A Nomogram for Predicting In-Hospital Mortality in Patients with Takotsubo Syndrome

Jun Chen The Third A liated Hospital of Zhejiang Chinese Medical University Yimin Wang The Third A liated Hospital of Zhejiang Chinese Medical University Xinyang Shou The Third A liated Hospital of Zhejiang Chinese Medical University Qiang Liu The Third A liated Hospital of Zhejiang Chinese Medical University Ziwei Mei (  lszxyymzw@163.com ) Lishui Central Hospital, the Fifth A liated Hospital of Wenzhou Medical College


Introduction
Takotsubo syndrome (TTS) is a clinical syndrome also viewed as "stress cardiomyopathy," "broken heart syndrome," or "transient apical ballooning" and is characterized by an acute and transient left ventricular dysfunction which always induced by physical or emotional stimulation 1 . Takotsubo syndrome is easily misdiagnosed as acute coronary syndrome (ACS), approximately 1%-3% of all patients who present with symptoms consistent with acute coronary syndrome and undergo coronary angiography have been identi ed to TTS 2 . Despite numerous studies reported that TTS was a reversible myocardial insu ciency and TTS patients always have a good prognosis during the hospitalization [3][4][5][6] . However, several serious complications may increase mortality to a rate, which is comparable with that of patients with ACS 7 . In the recent studies, the mortality rates have been reported higher than the previous studies 8 . Recently, one research based on the data from the SCAAR (Swedish Coronary Angiography and Angioplasty Registry) reported that the 30-day mortality rate in TTS was lower than ST-segment-elevation myocardial infarction (STEMI) but higher than non-STEMI (NSTEMI) 9 . Previous study on in-hospital outcomes of TTS patients discussed around risk factors abount mortality. Nevertheless, the predictive model of in-hospital mortality of TTS patients stay in ICU was not quite established. Therefore, assessment of the risk of in-hospital mortality in TTS patients by clinical predictive model can effectively provide a reference for subsequent hospitalization treatment and nursing. These measures would play an important role in improving the prognosis of patients with TTS.
At present, clinical prediction model is a good tool and can provide valuable guidance for clinical decision making. Nomogram as a special visual clinical prediction model that can calculates the risk scores for individuals, is convenient for clinicians to use in prognosis prediction. It has been widely used as a predictive method in the evaluation of prognosis in critically ill patients in recent years [10][11][12][13][14] . However, there was limited nomogram for predicting the hospitalization prognosis of patients with TTS. Therefore, we aimed to develop a nomogram for predicting the risk of in-hospital mortality in patients with TTS to identify the high risk of death patients early. This work would provide adequate reference for subsequent decision making, treatment and intensive care.

Data Source
This was a retrospective observational study, the primary data of our study was derived from MIMIC-IV database (version 1.0). MIMIC-IV database is an extensive database, and contained all medical record numbers corresponding to patients admitted to an intensive care unit (ICU) or the emergency department between 2008-2019 in the Beth Israel Deaconess Medical Center (BIDMC) 15 . The version 1.0 is the lasted version of MIMIC-IV database. One of our authors (C.J, certi cation ID: 8979131) gained permission to documented the database after online training at the National Institutes of Health (NIH). All methods were carried out in accordance with relevant guidelines to protect the privacy of patients.

Patients selection
We enrolled patients who aged more than 18 years old stay in ICU and diagnosed with TTS by the International Classi cation of Diseases version 9 ("42983") and version 10 diagnosis codes ("I5181") in the MIMIC-IV database. We excluded patients according to the following criteria: (I) No survival outcome data; (II) Being pregnancy and the postpartum condition; (III) Incomplete or unobtainable documented or other vital medical data records.
Clinical and laboratory data Extracted the basic information including age, height, weight and medical history such as diabetes, hypertension, chronic lung disease, myocardial infarction, heart failure, et al of TTS patients. Extracted vital signs information and the blood laboratory tests information of TTS patients as the rst document after patients admitted to the hospital within 24 hours. Vital signs data contained mean blood pressure (MBP), diastolic blood pressure (DBP), systolic blood pressure (SBP), body temperature (T), respiratory rate (RR), heart rate (HR), pulse oximetry derived oxygen saturation (spo2). Blood laboratory tests data consisted of hemoglobin, hematocrit, creatinine, anion gap, lactate, blood urea nitrogen (BUN), PH, white blood cell count, platelet count, chloride, glucose, prothrombin time (PT), serum potassium, serum sodium, and serum calcium. Therapies were also recorded, which contained the use of vasoactive drugs (norepinephrine), and continuous renal replacement therapy (CRRT) during the hospitalization. The sequential organ failure assessment (SOFA) score 16 of every patient also should be calculated. In our study, the endpoint was in-hospital mortality viewed as survival status at hospital discharge.

