Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery

Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery

+Model ARTICLE IN PRESS ACVD-958; No. of Pages 10 Archives of Cardiovascular Disease (2016) xxx, xxx—xxx Available online at ScienceDirect www.sc...

1MB Sizes 5 Downloads 47 Views

+Model

ARTICLE IN PRESS

ACVD-958; No. of Pages 10

Archives of Cardiovascular Disease (2016) xxx, xxx—xxx

Available online at

ScienceDirect www.sciencedirect.com

CLINICAL RESEARCH

Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery Imagerie de déformation myocardique pour évaluer la fonction longitudinale ventriculaire droite après chirurgie cardiaque pédiatrique Clement Karsenty a,b,∗, Khaled Hadeed a, Yves Dulac a, Florent Semet a, Xavier Alacoque c, Sophie Breinig b, Bertrand Leobon c, Philippe Acar a, Sebastien Hascoet a,d,e a

Paediatric Cardiology Unit, Children’s Hospital, CHU Toulouse, 31300 Toulouse, France Intensive Care Unit, Children’s Hospital, CHU Toulouse, 31300 Toulouse, France c Department of Congenital Heart Disease Surgery, Children’s Hospital, CHU Toulouse, 31300 Toulouse, France d INSERM UMR1048, Institut des Maladies Métaboliques et Cardiovasculaires, I2MC, Équipe 8, 31300 Toulouse, France e Pôle de Chirurgie des Cardiopathies Congénitales, M3C, Hôpital Marie-Lannelongue, 31300 Le Plessis-Robinson, France b

Received 11 April 2016; received in revised form 31 May 2016; accepted 6 September 2016

KEYWORDS Congenital heart disease;

Summary Background. — Right ventricular (RV) function is a prognostic marker of cardiac disease in children. Speckle tracking has been developed to assess RV longitudinal shortening, the dominant

Abbreviations: 2D, two-dimensional; CHD, congenital heart disease; CI, confidence interval; CPB, cardiopulmonary bypass; IQR, interquartile range; LV, left ventricular; LVEF, left ventricular ejection fraction; LV-PSS, left ventricular peak systolic strain; RV, right ventricular; RV-PSS, right ventricular peak systolic strain; TAPSE, tricuspid annular plane systolic excursion; TA Sa, tricuspid annular peak systolic velocity; TDI, tissue Doppler imaging. ∗ Corresponding author. Service de Cardiologie Pédiatrique, Hôpital des Enfants, 330, avenue de Grande-Bretagne, 31059 Toulouse Cedex 9, France. E-mail address: [email protected] (C. Karsenty). http://dx.doi.org/10.1016/j.acvd.2016.09.003 1875-2136/© 2016 Elsevier Masson SAS. All rights reserved.

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model ACVD-958; No. of Pages 10

ARTICLE IN PRESS

2

C. Karsenty et al. Speckle tracking; Right ventricular function

deformation during systole; little is known about its feasibility in children with congenital heart disease (CHD). Aims. — To evaluate the feasibility and reproducibility of RV two-dimensional (2D) strain assessed by speckle tracking in infants undergoing CHD surgery compared with conventional markers. Methods. — In this prospective single-centre study, RV peak systolic strain (RV-PSS) was measured using 2D speckle tracking in 37 consecutive children undergoing CHD surgery. Examinations were performed the day before surgery, a few hours after surgery and before discharge. Relationships with the z score of tricuspid annular plane systolic excursion (TAPSE) and tricuspid annular systolic velocity (TA Sa) were assessed. Results. — Median (interquartile range) age was 19 months (5—63); median weight was 9.2 kg (5.3—18.0). RV-PSS analysis was feasible in 92.9% (95% confidence interval [CI]: 86.0—97.1) of examinations. The coefficient of variation was 9.7% (95% CI: 7.4—11.9) for intraobserver variability and 15.1% (95% CI: 12.7—17.6) for interobserver variability. Correlations between RV-PSS and z score of TAPSE and TA Sa were strong (r = 0.71, P < 0.0001 and r = 0.70, P < 0.0001, respectively). RV-PSS was significantly reduced after surgery compared with baseline (—10.5 ± 2.9% vs. —19.5 ± 4.8%; P < 0.0001) and at discharge (—13.5 ± 4.0% vs. —19.5 ± 4.8%; P < 0.0001). Similar evolutions were observed with TAPSE and TA Sa (both P < 0.0001). Conclusion. — RV longitudinal strain by speckle tracking is a feasible and reproducible method of assessing perioperative evolution of RV function in children with CHD. © 2016 Elsevier Masson SAS. All rights reserved.

