Author + information
- Received November 11, 2013
- Revision received March 5, 2014
- Accepted March 6, 2014
- Published online October 1, 2014.
- Adelina Doltra, MD∗∗ (, )
- Bart Bijnens, PhD†,
- José M. Tolosana, MD∗,
- Roger Borràs, BSc∗,
- Malek Khatib, MD∗,
- Diego Penela, MD∗,
- Teresa Maria De Caralt, MD, PhD∗,
- María Ángeles Castel, MD, PhD∗,
- Antonio Berruezo, MD, PhD∗,
- Josep Brugada, MD, PhD∗,
- Lluís Mont, MD, PhD∗ and
- Marta Sitges, MD, PhD∗
- ∗Thorax Institute, Hospital Clinic, Institut d’Investigacions Biomèdiques August Pi i Sunyer, University of Barcelona, Barcelona, Spain
- †ICREA, Universitat Pompeu Fabra, Barcelona, Spain
- ↵∗Reprint requests and correspondence:
Dr. Adelina Doltra, Thorax Institute, Hospital Clinic, Villarroel 170, 08036 Barcelona, Spain.
Objectives Our aim was to identify “correctable abnormalities” using conventional grayscale and blood-pool Doppler echocardiography and evaluate their ability to predict both response and midterm survival.
Background Identification of mechanical abnormalities that may be corrected with cardiac resynchronization therapy (CRT) is useful for predicting echocardiographic response at 1-year follow-up.
Methods A total of 200 CRT patients were included. Clinical evaluation and echocardiography were performed before and after CRT to assess the presence of the mechanical abnormalities of interest (septal flash, abnormal ventricular filling, or exaggerated interventricular dependence). Response to CRT was defined as a reduction in left ventricular (LV) end-systolic volume (ESV) ≥15%. Four subgroups of extent of response were defined: LVESV reduction >26.68% (extensive remodeling); LVESV reduction 6.8% to 26.68% (slight remodeling); LVESV reduction <6.8% (no remodeling) and clinical response; and LVESV reduction <6.8% without clinical response or the occurrence of death or heart transplantation. Midterm cardiovascular survival was evaluated (mean follow-up 38 ± 19 months).
Results The presence of a correctable abnormality was independently associated with a better rate (odds ratio: 0.03 [95% confidence interval (CI): 0.01 to 0.10], p < 0.001) and extent of response to CRT (n = 59 [96.7%] for the extensive remodeling subgroup vs. n = 53 [85.5%] for the slight remodeling subgroup vs. n = 19 [47.5%] for the no remodeling with clinical response subgroup vs. n = 17 [45.9%] for the no remodeling without clinical response subgroup, p = 0.0001), as well as with increased midterm survival (hazard ratio: 0.11 [95% CI: 0.2 to 0.6]). Other independent predictors included creatinine level and LV end-systolic diameter for response; New York Heart Association functional class IV, creatinine, LV end-systolic diameter, and transmurality index for extent of response; and New York Heart Association functional class IV for cardiovascular mortality.
Conclusions The presence of a correctable abnormality evaluated by conventional echocardiography is associated with LV reverse remodeling and better survival at midterm follow-up. Clinical characteristics and myocardial viability also have an influence.
Although cardiac resynchronization therapy (CRT) is an established therapeutic option for patients with advanced heart failure, the significant rate of nonresponders has also been repeatedly demonstrated (1–3). This has led to intensive research on dyssynchrony indexes to improve response to CRT (4–6), which, despite promising initial results, have failed to show any benefit when applied in a multicenter setting such as the PROSPECT (Predictors of Response to CRT) study (7). Recently, a mechanism-focused approach has been proposed for selecting candidates to CRT, in which left ventricular (LV) dyssynchrony is only 1 of the mechanisms assessed (8). This approach would take into account all potential mechanical abnormalities amenable to correction with CRT, without relying on sophisticated and poorly reproducible methods.
However, differences between studies in the rate of CRT response may be partially explained by the lack of a single parameter to define response (9). Both clinical and echocardiographic criteria have been used, making results from different studies difficult to compare. Although a reduction in left ventricular end-systolic volume (LVESV) secondary to CRT has been linked with survival (10), not all patients remodel to the same extent, whereas others may experience a clinical benefit without reverse remodeling (11).