Statistical Analysis
The whole dataset was randomly divided into a training set and a test set with the proportion of 7:3. Mean±standard deviation (SD) documented normal distribution of continuous variables. Medians with upper and lower quartiles described unnormal distribution of continuous variables. Continuous variables were tested by T-test or Wilcoxon rank-sum test and categorical variables were analysed by chi-square test or Fisher's exact test for group comparisons. First, lasso regression was used to conduct preliminary screening of the predictors based on the whole study database, and screened out the predictors with large regression coe cients. Second, multivariate regression analysis was applyed to analyze the above screened predictors and established the nomogram in training dataset. The scores for each predictors were calculated based on coe cients of logistic regression variables in the model. We adopted receiver operating characteristic curve (ROC) to assess the discrimination of nomogram for in-hospital mortality of TTS patients. The area under the curve (AUC) of the ROC curve more than 0.7 was de ned as good discrimination. The degree of tting of the nomogram was assessed by calibration curve analysis which tested by Hosmer-Lemeshow test. The decision curve analysis (DCA) was conducted to evaluate the clinical utility of the nomogram through quantifying net bene ts against a range of threshold probabilities. The validation of nomogram capabilities were used by test set. These results were presented by odds ratio (OR) with 95% con dence intervals (CIs). All tests were two-tailed tests and p ≤ 0.05 was considered statistically signi cant. We used STATA 15.0 (Stata Corporation, College Station, Texas, USA) and R version 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria) for statistical analysis.

Results
The characteristics of study patients Totally, we included 368 eligible patients (85 males and 283 females) in our study, their average age was 66.29±16.09 years old, more information about missing data and data extraction process was shown in Supplementary Table S1 and Table S2. 48 patients (16 males and 32 females) died during the hospitalization, with the incidence of in-hospital death rate was 13.04%. Death group patients tend to older, and with higher value of anion gap, WBC count and SOFA score (As showed in Table1). The whole sample was randomly divided into a training dataset and a test dataset with the proportion of 7:3, and there were no signi cant difference in observed clinical variables between the training set and the validation set (As showed in Table 2).

Predict factors for TTS prognosis
We performed lasso regression analysis to make a preliminary selection of TTS prognosis predictors. (As showed in Figure 1). Nomogram for predicting the risk of in-hospital mortality in patients with TTS The nomogram included age, myocardial infarction history, gender, potassium, pt, WBC, hematocrit, anion gap, SOFA score as predictive factors for in-hospital mortality of TTS patients. For example, a 74-yearsold male without myocardial infarction history, his clinical data of admission as followed: PT: 30 seconds, potassium: 5 mmol/L, anion gap: 24 mEq/L, WBC: 10*10^9, hematocrit: 35% and SOFA score was 4 points. The corresponding score of each predictors were 6.5 points, 1.5 points, 2.8 points, 0.6 points, 0.7 points, 6.5 points, 1.5 points, 6.7 points, and 1.2 points respectively. Then his total score was about 28 points, and the risk of in-hospital mortality is 65%. (As showed in Figure 2).
Performance evaluation and validation of the nomogram ROC curve analysis for the training set showed that our risk prediction model has a good discrimination (AUC of ROC: 0.811, 95% Cl: 0.746-0.876, as showed in Figure 3(a)). We validated the prediction model in the test set, and the result showed that the present prediction model also has a good discrimination in test set (AUC of ROC: 0.793, 95% Cl: 0.724-0.862, as showed in Figure 3(b)). Furthermore, we performed calibration curve analysis to test the tting degree of nomogram. The calibration curve plot showed that predicted probabilities against observed death rates indicated good concordance (as showed in Figure  3(c)). In the test set, the calibration curve plot also showed good concordance between the predicted probabilities and the observed mortality (as showed in Figure 3(d)). In addition, we evaluate the clinical utility of the prediction model by decision curve analysis. DCA curve showed that our prediction model has a utility clinical practice (as showed in Figure 3(e) and Figure 3(f)).

