Prevalence and Prognostic Impact of Diabetes in Takotsubo Syndrome: Insights From the International, Multicenter GEIST Registry Free

Since its first systematic description more than two decades ago, Takotsubo syndrome (TTS) is increasingly recognized worldwide as an important differential diagnosis in patients with suspected acute coronary syndrome. The disease is characterized by typical, distinct ventricular contraction abnormalities that are not related to obstructive coronary artery disease and recover completely within several days to weeks (1–3). Postmenopausal women are primarily affected from TTS, and the occurrence is frequently triggered by physical or emotional stress (1). Recent studies demonstrated considerable short- and long-term adverse event rates with subsequent reevaluation of the disease (4–6). However, despite the progress achieved, several important aspects of TTS are still unclear and require additional investigations. These pending issues include the elusive pathophysiology of TTS and the identification of predictive factors for adverse outcome to guide monitoring and therapy. Recently, substantial research efforts were directed to the role of diabetes in patients with TTS. Experimental and clinical evidence exists that sympathetic overdrive and catecholamine excess are associated with, or can lead to, TTS (7–9). This explanation is currently the favored pathophysiological theory. Diabetic autonomic neuropathy can ameliorate catecholamine-mediated effects to the heart and has therefore been considered protective for the occurrence of TTS (10,11). Data suggesting a low prevalence of diabetes in TTS cohorts compared with the general population were interpreted in support of this concept (12). Furthermore, a previous single center study reported lower adverse event rates in TTS patients with diabetes compared with TTS patients without diabetes (13). This observation was also attributed to a potentially attenuated manifestation of TTS in patients with diabetes. However, other investigations could not confirm a beneficial outcome in TTS patients with diabetes and even reported increased mortality rates (4).

In view of these inconsistent findings, the aim of this study was to assess the prevalence and prognostic impact of diabetes in TTS in a large, international, multicenter study.

This study included 836 consecutive patients with TTS enrolled in the international, multicenter GErman Italian STress cardiomyopathy (GEIST) registry, which involves eight centers: University Heart Center Lübeck, Germany (n = 108); Heart Center Leipzig – University Hospital, Germany (n = 178); University Medical Centre Mannheim, Germany (n = 141); University Hospital of Foggia, Italy (n = 185); Casa Sollievo della Sofferenza Hospital, Italy (n = 30); San Paolo Hospital, Italy (n = 36); University Hospital of Palermo, Italy (n = 79); and University of Tor Vergata General Hospital, Italy (n = 79). The prospective inclusion of patients in the registry was performed according to recommended diagnostic criteria for TTS consisting of 1) transient regional wall motion abnormalities of the left or right ventricle, which are frequently preceded by a stressful trigger and usually extend beyond a single epicardial vascular distribution; 2) absence of culprit atherosclerotic coronary artery disease; 3) new and reversible electrocardiography abnormalities; 4) elevated cardiac troponin and serum natriuretic peptide levels; and 5) recovery of ventricular systolic function at follow-up (2,3). Exemplary images of the three most common contraction patterns are provided in Fig. 1. Recovery of left ventricular (LV) systolic function was documented 3–6 months after the acute event in all surviving patients.

Diabetes was defined as preexisting disease on admission, which was treated by either diet and lifestyle measures alone or the additional use of oral glucose-lowering medication and insulin, or newly diagnosed diabetes during the hospital stay for the TTS episode (fasting plasma glucose level ≥7.0 mmol/L [126 mg/dL], plasma glucose 2 h after a 75-g glucose load ≥11.1 mmol/L [200 mg/dL], or glycated hemoglobin ≥6.5%). The prognostic implications of diabetes were evaluated regarding all-cause mortality, which was assessed during regular outpatient visits or scheduled telephone contact with the patients, relatives, and treating physicians. All events were verified via medical records and evaluated by a clinical events committee.

The study was approved by the local ethics committees of the participating study centers, and all patients gave written informed consent before inclusion.

Categorical variables are presented as number (percentage) of patients and were compared with the χ² test. Continuous data were tested for normal distribution with the Shapiro-Wilk test. Since all continuous variables were nonnormally distributed, they are reported as median (interquartile range [IQR]) and were compared with the Mann-Whitney U test.

