Time-dependent S-wave areas predict sudden cardiac death risk: a prospective, multicentre registered study

Time-dependent S-wave areas predict sudden cardiac death risk: a prospective, multicentre registered study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Time-dependent S-wave areas predict sudden cardiac death risk: a prospective, multicentre registered study Xinfeng Wang, Wenhao Liu, Ziheng Zheng, Wenli Gu, Zhiteng Chen, and 16 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3490411/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Objectives The present study aimed to detect novel and time-dependent ECG parameters by analysing 24-h ECG data, especially the area under ECG waves. Design: Prospective, multicentre cohort study Setting: Four hospitals in China Participants: High risk of sudden cardiac death, including 43 survivors of sudden cardiac death (SCD) or patients who suffered haemodynamic disorder due to sustained ventricular tachycardia/ventricular fibrillation (SCDHR group), 138 patients with HF who did not experience sustained ventricular tachycardia/ventricular fibrillation but were diagnosed with dilated cardiomyopathy or ischaemic cardiomyopathy with LVEF ≤ 35% (HF group), and 108 healthy controls who presented with no heart disease (HC group). Exposure: Time-dependent ECG parameters by analysing 24-h ECG data Main outcome measures: The area under ECG waves was separately analysed to determine their associations with SCDHR and HF in the test set and was further examined in the validation set. Logistic regression analyses were performed. Results The multivariate logistic regression model for discriminating SCDHR patients and HCs indicated that the average area under the S-wave (inteS_mean) at 16:00–21:00 was positively associated with SCDHR (OR > 1, P-adjust < 0.050) and significantly (P value = 0.014) differed at 21:39 in the validation set. Similarly, the model for discriminating HF and HC indicated that the inteS_mean, minimum S-wave area (inteSm), and difference in S-wave and T-wave (inteST) were positively (OR > 1, P-adjust < 0.050) associated with HF in both the test set and validation set. Conclusions The time-dependent S-wave area-related ECG parameters (inteS_mean, inteSm, and inteST) are potentially early predictive factors for SCD risk. Trial registration: This study was registered on the website of http://register.clinicaltrails.gov/Organization . The Clinical Trials ID is NCT03485079. Biological sciences/Biological techniques/Electrophysiology/Electrocardiography – EKG Health sciences/Cardiology/Cardiovascular biology/Cardiovascular diseases/Arrhythmias Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Sudden cardiac arrest (SCA) is a major cause of mortality in the Western world, causing > 300,000 deaths annually in the USA 1 and accounting for 20% of SCD in Western societies. 2 According to the 2011 American College of Cardiology Foundation/American Heart Association (ACCF/AHA) 3 and 2003 American College of Cardiology/European Society of Cardiology (ACC/ESC) guidelines, 4 malignant arrhythmia is a critical characteristic of patients with a high risk of SCD (SDCHR). However, readily available methods to identify these individuals are lacking. 5 As SCD is primarily a result of electrical disturbance of the normal cardiac rhythm, 6 the widely available and low-cost ECG is a potentially attractive and noninvasive tool for SCD risk stratification. By using ECG data, most studies have attempted to detect the abnormal characteristics associated with SCD risk by comparing cases with controls. 7–9 The ECG parameters, including J-wave amplitude, 10 QRS prolongation, QT prolongation, T-wave alternans, and QRS fractionation, 11 have been well documented. To date, however, no individual ECG finding has adequately stratified patients with a risk for SCD. 12 Although recent studies have indicated that the combination of a few known ECG parameters result in improved SCD risk prediction, 6,13 the identification of optimal ECG risk markers remains a challenge. 14 Moreover, most of the studies are based on previously known ECG markers, 15–17 which lack extraction and use related data, such as the positioning of S and T peaks as well as peak areas, to make judgements. Additionally, patients with dilated cardiomyopathy or ischaemic cardiomyopathy are considered to have a high risk of SCD. 18 A previous study has indicated that approximately 30% of patients with dilated cardiomyopathy or ischaemic cardiomyopathy are at risk of developing SCD. 19 No studies have explored the differences in ECG parameters among patients with SCDHR or with a history of SCD. Another unknown of SCD is the episode timing. The Framingham cohort study first reported a morning peak of SCA. 20 A subsequent study, reporting approximately 1,000 events, has demonstrated that SCD episodes significantly increase 6:00 AM to 12:00 PM. 21 This result was later verified in prospective community studies of SCA and cohorts of patients implanted with ICDs. 22–26 Recently, a study involving 4,126 SCD cases has revealed that women are more likely to develop SCD at night. 27 Nevertheless, there is still a lack of dynamic risk assessment over time of the association between SCD and ECG markers. Thus, the time intervals in risk assessment that reliably predict individual risk for SCD remain unknown. In the present study, to identify time-dependent ECG parameters associated with SCD risk, we collected 24-hour ECG data, including patients who were survivors of SCD or with haemodynamic disorder due to sustained ventricular tachycardia/ventricular fibrillation (SCDHR), patients with heart failure (HF), and healthy controls (HCs). The ECG data of each sample were divided into one-minute-long fragments and were compared between SCDHR patients and HCs, between SCDHR patients and HF patients, and between HF patients and HCs. Multivariate logistic regression models were then built to reveal the time-dependent ECG parameters associated with SCDHR in HC, SCDHR in HF, and HF in HC. Finally, these ECG parameters were validated in an independent dataset. Methods Participants This study was a prospective, multicentre cohort study. We enrolled survivors of SCD or patients who suffered haemodynamic disorder due to sustained ventricular tachycardia/ventricular fibrillation (SCDHR group), patients with HF who did not experience sustained ventricular tachycardia/ventricular fibrillation but were diagnosed with dilated cardiomyopathy or ischaemic cardiomyopathy with LVEF ≤ 35% (HF group), and healthy controls who presented with no heart disease (HC group). The exclusion criteria of the participants were as follows: ( 1 ) had a high degree of atrioventricular block (≥ II) at resting heart rates; ( 2 ) were in NYHA IV at inclusion; ( 3 ) had an MI history within one month; ( 4 ) had undergone coronary revascularization within the preceding three months; ( 5 ) had advanced cerebrovascular or renal disease and suffered from any noncardiac condition with a high likelihood of death during the trial; ( 6 ) had a life expectancy of one year or less; and ( 7 ) had a history of valvular heart disease. The 24-hour ECG data regarding SCDHR, HF, and HC were collected in four hospitals (Sun Yat-sen Memorial Hospital, Ruijin Hospital of Shanghai Jiao Tong University, Drum Tower Affiliated Hospital of Nanjing University Medical School, and Zhongshan Hospital of Fudan University) in China. The numbers of patients with SCDHR and HF collected by each hospital are shown in Supplemental Table 1. This study was approved by the Institutional Review Boards of all hospitals mentioned above and received proper ethical oversight. Signed informed consent was provided by all participants. All patients and controls in this study had archived resting v3-e ECG data with a paper speed of 25 mm/s and calibration of 10 mm/mV. This study was registered on the website http://register.clinicaltrails.gov/Organization . The Clinical Trials ID is NCT03485079. Data filtering The ECG data of each sample were divided into one-minute-long time fragments. To filter the noise, each fragment was further divided into 60 pieces. The filtering principles in every second were as follows: ( 1 ) the maximum value in the one-second fragment was lower than 0.2 mV, ( 2 ) the maximum value was higher than 5 mV, or ( 3 ) the difference between the maximum and minimum value was greater than twice the median difference in this minute. If the data length was less than 30 seconds in a one-minute fragment, the fragment was discarded. Samples with ECG signal lengths less than 100 minutes were excluded. Calculation of ECG parameters In each one-minute-long time fragment, 42 parameters were calculated, including the following parameters: maximum and minimum peak values of waves (R-wave, T-wave, and S-wave); the maximum, minimum and average area under waves; the summation of peak values of waves; and the summation of the area under waves. Details of the ECG parameters are shown in Supplemental Table 2. Statistics Comparison of ECG parameters across samples The ECG parameters were first compared across the SCDHR, HF and HC groups. Differences between two groups were tested by Student’s t test. A P value less 0.0012 (0.05/42) was considered statistically significant. Statistical analysis to identify ECG parameters associated with SCDHR or HF Multivariate logistic regressions were used to detect the ECG parameters associated with SCDHR or HF. To construct the models, we randomly selected 70% of the samples collected from Sun Yat-sen Memorial Hospital as the training set. The remaining 30% of the samples from Sun Yat-sen Memorial Hospital and samples from other hospitals were used as an independent test dataset. To exclude redundant features and reduce multicollinearity, we used K-means 28 to cluster the 42 ECG parameters in the training set. The most fitted parameter in each cluster was kept as an independent parameter. The most fitted ECG parameter was the parameter indicated by the lowest Akaike information criterion in the univariable logistic regression model testing the association between SCDHR patients and HCs, between SCDHR patients and HF patients, or between HF patients and HCs. The features selected from all clusters were used to build a multivariate logistic regression model. We used an auto backward approach to select the features most likely associated with the SCDHR or HF. The performance of the multivariate logistic regression model was evaluated by the average area under the curve (AUC) of 5-fold cross-validation. Briefly, from the training set, we randomly selected 4/5 of the data for training, and the remaining 1/5 of the data was used for testing. This process was repeated five times, and the AUC of each fragment was calculated to detect the parameters most correlated with SCDHR or HF. To select time-dependent ECG parameters associated with SCDHR or HF, we constructed a multivariate logistic regression model that connected the models trained in different time fragments into one model by a greedy searching approach in the training set. First, the models of different time fragments were ranked according to the fivefold average AUC values. Then, the best-performing model was connected to the second-best model to construct the combined model. The connected model was evaluated by the average AUC of the fivefold cross-validation. This process was continued until all models of the 1440 time fragments were combined. The model yielding the best AUC and integrating more than two fragments was used for further analysis and finally tested in the independent test set. Measurements of the fitness of the regression models To evaluate the ECG parameters associated with SCDHR or HF, the fitness of the regression models was measured by AUC. The workflow is shown in Fig. 1 . Results Demographic and clinical characteristics The demographic and clinical data of patients with SCDHR, HF, and HC are shown in Table 1 . In total, 35.71% of patients with SCDHR, 28.57% of patients with HF, and 9.68% of HCs had a smoking history. Among the samples, 12.50% of patients with SCDHR and 19.85% of patients with HF had diabetes, whereas none of the HCs reported diabetes. Regarding other comorbidities, 39.02% of patients with SCDHR, 35.88% of patients with HF, and 9.02% of HCs had hypertension, and 9.30% of patients with SCDHR, 20.66% of patients with HF, and 0.81% of HCs had coronary artery disease. Table 1 The demographic, disease history and clinical data among patients with SCDHR, HF and HC. Clinical feature HC(n = 108) HF(n = 138) SCDHR(n = 43) Age (y) 46.07 ± 10.93 56.09 ± 13.78 54.88 ± 15 Male (%) 43.20 69.63 81.40 Diabetes (%) 0.00 19.85 12.50 Hypertension (%) 9.02 35.88 39.02 Coronary artery disease history (%) 0.81 20.66 9.30 Hyperlipidemia (%) 1.61 