Electrocardiograms are routinely obtained in clinical follow-up of patients with asymptomatic aortic stenosis (AS). The association with aortic valve, left ventricular (LV) response to long-term pressure load, and clinical covariates is unclear and the clinical value is thus uncertain. Data from clinical examination, electrocardiogram, and echocardiogram in 1,563 patients in the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) study were used. Electrocardiograms were Minnesota coded for arrhythmias and atrioventricular and intraventricular blocks; LV hypertrophy was assessed by Sokolow–Lyon voltage and Cornell voltage–duration criteria; and strain by T-wave inversion and ST-segment depression. Degree of AS severity was evaluated by echocardiography as peak aortic jet velocity and LV mass was indexed by body surface area. After adjustment for age, gender, LV mass index, heart rate, systolic and diastolic blood pressures, blood glucose, digoxin, antiarrhythmic drugs, drugs acting on the renin–angiotensin system, diuretics, β blockers and calcium receptor blockers; peak aortic jet velocity was significantly greater in patients with electrocardiographic strain (mean difference 0.13 m/s, p <0.001) and LV hypertrophy by Sokolow–Lyon voltage criteria (mean difference 0.12 m/s, p = 0.004). After similar adjustment, LV mass index was significantly greater in patients with electrocardiographic strain (mean difference 14.8 g/cm 2 , p <0.001) and LV hypertrophy by Sokolow–Lyon voltage criteria and Cornell voltage–duration criteria (mean differences 8.8 and 17.8 g/cm 2 , respectively, p <0.001 for the 2 comparisons). In multiple comparisons patients with electrocardiographic strain had increased peak aortic jet velocity, blood glucose, and uric acid, whereas patients with LV hypertrophy by Sokolow–Lyon voltage criteria were younger and patients with LV hypertrophy by Cornell voltage–duration criteria more often were women. In conclusion, electrocardiographic criteria for LV hypertrophy and strain are independently associated with peak aortic jet velocity and LV mass index. Moreover, clinical covariates differ significantly between patients with electrocardiographic strain and those with LV hypertrophy by Sokolow–Lyon voltage criteria and Cornell voltage–duration criteria.
Aortic valve replacement is the only known effective treatment for aortic stenosis (AS). However, the procedure is associated with several risks and surgery is in general restricted to symptomatic and asymptomatic patients with decreased left ventricular (LV) function or pathologic exercise test results. Patients with AS without these characteristics are screened with annual or biannual clinical, echocardiographic, and electrocardiographic examinations to evaluate the need for aortic valve replacement. During this period of watchful waiting, several factors including age, degree of calcification, lower LV systolic function, and co-morbid status pertain to the risk of AS progression and adverse outcome. In this regard, the electrocardiogram is appealing because it is a low-cost and widely available examination that can be interpreted without expert knowledge. However, the relation of electrocardiographic variables to the aortic valve, LV response to long-term pressure load and clinical covariates, and thus the ability to predict risk of adverse outcome has not been well studied. We hypothesized that electrocardiographic variables were closely associated with echocardiographic degree of AS severity as detected by standard echocardiographic examination and that electrocardiogram contains additive value in assessing the cardiovascular risk profile in asymptomatic patients with AS. To evaluate the potential value of electrocardiograms in contemporary patients with asymptomatic AS, this study investigated how electrocardiographic variables relate to clinical covariates, echocardiographically estimated peak aortic jet velocity, and LV mass indexed by body surface area in asymptomatic patients with AS.
