Author + information
- Received February 10, 2015
- Revision received April 23, 2015
- Accepted May 14, 2015
- Published online February 1, 2016.
- Julien Ternacle, MDa,b,c,d,e,f,
- Diane Bodez, MDa,b,c,d,e,f,
- Aziz Guellich, PhDa,b,c,d,e,f,
- Etienne Audureau, MD, PhDb,c,g,
- Stephane Rappeneau, PhDa,b,c,d,e,f,
- Pascal Lim, MD, PhDa,b,c,e,f,
- Costin Radu, MDa,b,e,f,g,
- Soulef Guendouz, MDa,b,c,d,e,f,
- Jean-Paul Couetil, MDb,e,h,
- Nicole Benhaiem, MDi,
- Luc Hittinger, MD, PhDa,b,c,d,e,f,
- Jean-Luc Dubois-Randé, MD, PhDa,b,c,d,e,f,
- Violaine Plante-Bordeneuve, MD, PhDa,b,f,j,
- Dania Mohty, MD, PhDk,
- Jean-François Deux, MD, PhDa,b,c,f,l and
- Thibaud Damy, MD, PhDa,b,c,d,e,f,∗ ()
- aUPEC, Créteil, France
- bMondor Amyloidosis Network, Créteil, France
- cDepartment of Cardiology, AP-HP, Henri-Mondor Teaching Hospital, Créteil, France
- dINSERM U955, GRC Amyloid Research Institute, Créteil, France
- eDHU ATVB, Créteil, France
- fINSERM Clinical Investigation Center 006, Créteil, France
- gDepartment of Public Health, AP-HP, Henri-Mondor Teaching Hospital and Clinical Epidemiology and Aging EA 4393, Créteil, France
- hDepartment of Cardiovascular Surgery, AP-HP, Henri-Mondor Teaching Hospital, Créteil, France
- iDepartment of Pathology, AP-HP, Henri-Mondor Teaching Hospital, Créteil, France
- jDepartment of Neurology, AP-HP, Henri-Mondor Teaching Hospital, Créteil, France
- kDepartment of Cardiology, Dupuytren Hospital, CHU Limoges, Pôle Cœur-Poumon-Rein, Limoges, France
- lDepartment of Radiology, AP-HP, Henri-Mondor Teaching Hospital, Créteil, France
- ↵∗Reprint requests and correspondence:
Prof. Thibaud Damy, Department of Cardiology, Henri Mondor University Hospital, 51 Av de Lattre de Tassigny, Créteil 94100, France.
Objectives The aim of this study was to compare left ventricular longitudinal strain (LS) evaluated by 2-dimensional echocardiography with cardiac magnetic resonance (CMR) in cardiac amyloidosis (CA), establish correlations between histological and imaging findings, and assess the prognostic usefulness of LS measurement and CMR.
Background CA is a condition with a poor prognosis due chiefly to 3 forms of amyloidosis: light-chain amyloidosis (AL), hereditary transthyretin (M-TTR), and wild-type transthyretin (WT-TTR). Two-dimensional echocardiography measurement of LS has been reported to detect early left ventricular systolic dysfunction. The pathophysiological underpinnings, regional distribution, and prognostic significance of LS in CA are unclear.
Methods All patients underwent echocardiography, and 53 underwent CMR. The native hearts of the 3 patients who received heart transplants were subjected to histological examination. For each of the 17 left ventricular segments in the American Heart Association model, we evaluated LS, late gadolinium enhancement (LGE) by CMR, and cardiac amyloid deposition. Univariate and multivariate analyses were performed at 6 months to identify variables associated with major adverse cardiac events (MACE).
Results We studied 79 patients with CA; 26 had AL, 36 M-TTR, and 17 WT-TTR. Mean LS was −10 ± 4%. Both LS and amyloid deposits showed a basal-to-apical gradient. The mean LS and number of segments with LGE were similar across the 3 CA types. LS correlated with LGE and amyloid burden (r = 0.72). LGE was seen in the 6 basal segments in all WT-TTR patients. During the median follow-up of 11 months (range 4 to 17 months), 36 (46%) patients experienced MACE. Independent predictors of MACE were apical LS (cutoff, −14.5%), N-terminal pro–B-type natriuretic peptide (cutoff, 4,000 ng/l), and New York Heart Association functional class III to IV heart failure.
Conclusions Basal-to-apical LS abnormalities are similar across CA types and reflect the amyloid burden. Apical LS independently predicts MACE.