The present study had 4 objectives: 1) to validate a simple approach to recognize mechanical abnormalities amenable to correction with CRT; 2) to compare its predictive value against other proposed parameters for CRT selection; 3) to investigate how this approach relates to the extent of response; and 4) to study whether it predicts survival.
A total of 200 patients underwent CRT implantation from January 2005 through December 2009. The indications for implanting a CRT device were taken from international guidelines: LV ejection fraction (EF) ≤35%, QRS ≥120 ms, and New York Heart Association (NYHA) functional class II to IV despite optimal medical therapy for at least 2 months. Candidates awaiting heart transplantation or with significant comorbidities that shortened life expectancy were excluded. The study protocol was accepted by our hospital’s ethics committee and conformed with the Declaration of Helsinki. Informed consent was obtained from all participants.
The study protocol included a baseline clinical and echocardiographic examination, which was repeated 24 to 72 h after CRT implantation and at 12-month follow-up. Clinical evaluation included NYHA functional class assessment and a 6-min walk test. Additionally, mortality was assessed (see the following text). In 82 patients, a baseline contrast-enhanced cardiac magnetic resonance (CMR) was also performed.
Patients received a CRT pacemaker or a CRT defibrillator, according to clinical indications. A total of 141 patients (70.5%) received a CRT defibrillator, whereas 59 (29.5%) received a CRT pacemaker. One lead was placed at the right ventricular apex, another was implanted through a distal cardiac vein into the posterolateral wall (or in the most laterally located available vein if necessary), and a third was placed in the right atrium if the patient had sinus rhythm. The coronary sinus was catheterized using a guiding catheter. All leads were implanted transvenously.
A comprehensive echocardiography-Doppler examination (Vivid 7, GE Healthcare, Milwaukee, Wisconsin) was performed. The echocardiography examination included 2-dimensional grayscale, color, and spectral blood-pool Doppler and tissue Doppler imaging. Three cardiac cycles were obtained for each acquisition. All studies were stored and post-processed offline using EchoPAC version 7.0.1 (GE Healthcare). LV end-diastolic diameter and left ventricular end-systolic diameter (LVESD) were measured from M-mode in the parasternal long-axis view, whereas LV end-diastolic volume, LVESV, and EF were quantified by the Simpson method.
The echocardiography examination also included the evaluation of mechanical abnormalities amenable to correction with CRT, as previously described (8). Patients were divided into 5 subgroups, according to the presence of these abnormalities (Figure 1):
• Septal flash (SF), defined as a fast contraction and relaxation (inward/outward motion) of the septum occurring during the isovolumetric contraction period (within the QRS width), as visualized either in 2-dimensional or M-mode images in the parasternal or 4-chamber views (SF—first subgroup).
• Abnormalities in LV filling in the absence of SF, including either the presence of fused E and A waves with diastolic mitral regurgitation (long atrioventricular [AV] delay—second subgroup) or a truncated A-wave (short left AV delay—third subgroup). The short left AV subgroup reflects a delayed left atrial mechanical activation; in this abnormality, a delayed interatrial conduction leads to a late left atrium activation, delayed left atrial contraction, and shortened emptying (12).
• Exaggerated right–left interaction without a concomitant SF or an abnormal filling (fourth subgroup), defined as the presence of an abnormal passive motion of the septum secondary to a prolonged interventricular delay (difference between the left and right pre-ejection periods, abnormal if >40 ms). Patients with prolonged interventricular delay were only classified as having exaggerated right–left interaction if neither SF nor abnormalities were present and a visual abnormal motion of the septum was observed, that is, septal motion related to the contraction of the free wall of the right ventricle and different from the SF.
• The fifth subgroup included those patients without any of the aforementioned abnormalities.
The presence and correction of correctable abnormalities with CRT was evaluated early after the implant (24 to 72 h after CRT implantation) and at 12-month follow-up.
Additionally, the LV filling time and the time difference of septal-to-lateral peak systolic velocities assessed by Doppler tissue imaging were determined (septal-to-lateral delay) (5). The LV filling time was measured from the onset of the E-wave to the end of the A-wave, and the R-R interval was measured to calculate the percentage of filling time relative to the cardiac cycle. The echocardiography examinations were analyzed by a cardiologist blinded to clinical and outcome data.