Methods
All data were from the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) study, a multicenter, double-blind, placebo-controlled study investigating whether randomizing 1,873 patients (45 to 85 years of age) with asymptomatic AS (defined as echocardiographic aortic valve thickening accompanied by Doppler-measured aortic peak flow velocity ≥2.5 and ≤4.0 m/s, normal LV systolic function, and absence of symptoms according to independent local investigators based on patient interviews) to intensive lipid lowering with simvastatin plus ezetimibe versus placebo could decrease the need for aortic valve replacement and risk of cardiovascular morbidity and mortality. The main outcome and the complete study protocol, study design, organization, clinical measurements, exclusion criteria (including clinically apparent coronary artery disease), and baseline characteristics have been published. As part of the SEAS study protocol, all patients were also enrolled in the SEAS electrocardiographic substudy. The present study used baseline data from the SEAS and SEAS electrocardiographic substudy to investigate the relation of electrocardiographic variables to clinical and echocardiographic covariates in asymptomatic patients with AS before receiving the study drug (the SEAS trial is registered at http://ClinicalTrials.gov , identifier NCT00092677 ).
All electrocardiograms were obtained at local study centers, labeled with the date and anonymous allocation number, and then sent to the SEAS Electrocardiographic Core Laboratory at The Heart Center, Rigshospitalet, Copenhagen, Denmark. A physician with several years of experience in electrocardiographic interpretation blinded to the randomization, echocardiographic- and clinical data, read and transferred all electrocardiograms directly to a database for statistical analysis. Only 12-lead electrocardiograms were used, with variables in or transformed to correspond to a 25-mm/s and 0.1-mm/mV formatted electrocardiogram. For cases in which no electrocardiogram was forwarded to the core laboratory, a letter was sent to the local investigator asking for a copy; if the electrocardiogram could not be provided, the patient was excluded from this substudy. Electrocardiograms were read using the Minnesota coding system for arrhythmias and atrioventricular and intraventricular blocks in accordance with recent suggestions. The following items were measured manually: (1) PR-interval duration (from start of P wave to beginning of QRS interval); (2) amplitude of ST-segment depression (assessed 80 ms after the J point) and T-wave inversion (defined as maximally negative T-wave amplitude; ST-T strain was not read if left bundle branch block was present, and T-wave inversion was censored in the presence of right bundle branch block; LV strain was computed by summing ST-segment depression and T-wave inversion in leads V 4 to V 6 and universal electrocardiographic strain by an equal procedure in all leads after previous reversal of values for lead aVR); (3) QT and RR intervals averaged over 3 adjacent beats to compute the corrected QT interval by the Bazett formula (corrected QT interval = QT interval [seconds]/ √ RR interval [seconds]); (4) QRS duration (lower detection limit 80 ms) and voltage amplitude; (5) LV hypertrophy determined as Cornell voltage–duration product ([R wave in lead aVL + S wave in lead V 3 + {6 mV in women}] × QRS duration) ≥2,440 mV/ms and/or Sokolow-Lyon voltage (R wave in leads V 5 to V 6 + S wave in lead V 1 ) ≥35 mV; and (6) electrocardiographic left atrial size estimated by P-wave duration in lead II plus maximal negative P-wave amplitude in precordial lead V 1 .
Echocardiographic study protocol, reading procedures, and reproducibility have been published. Briefly echocardiograms were read blinded at the SEAS Echocardiographic Core Laboratory at Haukeland University Hospital, Bergen, Norway. Peak aortic jet velocity was calculated by applying the continuity equation in accordance with recent recommendations. LV dimensions and wall thicknesses were measured on 2-dimensional images according to American Society of Echocardiography guidelines using an anatomically validated formula.
