Author's Accepted Manuscript
A prospective study of supraventricular activity and incidence of atrial fibrillation Linda S.B. Johnson M.D., Tord Juhlin M.D., Ph.D., Steen Juul-Möller M.D.,Ph.D., Bo Hedblad M.D., Ph. D., Peter M. Nilsson M.D., Ph.D., Gunnar Engström M.D., Ph.D.
www.elsevier.com/locate/buildenv
PII: DOI: Reference:
S1547-5271(15)00552-4 http://dx.doi.org/10.1016/j.hrthm.2015.04.042 HRTHM6259
To appear in:
Heart Rhythm
Cite this article as: Linda S.B. Johnson M.D., Tord Juhlin M.D., Ph.D., Steen JuulMöller M.D.,Ph.D., Bo Hedblad M.D., Ph.D., Peter M. Nilsson M.D., Ph.D., Gunnar Engström M.D., Ph.D., A prospective study of supraventricular activity and incidence of atrial fibrillation, Heart Rhythm, http://dx.doi.org/10.1016/j.hrthm.2015.04.042 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A prospective study of supraventricular activity and incidence of atrial fibrillation Authors: Linda S.B. Johnson, M.D.; TordJuhlin, M.D., Ph.D.; Steen Juul-Möller, M.D.,Ph.D.; Bo Hedblad, M.D., Ph.D.; Peter M Nilsson, M.D., Ph.D.;Gunnar Engström, M.D., Ph.D. Institution: All authors are at the Department of Clinical Sciences, Lund University, Skåne University Hospital, S-20502 Malmö, Sweden Number of words: 250 in abstract and 4895 in full text and with 30 references, 2 figures and 3 tables. Short title: Supraventricular activity and AF Conflict of interest: None Corresponding author: Linda Johnson email:
[email protected]
postal adress: Inga-Marie Nilssons väg 49 20502 Malmö Sweden tel:
+46-40 338681
Abstract Background: Atrial fibrillation (AF)episodes are thought to be startedby an electrical trigger reaching a susceptible atria. Such a trigger could be present long before the occurrence of sustained symptomaticarrhythmia. Objective: We sought to determine if supraventricular extrasystoles (SVES) and supraventricular tachycardias (SVT) measured at 24 h Holter ECG were associated with increased incidence of AF.
1
Methods: In 1998-2000, 389 individuals (44% men, mean age 65 years) were examined using 24 h Holter ECG. Six individuals withknown prevalent AF were excluded. After a mean follow-up of 10.3 years there were 45 cases of incident AF. Hazard ratios (HR) were computed using multivariable Cox regression adjusting for age, gender, systolic blood pressure, height, weight, smoking and HOMA-IR (homeostatic model assessment of insulin resistance). Results: Frequency of SVES as well as SVT episodes per hour were independent predictors of incident AF, HR per log unit 1.38, 95% confidence interval (CI) 1.14-1.68, p= 0.001; and HR 1.95, 95% CI 1.21-3.13, p=0.006, respectively. Further adjustment for education level, alcohol use, use of medication and physical activity did not substantially alter the results, nor did analysis using competing risks regression accounting for a competing risk of death. The maximum duration of SVT or the heart rate atSVT was not significantly associated with incidence of AF. Conclusion: SVES and SVT independently predict AF. The prognostic significance was similar for SVES, SVT and a combination of the two. Repeated efforts to detect AF could be of merit in individuals with frequent supraventricular activity.
Key words: Atrial fibrillation; supraventricular extra systoles; premature atrial contractions; supraventricular tachycardia; population.