4.51 4.76 Smoke history (%) 9.68 28.57 35.71 Drink (%) 4.03 14.29 19.05 Syncope (%) 0.81 0.75 29.27 Glucose (mmol/L) 4.97 ± 0.53 4.96 ± 0.83 4.76 ± 0.77 Uric acid (µmol/L) 373.12 ± 99.93 459.59 ± 144.43 430.51 ± 131.97 Total cholesterol (mmol/L) 5.18 ± 0.94 4.23 ± 1.04 4.29 ± 0.99 Triglyceride (mmol/L) 1.19 ± 0.48 1.22 ± 0.51 1.18 ± 0.48 HDL-C (mmol/L) 1.30 ± 0.26 1.08 ± 0.31 1.13 ± 0.31 LDL-C (mmol/L) 3.28 ± 0.70 2.62 ± 0.84 2.75 ± 0.8 apoA1 (mmol/L) 1.39 ± 0.19 1.11 ± 0.24 1.14 ± 0.25 apoB (mmol/L) 0.91 ± 0.20 0.81 ± 0.20 0.80 ± 0.22 CREA (µmol/L) 75.38 ± 13.77 92.08 ± 18.23 89.39 ± 17.16 Cystatin (mg/L) 0.82 ± 0.18 1.14 ± 0.26 1.02 ± 0.21 WBC (×10 9 /L) 5.99 ± 1.44 6.73 ± 1.95 7.66 ± 2.02 HGB (g/L) 137.63 ± 15.95 138.39 ± 18.30 138.25 ± 19.96 PLT (×10 9 /L) 248.68 ± 51.58 196.52 ± 57.56 223.24 ± 59.52 Neutrophil (%) 0.57 ± 0.10 4.53 ± 5.16 2.90 ± 4.52 NT-pro BNP (pg/ml) 46.85 ± 51.14 1774.78 ± 1311.25 837.00 ± 1127.69 LA (mm) 31.27 ± 3.41 43.89 ± 6.68 38.11 ± 7.94 IVSD (mm) 8.41 ± 1.23 9.18 ± 1.49 9.08 ± 1.48 LVDd (mm) 46.19 ± 3.67 63.75 ± 12.47 57.58 ± 10.75 LVPWd (mm) 8.28 ± 1.09 9.18 ± 1.37 8.97 ± 1.28 EF (%) 68 ± 4.3 36 ± 15.6 47 ± 17 HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; WBC, white blood cell; HGB, hemoglobin; LA, left atrium; IVSD, interventricular septal dimension; LVDd, left ventricular diastolic dimension; LVPWd, left ventricular posterior wall dimension; EF, ejection fraction. Analysis of individual ECG parameters among SCDHR, HF, and HC After excluding data with noise, the number of patients included in this study is shown in Table 2 . Differences in ECG parameters among SCDHR, HF, and HC were examined in each of the 1440 fragments. Table 2 Number of patients with SCDHR, HF and HC involved in this study after filtering noise. Sample group Hospital Groups SCDHR HF HC Training set Sun Yat-sen Memorial hospital 25 57 75 Independent test set Sun Yat-sen Memorial hospital 12 25 33 Zhongshan Hospital of Fudan University 1 8 Drum Tower Hospital, Nanjing University Medical School 1 5 Ruijin Hospital, Shanghai Jiao Tong University 4 43 When comparing the ECG parameters of patients with SCDHR and HCs in the training set, we found 20 ECG parameters that significantly differed (P value < 0.0012) in at least one time fragment (Fig. 2 D), which were concentrated from 17:00–22:00. Among them, four ECG parameters (“heart_rate”, “inteRS”, “t_T”, and “inteT_sum”) significantly differed between patients with SCDHR and HCs in more than 10 fragments, including two parameters (“t_T” and “inteT_sum”) related to T-waves within the time intervals ranging from 8:00–12:00. Supplemental Fig. 2 shows the P value distributions of these four features in the 1440 time segments. Comparison of the ECG parameters of SCDHR and HF indicated that a few time fragments contained significantly different ECG parameters between SCD and SCDHR (Fig. 2 B) in the training set, including 14 ECG parameters that significantly differed (P value < 0.0012) in at least one time fragment (Fig. 2 E). Among them, four ECG features, namely, "maxS", "inteRST", "inteST", and "t_S_sum", significantly differed in more than ten fragments. All of these features are related to the S-wave. Supplemental Fig. 3 shows the P value distributions of these four features in the 1440 time segments. The results shown in Fig. 2 E and Supplemental Fig. 3 suggest that S-wave-related ECG parameters are important in discriminating SCD from SCDHR. Compared to the ECG data collected from Sun Yat-sen Memorial Hospital, more time fragments were observed to contain ECG parameters that significantly differed between SCDHR and HF (Supplemental Fig. 1B) in the independent test set. The “heart_rate” was validated to play an important role in discriminating SCDHR and HF (Supplemental Fig. 1E). Analysis of the differences in ECG parameters between HF and HC demonstrated that significantly different signals were distributed at 21:00–5:00 (Fig. 2 C), and the same phenomenon was found in the independent test set (Supplemental Fig. 1C). Importantly, ten ECG parameters ("inteST", "inteSm", "inteS_mean", "mean_S", "inteS_sum", "maxS", "minS", "inteRST", "inteRS", and "t_RT") significantly differed in more than 20 fragments (Fig. 2 F). Interestingly, nine of these parameters are related to S-waves. Supplemental Fig. 4 shows the P value distributions of these ten parameters in the 1440 time segments, which indicated that those related to the area under the S-wave (inteST, inteSm, inteS_mean, and inteS_sum) significantly differed in almost all fragments from 20:00 to 6:00 on the next day. Multivariate logistic regression of ECG parameters To explore the association of ECG parameters with SCDHR and HF, multivariate logistic regression models were constructed. We first excluded the redundant ECG parameters by clustering them and selecting one parameter from each cluster. Briefly, the ECG parameters in each time segment were clustered into eight clusters (Supplemental Fig. 5). From each cluster, one ECG parameter showing the most significant association with outcome was selected to represent the cluster. The features from each cluster were used to perform a multivariate logistic regression analysis. ECG parameters associated with SCDHR patients and HCs The multivariate logistic regression model was trained at each time fragment in the training set to discriminate patients with SCDHR and HCs. To identify time-dependent ECG parameters associated with SCDHR, we combined the models into a unique regression model by the greedy feature selection approach and found that the combination of ten time fragments (16:06, 14:37, 21:39, 17:32, 17:28, 17:15, 14:18, 19:23, 13:50, and 9:31) and six ECG parameters ("inteS_mean", "J_up", "t_T", "mean_R", "T_minus", and "t_RT") yielded the best performance (AUC = 0.887) to discriminate patients with SCDHR and HCs (Supplemental Fig. 6A). This model was further tested in an independent dataset, which achieved an AUC of 0.747 (Fig. 3 A). We tested the associations of the "inteS_mean", "J_up", "t_T", "mean_R", "T_minus", and "t_RT" ECG parameters with SCDHR patients and HCs at ten time intervals using univariate logistic regression. The P value of the coefficient was adjusted by the Benjamini and Hochberg method. 29 As shown in Fig. 4 A, “inteS_mean”, “t_T”, “mean_R”, “T_minus”, and “t_RT” were significantly correlated with SCDHR in the training set. An increase in “inteS_mean” was positively correlated with SCDHR (OR = 1.595 ~ 2.253, P-adjusted < 0.050), and an increase in “t_T” was also positively correlated with SCDHR (OR = 1.278, P-adjusted = 0.034) (Fig. 4 A, Supplemental Table 4). In the independent test set, however, an increase in “inteS_mean” and “t_T” was positively but not significantly correlated with SCDHR (Fig. 4 B). Nevertheless, “inteS_mean” in the independent test set significantly differed between the patients with SCDHR and HCs in the time fragments at 21:39 (P-adjusted = 0.014), and “t_T” significantly differed between the patients with SCDHR and HCs at 19:23 (P-adjusted = 0.035) (Fig. 4 C). Thus, increases in “inteS_mean” at approximately 21:30 and “t_T” at approximately 19:30 are putatively associated with SCDHR. ECG abnormalities between SCDHR and HF A multivariate logistic regression model of each time fragment was constructed to explore the difference between SCDHR and HF. The model was trained to select ECG parameters and was evaluated by fivefold cross-validations using the training set. When the models were combined into a unique model, the multivariate logistic regression model for discriminating SCDHR and HF achieved the best performance (AUC 0.857) after integrating five ECG parameters (“inteTM”, “inteSm”, “inteT_mean”, “t_T”, and “t_S_sum”) at three time fragments (10:56, 14:38, and 14:51) in the training set (Fig. 3 B and Supplemental Fig. 6B). When this model was tested in the independent set, it achieved an AUC of 0.714. The associations of each ECG parameter and the difference between SCDHR and HF in each time fragment were examined by univariable logistic regression. However, none of these parameters were independently or significantly correlated with SCDHR in HF (Supplemental Fig. 7A and 7B). We next tested the difference in these ECG parameters (“inteTM”, “inteSm”, “inteT_mean”, “t_T”, and “t_S_sum”) at the three time intervals (10:56, 14:38, and 14:51) and found that only “inteSm” significantly differed between SCDHR and HF at 14:38 in both the training set and the independent test set (Supplemental Fig. 7C). ECG abnormalities between HF and HC A logistic regression model of ECG parameters in each time fragment was constructed to discriminate HF and HC. Fivefold cross-validation was used to select the ECG parameters achieving the best AUC in the training set. When the time segments were combined into a unique model, the best-performing model was obtained after combining seven time fragments (0:15, 1:06, 5:55, 23:33, 23:39, 23:40, and 23:46) and six ECG parameters (“inteR_sum”, “T_plus”, “inteSm”, “mean_R”, “inteS_mean”, and “inteST”), which achieved an AUC of 0.965 in the training set (Supplemental Fig. 6C). This model was tested in the independent test set and achieved an AUC of 0.842 (Fig. 3 C). The associations with HF were examined by logistic regression, which demonstrated that “inteSm”, “mean_R”, “inteS_mean”, and “inteST” were significantly correlated with HF at these seven time fragments in the training set (all P-adjusted < 0.050) (Fig. 5 A). These correlations were further validated in the independent test set (Fig. 5 B). The difference test results showed that only “inteSm” and “inteS_mean” significantly differed between HF and HC in all seven time fragments (0:15, 1:06, 5:55, 23:33, 23:39, 23:40, and 23:46) (all P-adjusted < 0.001) (Fig. 5 C). Discussion Major findings The analysis of 24-hour ECG data of patients with SCDHR and HCs revealed several important findings. First, we discovered that the S-wave is sensitive to heart health and that the S-wave-related parameters significantly differed in the pairwise comparison of patients with SCDHR, patients with HF, and HCs. Especially in the comparisons of SCDHR patients and HCs as well as HF patients and HCs, the S-wave area-related parameters, inteSm and inteS_mean, were positively correlated with an increased risk of SCDHR and HF at many time points. Previous studies have demonstrated that the S-wave upslope duration ratio is related to Brugada syndrome, 30,31 and a prolonged S-wave is frequently observed in patients with SCD carrying SCN5A mutations and patients with SCD with no heart disease history. 32 Thus, the present findings suggested that S-wave-related parameters, especially the S-wave area, are important for characterizing SCDHR. Potential mechanism We considered several mechanisms to explain our findings. First, S-wave areas may reflect an altered spatial sequence of ventricular depolarization. Various pathophysiological factors may account for increased S-wave areas, such as shortened action potential durations in ischaemic subepicardial myocyte layers. 33 These underlying conditions may indicate subclinical diseases and an elevated risk of SCD. 34 Second, a decreased S-wave area, partly resulting from a prolonged P-wave duration, may reflect an arrhythmogenic myocardial substrate in the atrium and ventricle. An increasing P-wave duration has been associated with the degree of interstitial left ventricular fibrosis on cardiac magnetic resonance imaging, 35 which may present with an increased S-wave area. Finally, the present study identified the influence of circadian rhythms on the occurrence of cardiovascular events. Circadian rhythms are important for regulating various physiological functions, such as heart rate and cardiovascular diseases. 36–38 Previous studies have reported a morning surge of sympathetic nervous system activity resulting in physiological changes that produce myocardial ischaemia. 39,40 In the present study, we discuss how circadian rhythms are related to SCD and elaborate on the importance of monitoring ECG at particular time intervals. Comparison with other studies Previous studies have revealed associations between T-wave inversion 41 and the QRS-T angle 42 as well as between the T-wave peak-to-end and SCD. 43 The present study suggested that the T-wave may be important for discriminating SCDHR from HF. Additionally, we observed that the ECG parameters were less significantly different between patients with SCDHR and patients with HF than between patients with SCDHR and HCs. The underlying reason may be due to the similar clinical characteristics between SCD and SCDHR, requiring an enlarged sample size to identify the difference. limitations of study One limitation of this study was that the independent test was based on a small sample size because the prevalence of SCDHR is relatively low compared to that of HF and it is difficult to collect SCD samples. Conclusions In this prospective, multicentre cohort study, we found that time-dependent S-wave area-related ECG parameters in 24-hour ECG (inteS_mean, inteSm, and inteST) are potentially early predictive factors for SCD risk. Summary boxes What is already known on this topic 1.Sudden cardiac death is primarily a result of electrical disturbance of the normal cardiac rhythm. 