Data were analyzed using SAS 9.1 (SAS Institute, Cary, North Carolina). Continuous variables are presented as mean ± SD and categorical variables as proportions. Continuous variables were tested for normality; ST-T segment strain was not normal and thus transformed to ranks with or without strain. To simplify the relation of continuous and categorical variables to echocardiographic degree of AS severity, variables were grouped by peak aortic jet velocity as mild (<3.0 m/s), moderate (≥3.0 to ≤4.0 m/s), and severe AS (>4.0 m/s) in accordance with guidelines from the European Society of Cardiology. Statistical significance of differences in continuous variables was tested using 1- and 2-way analyses of variance (with unweighted mean analysis owing to unequal sample sizes) and by trend tests and logistic regression for categorical data. Multivariable relations of electrocardiographic variables to peak aortic jet velocity and LV mass index were evaluated in 2 separate general linear models, checked for interactions, and adjusted for all variables with a significant univariate relation to echocardiographic AS severity and/or a presumed biological influence on evaluated parameters: (1) drugs acting on the renin–angiotensin system and blood glucose levels because of antifibrocalcific and profibrocalcific effects, respectively; (2) drugs with a β or calcium receptor blockade effect because of an influence on heart rate and atrioventricular conduction delay; (3) systolic and diastolic blood pressures because of confounding between hypertension- and AS-induced hypertrophy and strain; (4) digoxin because of an influence on heart rate, atrioventricular conduction delay, and appearance of ST-segment (Cohn effect); (5) antiarrhythmic drugs (defined as class IC or III) because of an influence on heart rate and prevalence of arrhythmias such as atrial fibrillation; and (6) in the multivariable model with LV mass index and vice-versa, peak aortic jet velocity to assess their separate relations to electrocardiographic changes. In all multivariable models electrocardiographic strain was, because of colinearity of ST-segment depression and T-wave inversion and a closer association of T-wave inversion to peak aortic jet velocity, defined as presence of T-wave inversion. To evaluate differences between patients with electrocardiographic strain versus LV hypertrophy by Sokolow–Lyon voltage and Cornell voltage–duration criteria, respectively, we assessed differences in Bonferroni-adjusted means of clinical and echocardiographic covariates in a general linear model with multiple comparisons of electrocardiographic strain and LV hypertrophy by Sokolow–Lyon voltage criteria and Cornell voltage–duration criteria. Interobserver variation was evaluated by paired t tests and kappa statistics. Because of randomization all models were adjusted by simvastatin/ezetimibe versus placebo arm. For all models a 2-tailed p value <0.05 was required for statistical significance.
Results
Electrocardiographic data were available for 1,563 patients (958 men, 61.4%; 605 women, 38.7%; mean age 67.4 ± 9.6 years) and echocardiographic data were available for 1,471 of these patients. Patients’ mean peak aortic jet velocity was 3.1 ± 0.5 m/s ( Table 1 ). Although the predefined inclusion criterion was peak aortic jet velocity ≥2.5 to ≤4.0 m/s, 84 patients had peak aortic jet velocity >4.0 m/s (severe AS) determined at echocardiographic proof reading in the core laboratory. Differences in clinical and echocardiographic covariates among mild, moderate, and severe AS are presented in Table 1 .