Cardiac magnetic resonance
At 24 to 72 h before CRT device implantation, CMR examinations were performed with a 1.5-T Signa magnetic resonance scanner (GE Healthcare). Delayed-enhancement (DE) images were obtained 10 min after injecting 0.2 mmol/kg body weight gadodiamide (Omniscan, GE Healthcare Buchler, Munich, Germany) using an inversion recovery gradient echocardiography sequence. The DE images were analyzed by an investigator blinded to clinical and echocardiography data, using a 17-segment model. Each LV segment was given a score from 0 to 4 according to the extent of DE (0 = no DE, 1 = DE 1% to 25% of wall thickness, 2 = 26% to 50%, 3 = 51% to 75%, and 4 = >75%). The sum of all scores was divided by 17 to obtain the transmurality index.
Extent of response
Considering the difficulty of classifying CRT response as a dichotomous variable and in order to analyze the extent of response achieved, we created 4 subgroups of patients with different degrees of response at 1-year follow-up, including clinical and echocardiographic evolution. The decrease in LVESV at 12 months was divided in tertiles to obtain reverse remodeling cutoff values. The subgroups were constituted as follows: patients in the first tertile and no clinical response, patients in the first tertile and clinical response, patients in the second tertile, and patients in the third tertile.
Patients were followed-up 12 months after device implantation at the outpatient clinic. Echocardiographic response to CRT was defined as a reduction of LVESV ≥15% at follow-up (11) in the absence of death or heart transplantation. Clinical responders were defined as alive, without heart transplantation, and with improvement of ≥20% in the 6-min walk test or of at least 1 NYHA functional class.
Mortality data were collected by reviewing outpatient clinical history or by phone interviews with relatives. Two cardiologists (J.M.T. and M.K.) reviewed the data and assigned the mode of cardiac death, by consensus. Deaths were categorized as cardiac, noncardiac, or unknown if the cause of death could not be determined. Cardiac mortality was defined as sudden (not preceded by heart failure or ischemic symptoms) or as the result of heart failure.
Continuous variables are presented as mean ± SD. To compare means of 2 variables, we used the Student t test. Categorical variables were expressed as total number (percentages) and compared between groups using the chi-square test. Logistic regression analysis was used to study the effects of baseline characteristics to evaluate the echocardiographic response at 1 year (transversal endpoint). Multinomial logistic regression was used to evaluate the association between baseline characteristics and extent of response. To evaluate cardiovascular mortality during the follow-up, we used Cox proportional hazards model analysis (longitudinal endpoint). Univariate regression models were conducted for each of the potential predictors. A p value <0.10 was used for the purpose of screening covariates. Backward stepwise selection algorithms were used for selecting the covariates into the multivariate regression model. At each step, the least significant variable was discarded from the model. Only covariates with a p value <0.10 remained in the final model. Odds ratios and hazard ratios with 95% confidence limits were also calculated. A 2-sided type I error of 5% was used for all tests. For the analysis of predictors of echocardiographic response, extent of response, and cardiovascular mortality, correctable mechanical abnormalities were analyzed as a binary variable (presence or absence of any of the correctable mechanisms). Kappa statistic was used to assess interobserver concordance. Statistical analysis was performed using R software for Windows version 2.14.1 (R Project for Statistical Computing, Vienna, Austria).
Table 1 summarizes the baseline characteristics of the studied population. No patients were excluded based on missing examinations or poor image quality. The mean follow-up was 38 ± 19 months. Of the 200 patients included, 105 (52.5%) responded to CRT at 1-year follow-up, according to echocardiographic volume criteria, and 95 (47.5%) were nonresponders. During follow-up, 46 (23%) patients died: 30 (65.2%) of cardiovascular causes, 8 (17.4%) of noncardiac causes, and 8 (17.4%) were recorded as unknown cause of death. The 8 patients with a noncardiovascular cause of death and those 3 that were lost to follow-up were excluded from mortality analysis. Only 1 death (in a patient that received a CRT pacemaker) was due to an arrhythmic event, whereas the remaining deaths were secondary to heart failure. Eleven patients died in the first 12 months after CRT implantation. Those patients were classified as echocardiography nonresponders, according to our pre-specified criteria. Those patients were more often in NYHA functional class IV (6 [54.5%] vs. 14 [7.4%], p < 0.0001), more often had atrial fibrillation (8 [72.7%] vs. 59 [31.2%], p = 0.008), had a significantly lower LVEF (19.45 ± 4.23% vs. 24.33 ± 6.40%, p = 0.013), and less often had a correctable abnormality was (3 [27%] vs. 145 [76.7%], p = 0.001).