2.Several ECG parameters are associated with sudden cardiac death risk. What this study adds 1.S-wave areas in certain time intervals were significantly associated with sudden cardiac death. 2.The application of these ECG parameters may be potentially useful for the early identification of patients with sudden cardiac death risk as independent predictive factors. Declarations Role of the funding source The work was partly funded by the National Key R&D Program of China (2020YFB0204803). This work was supported by the General Program of National Natural Science Foundation of China (NSFC) [Grants number: 82070237(JFW), 81870170 (JFW), 81970200 (YXC), 81770229 (YXC), 81970388 (YLZ), 81903299 (QC), 81801132 (HYZ) and 81971190 (HYZ)], the Guangzhou Health and Medical Collaborative Innovation Major Project [Grant number: 201803040010 (JFW)], Guangdong Provincial Laboratory of Regenerative Medicine and Health [Grant number: 198F041814 (JFW)], and the Natural Science Foundation of Guangdong [Grant number 2019A1515011682 (YLZ)]. The authors declare no relationships with industry. Conflicts of interest The authors have no conflicts of interest to declare. Ethics approval Institutional Review Boards of all hospitals mentioned above and received proper ethical oversight. Data sharing Study data can be made available upon request to the corresponding author. Transparency statement The lead author affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as originally planned (and, if relevant, registered) have been explained. References Aziz EF, Javed F, Pratap B, Herzog E. Strategies for the prevention and treatment of sudden cardiac death. Open Access Emerg Med 2010;2010:99-114. Wong CX, Brown A, Lau DH, et al. Epidemiology of Sudden Cardiac Death: Global and Regional Perspectives. Heart Lung Circ 2019;28:6-14. Wright RS, Anderson JL, Adams CD, et al. 2011 ACCF/AHA focused update of the Guidelines for the Management of Patients with Unstable Angina/Non-ST-Elevation Myocardial Infarction (updating the 2007 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines developed in collaboration with the American College of Emergency Physicians, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. J Am Coll Cardiol 2011;57:1920-59. Maron BJ, McKenna WJ, Danielson GK, et al. American College of Cardiology/European Society of Cardiology clinical expert consensus document on hypertrophic cardiomyopathy. A report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents and the European Society of Cardiology Committee for Practice Guidelines. J Am Coll Cardiol 2003;42:1687-713. Waks JW, Sitlani CM, Soliman EZ, et al. Global Electric Heterogeneity Risk Score for Prediction of Sudden Cardiac Death in the General Population: The Atherosclerosis Risk in Communities (ARIC) and Cardiovascular Health (CHS) Studies. Circulation 2016;133:2222-34. Terho HK, Tikkanen JT, Kentta TV, et al. Electrocardiogram as a predictor of sudden cardiac death in middle-aged subjects without a known cardiac disease. Int J Cardiol Heart Vasc 2018;20:50-5. Teodorescu C, Reinier K, Uy-Evanado A, et al. Prolonged QRS duration on the resting ECG is associated with sudden death risk in coronary disease, independent of prolonged ventricular repolarization. Heart Rhythm 2011;8:1562-7. Pieroni M, Bellocci F, Crea F. Sudden cardiac arrest associated with early repolarization. N Engl J Med 2008;359:761-2; author reply 2. Kim SH, Kim DY, Kim HJ, et al. Early repolarization with horizontal ST segment may be associated with aborted sudden cardiac arrest: a retrospective case control study. BMC Cardiovasc Disord 2012;12:122. Rosso R, Kogan E, Belhassen B, et al. J-point elevation in survivors of primary ventricular fibrillation and matched control subjects: incidence and clinical significance. J Am Coll Cardiol 2008;52:1231-8. Lambiase PD, Young WJ. Sudden cardiac death risk prediction - As easy as ECG? Int J Cardiol 2018;276:152-3. Perez-Riera AR, Barbosa-Barros R, Shenasa M. Electrocardiographic Markers of Sudden Cardiac Death (Including Left Ventricular Hypertrophy). Card Electrophysiol Clin 2017;9:605-29. Aro AL, Reinier K, Rusinaru C, et al. Electrical risk score beyond the left ventricular ejection fraction: prediction of sudden cardiac death in the Oregon Sudden Unexpected Death Study and the Atherosclerosis Risk in Communities Study. Eur Heart J 2017;38:3017-25. Wellens HJ, Schwartz PJ, Lindemans FW, et al. Risk stratification for sudden cardiac death: current status and challenges for the future. Eur Heart J 2014;35:1642-51. Perez-Alday EA, Li-Pershing Y, Bender A, et al. Importance of the heart vector origin point definition for an ECG analysis: The Atherosclerosis Risk in Communities (ARIC) study. Comput Biol Med 2018;104:127-38. Lyon A, Ariga R, Minchole A, et al. Distinct ECG Phenotypes Identified in Hypertrophic Cardiomyopathy Using Machine Learning Associate With Arrhythmic Risk Markers. Front Physiol 2018;9:213. Holkeri A, Eranti A, Haukilahti MAE, et al. Predicting sudden cardiac death in a general population using an electrocardiographic risk score. Heart 2020;106:427-33. Adabag AS, Luepker RV, Roger VL, Gersh BJ. Sudden cardiac death: epidemiology and risk factors. Nat Rev Cardiol 2010;7:216-25. Akhtar M, Elliott PM. Risk Stratification for Sudden Cardiac Death in Non-Ischaemic Dilated Cardiomyopathy. Curr Cardiol Rep 2019;21:155. Willich SN, Levy D, Rocco MB, Tofler GH, Stone PH, Muller JE. Circadian variation in the incidence of sudden cardiac death in the Framingham Heart Study population. Am J Cardiol 1987;60:801-6. Levine RL, Pepe PE, Fromm RE, Jr., Curka PA, Clark PA. Prospective evidence of a circadian rhythm for out-of-hospital cardiac arrests. JAMA 1992;267:2935-7. Bagai A, McNally BF, Al-Khatib SM, et al. Temporal differences in out-of-hospital cardiac arrest incidence and survival. Circulation 2013;128:2595-602. Fukuda T, Ohashi-Fukuda N, Kondo Y, Sera T, Doi K, Yahagi N. Epidemiology, Risk Factors, and Outcomes of Out-of-Hospital Cardiac Arrest Caused by Stroke: A Population-Based Study. Medicine (Baltimore) 2016;95:e3107. Englund A, Behrens S, Wegscheider K, Rowland E. Circadian variation of malignant ventricular arrhythmias in patients with ischemic and nonischemic heart disease after cardioverter defibrillator implantation. European 7219 Jewel Investigators. J Am Coll Cardiol 1999;34:1560-8. Peters RW, McQuillan S, Gold MR. Interaction of septadian and circadian rhythms in life-threatening ventricular arrhythmias in patients with implantable cardioverter-defibrillators. Am J Cardiol 1999;84:555-7. Ruwald MH, Moss AJ, Zareba W, et al. Circadian distribution of ventricular tachyarrhythmias and association with mortality in the MADIT-CRT trial. J Cardiovasc Electrophysiol 2015;26:291-9. Ramireddy A, Chugh HS, Reinier K, et al. Sudden cardiac death during nighttime hours. Heart Rhythm 2021;18:778-84. Hartigan JA, Wong MA. Algorithm AS 136: A K-Means Clustering Algorithm. Appl Stat 1979;28:100-8. Haynes W. Benjamini–Hochberg Method. In: Dubitzky W, Wolkenhauer O, Cho K-H, Yokota H, editors. Encyclopedia of Systems Biology. New York, NY: Springer New York; 2013. p. 78. Subramanian M, Prabhu MA, Rai M, et al. A novel prediction model for risk stratification in patients with a type 1 Brugada ECG pattern. J Electrocardiol 2019;55:65-71. Michowitz Y, Milman A, Andorin A, et al. Characterization and Management of Arrhythmic Events in Young Patients With Brugada Syndrome. J Am Coll Cardiol 2019;73:1756-65. Vanninen SUM, Nikus K, Aalto-Setala K. Electrocardiogram changes and atrial arrhythmias in individuals carrying sodium channel SCN5A D1275N mutation. Ann Med 2017;49:496-503. Rautaharju PM, Zhou SH, Gregg RE, Startt-Selvester RH. Heart rate, gender differences, and presence versus absence of diagnostic ST elevation as determinants of spatial QRS|T angle widening in acute coronary syndrome. Am J Cardiol 2011;107:1744-50. Whang W, Shimbo D, Levitan EB, et al. Relations between QRS| T angle, cardiac risk factors, and mortality in the third National Health and Nutrition Examination Survey (NHANES III). Am J Cardiol 2012;109:981-7. Win TT, Venkatesh BA, Volpe GJ, et al. Associations of electrocardiographic P-wave characteristics with left atrial function, and diffuse left ventricular fibrosis defined by cardiac magnetic resonance: The PRIMERI Study. Heart Rhythm 2015;12:155-62. Kollias GE, Stamatelopoulos KS, Papaioannou TG, et al. Diurnal variation of endothelial function and arterial stiffness in hypertension. J Hum Hypertens 2009;23:597-604. Degaute JP, van de Borne P, Linkowski P, Van Cauter E. Quantitative analysis of the 24-hour blood pressure and heart rate patterns in young men. Hypertension 1991;18:199-210. Crnko S, Du Pré BC, Sluijter JP, Van Laake LW. Circadian rhythms and the molecular clock in cardiovascular biology and disease. Nat Rev Cardiol 2019;16:437-47. Muller JE, Tofler GH, Stone PH. Circadian variation and triggers of onset of acute cardiovascular disease. Circulation 1989;79:733-43. Willich S, Maclure M, Mittleman M, Arntz H-R, Muller J. Sudden cardiac death. Support for a role of triggering in causation. Circulation 1993;87:1442-50. Rautaharju PM, Surawicz B, Gettes LS. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: part IV: the ST segment, T and U waves, and the QT interval a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society Endorsed by the International Society for Computerized Electrocardiology. J Am Coll Cardiol 2009;53:982-91. Aro AL, Huikuri HV, Tikkanen JT, et al. QRS-T angle as a predictor of sudden cardiac death in a middle-aged general population. Europace 2012;14:872-6. Panikkath R, Reinier K, Uy-Evanado A, et al. Prolonged Tpeak-to-tend interval on the resting ECG is associated with increased risk of sudden cardiac death. Circ Arrhythmia Electrophysiol 2011;4:441-7. Additional Declarations There is NO Competing Interest. Supplementary Files SupplementalAppendix.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3490411","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":271389912,"identity":"c2c8afba-c84f-49ef-9be4-c11443153244","order_by":0,"name":"Xinfeng Wang","email":"data:image/png;base64,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","orcid":"","institution":"","correspondingAuthor":true,"prefix":"","firstName":"Xinfeng","middleName":"","lastName":"Wang","suffix":""},{"id":271389913,"identity":"65b39865-70f0-4b33-8791-22a892c34291","order_by":1,"name":"Wenhao Liu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Wenhao","middleName":"","lastName":"Liu","suffix":""},{"id":271389914,"identity":"2643fbc4-ce5b-4264-93aa-997b2d8e335e","order_by":2,"name":"Ziheng Zheng","email":"","orcid":"","institution":"Sun Yat-sen Memorial Hospital, Sun Yat-sen University","correspondingAuthor":false,"prefix":"","firstName":"Ziheng","middleName":"","lastName":"Zheng","suffix":""},{"id":271389915,"identity":"9cf5ff3a-21be-4ec7-8753-790d2ac4af41","order_by":3,"name":"Wenli Gu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Wenli","middleName":"","lastName":"Gu","suffix":""},{"id":271389916,"identity":"0276f348-0e41-49b1-a920-7691fba0f836","order_by":4,"name":"Zhiteng Chen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Zhiteng","middleName":"","lastName":"Chen","suffix":""},{"id":271389917,"identity":"0fbd0523-6ab0-4b0e-a21c-f676be3b2c6d","order_by":5,"name":"Qi Guo","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Qi","middleName":"","lastName":"Guo","suffix":""},{"id":271389918,"identity":"16dff075-71ab-4803-9bec-2ec58c552824","order_by":6,"name":"Liqun Wu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Liqun","middleName":"","lastName":"Wu","suffix":""},{"id":271389919,"identity":"7b5dd797-0504-40fe-ab6b-119ca5942a6d","order_by":7,"name":"Wei Xu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Wei","middleName":"","lastName":"Xu","suffix":""},{"id":271389920,"identity":"8239d5b5-3de5-45dd-b3b8-feb2de45cb93","order_by":8,"name":"Yangang Su","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yangang","middleName":"","lastName":"Su","suffix":""},{"id":271389921,"identity":"010b49a8-b0a9-4810-b79a-dde3e12e0cfd","order_by":9,"name":"Keping Chen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Keping","middleName":"","lastName":"Chen","suffix":""},{"id":271389922,"identity":"e25cbf21-45ef-4c10-8dd8-8c9ea359741a","order_by":10,"name":"MaoXiong