| Variable | Total (n = 1,471) | Mild (n = 681) | Moderate (n = 706) | Severe (n = 84) | p Value |
|---|---|---|---|---|---|
| Age (years) | 67.5 ± 9.6 | 67.3 ± 9.6 | 67.5 ± 9.6 | 68.6 ± 8.9 | 0.47 |
| Men | 61.5% | 60.7% | 61.8% | 65.5% | 0.43 |
| Systolic blood pressure (mm Hg) | 144.8 ± 20.2 | 144.5 ± 19.6 | 145.4 ± 21.2 | 145.6 ± 18.3 | 0.71 |
| Diastolic blood pressure (mm Hg) | 82.0 ± 10.3 | 81.7 ± 10.5 | 82.4 ± 10.7 | 80.9 ± 10.0 | 0.29 |
| Peak aortic jet velocity (m/s) | 3.1 ± 0.5 | 2.6 ± 0.3 | 3.4 ± 0.3 | 4.2 ± 0.1 | — |
| Left ventricular mass index (g/cm 2 ) | 100.2 ± 30.7 | 96.7 ± 29.9 | 102.4 ± 30.0 | 110.3 ± 34.9 | <0.001 |
| Body mass index (kg/m 2 ) | 26.9 ± 4.4 | 27.0 ± 4.3 | 27.0 ± 4.5 | 26.4 ± 4.3 | 0.54 |
| High-density lipoprotein cholesterol (mmol/L) | 1.5 ± 0.4 | 1.5 ± 0.4 | 1.5 ± 0.4 | 1.5 ± 0.4 | 0.62 |
| High-density lipoprotein cholesterol (mg/dl) | 57.9 ± 15.4 | 57.9 ± 15.4 | 57.9 ± 15.4 | 57.9 ± 15.4 | 0.62 |
| Low-density lipoprotein (mmol/L) | 3.6 ± 0.9 | 3.6 ± 0.9 | 3.5 ± 0.9 | 3.6 ± 1.0 | 0.87 |
| Low-density lipoprotein (mg/dl) | 139.0 ± 23.2 | 139.0 ± 34.7 | 135.1 ± 34.7 | 139.0 ± 38.6 | 0.87 |
| Creatinine (μmol/L) † | 93.6 ± 15.7 | 94.3 ± 15.8 | 93.5 ± 15.8 | 93.9 ± 13.2 | 0.67 |
| Uric acid (μmol/L) † | 351.8 ± 76.6 | 352.1 ± 79.3 | 351.9 ± 75.7 | 361.1 ± 67.9 | 0.60 |
| Glucose (mmol/L) † | 5.3 ± 0.8 | 5.3 ± 0.8 | 5.3 ± 0.9 | 5.2 ± 0.8 | 0.30 |
| Heart rate (beats/min) | 65.1 ± 11.2 | 64.2 ± 11.1 | 65.9 ± 11.3 | 65.5 ± 11.3 | 0.01 |
| Digoxin | 2.9% | 2.8% | 2.8% | 3.6% | 0.78 |
| Antiarrhythmic drugs | 2.0% | 1.9% | 2.1% | 1.2% | 0.94 |
| Renin–angiotensin system inhibitors | 41.9% | 42.9% | 39.5% | 52.4% | 0.89 |
| β Blockers | 50.9% | 45.4% | 55.2% | 58.3% | <0.001 |
| Calcium antagonist | 29.1% | 28.9% | 28.1% | 39.3% | 0.35 |
| Diuretics | 46.7% | 43.6% | 48.6% | 56.0% | 0.01 |
⁎ Mild = peak aortic jet velocity >2.5 to <3.0 m/s; moderate = peak aortic jet velocity ≥3.0 to ≤4.0 m/s; severe = peak aortic jet velocity >4.0 m/s.
† To convert values to milligrams per deciliter, divide creatine value by 88.4, uric acid value by 59.48, and glucose value by 0.0555.
Interobserver variability analyses of PR and QRS durations, Sokolow–Lyon voltage, and Cornell voltage–duration product are presented in Table 2 . Kappa values for determining ST-segment depression and T-wave inversion were 0.88 and 1.0 (p >0.05 for difference for the 2 comparisons), respectively.
| Variable | Interobserver |
|---|---|
| Sokolow–Lyon voltage (mV) | 0.31 ± 0.95 |
| Cornell voltage–duration product (mV × ms) | 56.7 ± 145.1 |
| PR interval (ms) | 1.5 ± 10.4 |
| QRS duration (ms) | 2.5 ± 6.8 |
Electrocardiographic changes in mild, moderate, and severe AS are presented in Table 3 . In univariate logistic regression ( Table 4 ), presence of LV hypertrophy by voltage of R wave in leads V 5 to V 6 plus S wave in lead V 1 related to a 2.0-fold higher risk of echocardiographic severe AS (95% confidence interval 1.19 to 3.24, p <0.01), whereas T-wave inversion and ST-segment strain related to a 2.1-fold higher risk of severe AS (95% confidence interval 1.31 to 3.52 and 1.29 to 3.46, respectively, p <0.01 for the 2 comparisons; Figure 1 ).
