Summary
Background
Left ventricular (LV) dysfunction may complicate paediatric cardiac surgery with cardiopulmonary bypass, notably after long aortic cross-clamping (ACC). Assessment of occult myocardial injury by conventional echocardiographic variables may be difficult in the postoperative period.
Aims
To evaluate the feasibility of two-dimensional (2D) strain in the postoperative period, and to assess the effect of ACC duration on this variable.
Methods
Thirty-three paediatric patients (age < 18 years) with congenital heart disease undergoing cardiac surgery with cardiopulmonary bypass were included in this prospective single-centre study. Daily echocardiography was performed from the day before surgery to the fifth postoperative day. LV ejection fraction and LV 2D strain were measured. The cohort was divided into three groups according to ACC duration (group 1: < 30 minutes; group 2: 30–80 minutes; group 3: > 80 minutes).
Results
Mean age and weight were 4.2 ± 2.5 years and 15.1 ± 5.2 kg, respectively. Feasibilities of longitudinal, circumferential and radial strains were good, and quite similar to conventional variables. Compared with conventional variables, intra- and interobserver agreements regarding 2D strain were better ( r = 0.916, P < 0.001 and r = 0.855, P < 0.001 for longitudinal strain versus r = 0.156, P = 0.54 and r = 0.064, P = 0.80 for LV ejection fraction by Simpson’s method). Postoperative evolution of longitudinal and circumferential strains was significantly different between the three groups ( P < 0.001), whereas there was no difference using conventional variables.
Conclusion
Postoperative LV 2D strain is a feasible and reproducible method. Strain measurements seem to indicate correlation with ACC duration.
Résumé
Contexte
La chirurgie cardiaque pédiatrique avec circulation extracorporelle peut être compliquée de dysfonction ventriculaire gauche, notamment après une longue durée de clampage aortique. L’évaluation de l’atteinte cardiaque par les paramètres échographiques conventionnels peut être difficile en période postopératoire.
Objectifs
L’objectif de l’étude était d’évaluer la faisabilité du 2D strain en période postopératoire et d’évaluer l’impact de la durée de clampage aortique sur ce paramètre.
Méthodes
Trente-trois patients âgés de moins de 18 ans, présentant une cardiopathie congénitale nécessitant une chirurgie cardiaque avec circulation extracorporelle, ont été inclus dans cette étude prospective monocentrique. Une échographique quotidienne a été réalisée la veille de la chirurgie au 5 e jour postopératoire. La fraction d’éjection ventriculaire gauche et le 2D strain du ventricule gauche ont été mesurés. La cohorte a été divisée en trois groupes en fonction de la durée de clampage aortique (< 30 min pour le groupe 1, 30–80 min pour le groupe 2, > 80 min pour le groupe 3).
Résultats
L’âge moyen et le poids moyen étaient respectivement de 4,2 ± 2,5 ans et 15,1 ± 5,2 kg. Le 2D strain longitudinal, circonférentiel et radial avait une bonne faisabilité, similaire à celle des paramètres conventionnels. La variabilité intra- et inter-observateur était meilleure que celle des paramètres conventionnels ( r = 0,916, p < 0,001 et r = 0,855, p < 0,001 pour le strain longitudinal versus r = 0,156, p = 0,54 et r = 0,064, p = 0,80 pour le Simpson). L’évolution postopératoire du strain longitudinal et circonférentiel était significativement différente entre les trois groupes ( p < 0,001).
Conclusion
Le 2D strain est une méthode faisable et reproductible en postopératoire. L’évolution postopératoire de ce paramètre semble être corrélée à la durée de clampage aortique.
Background
Cardiopulmonary bypass (CPB) is routinely used in the surgical repair of many congenital heart diseases. CPB is one of the causes of postoperative left ventricular (LV) dysfunction , which can result in low cardiac output syndrome and multiple organ failure. Therefore, reliable assessment of postoperative LV function is necessary. Echocardiography is the gold standard in the cardiac intensive care unit (CICU), using visual assessment and measurement of LV ejection fraction (LVEF) using the methods of Teichholz or Simpson. However, these methods lack reliability and reproducibility, especially when the cardiac anatomy is abnormal .
Analysis of two-dimensional (2D) strain by speckle tracking is a 2D echocardiographic technique that relies on a frame-by-frame follow-up of acoustic markers, called “speckles”, through the myocardium . Speckle displacement represents myocardial motion, and ventricular function can be studied in several dimensions (longitudinal, radial and circumferential). The value of this technique has been demonstrated in several cardiovascular diseases in adults and children, for diagnosis, prognosis and treatment ; however, its use in the CICU after paediatric cardiac surgery has not been studied so far.
The main objective of this study was to evaluate the feasibility and reproducibility of LV 2D strain after congenital heart disease surgery with CPB. The second objective was to assess the effect of aortic cross-clamping (ACC) duration on this variable.