Amyloidosis is a systemic disease in which extracellular aggregates of insoluble beta-fibrillar protein deposit within various organs (1). The 3 main types of amyloidosis associated with cardiac involvement are light-chain amyloidosis (AL), hereditary transthyretin amyloidosis (M-TTR), and wild-type transthyretin amyloidosis (WT-TTR) (2).
Cardiac amyloidosis (CA) carries a poor prognosis. Amyloid is deposited throughout the heart, leading to thickening of the walls and to impairments in systolic and diastolic left ventricular (LV) function. However, the pathophysiology of the regional contractile abnormalities seen in CA and their potential relationships with the amyloid deposits are unclear.
Several recent studies established that myocardial longitudinal strain (LS) and strain rate determination by 2-dimensional (2D) speckle-tracking echocardiography was effective in detecting subtle systolic function impairments in a variety of diseases (3) and in evaluating the prognosis of CA (4–6). However, most studies of CA were confined to patients with AL, and whether their findings are applicable to transthyretin amyloidosis remains unclear. Cardiac magnetic resonance (CMR) also provides information on systolic function myocardial tissue characteristics. Late gadolinium enhancement (LGE) during CMR has been proven helpful for estimating the cardiac amyloid burden (7,8).
Combined echocardiography and CMR may help to elucidate the relationship between cardiac systolic function and myocardial tissue composition in various types of CA (7,9). However, CMR is costly and of limited availability and cannot be repeated in all patients. An echocardiographic tool capable of predicting LGE would help to identify those patients most likely to benefit from CMR, as well as patients who have a poor prognosis.
The primary objective of this study was to compare 2D LS echocardiography with CMR in patients with CA. Our secondary objectives were to assess correlations between histological and imaging study (echocardiography and CMR) evidence of amyloid deposition and to evaluate the prognostic usefulness of echocardiography and CMR interpreted in the light of the clinical and laboratory features.
All patients provided written informed consent. The study was approved by our local ethics committee (Créteil) and by the French Comité National Informatique et Liberté (CNIL no. 1431858). Data collection was approved by DIRC Ile de France (DC 2009-930).
Diagnosis of amyloidosis and CA
M-TTR amyloidosis was diagnosed based on identification of a TTR gene mutation combined with Congo red staining and anti-TTR antibody labeling of an endomyocardial or extracardiac biopsy specimen. The diagnosis of AL amyloidosis relied on high monoclonal protein levels in serum and/or urine and on an endomyocardial or extracardiac biopsy showing both Congo red staining and labeling with specific anti-κ or anti-λ light-chain antibodies. Amyloidosis in WT-TTR was diagnosed when endomyocardial biopsy exhibited both red Congo staining and labeling with anti-TTR antibodies or positive staining on extracardiac biopsy associated with a strong cardiac uptake of technetium 99 bisphosphonate during scintigraphy in absence of any TTR mutation. Cardiac involvement in patients with amyloidosis was defined as an interventricular septum thickness (IVST) ≥12 mm.
Among consecutive patients with suspected amyloidosis referred to the multidisciplinary Henri Mondor Amyloidosis Network between July 2009 and September 2013 and evaluated prospectively, those given a definite diagnosis of CA were included in the study provided that they had 3 available apical echocardiography views allowing global LVLS measurement. Exclusion criteria were amyloidosis types other than AL, M-TTR, and WT-TTR; severe aortic stenosis (aortic valve area <1 cm2); history of myocardial infarction; and history of intravenous chemotherapy or stem cell transplantation before the baseline echocardiogram.
Echocardiography was performed at baseline and analyzed offline by 2 experienced cardiologists (J.T. and D.B.) blinded to amyloidosis type. All acquisitions were recorded digitally over 3 consecutive cycles and were stored in raw format.
The LS values of each of the 17 LV segments defined by the American Heart Association were obtained, and LV global LS was calculated as the mean of these 17 values. Average LS was also calculated for 3 LV sections (apical, midcavity, and basal) (10). Mean basal free-wall strain was defined as the mean of the values for the basal inferior, inferolateral, and anterolateral segments. The relative “apical sparing” was calculated by the following equation: average apical LS / (average basal LS + average mid-LS) at the end of the study (11) (Online Appendix).