2
Abbreviations: AF = Atrial Fibrillation SVES= Supraventricular extrasystoles ECG= Electrocardiogram SVT= Supraventricular tachycardia ESVEA= Excessive supraventricular ectopic activity MDCS= Malmö Diet and Cancer Study HOMA-IR= Homeostatic model assessment of insulin resistance CAD= Coronary artery disease HF= Heart failure SBP= Systolic blood pressure SD= Standard deviation HR= Hazard ratio SHR= Subhazard ratio CHS= Cardiovascular Health Study
3
Introduction
Atrial fibrillation (AF) is a common arrhythmia, especially among the elderly,1 and leads to increased mortality and morbidity. Known risk factors include age, male gender, height and weight, elevated systolic blood pressure, and other cardiac and pulmonary diseases as well as life-style factors such as smoking, and low physical activity. 2-10 AF is thought to be caused by an electrical trigger reaching a receptive substrate.11 Such a trigger could arise from the pulmonary veins, or ectopic foci in the atria, and be present in an individual for a long time before the initiation of fibrillation. The natural history of AF, with short and often asymptomatic episodes gradually progressing to longer periods of sustained, and eventually permanent AF,12, 13 lends some support to this idea. An eventual trigger could thus have a predictive value, and perhaps explain some of the attributable risk of AF that is currently unaccounted for. 3, 10, 14
Supraventricular extrasystoles (SVES) are a common finding at electrocardiograms (ECG’s), 15 and have previously been considered harmless. Recent studies have linked increased supraventricular activity to an increased risk of atrial fibrillation, stroke and cardiovascular death,16-20 and in light of this, further studies on the subject arewarranted. SVES occur as single premature beats, but also as short and often asymptomatic supraventricular tachycardias (SVT). Previous studies have shown that SVES is a predictor for AF, and that the addition of SVES improves the risk discrimination of the Framingham risk equation model,19. Furthermore, individuals with more than 30 SVES per hour or runs of consecutive SVES lasting 20 beats or longer, termed excessive supraventricular ectopic activity (ESVEA), 17have been shown to be at higher risk of hospitalization for AF, stroke and
4
death. The association between SVT at 24 h ECG and AF has to our knowledge not been fully studied previously, nor has it been determined which, if any, of the above measurements of increased ectopic supraventricular activity has the best predictive power.
The Malmö Diet and Cancer cohort is a large prospective cohort consisting of individuals from the city of Malmö in southern Sweden. Out of this cohort 389 randomly selected individuals (aged 53 to 74 years, 45% men) underwent 24 h ECG monitoring at a follow-up investigation. We sought to determine if SVES and SVT were independent predictors of atrial fibrillation in this sub-study, as well as to examine whether there could be predictive value in the duration of SVT as well as the rate of SVT. We also aimed to determine which measure of ectopic supraventricular activity had the best predictive power.
Methods
Study population The current study population is derived from the larger Malmö Diet and Cancer Study (MDCS), which included 30,447individuals (aged 44 to 73 years, 40 % men) examined in 1991-1996. The study has been described in more detail elsewhere.10From this population a cardiovascular study was conducted in a random subsample of n=6103 men and women. A subpopulation was randomly invitedfrom the cardiovascular study based on HOMA-IR (homeostatic model assessment of insulin resistance) for further follow up studies, in the years 1999-2000. HOMA-IR was calculated as fasting insulin (mIU/L) times fasting blood glucose (mmol/L) divided by 22.5.21High HOMA-IR (defined as the sex-specific 75th
5
percentile, i.e. 1.80 for women and 2.12 for men) was slightly oversampled (15% from each of quartile 1 and 2, 30% from quartile 3 and 40% from quartile 4).22, 23 Out of these individuals, 389 individuals were randomly selected for 24 h ECG monitoring. Six of these subjects had a history of AF and were excluded from analysis. Thus, the study population thus consists of 383individuals (aged 53 to 74 years, 45% men).The derivation of the study selection is described in Fig 1.
Data collection At the MDCS baseline and follow-up investigations, individuals underwent physical examination and blood-sampling after overnight fasting. Blood-samples were analysed using standard laboratory procedures at the Malmö University Hospital. A questionnaire was administered at baseline and follow-up; from this smoking status, physical activity score, education grade (classed as 1 year) and alcohol use were derived. Low education was defined as 9 years of schooling or less. Alcohol use was self-reported in the context of a 7-day dietary registration. Physical activity score was calculated through a modification of the Minnesota Leisure Time Physical Activity Questionnaire 24 and has been described previously.25 Individuals in the lowest, sex-specific quartile of physical activity were defined as having a low physical activity. The variables alcohol use and physical activity are derived from the baseline questionnaire. Height and weight were measured standing in light indoor clothing and without shoes. Blood pressure was measured after 10 min supine rest using a modifiable cuff sphygmomanometer.
