Takotsubo syndrome (TS) is an acute cardiomyopathy that is triggered by physical or emotional stress and has an estimated incidence of up to 1/6,700 surgical procedures,1 representing up to 10% of acute coronary syndrome incidence after non-cardiac surgery.2 It is characterized by reversible left ventricular systolic dysfunction in the absence of coronary arterial occlusion, with a 0–8% rate of in-hospital mortality.3,4,5,6,7,8,9,10,11

Worldwide, the uniqueness of the syndrome has given rise to several investigations on the events that trigger this cardiomyopathy, aided by available literature on the subject in the form of case reports, case series, and national or regional registries. Ghadri et al. proposed a novel classification for this syndrome: TS related to emotional stress (class I); TS related to physical stress (class II), which includes TS secondary to physical activities or medical conditions or procedures (class IIa) and TS secondary to neurologic disorders (IIb); and TS without an identifiable triggering factor (class III).12 It is now recognized that physical stress is the most common trigger of TS.13,14,15

Regarding TS due to physical stress, some studies have differentiated perioperative TS (pTS), occurring in the context of surgical, interventional, or anesthetic procedures, from non-pTS (npTS), which occurs in the context of an acute medical illness or critical condition.16, 17 Nevertheless, the vast majority of studies on this subject have focused on npTS, probably because of its higher incidence and larger available cohorts, but evidence suggests that pTS may not share many of the demographic or clinical characteristics of npTS. Additionally, patients in the perioperative setting may encounter different types of physical stressors such as pain, anxiety, residual neuromuscular blockade, and exogenous administration of catecholamines.1, 18

The primary aim of the study was to describe the baseline characteristics and surgical/anesthesia-related triggering events for a group of patients diagnosed with pTS. The secondary aim was to describe their clinical course and in-hospital outcomes, as well as comparing these factors between patients with pTS or npTS.

Methods

We retrospectively reviewed a single-centre cohort of patients diagnosed with TS that was triggered by physical stressors. Ethical approval was provided by the Ethical Committee of the Hospital Italiano de Buenos Aires in Buenos Aires, Argentina (Chairperson: Dr. Augusto Pérez, Ethical Committee N° 3382) on 9 October 2017. Requirement for written informed consent was waived by the ethics committee because of the retrospective nature of the study.

Study population

We included all consecutive adult patients who were admitted to Hospital Italiano de Buenos Aires between 1 June 2008 and 30 November 2017 and diagnosed with TS according to the revised criteria of the European Society of Cardiology during hospitalization.3

Data sources

The Research Department of our institution, specifically the Information Management Area, conducted the search. They searched the electronic medical records for patients who were diagnosed with TS during this period. The electronic medical records of our institution have a specific field in which physicians describe all relevant diseases diagnosed during hospitalization. A total of 51 adult patients were diagnosed with “physically triggered TS”, which was defined as TS related to physical stress (particularly, TS secondary to medical conditions or procedures) during the study period. Of these, 21 were identified as cases of pTS since diagnosis occurred during surgical, interventional, or anesthetic procedures or within a 30-day postoperative period; the remaining 30 patients were diagnosed with npTS, given that they presented probable causes of physical stress (acute medical conditions, asthma, exacerbations of chronic obstructive pulmonary disease, sepsis, or anaphylaxis) but did not fulfill the criteria for pTS. We identified the diagnosis time using electrocardiography, troponin dosage, and echocardiography.

We obtained baseline data from the electronic medical records, including age, sex, comorbidities (e.g., diabetes, arterial hypertension, dyslipidemia, tobacco smoking, myocardial infarction, cerebrovascular disease, peripheral vascular disease, cancer), and medications (use of antipsychotics, antidepressants, mood stabilizers, and beta-receptor blockers over the previous 6 months). We obtained the following data regarding the clinical and biochemical characteristics of TS at diagnosis: clinical presentation (evidence of heart failure [hypotension, dyspnea, hypoxia, and pulmonary edema], symptoms of acute coronary syndrome, or unexplained syncope/cardiac arrest), electrocardiographic abnormalities (ST-segment elevation or other changes, including ST depression, T wave inversion, and left bundle block), echocardiographic TS pattern (classical apical, midventricular, basal or inverted pattern, global hypokinesia and sectorial hypokinesia), initial left ventricular ejection fraction (LVEF), ultrasensitive troponin value at diagnosis calculated as a factor × upper limit of the normal range (ULN), pro B-type natriuretic peptide value at diagnosis, peak ultrasensitive troponin value as a factor × ULN, ultrasensitive troponin increase factor during hospitalization (ratio between troponin value at diagnosis and the maximum troponin value during hospitalization) and evaluation of obstructive coronary artery disease (coronary angiography, thallium perfusion scan, cardiovascular magnetic resonance imaging, or none).