MOTS CLÉS

Résumé Contexte. — La fonction ventriculaire droite (VD) est un déterminant pronostique des cardiopathies de l’enfant. L’analyse par speckle tracking permet d’évaluer le raccourcissement longitudinal des fibres VD qui correspond à la composante systolique dominante. La faisabilité de ce nouveau paramètre a été peu étudiée en cardiopédiatrie. Objectifs. — Nous avons évalué la faisabilité et la reproductibilité du strain VD par speckle tracking chez des enfants opérés d’une cardiopathie congénitale. Méthodes. — Le pic systolique de strain (PSS) de la paroi libre du VD a été mesuré prospectivement par speckle tracking chez 37 enfants inclus consécutivement. Trois échocardiographies ont été réalisées : le jour précédant la chirurgie, quelques heures après la chirurgie et en fin d’hospitalisation. Le PSS a été comparée à des marqueurs de fonction systolique conventionnels. Résultats. — L’âge médian (ITQ) était de 19 mois (5—63) avec un poids médian de 9,2 kg (5,3—18,0). L’analyse du PSS-VD était possible pour 92,9 % (IC 95 % : 86,0—97,1) des acquisitions. Le coefficient de variation intra-observateur était de 9,7 % (IC 95 % : 7,4—11,9) et inter-observateur de 15,1 % (IC 95 % : 12,7—17,6). Les corrélations entre le PSS-VD, le z score du TAPSE et l’Onde S tricuspide étaient bonnes (r = 0,71, p < 0,0001 ; r = 0,70, p < 0,0001). En postopératoire, le PSS-VD était diminué par rapport à l’échocardiographie préopératoire (—10,5 ± 2,9 % vs —19,5 ± 4,8 % ; p < 0,0001) ainsi qu’en fin d’hospitalisation (—13,5 ± 4,0 % vs —19,5 ± 4,8 % ; p < 0,0001). Conclusion. — La mesure du strain longitudinal VD par speckle tracking est faisable et reproductible pour évaluer l’évolution de la fonction systolique VD après chirurgie cardiaque pédiatrique. © 2016 Elsevier Masson SAS. Tous droits r´ eserv´ es.

Cardiopathie congénitale ; Strain ; Fonction ventriculaire droite

Background Right ventricular (RV) function is a prognostic marker of cardiac disease. RV function is a strong predictor of outcome after myocardial infarction [1], in patients with heart failure [2,3] and after cardiac surgery in adults [4,5].

RV function is currently assessed by echocardiography, using tissue Doppler imaging (TDI) and tricuspid annular plane systolic excursion (TAPSE). Using these markers, significant impairment of RV contractility in the early postoperative period of cardiac surgery has been observed [6], lasting over months [7], including after surgery for adult

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model ACVD-958; No. of Pages 10

ARTICLE IN PRESS

Right ventricular function by 2D strain after cardiac paediatric surgery congenital heart disease (CHD) [8]. In children, TDI detects impairment of RV performance shortly after surgery for CHD [9,10]. Myocardial deformation measured by TDI or speckle tracking (two-dimensional [2D] strain) techniques seems to provide incremental information [11]. Strain analysis obtained from tissue velocities appears to be a sensitive tool for monitoring the infant heart after cardiopulmonary bypass (CPB) [12]; however, strain was measured by TDI rather than by speckle tracking. The latter seems more reliable, as it assesses the frame-to-frame movement of speckles. Compared with TDI, speckle tracking is semiautomated, angle-independent [13], more reproducible, and thus appears more robust [14]. Little is known about the evolution of RV systolic 2D strain during the perioperative care of children with CHD. The aims of this study were to evaluate the feasibility and reproducibility of RV 2D strain by speckle tracking compared with conventional echocardiographic markers, and to assess its evolution in infants and neonates undergoing surgery for CHD.

Methods We prospectively included 37 patients with CHD aged 0—18 years undergoing CHD surgery in the Children’s Hospital, University of Toulouse, between June 2013 and October 2013. Exclusion criteria were age > 18 years and univentricular CHD. Informed verbal consent was obtained from each patient or their legal representatives after they had received a full explanation of the procedure given. A written consent form was not required, according to French law, given that the echocardiographic evaluation was part of the regular management of the children, and was required by their medical condition. The database was declared to the National Commission for Data Processing and Freedoms (No. 1673449). No additional examination was performed for the sole purpose of the study.

Echocardiography and data analysis Echocardiography (IE33 Ultrasound System; Philips Medical Systems, Andover, MA, USA) was performed in all patients, preoperatively the day before the surgery (PREOP), on the day of surgery in the first 3 hours after admission to the paediatric cardiology intensive care unit (EARLY) and at hospital discharge (DISCHARGE). Examinations were performed with the child in the decubitus position. S8-1 or S5-1 matrix probes were used, depending on the age of the patient. The frame rate was kept between 70 and 100 Hz, and an electrocardiogram was connected to detect systole and diastole. Examinations started with conventional bidimensional echocardiography.

Echocardiographic assessment of right and left ventricular global function Data on variables of RV function were collected. TAPSE was acquired by the conventional M-mode method from the lateral point of the tricuspid valve in a standard apical fourchamber view [15], and was expressed according to the z

3

score because of the important effect of age [16]. Tricuspid annular peak systolic velocity (TA Sa) was obtained from a modified apical four-chamber view in TDI [17]. Left ventricular ejection fraction (LVEF) was measured from the apical four-chamber view by the monoplane Simpson method, and fractional shortening was obtained by the M-mode method from the parasternal long-axis view [18]. We focused on substantial factors that could influence the evolution of RV variables. Volume overload is known to influence RV variables, so we assessed RV function in patients with preoperative volume overload (those with atrial septal defect or pulmonary and associated tricuspid regurgitation) and patients without preoperative volume overload. Also, the type of surgery can be linked to RV dysfunction, so we compared patients who required surgery with and without CBP.

