QT and Tpeak‐Tend interval variability: Predictive electrical markers of hospital stay length and mortality in acute decompensated heart failure. Preliminary data

Abstract Background As previously reported, an increased repolarization temporal imbalance induces a higher risk of total/cardiovascular mortality. Hypothesis The aim of this study was to assess if the electrocardiographic short period markers of repolarization temporal dispersion could be predictive of the hospital stay length and mortality in patients with acutely decompensated chronic heart failure (CHF). Method Mean, standard deviation (SD), and normalized variance (VN) of QT (QT) and Tpeak‐Tend (Te) were obtained on 5‐min ECG recording in 139 patients hospitalized for acutely decompensated CHF, subgrouping the patients for hospital length of stay (LoS): less or equal 1 week (≤1 W) and those with more than 1 week (>1 W). Results We observed an increase of short‐period repolarization variables (TeSD and TeVN, p < .05), a decrease of blood pressure (p < .05), lower ejection fraction (p < .05), and higher plasma level of biomarkers (NT‐proBNP, p < .001; Troponin, p < .05) in >1 W LoS subjects. 30‐day deceased subjects reported significantly higher levels of QTSD (p < .05), Te mean (p < .001), TeSD (p < .05), QTVN (p < .05) in comparison to the survivors. Multivariable Cox regression analysis reported that TeVN was a risk factor for longer hospital stay (hazard ratio: 1.04, 95% confidence limit: 1.01–1.08, p < .05); whereas, a longer Te mean was associated with higher mortality risk (hazard ratio: 1.02, 95% confidence limit: 1.01–1.03, p < .05). Conclusion A longer hospital stay is considered a clinical surrogate of CHF severity, we confirmed this finding. Therefore, these electrical and simple parameters could be used as noninvasive, transmissible, inexpensive markers of CHF severity and mortality.


| INTRODUCTION
The prevalence of chronic heart failure (CHF) is approximately 10% among the over 70 years-old elderly 1 and there are 600 000 and 500 000 new cases per year, respectively, in Europe and in the United States. 2 Therefore, this complex, progressive clinical syndrome frequently causes hospital readmission, high levels of mortality rates, and a huge amount of healthcare costs. Particularly, it was previously observed that the longer the hospital stay, the poorer the prognosis of the patients was. [3][4][5][6] In some previous studies, in acutely decompensated CHF patients, some noninvasive electrical markers, based on short-term repolarization temporal dispersion (such as QT, QTe, Te variance normalized [VN]) were found capable of specifically selecting high risk cardiovascular and all-cause mortality patients, confirming a strict relationship with some biomarkers of decompensated CHF. 7,8 Predicting the length of hospitalization and the possibility to benefit from the intensive care and specific therapies during the acute phase of his illness could be highly helpful for health-care systems, to concentrate their energies in terms of personnel and resources. Consequently, the duration and variability of QT and T peak-end (Te, expressed as standard deviation, SD) intervals as markers of mortality and severity of CHF, expressed as hospital length of stay 3,9 were analyzed.

| Patients and protocol
Consecutive in-patients were enrolled when admitted to the Department of Geriatric Medicine, from January 2019 to September 2020, due to decompensated CHF in NYHA functional class IV at the enrollment. Patients were defined as decompensated following the European Society of Cardiology guidelines. 1

| Offline data analysis
Five minutes ECG (Miocardio Event™) II lead signals were acquired, digitalized at a sampling frequency of 500 Hz, and wirelessly transmitted to a cloud platform for data storage via mobile phone 7,8 Subsequently, all digitized signal recordings were downloaded by the cloud and checked by a single physician (G. P.) blinded to subjects' circumstances. Therefore, we measured the following intervals from the respective time series in ECG recordings: RR, QT, and Te; QT was obtained by measuring the interval from the onset of the Q-wave to the T-wave end; Te was obtained from the T peak to the end of T wave. To identify the repolarization intervals, we used software originally proposed by Berger 10 and validated in other studies. [11][12][13][14] We, therefore, calculated mean, variance, and SD (QTSD and TeSD) values for each of these repolarization phase intervals and, finally, we Software for data analysis was designed and produced by our research group with the LabView program (National Instruments).