Wu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"MaoXiong","middleName":"","lastName":"Wu","suffix":""},{"id":271389923,"identity":"15842db2-5ea9-4b3d-a6a1-1e91255580bd","order_by":11,"name":"Qian Chen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Qian","middleName":"","lastName":"Chen","suffix":""},{"id":271389924,"identity":"334aefb0-e1a8-422b-b1b6-c1c4d397b034","order_by":12,"name":"Xiangkun Xie","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xiangkun","middleName":"","lastName":"Xie","suffix":""},{"id":271389925,"identity":"2da18646-0c26-48fa-8d74-41e39fc7c711","order_by":13,"name":"Xiaoying Wu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xiaoying","middleName":"","lastName":"Wu","suffix":""},{"id":271389926,"identity":"dc0de147-8961-485f-a497-f38dd0eaa2ab","order_by":14,"name":"Xiao Liu","email":"","orcid":"https://orcid.org/0000-0001-8252-3435","institution":"Sun Yat-Sen University","correspondingAuthor":false,"prefix":"","firstName":"Xiao","middleName":"","lastName":"Liu","suffix":""},{"id":271389927,"identity":"739184c9-9b68-4db1-ae23-6b3bf727a58d","order_by":15,"name":"Hongwei Li","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Hongwei","middleName":"","lastName":"Li","suffix":""},{"id":271389928,"identity":"d16f97cd-af5e-4a26-98b0-dbe35fcf79a1","order_by":16,"name":"Yuedong Yang","email":"","orcid":"https://orcid.org/0000-0002-6782-2813","institution":"Sun Yat-sen University","correspondingAuthor":false,"prefix":"","firstName":"Yuedong","middleName":"","lastName":"Yang","suffix":""},{"id":271389929,"identity":"cdf108a5-138d-4544-8a01-fc35e8890eac","order_by":17,"name":"Yuling Zhang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yuling","middleName":"","lastName":"Zhang","suffix":""},{"id":271389930,"identity":"c95f86d0-b300-4ba7-ac08-6f3aab23f6a5","order_by":18,"name":"Yangxin Chen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yangxin","middleName":"","lastName":"Chen","suffix":""},{"id":271389931,"identity":"cefbacfd-6e40-4522-8354-ef3cccb13b30","order_by":19,"name":"Mengling Qi","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Mengling","middleName":"","lastName":"Qi","suffix":""},{"id":271389932,"identity":"41aacf58-9f96-4b37-b15f-bcabf80fce69","order_by":20,"name":"Huiying Zhao","email":"","orcid":"","institution":"Sun Yat-sen University","correspondingAuthor":false,"prefix":"","firstName":"Huiying","middleName":"","lastName":"Zhao","suffix":""}],"badges":[],"createdAt":"2023-10-25 12:26:39","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3490411/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3490411/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":50924691,"identity":"9ac70d9b-c30f-415a-bede-76c3f93084fd","added_by":"auto","created_at":"2024-02-09 17:05:23","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":103642,"visible":true,"origin":"","legend":"\u003cp\u003eOverall workflow and main results of this study. SCDHR, high risk of sudden cardiac death; HC, healthy control; inteS_mean, the average area under the S-wave; inteSm, the minimum area under the S-wave; inteST, the difference in the S-wave and T-wave.\u003c/p\u003e","description":"","filename":"figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-3490411/v1/6185bec7cb84b9572beee12d.png"},{"id":50924692,"identity":"d4e8b732-b73b-4996-9b67-88133dfebe1a","added_by":"auto","created_at":"2024-02-09 17:05:23","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":223108,"visible":true,"origin":"","legend":"\u003cp\u003eECG parameters and time fragments in the training set. A. The number of features showing significant differences between patients with SCDHR and HCs in time fragments. B. The number of features showing significant differences between SCDHR patients and HF patients in time-fragments. C. The number of features showing significant differences between HF patients and HCs in the time fragments. D. The number of time fragments for features showing significant differences between SCDHR patients and HCs. E. The number of time fragments for features showing significant differences between SCDHR patients and HF patients. F. The number of time fragments for features showing significant differences between HF patients and HCs. A \u003cem\u003eP value\u003c/em\u003e\u0026lt;0.0012 (0.05/42) was considered statistically significant.\u003c/p\u003e","description":"","filename":"figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-3490411/v1/2c94d9fe9707c0282ca495fc.png"},{"id":50924695,"identity":"0ca20202-2c88-4765-b089-c14465f7372c","added_by":"auto","created_at":"2024-02-09 17:05:24","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":40182,"visible":true,"origin":"","legend":"\u003cp\u003eMultivariate logistic regression models for discriminating SCDHR patients, HF patients, and HCs. A. The ROC curves of the combined model in discriminating SCDHR patients and HCs. B. The ROC curves of the combined model in discriminating SCDHR patients and HF patients. C. The ROC curves of the combined model in discriminating HF patients and HCs. ROC, receiver operating characteristic curve; Test, independent test set; Training, training set.\u003c/p\u003e","description":"","filename":"figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-3490411/v1/a686210787777897637317a8.png"},{"id":50924697,"identity":"31537308-5c9e-408e-8d20-32757278be1f","added_by":"auto","created_at":"2024-02-09 17:05:24","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":603973,"visible":true,"origin":"","legend":"\u003cp\u003eThe OR values of combined features between SCDHR patients and HCs. A. The OR values of individual features for SCDHR in the training set at significant time-fragments. B. The OR values of individual features associated with SCDHR in the independent dataset at multiple time points. An OR\u0026lt;1.0 indicates that the ECG parameters were negatively correlated with SCDHR, while an OR \u0026gt;1.0 indicates that the ECG parameters were positively correlated with SCDHR. The OR values indicate the change in SCDHR when the feature changes by 1% unit. C. The difference in these features in SCDHR patients and HCs at these time-fragments was tested by Student’s t test. The title of each subplot indicates the time-fragment of this comparison. OR, odds ratio; CI: confidence interval.\u003c/p\u003e","description":"","filename":"figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-3490411/v1/2d93f018c718234787b35e6e.png"},{"id":50924696,"identity":"9321e301-aaf1-4e60-8ed6-3690cfe4302c","added_by":"auto","created_at":"2024-02-09 17:05:24","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":563821,"visible":true,"origin":"","legend":"\u003cp\u003eThe OR values of combined features between HF patients and HCs. A. The OR values of individual features for HF patients in the training set at significant time-fragments. B. The OR values of individual features associated with HF patients and HCs in the independent set at multiple time points. The OR values indicate the change in risk when the feature changes by 1% unit. C. The difference in these features in the HF patients and HCs at these time-fragments was tested by Student’s t test. The title of each subplot indicates the time-fragment of this comparison. OR, odds ratio; CI: confidence interval.\u003c/p\u003e","description":"","filename":"figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-3490411/v1/ef3bfcd87a901cef369f2d8e.png"},{"id":50924694,"identity":"331e8234-3ccb-48f2-866b-1130a174f85e","added_by":"auto","created_at":"2024-02-09 17:05:24","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":580510,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCentral Illustration. \u003c/strong\u003eSchematic of revealing ECG parameters associated with SCDHR or HF. First, the ECG data of each individual were scattered into one-minute-long time fragment; Secondly, the multi-variate logistic regression was used to select critical ECG parameters; Third, the associations of individual ECG parameters with SCD or HF were tested by univariable logistic regression. Finally, the results the time-dependent S-wave area-related ECG parameters (inteS_mean, inteSm, and inteST) are potentially early predictive factors for SCD risk.\u003c/p\u003e\n\u003cp\u003eSCDHR, high risk of sudden cardiac death.\u003cem\u003e P-value\u003c/em\u003e \u0026lt;0.0012 (0.05/42) was considered statistically significant.\u003c/p\u003e","description":"","filename":"centralfigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-3490411/v1/5f3687a4ab01a62654c8b851.png"},{"id":51495569,"identity":"a33b5992-4ed3-42d5-9741-6b60d6e43588","added_by":"auto","created_at":"2024-02-22 15:29:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1929847,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3490411/v1/34386a0b-99cd-44fb-852e-d4e1fbe343ea.pdf"},{"id":50924693,"identity":"9a45e887-36f8-4e2e-a184-c2d59b722405","added_by":"auto","created_at":"2024-02-09 17:05:24","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1142034,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"SupplementalAppendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-3490411/v1/88f362ad3a9aafb7ce1f252f.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Time-dependent S-wave areas predict sudden cardiac death risk: a prospective, multicentre registered study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSudden cardiac arrest (SCA) is a major cause of mortality in the Western world, causing\u0026thinsp;\u0026gt;\u0026thinsp;300,000 deaths annually in the USA\u003csup\u003e1\u003c/sup\u003e and accounting for 20% of SCD in Western societies.\u003csup\u003e2\u003c/sup\u003e According to the 2011 American College of Cardiology Foundation/American Heart Association (ACCF/AHA)\u003csup\u003e3\u003c/sup\u003e and 2003 American College of Cardiology/European Society of Cardiology (ACC/ESC) guidelines,\u003csup\u003e4\u003c/sup\u003e malignant arrhythmia is a critical characteristic of patients with a high risk of SCD (SDCHR). However, readily available methods to identify these individuals are lacking.\u003csup\u003e5\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eAs SCD is primarily a result of electrical disturbance of the normal cardiac rhythm,\u003csup\u003e6\u003c/sup\u003e the widely available and low-cost ECG is a potentially attractive and noninvasive tool for SCD risk stratification. By using ECG data, most studies have attempted to detect the abnormal characteristics associated with SCD risk by comparing cases with controls.\u003csup\u003e7\u0026ndash;9\u003c/sup\u003e The ECG parameters, including J-wave amplitude,\u003csup\u003e10\u003c/sup\u003e QRS prolongation, QT prolongation, T-wave alternans, and QRS fractionation,\u003csup\u003e11\u003c/sup\u003e have been well documented. To date, however, no individual ECG finding has adequately stratified patients with a risk for SCD.\u003csup\u003e12\u003c/sup\u003e Although recent studies have indicated that the combination of a few known ECG parameters result in improved SCD risk prediction,\u003csup\u003e6,13\u003c/sup\u003e the identification of optimal ECG risk markers remains a challenge.\u003csup\u003e14\u003c/sup\u003e Moreover, most of the studies are based on previously known ECG markers,\u003csup\u003e15\u0026ndash;17\u003c/sup\u003e which lack extraction and use related data, such as the positioning of S and T peaks as well as peak areas, to make judgements. Additionally, patients with dilated cardiomyopathy or ischaemic cardiomyopathy are considered to have a high risk of SCD.\u003csup\u003e18\u003c/sup\u003e A previous study has indicated that approximately 30% of patients with dilated cardiomyopathy or ischaemic cardiomyopathy are at risk of developing SCD.\u003csup\u003e19\u003c/sup\u003e No studies have explored the differences in ECG parameters among patients with SCDHR or with a history of SCD. Another unknown of SCD is the episode timing. The Framingham cohort study first reported a morning peak of SCA.\u003csup\u003e20\u003c/sup\u003e A subsequent study, reporting approximately 1,000 events, has demonstrated that SCD episodes significantly increase 6:00 AM to 12:00 PM.\u003csup\u003e21\u003c/sup\u003e This result was later verified in prospective community studies of SCA and cohorts of patients implanted with ICDs.\u003csup\u003e22\u0026ndash;26\u003c/sup\u003e Recently, a study involving 4,126 SCD cases has revealed that women are more likely to develop SCD at night.\u003csup\u003e27\u003c/sup\u003e Nevertheless, there is still a lack of dynamic risk assessment over time of the association between SCD and ECG markers. Thus, the time intervals in risk assessment that reliably predict individual risk for SCD remain unknown.