2D strain At the end of the examination, three loops were recorded in the apical four-chamber view, focused on the left ventricle and then the right ventricle. A modified apical view was used to obtain a complete image of the right ventricle [15]. Special attention was given to optimize gain and compression, and to obtain a central and complete picture of the left ventricle and then of the free wall of the right ventricle. Examinations were recorded on Digital Imaging and Communications in Medicine (DICOM) compact discs for off-line analysis using QLab 10 software (Philips Medical Systems, Andover, MA, USA). To measure longitudinal global left ventricular (LV) and RV peak systolic strain, the endocardial border was traced manually by the operator. Basal points were placed just above the atrioventricular valve. The software automatically generated a second line at the level of the epicardium, delineating a region of interest, including the entire wall between endocardial and epicardial border. The accuracy of the tracking of the region of interest during the cardiac cycle was checked visually on a dynamic loop. In case of discrepancy, manual corrections were applied. The measurement of strain was performed automatically. The analysis was considered as acceptable if the software validated the measurement for each cardiac segment. If a segment was not accepted, the limit of the region of interest was retraced. The strain values were averaged to determine global longitudinal strain. RV peak systolic strain (RV-PSS) was calculated by averaging only strain curves from the RV free wall (threesegment model: basal, median and apical), excluding the septum to avoid interaction with LV function [19].

Reproducibility Interobserver and intraobserver variabilities of longitudinal LV and RV strain variables were tested in 30 examinations from 10 randomly selected subjects for each echocardiographic assessment (PREOP, EARLY and DISCHARGE). Loop acquisitions were performed by the same operator in all patients. To assess interobserver variability, the dataset was analysed off-line by a second operator with the same level of experience, blinded to the previous results. To evaluate intraobserver variability, the first operator repeated

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model ACVD-958; No. of Pages 10

ARTICLE IN PRESS

4

C. Karsenty et al.

the measurements at the end of the data analysis process, blinded to the previous results.

Table 1

General data.

Variable

Statistical analysis Data collected were demographic variables, including body weight, length, sex, body surface area, age, heart rate and type of CHD. Surgery data included CPB time, ultrafiltration rate and aortic cross-clamp time. Postoperative data included mechanical ventilation duration, lactate peak, troponin peak and inotrope duration. Echocardiographic variables included z score of TAPSE, RV-PSS, TA Sa, LVEF, LV fractional shortening and LV peak systolic strain (LV-PSS). Data are expressed as median and interquartile range (IQR) or number (percentage). Because the number of subjects in the study was small, and many of the variables did not fit normal distributions, non-parametric tests were used to compare data. In particular, the Mann-Whitney U test was used to compare RV-PSS among subgroups, according to CHD type. Spearman’s coefficients were used to assess correlations between systolic function variables. Relationships between RV and LV 2D strain and outcome or myocardial injury were further assessed by linear regression analysis. The Wilcoxon matched-pairs signed-rank test was used to assess the evolution of systolic function variables after surgery compared with preoperative data. Intraobserver and interobserver variabilities for the different strain variables were assessed using the Bland-Altman approach, with calculation of the mean bias (average difference between measurements) and the lower and upper limits of agreement (95% limits of agreement of mean bias) [20]. Statistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software,Inc., La Jolla, CA, USA). P values < 0.05 were considered to indicate statistical significance.

Age (month) Female sex Weight (kg) BSA (m2 ) Heart rate Type of congenital heart disease ASD VSD ToF Coarctation TGA Subaortic membrane PA-VSD AVSD Pulmonary tricuspid regurgitation Double aortic arch CBP duration (minutes) Clamping duration (minutes) Mechanical ventilation duration (hours) Ultrafiltration (mL/kg/ min) Lactate peak (mmol/L) Troponin peak (␮g/L) Inotrope duration (day) Days in ICU Discharge echography delay (day)

19 (5—63) 18 (48.6) 9.2 (5.3—18.0) 0.41 (0.30—0.72) 117.0 (91.0—140.5) 10 (27.0) 8 (21.6) 6 (16.2) 5 (13.5) 2 (5.4) 2 (5.4) 1 (2.7) 1 (2.7) 1 (2.7) 1 (2.7) 70 (43—93) 52 (22—72) 10 (4—72) 0.62 (0.23—0.95) 2.9 (2.2—5.0) 27.9 (10.5—55.9) 2.6 ± 3.5 3 (2—7.3) 6 (4—8)

Data expressed as median (interquartile range), number (%) or mean ± standard deviation. ASD: atrial septal defect; AVSD: atrioventricular septal defect; BSA: body surface area; CPB: cardiopulmonary bypass; ICU: intensive care unit; PA-VSD: pulmonary stenosis and VSD; TGA: transposition of the great arteries; ToF: tetralogy of Fallot; VSD: ventricular septal defect.

