The utility of echocardiographic right ventricular parameters in predictingresponse to cardiac resynchronization therapy in patientswith heart failure: a proof of concept

Abstract

Aim Despite the guideline-based selection of patients, failure to respond to cardiac resynchronization therapy (CRT) remains common. We aimed to assess the utility of right ventricular (RV) echocardiographic parameters to predict response to CRT. Methods A prospective study of 30 patients who were eligible for CRT. Echocardiographic RV parameters were assessed at baseline and 6 months post-CRT. CRT response was defined as a ≥ 15% reduction in left ventricular end-systolic volume. Results Statistical analysis revealed that 66.7% were CRT-responders. CRT-responders had higher fractional area change (40.9 ± 6.4 vs 30.4 ± 10.1%; p = 0.002), tricuspid annular plane systolic excursion (22.1 ± 4.9 vs 15.2 ± 3.9 mm; p = 0.001), tissue Doppler–derived tricuspid lateral annular systolic velocity (S’) (12.8 ± 2.3 vs 8.4 ± 1.8 mm; p < 0.0001), and RV myocardial performance index (0.41 ± 0.07 vs 0.54 ± 0.09, p < 0.0001) at baseline compared with non-responders. On multivariate analysis, pre-CRT S’ was the single independent predictor of CRT response (OR: 3.21, 95% CI: 1.31 to 7.82, p = 0.01). S’ >8 cm/s was 100% sensitive and 70% specific in predicting response to CRT (AUC = 0.91, p < 0.0001). Conclusion Echocardiography-derived S’ >8cm/s could be considered as a highly sensitive and specific predictor of response to CRT.

Keywords

Introduction

Cardiac resynchronization therapy (CRT) is an established treatment for patients with systolic heart failure (HF) and ventricular electrical dyssynchrony.^1,2 Despite technological advances in device implantation and programming, nearly one-third of HF patients fail to demonstrate a satisfactory response to CRT.³ Identification of clinical and imaging parameters that could predict patient’s response to CRT could facilitate appropriate patient selection for these procedures and improve efficacy. Prior studies have shown that right ventricular (RV) dysfunction is a poor prognostic factor in patients with moderate to severe chronic HF.^4,5 Furthermore, RV dysfunction is associated with worse outcomes in patients undergoing CRT.⁶ However, assessment of RV parameters to aid selection of patients for CRT has not translated into clinical practice. This is partly due to randomized studies focusing primarily on LV indices, lack of experience with echocardiographic assessment tools to appropriately assess the RV, and the paucity of studies providing normal reference values of RV size and function.⁷ In the current proof-of-concept prospective study, we sought to evaluate the feasibility of various echocardiographic RV function parameters to predict the response of HF patients to CRT.

