β-blockers and outcomes of Takotsubo syndrome: need more clinical data

Takotsubo syndrome (TTS) has gained more awareness and attention in clinical practice in the last decade. Prior studies revealed important insights into the patient characteristics and outcomes of TTS. 2 One of the most notable facts is that the prognosis of TTS is not as benign as initially expected and is comparable to acute coronary syndrome. 2 Therefore, it is no surprise that physicians and researchers investigate a potential treatment option to improve the prognosis of TTS. Although the pathophysiology of TTS remains to be fully elucidated, catecholamines appear to play a critical role, as evidenced by the fact that TTS is frequently triggered by acute emotional or physical stress along with excess plasma catecholamine levels. 3 As a result, β-blockers (BBs) have been empirically considered a reasonable therapy for TTS in the absence of randomised clinical trials. In the current issue of Heart, Silverio et al examined the association between BBs and longterm survival using 825 patients in the Takotsubo Italian Network registry. The authors demonstrated that BB prescription at discharge was significantly associated with lower allcause mortality after TTS (6.8% vs 13.6%; adjusted hazard ratio (aHR)=0.563, 95% confidence interval (CI)=0.356 to 0.889, p=0.014) during a median followup of 24 months, particularly with lower noncardiac mortality (4.9% vs 10.7%; aHR=0.525, 95% CI=0.309 to 0.893, p=0.018) rather than cardiac mortality (1.8% vs 3.0%; aHR=0.699, 95% CI=0.284 to 1.722, p=0.436). Also, the effect modification was observed in patients with hypertension and those who developed cardiogenic shock during the acute phase (p for interaction <0.05). Meanwhile, there was no significant association between BB prescription and TTS recurrence. The authors are to be congratulated on their contribution to current literature on the topic. Nonetheless, several discussions need to be raised about the results and potential limitations of the study. One may expect that BBs theoretically have potential in reducing cardiac mortality after TTS (maybe through the facilitated recovery from cardiac dysfunction or the prevention of fatal ventricular arrhythmia or other cardiac events), but not noncardiac mortality. However, contrary to this expectation, Silverio et al demonstrated that BB prescription at discharge was significantly associated with lower noncardiac mortality, but not cardiac mortality. The statistically insignificant association between BB prescription and cardiac mortality may be at least partly due to the low cardiac mortality rate (2.3%), as Silverio et al discussed. Notably, however, their data revealed that the estimated risk reduction in allcause mortality by BB use was driven largely by the reduction in noncardiac mortality. How could BB use be strongly associated with lower (nearly half) noncardiac mortality after TTS? As Silverio et al speculated, there might be several possible mechanisms for it. Meanwhile, it might also be due to residual biases or/and confounders that could not be handled in the observational study. For example, among the elderly population with TTS, patients could be frail, and BBs might be less likely to be prescribed at discharge in more frail patients because of the uncertain benefit and the concern for adverse events related to BBs. This potentially could bias the results towards a better prognosis in BB users, although it is conjectural. Furthermore, in the study by Silverio et al, the only available data on BB use after TTS was whether BB was prescribed or not at hospital discharge, without any data on postdischarge continuation or discontinuation of BBs, which is a crucial limitation of this study. Therefore, we agree with Silverio et al that the present study should be considered hypothesisgenerating. At this point, BBs should not be prescribed for TTS solely based on the present results unless more data are available to support the findings of the present study. The study design and main results of eight original studies (sample size >200) regarding BB use for TTS 4 6–11 are summarised in table 1. All were nonrandomised studies. Among them, two studies focused on the inhospital outcomes, both demonstrating no significant association between BB use and better outcomes. 7 The remaining six studies examined the effectiveness of BB on longer followup outcomes, demonstrating inconsistent results. 4 8–11 In addition to the observational design, several limitations and considerations need to be acknowledged in these studies (table 2): (1) the effect of BBs for TTS might be different between preadmission users (ie, patients who already took BBs before admission with TTS) and new users (ie, patients who start to take BBs after admission with TTS); (2) the effect of BBs may not be the ‘class effect’ but rather different across BB subclasses; (3) medication continuation/discontinuation and adherence could modify the impact of BBs on prognosis; (4) given the high age of patients with TTS, the effectiveness of BBs needs to be weighed against their potential adverse effects (eg, bradycardia, hypotension, or thoserelated injuries or hospitalisations); (5) whether TTS is complicated by left ventricular outflow tract obstruction (prevalence 10%–25%) or not could be a key factor, especially in assessing acutephase outcomes because BBs appear effective in reducing the gradients in the outflow tract ; (6) last but not least, patients’ general conditions (eg, frailty) could be confounders in outcome assessment in this elderly population and thus should be adjusted for a fair comparison between BB users and nonusers unless it is a randomised comparison. In these contexts, all eight studies in table 2 seem to lack some essential data that could act as effect modifiers or unmeasured confounders, suggesting that they should be considered hypothesisgenerating. As for outcomes, noncardiac deaths accounted for 76% (60/79) of all deaths during the followup in the study by Silverio et al, which is comparable to the RETAKO study (72%, 39/54). This finding suggests that the majority of deaths in patients experiencing TTS are noncardiac. In addition, TTS manifests a rapid recovery from cardiac dysfunction within days to weeks from the onset in most cases. Thus, although TTS is considered an acute heart failure syndrome mimicking acute coronary syndrome, it may be worth exploring a therapeutic strategy to reduce the noncardiac mortality among patients with TTS. If we investigate the effectiveness of cardiovascular drugs such as BBs, it may be reasonable to focus on patients Department of Cardiovascular Medicine, Heart, Vascular and Thoracic Institute, Cleveland Clinic, Cleveland, Ohio, USA Department of Cardiovascular Medicine, Chiba University Graduate School of Medicine, Chiba, Japan