The in-hospital outcomes reviewed included days of in-patient stay, days of intensive care unit (ICU) stay, in-hospital mortality, and LVEF at discharge. Other in-hospital outcomes considered were treatment requirements such as vasopressor/inotropic support, mechanical ventilation, and dialytic therapy. We also gathered data regarding severe complications such as pulmonary edema (from chest radiograph), cardiogenic shock requiring ventricular assistance, unexplained syncope/cardiac arrest, atrial fibrillation, left ventricular outflow tract obstruction, mitral regurgitation, and thrombus formation.

Additionally, the American Society of Anesthesiologists (ASA) Physical Status, procedure/surgery type, procedure immediacy (elective or urgent/emergency), type of surgery according to cardiovascular risk stratification in non-cardiac surgery,19 and type of anesthesia (sedation, general anesthesia, regional anesthesia or combined regional/general anesthesia) were recorded for pTS cases.

Data analysis

We performed descriptive analyses using the mean (standard deviation) (normal distribution) and median [interquartile range (IQR)] (non-normal distribution) for continuous variables. Numbers and percentages were used for categorical variables. We compared qualitative variables derived from each group using the Chi square test or Fisher’s exact test in cases involving expected low counts. The Student’s t test was used to analyze normally distributed quantitative data, while the Wilcoxon rank-sum test was used to analyze non-normally distributed quantitative data. We used STATA 13 (StataCorp, College Station, TX, USA) for statistical analyses, and the significance level was set at P < 0.05.

Results

A total of 21 patients with a diagnosis of pTS were identified between 1 June 2008 and 30 November 2017. During this period, 305,906 procedures under anesthesia were performed in our institution. This calculates to an incidence of 0.007% or approximately 1/15,000 surgical procedures. Table 1 summarizes the baseline characteristics of the patients with pTS. The median patient age was 74 yr and the majority were women. Most patients (71%) were categorized as ASA Physical Status II–III (Table 1). Eight (38%) patients presented with current malignancy, including pancreatic cancer (n = 1), bladder cancer (n = 1), hepatocellular carcinoma (n = 1), colorectal cancer (n = 2), atrial myxoma (n = 1), and laryngeal cancer (n = 2). None of the patients were receiving active chemotherapy at the time of surgery.

Table 1 Baseline characteristics of the patients with perioperative Takotsubo syndrome
Full size table

Of the 21 surgeries, 10 (48%) were classified as low-to-intermediate cardiac-risk procedures, including gastrointestinal endoscopy (n = 2), colovesical fistula repair (n = 1), larynx surgery (n = 3), breast surgery (n = 1), laparoscopic cholecystectomy (n = 1), subdural hematoma drainage (n = 1), and wound surgical toilette (n = 1), while 10 (48%) were classified as high-cardiac-risk surgeries, including pancreatectomy (n = 1), laparoscopic Heller myotomy associated with fundoplication (n = 1), bowel resection (n = 1), liver transplantation (n = 2), kidney transplantation (n = 1), ultralow anterior resection with liver metastasectomy (n = 2), laparotomy with gastric suture (n = 1, perforated gastric ulcer), and total colectomy (n = 1). One (5 %) patient underwent cardiac surgery because of an atrial myxoma. General anesthesia was used in 19 (90%) patients. In 29% of patients, pTS was diagnosed during the intraoperative period, 33% postoperatively until day 3, and 38% beyond day 3 (Table 1).

Congestive heart failure was the most frequent clinical presentation, in 12/21 (57%) patients (Table 2). Nine (43%) patients presented with ST-segment elevation on electrocardiography, while the remaining 12 (57%) patients presented with other electrocardiographic abnormalities such as ST-segment depression or T wave inversion. The median (IQR) LVEF at diagnosis was 35 (35–42)%, and only 11 (52%) patients presented with a typical apical TS echocardiographic pattern; focal hypokinesia, inverted or basal pattern, and global hypokinesia were found in 6/21 (28%), 2/21 (10%), and 1/21 (5%) patients, respectively. One (5%) patient presented with a normal echocardiographic pattern. Troponin increased by a median (IQR) factor of 1.2 (1.0–1.7) during hospitalization (Table 2). Coronary artery angiography was performed for 12/21 (57%) patients. The remaining 9/21 (43%) patients did not undergo coronary artery angiography because of either hemodynamic instability or family refusal during the acute phase of pTS. Nonetheless, these nine patients showed a complete recovery of LVEF (> 55%) within three months following the event, as documented by echocardiography. One (5%) patient underwent a cardiac magnetic resonance after the acute phase. None of these patients presented an ultrasensitive troponin increase factor of > 1.8 during hospitalization.