Materials and Methods

We conducted a prospective, open-labeled, single-center study that enrolled a cohort of 30 consecutive patients who presented to our tertiary medical center between January and December 2016 with a diagnosis of chronic systolic HF and met guidelinebased indications for CRT implantation. In accordance with the current guidelines.⁸ inclusion criteria included symptomatic HF despite optimal pharmacological therapy for at least 3 months prior to inclusion, with New York Heart Association (NYHA) functional class II-IV, ejection fraction (EF) ≤ 35% and QRS duration ≥ 120 ms (in presence of LBBB) or ≥ 150 ms (in absence of LBBB). Exclusion criteria were patients with: life expectancy less than 1 year such as advanced malignancy or dementia, acute HF requiring inotropic support, permanent atrial fibrillation (AF), and previous pacemaker implantation. Written informed consent was obtained from every patient after meeting our inclusion criteria, and the study protocol was reviewed and approved by our local institutional human research committee as it conforms to the Declaration of Helsinki. Data gathered at baseline (pre-CRT implantation) included following: demographic data, HF clinical status according to NYHA functional class and Minnesota living with heart failure questionnaire (MHFQ), electrocardiogram (ECG) analysis and transthoracic echocardiography (TTE). At 6-months follow up, patients were re-evaluated at a clinic visit for HF clinical status, ECG and repeat TTE for assessment of response to CRT. Physicians evaluating patients at follow-up were blinded to the pre-CRT data to avoid potential bias. The main outcomes assessed after CRT implantation were: a) response to CRT by echocardiographic parameters, defined as ≥ 15% reduction in LV end-systolic volume, and b) clinical improvement assessed by NYHA functional class and MHFQ, at 6-month follow-up.^9,10 After determining CRT-responders, we examined the changes of RV echocardiographic parameters with CRT placement compared with non-responders.Echocardiographic ImagingTTE examination with machine-integrated ECG recording was performed, preferably with the patients lying in the left lateral decubitus position, using a Vivid 5, S5 or 7 machines (General Electric Vingmed Ultrasound, Horten, Norway) with an M4S matrix sector array probe (frequency of 2.5 MHz). Standard images were obtained in the parasternal (long- and short-axis) and apical (2- and 4-chamber) views. RV parameters evaluated included following: RV longitudinal and transverse diameters, tricuspid annular plane systolic excursion (TAPSE), tissue Doppler-derived tricuspid lateral annular systolic velocity (S’), Percentage RV fractional area of change (FAC), RV myocardial performance index (MPI), RV systolic pressure (RVSP), and tricuspid regurgitation (TR) severity. LV parameters evaluated included LV end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), LVEF and mitral regurgitation (MR) severity. We followed the current guidelines in echocardiographic assessment of right and left heart in adults.^7,11,12 All echocardiographic analysis was conducted by two independent and experienced cardiologists. The detailed description of the definitions and echocardiographic assessment methods are reported in the supplemental appendix (1). CRT-P pacemakers were successfully implanted in all patients under fluoroscopic guidance. Pacing leads were implanted via the left axillary/subclavian vein approach. The right atrial lead was positioned in the right atrial appendage and the RV lead tip was placed at the apex of the right ventricle. The LV lead tip was positioned in the posterolateral cardiac vein (n= 27), posterior vein (n= 2) and lateral vein (n= 1). All implanted devices were QUADRA ALLURE MP™ RF CRT-P (St. Jude Medical, Minnesota, USA).Statistical AnalysisCategorical variables were presented as number and percentages and compared by using the Chi-square test. Continuous variables were presented as the mean and standard deviation and compared using Independent t-test when the data were parametric and Mann-Whitney test when the data were nonparametric. Receiver operating characteristic curve (ROC) was used to assess the cutoff point with the highest sensitivity and specificity. Logistic regression analysis was used to identify the predictors of CRT response. Predictors associated with a P value <0.1 on univariate analysis were entered a stepwise multivariate logistic regression model to identify independent predictors of reverse remodeling. The p-value of <0.05 was considered statistically significant. Statistical analysis was performed using SPSS 16 for Windows (SPSS Inc., Chicago, Illinois, USA).