Cardiac magnetic resonance
Cardiac magnetic resonance (CMR) was performed within 5 days after echocardiography. All acquisitions were recorded as previously reported (8) and analyzed by 2 experienced radiologists (J.F.D. and J.T.) blinded to the patients’ clinical data. LGE was defined as abnormal myocardial hyperintensity on inverse recovery sequences 10 min after an intravenous injection of 0.2 mmol/kg gadolinium. The optimal inversion recovery time was defined as previously described (8). The image acquisition time ranged from 12 to 20 s depending on the heart rate. Five phase-sensitive inversion recovery images were acquired in the short-axis plane encompassing the left ventricle. A single slice was acquired in the 4-chamber and 2-chamber views. Each of the 17 LV segments was evaluated separately for LGE. The total number and names of the involved segments were specified for the entire left ventricle (LV LGE) and at each of 3 LV levels (apical, midcavity, and basal) (Online Appendix).
The percentage of segments with positive LGE was calculated for each LV region. Mean basal free-wall LGE was defined as the mean of the number of segments involved in the basal inferior, inferolateral, and anterolateral segments.
The hearts from the 3 patients who underwent heart transplantation (2 with AL and 1 homozygous for the TTR Val122Ile variant) were examined. Each heart was cut in 4 parts: basal, median, apical, and apex. Each part was then subdivided according to the American Heart Association 17-segment model. A slice of 5 mm at the middle of each segment was then cut and embedded in paraffin. Sections 6 μm thick were stained with Congo red to detect amyloid deposits. Amyloid deposits were identified as areas producing apple-green birefringence under polarized light. A standard light microscope was used to image fields covering each entire section at 5× magnification. Amyloid deposits were quantified using ImageJ software (National Institutes of Health, 2004). Stained surface areas were measured and expressed as a percentage of the total surface area of each segment. The assessor was blinded to the patient’s clinical data.
Outcomes and follow-up
Prospective follow-up began at the completion of the first echocardiogram and consisted of regular physician visits. The composite primary endpoint was the occurrence of a major adverse cardiac event (MACE), i.e., death, heart transplantation, or new-onset acute heart failure defined as hospital admission and intravenous diuretic treatment. No patient was lost to follow-up during the study.
Continuous variables were reported as mean ± SD or median (25th and 75th percentiles) depending on distribution. Dichotomous data were expressed as percentages. Differences in frequencies of quantitative variables were compared using the Pearson chi-square test with Yate’s correction. For continuous data, the Mann-Whitney test was used to compare 2 groups and the Kruskal-Wallis test to compare 3 groups. No adjustments were made for multiple pairwise comparisons among the 3 groups of cardiac amyloidosis. Correlation analyses were used to compare the relationship between echocardiographic parameters between themselves (LV global LS, left ventricular ejection fraction [LVEF], and IVST) or the percentage of positive LGE segment with amyloid deposits by comparing Pearson correlation coefficients. Early prognostic markers were identified by univariate analysis of baseline characteristics in patients who did and did not reach the primary endpoint (MACE) within 6 months. To avoid bias, patients without MACE who were followed for <6 months were excluded from this analysis.
Univariate analysis was performed. Variables with p values <0.05 were then entered into multivariate Cox proportional hazard models for the overall follow-up. Time-to-event analyses were then performed using backward stepwise selection to identify independent prognostic factors including all the patients in the analysis. We built 4 types of models, based on clinical (model 1), echocardiographic (models 2 and 2′), CMR (model 3), and laboratory (model 4) data. We then assessed 2 models including all variables associated with significant p values in the previous models. The hazards ratios and their 95% confidence intervals were determined using the Wald method.
To identify the best cutoffs for the independent predictors of MACE identified by the multivariate analysis, we constructed receiver-operating characteristic curves using data from the first 6 months of follow-up (see previous text), and we computed the areas under the curves. The cutoffs producing the best sensitivity and specificity for predicting MACE within 6 months were identified using Youden’s test. We then plotted cumulative event curves using the Kaplan-Meier survival method for the overall follow-up and performed a log-rank analysis to test the differences between curves for statistical significance. The statistical analysis was performed using SPSS version 19 software (SPSS Inc., Chicago, Illinois). Values of p < 0.05 were considered statistically significant.
Of the 169 patients referred for suspected amyloidosis, 79 were given a definite diagnosis of CA and were included in the study (Figure 1). In the M-TTR group, the main TTR mutations were Val30Met (31%), Val122Ile (22%), Ser77Tyr (14%), Ser77Phe (6%), and Val107Ile (6%). There were no major differences in clinical characteristics across the 3 types of amyloidosis except for age (Table 1). Preserved (LVEF >50%) was observed in 59.5% of the overall population. Restricted LV filling pattern (E/A ≥2) was found in 45% of all CA. Ischemic and increased filling pressure myocardial biomarkers (i.e., troponin and N-terminal pro–B-type natriuretic peptide [NT-proBNP]) were similarly increased in the 3 CA groups.