6
All patients underwent a 24 hour 3-lead (X,Y,Z coupling) Holter ECG recording, with 256 Hz sampling rate, for arrhythmia detection. The equipment used was Spacelabs Healthcare Lifecard CF 12 bits Digital ECG recorder. Findings were classed using the Pathfinder SL analysis tool. All arrhythmias were confirmed visually before they were included in the Holter report. Mean analysis time was 22.9 hours, with a standard deviation of 2.9 hours. SVES was defined as a single supraventricular extra systole, preceded and followed by a normal sinus beat. Any period with three or more consecutive SVES was defined as a SVT. The term ESVEA has previously been described by Binici et al; 17 individuals with either more than 30 SVES/h or any SVT lasting for more than 20 consecutive beats, or both, were classed as having ESVEA.
Endpoint retrieval from Local and National Registries The endpoint was clinicalAF diagnosed in in-patients or out-patients in hospital ward. The Swedish Registers for in- and outpatients, which is administered by The Swedish National Board of Health and Welfare were used for case retrieval. The AF diagnosis in MDCS has recently been validated and found to be of high quality. 10 Participants were followed until first episode of AF (diagnosis codes 427D for the 9th revision of International Classification of Diseases, ICD-9, and I48 for the 10th revision, ICD-10), or until censoring by death or emigration from Sweden. Follow-up ended at 31st of December 2010. AF and atrial flutter have not been distinguished due to the similarities of these diagnoses. 26 The regional ethics review board has approved the study.
7
Statistics
All data was analysed using Stata for windows version 12.1. Continuous variables were assessed for skewness and all skewed variables (24 h ECG measurements, alcohol use and HOMA-IR) were log-transformed with the natural logarithm after the addition of a small constant (one) to the 24 h ECG variables, due to many subjects without SVES. In addition to this the number of SVES per 24 h was analysed as quartiles.
The association between 24 h ECG variables and risk of AF was analysed using both Cox regression and competing risks regression as described by Fine and Gray, 27 in order to assess the influence of the competing risk of death. The proportional hazards assumption was checked using –log-log plots as well as plots of expected and predicted failures. Three models were used; Model 1 was adjusted for age and gender. Model 2 included also adjustment for systolic blood pressure, height, weight, current smoking and HOMA-IR. Finally, Model 3 included further adjustment for physical activity score, education level and alcohol consumption. In order to determine whether any association between ectopic supraventricular activity and AF was influenced by coronary artery disease (CAD) or heart failure (HF) we also performed a sub-analysis of only those individuals in whom these diseases were not diagnosed at screening, and with censoring of individuals at the time of incident CAD or HF, if this happened before a diagnosis of AF.
The predictive power of SVES, SVT and ESVEA was analysed using both likelihood ratio test and Harrell’s C-statistic with 95% confidence interval.
8
Results Mean age at baseline ± standard deviation (SD) was 64.6± 5.9 years, and the median age 65.4 years. Fifty-five percent of the study population was female. Mean systolic blood pressure (SBP) ± SD was 143.7±18.5 mmHg and 23 percent of the population were smokers at the time of investigation. Baseline characteristics are otherwise reported in Table 1. Mean follow-up was 10.3 years. During that time 45 individuals developed AF. Those who received a diagnosis of AF did so after an average follow-up of 7.1 years, at a mean age of 75.6 ±4.5 years.
Results from the Cox regression models assessing the association between 24h ECG parameters and risk of incident AF are given in Table 2 and the results from the competing risks models are given in Table 3. Low physical activity, alcohol consumption and low education level were not significantly associated with incident AF and this in consideration with the study size led us to consider Model 2 the most appropriate model. The number of SVES/24 h was a statistically significant predictor of incident AF; hazard ratio (HR) 1.38, 95% CI 1.14-1.68, p=0.001 per log unit, after adjustment for Model 2 covariates. There was no substantial difference between results of competing risks regression analysis; subhazard ratio (SHR) 1.39, 95% CI 1.16-1.68, p< 0.0001 per log unit. Unadjusted Kaplan-Meier curves representing the risk of incident AF by quartiles of SVES are presented in Fig 2. The p-value for trend across quartiles was 0.015.