Results

Baseline clinical and Echocardiographic Characteristics The mean age of the study population was 52.8 ± 15.4 years, 80% men. Sixty percent of patients had non-ischemic etiology for HF. Patients with LBBB represented 76.6% of the study cohort with a mean QRS duration of 148.1 ± 10.1 msec. The majority of patients (80%) had NYHA class III symptoms. Mean indexed LVEDV was 223.9 ± 73.1 ml/m2 , mean indexed LVESV was 166.1 ± 60.7 ml/m2 and mean EF was 25.9 ± 5.9%. The mean basal transverse RV diameter was 43.8 ± 9.5 mm, mid RV diameter was 28.9 ± 6.4 mm, and longitudinal diameter was 82.7 ± 13.6 mm. RV systolic parameters showed a mean S’ of 11.3 ± 3 cm/s, FAC of 37.4 ± 9.1%, TAPSE of 14.8 ± 5.6 mm, and MPI of 0.46 ± 0.1. Forty percent of patients had at least moderate of tricuspid regurgitation. Details of baseline clinical and echocardiographic characteristics of the study population are shown in Table 1. Clinical and Echocardiographic Outcomes** At 6-months follow-up, CRT was associated with a significant improvement in MHFQ (70.8 ± 15.1 vs 43.8 ± 22.9; p < 0.0001), and mean NYHA class (3.2 ± 0.4 vs 2.1 ± 0.7; p < 0.0001) in the entire cohort. There was significant reduction in post-CRT QRS duration (148.1 ± 10.4 vs 128.3 ± 8.2 msec, p < 0.001). Echocardiographic LV parameters post CRT showed significant reduction in LVESV (138.3 ± 56.4 ml vs 166.1 ± 60.7 ml; p = 0.002), and significant increase in LVEF (25.9 ± 5.9% vs 33.8 ± 10.7%; p < 0.0001) compared with pre-CRT parameters. Successful response to CRT (defined by a reduction of ≥ 15% in LVESV) was achieved in 20 patients (66.7%). Regarding echocardiographic RV parameters, there was a significant reduction in RV basal (43.8 ± 9.5 vs 39.7 ± 11.8 mm; p < 0.0001), and longitudinal dimensions (82.7 ± 13.6 vs 75.6 ± 15.4 mm; p < 0.001) with CRT compared to baseline. CRT implantation was also associated with significant improvement in RV systolic function parameters including S’ (11.3 ± 3 vs 12.8 ± 4.2 cm/s; p < 0.0001), TAPSE (14.8 ± 5.6 vs 21.5 ± 6.9 mm; p = 0.002), and FAC (37.4 ± 9.1% vs 41.0 ± 12.8%; p = 0.01). No significant difference was observed in post-CRT RV systolic pressure (39.3 ± 10.5 vs 38.8 ± 9.7 mmHg; p= 0.16), MPI (0.46 ± 0.1 vs 0.44 ± 0.12, p= 0.15) and degree of TR (1.53 ± 0.7 vs 1.56 ± 0.7; p= 0.88) compared with pre-CRT parameters. Comparison of clinical and echocardiographic parameters before and after CRT for the study population is reported in Table 2. Predictors of CRT Response Patients were classified as CRT-responders (≥ 15% reduction in LVESV at 6-months follow-up) and non-responders. CRTresponders had a higher frequency of baseline LBBB (90% vs 50%, p = 0.03), and more prolonged QRS duration (151.9 ± 8.1 vs 141 ± 11 msec; p = 0.005) compared with non-responders. In addition, CRT-responders had lower MHFQ at baseline compared with non-responders (66.9 ± 14.7 vs 78.7 ± 13.1, p = 0.04), with no significant difference in NYHA functional class (3.2 ± 0.37 vs 3.3 ± 0.48, p = 0.37). There was no significant correlation between baseline LV echocardiographic parameters and response to CRT (Table 1). There was significant difference