This experimental protocol was part of a larger study and followed the guidelines of the ethical use of animals of the University of Liege (approval number 655), in accordance with the European directive 2010/63/EU. Ten adult nulliparous female Saanen goats were included in this study. They were aged 22 to 28 months (mean age: 24.7 ± 2.1 months), they weighed 51 to 80 kg (mean body weight: 65.1 ± 8.3 kg) and they were accustomed to handling. Included animals were bought from a private breeding farm producing goat milk located less than 20km from the University. All animals were considered healthy based on medical history and the absence of abnormalities on physical examination, cardiac auscultation, electrocardiography, haematology and standard biochemistry panel. A complete standard two-dimensional, time-motion mode and Doppler echocardiography (from right and left side) was performed on each goat before starting the protocol to ensure that the studied goats were free of any cardiac disease. Studied goats were housed in groups of three or four in an enclosed barn with natural ventilation, in accordance with appendix A of the European convention for the protection of vertebrate animals used for experimental and other scientific purposes (ETS NO. 123). All goats were fed with hay ad libitum. Seven of the ten studied goats were randomly chosen to be accustomed to perform a standardized exercise on treadmill, which consisted of walking on a treadmill (Domyos TC290, Decathlon, Villeneuve d'Ascq, France) for 6 minutes at 4 km/h and at 5% slope as previously described .
Before imaging, the hair was shaved from the 3rd to the 5th right intercostal space just caudal to the right triceps muscle mass, from 3 to 5 cm below the right olecranon to 5 to 10 cm above it. The shaved area was then copiously rinsed with water and acoustic coupling was obtained using ultrasound gel. First repeatability and variability of the technique were tested at rest and the echocardiographic protocol was repeated three times at one-day intervals by the same observer (AAL) on the ten goats. After this study of repeatability, the same echocardiographic protocol was performed on seven of the goats by the same observer (AAL) immediately before and after a standardized exercise on treadmill. The exercise stress echocardiography was performed between 1 week and 4 weeks after the repeatability study. Then, MI was induced in five of the goats. These goats were anaesthetised with 0.3 mg/kg BW of midazolam 5 mg/3 ml (Dormicum, Roche, Bruxelles, Belgium) and 10 mg/kg BW of ketamine 50 mg/ml (Ketalar, Pfizer, Bruxelles, Belgium) using a catheter in the right jugular vein. Tracheal intubation was performed using a 9.0 mm or 10.0 mm ID endotracheal tube and the goats were ventilated with a tidal volume of 15 ml/kg BW with a respirator using a mixture of 2% isoflurane in oxygen. The left common carotid artery (arteria carotis communis) was punctured and canulated with a sterile 6 French introducer sheath (Cook Medical, Bloomington, IN, USA) to allow introduction of a Judkins left 6 French catheter (Medtronic, Heerlen, Pays-Bas) under cine-fluoroscopic guidance (Pulsera, Philips, Eindhoven, Pays-Bas) into the right and left coronary trunks. As soon as the coronary branches from the left coronary artery were identified, infarction was performed using a coil embolization. A microguide (Cook Medical, Bloomington, IN, USA) was advanced in the left circumflex coronary artery (ramus circumflexus sinister) and placed into the left marginal branch of the left circumflex artery (ramus marginis ventricularis sinister), into the posterior descending branch of the left circumflex artery (ramus interventricularis subsinuosus), or just upstream from these two vessels. A coil delivering microcatheter (Cook Medical, Bloomington, IN, USA) was placed into the selected location and a platinum Nester embolization microcoil 30x2mm or 30x3mm (Cook Medical, Bloomington, IN, USA) was pushed into the vessel using a wire coil pusher (Cook Medical, Bloomington, IN, USA). Ten minutes after coil embolization, total coronary occlusion of the left marginal branch (in 1 goat) or of the posterior descending branch (in 2 goats) or of the left circumflex artery just upstream from these two vessels (in 2 goats) was confirmed by angiography. When cardiopulmonary status of the goat was stable, the left common carotid artery, the subcutaneous tissue and the skin were sutured. Then the goat was awakened in sternal position and endotracheal tube was removed as soon as a gag reflex was present. After a 24h post-operative intensive care follow-up including continuous ECG and pain management with intravenous administration of 1.1 mg/kg BW of flunixin (Finadyne, Intervet, Bruxelles, Belgium), 2DST echocardiography was performed on unanaesthetised standing goats to assess average and regional myocardial function. At the end of the experiments, all goats were deeply sedated with intravenous administration of xylazine 0.2 mg/kg BW (Proxylaz, Prodivet, Eynatten, Belgium) then euthanized with intravenous administration of 10 ml of an association of embutramide 200 mg/ml, mebezonium iodide 50 mg/ml and tetracaine hydrochloride 5 mg/ml (T 61, Intervet, Bruxelles, Belgium).
An ultrasound system (Vivid i, Software version 9.1.0, General Electric Healthcare Europe GmbH, Diegem, Belgium) equipped with a 1.5-3.6 MHz phased array transducer (GE 3S-RS probe, General Electric Healthcare Europe GmbH, Diegem, Belgium) was used to perform the echocardiography. All examinations were performed by the same observer (AAL) on standing animals with the right forelimb extended by an assistant as far forward as tolerated by the goat. Echocardiographic images were recorded digitally as cine-loops. An imaging depth between 11 and 15cm, and a frame rate between 54 and 115 frames/second in 2D-mode, was used for all examinations. 2D image terminology and orientation recommended by the Echocardiography Committee of The Specialty of Cardiology, American College of Veterinary Internal Medicine, were used . At each examination, a right parasternal long axis four chambers view with chordae tendinae and mitral valve clearly visible was obtained in 2D-mode. From this view, the transducer was turned clockwise until obtaining a 2D-mode right parasternal short-axis view of the left ventricle at the level of the papillary muscles and several cine-loops of this view were recorded. Care was taken to avoid oblique views and to allow good endocardial delineation.