| Statistical analysis
All variables with normal distribution were expressed as means ± SD whereas non-normally distributed variables were as median and interquartile range (i.r.). Categorical variables were analyzed with the χ 2 test. Unpaired Student's t-test was used to compare data for the normally distributed variables; on the contrary, Mann-Whitney tests were used to compare non-normally distributed variables (as evaluated by the Kolmogorov-Smirnov test). Uni-and multivariable forward (A. Wald) Cox proportion-hazard regression analysis was used to determine the association between repolarization variables (covariates) and some dichotomized variables (dependent variables).

| DISCUSSION
The major new finding of the present work was that some simple, noninvasive, inexpensive, easily repeatable electrical markers, based on short-term repolarization variability, were associated with the inhospital LoS and mortality. In particular, TeVN and Te mean were respectively predictive of the LoS and 30-day mortality, when obtained at hospital admission. In-hospital LoS was considered a marker of severity of decompensated CHF. 3,9,18 Although, in fact, the mortality prevalence in our study was similar among the two studied groups, whereas, in the >1 W LoS group the major part of low ejection fraction and impaired blood pressure, longer repolarization temporal dispersion, higher level of CHF biomarkers (NT-proBNP and troponin) cases were found. In other words, clinical heart failure severity was connected to the in-hospital LoS.
The pathophysiological basis of Te segment prolongation is controversial; while Te temporal dispersion (TeVN) is still little studied, some authors believe that this interval represents a noninvasive marker of transmural dispersion of repolarization in the left ventricle, 19 but other authors oppose this interpretation without offering alternative explanations. 20 If the pathophysiologic basis of the Te is not clear, the clinical use of this marker was not disputable.
In fact, a recent meta-analysis by Tse and coll, 21  result is a prolonged temporal dispersion of QT and Te. 14-17 Recent observations have highlighted that sympathetic hyperactivity is able to induce an increase in QTe variability index, QTeVN, TeVN, TeSD, in normal 17 and pathological 23,24 conditions. Thus, the morphological and structural alterations of the myocardium resulting in specific modifications of these electrocardiographic markers. This allows for the patient's telemonitoring over time, as when the clinical and morpho-structural improvement occurs (no more swelling nor alteration of the functionality of the ion channels), a relative improvement of the electrocardiographic parameters should be observed.
In conclusion, patients with acute decompensated CHF and ≥1 W length of in-hospital stay showed elevated short-term repolarization markers. In particular, these patients had higher TeVN, TeSD, and higher levels of CHF biomarkers.
Finally, in the subgroup of re-evaluated, survived patients, the substantial stability of both QT mean and Te mean from admission to discharge suggests, once again, the predictive capability of these ECG markers. As shown in Table 3, Te mean showed a noticeable difference in the hospital admission between survived and deceased patients; therefore, Te mean tends to be able to identify patients with worse prognoses.
Obviously, the present study supplies a working hypothesis, that needed to be confirmed in larger and case-control studies.
Nevertheless, it could be hypothesized that these markers could be used as a noninvasive, transmissible, inexpensive, electric, surrogate marker of hospital length of stay, of CHF clinical severity, suggesting higher mortality risk.

| LIMITATIONS
The main limitation of this study is the small sample size. This aspect can be overcome by continuing the enrollment and the involvement of more centers. With an adequate sample size, it will be possible to verify the predictive power of the variables under examination even in subjects with less severe degrees of decompensated CHF.
In addition, an actual limitation of the study is the absence of patients treated with SGLT2 inhibitors. The sample was in fact studied before the recent indications provided by the European Society of Cardiology guidelines on the use in class I evidence A of these drugs in subjects with heart failure. 25 The use of these drugs could in fact have some impact on the clinical presentation and duration of hospitalizations. It will be the authors' responsibility to deepen this data in a further study.
Furthermore, the geographic variability, already described in the past, 26 in the duration of hospitalization and rehospitalization of patients suffering from heart failure is a relevant datum. In this study, carried out in a single center, it is not possible to highlight discrepancies from this point of view. The involvement of further centers could ensure a better definition of the geographical impact.