in degree of MR severity between the two groups (1.6 ± 0.8 vs 2.3 ± 0.7; p = 0.027). On analysis of echocardiographic RV parameters, CRTresponders demonstrated smaller basal and transverse diameters (40.7 ± 8.6 vs 50.1 ± 8.4 mm; p = 0.009; and 27 ± 6.1 vs 32.6 ± 5.9 mm, p = 0.025, respectively), as well as higher FAC (40.9 ± 6.4 vs 30.4 ± 10.1%; p = 0.002), TAPSE (22.1 ± 4.9 vs 15.2 ± 3.9 mm; p = 0.001), S’ (12.8 ± 2.3 vs 8.4 ± 1.8 mm; p < 0.0001), and MPI (0.41 ± 0.07 vs 0.54 ± 0.09; p < 0.0001) at baseline compared with non-responders. Table 1 illustrates the correlation between baseline parameters (clinical and echo cardiographic) and response to CRT. The degree of change in echocardiographic RV parameters in CRT-responders and nonresponders is presented in Table 3. As shown, CRT-responders had significant positive remodeling in the RV when compared to non-responders at 6-month follow-up.NICM= non-ischemic cardiomyopathy, ICM= ischemic cardiomyopathy, NYHA= New York heart association, HF= heart failure, MHFQ= Minnesota HF questionnaire, LBBB= left bundle branch block, LVEDD= left ventricular end-diastolic diameter, LVESD= Left ventricular end-systolic diameter, LVEDV= left ventricular end-diastolic volume, LVESV= Left ventricular end-systolic volume, EF= Ejection fraction, MR= Mitral regurgitation, RVD= Right ventricular diameter FAC= Fractional area change, TAPSE= Tricuspid annular systolic plane excursion, S’= Tissue Doppler–derived tricuspid lateral annular systolic velocity, MPI= Myocardial performance index, RVSP= Right ventricular systolic pressure, TR= Tricuspid regurgitation.NYHA= New York heart association, HF= heart failure, MHFQ= Minnesota HF questionnaire, LVEDD= left ventricular end-diastolic diameter, LVESD= Left ventricular end-systolic diameter, LVEDV= left ventricular end-diastolic volume, LVESV= Left ventricular end-systolic volume, EF= Ejection fraction, MR= Mitral regurgitation, RVD= Right ventricular diameter FAC= Fractional area change, TAPSE= Tricuspid annular systolic plane excursion, S’= Tissue Doppler–derived tricuspid lateral annular systolic velocity, MPI= Myocardial performance index, RVSP= Right ventricular systolic pressure, TR= Tricuspid regurgitation On a further ROC curve analysis, FAC >30% (100% sensitivity and 80% specificity; p = 0.003), TAPSE >18 mm (90% sensitivity and 80% specificity; p < 0.0001), S’ >8 cm/s (100% sensitivity and 70% specificity; p< 0.0001) and MPI < 0.52 (100% sensitivity and 70%; p < 0.0001) were identified as the baseline cut-off values for such parameters to predict appropriate CRT response (Figure 1). Multivariate logistic regression analysis including all the significant baseline clinical and echocardiographic parameters in a step-wise fashion revealed baseline RV S’ as the single independent predictor of successful response to CRT (odds ratio (OR): 3.21, 95% CI: 1.32 to 7.82, p= 0.01). Based on this finding, we chose S’ > 8 cm/s as a cutoff to stratify the population into 2 groups. At baseline, 23 out of 30 patients (76.6%) had an S’ value >8 cm/s. CRT response was observed in 20 patients with S’ value >8 cm/s versus none with S’ ≤ 8 cm/s.