Discussion
In the retrospective observational study, we found that the TTS patients have a hospital mortality rate of 13.04%. This rate of in-hospital mortality in our study was higher than the rate 2-5% in previous studies [3][4][5][6] , this was due to our study only included the TTS patients admitted to ICU, these patients tend to have more serious and complicated illness condition than the TTS patients in previous studies reported. Our study demonstrated that age, myocardial infarction history, PT, WBC, hematocrit, anion gap, SOFA score were important predictive factors for in-hospital mortality of severe TTS patients. Despite gender (male) and serum potassium concentration in our study without statistical difference (as shown in Table 3), several previous studies have identi ed that gender (male) have a higher in-hospital mortality compared with female [17][18] , thus, we also included gender as a predictor in the nomogram. In addition, multiple studies have reported that serum potassium concentration was associated with poor prognosis of acute heart failure patients [19][20] . Considering that almost all the TTS patients have moderate to severe cardiac insu ciency, thus, we included serum potassium in the nal nomogram. The nomogram based on above predictors showed a good predictive ability. Gender is closely related to the onset and prognosis of TTS patients. Male with TTS were reported to have a higher in-hospital mortality compared with female 17,18,21 and were viewed as an independent risk factor for in-hospital death in TTS patients in previous study 22 . In addition, male were related to adverse composite events consist of cardiovascular death, severe pump failure, and fatal ventricular arrhythmias (odds ratio 4.32, 95% CI 1.41-13.6) reported by a multicenter registry study of TTS 23 . The reason for male gender is more likely to be in-hospital death probably is that men are more possible to exit comorbidities such as COPD, coronary artery disease, a higher peak troponin-I level and to burden the severity of myocardial injury than women. Moreover, men were more likely to have a physical stressor compared with women, TTS patients induced by physicology were more likely to have malignancy, lower hemoglobin, higher serum creatinine and died in the hospital than unphysicology 24 .
For women, their organism release estrogen which could mitigates myocardial damage. The protective action for myocardial injury performed by estrogen was demonstrated by several studies in animal experimental models 25 . In our study, 23.1% of our TTS patients were male and the in-hospital death rate was 37.65% (32/85) in males. On the contrary, the in-hospital death rate was 5.65% (16/283) in females.
It seems that males have higher in-hospital mortality compared with female (37.65% vs 5.65%, P<0.001).
This result was same as the study performed by Yoshihiro Sobue et al 22 . However, to our surprise, gender was not found to be a risk factor for in-hospital death in either univariate or multivariate regression analyses. Recently, Budnik M et al study found that inhospital outcomes (arrhythmias, mechanical complications, cardiogenic shock, mortality rate) were similar in both male and female groups after adjusted other confounding factors 26 . These results suggested that gender contribute for the in-hospital death of TTS should be further research.
The sequential organ failure assessment (SOFA) score was applied to describe the time course of multiple organ dysfunction. Recent studies showed that SOFA score was associated with survival in severe cases 27 . Previous study revealed that increase SOFA during the rst few days of ICU admission is a good indicator of prognosis. Both the mean and highest SOFA scores are particularly useful predictors of outcome 28 . In our study, the SOFA score of in-hospital death patients was statistically higher than survival patients (7.27 vs 5.78, P=0.016). Multivariate logistic regression analysis demonstrates that sofa score was an independent predictive factor for in-hospital mortality in TTS patients. Therefore, SOFA score could be a good predictor for prognosis of TTS patients in intensive care units.
We included myocardial infarction history as predictor in the nomogram and found for the rst time These parameters have important clinical value, and future studies can add cardiac hyperparameters on the basis of our study to increase the prediction e ciency of the prediction model.

Study limitation
Several limitations must be acknowledged. First, we aimed to establish a rapid and simple nomogram for predicting in-hospital mortality in TTS patients. The situation of TTS patients was very complex and there were many factors related to the prognosis of their hospitalization. We only included the clinical data of the patients within 24 hours after admission, and did not consider the intervention measures at the onset or after hospitalization, which has certain limitations; Second, our study did not further analyze the causes of TTS, such as physical triggers or psychological triggers. Future studies can conduct a subgroup analysis of speci c causes of cardiac arrest for provide more evidence for this term.

Conclusions
We have established a risk prediction model by using admission characteristics of TTS patients could help identify in-hospital mortality in TTS patients.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. TableS1andTableS2.docx