Results Study population Of the 37 CHD operations, six (16.2%) were performed in the neonatal period. Demographics, surgery data and distribution of CHD are reported in Table 1. Median age was 19 months (IQR 5—63 months). Median weight was 9.2 kg (IQR 5.3—18.0 kg). There were no surgical deaths, no deep hypothermic cardiac arrest and no pacing.

Feasibility of measurements Accepted measurements of RV-PSS and LV-PSS were obtained in 86.7% (95% confidence interval [CI]: 69.3—96.2) of patients on the day of the surgery, and in 92.9% (95% CI: 86.0—97.1) and 91.9% (95% CI: 84.7—96.4) of patients, respectively, for overall measurements. The feasibility of every measurement is detailed in Table 2. Postoperative 2D strain was not feasible in three patients, all aged > 5 years.

Reproducibility of measurements Coefficients of variation for intraobserver variability were 9.7% (95% CI: 7.4—11.9) for RV-PSS and 5.0% (95% CI:

3.7—6.3) for LV-PSS. Corresponding data for interobserver variability are 15.1% (95% CI: 12.7—17.6) and 10.5% (95% CI: 7.7—13.3), respectively. Bland-Altman analysis confirmed the lack of significant bias, with relatively narrow 95% limits of agreement for RVPSS and LV-PSS (Fig. 1). The mean bias for intraobserver RV-PSS was 0.3% (Fig. 1). Bland-Altman analysis indicated that the bias tended to increase slightly with increasing RV-PSS values. The magnitude of intraobserver differences observed was moderate, as reflected in 95% of values ranging between —3.2% and 3.7%. The mean bias for interobserver RV-PSS was 0.8% (Fig. 1). Bland-Altman analysis indicated that the bias tended to increase slightly with increasing RV-PSS values. The magnitude of interobserver differences observed was more pronounced, as reflected in 95% of values ranging between —4.4% and 5.7%. The mean bias for intraobserver LV-PSS was 0.0% (Fig. 1). Bland-Altman analysis indicated that the bias tended to be relatively identical, whatever the LV-PSS value. The magnitude of intraobserver differences observed was low, as reflected in 95% of values ranging between —2.2% and 2.2%.

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model

ARTICLE IN PRESS

ACVD-958; No. of Pages 10

Right ventricular function by 2D strain after cardiac paediatric surgery Table 2

5

Feasibility of systolic right and left ventricular function variables. RV function variables

PREOPa EARLYb DISCHARGEc OVERALLd (95% CI)

LV function variables

RV-PSS

TAPSE

TA Sa

LV-PSS

LVEF

FS

97.3 86.7 93.8 92.9 (86.0—97.1)

100 100 100 100.0 (96.3—100.0)

100 100 100 100.0 (96.3—100.0)

94.6 86.7 93.8 91.9 (84.7—96.4)

94.6 86.7 96.9 92.9 (86.0—97.1)

100 100 100 100.0 (95.8—100.0)

Data are expressed as percentage. CI: confidence interval; FS: fractional shortening; LV: left ventricular; LVEF: left ventricular ejection fraction; LV-PSS: left ventricular peak systolic strain; RV: right ventricular; RV-PSS: right ventricular peak systolic strain; TA Sa: tricuspid annular peak systolic velocity. a Baseline values. b Values on operation day in the intensive care unit. c Values before discharge at the end of the hospitalization. d The mean of the three examination values.

The mean bias for interobserver LV-PSS was 1.1% (Fig. 1). Bland-Altman analysis indicated that the bias tended to be relatively identical, whatever the LV-PSS value. The magnitude of interobserver differences observed was more pronounced, as reflected in 95% of values ranging between —2.8% and 4.7%.

Comparison between RV-PSS and conventional RV systolic function variables Linear regression analysis between RV-PSS, z score of TAPSE and TA Sa is presented in Fig. 2. The relationships between RV-PSS and z score of TAPSE, and between RV-PSS and TA

Sa were good (r = 0.71, P < 0.0001 and r = 0.70, P < 0.0001, respectively). Median RV-PSS was —23.0% (95% CI: —20.2 to —25.6) in atrial septal defect patients versus —17.5% (95% CI: —16.5 to —20.2) in other patients (P = 0.02). RV-PSS was higher in children with atrial septal defect than in children with tetralogy of Fallot: —23.0% (95% CI: —20.2 to —25.6) vs. —16.4% (95% CI: —12.9 to —19.9) (P = 0.007).

Evolution of RV and LV systolic variables after surgery The z score of TAPSE and TA Sa values were significantly lower in postoperative care compared with baseline: z

Figure 1. Intra- and interobserver reproducibility of right ventricular peak systolic strain (RV-PSS) and left ventricular peak systolic strain (LV-PSS) measurements assessed by Bland-Altman analysis. SD: standard deviation.