Discussion

In the current single-center, prospective study of 30 patients, we demonstrated baseline echocardiographic RV function parameters to be significant predictors of CRT response in HF. Our study showed that at follow-up of 6 months, CRT resulted in significant improvement in the clinical symptoms, NYHA class, MHFQ score and echocardiographic measures of LV and RV function. Baseline criteria that were associated with improved CRT response included LBBB, increased QRS duration, lower MR severity, smaller RV basal and transverse diameters, and higher RV FAC, TAPSE, S’ and MPI. However, RV S’ was the single independent predictor of CRT response on multivariate analysis. An S’ cut off value of >8 cm/s was 100% sensitive, and 70% specific in predicting a satisfactory CRT response. CRT is the recommended line of treatment for patients with systolic HF on maximally tolerated guideline-based medical therapy and has been proven to reinstitute ventricular synchrony and improve LV function.^1,2 The effect of CRT on RV function is, however, still a topic of debate. Significant reverse remodeling, reduction of RV diameters, as well as improvement of RV function were demonstrated in prior studies.^10,13 The effect of CRT on RV systolic function was observed earlier to the changes in RV dimensions.¹⁴ Other studies failed to find a significant improvement in the RV function in conjunction with the observed improvement in LV function.¹⁵ We aimed in this study to perform a comprehensive study of RV parameters and were able to demonstrate a significant improvement in the parameters of RV function in patients who responded to CRT. It is estimated that approximately 30% of HF patients who receive a CRT device fail to show an appropriate response. Thus, novel methods to improve the CRT selection criteria and predict successful response are crucial, and RV parameters seem to be promising. Prior studies have demonstrated a significant correlation between echocardiographic indices of baseline RV
function and the degree of LV reverse remodeling observed at 6-months follow up after CRT implantation in patients with HF.¹⁶ Another study described pre-implant reduced RV systolic function to be a predictor of poor response to CRT.¹⁷ TAPSE > 17 mm was also correlated with improved CRT response, however, only with the sensitivity of 68% and specificity of 54% in one study, and sensitivity of 64% and specificity of 60% in another.¹⁸ In our study, we show for the first time that tissue Doppler-derived tricuspid annular systolic velocity is the most significant predictor of CRT response with a high degree of sensitivity and specificity. The assessment of RV dimensions and function is clinically challenging owing to the complex geometry of the RV and difficult visualization on standard echocardiography.^4,5 Cardiac magnetic resonance (CMR) imaging has emerged as an accurate modality for RV assessment, but not feasible for all patients.¹⁹ On contrary, RV S’ is a reliable, reproducible, and easy to measure, way of RV function assessment.⁷ Furthermore, it highly correlates with volumetric quantification of the RV systolic function on MRI.²⁰ compared with 2D and 3D echocardiographic estimates of RV size and systolic function which display only a moderate correlation with MRI measurements of such parameters.²¹ Despite these facts, the implementation of tissue Doppler-derived tricuspid lateral annular systolic velocity, known as RV S’, as a predictor of CRT response was clearly under-represented in clinical trials. To the best of our knowledge, our study is the first to-date to shed the light on RV S’ as a reliable predictor of CRT response. The S’ > 8 cm/s represents the highest sensitivity (100%), and specificity (70%) test, known in literature so far, to correlate with a successful CRT response. Such high sensitivity may offer physicians a possible reliable parameter to identify patients that will most probably not show adequate benefit to CRT device placement.Strengths and LimitationsStrengths of our study include a prospective study design, systematic assessment of all available 2D RV echocardiographic parameters and analysis of a composite outcome of imaging
and clinical parameters. However, we acknowledge several limitations in our study. First, a small number of patients were included in a single center. Since our study was only performed as
a proof-of-concept, we highly encourage multicenter randomized clinical trials to validate these findings on a larger group of patients. Second, the study was open labeled however to reduce the risk of bias, the follow-up evaluation for patients was blinded from their baseline characteristics. Finally, a more accurate parameter to assess the RV size and function would be better if it were based on MRI. However, in our study, we had in mind to utilize simple, cheap and readily available parameters derived from the echocardiography.