LV 2D echocardiography LS
LVLS was severely impaired (overall mean, −10 ± 4%) in all 3 amyloidosis groups (Table 2). LVLS correlated significantly with LVEF (r = 0.66) (Figure 2A) and IVST (r = 0.36); IVST values differed somewhat according to the type of amyloidosis (Figure 2B). The global LS-IVST slope coefficient of the regression line seems to be greater in the AL group than in the other 2 groups, suggesting worse impairment of LV contractility in AL than in M-TTR or WT-TTR for a given cardiac amyloid infiltration burden. Myocardial deformation analysis showed a basal-to-apical gradient (Figure 3A). However, 52% of patients overall did not exhibit apical sparing (apical LS/[basal LS + midcavity LS] <1) (56% in the AL group, 56% in the M-TTR group, and 38% in the WT-TTR group).
Correlations linking LS and CMR data to histopathological findings
Table 3 reports the LGE data from the 53 patients who underwent CMR. LGE was seen in the left ventricle in most patients and most LV segments. LV LGE was most common in the basal section (Figure 3B), and all patients with LGE had this abnormality in the basal inferolateral segment.
LS correlated with the percentage of LV segments exhibiting LGE in each of the 3 LV sections (basal, midcavity, and apical), as well as in the 3 segments forming the basal free wall (Figure 4A). Inferior wall LS was accurate in predicting LV LGE (area under the curve, 0.97; p = 0.002; 95% confidence interval: 0.92 to 0.99). LGE was seen consistently in all 6 basal segments in the WT-TTR group. The mean number of segments with LGE in the basal free wall was significantly lower in the AL group than in the other 2 groups (p = 0.012). In each of the LV sections, the percentage of LV segments exhibiting LGE correlated negatively with the mean LS (Figure 4A).
For each of the 17 segments, LS correlated strongly (mean r = 0.72) with the amyloid deposition measured histologically in the 3 explanted hearts (Figure 4B). In these 3 hearts, all LV segments exhibited LGE.
Identification of prognostic factors
Median follow-up was 11 months (range 4 to 17 months) overall and 13 months (range 7 to 23 months) in survivors. At least 1 MACE occurred during the overall follow-up in 36 patients (46%). Regarding the first 6 months, 19 patients (26%) experienced a first MACE; 8 of whom died and 11 had acute heart failure. Cause of death was refractory heart failure in 5 patients, pulmonary embolism in 1 patient, and cardiac arrest in 2 patients. After this episode of decompensation, as a second event, 6 died within the first 6 months of follow-up, 2 underwent heart transplantation, and 1 experienced a second episode of acute heart failure. One patient underwent heart transplantation after the first 6 months of follow-up.
Table 4 reports the results of the univariate analysis. Variables with a p value <0.05 were included in the multivariate model (Table 5). Independent predictors of MACE for the overall follow-up were New York Heart Association (NYHA) functional classes III to IV heart failure, apical LS, LVEF estimated by echocardiography, and NT-proBNP level. The best cutoff values were −14.5% for apical LS (area under the curve, 0.78; 95% confidence interval, 0.65 to 0.9) and 4,000 ng/l−1 for NT-proBNP (area under the curve: 0.76; 95% confidence interval: 0.63 to 0.88). All p values were <0.05.
Figure 5 reports the relative Kaplan-Meier analyses for the 3 variables independently associated with MACE according to the cutoffs identified by receiver-operating characteristic curve analysis. Patients with NYHA functional classes III to VI heart failure, apical LS >−14.5%, or NT-proBNP >4,000 ng/l−1 were at greatest risk of MACE.
Our study provides information on the usefulness of speckle-tracking echocardiography and CMR in the 3 main types of CA and suggests hypotheses regarding the pathophysiology of amyloidosis-induced LV dysfunction. Global and regional LS values were similarly impaired, and a basal-to-apical gradient was found in all 3 types of CA. LVLS impairment reflected the amyloid burden and LGE during CMR was associated with LVLS impairment. Apical LS, NT-proBNP, and NYHA functional class independently predicted the occurrence of MACE.