Table 2 Clinical features, laboratory, and imaging data of the patients with perioperative Takotsubo syndrome
Full size table

Complete recovery of LVEF was formally documented by echocardiography in 19/21 (90%) patients, in a median (IQR) time period of 7 (4–41) days. The remaining two patients showed clinical recovery but did not undergo echocardiography at the institution.

Complications, including unexplained pulmonary edema (n = 3), cardiogenic shock (n = 1), syncope/cardiac arrest (n = 5), atrial fibrillation (n = 5), left ventricular outflow tract obstruction (n = 1), mitral regurgitation (n = 2), and thrombus ventricular formation (n = 1) occurred in 18/21 (86%) patients (Table 3). The patient who experienced cardiogenic shock requiring extracorporeal membrane oxygenation had undergone cardiac surgery (myxoma resection) and showed spontaneous, complete recovery of LVEF. The median (IQR) ICU stay was 8 (5.5–15.5) days. Two (10%) patients died during the in-hospital stay; one of them presented with pTS on the second postoperative day after surgical repair of an intestinal perforation; death occurred on postoperative day 32 because of the progression of the primary oncological illness (advanced bladder carcinoma with refractory renal injury). The other patient presented with intraoperative pTS during laparotomy for gastric perforation repair. Although an improvement of LVEF was observed from severe to moderate deterioration, the patient died on postoperative day 6 because of septic shock associated with peritonitis.

Table 3 In-hospital outcomes of the patients with perioperative Takotsubo syndrome
Full size table

When patients with pTS were compared with a cohort of 30 patients with npTS, patients with pTS were found to be significantly younger than those with npTS (median [IQR] age, 74 [55–82] yr vs 84 [76–88] yr; mean difference, 10 yr; 95% confidence interval [CI], 0 to 20; P = 0.005) and they presented with a worse LVEF at diagnosis (35 [35–42]% vs 42 [38–47]%; mean difference, 8%; 95% CI, 2 to 14; P = 0.01). Regarding the exclusion of coronary artery disease, 22/30 (73%) patients with npTS underwent coronary angiography, one (3%) underwent cardiac magnetic resonance imaging, and one (3%) underwent thallium perfusion. Baseline characteristics between pTS and npTS patients did not present statistically significant differences. When comparing in-hospital outcomes between these groups, pTS patients presented higher inotrope use (71% vs 40%, P = 0.03), higher incidence of mechanical ventilation requirement (10% vs 0%, P = 0.01), and longer in-hospital stay (7 days vs 5 days, P = 0.045), but similar mortality.

Discussion

Over seven years and four months, 21 patients were diagnosed with pTS at our institution, which represents a relatively large population of patients with this uncommon condition. These presented with a low LVEF at diagnosis, which improved at discharge in most patients. Nevertheless, a large proportion of patients had severe complications and a non-negligible mortality rate was found. Patients with pTS were younger and presented with lower LVEFs at diagnosis than npTS patients did.

Our study describes one of the largest case series of pTS from a single institution; pTS represented 42% of all patients at our institution who had TS triggered by physical events, while in the study by Lee et al. pTS accounted for about 29% of hospital cases of physically triggered stress-induced cardiomyopathy.17 Our calculated incidence of pTS is about 1/15,000 surgical cases, which is somewhat lower than the 1/6,700 hypothesized by Hessel.1 As noted in other studies, the number of patients with pTS represented only a portion of all patients with TS at our institution, considering that all those patients with TS triggered by emotional events were not included in our analyses.13, 16 A diagnosis of pTS was more common in women (76%) as reported in other pTS series1, 17, 18 but was lower than noted in the International Takotsubo Registry.13 The patients in our study population were older than those in other published pTS series.1, 18 One possible reason for this is the large proportion of elderly patients receiving care at our institution.

