepared using the Kaplan-Meier method, and comparison between groups performed with the long-rank test. The Cox proportional hazard regression multivariate model was used to evaluate and select the variables independently associated with the presence of long-term events. An excessive number of variables in the multivariate analysis were eluded using a prespecified model that included those known to be associated with cardiovascular prognosis and regarding the univariate results. Thus, age (quantitative), gender, diabetes mellitus, smoking habit, obesity, dyslipidaemia, and hypertension (present in discharge reports), previous functional class, primary or secondary Takotsubo form, first left ventricle ejection fraction (LVEF) (<30%, 30–45%, 45–55%, >55%), maximum Killip degree and peak creatine kinase (quantitative in stratus), were included as covariates in the final models for death, MACE, readmission and recurrences. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated by backwards (Wald) stepwise regression analysis. Comparisons were considered significant in presence of a two-tail p value⩽0.05.

Results

Finally, 328 patients from 23 hospitals throughout the entire peninsula were included, mainly women (90.2%), with a mean age of 69.7 years (see Figure 1). Patients were divided, regarding their trigger, as 265 primary TKS forms and 63 secondary TKS forms after a severe concurrent physical stress. Patient’s features and comparison between those groups are shown in Table 1. Age, gender, previous functional class and cardiovascular risk profile displayed no differences between groups, before admission. However, the clinical presentation was rather different. Primary forms suffered chest pain as the main complaint (89.4% vs 50.7%, p<0.0001) with frequent vegetative symptoms.

Figure 1.

Case distribution (recruitment) by region

Open in new tabDownload slide

The secondary TKS cohort presented a trend to develop more advanced heart failure symptoms, including cardiogenic shock. In addition, maximum levels of creatine kinase, intensive care unit and in-hospital stay, reached statistically significant differences (Table 1). However, ST elevation percentage at admission, maximum QTc and onset LVEF or follow-up LVEF disclosed no differences between cohorts.

Regarding treatment before admission, there were no differences either, see Appendix 1. During admission we were able to find differences related to more intensive antithrombotic and anxiolytic drug use in the primary TKS group. Nevertheless, inotropic and mechanical ventilation use rates were higher in the secondary TKS group. Finally, after discharge, there were again differences with a more frequent prescription of beta-blockers and statins in the primary TKS group.

Table 2 displays events during follow up. Evolutive complications were higher in the secondary cohort. There were 16 deaths in the overall group (two sudden deaths; four due to sepsis; two neurological; two digestive; four respiratory; one pulmonary embolism; one neoplasia). Figure 2 shows the Kaplan-Meier graphs regarding death and the combined event variable during follow up and Table 3 illustrates the multivariate assessment for death and MACE. As depicted, secondary type presented prognostic relevance, remaining in the Wald multivariate equation for death (HR: 3.41; 95% CI 1.15–10.17, p=0.02), MACE (HR: 1.61, 95% CI 1.01–2.60, p=0.04), recurrences (HR: 1.85; 95% CI 1.06–3.22, p=0.02) and readmission (HR: 1.58; 95% CI: 0.95–2.62, p=0.07).

Figure 2.

Kaplan-Meier (event-free survival) graphs: (a) analysis regarding any cause death; (b) shows the evolution on the combined event variable (MACE) (any cause death/recurrence/readmission because cardiovascular causes); (c) displays recurrences between groups; (d) graph showing cardiology readmissions during long term follow-up

Open in new tabDownload slide

Discussion

TKS is a new condition described in the 1990s in Japan.^1,14 Various diagnostic criteria and several pathophysiologic explanations have been proposed, most of them related to stressful conditions.^2–6,8,14 Although related to blood-circulating catecholamine levels, its definitive pathophysiology remains elusive. Moreover, although TKS has been described in many circumstances, there is a lack of a clear classification beyond typical and atypical forms or with uni or biventricular involvement.^6,13,16

In this regard, to the best of our knowledge, this article constitutes the first attempt to establish a prognostic classification into two simple working categories. Based on a large multicentric series, mainly Caucasian, with a thorough and long-term follow-up,¹⁵ we classified as primary forms those without a clear trigger or complaining of an intense stressful setting. On the other hand, secondary forms would be those associated with other adjuvant acute organic conditions (surgery, bronchospasm, severe trauma, sepsis, etc.). As we showed in this study (Table 1), clinical presentation is fairly different between groups as well. However, differences in reported symptoms at presentation may have been influenced by associated conditions.

Thus, we propose a simple nomenclature for TKS in order to unify definitions and facilitate further research, following modified Mayo criteria as a diagnostic set⁵ (Figure 3). Moreover, we have previously shown that a simple tool such as Killip degree has diagnostic value. Thus, an assessment on this issue would be warranted as well from a practical point of view.² In this sense, although we did not address this issue, other specific severity scores, such Acute Physiology and Chronic Health Evaluation (APACHE), Simplified Acute Physiology Score (SAPS), Charlson or Multiple Organ Dysfunction Score (MODS) could provide an orientation on prognostic grounds.

Figure 3.