The number of SVT/24 h was likewise associated with increased risk of AF; HR 1.95, 1.21-3.13, p=0.006 per log unit after adjustment for Model 2 covariates using Cox regression, and SHR 1.99, 95% CI 1.40-2.82, p
A prospective study of supraventricular activity and incidence of atrial fibrillation Linda S.B. Johnson M.D., Tord Juhlin M.D., Ph.D., Steen Juul-Möller M.D.,Ph.D., Bo Hedblad M.D., Ph. D., Peter M. Nilsson M.D., Ph.D., Gunnar Engström M.D., Ph.D.
www.elsevier.com/locate/buildenv
PII: DOI: Reference:
S1547-5271(15)00552-4 http://dx.doi.org/10.1016/j.hrthm.2015.04.042 HRTHM6259
To appear in:
Heart Rhythm
Cite this article as: Linda S.B. Johnson M.D., Tord Juhlin M.D., Ph.D., Steen JuulMöller M.D.,Ph.D., Bo Hedblad M.D., Ph.D., Peter M. Nilsson M.D., Ph.D., Gunnar Engström M.D., Ph.D., A prospective study of supraventricular activity and incidence of atrial fibrillation, Heart Rhythm, http://dx.doi.org/10.1016/j.hrthm.2015.04.042 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A prospective study of supraventricular activity and incidence of atrial fibrillation Authors: Linda S.B. Johnson, M.D.; TordJuhlin, M.D., Ph.D.; Steen Juul-Möller, M.D.,Ph.D.; Bo Hedblad, M.D., Ph.D.; Peter M Nilsson, M.D., Ph.D.;Gunnar Engström, M.D., Ph.D. Institution: All authors are at the Department of Clinical Sciences, Lund University, Skåne University Hospital, S-20502 Malmö, Sweden Number of words: 250 in abstract and 4895 in full text and with 30 references, 2 figures and 3 tables. Short title: Supraventricular activity and AF Conflict of interest: None Corresponding author: Linda Johnson email:
[email protected]
postal adress: Inga-Marie Nilssons väg 49 20502 Malmö Sweden tel:
+46-40 338681
Abstract Background: Atrial fibrillation (AF)episodes are thought to be startedby an electrical trigger reaching a susceptible atria. Such a trigger could be present long before the occurrence of sustained symptomaticarrhythmia. Objective: We sought to determine if supraventricular extrasystoles (SVES) and supraventricular tachycardias (SVT) measured at 24 h Holter ECG were associated with increased incidence of AF.
1
Methods: In 1998-2000, 389 individuals (44% men, mean age 65 years) were examined using 24 h Holter ECG. Six individuals withknown prevalent AF were excluded. After a mean follow-up of 10.3 years there were 45 cases of incident AF. Hazard ratios (HR) were computed using multivariable Cox regression adjusting for age, gender, systolic blood pressure, height, weight, smoking and HOMA-IR (homeostatic model assessment of insulin resistance). Results: Frequency of SVES as well as SVT episodes per hour were independent predictors of incident AF, HR per log unit 1.38, 95% confidence interval (CI) 1.14-1.68, p= 0.001; and HR 1.95, 95% CI 1.21-3.13, p=0.006, respectively. Further adjustment for education level, alcohol use, use of medication and physical activity did not substantially alter the results, nor did analysis using competing risks regression accounting for a competing risk of death. The maximum duration of SVT or the heart rate atSVT was not significantly associated with incidence of AF. Conclusion: SVES and SVT independently predict AF. The prognostic significance was similar for SVES, SVT and a combination of the two. Repeated efforts to detect AF could be of merit in individuals with frequent supraventricular activity.
Key words: Atrial fibrillation; supraventricular extra systoles; premature atrial contractions; supraventricular tachycardia; population.