Methods
Study population
A prospective single-centre study was conducted in our paediatric CICU from 1st July 2012 to 1st November 2012. All patients aged < 18 years who required congenital heart disease surgery with CPB were included. Exclusion criteria were: age > 18 years; univentricular heart; cardiac surgery without CPB; and a need for postoperative circulatory assistance. For each patient, preoperative data were collected: age; sex; weight; height; congenital heart disease; and surgical repair. Pre- and postoperative data were also collected, including CPB and ACC durations; ultrafiltration volume; biological variables (B-type natriuretic peptide, lactate and troponin); duration of hospitalization; and durations of respiratory and inotropic support. The different inotropic drugs used were milrinone (0.5–1 μg/kg/min), adrenaline (0.05–0.1 μg/kg/min) and levosimendan (0.2 μg/kg/min in one 24-hour perfusion), using routine protocols.
Transthoracic echocardiographic assessment
For every patient, standard echocardiography was performed the day before surgery, 4 hours after the end of surgery and every day until the fifth postoperative day (POD), using a commercially available ultrasound imaging system (Vivid 7 preoperatively and Vivid I postoperatively; GE Medical, Milwaukee, WI, USA) and a 7S or 3S probe, according to the size of the patient. The frame rate was kept between 80 and 110 Hz, and the electrocardiogram was connected to detect systole, diastole and global systolic peak of 2D strain. The best views were selected to meet the following criteria: several cardiac cycles recorded; and a central and complete picture with a good image quality. LVEF by Teichholz was measured in the parasternal long-axis view in M-mode; LVEF by Simpson was measured in the four-chamber view in 2D mode. Longitudinal strain was measured in the four-chamber view. Circumferential and radial strains were measured in the short-axis view at the level of the mitral valve or the papillary muscles of the mitral valve ( Fig. 1 ). All DICOM data were analysed offline with EchoPAC software (GE Medical, Milwaukee, WI, USA). The end of the systole was defined by the closure of the aortic valve in the parasternal long-axis view. To measure global systolic peak of longitudinal, circumferential and radial strains, the limit of the endocardium was traced by the operator at the end of the systole. The software automatically generated a second line at the level of the epicardium, delineating a region of interest, which could be adjusted by the operator. The measurement of strain was automatically performed, and the analysis was accepted if the software validated the measurement for each cardiac segment (septal, lateral, inferior, posterior, anterior and anteroseptal). If a segment was not accepted, the limit of the endocardium was retraced. To determine global strain, the strain values were averaged for the four-chamber and short-axis views. Strain values are dimensionless and are expressed as percentages. Negative strain values reflect shortening, while positive strain values reflect lengthening or thickening.
Statistical analysis
Continuous variables are presented as means ± standard deviations. Regarding the diversity of congenital heart diseases, we divided our cohort into three groups, using ACC duration, which corresponds to the duration of ischaemia, which is more likely to cause myocardial injury. ACC duration was < 30 minutes in group 1, 30–80 minutes in group 2 and > 80 minutes in group 3. Non-parametric tests (Kruskal–Wallis and Mann–Whitney) were used to compare clinical and biological variables between the three groups. Inter- and intraobserver variabilities of echocardiographic measurements were assessed for 20 randomly selected patients (images acquired on POD 1). Intraobserver variability was assessed by a paediatric cardiology fellow, performing offline analysis on the same patients, 4 weeks apart, to reduce recall bias. Interobserver variability was assessed by a second investigator (paediatric cardiology consultant), who was unaware of the previous results, and performed the investigation on the same 20 patients. Intra- and interobserver agreements were assessed by calculating Spearman’s correlation coefficient, accepting P < 0.05 as significant. Bland–Altman and intraclass correlation coefficient methods were used to further examine the agreement between the measurements of the two investigators. The relationships between 2D strain measurements, ACC duration and clinical variables were evaluated by calculating Spearman’s correlation coefficient, accepting P < 0.05 as significant. We performed a two-factor analysis of variance (ANOVA) test (group and time) to analyse the evolution of LVEF and 2D strain measurements between the three groups. To analyse the evolution of echographic variables after atrial septal defect closure, a non-parametric test (Wilcoxon test) was performed, accepting P < 0.05 as significant. The software used for analysis was SPSS 17.0 (SPSS Inc., Chicago, IL, USA).
Ethics
Informed verbal consent was obtained from legal representatives for every patient. The study was approved by our local ethics committee. A written consent form was not required according to French law, as the studies were part of the regular management of the children.
Results
Population characteristics
From 40 children who underwent surgery with CPB during the period, 33 patients were included in the study. Seven patients were excluded: two with univentricular heart; two requiring postoperative circulatory assistance; one with infectious endocarditis; one death in theatre; and one with no ACC. The characteristics of the patients are summarized in Table 1 , and the congenital heart diseases encountered are reported in Table 2 . No patient had preoperative LV dysfunction. Inotropic support was started at the end of CPB in 26 patients. Group 1 only comprised children who underwent atrial septal defect closure. Significant differences were noted between groups regarding the durations of mechanical ventilation and inotropic support, perioperative lactate concentration and troponin concentration on POD 0.