Mechanical abnormalities and response to CRT
A total of 122 patients (61%) corrected the baseline abnormality with CRT. In most of them, the correction was already present in the early echocardiogram (84%), whereas in the remaining 16%, an early improvement was observed, followed by a resolution at 12 months.
Of 200 patients included, 106 (53%) presented with SF at baseline; the rate of echocardiographic response at 1 year in this subgroup was 80.2% (n = 85), and CRT corrected this abnormality in 93% of them. Of the 21 nonresponders, SF persisted in 7 (33.3%) at follow-up, and the 14 remaining patients had either a small SF at baseline or an advanced disease (renal insufficiency or severely dilated hearts). The long-AV group contained 26 (13%) patients, with a response rate of 38.5% (10), and the abnormality was corrected in all of them; of the 16 nonresponders, it was not corrected in 7 (43.8%). The short left AV subgroup had 11 patients and a response rate of 63.6% (7); of the 4 nonresponders (36.4%), short left AV persisted in 3 (75%) at follow-up. Only 5 patients (2.5%) were classified as having an exaggerated right–left interaction; none of them responded, despite improvement in interventricular dyssynchrony in 3 (60%). The remaining 52 patients (26%) constituted the fifth subgroup (others), and 49 (94.2%) were nonresponders. Three (5.8%) patients responded despite no mechanical abnormality being identified at baseline (Figure 2).
To assess the interobserver reproducibility of classifying patients according to the underlying correctable abnormality, 40 random patients were reanalyzed by a second observer blinded to the results of the first observer. The concordance between the 2 observers was very good (Kappa = 0.76, p < 0.0001, misclassification error 0.17). When the concordance was analyzed according to the presence of any correctable abnormality versus no correctable abnormality, we obtained also a very good interobserver concordance (Kappa = 0.78, p < 0.0001, misclassification error 0.07). Of note, there was an absolute agreement between the 2 observers regarding the presence of SF (Kappa = 1, misclassification error 0).
The variables analyzed are summarized in Table 2. Univariate predictors of echocardiographic response at 1 year were female sex, NYHA functional class IV, atrial fibrillation, presence of a correctable mechanical abnormality, baseline plasma creatinine level, LV diameters, and baseline QRS. In the multivariate analysis, the independently related variables were the presence of a correctable abnormality, creatinine, and LVESD (Figure 3).
Extent of response
The decrease in LVESV at 12-month follow-up was divided into tertiles, obtaining the following cutoff values: reduction of LVESV >26.68% versus baseline; reduction of LVESV between 26.68% and 6.8%; and volumetric reduction <6.8% or LVESV increase at follow-up. The 4 subgroups of response were: 1) LVESV reduction >26.68%; 2) LVESV reduction of 6.8% to 26.68%; 3) LVESV reduction <6.8% but with clinical response (as previously defined); and 4) LVESV reduction <6.8% without clinical response or the occurrence of death or heart transplantation at 1-year follow-up.
Table 3 and Figure 4 summarize our results. The extent of response was significantly related to NYHA functional class IV, baseline plasma creatinine level, baseline LVESD, CMR transmurality index, and the presence of a correctable mechanical abnormality.
Midterm cardiovascular survival
In the univariate analysis, baseline creatinine level, atrial fibrillation, NYHA functional class IV, LVEF, transmurality index, and the presence of a correctable mechanical abnormality were associated with midterm survival. Only the presence of NYHA functional class IV and a correctable mechanical abnormality were independent predictors under multivariate analysis (Table 4). Figure 5 shows the Kaplan-Meier survival curves according to the presence of any correctable mechanical abnormality and the impact of its correction.
In the present study, a mechanical abnormality correctable with CRT and easily recognized on standard echocardiography was a key factor in determining CRT response and was independently associated with a high likelihood of response. Response is not restricted to LV “dyssynchrony”; it includes other correctable mechanical abnormalities. Some of these (SF and short left AV) are more strongly correlated with response than others. Conversely, patients without any of the mechanical abnormalities studied are most likely nonresponders. Other results were that the presence of a transmural scar determines the extent of reverse remodeling, and the presence of a correctable mechanical abnormality is associated with better survival at midterm follow-up, provided that it is corrected with CRT.