Baseline characteristics and outcome were compared between TTS patients with diabetes and TTS patients without diabetes. Mortality rates were illustrated with the Kaplan-Meier method and compared with log-rank testing. Univariate and stepwise multivariate Cox regression analysis was performed to obtain hazard ratios (HRs) and 95% CIs. Multivariate testing included only variables with a significant P value in univariate analysis. Statistical analyses were performed with SPSS (version 17.0; SPSS, Chicago, IL). A two-sided P value ≤0.05 was considered statistically significant.

Of the 836 consecutive patients included in the GEIST registry, 10 (1.2%) did not have diabetes status available, resulting in a final study population of 826 patients with TTS for the present analysis. The baseline clinical characteristics provided in Table 1 reflect a typical TTS cohort, with a predominance of postmenopausal women. The acute event was frequently preceded by a stressful trigger, and apical ballooning was the most prevalent contraction pattern. LV ejection fraction was moderately impaired at acute presentation and recovered during follow-up. The prevalence of diabetes was 21.1% (n = 174), and all cases of diabetes were type 2 diabetes.

TTS patients with diabetes were older (P < 0.001), with a higher prevalence of hypertension (P < 0.001) and physical triggers (P = 0.041), and the proportion of patients who were male among TTS patients with diabetes was higher (P = 0.003), compared with patients without diabetes (Table 1). Furthermore, typical apical ballooning was observed more frequently among patients with diabetes (P = 0.010), which resulted in a more severely impaired LV ejection fraction (P = 0.008), a higher rate of pulmonary edema (P = 0.032), and, consequently, a longer hospital stay (P = 0.009). Recovery of systolic LV function was similar in both groups and not affected by diabetes status (P = 0.198).

Follow-up data were available in 755 patients (91.4%) after a median of 2.5 years (IQR 0.5–4.9) and showed an all-cause mortality of 19.6% (n = 148). Stratification according to the presence of diabetes revealed significantly higher mortality rates in TTS patients with diabetes compared with TTS patients without diabetes (31.4% vs. 16.5%; HR 2.02 [95% CI 1.43–2.84]; P < 0.001). As illustrated in the Kaplan-Meier plot (Fig. 2), survival curves initially show a parallel course before starting to diverge several months after the initial event, which results in a significant difference >1 year after the acute TTS episode. Accordingly, 28-day mortality did not differ significantly between groups (6.4% vs. 5.7%; HR 1.11 [95% CI 0.55–2.25]; P = 0.772).

After adjustment for other clinical risk factors in multivariate Cox regression analysis, diabetes remained a significant predictor of increased mortality (HR 1.66 [95% CI 1.16–2.39]; P = 0.006) (Table 2). In addition, age, male sex, the presence of malignancies, LV ejection fraction, the occurrence of cardiogenic shock, and preceding emotional trigger were identified as determinants of outcome.

This study investigated the prevalence and prognostic relevance of concomitant diabetes in one of the largest TTS populations studied to date. The main results are as follows: 1) diabetes can be found in approximately one-fifth of TTS patients, 2) short-term outcome is not influenced by diabetes despite higher rates of pulmonary edema and longer hospitalization, 3) longer-term mortality is significantly higher in TTS patients with diabetes than in TTS patients without diabetes, and 4) diabetes is an independent predictor of longer-term mortality in TTS irrespective of other risk factors (Table 3).

Diabetic cardiovascular autonomic neuropathy occurs in approximately one-third of patients with type 2 diabetes and causes damage to the autonomic nerve fibers that innervate the heart and blood vessels (10,11). The potential clinical manifestation ranges from unspecific symptoms (e.g., abnormalities in heart rate variability, exercise intolerance, or orthostatic dysregulation) to an increased risk of cardiovascular events and mortality (10,11). Catecholamine hyposecretion and amelioration of adrenergic effects to the heart were postulated in cases of autonomic neuropathy and led to the association between diabetes and the pathophysiology of TTS (12,14). Although the exact etiology of the disease is still elusive, convincing evidence exists that catecholamine excess and sympathetic overstimulation seem to play a major causative role (7–9). The observation of a low diabetes prevalence in some TTS populations in conjunction with the aforementioned pathophysiological pathways was interpreted in support of a protective effect of diabetes for the occurrence of TTS (12). However, although this concept might be theoretically conceivable, it has several shortcomings. The reported diabetes frequency among TTS patients has shown a wide range: from 1.6% to 25.5% (12,15,16). Diabetes rates of 1.6% in patients with an average age of 65 years (15) seem extremely low even when one expects a one-third reduction in TTS related to patients with diabetes with autonomic neuropathy. A potential explanation could be the use of National (Nationwide) Inpatient Sample (NIS) databases and ICD codes (15), which is an error-prone approach, given the possibility of incomplete data entry. Using prospective data collection, our analysis showed diabetes in 21.1% of patients with TTS, which is slightly higher than the expected sex- and age-adjusted rates in the general population of the participating countries, Italy and Germany (17,18). Therefore, our findings do not support the hypothesized protective role of diabetes for the development of TTS. Apart from that, low diabetes prevalence alone is not sufficient to prove the prevention of TTS on the basis of autonomic neuropathy. Verification of a causal connection requires a more sophisticated approach including experimental models (19).