\u003c/p\u003e \u003cp\u003eIn the present study, to identify time-dependent ECG parameters associated with SCD risk, we collected 24-hour ECG data, including patients who were survivors of SCD or with haemodynamic disorder due to sustained ventricular tachycardia/ventricular fibrillation (SCDHR), patients with heart failure (HF), and healthy controls (HCs). The ECG data of each sample were divided into one-minute-long fragments and were compared between SCDHR patients and HCs, between SCDHR patients and HF patients, and between HF patients and HCs. Multivariate logistic regression models were then built to reveal the time-dependent ECG parameters associated with SCDHR in HC, SCDHR in HF, and HF in HC. Finally, these ECG parameters were validated in an independent dataset.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003eThis study was a prospective, multicentre cohort study. We enrolled survivors of SCD or patients who suffered haemodynamic disorder due to sustained ventricular tachycardia/ventricular fibrillation (SCDHR group), patients with HF who did not experience sustained ventricular tachycardia/ventricular fibrillation but were diagnosed with dilated cardiomyopathy or ischaemic cardiomyopathy with LVEF\u0026thinsp;\u0026le;\u0026thinsp;35% (HF group), and healthy controls who presented with no heart disease (HC group). The exclusion criteria of the participants were as follows: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) had a high degree of atrioventricular block (\u0026ge;\u0026thinsp;II) at resting heart rates; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) were in NYHA IV at inclusion; (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) had an MI history within one month; (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e) had undergone coronary revascularization within the preceding three months; (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e) had advanced cerebrovascular or renal disease and suffered from any noncardiac condition with a high likelihood of death during the trial; (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e) had a life expectancy of one year or less; and (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e) had a history of valvular heart disease.\u003c/p\u003e \u003cp\u003eThe 24-hour ECG data regarding SCDHR, HF, and HC were collected in four hospitals (Sun Yat-sen Memorial Hospital, Ruijin Hospital of Shanghai Jiao Tong University, Drum Tower Affiliated Hospital of Nanjing University Medical School, and Zhongshan Hospital of Fudan University) in China. The numbers of patients with SCDHR and HF collected by each hospital are shown in Supplemental Table\u0026nbsp;1. This study was approved by the Institutional Review Boards of all hospitals mentioned above and received proper ethical oversight. Signed informed consent was provided by all participants. All patients and controls in this study had archived resting v3-e ECG data with a paper speed of 25 mm/s and calibration of 10 mm/mV.\u003c/p\u003e \u003cp\u003eThis study was registered on the website \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://register.clinicaltrails.gov/Organization\u003c/span\u003e\u003cspan address=\"http://register.clinicaltrails.gov/Organization\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The Clinical Trials ID is NCT03485079.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eData filtering\u003c/h2\u003e \u003cp\u003eThe ECG data of each sample were divided into one-minute-long time fragments. To filter the noise, each fragment was further divided into 60 pieces. The filtering principles in every second were as follows: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) the maximum value in the one-second fragment was lower than 0.2 mV, (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) the maximum value was higher than 5 mV, or (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) the difference between the maximum and minimum value was greater than twice the median difference in this minute. If the data length was less than 30 seconds in a one-minute fragment, the fragment was discarded. Samples with ECG signal lengths less than 100 minutes were excluded.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eCalculation of ECG parameters\u003c/h2\u003e \u003cp\u003eIn each one-minute-long time fragment, 42 parameters were calculated, including the following parameters: maximum and minimum peak values of waves (R-wave, T-wave, and S-wave); the maximum, minimum and average area under waves; the summation of peak values of waves; and the summation of the area under waves. Details of the ECG parameters are shown in Supplemental Table\u0026nbsp;2.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistics\u003c/h2\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003eComparison of ECG parameters across samples\u003c/h2\u003e \u003cp\u003eThe ECG parameters were first compared across the SCDHR, HF and HC groups. Differences between two groups were tested by Student\u0026rsquo;s t test. A \u003cem\u003eP value\u003c/em\u003e less 0.0012 (0.05/42) was considered statistically significant.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis to identify ECG parameters associated with SCDHR or HF\u003c/h2\u003e \u003cp\u003eMultivariate logistic regressions were used to detect the ECG parameters associated with SCDHR or HF. To construct the models, we randomly selected 70% of the samples collected from Sun Yat-sen Memorial Hospital as the training set. The remaining 30% of the samples from Sun Yat-sen Memorial Hospital and samples from other hospitals were used as an independent test dataset.\u003c/p\u003e \u003cp\u003eTo exclude redundant features and reduce multicollinearity, we used K-means\u003csup\u003e28\u003c/sup\u003e to cluster the 42 ECG parameters in the training set. The most fitted parameter in each cluster was kept as an independent parameter. The most fitted ECG parameter was the parameter indicated by the lowest Akaike information criterion in the univariable logistic regression model testing the association between SCDHR patients and HCs, between SCDHR patients and HF patients, or between HF patients and HCs.\u003c/p\u003e \u003cp\u003eThe features selected from all clusters were used to build a multivariate logistic regression model. We used an auto backward approach to select the features most likely associated with the SCDHR or HF. The performance of the multivariate logistic regression model was evaluated by the average area under the curve (AUC) of 5-fold cross-validation. Briefly, from the training set, we randomly selected 4/5 of the data for training, and the remaining 1/5 of the data was used for testing. This process was repeated five times, and the AUC of each fragment was calculated to detect the parameters most correlated with SCDHR or HF.\u003c/p\u003e \u003cp\u003eTo select time-dependent ECG parameters associated with SCDHR or HF, we constructed a multivariate logistic regression model that connected the models trained in different time fragments into one model by a greedy searching approach in the training set. First, the models of different time fragments were ranked according to the fivefold average AUC values. Then, the best-performing model was connected to the second-best model to construct the combined model. The connected model was evaluated by the average AUC of the fivefold cross-validation. This process was continued until all models of the 1440 time fragments were combined. The model yielding the best AUC and integrating more than two fragments was used for further analysis and finally tested in the independent test set.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eMeasurements of the fitness of the regression models\u003c/h2\u003e \u003cp\u003eTo evaluate the ECG parameters associated with SCDHR or HF, the fitness of the regression models was measured by AUC. The workflow is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eDemographic and clinical characteristics\u003c/h2\u003e \u003cp\u003eThe demographic and clinical data of patients with SCDHR, HF, and HC are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. In total, 35.71% of patients with SCDHR, 28.57% of patients with HF, and 9.68% of HCs had a smoking history. Among the samples, 12.50% of patients with SCDHR and 19.85% of patients with HF had diabetes, whereas none of the HCs reported diabetes. Regarding other comorbidities, 39.02% of patients with SCDHR, 35.88% of patients with HF, and 9.02% of HCs had hypertension, and 9.30% of patients with SCDHR, 20.66% of patients with HF, and 0.81% of HCs had coronary artery disease.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe demographic, disease history and clinical data among patients with SCDHR, HF and HC.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical feature\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHC(n\u0026thinsp;=\u0026thinsp;108)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHF(n\u0026thinsp;=\u0026thinsp;138)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSCDHR(n\u0026thinsp;=\u0026thinsp;43)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (y)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e46.07\u0026thinsp;\u0026plusmn;\u0026thinsp;10.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e56.09\u0026thinsp;\u0026plusmn;\u0026thinsp;13.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e54.88\u0026thinsp;\u0026plusmn;\u0026thinsp;15\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e43.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e69.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e81.40\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiabetes (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e12.50\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHypertension (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e35.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e39.02\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCoronary artery disease history (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e9.30\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHyperlipidemia (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.76\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSmoke history (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e28.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e35.71\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDrink (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e19.05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSyncope (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e29.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlucose (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.97\u0026thinsp;\u0026plusmn;\u0026thinsp;0.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.96\u0026thinsp;\u0026plusmn;\u0026thinsp;0.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.76\u0026thinsp;\u0026plusmn;\u0026thinsp;0.77\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUric acid (\u0026micro;mol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e373.12\u0026thinsp;\u0026plusmn;\u0026thinsp;99.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e459.59\u0026thinsp;\u0026plusmn;\u0026thinsp;144.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e430.51\u0026thinsp;\u0026plusmn;\u0026thinsp;131.97\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal cholesterol (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.18\u0026thinsp;\u0026plusmn;\u0026thinsp;0.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.23\u0026thinsp;\u0026plusmn;\u0026thinsp;1.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.29\u0026thinsp;\u0026plusmn;\u0026thinsp;0.99\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTriglyceride (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.19\u0026thinsp;\u0026plusmn;\u0026thinsp;0.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.22\u0026thinsp;\u0026plusmn;\u0026thinsp;0.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.18\u0026thinsp;\u0026plusmn;\u0026thinsp;0.48\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHDL-C (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.30\u0026thinsp;\u0026plusmn;\u0026thinsp;0.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.08\u0026thinsp;\u0026plusmn;\u0026thinsp;0.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.13\u0026thinsp;\u0026plusmn;\u0026thinsp;0.31\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLDL-C (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.28\u0026thinsp;\u0026plusmn;\u0026thinsp;0.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.62\u0026thinsp;\u0026plusmn;\u0026thinsp;0.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.75\u0026thinsp;\u0026plusmn;\u0026thinsp;0.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eapoA1 (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.39\u0026thinsp;\u0026plusmn;\u0026thinsp;0.