A mechanistic approach to select CRT candidates
Besides technical limitations, another explanation for the negative results of the PROSPECT study may be that only “single parameters” to define dyssynchrony were evaluated, each of them related to only 1 aspect of “dyssynchrony.” We tested an approach that was proposed initially by Parsai et al. (8). This approach takes into account LV, AV, and interventricular dyssynchrony. Similar to the findings of Parsai et al. (8), our study demonstrated that the presence of a correctable mechanical abnormality is almost mandatory to obtain a positive response with CRT. Patients without any of them were largely nonresponders (1-year nonresponse rate 94.2%) and had higher midterm mortality.
Nonetheless, mechanical abnormalities are not all equally related to response: although the presence of SF is strongly associated with response, patients with a long AV or abnormal right–left interaction are mainly nonresponders, as demonstrated by the fact that from the 45.9% of nonresponders that had a mechanical abnormality at baseline, only 29.4% were having a SF, whereas the others were filling abnormalities or exaggerated right–left interaction. These findings suggest that the presence of SF is the mechanical abnormality most easily amenable to be corrected with CRT and most related to a clear response. That suggests that although a mechanical abnormality seems to be necessary to obtain CRT response, the probability of finally achieving it is also determined by the underlying mechanical abnormality and myocardial substrate. Other studies have also demonstrated the ability of abnormal septal motion to predict CRT response. Voigt et al. (13) referred to it as apical rocking, and demonstrated a high rate of response when it was present. Also, low-dose dobutamine increased the amplitude of apical rocking, which correlated strongly with reverse remodeling and survival (14), Additionally, Risum et al. (15) differentiated a classic SF pattern (characterized by an early contraction of the septum or anteroseptal wall and early pre-stretching and late contraction of the opposite ventricular wall) from a nonclassical pattern, and demonstrated that the classical pattern was highly predictive of response. Our data add more evidence suggesting a link between SF and high probability of response to CRT and additionally emphasize the importance of taking into account other mechanical abnormalities amenable to being corrected with CRT such as the case of AV filling abnormalities.
In a multivariate analysis, we demonstrated that the presence of these mechanical abnormalities is an independent predictor of echocardiographic response and midterm cardiovascular mortality, along with creatinine level and LVESD (reflecting “severity” of the disease). The main advantage of the proposed approach is that it would prevent the implant of a CRT device in patients without any mechanical abnormality and, thus, a strong probability of nonresponse (94.2%).
Extent of response
When the patients in our population were divided according to the degree of response achieved with CRT, the presence of a mechanical abnormality was also a significant difference between subgroups; other significant differences included transmurality index, NYHA functional class IV, baseline creatinine level, and baseline LVESD. Interestingly, the transmurality index was not significantly related to CRT response when response was analyzed dichotomously (i.e., based on LVESV reduction ≥15%). According to our results then, transmurality index would determine not the response per se, but rather, the extent of response. A patient would be able to respond to CRT if some degree of dyssynchrony was present (not necessarily at the LV level), but the extent of response achieved would be determined by the amount of myocardium that can actually be remodelled. This is in concordance with previous reports indicating a lesser extent of reverse remodeling in CRT patients with ischemic cardiomyopathy (16–18).
Influence on survival
The presence of a correctable mechanical abnormality was independently predictive of a better outcome at midterm follow-up, with reduced cardiovascular mortality at multivariate analysis. Prior studies have also demonstrated a relationship between dyssynchrony and outcome, but in contrast to the present work, had analyzed only 1 aspect of dyssynchrony (19). Our results indicate that the presence of a correctable mechanical abnormality not only detects patients with a higher probability of reverse remodeling, but also has a real impact on survival. However, the extent of response will be variable depending on other baseline parameters such as myocardial substrate (viability), underlying disease (renal insufficiency), and clinical status.
In the present study, as in previous reports (20), the division between the different subgroups of extent of response is arbitrary; therefore, slightly different results might be obtained using other cutoff values. However, changing the threshold for an absolute dichotomous echocardiographic response to CRT could produce the same effect. In our clinical practice, not all patients benefit equally from CRT, and we believe studying subgroups of extent of response is more informative than a simple cutoff value.