2
Abbreviations: AF = Atrial Fibrillation SVES= Supraventricular extrasystoles ECG= Electrocardiogram SVT= Supraventricular tachycardia ESVEA= Excessive supraventricular ectopic activity MDCS= Malmö Diet and Cancer Study HOMA-IR= Homeostatic model assessment of insulin resistance CAD= Coronary artery disease HF= Heart failure SBP= Systolic blood pressure SD= Standard deviation HR= Hazard ratio SHR= Subhazard ratio CHS= Cardiovascular Health Study
3
Introduction
Atrial fibrillation (AF) is a common arrhythmia, especially among the elderly,1 and leads to increased mortality and morbidity. Known risk factors include age, male gender, height and weight, elevated systolic blood pressure, and other cardiac and pulmonary diseases as well as life-style factors such as smoking, and low physical activity. 2-10 AF is thought to be caused by an electrical trigger reaching a receptive substrate.11 Such a trigger could arise from the pulmonary veins, or ectopic foci in the atria, and be present in an individual for a long time before the initiation of fibrillation. The natural history of AF, with short and often asymptomatic episodes gradually progressing to longer periods of sustained, and eventually permanent AF,12, 13 lends some support to this idea. An eventual trigger could thus have a predictive value, and perhaps explain some of the attributable risk of AF that is currently unaccounted for. 3, 10, 14
Supraventricular extrasystoles (SVES) are a common finding at electrocardiograms (ECG’s), 15 and have previously been considered harmless. Recent studies have linked increased supraventricular activity to an increased risk of atrial fibrillation, stroke and cardiovascular death,16-20 and in light of this, further studies on the subject arewarranted. SVES occur as single premature beats, but also as short and often asymptomatic supraventricular tachycardias (SVT). Previous studies have shown that SVES is a predictor for AF, and that the addition of SVES improves the risk discrimination of the Framingham risk equation model,19. Furthermore, individuals with more than 30 SVES per hour or runs of consecutive SVES lasting 20 beats or longer, termed excessive supraventricular ectopic activity (ESVEA), 17have been shown to be at higher risk of hospitalization for AF, stroke and
4
death. The association between SVT at 24 h ECG and AF has to our knowledge not been fully studied previously, nor has it been determined which, if any, of the above measurements of increased ectopic supraventricular activity has the best predictive power.
The Malmö Diet and Cancer cohort is a large prospective cohort consisting of individuals from the city of Malmö in southern Sweden. Out of this cohort 389 randomly selected individuals (aged 53 to 74 years, 45% men) underwent 24 h ECG monitoring at a follow-up investigation. We sought to determine if SVES and SVT were independent predictors of atrial fibrillation in this sub-study, as well as to examine whether there could be predictive value in the duration of SVT as well as the rate of SVT. We also aimed to determine which measure of ectopic supraventricular activity had the best predictive power.
Methods
Study population The current study population is derived from the larger Malmö Diet and Cancer Study (MDCS), which included 30,447individuals (aged 44 to 73 years, 40 % men) examined in 1991-1996. The study has been described in more detail elsewhere.10From this population a cardiovascular study was conducted in a random subsample of n=6103 men and women. A subpopulation was randomly invitedfrom the cardiovascular study based on HOMA-IR (homeostatic model assessment of insulin resistance) for further follow up studies, in the years 1999-2000. HOMA-IR was calculated as fasting insulin (mIU/L) times fasting blood glucose (mmol/L) divided by 22.5.21High HOMA-IR (defined as the sex-specific 75th
5
percentile, i.e. 1.80 for women and 2.12 for men) was slightly oversampled (15% from each of quartile 1 and 2, 30% from quartile 3 and 40% from quartile 4).22, 23 Out of these individuals, 389 individuals were randomly selected for 24 h ECG monitoring. Six of these subjects had a history of AF and were excluded from analysis. Thus, the study population thus consists of 383individuals (aged 53 to 74 years, 45% men).The derivation of the study selection is described in Fig 1.
Data collection At the MDCS baseline and follow-up investigations, individuals underwent physical examination and blood-sampling after overnight fasting. Blood-samples were analysed using standard laboratory procedures at the Malmö University Hospital. A questionnaire was administered at baseline and follow-up; from this smoking status, physical activity score, education grade (classed as 1 year) and alcohol use were derived. Low education was defined as 9 years of schooling or less. Alcohol use was self-reported in the context of a 7-day dietary registration. Physical activity score was calculated through a modification of the Minnesota Leisure Time Physical Activity Questionnaire 24 and has been described previously.25 Individuals in the lowest, sex-specific quartile of physical activity were defined as having a low physical activity. The variables alcohol use and physical activity are derived from the baseline questionnaire. Height and weight were measured standing in light indoor clothing and without shoes. Blood pressure was measured after 10 min supine rest using a modifiable cuff sphygmomanometer.