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model ACVD-958; No. of Pages 10

ARTICLE IN PRESS

6

C. Karsenty et al. they remained significantly lower compared with baseline (Fig. 3). In patients with (n = 11) and without (n = 26) preoperative volume overload we observed a significant decrease in all RV variables (Fig. 4). Nevertheless, without volume overload we observed a normal range at baseline (z score of TAPSE between —2 and +2 standard deviations: 1.45, 95% CI: 0.8—2.7) and still a significant decrease in all RV variables (Fig. 4). Surgery without CPB (n = 7) did not lead to RV failure compared with surgery with CBP (n = 30) (Fig. 5). LVEF and LV-PSS were moderately but significantly lower in postoperative care compared with baseline: 55.4% (95% CI: 53.2—57.6) vs 62.6% (95% CI: 60.5—64.5) (P < 0.0001) and —17.0% (95% CI: —15.4 to —17.7) vs. —21.0% (95% CI: —19.8 to —21.8) (P < 0.0001), respectively. LVEF values at the end of hospitalization were not significantly different compared with baseline (59.8%, 95% CI: 57.9—62.1 vs. 62.6%, 95% CI: 60.5—64.5; P = 0.22), unlike LV-PSS, which remained slightly lower at discharge (—18.0%, 95% CI: —16.1 to —18.5 vs. —21.0%, 95% CI: —19.8 to —21.8; P < 0.0006) (Fig. 3). Fractional shortening was not different between the three echocardiograms, with large dispersion, especially for EARLY time (33.5 ± 10.3%). A lower decrease in LV-PSS than RV-PSS at an early stage compared with baseline was observed (20.1 ± 3.4% vs 40.6 ± 4.7%; P < 0.001). There were no significant relationships between RV or LV longitudinal 2D strain and myocardial injury, such as CPB duration, clamping duration, lactate peak and troponin peak. No link between RV or LV longitudinal 2D strain and outcome (length of inotropes and mechanical duration or intensive care unit and hospital length of stay) was found.

Discussion

Figure 2. Relationship between right ventricular peak systolic strain (RV-PSS), z score of tricuspid annular plane systolic excursion (TAPSE) and tricuspid annular peak systolic velocity (TA Sa) by linear regression analysis.

score of TAPSE—5.7 (95% CI: —4.6 to —6.8) vs 2.5 (95% CI: 1.7—3.3) (P < 0.0001); TA Sa 6.4 cm/s (95% CI: 5.8—7.9) vs 12.5 cm/s (95% CI: 11.6—13.9) (P < 0.0001). Likewise, RV-PSS was decreased after surgery compared with preoperative values: —10.7% (95% CI: —9.3 to —11.7) vs. —19.7% (95% CI: —17.9 to —21.1) (P < 0.0001). At the end of hospitalization, there was a trend towards recovery of all RV variables, but

To our knowledge, this is the first study to describe the use of 2D strain measured by speckle tracking for the assessment of RV function after paediatric cardiac surgery. Speckle tracking was used previously in children with a high feasibility for the left ventricle (up to 90%) [21—23]. In our study, we have shown that the feasibility of 2D strain measurements was good, even in the postoperative period. However, despite a standardized validated method [19], RV-PSS feasibility remains lower than that of conventional variables (z score of TAPSE and TA Sa). Echocardiography picture quality and region of interest tracking may be more affected by postoperative factors (dorsal decubitus with mechanical ventilation, sternotomy wound with thoracic dressing and chest tubes). TAPSE and TA Sa measured at the tricuspid annulus seem to be easier to measure [24] compared with RV-PSS, as precise delineation of the myocardium is required for 2D strain analysis by speckle tracking. We observed that RV-PSS reproducibility was good, with low bias, narrow limits of agreement and an acceptable coefficient of variation. Similar results have been observed in healthy adults [25] and premature infants [26]. Furthermore, the level of experience of the operator does not

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model ACVD-958; No. of Pages 10

ARTICLE IN PRESS

Right ventricular function by 2D strain after cardiac paediatric surgery

7

Figure 3. Box plot displaying right ventricular (RV) and left ventricular (LV) function variables (median, 25th—75th percentile) at baseline (PREOP), on operation day in the intensive care unit (EARLY) and at discharge echocardiogram (DISCHARGE). FS: fractional shortening; LVEF: left ventricular ejection fraction; LV-PSS: left ventricular peak systolic strain; ns: not significant; RV-PSS: right ventricular peak systolic strain; TAPSE: tricuspid annular peak systolic excursion; TA Sa: tricuspid annular peak systolic velocity. * P < 0.05; ** P < 0.001; *** P < 0.0001.