Offline 2DST analysis
All 2DST analyses were performed blindly, in a random order, by the same observer (AAL) using the 2D strain application of a specific software (Echo Pac System for Vivid i, Software version 108.1.5, General Electric Healthcare Europe GmbH, Diegem, Belgium) as previously described by several authors [14-16, 19, 20]. All variables were measured five times on five different non consecutive cycles. For each measurement, a 2D-mode right parasternal short-axis cine-loop during one cardiac cycle and with an optimal image quality was selected and the semi-automated 2D strain application was started using the SAX-PM option. The regions of interest (ROI) were manually traced starting at the anterior interventricular septum and following the endocardial border in a clockwise manner. After careful endocardial border definition, 2D-strain application was launched and enabled an automated radial and circumferential LV strain analysis. The software divides the myocardium into six segments according to the human guidelines (Ant: anterior; Antsept: anteroseptal; Inf: inferior; Lat: lateral; Post: posterior; Sept: septal), and allows assessment of the tracking quality (Fig. 1). After the operator has visually verified the tracking, the analysis was approved and six curves, each corresponding to each ROI segment, were analyzed. Peak strain and strain rate measurements were performed automatically by the software and provided in a table. Circumferential strain (Sc), radial strain (Sr), circumferential strain rate (SRc), radial strain rate (SRr), radial displacement (Dr), rotation (Rot) and rotation rate (RotR) were measured. For each variable one or several peaks were measured for each ROI segment. Each peak value was automatically calculated by the program then manually adapted when needed. Measurements for each ROI segment included: one systolic peak for Sc and Sr (Sc-S and Sr-S); 3 peaks for SRc, SRr and RotR, comprising one systolic peak (SRc-S, SRr-S and RotR-S), one early diastolic peak (SRc-E, SRr-E and RotR-E) and one late diastolic peak (SRc-A, SRr-A and RotR-A); and one single systolic peak for DR and Rot (DR-S and Rot-S) (Fig. 2). The average peak values of each 2DST variables corresponding to the mean of the 6 ROI segments measurements was calculated manually. The time of the aortic valve closure was automatically calculated by the software since this calculation was demonstrated to be reliable in goats for 2DST short axis analyses at the papillary muscle level .
Heart rate (HR) was calculated from the electrocardiograms from five cineloops containing five successive cardiac cycles and including the ones used for the offline measurements. Echocardiographic recordings at rest were only measured when HR was below 110 bpm. Exercise stress echocardiography was performed immediately after the standardized treadmill exercise test, images were recorded within 2 min after exercise cessation and measurements after exercise were performed only when HR was between 130 and 140 bpm.
Statistical analyses were performed using a computer statistical software (Statistical Analysis System, version 9.1, SAS Institute Inc, Cary, NC, USA) and a standard computer software (Microsoft Office Excel 2003, Microsoft corp, Redmond, WA, USA). The repeatability of the measurements was established by analysis of variance with respect to goat and time factors (two-way ANOVA) and by calculating coefficients of variation both between cycles and between days for each parameter. A two-way ANOVA considering goats, days and interaction between goats and days as factors, allowed the determination of between-day differences of the measurements. Moreover observed means, standard deviations (SD), least square means, and standard errors (SE), were automatically calculated for each variable and each day. The within-goat within-day between-cycle variability was evaluated using the coefficient of variation (Between-cycle CV) measured from SD and observed means obtained from the two-way ANOVA for each variable. The within-goat between-day variability (Between-day CV), representing the variability of the same repeated measurements on the same goat, without taking into account the day of examination (i.e. the measurements of each parameters for each goat were pooled without considering the day of the measurements), was measured from SD and observed means obtained in a one-way ANOVA considering only goats as factor. Degree of variability of each measurement was defined as applied in previous studies [16, 19]: variables with a CV inferior to 15% were considered to have low variability, those with a CV between 15% and 25% were considered to have moderate variability, and those with a CV superior to 25% were considered to have high variability. In addition to the CV, absolute variability was obtained by calculating the interval of confidence of the mean for each variable. This interval of confidence was defined as the interval within the absolute value of the mean of the studied population had 95% of probability to be included. Superior and inferior limits of this interval were calculated as follows: observed mean + 2.262 x SE and observed mean – 2.262 x SE. Measurements were considered repeatable if both a non-significant result of the two-way ANOVA and a low or moderate variability were observed.
Concerning stress echocardiography analysis, observed means, SD, least square means, and SE were calculated for each variable before and after exercise. A two-way ANOVA considering goats, exercise, and interaction between goats and exercise as factors, allowed the comparison between measurements obtained at rest and immediately after exercise.
One to three days after stress-echocardiography, five of the studied goats underwent the experimental model of ischemic cardiomyopathy procedure. Echocardiography performed at rest before exercise test was considered as baseline. Then the same echocardiographic protocol was performed on the five goats by the same observer at 24h after MI induction and each variable was measured five times on five different non consecutive cardiac cycles. Observed means, SD, least square means, and SE were calculated for each variable. A two-way ANOVA considering goats, MI, and interaction between goats and MI as factors, allowed the comparison between variable obtained during baseline and after MI creation.
For all statistical analyses, a P-value inferior to 0.05 was considered significant.