Global and regional LVLS in the 3 types of CA
CA is usually diagnosed late. Noninvasive imaging methods that provide LS estimates on the basis of myocardial deformation can detect regional myocardial dysfunction (3). In most of the available studies of amyloidosis, LS was estimated from myocardial velocities measured using tissue Doppler imaging. Limitations of tissue Doppler imaging include angle dependency, interobserver and intraobserver variability, and the occurrence of artifacts (12). Two-dimensional speckle-tracking echocardiography has been found more sensitive and more reproducible than Doppler imaging and has shown promise for the early diagnosis of CA (4,13). Most studies of imaging techniques for establishing the diagnosis and prognosis of CA were confined to patients with AL (5,6). We reported recently that global LS determined using 2D speckle-tracking echocardiography was capable of detecting early LV systolic dysfunction in patients with M-TTR CA, even those who were free of symptoms and had normal LVEF values (14). In the present study, LVEF was within the normal range in more than one-half of the patients.
The relative apical sparing first reported in AL CA (15) was also noted in our patients with M-TTR CA and WT-TTR CA, in keeping with a recent study (16). However, this study (16) also showed greater global LS impairment in AL and WT-TTR CA than in M-TTR CA, whereas LS values showed no significant differences across the 3 amyloidosis types in our population. This discrepancy may be ascribable differences in CA severity between the 2 populations.
Amyloid burden correlates with impaired LS and LGE
We demonstrated a strong negative correlation between the amyloid burden measured by histopathology and segmental LVLS in all 3 types of amyloidosis. Amyloid deposits were more abundant in the basal and midcavity sections of the explanted hearts. In these sections, LGE was more marked and was associated with LS impairment, indicating regional contractile dysfunction. All patients with impaired LS in at least 1 basal segment exhibited LGE. Thus, LS values in the basal segments may predict CMR abnormalities.
The pathophysiology of the basal-to-apical LS gradient in CA is unclear. Current hypotheses involve higher wall stress (17,18) and a greater tendency toward apoptosis and remodeling (18) in the basal segments related to turbulent flow in the LV outflow tract. The greater amyloid deposition in the basal segments may be explained by the high mechanical displacement of cardiac myocytes in these segments. Phelan et al. (11) have shown that the LV wall thickness determined by CMR at the basal and midcavity segments increased more than at the apex in patients with CA, suggesting relatively less amyloid deposition in the apex (11). The regional differences in LGE reflect differences in the topographic distribution of the amyloid deposits, as established by the comparison of histological and CMR findings in 3 of our patients and a previously described patient (9). However, in another study, there was no association between the extent of LGE and amyloidosis load. Of note, in the Hosch et al. (19) study, both amyloidosis and LGE were evaluated visually on a 4-point scale (0 to 3).
Extracellular matrix remodeling associated with amyloid deposition may contribute to impaired contractility (20,21). The mechanical effect of amyloid deposits and the direct toxic effects of amyloid on cells may vary across CA types (1,2,20–22). Compared with the 2 transthyretin groups in our study, the AL group had less amyloid deposition as assessed based on LV wall thickness and less LGE in the basal section but similar alterations in LVLS. This suggests that amyloid AL deposits have greater toxicity than the 2 other forms. In agreement, Brenner et al. (21) reported that human amyloidogenic light-chain oligomers could directly impair contractility through oxidative stress in the cardiomyocytes.
Prognostic factors in CA
Independent prognostic factors in our study were NYHA functional classes III to IV, NT-proBNP >4,000 ng/l, and apical LS >−14.5%. NYHA functional classes III to IV was 1 of the factors independently associated with death in 2 studies including patients with AL (23,24).
NT-proBNP is an independent predictor of poor cardiac outcomes, particularly in AL (25–27). In patients with M-TTR, increased in NT-proBNP is associated with early cardiac involvement (28). Furthermore, NT-proBNP >3,400 ng/l independently predicted death in 58 patients with transthyretin CA due to the Thr60Ala mutation (29). NT-proBNP is released in response to ventricular pressure and stretch (30–32). In CA, NT-proBNP release may be related to increased LV filling pressure and to cardiomyocyte compression and stretching by amyloid deposits (30).
In our study, apical LS, but not LS in the basal or midcavity sections or global LS, independently predicted MACE. Given that the apex is relatively spared, LS impairment at the apical section may indicate severe amyloid deposition and tissue remodeling. Our results are at variance with those of a study in AL-CA, in which mean basal LS was the only independent predictor of death (6). However, in this study, LS was computed using Doppler imaging and included only AL patients. In another study of AL-CA, midseptum LS >−11%, together with NYHA functional class, independently predicted death (33). These discrepancies may be ascribable to differences in CA severity across studies or different echocardiographic technics.