seem to increase the variability of 2D strain measurements [27]. Likewise, correlations between RV-PSS and z score of TAPSE and TA Sa were good. TAPSE and TA Sa reflect RV function through tricuspid annular motion. RV-PSS is rather a global marker of RV free wall contractility. Furthermore, this added value is enhanced by the intrinsic capacity of 2D strain, including its angle independency. We focused on the analysis of the RV free wall in 2D strain, because this part of the RV contributes to 80% of the RV stroke volume, and inclusion of the septal wall in the RV global strain calculation would overlap with LV longitudinal global strain [28]. We observed a significant decrease in RV function after cardiac surgery, without complete recovery at hospital discharge. This partial recovery of ventricular performance has been noticed in several studies, with several different subgroups of CHD patients, such as great vessels transposition and tetralogy of Fallot surgery [29,30]. Even in long-term follow-up of children with ventricular septal defect, Klitsie et al. found that TAPSE and TA Sa remained impaired for up to 20 months after surgery [31]. As expected, CPB appears as a major determinant of RV dysfunction, as surgery without CPB does not lead to a

significant difference in pre- and postoperative variables. The systemic inflammatory response is a key factor in this adverse effect [32]. Interestingly, RV postoperative dysfunction is more pronounced than LV postoperative dysfunction. This highlights a different response and adaptation of the right ventricle compared with the left ventricle. RV cardiomyocytes seem to be more sensitive to CPB. Little is known about the physiopathology of RV failure and remodelling compared with LV RV failure and remodelling [33]. Until recently, PSS was thought to be less load dependent than other conventional systolic variables, and to reflect only intrinsic myocardial contractility. We observed higher RV-PSS values in children with atrial septal defect, resulting in right atrial and RV volume overload. The decrease in RV variables is stronger in patients with RV volume overload. This finding supports recent studies about load dependency of longitudinal systolic 2D strain [34—36]. Nevertheless, patients without RV volume overload had normal values at baseline, with a significant postoperative decrease. So, the RV failure observed in our study is not just caused by load relief. In experimental and clinical studies, strain rate appears to be a more robust measure of contractility, which is less influenced by changes in cardiac load and structure

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model ACVD-958; No. of Pages 10

ARTICLE IN PRESS

8

C. Karsenty et al.

Figure 4. Box plot displaying right ventricular function variables (median, 25th—75th percentile) for patients with preoperative right volume overload (n = 11) versus patients without preoperative volume overload (n = 26) at baseline (PREOP), on operation day in the intensive care unit (EARLY) and at discharge echocardiogram (DISCHARGE). ns: not significant; RV-PSS: right ventricular peak systolic strain; TAPSE: tricuspid annular peak systolic excursion; TA Sa: tricuspid annular peak systolic velocity. ** P < 0.001; *** P < 0.0001.

[37,38]. However, it is still a research tool at present, and its feasibility in clinical practice remains to be demonstrated. Thus, because of the complex RV geometry and high susceptibility of RV function to loading conditions, accurate assessment of RV function remains challenging, and requires an integrated approach using multiple echocardiographic variables, including 2D strain [39]. Study limitations: we did not measure RV fractional area change, which is a validated marker of RV function in adults [4]; it also requires accurate delineation of RV borders and a non-automated analysis. This variable assesses radial and longitudinal global RV function, and is less sensitive at detecting early impairment of RV function, which seems to first affect the longitudinal component [28]. In children, a weak correlation between magnetic resonance imagingderived RV systolic performance variables and RV fractional area change has been observed, thus discouraging its use [40]. We were unable to compare echocardiographic 2D strain measurements with cardiac magnetic resonance imaging, which is considered to be the gold standard for myocardial strain values [13]. Indeed magnetic resonance imaging allows accurate follow-up of ventricular performance, even in CHD patients with complex anatomy [41]. However, in

the postoperative care of paediatric CHD surgery, magnetic resonance imaging is not easily applicable routinely; it requires transfer to a radiology unit, which increases the risk of adverse events in postoperative care. Transthoracic echocardiography is an accessible and robust tool for assessing both RV and LV function before and after cardiac surgery. The wide variation in age and cardiac defects of the children included in our study was associated with a wide variation in the type and grade of complexity of surgical procedures. Previous findings have demonstrated that the type of cardiac surgery affects postoperative echocardiographic variables [42]. Given that the number of patients with specific lesions was small, it was not possible to draw conclusions about specific subsets of patients. However, one goal of the study was to assess 2D strain feasibility and reliability in a heterogeneous population of patients, reflecting usual care in a tertiary centre. During the study period in our department, RV and LV-PSS measured by speckle tracking required off-line analysis on dedicated software. Nowadays, 2D strain by speckle tracking can be measured directly on the echocardiography machine in a semi-automated fashion, reducing analysis duration [43].

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model ACVD-958; No. of Pages 10

ARTICLE IN PRESS

Right ventricular function by 2D strain after cardiac paediatric surgery

9

Figure 5. Box plot displaying right ventricular function variables (median, 25th—75th percentile) for patients requiring surgery without cardiopulmonary bypass (CPB) (n = 7) versus surgery with CPB (n = 30) at baseline (PREOP), on operation day in the intensive care unit (EARLY) and at discharge echocardiogram (DISCHARGE). ns: not significant; RV-PSS: right ventricular peak systolic strain; TAPSE: tricuspid annular peak systolic excursion; TA Sa: tricuspid annular peak systolic velocity. ** P < 0.001; *** P < 0.0001.