CMR was not performed in the whole cohort. Additionally, we did not quantify scar extent, which would give more precise information on the relationship between scar and response, and we did not analyze the effect of scar location and the relationship of scar with the other factors of response. Therefore, conclusions regarding the effect of scar on (the amount of) CRT response should be further investigated in larger studies. Additionally, we did not perform a dobutamine challenge to our patients, which has been used to depict myocardial contractile reserve (21) and also to induce a potentially correctable SF that is subtle at rest and might be misdiagnosed (14,22). Finally, we are aware that the reproducibility of the outcome measurements (i.e., LV volumes and 6-min walking test performance) might be imperfect. However, even when we considered several groups of combined and progressive responses, we could find a strong relationship between response and the presence of a correctable abnormality, supporting our hypothesis and the validity of this approach (23,24).
The number of mortality events at follow-up in our study population was relatively small; this small number of events constitutes a limitation in the survival multivariate analysis, and should be taken into account when interpreting its results. Also, we did not include appropriate implantable cardioverter-defibrillator therapies and shocks in the mortality analysis. Different results could have been obtained had this variable been included.
The presence of reverse remodeling may have contributed to the correction of baseline abnormalities, especially in those few patients in which correction was mainly observed at 12-month echocardiogram. However, even in those patients, an immediate abnormality improvement was observed, and therefore, the correction can be mostly attributed to CRT. Finally, the number of patients having mechanical abnormalities other than SF was small. Conclusions regarding those subgroups should also be taken with caution and warrant confirmation in further larger studies.
Our study supports the easy evaluation of a potentially correctable mechanical abnormality with conventional echocardiography that would lead to more adequate selection of candidates to CRT. The presence of a correctable abnormality is associated with LV reverse remodeling and survival at midterm follow-up. However, the extent of LV reverse remodeling is also influenced by other factors related to the clinical evolution of the disease (NYHA functional class) and the status of the underlying myocardium (myocardial viability) that should be taken into account when selecting potential candidates for CRT.
This work was supported by grants from the CENIT program of the Centro de Desarrollo Tecnológico Industrial (cvREMOD project), the Agència de Gestió d’Ajuts Universitaris i de Recerca (AGAUR) program of the Generalitat de Catalunya [2009-SGR-1104], the Spanish Society of Cardiology (2010), and the Spanish Government [REDINSCOR RD06/0003/0008]. Dr. Doltra has received a personal grant from Fundació Clínic. Dr. Mont is a consultant to, has received lecture fees, and has received research funding from Boston Scientific, Medtronic, St. Jude, and Sorin. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- cardiac magnetic resonance
- cardiac resynchronization therapy
- delayed enhancement
- ejection fraction
- left ventricle/ventricular
- left ventricular end-systolic diameter
- left ventricular end-systolic volume
- New York Heart Association
- septal flash
- Received November 11, 2013.
- Revision received March 5, 2014.
- Accepted March 6, 2014.
- American College of Cardiology Foundation
- St John Sutton M.G.,
- Plappert T.,
- Abraham W.T.,
- et al.
- Bax J.J.,
- Bleeker G.B.,
- Marwick T.H.,
- et al.
- Delgado V.,
- Ypenburg C.,
- van Bommel R.J.,
- et al.
- Chung E.S.,
- Leon A.R.,
- Tavazzi L.,
- et al.
- Parsai C.,
- Bijnens B.,
- Sutherland G.R.,
- et al.
- Yu C.M.,
- Bleeker G.B.,
- Fung J.W.,
- et al.
- Voigt J.U.,
- Schneider T.M.,
- Korder S.,
- et al.
- Stankovic I.,
- Aarones M.,
- Smith H.J.,
- et al.
- Sutton M.G.,
- Plappert T.,
- Hilpisch K.E.,
- Abraham W.T.,
- Hayes D.L.,
- Chinchoy E.
- Knappe D.,
- Pouleur A.C.,
- Shah A.M.,
- et al.
- Ypenburg C.,
- van Bommel R.J.,
- Borleffs C.J.,
- et al.
- Parsai C.,
- Baltabaeva A.,
- Anderson L.,
- Chaparro M.,
- Bijnens B.,
- Sutherland G.R.
- Marwick T.H.,
- Narula J.