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.11\u0026thinsp;\u0026plusmn;\u0026thinsp;0.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.14\u0026thinsp;\u0026plusmn;\u0026thinsp;0.25\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eapoB (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.91\u0026thinsp;\u0026plusmn;\u0026thinsp;0.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.81\u0026thinsp;\u0026plusmn;\u0026thinsp;0.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.80\u0026thinsp;\u0026plusmn;\u0026thinsp;0.22\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCREA (\u0026micro;mol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e75.38\u0026thinsp;\u0026plusmn;\u0026thinsp;13.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e92.08\u0026thinsp;\u0026plusmn;\u0026thinsp;18.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e89.39\u0026thinsp;\u0026plusmn;\u0026thinsp;17.16\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCystatin (mg/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.82\u0026thinsp;\u0026plusmn;\u0026thinsp;0.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.14\u0026thinsp;\u0026plusmn;\u0026thinsp;0.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.02\u0026thinsp;\u0026plusmn;\u0026thinsp;0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWBC (\u0026times;10\u003csup\u003e9\u003c/sup\u003e/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.99\u0026thinsp;\u0026plusmn;\u0026thinsp;1.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.73\u0026thinsp;\u0026plusmn;\u0026thinsp;1.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.66\u0026thinsp;\u0026plusmn;\u0026thinsp;2.02\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHGB (g/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e137.63\u0026thinsp;\u0026plusmn;\u0026thinsp;15.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e138.39\u0026thinsp;\u0026plusmn;\u0026thinsp;18.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e138.25\u0026thinsp;\u0026plusmn;\u0026thinsp;19.96\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePLT (\u0026times;10\u003csup\u003e9\u003c/sup\u003e/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e248.68\u0026thinsp;\u0026plusmn;\u0026thinsp;51.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e196.52\u0026thinsp;\u0026plusmn;\u0026thinsp;57.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e223.24\u0026thinsp;\u0026plusmn;\u0026thinsp;59.52\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNeutrophil (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.57\u0026thinsp;\u0026plusmn;\u0026thinsp;0.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.53\u0026thinsp;\u0026plusmn;\u0026thinsp;5.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.90\u0026thinsp;\u0026plusmn;\u0026thinsp;4.52\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNT-pro BNP (pg/ml)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e46.85\u0026thinsp;\u0026plusmn;\u0026thinsp;51.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1774.78\u0026thinsp;\u0026plusmn;\u0026thinsp;1311.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e837.00\u0026thinsp;\u0026plusmn;\u0026thinsp;1127.69\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLA (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e31.27\u0026thinsp;\u0026plusmn;\u0026thinsp;3.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43.89\u0026thinsp;\u0026plusmn;\u0026thinsp;6.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e38.11\u0026thinsp;\u0026plusmn;\u0026thinsp;7.94\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIVSD (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.41\u0026thinsp;\u0026plusmn;\u0026thinsp;1.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.18\u0026thinsp;\u0026plusmn;\u0026thinsp;1.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e9.08\u0026thinsp;\u0026plusmn;\u0026thinsp;1.48\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVDd (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e46.19\u0026thinsp;\u0026plusmn;\u0026thinsp;3.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e63.75\u0026thinsp;\u0026plusmn;\u0026thinsp;12.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e57.58\u0026thinsp;\u0026plusmn;\u0026thinsp;10.75\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVPWd (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.28\u0026thinsp;\u0026plusmn;\u0026thinsp;1.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.18\u0026thinsp;\u0026plusmn;\u0026thinsp;1.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8.97\u0026thinsp;\u0026plusmn;\u0026thinsp;1.28\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEF (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e68\u0026thinsp;\u0026plusmn;\u0026thinsp;4.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36\u0026thinsp;\u0026plusmn;\u0026thinsp;15.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e47\u0026thinsp;\u0026plusmn;\u0026thinsp;17\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eHDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; WBC, white blood cell; HGB, hemoglobin; LA, left atrium; IVSD, interventricular septal dimension; LVDd, left ventricular diastolic dimension; LVPWd, left ventricular posterior wall dimension; EF, ejection fraction.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eAnalysis of individual ECG parameters among SCDHR, HF, and HC\u003c/h2\u003e \u003cp\u003eAfter excluding data with noise, the number of patients included in this study is shown in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Differences in ECG parameters among SCDHR, HF, and HC were examined in each of the 1440 fragments.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eNumber of patients with SCDHR, HF and HC involved in this study after filtering noise.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSample group\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eHospital\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e \u003cp\u003eGroups\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSCDHR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHF\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHC\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTraining set\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSun Yat-sen Memorial hospital\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e75\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eIndependent test set\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSun Yat-sen Memorial hospital\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e33\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eZhongshan Hospital of Fudan University\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDrum Tower Hospital, Nanjing University Medical School\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRuijin Hospital, Shanghai Jiao Tong University\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e43\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eWhen comparing the ECG parameters of patients with SCDHR and HCs in the training set, we found 20 ECG parameters that significantly differed (P value\u0026thinsp;\u0026lt;\u0026thinsp;0.0012) in at least one time fragment (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD), which were concentrated from 17:00\u0026ndash;22:00. Among them, four ECG parameters (\u0026ldquo;heart_rate\u0026rdquo;, \u0026ldquo;inteRS\u0026rdquo;, \u0026ldquo;t_T\u0026rdquo;, and \u0026ldquo;inteT_sum\u0026rdquo;) significantly differed between patients with SCDHR and HCs in more than 10 fragments, including two parameters (\u0026ldquo;t_T\u0026rdquo; and \u0026ldquo;inteT_sum\u0026rdquo;) related to T-waves within the time intervals ranging from 8:00\u0026ndash;12:00. Supplemental Fig.\u0026nbsp;2 shows the P value distributions of these four features in the 1440 time segments.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eComparison of the ECG parameters of SCDHR and HF indicated that a few time fragments contained significantly different ECG parameters between SCD and SCDHR (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB) in the training set, including 14 ECG parameters that significantly differed (P value\u0026thinsp;\u0026lt;\u0026thinsp;0.0012) in at least one time fragment (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE). Among them, four ECG features, namely, \"maxS\", \"inteRST\", \"inteST\", and \"t_S_sum\", significantly differed in more than ten fragments. All of these features are related to the S-wave. Supplemental Fig.\u0026nbsp;3 shows the P value distributions of these four features in the 1440 time segments. The results shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE and Supplemental Fig.\u0026nbsp;3 suggest that S-wave-related ECG parameters are important in discriminating SCD from SCDHR.\u003c/p\u003e \u003cp\u003eCompared to the ECG data collected from Sun Yat-sen Memorial Hospital, more time fragments were observed to contain ECG parameters that significantly differed between SCDHR and HF (Supplemental Fig.\u0026nbsp;1B) in the independent test set. The \u0026ldquo;heart_rate\u0026rdquo; was validated to play an important role in discriminating SCDHR and HF (Supplemental Fig.\u0026nbsp;1E).\u003c/p\u003e \u003cp\u003eAnalysis of the differences in ECG parameters between HF and HC demonstrated that significantly different signals were distributed at 21:00\u0026ndash;5:00 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC), and the same phenomenon was found in the independent test set (Supplemental Fig.\u0026nbsp;1C). Importantly, ten ECG parameters (\"inteST\", \"inteSm\", \"inteS_mean\", \"mean_S\", \"inteS_sum\", \"maxS\", \"minS\", \"inteRST\", \"inteRS\", and \"t_RT\") significantly differed in more than 20 fragments (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF). Interestingly, nine of these parameters are related to S-waves. Supplemental Fig.\u0026nbsp;4 shows the P value distributions of these ten parameters in the 1440 time segments, which indicated that those related to the area under the S-wave (inteST, inteSm, inteS_mean, and inteS_sum) significantly differed in almost all fragments from 20:00 to 6:00 on the next day.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003eMultivariate logistic regression of ECG parameters\u003c/b\u003e\u003c/h2\u003e \u003cp\u003eTo explore the association of ECG parameters with SCDHR and HF, multivariate logistic regression models were constructed. We first excluded the redundant ECG parameters by clustering them and selecting one parameter from each cluster. Briefly, the ECG parameters in each time segment were clustered into eight clusters (Supplemental Fig.\u0026nbsp;5). From each cluster, one ECG parameter showing the most significant association with outcome was selected to represent the cluster. The features from each cluster were used to perform a multivariate logistic regression analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eECG parameters associated with SCDHR patients and HCs\u003c/h2\u003e \u003cp\u003eThe multivariate logistic regression model was trained at each time fragment in the training set to discriminate patients with SCDHR and HCs. To identify time-dependent ECG parameters associated with SCDHR, we combined the models into a unique regression model by the greedy feature selection approach and found that the combination of ten time fragments (16:06, 14:37, 21:39, 17:32, 17:28, 17:15, 14:18, 19:23, 13:50, and 9:31) and six ECG parameters (\"inteS_mean\", \"J_up\", \"t_T\", \"mean_R\", \"T_minus\", and \"t_RT\") yielded the best performance (AUC\u0026thinsp;=\u0026thinsp;0.887) to discriminate patients with SCDHR and HCs (Supplemental Fig.\u0026nbsp;6A). This model was further tested in an independent dataset, which achieved an AUC of 0.747 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe tested the associations of the \"inteS_mean\", \"J_up\", \"t_T\", \"mean_R\", \"T_minus\", and \"t_RT\" ECG parameters with SCDHR patients and HCs at ten time intervals using univariate logistic regression. The P value of the coefficient was adjusted by the Benjamini and Hochberg method.