Conclusions RV 2D strain measured by speckle tracking is a feasible and reproducible method for assessing RV function in the postoperative care of paediatric CHD surgery. Correlations between RV-PSS and conventional systolic variables were strong. After surgery with CPB, RV dysfunction was pronounced, regardless of the preoperative level of load. Future studies in CHD patients are needed to assess its prognostic value.

Disclosure of interest

[4]

[5]

[6]

The authors declare that they have no competing interest.

References [1] Zornoff LA, Skali H, Pfeffer MA, et al. Right ventricular dysfunction and risk of heart failure and mortality after myocardial infarction. J Am Coll Cardiol 2002;39:1450—5. [2] 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;32:948—54. [3] Gavazzi A, Berzuini C, Campana C, et al. Value of right ventricular ejection fraction in predicting short-term prognosis of

[7]

[8]

[9]

patients with severe chronic heart failure. J Heart Lung Transplant 1997;16:774—85. Haddad F, Denault AY, Couture P, et al. Right ventricular myocardial performance index predicts perioperative mortality or circulatory failure in high-risk valvular surgery. J Am Soc Echocardiogr 2007;20:1065—72. Maslow AD, Regan MM, Panzica P, Heindel S, Mashikian J, Comunale ME. Precardiopulmonary bypass right ventricular function is associated with poor outcome after coronary artery bypass grafting in patients with severe left ventricular systolic dysfunction. Anesth Analg 2002;95:1507—18 [table of contents]. Brookes CI, White PA, Bishop AJ, Oldershaw PJ, Redington AN, Moat NE. Validation of a new intraoperative technique to evaluate load-independent indices of right ventricular performance in patients undergoing cardiac operations. J Thorac Cardiovasc Surg 1998;116:468—76. Wranne B, Pinto FJ, Hammarstrom E, St Goar FG, Puryear J, Popp RL. Abnormal right heart filling after cardiac surgery: time course and mechanisms. Br Heart J 1991;66:435—42. Schuuring MJ, Bolmers PP, Mulder BJ, et al. Right ventricular function declines after cardiac surgery in adult patients with congenital heart disease. Int J Cardiovasc Imaging 2012;28:755—62. Vassalos A, Lilley S, Young D, et al. Tissue Doppler imaging following paediatric cardiac surgery: early patterns of change and relationship to outcome. Interact Cardiovasc Thorac Surg 2009;9:173—7.

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003

+Model ACVD-958; No. of Pages 10

ARTICLE IN PRESS

10 [10] Klitsie LM, Hazekamp MG, Roest AA, et al. Tissue Doppler imaging detects impaired biventricular performance shortly after congenital heart defect surgery. Pediatr Cardiol 2013;34:630—8. [11] Abraham TP, Dimaano VL, Liang HY. Role of tissue Doppler and strain echocardiography in current clinical practice. Circulation 2007;116:2597—609. [12] Pauliks LB, Valdes-Cruz LM, Perryman R, Scholl FG. Right ventricular wall-motion changes after infant open heart surgery — a tissue Doppler study. Echocardiography 2014;31:209—17. [13] Amundsen BH, Helle-Valle T, Edvardsen T, et al. Noninvasive myocardial strain measurement by speckle tracking echocardiography: validation against sonomicrometry and tagged magnetic resonance imaging. J Am Coll Cardiol 2006;47:789—93. [14] Dandel M, Lehmkuhl H, Knosalla C, Suramelashvili N, Hetzer R. Strain and strain rate imaging by echocardiography — basic concepts and clinical applicability. Curr Cardiol Rev 2009;5:133—48. [15] Hammarstrom E, Wranne B, Pinto FJ, Puryear J, Popp RL. Tricuspid annular motion. J Am Soc Echocardiogr 1991;4:131—9. [16] Koestenberger M, Ravekes W, Everett AD, et al. Right ventricular function in infants, children and adolescents: reference values of the tricuspid annular plane systolic excursion (TAPSE) in 640 healthy patients and calculation of z score values. J Am Soc Echocardiogr 2009;22:715—9. [17] Meluzin J, Spinarova L, Bakala J, et al. Pulsed Doppler tissue imaging of the velocity of tricuspid annular systolic motion; a new, rapid, and non-invasive method of evaluating right ventricular systolic function. Eur Heart J 2001;22:340—8. [18] Lopez L, Colan SD, Frommelt PC, et al. Recommendations for quantification methods during the performance of a pediatric echocardiogram: a report from the Pediatric Measurements Writing Group of the American Society of Echocardiography Pediatric and Congenital Heart Disease Council. J Am Soc Echocardiogr 2010;23:465—95 [quiz 576—7]. [19] Voigt JU, Pedrizzetti G, Lysyansky P, et al. Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging. J Am Soc Echocardiogr 2015;28:183—93. [20] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307—10. [21] Koopman LP, Slorach C, Hui W, et al. Comparison between different speckle tracking and color tissue Doppler techniques to measure global and regional myocardial deformation in children. J Am Soc Echocardiogr 2010;23:919—28. [22] Marcus KA, Mavinkurve-Groothuis AM, Barends M, et al. Reference values for myocardial two-dimensional strain echocardiography in a healthy pediatric and young adult cohort. J Am Soc Echocardiogr 2011;24:625—36. [23] Perdreau E, Seguela PE, Jalal Z, et al. Postoperative assessment of left ventricular function by two-dimensional strain (speckle tracking) after paediatric cardiac surgery. Arch Cardiovasc Dis 2016. [24] 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 endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and the Canadian Society of Echocardiography. J Am Soc Echocardiogr 2010;23:685—713 [quiz 86—8]. [25] Marwick TH. The deconvolution of diastole. J Am Coll Cardiol 2009;54:47—8.

