Author + information
- Received November 17, 2014
- Revision received December 22, 2014
- Accepted December 26, 2014
- Published online March 1, 2015.
- Yasufumi Nagata, MD∗,
- Masaaki Takeuchi, MD∗∗ (, )
- Victor Chien-Chia Wu, MD∗,†,
- Masaki Izumo, MD‡,
- Kengo Suzuki, MD‡,
- Kimi Sato, MD§,
- Yoshihiro Seo, MD§,
- Yoshihiro J. Akashi, MD‡,
- Kazutaka Aonuma, MD§ and
- Yutaka Otsuji, MD∗
- ∗Second Department of Internal Medicine, University of Occupational and Environmental Health, School of Medicine, Kitakyushu, Japan
- †Department of Cardiology, Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Taipei, Taiwan
- ‡Division of Cardiology, Department of Internal Medicine, St. Marianna University School of Medicine, Kawasaki, Japan
- §Cardiovascular Division, University of Tsukuba, Tsukuba, Japan
- ↵∗Reprint requests and correspondence:
Dr. Masaaki Takeuchi, Second Department of Internal Medicine, University of Occupational and Environmental Health, School of Medicine, Iseigaoka, Yahatanishi-ku, Kitakyushu 807-8555, Japan.
Objectives The objective of this study was to determine which strain component assessed by 2-dimensional speckle-tracking echocardiography (2DSTE) and 3-dimensional speckle-tracking echocardiography (3DSTE) was the most powerful predictor for future major adverse cardiac events (MACE) in asymptomatic patients with severe aortic stenosis (AS).
Background Ongoing debate exists regarding the appropriateness of early surgery in asymptomatic severe AS and preserved left ventricular ejection fraction (LVEF). Previous studies showed that 2-dimensional global longitudinal strain (2DGLS) was a significant predictor in asymptomatic severe AS patients. However, the prognostic utility of 3DSTE-derived multidirectional strain parameters has not been investigated in these patients.
Methods We enrolled 104 asymptomatic severe AS patients (indexed aortic valve area <0.6 cm2/m2) and preserved LVEF and performed strain analysis using both 2DSTE and 3DSTE. Two-dimensional and 3-dimensional global longitudinal, circumferential, and radial strain and global 3-dimensional strain were measured in each patient. All patients were followed to record MACE.
Results During a median follow-up of 373 days, MACE developed in 33 patients (32%). 2DGLS (−14.7 ± 3.3 vs. −16.3 ± 3.3, p = 0.0168), 3DGLS (−13.5 ± 2.5 vs. −16.1 ± 2.4, p < 0.0001) and 3-dimensional global radial strain (3DGRS) (35.9 ± 4.5 vs. 38.1 ± 4.4, p = 0.0209) were significantly impaired in patients with MACE compared with those without MACE. Kaplan-Meier analysis showed 2DGLS (cutoff: −17.0%), 3DGLS (cutoff: −14.5%), and 3DGRS (cutoff: 39.0%) provide a significant difference in MACE rate. Receiver-operating characteristic analysis revealed that the area under the curve of 3DGLS for MACE (0.78) was significantly larger than that of 2DGLS (0.62, p = 0.0044) and 3DGRS (0.66, p = 0.0069). Separate multivariate analysis revealed 3DGLS was only significant as independent predictor for future MACE after correcting for mean pressure gradient and left ventricular mass index.
Conclusions 3DGLS is the most robust index for predicting future adverse cardiac events in asymptomatic severe AS patients with preserved LVEF.
Calcific aortic stenosis (AS) is the most common form of valvular heart disease in developed countries, and the number of affected individuals is rapidly growing as life expectancy increases (1,2). Symptomatic severe AS patients and asymptomatic severe AS patients with impaired left ventricular (LV) function have Class I indications for aortic valve replacement (AVR) according to 2014 AHA/ACC guideline for the management of patients with valvular heart disease (2,3). However, there are ongoing debates regarding early surgery or watchful waiting in asymptomatic severe AS and preserved LV ejection fraction (LVEF) (4–6). Subendocardial fibrosis caused by increased mechanical stress imposed by AS itself is observed, even in normal LVEF (7). Moreover, LVEF is not a sensitive marker for detecting subclinical LV dysfunction (8). Therefore, a more sophisticated approach to evaluating LV mechanics is mandatory for selecting high-risk asymptomatic patients with severe AS and preserved patients who should undergo early surgical intervention. Among several proposed echocardiographic parameters for predicting adverse outcomes, global longitudinal strain (GLS) assessed by 2-dimensional (2D) speckle-tracking analysis can provide useful prognostic information (9–11). However, 2D strain measurements have the inherent limitation of losing speckles from out-of-plane cardiac motion (12). Three-dimensional speckle-tracking echocardiography (3DSTE), on the other hand, has the capability to overcome these drawbacks (13,14). However, it has not been determined which directions of strain components measured by 3DSTE are the best predictors of future prognosis in this subset of patients. Accordingly, we aimed to investigate the impairment of multidirectional strain components using 3DSTE in asymptomatic patients with severe AS and preserved ejection fraction and to elucidate which strain assessed by both 2-dimensional speckle-tracking echocardiography (2DSTE) and 3DSTE is the more robust predictor of future major adverse cardiac events (MACE) in these patients.