| Variable | Total (n = 1,471) | Mild (n = 681) | Moderate (n = 706) | Severe (n = 84) | p Value |
|---|---|---|---|---|---|
| Sokolow–Lyon voltage (mV) | 26.6 ± 9.5 | 25.6 ± 9.2 | 27.2 ± 9.4 | 29.6 ± 11.1 | <0.001 |
| Cornell voltage–duration product (mV × ms) | 1,743.3 ± 878.9 | 1,677.9 ± 874.9 | 1,801.0 ± 885.0 | 1,792.3 ± 891.7 | 0.03 |
| PR interval (ms) | 175.5 ± 30.3 | 174.3 ± 28.4 | 176.4 ± 32.6 | 179.5 ± 25.7 | 0.22 |
| QRS duration (ms) | 88.2 ± 14.1 | 87.9 ± 14.4 | 88.2 ± 13.6 | 90.1 ± 16.5 | 0.41 |
| Corrected QT interval (ms/ √ s) | 410.3 ± 22.5 | 408.6 ± 23.7 | 411.4 ± 21.4 | 412.2 ± 23.9 | 0.06 |
| Electrocardiographic left atrial size (mm + ms) | 96.2 ± 12.9 | 95.6 ± 13.9 | 96.4 ± 13.1 | 96.8 ± 11.3 | 0.72 |
| Left bundle branch block | 2.9% | 3.5% | 2.1% | 3.6% | 0.23 |
| Right bundle branch block | 4.0% | 3.4% | 4.3% | 7.1% | 0.12 |
| ST-segment depression in leads V 4 –V 6 | 17.1% | 12.5% | 20.3% | 27.3% | <0.001 |
| T-wave inversion in leads V 4 –V 6 | 24.1% | 18.5% | 27.9% | 39.4% | <0.001 |
| ST-segment depression in any lead | 22.2% | 17.1% | 25.5% | 36.5% | <0.001 |
| T-wave inversion in any lead | 38.8% | 32.6% | 43.3% | 52.2% | <0.001 |
| Atrial fibrillation | 3.7% | 3.7% | 3.8% | 2.4% | 0.81 |
| Premature ventricular beats | 2.3% | 1.6% | 2.8% | 2.4% | 0.19 |
| Normal electrocardiogram | 42.5% | 49.4% | 36.0% | 40.6% | <0.001 |
⁎ Mild = peak aortic jet velocity >2.5 to <3.0 m/s; moderate = peak aortic jet velocity ≥3.0 to ≤4.0 m/s; severe = peak aortic jet velocity >4.0 m/s.
| Variable | Prevalence in Nonsevere AS (n = 1,387) | Prevalence in Severe AS (n = 84) | OR for Severe AS (95% CI) | p Value |
|---|---|---|---|---|
| Normal electrocardiogram | 42.6% | 40.6% | 0.9 (0.56–1.51) | 0.74 |
| Atrial fibrillation | 3.8% | 2.4% | 0.6 (0.15–2.61) | 0.52 |
| At least first-degree of atrioventricular block | 5.5% | 4.8% | 0.9 (0.31–2.41) | 0.78 |
| Left bundle branch block | 2.8% | 3.6% | 1.3 (0.39–4.23) | 0.69 |
| Right bundle branch block | 3.8% | 7.1% | 1.9 (0.81–4.64) | 0.14 |
| Premature ventricular beats | 2.2% | 2.4% | 1.1 (0.25–4.53) | 0.93 |
| T-wave inversion in leads V 4 –V 6 | 23.3% | 39.4% | 2.1 (1.31–3.52) | 0.002 |
| ST-segment depression in leads V 4 –V 6 | 16.5% | 27.3% | 2.1 (1.29–3.46) | 0.003 |
| Left ventricular hypertrophy by Cornell voltage–duration | 14.2% | 17.5% | 1.3 (0.71–2.32) | 0.42 |
| Left ventricular hypertrophy by Sokolow–Lyon voltage | 16.3% | 27.7% | 2.0 (1.19–3.24) | 0.008 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
