*2.2. Endpoints*

The primary endpoint was the recurrence of TCM, defined as a new episode of acute HF with reduced ejection fraction along with the evidence of atrial or ventricular arrhythmia, occurring after complete clinical (no HF symptoms) and echocardiographic (LVEF ≥ 50%) recovery of the original episode.

The secondary endpoints were: Death from all causes, major adverse cardiovascular events (defined as non-fatal stroke, non-fatal myocardial infarction and cardiovascular death), and cardiovascular hospitalizations. Cardiovascular hospitalizations were defined as any hospitalization longer than 12 h for one or more of the following reasons: Acute coronary syndrome, unstable angina, HF, atrial or ventricular arrhythmia, valvular heart disease, infective endocarditis, myocarditis, pericarditis, aortic disease, pulmonary embolism, stroke/transient ischemic attack, syncope, cardiovascular-related elective and urgent procedures and complications of such procedures.

#### *2.3. Data Collection*

Two expert physicians were responsible for the prospective data collecting regarding the patients' demographics, risk factors, medical history and treatment. Continuous 12-lead ECG monitoring (Mortara Rangoni, Arezzo, Italy) was used to assess the heart rate during hospitalization and underlying arrhythmias. The blood samples for brain natriuretic peptide (BNP) and troponin I were collected on admission and at discharge. Echocardiographic examinations were performed with a monoplane ultrasound probe 4 MHz (M4S) of Vivid 7 Pro (GE Medical Systems, Milwaukee). The digital loops were captured, recording at least three consecutive beats, and analysed off-line using a dedicated software

(EchoPAC 13.0; GE Medical Systems, Milwaukee) according to the most recent recommendations. A complete echocardiogram was performed on admission, at discharge, at a 3-month follow-up. Serial echocardiograms were then performed at least every 6 months until complete recovery of LVEF. The echocardiographic loops were obtained with the patient supine and in the left lateral decubitus at the end of a normal breath, minimizing the depth in order to optimize the frame rate (40–80 fps).s LVEF was calculated by the Simpson biplane method. All echo exams were reviewed by two authors (G.S. and F.G.), who were responsible for the off-line analysis and collected all measurements blinded to the recurrence or other clinical endpoints. The inter-operator coefficient of variations for LVEF was 3.2% and the intra-operator coefficient of variation was 2.4%.

To allow for the comparability of drug regimens across the patients taking many different medications, a treatment intensity score (TIS) was calculated. As previously reported, [15] the recorded daily dose taken by the patient was divided by the maximum recommended daily dose to obtain a proportional dose for that medication, called intensity. The maximum recommended daily doses were set by the European and American guidelines [9–12].

#### *2.4. Statistical Analysis*

All continuous variables were checked for normality through the Kolmogorov-Smirnov test. The normally-distributed variables were described by the mean and standard deviation and compared by analysis of variance. The not-normally-distributed variables were described as the median and 1st–3rd IQR and compared by non-parametric tests. The categorical variables were described as the absolute and relative values, and compared by chi-square test or Fisher exact test, as appropriate.

The Kaplan-Meier analysis was used in order to assess the time free from primary and secondary endpoints. The association between the individual variables and the risk of TCM recurrence and cardiovascular hospitalization was investigated by using univariate Cox proportional hazards models. The variables that showed an association with each endpoint with a significance level < 0.1 on univariate analyses were entered into the multivariable Cox proportional hazards model. The independent risk factors for each endpoint were then presented as hazard ratios (HRs) and 95% confidence intervals (CIs).

The linearity assumption of the relationship between the independent continuous risk factors and the outcome of interest was represented using restricted cubic splines with three knots located to the 10th, 50th, and 90th percentiles according to the Harrell rule, and assessed by the Wald test for linearity.

SPSS 22.0 for Windows (SPSS Inc., Chicago, IL, USA) and R (R Foundation for Statistical Computing, Vienna, Austria) were used for statistical analysis. The values of *p* < 0.05 (two-tailed) were considered as statistically significant.

#### **3. Results**

The population included 107 patients (68 males, mean age 66.7 ± 14.5 years). The patients' characteristics are summarized in Table 1. The median follow-up was 22.6 months (1st–3rd quartile 10.0–40.0 months). The median hospitalization time (i.e., the first phase of the selection process) was 7 days (1st–3rd quartile 4–11 days). The median time to TCM diagnosis confirmation (i.e., the second phase of the selection process) was 72 days (1st–3rd quartile 48–130 days). Eighty-three patients (77.6%) were diagnosed with atrial fibrillation (AF) as underlying arrhythmia, and 16 (15.0%) with atrial flutter. Other triggering arrhythmias included non-sustained ventricular tachycardia (4, 3.7%), paroxysmal supraventricular tachycardia (1, 0.9%) and premature ventricular contractions (PVCs) (3, 2.8%).


*J. Clin. Med.* **2019**, *8*, 1411

AF: atrial fibrillation; BMI: body mass index; BNP: brain natriuretic peptide; CKD: chronic kidney disease; COPD: chronic obstructive pulmonary disease; CV:

indexed left atrial volume; LVEF: left ventricular ejection fraction; NYHA: New York Heart Association; OSAS: obstructive sleep apnoea.

cardiovascular;

 iLAV:

During the follow-up, 17 patients experienced at least one recurrence (15.8% of all patients) and 51 were hospitalized for cardiovascular reasons (47.7%). Among the major adverse cardiovascular events, two patients suffered from thromboembolic events (1.8%) and one died from cardiovascular causes (0.9%). No non-fatal myocardial infarctions were reported. The annual incidence of recurrence was 8.4% per year, 0.9% per year for thromboembolic events and 0.4% per year for cardiovascular mortality.

The only death occurred in a 58-year old male, one year and three months after recovery of both NYHA class and LVEF. The patient had no evidence of progression to any kind of structural heart disease and died suddenly after an out-of-hospital cardiac arrest due to idiopathic ventricular fibrillation. One transient ischemic attack and one non-fatal stroke occurred in two different patients after 3 and 832 days, respectively. The treatment strategies at discharge are described in Table 2.


**Table 2.** Treatment strategies at discharge.

\* The mean therapeutic index was calculated only in those patients who were administered the drug at least until discharge. \*\* For loop diuretics we considered the total dose per day as equivalents of furosemide. ACE-I: angiotensin converting enzyme inhibitor; ARB: angiotensin II receptor blocker; CCB: calcium-channel blocker; MRA: mineralocorticoid receptor antagonist; WCD: wearable cardioverter-defibrillator.