The limitations are the single-center design of our study, the number of patients in each amyloidosis subgroup. Two-dimensional strain imaging is not standardized, and the LS cutoff identified in our study may not apply to imaging with systems other than that used in our study (Echopac GE, GE Healthcare, Horten, Norway) (28). We chose 10 min for LGE because there is no recommended timing in CA. In the literature, this ranges from 5 to 20 min (5,34–38). Studies are needed to determine the optimal time for this disease. We used a semiquantitative method to measure segmental LGE, as no standardized quantitative technique is available (14). Finally, our histopathological study was limited to 3 patients, although it included 51 segments. Studies including more explanted hearts are needed to confirm our histopathological data.
Our study adds new information on the effect of the quantity and location of amyloid deposits, their relation with structural CMR findings, and their effect on LV longitudinal function. We also identified factors associated with the prognosis of all 3 major forms of CA.
COMPETENCY IN MEDICAL KNOWLEDGE 1: Amyloidosis results from accumulation of insoluble proteins in the extracellular matrix in various organs including the heart. The 3 major types involved in this process are immunoglobulin light chains (AL), m-TTR, and WT-TTR.
COMPETENCY IN MEDICAL KNOWLEDGE 2: LV dysfunction evaluated by LS with a basal-to-apical gradient and LGE on CMR are common in CA.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: CA has a poor prognosis and should be considered in the differential diagnosis of patients with increased LV wall thickness. When suspected, patients should be referred for genetic testing, bone scintigraphy, and biopsy analysis.
TRANSLATIONAL OUTLOOK: This study investigated the relationship between LS and the severity of amyloid deposition according to the American Heart Association 17-segment model. LS was impaired in the LV free wall and correlated with the severity of amyloid deposition. LS provides an indirect evaluation of the extent of LV amyloid infiltration. The best LS segment for prognostication was the apical section (cutoff, −14.5%). In future, additional research using new imaging techniques should be undertaken to identify and validate specific diagnostic markers.
The study was funded by Henri-Mondor Teaching Hospital. Prof. Damy has received grants and consultant fees from Pfizer and consultant fees from Alnylam. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- light-chain amyloidosis
- cardiac amyloidosis
- cardiac magnetic resonance
- interventricular septum thickness
- late gadolinium enhancement
- longitudinal strain
- left ventricular
- left ventricular ejection fraction
- major adverse cardiac event(s) (death, heart transplantation, or new-onset acute heart failure)
- hereditary transthyretin amyloidosis
- N-terminal pro–B-type natriuretic peptide
- New York Heart Association
- wild-type transthyretin amyloidosis (senile amyloidosis)
- Received February 10, 2015.
- Revision received April 23, 2015.
- Accepted May 14, 2015.
- American College of Cardiology Foundation
- Fontana M.,
- Banypersad S.M.,
- Treibel T.A.,
- et al.
- Koyama J.,
- Falk R.H.
- Syed I.S.,
- Glockner J.F.,
- Feng D.,
- et al.
- Perugini E.,
- Rapezzi C.,
- Piva T.,
- et al.
- Cerqueira M.D.,
- Weissman N.J.,
- Dilsizian V.,
- et al.
- Phelan D.,
- Collier P.,
- Thavendiranathan P.,
- et al.
- Tedford R.J.,
- Hassoun P.M.,
- Mathai S.C.,
- et al.
- Quarta C.C.,
- Solomon S.D.,
- Uraizee I.,
- et al.
- Jiang L.,
- Huang Y.,
- Hunyor S.,
- dos Remedios C.G.
- Liao R.,
- Jain M.,
- Teller P.,
- et al.
- Brenner D.A.,
- Jain M.,
- Pimentel D.R.,
- et al.
- Shi J.,
- Guan J.,
- Jiang B.,
- et al.
- Palladini G.,
- Campana C.,
- Klersy C.,
- et al.
- Dispenzieri A.,
- Gertz M.A.,
- Kyle R.A.,
- et al.
- Kristen A.V.,
- Giannitsis E.,
- Lehrke S.,
- et al.
- Sattianayagam P.T.,
- Hahn A.F.,
- Whelan C.J.,
- et al.
- Koyama J.,
- Ray-Sequin P.A.,
- Falk R.H.
- Austin B.A.,
- Tang W.H.,
- Rodriguez E.R.,
- et al.
- White J.A.,
- Kim H.W.,
- Shah D.,
- et al.
- Karamitsos T.D.,
- Piechnik S.K.,
- Banypersad S.M.,
- et al.