We enrolled patients with AS who underwent transthoracic 3-dimensional (3D) echocardiography from 3 cardiovascular institutions. 3D echocardiography dataset acquisitions were performed at the time of clinically indicated echocardiography examinations in each hospital from January 2011 to February 2014. At the same time, attending physicians performed careful history taking regarding AS-related symptoms in these patients. The inclusion criteria in this study were patients with severe AS defined as indexed aortic valve area (iAVA) <0.6 cm2/m2, preserved LVEF (LVEF >50%), and no recorded symptoms relating to AS. We excluded patients who had <2 months of follow-up and those who had underwent AVR within 2 months after echocardiographic examination. Clinical characteristics including hypertension, diabetes mellitus, hyperlipidemia, chronic kidney disease, and coronary artery disease were documented at the time of echocardiography examination. Chronic kidney disease was defined as an estimated glomerular filtration rate <60 ml/min/1.73 m2. The study was approved by the Ethics Committee at each hospital, and informed consent was obtained from all subjects.
Comprehensive 2D and Doppler echocardiography were performed using a commercially available ultrasound machine and transducer (iE33, Philips Medical System, Andover, Massachusetts or Vivid 7 or E9, GE Healthcare, Horten, Norway). Views from 3 short axes (basal, middle, and apical) and 3 apical axes (apical 4-chamber, 2-chamber, and long-axis views) that encompassed the whole part of the left ventricle were acquired. Pulsed-wave Doppler examination of LV inflow and outflow and tissue Doppler examination of the mitral annulus were performed according to the American Society of Echocardiography recommendations (15). Flow velocity across the aortic valve was measured at multiple transducer positions using continuous-wave Doppler and peak and mean pressure gradients (PGs) were calculated with a simplified Bernoulli equation. Aortic valve area (AVA) was calculated using the continuity equation. LV volume measurements were performed using 3DSTE.
3D full-volume datasets were acquired from the apical transducer position during held respiration by a fully sampled matrix array transducer (X5-1/X3-1, Philips Medical Systems or 4V, GE Healthcare). To ensure the inclusion of the entire left ventricle within the pyramidal scan volume with a relatively high volume rate, datasets throughout 1 cardiac cycle were acquired using the wide-angle mode, wherein multiple wedge-shaped subvolumes were acquired with electrocardiographic gating during a single 5-s to 7-s breath-hold (16).
2D speckle-tracking echocardiography
2D speckle-tracking analysis was performed using vendor-independent 2D speckle-tracking software (2D Cardiac Performance Analysis, TomTec Imaging Systems, Unterschleissheim, Germany). Radial and circumferential strains were determined by endocardial tracing in the 3 levels of short-axis views. Longitudinal strain was measured by manual tracing of the endocardial border in the 3 apical views. After speckle-tracking analysis of the LV endocardium on a frame-by-frame basis during 1 cardiac cycle, the software provides regional strain curves of 6 segments (4 segments in the apical short-axis view) in each view, from which the peak regional strain value was determined. Global strain was defined as the peak strain value from the averaged strain curve that was generated from 16 segmental strain curves (global radial strain [GRS], global circumferential strain [GCS], or 18 segmental strain curves [GLS]). Adequacy of tracking was verified visually, and if the tracking was deemed to be suboptimal, manual adjustment of endocardial border was performed. If tracking was still judged to be unsatisfactory, subjects were excluded from the analysis.