6
All patients underwent a 24 hour 3-lead (X,Y,Z coupling) Holter ECG recording, with 256 Hz sampling rate, for arrhythmia detection. The equipment used was Spacelabs Healthcare Lifecard CF 12 bits Digital ECG recorder. Findings were classed using the Pathfinder SL analysis tool. All arrhythmias were confirmed visually before they were included in the Holter report. Mean analysis time was 22.9 hours, with a standard deviation of 2.9 hours. SVES was defined as a single supraventricular extra systole, preceded and followed by a normal sinus beat. Any period with three or more consecutive SVES was defined as a SVT. The term ESVEA has previously been described by Binici et al; 17 individuals with either more than 30 SVES/h or any SVT lasting for more than 20 consecutive beats, or both, were classed as having ESVEA.
Endpoint retrieval from Local and National Registries The endpoint was clinicalAF diagnosed in in-patients or out-patients in hospital ward. The Swedish Registers for in- and outpatients, which is administered by The Swedish National Board of Health and Welfare were used for case retrieval. The AF diagnosis in MDCS has recently been validated and found to be of high quality. 10 Participants were followed until first episode of AF (diagnosis codes 427D for the 9th revision of International Classification of Diseases, ICD-9, and I48 for the 10th revision, ICD-10), or until censoring by death or emigration from Sweden. Follow-up ended at 31st of December 2010. AF and atrial flutter have not been distinguished due to the similarities of these diagnoses. 26 The regional ethics review board has approved the study.
7
Statistics
All data was analysed using Stata for windows version 12.1. Continuous variables were assessed for skewness and all skewed variables (24 h ECG measurements, alcohol use and HOMA-IR) were log-transformed with the natural logarithm after the addition of a small constant (one) to the 24 h ECG variables, due to many subjects without SVES. In addition to this the number of SVES per 24 h was analysed as quartiles.
The association between 24 h ECG variables and risk of AF was analysed using both Cox regression and competing risks regression as described by Fine and Gray, 27 in order to assess the influence of the competing risk of death. The proportional hazards assumption was checked using –log-log plots as well as plots of expected and predicted failures. Three models were used; Model 1 was adjusted for age and gender. Model 2 included also adjustment for systolic blood pressure, height, weight, current smoking and HOMA-IR. Finally, Model 3 included further adjustment for physical activity score, education level and alcohol consumption. In order to determine whether any association between ectopic supraventricular activity and AF was influenced by coronary artery disease (CAD) or heart failure (HF) we also performed a sub-analysis of only those individuals in whom these diseases were not diagnosed at screening, and with censoring of individuals at the time of incident CAD or HF, if this happened before a diagnosis of AF.
The predictive power of SVES, SVT and ESVEA was analysed using both likelihood ratio test and Harrell’s C-statistic with 95% confidence interval.
8
Results Mean age at baseline ± standard deviation (SD) was 64.6± 5.9 years, and the median age 65.4 years. Fifty-five percent of the study population was female. Mean systolic blood pressure (SBP) ± SD was 143.7±18.5 mmHg and 23 percent of the population were smokers at the time of investigation. Baseline characteristics are otherwise reported in Table 1. Mean follow-up was 10.3 years. During that time 45 individuals developed AF. Those who received a diagnosis of AF did so after an average follow-up of 7.1 years, at a mean age of 75.6 ±4.5 years.
Results from the Cox regression models assessing the association between 24h ECG parameters and risk of incident AF are given in Table 2 and the results from the competing risks models are given in Table 3. Low physical activity, alcohol consumption and low education level were not significantly associated with incident AF and this in consideration with the study size led us to consider Model 2 the most appropriate model. The number of SVES/24 h was a statistically significant predictor of incident AF; hazard ratio (HR) 1.38, 95% CI 1.14-1.68, p=0.001 per log unit, after adjustment for Model 2 covariates. There was no substantial difference between results of competing risks regression analysis; subhazard ratio (SHR) 1.39, 95% CI 1.16-1.68, p< 0.0001 per log unit. Unadjusted Kaplan-Meier curves representing the risk of incident AF by quartiles of SVES are presented in Fig 2. The p-value for trend across quartiles was 0.015.
The number of SVT/24 h was likewise associated with increased risk of AF; HR 1.95, 1.21-3.13, p=0.006 per log unit after adjustment for Model 2 covariates using Cox regression, and SHR 1.99, 95% CI 1.40-2.82, p