\u003csup\u003e29\u003c/sup\u003e As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, \u0026ldquo;inteS_mean\u0026rdquo;, \u0026ldquo;t_T\u0026rdquo;, \u0026ldquo;mean_R\u0026rdquo;, \u0026ldquo;T_minus\u0026rdquo;, and \u0026ldquo;t_RT\u0026rdquo; were significantly correlated with SCDHR in the training set. An increase in \u0026ldquo;inteS_mean\u0026rdquo; was positively correlated with SCDHR (OR\u0026thinsp;=\u0026thinsp;1.595\u0026thinsp;~\u0026thinsp;2.253, P-adjusted\u0026thinsp;\u0026lt;\u0026thinsp;0.050), and an increase in \u0026ldquo;t_T\u0026rdquo; was also positively correlated with SCDHR (OR\u0026thinsp;=\u0026thinsp;1.278, P-adjusted\u0026thinsp;=\u0026thinsp;0.034) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, Supplemental Table\u0026nbsp;4). In the independent test set, however, an increase in \u0026ldquo;inteS_mean\u0026rdquo; and \u0026ldquo;t_T\u0026rdquo; was positively but not significantly correlated with SCDHR (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). Nevertheless, \u0026ldquo;inteS_mean\u0026rdquo; in the independent test set significantly differed between the patients with SCDHR and HCs in the time fragments at 21:39 (P-adjusted\u0026thinsp;=\u0026thinsp;0.014), and \u0026ldquo;t_T\u0026rdquo; significantly differed between the patients with SCDHR and HCs at 19:23 (P-adjusted\u0026thinsp;=\u0026thinsp;0.035) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Thus, increases in \u0026ldquo;inteS_mean\u0026rdquo; at approximately 21:30 and \u0026ldquo;t_T\u0026rdquo; at approximately 19:30 are putatively associated with SCDHR.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eECG abnormalities between SCDHR and HF\u003c/h2\u003e \u003cp\u003eA multivariate logistic regression model of each time fragment was constructed to explore the difference between SCDHR and HF. The model was trained to select ECG parameters and was evaluated by fivefold cross-validations using the training set. When the models were combined into a unique model, the multivariate logistic regression model for discriminating SCDHR and HF achieved the best performance (AUC 0.857) after integrating five ECG parameters (\u0026ldquo;inteTM\u0026rdquo;, \u0026ldquo;inteSm\u0026rdquo;, \u0026ldquo;inteT_mean\u0026rdquo;, \u0026ldquo;t_T\u0026rdquo;, and \u0026ldquo;t_S_sum\u0026rdquo;) at three time fragments (10:56, 14:38, and 14:51) in the training set (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB and Supplemental Fig.\u0026nbsp;6B). When this model was tested in the independent set, it achieved an AUC of 0.714. The associations of each ECG parameter and the difference between SCDHR and HF in each time fragment were examined by univariable logistic regression. However, none of these parameters were independently or significantly correlated with SCDHR in HF (Supplemental Fig.\u0026nbsp;7A and 7B). We next tested the difference in these ECG parameters (\u0026ldquo;inteTM\u0026rdquo;, \u0026ldquo;inteSm\u0026rdquo;, \u0026ldquo;inteT_mean\u0026rdquo;, \u0026ldquo;t_T\u0026rdquo;, and \u0026ldquo;t_S_sum\u0026rdquo;) at the three time intervals (10:56, 14:38, and 14:51) and found that only \u0026ldquo;inteSm\u0026rdquo; significantly differed between SCDHR and HF at 14:38 in both the training set and the independent test set (Supplemental Fig.\u0026nbsp;7C).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eECG abnormalities between HF and HC\u003c/h2\u003e \u003cp\u003eA logistic regression model of ECG parameters in each time fragment was constructed to discriminate HF and HC. Fivefold cross-validation was used to select the ECG parameters achieving the best AUC in the training set. When the time segments were combined into a unique model, the best-performing model was obtained after combining seven time fragments (0:15, 1:06, 5:55, 23:33, 23:39, 23:40, and 23:46) and six ECG parameters (\u0026ldquo;inteR_sum\u0026rdquo;, \u0026ldquo;T_plus\u0026rdquo;, \u0026ldquo;inteSm\u0026rdquo;, \u0026ldquo;mean_R\u0026rdquo;, \u0026ldquo;inteS_mean\u0026rdquo;, and \u0026ldquo;inteST\u0026rdquo;), which achieved an AUC of 0.965 in the training set (Supplemental Fig.\u0026nbsp;6C). This model was tested in the independent test set and achieved an AUC of 0.842 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). The associations with HF were examined by logistic regression, which demonstrated that \u0026ldquo;inteSm\u0026rdquo;, \u0026ldquo;mean_R\u0026rdquo;, \u0026ldquo;inteS_mean\u0026rdquo;, and \u0026ldquo;inteST\u0026rdquo; were significantly correlated with HF at these seven time fragments in the training set (all P-adjusted\u0026thinsp;\u0026lt;\u0026thinsp;0.050) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). These correlations were further validated in the independent test set (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). The difference test results showed that only \u0026ldquo;inteSm\u0026rdquo; and \u0026ldquo;inteS_mean\u0026rdquo; significantly differed between HF and HC in all seven time fragments (0:15, 1:06, 5:55, 23:33, 23:39, 23:40, and 23:46) (all P-adjusted\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eMajor findings\u003c/h2\u003e \u003cp\u003eThe analysis of 24-hour ECG data of patients with SCDHR and HCs revealed several important findings. First, we discovered that the S-wave is sensitive to heart health and that the S-wave-related parameters significantly differed in the pairwise comparison of patients with SCDHR, patients with HF, and HCs. Especially in the comparisons of SCDHR patients and HCs as well as HF patients and HCs, the S-wave area-related parameters, inteSm and inteS_mean, were positively correlated with an increased risk of SCDHR and HF at many time points. Previous studies have demonstrated that the S-wave upslope duration ratio is related to Brugada syndrome,\u003csup\u003e30,31\u003c/sup\u003e and a prolonged S-wave is frequently observed in patients with SCD carrying \u003cem\u003eSCN5A\u003c/em\u003e mutations and patients with SCD with no heart disease history.\u003csup\u003e32\u003c/sup\u003e Thus, the present findings suggested that S-wave-related parameters, especially the S-wave area, are important for characterizing SCDHR.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003ePotential mechanism\u003c/h2\u003e \u003cp\u003eWe considered several mechanisms to explain our findings. First, S-wave areas may reflect an altered spatial sequence of ventricular depolarization. Various pathophysiological factors may account for increased S-wave areas, such as shortened action potential durations in ischaemic subepicardial myocyte layers.\u003csup\u003e33\u003c/sup\u003e These underlying conditions may indicate subclinical diseases and an elevated risk of SCD.\u003csup\u003e34\u003c/sup\u003e Second, a decreased S-wave area, partly resulting from a prolonged P-wave duration, may reflect an arrhythmogenic myocardial substrate in the atrium and ventricle. An increasing P-wave duration has been associated with the degree of interstitial left ventricular fibrosis on cardiac magnetic resonance imaging,\u003csup\u003e35\u003c/sup\u003e which may present with an increased S-wave area. Finally, the present study identified the influence of circadian rhythms on the occurrence of cardiovascular events. Circadian rhythms are important for regulating various physiological functions, such as heart rate and cardiovascular diseases.\u003csup\u003e36\u0026ndash;38\u003c/sup\u003e Previous studies have reported a morning surge of sympathetic nervous system activity resulting in physiological changes that produce myocardial ischaemia.\u003csup\u003e39,40\u003c/sup\u003e In the present study, we discuss how circadian rhythms are related to SCD and elaborate on the importance of monitoring ECG at particular time intervals.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eComparison with other studies\u003c/h2\u003e \u003cp\u003ePrevious studies have revealed associations between T-wave inversion\u003csup\u003e41\u003c/sup\u003e and the QRS-T angle\u003csup\u003e42\u003c/sup\u003e as well as between the T-wave peak-to-end and SCD.\u003csup\u003e43\u003c/sup\u003e The present study suggested that the T-wave may be important for discriminating SCDHR from HF. Additionally, we observed that the ECG parameters were less significantly different between patients with SCDHR and patients with HF than between patients with SCDHR and HCs. The underlying reason may be due to the similar clinical characteristics between SCD and SCDHR, requiring an enlarged sample size to identify the difference.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003elimitations of study\u003c/h2\u003e \u003cp\u003eOne limitation of this study was that the independent test was based on a small sample size because the prevalence of SCDHR is relatively low compared to that of HF and it is difficult to collect SCD samples.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eIn this prospective, multicentre cohort study, we found that time-dependent S-wave area-related ECG parameters in 24-hour ECG (inteS_mean, inteSm, and inteST) are potentially early predictive factors for SCD risk.\u003c/p\u003e"},{"header":"Summary boxes","content":"\u003ch4\u003e\u003cstrong\u003eWhat is already known on this topic\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003e1.Sudden cardiac death\u0026nbsp;is primarily a result of electrical disturbance of the normal cardiac rhythm.\u003c/p\u003e\n\u003cp\u003e2.Several ECG parameters are associated with\u0026nbsp;sudden cardiac death\u0026nbsp;risk.\u0026nbsp;\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eWhat this study adds\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003e1.S-wave areas in certain time intervals were significantly associated with sudden cardiac death.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.The application of these ECG parameters may be potentially useful for the early identification of patients with sudden cardiac death risk as independent predictive factors.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eRole of the funding source\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe work was partly funded by the National Key R\u0026amp;D Program of China (2020YFB0204803). This work was supported by the General Program of National Natural Science Foundation of China (NSFC) [Grants number: 82070237(JFW), 81870170 (JFW), 81970200 (YXC), 81770229 (YXC), 81970388 (YLZ), 81903299 (QC), 81801132 (HYZ) and 81971190 (HYZ)], the Guangzhou Health and Medical Collaborative Innovation Major Project [Grant number: 201803040010 (JFW)], Guangdong Provincial Laboratory of Regenerative Medicine and Health [Grant number: 198F041814 (JFW)], and the Natural Science Foundation of Guangdong [Grant number 2019A1515011682 (YLZ)].\u0026nbsp;The authors declare no relationships with industry.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no conflicts of interest to declare.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInstitutional Review Boards of all hospitals mentioned above and received proper ethical oversight.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData sharing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStudy data can be made available upon request to the corresponding author.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTransparency statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe lead author affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as originally planned (and, if relevant, registered) have been explained.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAziz EF, Javed F, Pratap B, Herzog E. Strategies for the prevention and treatment of sudden cardiac death. \u003cem\u003eOpen Access Emerg Med\u003c/em\u003e 2010;2010:99-114.\u003c/li\u003e\n\u003cli\u003eWong CX, Brown A, Lau DH, et al. Epidemiology of Sudden Cardiac Death: Global and Regional Perspectives. \u003cem\u003eHeart Lung Circ\u003c/em\u003e 2019;28:6-14.\u003c/li\u003e\n\u003cli\u003eWright RS, Anderson JL, Adams CD, et al. 2011 ACCF/AHA focused update of the Guidelines for the Management of Patients with Unstable Angina/Non-ST-Elevation Myocardial Infarction (updating the 2007 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines developed in collaboration with the American College of Emergency Physicians, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e 2011;57:1920-59.\u003c/li\u003e\n\u003cli\u003eMaron BJ, McKenna WJ, Danielson GK, et al. American College of Cardiology/European Society of Cardiology clinical expert consensus document on hypertrophic cardiomyopathy. A report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents and the European Society of Cardiology Committee for Practice Guidelines. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e 2003;42:1687-713.\u003c/li\u003e\n\u003cli\u003eWaks JW, Sitlani CM, Soliman EZ, et al. Global Electric Heterogeneity Risk Score for Prediction of Sudden Cardiac Death in the General Population: The Atherosclerosis Risk in Communities (ARIC) and Cardiovascular Health (CHS) Studies. \u003cem\u003eCirculation\u003c/em\u003e 2016;133:2222-34.\u003c/li\u003e\n\u003cli\u003eTerho HK, Tikkanen JT, Kentta TV, et al. Electrocardiogram as a predictor of sudden cardiac death in middle-aged subjects without a known cardiac disease. \u003cem\u003eInt J Cardiol Heart Vasc\u003c/em\u003e 2018;20:50-5.