3D speckle-tracking echocardiography
A 3D volumetric analysis and 3D strain measurements of the left ventricle were performed using 3DSTE. 3D full-volume datasets were analyzed using vendor-independent 3D speckle-tracking software (4D LV Analysis, version 3.1.2, TomTec Imaging Systems) by an experienced investigator (Figure 1). After importing 3D full-volume datasets, the apical 4-chamber, 2-chamber, and long-axis views and 3 short-axis views at end-diastole were automatically extracted. Nonforeshortened apical views were identified to select the point of the apex and the center of the mitral annular line connecting both sides of the mitral annulus with largest LV long-axis dimensions after which the 3D endocardial surface was automatically reconstructed. The papillary muscles were included in the LV cavity. Manual adjustments of the endocardial surface were performed when necessary. The same procedure was performed at the end-systolic frame. Subsequently, the software performed 3D speckle-tracking analysis throughout the cardiac cycle. For LV volume measurements, LV end-diastolic volume and LV end-systolic volume were obtained as the largest and smallest volume, respectively, from the time-volume curve. LV stroke volume and LVEF were calculated as: LV end-diastolic volume − LV end-systolic volume and LV stroke volume/LV end-diastolic volume, respectively. For the determination of LV mass, epicardial surface delineation was initialized at the end-diastolic frame. The LV mass was calculated as: (LV epicardial volume − LV endocardial volume) × 1.05 (g/ml) (17). For 3D strain analysis, the left ventricle was automatically divided into 16 segments using standard segmentation schemes. The software provided averaged longitudinal, circumferential, radial, and 3D strain time curves from each segmental strain curve, from which the peak global strain was determined. 3D strain describes the tangential deformation and is calculated as the vector sum of the longitudinal and circumferential strain components, ignoring the radial component. Because the software did not provide an automated evaluation of the adequacy of image-tracking capabilities, the accuracy of tracking was visually evaluated on the 2D images extracted from 3D datasets. When tracking was deemed to be inadequate, the endocardial surface was manually readjusted as necessary. If tracking was still judged inadequate, subjects were excluded from the analysis.
Follow-up information was obtained regularly in the outpatient clinic. Telephone contact with patients, physicians, and next of kin was performed if the patients had been treated at another hospital. If symptoms developed, AS severity rapidly progressed, or decreased LVEF (<50%), the attending physician would refer the patient for AVR. Because each physician did not have the strain measurements information, depressed strain values did not have impact on the decision. The primary endpoint was MACE including cardiac death, sustained ventricular tachyarrhythmia, AVR, and hospital admission for heart failure within 2 years of follow-up.
Intraobserver and interobserver variability
Intraobserver and interobserver variability in measurements of all components of 3D global strains was assessed in 15 randomly selected subjects, and the reported percentage of variability was defined as the absolute difference in the percentage of the mean of repeated measurements and intraclass correlation coefficient.
Continuous variables were expressed as the mean ± SD or as the median (interquartile range) according to the data distribution. Categorical variables were presented as numbers and proportions. All statistical analyses were performed using commercially available software (JMP version 11.0, SAS Institute Inc., Cary, North Carolina). Differences in measurements between the 2 groups were assessed using Student t test for continuous variables and the chi-square test or Fisher exact test for categorical variables whenever appropriate. Receiver-operating characteristic (ROC) analysis was performed to investigate the sensitivity and specificity for MACE of parameters and to determine the best cutoff value of each variable for MACE. A Kaplan-Meier survival analysis was used to plot MACE. The log-rank test was used to evaluate the differences between the 2 groups. Univariate analysis was performed to determine the significant predictors of MACE. For multivariate analysis, a separate Cox proportional hazard model, including 1 of each global strain parameter, was used to identify the independent variables for predicting future MACE. A p value <0.05 was considered significant.
Of 429 AS patients for whom 3D echocardiography data were acquired (St. Marianna University, 126 patients; University of Tsukuba, 78 patients; and University of Occupational and Environmental Health, 225 patients), 133 patients met inclusion criteria. Subsequently, 29 patients were excluded from analysis due to poor tracking of 3DSTE (n = 9) or a follow-up period of <2 months (n = 20), leaving 104 patients as the final study population (Figure 2). Baseline clinical and echocardiographic parameters are shown in Tables 1 and 2⇓. The majority of patients had some cardiovascular risk factors. The iAVA was 0.42 ± 0.10 cm2/m2 and the mean PG was 39 ± 17 mm Hg. 2D speckle-tracking analysis could not be performed in 2 patients for longitudinal strain and 6 patients for circumferential and radial strain due to unreliable tracking. 3D speckle-tracking analysis was possible in all patients. The mean values of 2DGLS, 2DGCS, and 2DGRS were −15.8 ± 3.4%, −26.8 ± 6.0%, and 34.1 ± 12.2%, respectively. Corresponding values of 3DGLS, 3DGCS, global 3-dimensional strain (G3DS), and 3DGRS were −15.3 ± 2.7%, −30.6 ± 4.0%, −35.0 ± 4.0%, and 37.4 ± 4.5%, respectively. A weak but significant correlation of global strain values between the 2 modalities was noted (GLS: r = 0.56, p < 0.0001; GCS: r = 0.31, p = 0.0021; GRS: r = 0.29, p = 0.0042).