Additionally, we found a high proportion (38%) of patients with cancer. Although there are several hypotheses regarding the mechanisms underlying the association between TS and cancer, the similarity of pathophysiologic mechanisms between malignancies and TS in terms of activation of the sympathetic system with an increase of catecholamine circulation is the most widely accepted one.19,20,21

Most of the procedures involved in the pTS cases were elective (62%), with wide variation regarding the cardiac risk classification. Consequently, although we cannot confirm that pTS occurs in relation to any specific type of surgery based on these findings alone, it is important to note that it can occur in association with a broad range of anesthetic/surgical procedures. The predominant anesthetic technique was general anesthesia, which is also consistent with the aforementioned pTS series.1, 18

The most frequently reported clinical presentation in our patients with pTS was heart failure, which is also in concordance with the study by Hessel et al.1 Although it is known that chest pain is the most frequent clinical presentation of any form of TS syndrome,13 it is important to consider that patients who experience pTS during the intraoperative or early postoperative period do not experience or are not physically able to report chest pain because of high-dose opioid administration, regional anesthesia, or unconsciousness. This highlights the importance of a high index of suspicion regarding this diagnosis and the role of echocardiography in the perioperative setting. Only 52% of patients presented with the typical apical ballooning pattern of regional wall motion abnormality; 10% exhibited the inverted/basal pattern. This is much higher than that found in npTS cases (typically reported as about 2%), but similar to that reported in other studies of pTS cases.1, 18Although coronary angiography is the gold standard tool to confirm TS,5 we have considered the International Takotsubo (InterTAK) Diagnostic Criteria (Table 4).10 As mentioned above, only 60% of patients with pTS underwent coronary angiography. Lee et al. had mentioned,17 in relation to pTS, that coronary angiography could sometimes be impractical, indicating the need for other diagnostic tools for TS in this specific context. As Awardal et al. suggest, patients who experience TS in the perioperative setting need a single set of criteria.24

Table 4 International Takotsubo Diagnostic Criteria (InterTAK Diagnostic Criteria)
Full size table

In accordance with the prior series,1, 17, 18 the LVEF at diagnosis, which is an important negative prognostic factor for this syndrome, was lower in our study than that reported in the International Registry,13, 25 but similar to the average initial LVEF of approximately 31% reported in two other reviews of perioperative cases.1, 18 Notably, a high proportion of patients required inotropic/vasopressor support; this, according to the most widely accepted physiopathological hypothesis describing an initial activation of beta receptors followed by the activation of a negative inotropic pathway in TS, could be a controversial treatment.26,27,28 Besides, there are some studies that suggest a deleterious clinical effect in patients with TS who receive catecholamines.29, 30 Only one patient in our series required mechanical ventricular support (extracorporeal membrane oxygenation) because of cardiogenic refractory shock. This option should be considered in the early treatment of TS.31 More than half of the patients presented with severe complications, similar to what had been previously reported.32 We found a large proportion of patients who experienced unexplained syncope and cardiac arrest; cardiac arrest was reported in a proportion as high as 6%. In the InterTAK registry, 5.9% of patients presented cardiac arrest, and a 60-day mortality rate of 40.5% was noted.33

Mortality was higher than that reported in some studies, probably because of the severity of patients' baseline illness, as presented in the results section.1, 17, 18 Furthermore, we found a similar mortality rate to that described by Brinjikji et al.34 Nevertheless, none of our patients died from cardiogenic causes.

Compared with patients with npTS, those with pTS were younger and had worse LVEF at diagnosis. Moreover, a higher proportion of patients with npTS underwent coronary angiography, cardiac magnetic resonance imaging or thallium scan to rule out coronary disease, probably enabled by a better clinical status.

Our study has several limitations. First, a notable proportion of patients did not undergo coronary angiography. Although these patients showed complete recovery of LVEF, they did not undergo the gold standard diagnostic tool for TS. Second, the reliability of our data collection and information may be a concern. Our institution provides physician access to a database of complete electronic records, and it is possible to obtain lists of patients with certain diagnoses over a specific time period through a searching tool. Hence, we have confidence regarding our data. Nevertheless, we acknowledge there could be inaccurate information regarding symptoms described by physicians, specific time of diagnosis, and treatments. In accordance with this, we can not confirm the time up to recovery of LVEF; we can only estimate the time until the first echocardiography that documented a normal LVEF (> 55%). All patients who had a diagnosis of TS in their in-hospital stay summary were included in our study. Nevertheless, we cannot assume there was no missing data for patients with incomplete clinical records. Third, our institution is a tertiary hospital with a large proportion of high-complexity procedures, although almost half of the patients diagnosed with pTS underwent surgeries with a low risk of cardiovascular events.

In conclusion, our study highlights that the occurrence of pTS, a complex syndrome that is triggered by physical stress, could be associated with variable levels of surgical/procedural stress. In addition, although most patients showed a recovery of LVEF, a non-negligible proportion of patients had complications, required intensive care, and experienced mortality.

For patients with probable pTS, anesthesiologists should be aware of the severity of TS in this context. Further investigations regarding specific diagnostic criteria for pTS are necessary.