\u003c/li\u003e\n\u003cli\u003eTeodorescu C, Reinier K, Uy-Evanado A, et al. Prolonged QRS duration on the resting ECG is associated with sudden death risk in coronary disease, independent of prolonged ventricular repolarization. \u003cem\u003eHeart Rhythm\u003c/em\u003e 2011;8:1562-7.\u003c/li\u003e\n\u003cli\u003ePieroni M, Bellocci F, Crea F. Sudden cardiac arrest associated with early repolarization. \u003cem\u003eN Engl J Med\u003c/em\u003e 2008;359:761-2; author reply 2.\u003c/li\u003e\n\u003cli\u003eKim SH, Kim DY, Kim HJ, et al. Early repolarization with horizontal ST segment may be associated with aborted sudden cardiac arrest: a retrospective case control study. \u003cem\u003eBMC Cardiovasc Disord\u003c/em\u003e 2012;12:122.\u003c/li\u003e\n\u003cli\u003eRosso R, Kogan E, Belhassen B, et al. J-point elevation in survivors of primary ventricular fibrillation and matched control subjects: incidence and clinical significance. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e 2008;52:1231-8.\u003c/li\u003e\n\u003cli\u003eLambiase PD, Young WJ. Sudden cardiac death risk prediction - As easy as ECG? \u003cem\u003eInt J Cardiol\u003c/em\u003e 2018;276:152-3.\u003c/li\u003e\n\u003cli\u003ePerez-Riera AR, Barbosa-Barros R, Shenasa M. Electrocardiographic Markers of Sudden Cardiac Death (Including Left Ventricular Hypertrophy). \u003cem\u003eCard Electrophysiol Clin\u003c/em\u003e 2017;9:605-29.\u003c/li\u003e\n\u003cli\u003eAro AL, Reinier K, Rusinaru C, et al. Electrical risk score beyond the left ventricular ejection fraction: prediction of sudden cardiac death in the Oregon Sudden Unexpected Death Study and the Atherosclerosis Risk in Communities Study. \u003cem\u003eEur Heart J\u003c/em\u003e 2017;38:3017-25.\u003c/li\u003e\n\u003cli\u003eWellens HJ, Schwartz PJ, Lindemans FW, et al. Risk stratification for sudden cardiac death: current status and challenges for the future. \u003cem\u003eEur Heart J\u003c/em\u003e 2014;35:1642-51.\u003c/li\u003e\n\u003cli\u003ePerez-Alday EA, Li-Pershing Y, Bender A, et al. Importance of the heart vector origin point definition for an ECG analysis: The Atherosclerosis Risk in Communities (ARIC) study. \u003cem\u003eComput Biol Med\u003c/em\u003e 2018;104:127-38.\u003c/li\u003e\n\u003cli\u003eLyon A, Ariga R, Minchole A, et al. Distinct ECG Phenotypes Identified in Hypertrophic Cardiomyopathy Using Machine Learning Associate With Arrhythmic Risk Markers. \u003cem\u003eFront Physiol\u003c/em\u003e 2018;9:213.\u003c/li\u003e\n\u003cli\u003eHolkeri A, Eranti A, Haukilahti MAE, et al. Predicting sudden cardiac death in a general population using an electrocardiographic risk score. \u003cem\u003eHeart\u003c/em\u003e 2020;106:427-33.\u003c/li\u003e\n\u003cli\u003eAdabag AS, Luepker RV, Roger VL, Gersh BJ. Sudden cardiac death: epidemiology and risk factors. \u003cem\u003eNat Rev Cardiol\u003c/em\u003e 2010;7:216-25.\u003c/li\u003e\n\u003cli\u003eAkhtar M, Elliott PM. Risk Stratification for Sudden Cardiac Death in Non-Ischaemic Dilated Cardiomyopathy. \u003cem\u003eCurr Cardiol Rep\u003c/em\u003e 2019;21:155.\u003c/li\u003e\n\u003cli\u003eWillich SN, Levy D, Rocco MB, Tofler GH, Stone PH, Muller JE. Circadian variation in the incidence of sudden cardiac death in the Framingham Heart Study population. \u003cem\u003eAm J Cardiol\u003c/em\u003e 1987;60:801-6.\u003c/li\u003e\n\u003cli\u003eLevine RL, Pepe PE, Fromm RE, Jr., Curka PA, Clark PA. Prospective evidence of a circadian rhythm for out-of-hospital cardiac arrests. \u003cem\u003eJAMA\u003c/em\u003e 1992;267:2935-7.\u003c/li\u003e\n\u003cli\u003eBagai A, McNally BF, Al-Khatib SM, et al. Temporal differences in out-of-hospital cardiac arrest incidence and survival. \u003cem\u003eCirculation\u003c/em\u003e 2013;128:2595-602.\u003c/li\u003e\n\u003cli\u003eFukuda T, Ohashi-Fukuda N, Kondo Y, Sera T, Doi K, Yahagi N. Epidemiology, Risk Factors, and Outcomes of Out-of-Hospital Cardiac Arrest Caused by Stroke: A Population-Based Study. \u003cem\u003eMedicine (Baltimore)\u003c/em\u003e 2016;95:e3107.\u003c/li\u003e\n\u003cli\u003eEnglund A, Behrens S, Wegscheider K, Rowland E. Circadian variation of malignant ventricular arrhythmias in patients with ischemic and nonischemic heart disease after cardioverter defibrillator implantation. European 7219 Jewel Investigators. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e 1999;34:1560-8.\u003c/li\u003e\n\u003cli\u003ePeters RW, McQuillan S, Gold MR. Interaction of septadian and circadian rhythms in life-threatening ventricular arrhythmias in patients with implantable cardioverter-defibrillators. \u003cem\u003eAm J Cardiol\u003c/em\u003e 1999;84:555-7.\u003c/li\u003e\n\u003cli\u003eRuwald MH, Moss AJ, Zareba W, et al. Circadian distribution of ventricular tachyarrhythmias and association with mortality in the MADIT-CRT trial. \u003cem\u003eJ Cardiovasc Electrophysiol\u003c/em\u003e 2015;26:291-9.\u003c/li\u003e\n\u003cli\u003eRamireddy A, Chugh HS, Reinier K, et al. Sudden cardiac death during nighttime hours. \u003cem\u003eHeart Rhythm\u003c/em\u003e 2021;18:778-84.\u003c/li\u003e\n\u003cli\u003eHartigan JA, Wong MA. Algorithm AS 136: A K-Means Clustering Algorithm. \u003cem\u003eAppl Stat\u003c/em\u003e 1979;28:100-8.\u003c/li\u003e\n\u003cli\u003eHaynes W. Benjamini\u0026ndash;Hochberg Method. In: Dubitzky W, Wolkenhauer O, Cho K-H, Yokota H, editors. Encyclopedia of Systems Biology. New York, NY: Springer New York; 2013. p. 78.\u003c/li\u003e\n\u003cli\u003eSubramanian M, Prabhu MA, Rai M, et al. A novel prediction model for risk stratification in patients with a type 1 Brugada ECG pattern. \u003cem\u003eJ Electrocardiol\u003c/em\u003e 2019;55:65-71.\u003c/li\u003e\n\u003cli\u003eMichowitz Y, Milman A, Andorin A, et al. Characterization and Management of Arrhythmic Events in Young Patients With Brugada Syndrome. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e 2019;73:1756-65.\u003c/li\u003e\n\u003cli\u003eVanninen SUM, Nikus K, Aalto-Setala K. Electrocardiogram changes and atrial arrhythmias in individuals carrying sodium channel SCN5A D1275N mutation. \u003cem\u003eAnn Med\u003c/em\u003e 2017;49:496-503.\u003c/li\u003e\n\u003cli\u003eRautaharju PM, Zhou SH, Gregg RE, Startt-Selvester RH. Heart rate, gender differences, and presence versus absence of diagnostic ST elevation as determinants of spatial QRS|T angle widening in acute coronary syndrome. \u003cem\u003eAm J Cardiol\u003c/em\u003e 2011;107:1744-50.\u003c/li\u003e\n\u003cli\u003eWhang W, Shimbo D, Levitan EB, et al. Relations between QRS| T angle, cardiac risk factors, and mortality in the third National Health and Nutrition Examination Survey (NHANES III). \u003cem\u003eAm J Cardiol\u003c/em\u003e 2012;109:981-7.\u003c/li\u003e\n\u003cli\u003eWin TT, Venkatesh BA, Volpe GJ, et al. Associations of electrocardiographic P-wave characteristics with left atrial function, and diffuse left ventricular fibrosis defined by cardiac magnetic resonance: The PRIMERI Study. \u003cem\u003eHeart Rhythm\u003c/em\u003e 2015;12:155-62.\u003c/li\u003e\n\u003cli\u003eKollias GE, Stamatelopoulos KS, Papaioannou TG, et al. Diurnal variation of endothelial function and arterial stiffness in hypertension. \u003cem\u003eJ Hum Hypertens\u003c/em\u003e 2009;23:597-604.\u003c/li\u003e\n\u003cli\u003eDegaute JP, van de Borne P, Linkowski P, Van Cauter E. Quantitative analysis of the 24-hour blood pressure and heart rate patterns in young men. \u003cem\u003eHypertension\u003c/em\u003e 1991;18:199-210.\u003c/li\u003e\n\u003cli\u003eCrnko S, Du Pr\u0026eacute; BC, Sluijter JP, Van Laake LW. Circadian rhythms and the molecular clock in cardiovascular biology and disease. \u003cem\u003eNat Rev Cardiol\u003c/em\u003e 2019;16:437-47.\u003c/li\u003e\n\u003cli\u003eMuller JE, Tofler GH, Stone PH. Circadian variation and triggers of onset of acute cardiovascular disease. \u003cem\u003eCirculation\u003c/em\u003e 1989;79:733-43.\u003c/li\u003e\n\u003cli\u003eWillich S, Maclure M, Mittleman M, Arntz H-R, Muller J. Sudden cardiac death. Support for a role of triggering in causation. \u003cem\u003eCirculation\u003c/em\u003e 1993;87:1442-50.\u003c/li\u003e\n\u003cli\u003eRautaharju PM, Surawicz B, Gettes LS. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: part IV: the ST segment, T and U waves, and the QT interval a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society Endorsed by the International Society for Computerized Electrocardiology. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e 2009;53:982-91.\u003c/li\u003e\n\u003cli\u003eAro AL, Huikuri HV, Tikkanen JT, et al. QRS-T angle as a predictor of sudden cardiac death in a middle-aged general population. \u003cem\u003eEuropace\u003c/em\u003e 2012;14:872-6.\u003c/li\u003e\n\u003cli\u003ePanikkath R, Reinier K, Uy-Evanado A, et al. Prolonged Tpeak-to-tend interval on the resting ECG is associated with increased risk of sudden cardiac death. \u003cem\u003eCirc Arrhythmia Electrophysiol\u003c/em\u003e 2011;4:441-7.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-3490411/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3490411/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjectives\u003c/h2\u003e \u003cp\u003eThe present study aimed to detect novel and time-dependent ECG parameters by analysing 24-h ECG data, especially the area under ECG waves.\u003c/p\u003e\u003ch2\u003eDesign:\u003c/h2\u003e \u003cp\u003eProspective, multicentre cohort study\u003c/p\u003e\u003ch2\u003eSetting:\u003c/h2\u003e \u003cp\u003eFour hospitals in China\u003c/p\u003e\u003ch2\u003eParticipants:\u003c/h2\u003e \u003cp\u003eHigh risk of sudden cardiac death, including 43 survivors of sudden cardiac death (SCD) or patients who suffered haemodynamic disorder due to sustained ventricular tachycardia/ventricular fibrillation (SCDHR group), 138 patients with HF who did not experience sustained ventricular tachycardia/ventricular fibrillation but were diagnosed with dilated cardiomyopathy or ischaemic cardiomyopathy with LVEF\u0026thinsp;\u0026le;\u0026thinsp;35% (HF group), and 108 healthy controls who presented with no heart disease (HC group).\u003c/p\u003e\u003ch2\u003eExposure:\u003c/h2\u003e \u003cp\u003eTime-dependent ECG parameters by analysing 24-h ECG data\u003c/p\u003e\u003ch2\u003eMain outcome measures:\u003c/h2\u003e \u003cp\u003eThe area under ECG waves was separately analysed to determine their associations with SCDHR and HF in the test set and was further examined in the validation set. Logistic regression analyses were performed.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe multivariate logistic regression model for discriminating SCDHR patients and HCs indicated that the average area under the S-wave (inteS_mean) at 16:00\u0026ndash;21:00 was positively associated with SCDHR (OR\u0026thinsp;\u0026gt;\u0026thinsp;1, P-adjust\u0026thinsp;\u0026lt;\u0026thinsp;0.050) and significantly (P value\u0026thinsp;=\u0026thinsp;0.014) differed at 21:39 in the validation set. Similarly, the model for discriminating HF and HC indicated that the inteS_mean, minimum S-wave area (inteSm), and difference in S-wave and T-wave (inteST) were positively (OR\u0026thinsp;\u0026gt;\u0026thinsp;1, P-adjust\u0026thinsp;\u0026lt;\u0026thinsp;0.050) associated with HF in both the test set and validation set.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThe time-dependent S-wave area-related ECG parameters (inteS_mean, inteSm, and inteST) are potentially early predictive factors for SCD risk.\u003c/p\u003e\u003ch2\u003eTrial registration:\u003c/h2\u003e \u003cp\u003eThis study was registered on the website of \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://register.clinicaltrails.gov/Organization\u003c/span\u003e\u003cspan address=\"http://register.clinicaltrails.gov/Organization\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The Clinical Trials ID is NCT03485079.\u003c/p\u003e","manuscriptTitle":"Time-dependent S-wave areas predict sudden cardiac death risk: a prospective, multicentre registered study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-02-09 17:05:19","doi":"10.21203/rs.3.rs-3490411/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"66254022-4a55-4d54-a4b7-d3990ece5ba5","owner":[],"postedDate":"February 9th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":28610677,"name":"Biological sciences/Biological techniques/Electrophysiology/Electrocardiography \u0026#x2013; EKG"},{"id":28610678,"name":"Health sciences/Cardiology/Cardiovascular biology/Cardiovascular diseases/Arrhythmias"}],"tags":[],"updatedAt":"2024-02-22T15:21:32+00:00","versionOfRecord":[],"versionCreatedAt":"2024-02-09 17:05:19","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3490411","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3490411","identity":"rs-3490411","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}