Conclusion

In summary, the current study identifies RV S’ (with cut-off value > 8cm/s) as a feasible and independent predictor of successful CRT response in HF patients.Clinical PerspectivesCompetency in Medical KnowledgeThe utility of echocardiographic RV parameters on response to CRT in patients with HF is not well known. In a prospective study design, we demonstrate that various baseline RV function parameters significantly predict a positive response to CRT, irrespective of LV echocardiographic parameters. Among all available measures of RV function, tissue Doppler-derived tricuspid annular systolic velocity had the best predictive value of CRT response on ROC analysis. In addition, all patients with a positive CRT response had a concomitant improvement in RV function by TTE.Translational OutlookThe findings of this prospective study are directly applicable to optimizing appropriate patient selection for CRT. Further larger studies validating these findings are encouraged.

Declarations

References

1. Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronization in chronic heart failure. N Engl J Med. 2002;346(24):1845-1853. doi:10.1056/nejmoa013168
   Article · PubMed · Google Scholar
2. Bristow MR, Saxon LA, Boehmer J, et al. Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med. 2004;350(21):2140-2150. doi:10.1056/nejmoa032423
   Article · PubMed · Google Scholar
3. Lunati M, Magenta G, Cattafi G, et al. Clinical relevance of systematic CRT device optimization. J Atr Fibrillation. 2014;7(2):1077. doi:10.4022/jafib.1077
   Article · PubMed · Google Scholar
4. De Groote P, Millaire A, Foucher-Hossein C, et al. Right ventricular ejection fraction is an independent predictor of survival in patients with moderate heart failure. J Am Coll Cardiol. 1998;32(4):948-954. doi:10.1016/s0735-1097(98)00337-4
   Article · PubMed · Google Scholar
5. Ghio S, Gavazzi A, Campana C, et al. Independent and additive prognostic value of right ventricular systolic function and pulmonary artery pressure in patients with chronic heart failure. J Am Coll Cardiol. 2001;37(1):183-188. doi:10.1016/s0735-1097(00)01102-5
   Article · PubMed · Google Scholar
6. Ghio S, Gavazzi A, Campana C, et al. Independent and additive prognostic value of right ventricular systolic function and pulmonary artery pressure in patients with chronic heart failure. J Am Coll Cardiol. 2001;37(1):183-188. doi:10.1016/s0735-1097(00)01102-5
   Article · PubMed · Google Scholar
7. Rudski LG, Lai WW, Afilalo J, et al. Guidelines for the echocardiographic assessment of the right heart in adults: a report from the American Society of Echocardiography. J Am Soc Echocardiogr. 2010;23(7):685-713. doi:10.1016/j.echo.2010.05.010
   Article · PubMed · Google Scholar
8. Brignole M, Auricchio A, Baron-Esquivias G, et al. 2013 ESC guidelines on cardiac pacing and cardiac resynchronization therapy. Eur Heart J. 2013;34(29):2281-2329. doi:10.1093/eurheartj/eht150
   Article · PubMed · Google Scholar
9. Chung ES, Leon AR, Tavazzi L, et al. Results of the Predictors of Response to CRT trial. Circulation. 2008;117(20):2608-2616. doi:10.1161/circulationaha.107.743120
   Article · PubMed · Google Scholar
10. Bleeker GB, Bax JJ, Fung JW, et al. Clinical versus echocardiographic parameters to assess response to cardiac resynchronization therapy. Am J Cardiol. 2006;97(2):260-263. doi:10.1016/j.amjcard.2005.08.030
    Article · PubMed · Google Scholar
11. Lang RM, Badano LP, Mor-Avi V, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr. 2015;28(1):1-39. doi:10.1016/j.echo.2014.10.003
    Article · PubMed · Google Scholar
12. Lancellotti P, Tribouilloy C, Hagendorff A, et al. Recommendations for the echocardiographic assessment of native valvular regurgitation: an executive summary from the European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc Imaging. 2013;14(7):611-644. doi:10.1093/ehjci/jet105
    Article · PubMed · Google Scholar
13. Rajagopalan N, Suffoletto MS, Tanabe M, et al. Right ventricular function following cardiac resynchronization therapy. Am J Cardiol. 2007;100(9):1434-1436. doi:10.1016/j.amjcard.2007.06.037
    Article · PubMed · Google Scholar
14. Donal E, Vignat N, De Place C, et al. Acute effects of biventricular pacing on right ventricular function assessed by tissue Doppler imaging. Europace. 2007;9(2):108-112. doi:10.1093/europace/eul175
    Article · PubMed · Google Scholar
15. Boriani G, Fallani F, Martignani C, et al. Cardiac resynchronization therapy: effects on left and right ventricular ejection fraction during exercise. Pacing Clin Electrophysiol. 2005;28(suppl 1):S11-S14. doi:10.1111/j.1540-8159.2005.00005.x
    Article · PubMed · Google Scholar
16. Scuteri L, Rordorf R, Marsan NA, et al. Relevance of echocardiographic evaluation of right ventricular function in patients undergoing cardiac resynchronization therapy. Pacing Clin Electrophysiol. 2009;32(8):1040-1049. doi:10.1111/j.1540-8159.2009.02436.x
    Article · PubMed · Google Scholar
17. Edner M, Ring M, Henriksson P. Pre-implant right ventricular function might be an important predictor of the response to cardiac resynchronization therapy. Cardiovasc Ultrasound. 2011;9(1):28. doi:10.1186/1476-7120-9-28
    Article · PubMed · Google Scholar
18. Cappelli F, Cristina Porciani M, Ricceri I, et al. Tricuspid annular plane systolic excursion evaluation improves selection of cardiac resynchronization therapy patients. Clin Cardiol. 2010;33(9):578-582. doi:10.1002/clc.20806
    Article · PubMed · Google Scholar
19. Beygui F, Furber A, Delépine S, et al. Routine breath-hold gradient echo MRI-derived right ventricular mass, volumes and function: accuracy, reproducibility and coherence study. Int J Cardiovasc Imaging. 2004;20(6):509-516. doi:10.1007/s10554-004-1097-7
    Article · PubMed · Google Scholar
20. Li Y, Wang Y, Zhai Z, et al. Relationship between echocardiographic and cardiac magnetic resonance imaging-derived measures of right ventricular function in patients with chronic thromboembolic pulmonary hypertension. Thromb Res. 2015;135(4):602-606. doi:10.1016/j.thromres.2015.01.008
    Article · PubMed · Google Scholar
21. Kjaergaard J, Petersen CL, Kjaer A, Schaadt BK, Oh JK, Hassager C. Evaluation of right ventricular volume and function by 2D and 3D echocardiography compared to MRI. Eur J Echocardiogr. 2006;7(6):430-438. doi:10.1016/j.euje.2005.10.009
    Article · PubMed · Google Scholar

Tables

Additional Information

Publisher’s Note
Bayrakol MP remains neutral with regard to jurisdictional and institutional claims.

Rights and Permissions