A proposed working classification for Takotsubo syndrome (TKS). Proved TKS is made when left ventricular (LV) motion abnormalities recover completely. If the patient dies before (i.e. in hospital), probably this criterion can be omitted if the other Mayo criteria are fulfilled

Open in new tabDownload slide

Though in general, after the acute phase, TKS over-all carries a good prognosis, our data suggest that the secondary patients present more complications (short and long-term). We feel that, although initially, the patients’ profiles seem to be not so different (Table 1) between both groups, an acute disease vigorous enough (sepsis, trauma, surgery, effort) to be able to provoke intense myocardial stunning, such as TKS is, probably has to mark a prognostic influence.

Thus, apart from Killip degree and other points important in the acute setting (LVEF, biomarkers peak, symptoms), and some clinical features, such as previous functional class,² considering the type of TKS would have prognostic implications. Although we could not find several differences between clinical profiles (Table 1) it seems reasonable to think that ‘secondary’ form patients may have a worse initial situation due to trauma and sepsis compared to patients with primary TKS and no concomitant disease.

Although, by definition, TKS is intended to be transient and benign, even after discharge, those patients with secondary forms are probably more fragile and need closer surveillance, both during hospitalisation and during outpatient follow-up, at least in the first years. After a while, we observed that the event-free survival Kaplan-Meier rates tend to be parallel (death) or convergent (rest of events). This finding is difficult to interpret accurately because of the small population at risk at that moment (>80 months) in our Kaplan-Meier analysis, and could reflect the natural evolution of an aged population (mean >69 years in both groups).

Several papers have explored mortality after TKS. Recently, Brinjikji et al. studied in-hospital mortality in 24,701 patients with TKS in the National (US) Inpatient database samples and concluded that critical illness was the main driver of in-hospital mortality and that male TKS suffered from significantly more mortality.¹⁷ In addition, a Japanese group recently showed that in-hospital TKS was associated with more severe clinical background and poorer short-term prognosis than, the most frequent, out-of-hospital TKS.¹⁸ Our results, obtained from a primary source basis and with a clinical follow-up, agree with that.

These findings could have management implications. Thus, primary forms in a stable Killip I-II patient,² would probably only need a short stay and do well, even outside the intensive care unit setting. On the contrary, probably secondary forms would benefit from an aggressive and, if possible, cause-directed approach.

Limitations

We cannot be sure that all patients enrolled in our registry were consecutive in their respective centres and time frames. However, since we were able to include a considerable amount of cases, we think that this limitation is only a minor one if we consider previously published works. Regarding the high prevalence of recurrences, probably there is a large list of causes involved in TKS development, and since this diagnosis was established by each attending doctor, the diagnostic criteria would vary. Moreover, the management of these patients was the current management at that time in every hospital and it depended on the local physician. So, in this way, we believe our results reflect real clinical practice more closely.

Conclusion

Patients with secondary TKS could present worse short and long-term prognosis in terms of mortality, recurrences and readmissions. This fact warrants thorough surveillance and aggressive management during in-hospital stay but during outpatient follow-up as well, since secondary forms are not so benign. We propose a simple working nomenclature for TKS.

Acknowledgement

All of the authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation, and have approved the manuscript. The authors wish to acknowledge all RETAKO contributors: Jaume Figueras (Servicio de Cardiología, Hospital Valld’Hebron, Barcelona), Francisco Ridocci-Soriano (Servicio de Cardiología, Consorcio Hospital General Universitario de Valencia, Valencia), Alberto Duran-Cambra (Servicio de Cardiología, Hospital Sant Pau, Barcelona), José R. Ruiz-Arroyo (Servicio de Cardiología, Hospital Clínico Lozano Blesa, Zaragoza), Martín J. García-González (Servicio de Cardiología, Hospital Universitario de Canarias, Sta. Cruz de Tenerife), Manuel de Mora Martín (Servicio de Cardiología, Hospital Carlos Haya, Málaga), Álvaro Aceña and Hans P. Gaebelt (Servicio de Cardiología, Hospital Fundación Jiménez Díaz, Madrid), Irene Santos-Pardo (Servicio de Cardiología, Hospital Universitario GermansTrias i Pujol, Badalona, Barcelona), José M. García Acuña and Milagros Pedreira Pérez (Servicio de Cardiología, Hospital Clínico Universitario de Santiago, Santiago de Compostela), Andrea Moreno-Arcinegas, Jaime Benítez-Peyrat, Álvaro-León Moreno Reig and Luis Ruiz Valdepeas Herrero (Servicio de Cardiología, Hospital General Universitario de Ciudad Real, Ciudad Real), Ernesto Valero-Picher (Servicio de Cardiología, Hospital Clínico de Valencia, Valencia), Pedro L. Sánchez and Javier Jiménez-Candil (Servicio de Cardiología, Hospital Universitario de Salamanca, Salamanca), Ramón Bascompte and Fernando Worner (Servicio de Cardiología, Hospital Universitario Arnau de Vilanova, Lérida), María C. Manzano Nieto (Servicio de Cardiología, Hospital Severo Ochoa, Leganés, Madrid), Javier García (Servicio de Cardiología, Hospital Príncipe de Asturias, Alcalá de Henares, Madrid), Juan C. García-Rubira and Manuel Lobo (Servicio de Cardiología, Hospital Virgen de la Macarena, Sevilla).

Conflict of interest

The authors declare that there is no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. The retako website (www.retako.es) was funded by Astra Zeneca.

References