Several recent studies have consistently shown an unfavorable prognosis after TTS, with substantial short- and long-term mortality rates (4–6). Deaths were directly related to TTS-associated adverse events (e.g., arrhythmias or cardiogenic shock) (20–23) as well as concomitant noncardiovascular factors (e.g., malignancies) (24,25). Efforts to identify determinants of outcome included diabetes and showed conflicting results. While diabetes was predictive of long-term mortality in one study (4), another investigation reported a lower 1-year event rate in TTS patients with diabetes compared with that in TTS patients without diabetes (P = 0.04 [composite of all-cause death, life-threatening arrhythmias, thromboembolic events, rehospitalization for heart failure, TTS recurrence, and stroke]) (13). Different end point definitions and follow-up durations could at least in part explain divergent findings in these studies. In theory, lower event rates in case of concomitant diabetes could again be attributed to diabetic cardiovascular autonomic neuropathy resulting in an attenuated course of TTS. However, the present multicenter study with adequate statistical power (large sample size and sufficient number of events) clearly rebuts a “diabetes paradox” in TTS. The frequency of short-term TTS-associated complications was not decreased in patients with diabetes. On the contrary, TTS patients with diabetes showed a higher incidence of pulmonary edema, probably as a consequence of the lower initial LV ejection fraction. Consistently, typical apical ballooning, which has been associated with more severely impaired systolic function, was also more prevalent in the diabetes cohort (26). Nevertheless, short-term outcome did not differ between TTS patients with diabetes and TTS patients without diabetes. In the long run, however, mortality in TTS patients with diabetes significantly exceeded that of TTS patients without diabetes. Therefore, the long-term deleterious effects of diabetes seem more pivotal for the observed difference than the impact of TTS itself, albeit this hypothesis cannot be definitely proven by our study because of missing data regarding causes of death. In view of the independent prognostic value observed in the current study, diabetes should be recognized as an important risk factor and a therapeutic target to optimize long-term outcome in patients with TTS.

Although the current study has several strengths (e.g., multicenter approach, large cohort of consecutive TTS patients, prospective patient identification by experienced investigators, “hard” clinical end point), some limitations should be acknowledged. The analysis is obviously limited by its registry-based design, with differences regarding the recorded parameters between centers. Missing data concerning the exact causes of death hamper a further evaluation of the impact on cardiovascular and noncardiovascular deaths. Furthermore, information regarding medical treatment of diabetes, actual glycemic status, and long-term blood glucose control (glycated hemoglobin) were not available in the entire population and therefore could not be analyzed.

Diabetes is not uncommon in patients with TTS and is associated with significantly elevated longer-term mortality rates. Furthermore, it is an independent predictor of outcome irrespective of further risk factors. Therefore, diabetes might represent an additional parameter for optimized risk stratification after TTS.

Acknowledgments. The authors thank the following collaborators of the GEIST registry: Luisa De Gennaro, San Paolo Hospital; Francesca Guastafierro and Nicola Tarantino, Department of Medical and Surgery Sciences, University of Foggia; and Fabiana Romeo, Annalisa Viele, and Pasquale La Prova, Division of Cardiology, University of Rome Tor Vergata.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. T.S. and I.E.-B. designed the study. F.S., I.E.-B., N.D.B., and I.A. contributed to the design of the study. All authors collected data. T.S. and I.E.-B. analyzed data and wrote the manuscript. All authors discussed the results and implications and commented on the manuscript at all stages. T.S. and I.E.-B. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.