C. Karsenty et al. [26] Levy PT, Holland MR, Sekarski TJ, Hamvas A, Singh GK. Feasibility and reproducibility of systolic right ventricular strain measurement by speckle-tracking echocardiography in premature infants. J Am Soc Echocardiogr 2013;26:1201—13. [27] Dragulescu A, Mertens LL. Developments in echocardiographic techniques for the evaluation of ventricular function in children. Arch Cardiovasc Dis 2010;103:603—14. [28] Haddad F, Hunt SA, Rosenthal DN, Murphy DJ. Right ventricular function in cardiovascular disease, part I: anatomy, physiology, aging, and functional assessment of the right ventricle. Circulation 2008;117:1436—48. [29] Pettersen E, Fredriksen PM, Urheim S, et al. Ventricular function in patients with transposition of the great arteries operated with arterial switch. Am J Cardiol 2009;104:583—9. [30] Farah MC, Castro CR, Moreira Vde M, et al. The impact of preexisting myocardial remodeling on ventricular function early after tetralogy of Fallot repair. J Am Soc Echocardiogr 2010;23:912—8. [31] Klitsie LM, Kuipers IM, Roest AA, et al. Disparity in right vs left ventricular recovery during follow-up after ventricular septal defect correction in children. Eur J Cardiothorac Surg 2013;44:269—74. [32] Asimakopoulos G. Systemic inflammation and cardiac surgery: an update. Perfusion 2001;16:353—60. [33] Haddad F, Doyle R, Murphy DJ, Hunt SA. Right ventricular function in cardiovascular disease, part II: pathophysiology, clinical importance, and management of right ventricular failure. Circulation 2008;117:1717—31. [34] Choi JO, Shin DH, Cho SW, et al. Effect of preload on left ventricular longitudinal strain by 2D speckle tracking. Echocardiography 2008;25:873—9. [35] Burns AT, La Gerche A, D’Hooge J, MacIsaac AI, Prior DL. Left ventricular strain and strain rate: characterization of the effect of load in human subjects. Eur J Echocardiogr 2010;11:283—9. [36] Yotti R, Bermejo J, Benito Y, et al. Validation of noninvasive indices of global systolic function in patients with normal and abnormal loading conditions: a simultaneous echocardiography pressure-volume catheterization study. Circ Cardiovasc Imaging 2014;7:164—72. [37] Ferferieva V, Van den Bergh A, Claus P, et al. The relative value of strain and strain rate for defining intrinsic myocardial function. Am J Physiol Heart Circ Physiol 2012;302:H188—95. [38] Schlangen J, Petko C, Hansen JH, et al. Two-dimensional global longitudinal strain rate is a preload independent index of systemic right ventricular contractility in hypoplastic left heart syndrome patients after Fontan operation. Circ Cardiovasc Imaging 2014;7:880—6. [39] Khraiche D, Ben Moussa N. Assessment of right ventricular systolic function by echocardiography after surgical repair of congenital heart defects. Arch Cardiovasc Dis 2016;109:113—9. [40] Srinivasan C, Sachdeva R, Morrow WR, Greenberg SB, Vyas HV. Limitations of standard echocardiographic methods for quantification of right ventricular size and function in children and young adults. J Ultrasound Med 2011;30:487—93. [41] van der Hulst AE, Roest AA, Westenberg JJ, Kroft LJ, de Roos A. Cardiac MRI in postoperative congenital heart disease patients. J Magn Reson Imaging 2012;36:511—28. [42] Pedersen TA, Andersen NH, Knudsen MR, Christensen TD, Sorensen KE, Hjortdal VE. The effects of surgically induced right bundle branch block on left ventricular function after closure of the ventricular septal defect. Cardiol Young 2008;18:430—6. [43] Medvedofsky D, Addetia K, Hamilton J, Leon Jimenez J, Lang RM, Mor-Avi V. Semi-automated echocardiographic quantification of right ventricular size and function. Int J Cardiovasc Imaging 2015;31:1149—57.

Please cite this article in press as: Karsenty C, et al. Two-dimensional right ventricular strain by speckle tracking for assessment of longitudinal right ventricular function after paediatric congenital heart disease surgery. Arch Cardiovasc Dis (2016), http://dx.doi.org/10.1016/j.acvd.2016.09.003