The median follow-up was 373 days (interquartile range: 163 to 495 days). A total of 33 patients reached the primary endpoints until 2 years, including 4 cardiac deaths, 1 ventricular fibrillation, 11 heart failures requiring hospital admission, and 17 AVRs due to symptom development (n = 13), depressed LVEF (n = 2), and rapid hemodynamic progression (n = 2). Event-free survival in the overall population was 72 ± 5% and 58 ± 7% at 1- and 2-year follow up, respectively. Table 3 shows baseline echocardiographic parameters and strain values between patients with and without events. There was no significant difference in clinical parameters between the 2 groups. Peak velocity and mean PG were significantly higher, whereas iAVA was significantly smaller in patients with MACE compared with those with no MACE. Although no significant differences in LV volume were noted, maximal left atrial (LA) volume index was significantly larger in patients with MACE. Among 2DSTE-derived strain parameters, 2DGLS showed significant differences between the 2 groups. Regarding 3D strains, 3DGLS and 3DGRS were significantly impaired in patients with MACE compared with those without MACE.
Associations of outcomes
In all 7 global strain components determined by 2DSTE and 3DSTE, the area under the curve (AUC) calculated by ROC analysis for future MACE was largest in 3DGLS (0.78), followed by 3DGRS (0.66) and 2DGLS (0.62). The AUC of 3DGLS was significantly larger than that of the 6 other global strains. 3DGLS had also the largest AUC among other traditional echocardiographic parameters (Figure 3), and the AUC of 3DGLS was significantly larger than that of iAVA, LVEF, LV mass index, and maximal LA volume index. ROC analysis revealed a 3DGLS cutoff value of −14.5% had a sensitivity of 76% and a specificity of 77% for predicting future MACE. Corresponding analysis showed that a 3DGRS cutoff value of 39.0% had a sensitivity of 82% and a specificity of 44% for MACE. A 2DGLS cutoff value of −17.0% had a sensitivity of 85% and a specificity of 48% for predicting future MACE. Figure 4 depicts a Kaplan-Meier survival curve for the 2 groups classified by cutoff values of 2DGLS, 3DGLS, and 3DGRS. All 3 cutoff criteria had significant predictive power for MACE.
Table 4 shows the results of univariate and multivariate analyses of clinical and echocardiographic variables. The significant univariate factors associated with MACE (p < 0.1) were iAVA, peak velocity, mean PG, valvuloarterial impedance, stroke volume index, LV mass index, maximal LA volume index, 2DGLS, 3DGLS, and 3DGRS. The independent associations of outcome were analyzed using multivariate Cox proportional hazards models. To avoid problems of colinearity and overfitting the data, the LV mass index and mean PG were selected from the parameters regarding AS severity, and each strain parameter was evaluated in separate models (Table 4). Multivariate analysis revealed that only 3DGLS remained statistically significant as a predictor of future MACE.
Because the median value of the mean PG was 35 mm Hg, more than one-half of patients had a mean PG <40 mm Hg. To determine the prognostic value of 2DGLS/3DGLS and 3DGRS regarding the status of the mean PG, we divided patients into 2 groups according to the well-established cutoff value of the mean PG, where a low PG AS is <40 mm Hg and a high PG is ≥40 mm Hg (18). Both 2DGLS and 3DGLS were significant predictors of future MACE in both groups of patients with high or low PG severe AS. 3DGRS was a significant predictor of MACE in a group of patients with a low PG AS (Figure 5). These results showed that both 2DGLS and 3DGLS manifested a significant incremental power over the mean PG for predicting future cardiovascular events.
The intraobserver variability for the 3D strain was as follows: 3DGLS, 4.4 ± 2.7%; 3DGCS, 4.4 ± 3.5%; 3DGRS, 3.9 ± 2.4%; and G3DS, 4.9 ± 3.2%. The corresponding interobserver variability was 5.2 ± 4.5%, 6.5 ± 6.3%, 5.5 ± 4.9%, and 7.1 ± 4.5%, respectively. The intraobserver intraclass correlation coefficients for 3DGLS, 3DGCS, 3DGRS, and G3DS were 0.964, 0.898, 0.936, and 0.861, respectively. The corresponding interobserver intraclass correlation coefficients were 0.954, 0.673, 0.860, and 0.686, respectively.
To the best of our knowledge, this is the first study to evaluate the prognostic impact of 3DSTE-derived LV deformation parameters in asymptomatic severe AS patients with preserved LVEF and to directly compare its utility with that of 2DSTE-derived strains. The major findings of this study were as follows: 1) 2DGLS, 3DGLS, and 3DGRS were significantly impaired in patients with future MACE compared with those without MACE; 2) 2DGLS, 3DGLS, and 3DGRS could stratify a group of patients at high-risk of MACE; 3) only 3DGLS was powerful enough as an independent predictor on multivariate analysis; and 4) subgroup analysis according to the status of the mean PG demonstrated that both 2DGLS and 3DGLS could predict future MACE in low and high PG severe AS patients.
Of the population with severe AS, symptomatic and asymptomatic patients with LV dysfunction are Class I indications for AVR (3). However, management of asymptomatic patients with severe AS but preserved LVEF remains controversial (4,5). The natural history of patients with asymptomatic severe AS showed a 1-year event rate ranging from 20% to 43% (5,6,19). Although the occurrence of sudden death without preceding symptoms is uncommon, sudden death presented in ∼1% to 6% of asymptomatic severe AS patients annually (6,20). Various studies had sought to identify high-risk patients in this group using echocardiographic parameters, cardiac magnetic resonance–determined myocardial fibrosis, and certain biomarkers such as brain natriuretic peptide (19). Among them, 2D strain analysis has emerged to show potential for both quantifying LV mechanics and providing prognostic information (8,9,21–23). Previous studies reported that 2D global strains improved after AVR in patients with AS, even though the LVEF did not change considerably (21,22). In particular, 2DGLS has been shown to be the most useful index to reflect AS severity and symptomatic status (23). Lancellotti et al. (9) verified that 2DGLS could differentiate high-risk patients from low-risk patients for future cardiac events. Yingchoncharoen et al. (10) also demonstrated that 2DGLS has a significant incremental value over other clinical and echocardiographic parameters for predicting future cardiovascular events. Because 3DSTE theoretically overcomes the limitations of 2DSTE, such as loss of speckles during through-plane motion (12), a direct comparison of strain values by 2DSTE and 3DSTE for predicting future MACE is imperative and clinically important.
The 1-year event rate was 28%, which was in agreement with previous studies (5,6,19). Analyses of 3DSTE-determined multidirectional strains showed that 3DGLS and 3DGRS, but not 3DGCS and G3DS, were significantly impaired in patients with future MACE compared with those without. Our findings are in line with previous observations that the impairment of longitudinal function is closely coupled with subendocardial myocardial fibrosis (24). During disease progression, myocardial fibrosis gradually develops, starting at the subendocardial layers and progressing toward the outer myocardium. These alterations could affect LV systolic and diastolic function, contributing to the development of typical AS-related symptoms and their prognosis (24). Because the early stage of myocardial fibrosis is observed at the subendocardial layer, only longitudinal function such as GLS could detect this abnormality. This abnormality is not well represented by LVEF that is mainly related to global radial thickening. Circumferential function also could not detect this abnormality because it is related to midwall function.
The reason for the significant reduction in 3DGRS in patients with events could relate to the algorithm of 3DSTE software used in this study. To analyze 3DGRS, the software performs speckle tracking only in the subendocardial layer; thus, the strain value represents subendocardial rather than transmural function. Because radial thickening increases progressively from the epicardium toward the subendocardium, assessment of endocardial radial strain is expected to detect subendocardial dysfunction.
We identified several univariate echocardiographic variables including iAVA, peak velocity, mean PG, valvuloarterial impedance, stroke volume index, LV mass index, maximal LA volume index, 2DGLS, 3DGLS, and 3DGRS, which were significantly associated with MACE (p < 0.1). The prognostic value of all these parameters except 3DGLS and 3DGRS was already reported (4,9,19). We found that 3DGLS and 3DGRS were useful to identify high-risk patients for future MACE. ROC analysis revealed that 3DGLS has a significantly larger AUC for predicting MACE compared with 2DGLS and 3DGRS. Using a multivariate Cox hazard model, only 3DGLS was still a significantly powerful predictor of future MACE. Subgroup analysis in patients with low PG and with high PG severe AS showed that both 2DGLS and 3DGLS are a significant predictor of future MACE compared with 2DGLS. In summary, our study identified 3DGLS as the most robust index for predicting future adverse cardiac events in asymptomatic severe AS patients.
There are multiple markers for predicting adverse events in patients with asymptomatic severe AS with preserved LVEF, including increased brain natriuretic peptide, decreased 2DGLS, exercise hemodynamic markers of prognosis (25,26), and severity of aortic valve calcification (6). In this study, the reduction of 3DGLS also identified a high risk in asymptomatic severe AS patients with preserved LVEF. Currently no studies have been performed to determine the appropriateness of applying these parameters for the management of asymptomatic severe AS patients. Thus, multiple prognostic markers should be collected at the time of assessment, and if several parameters had already become abnormal, it might be better to recommend that the patient undergo early AVR surgery. Further prospective study is necessary to investigate whether a 3DGLS-directed strategy is useful to reduce sudden cardiac death in a large number of patients with asymptomatic severe AS.
First, only patients with good echocardiographic images that could be adequately analyzed for 3DSTE were enrolled; hence, there could be a selection bias. Second, the follow-up period was relatively short, with a median follow-up duration of 373 days. Third, due to limited physical activities in elderly patients, AS-related symptoms may not develop. It is also quite difficult to perform the exercise stress test in all elderly patients to assess their symptoms because of comorbidities such as orthopedic disease. Even in this setting, our results validated that 2D/3DGLS and 3DGRS would be useful to stratify high-risk patients with a poor prognosis. Fourth, the small number of cardiac deaths prompted us to also include AVR and admission for heart failure in MACE as a primary endpoint. The decision to perform AVR was made by individual attending physicians. Although physicians could determine this referral based on varying rationales for intervention, AVR was mainly dictated by the onset of AS-related symptoms, which is currently a Class I indication of intervention (3).
In asymptomatic patients with severe AS and preserved LVEF, multivariate analysis revealed that 3DGLS was a powerful independent predictor of future MACE. After ROC and subgroup analysis, we concluded that 3DGLS is the most robust index for predicting future adverse cardiac events in asymptomatic severe AS patients with preserved LV function.
The authors thank Drs. Kyoko Otani, Yuichiro Kado, Kei Mizukoshi, and Tomoko Ishizu for collecting datasets.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- 3-dimensional global strain
- aortic stenosis
- area under the curve
- aortic valve area
- aortic valve replacement
- global circumferential strain
- global longitudinal strain
- global radial strain
- indexed aortic valve area
- left atrial
- left ventricular
- left ventricular ejection fraction
- major adverse cardiac event(s)
- pressure gradient
- receiver-operating characteristic
- speckle-tracking echocardiography
- Received November 17, 2014.
- Revision received December 22, 2014.
- Accepted December 26, 2014.
- American College of Cardiology Foundation
- Freeman R.V.,
- Otto C.M.
- Otto C.M.
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.
- Lancellotti P.,
- Magne J.,
- Donal E.,
- et al.
- Rahimtoola S.H.
- Rudolph A.,
- Abdel-Aty H.,
- Bohl S.,
- et al.
- Ng A.C.T.,
- Delgado V.,
- Bertini M.,
- et al.
- Lancellotti P.,
- Donal E.,
- Magne J.,
- et al.
- Yingchoncharoen T.,
- Gibby C.,
- Rodriguez L.L.,
- Grimm R.A.,
- Marwick T.H.
- Dumesnil J.G.,
- Pibarot P.,
- Carabello B.
- Bhattacharyya S.,
- Hayward C.,
- Pepper J.,
- Senior R.
- Pellikka P.A.,
- Sarano M.E.,
- Nishimura R.A.,
- et al.
- Delgado V.,
- Tops L.F.,
- van Bommel R.J.,
- et al.
- Miyazaki S.,
- Daimon M.,
- Miyazaki T.,
- et al.
- Weidemann F.,
- Herrmann S.,
- Störk S.,
- et al.
- Marechaux S.,
- Hachicha Z.,
- Bellouin A.,
- et al.