The experiment was performed in an open chest porcine model. Data for the present study were collected from experiment series published in two previous articles [6, 24]. The open model facilitated recording of high quality echocardiographic images needed for the study, otherwise difficult to obtain in the pigs.
The study was approved by Nationals Animal Research Authority of Norway ( trial registration number: FOTS 3866) and carried out in accordance to the European Convention for the protection of vertebrate animals used for experimental and other scientific purposes, the European Union Directive 2010/63/EU .
The animals (n = 14, Norwegian land race swine, mean weight 52 ± 4.3 kg) were fasting overnight aside from free water access. Before transportation to the laboratory they were pre-medicated by an intramuscular injection with ketamine (20 mg/kg), azaperone (3 mg/kg) and atropine (20 µg/kg). Anaesthesia was induced with intravenous pentobarbital (3 mg/kg) and morphine (2 mg/kg), and maintained with morphine-infusion (1-2 mg/kg/h) and isoflurane-inhalation (1 to 1.5 %). Neuromuscular blocking agents were not used. Level of anaesthesia was monitored by continually hemodynamic measurements and observation of clinical signs. When the protocol was finished, the animals were euthanatized by bolus infusion of 80 mmol potassium chloride and 1000 mg pentobarbital. The animals were mechanically ventilated via a tracheostomy tube by a fraction of air/oxygen (FiO2 0.4), with tidal volumes of 10 to 15 ml/kg and surgically prepared with sternotomy as previously reported [6, 24]. Three-lead ECG was obtained by surface leads. Pacemaker leads were sutured to the right atrium. For temperature control, a water circulated catheter (Cool Line; Zoll, Chelmsford, MA, USA), was inserted into the inferior vena cava by cannulation of the left femoral vein and connected to the thermal regulation system (Coolgard 3000; Zoll). A pulmonary artery catheter (Swan-Ganz CCO; Edwards Lifesciences, Irvin, CA, USA) was inserted through the right jugular vein to measure central temperature and hemodynamic variables. Left ventricular pressure was measured by a micro-manometer pressure transducer (MPR-500; Millar Instruments, Houston, TX) placed via the right carotid artery. A Vivid 7 scanner (GE Vingmed Ultrasound, Horten, Norway) with 2.5 / 2.75 MHz probe directly on the heart, was used for echocardiographic and Doppler recordings.
Hemodynamic data regarding systolic and diastolic function obtained from the experiment has been published [6, 24], but is included as background data in the current manuscript. All data presented concerning the electromechanical relations, has been analysed for this study.
Measurements and analysis
Measurements were obtained at body temperature 38 °C and 33 °C. 33 °C was chosen representing the median temperature according to the current recommended guidelines at that time, with a targeted temperature of 32-34 °C . In addition, the pigs’ baseline temperature was 38 °C, and a drop to 33 °C corresponded to a drop from 37 °C to 32 °C in humans. Measurements were made at both spontaneous heart rate (HR) and during atrial pacing, 100 beats/min (bpm). All measurements and recordings were made in three echocardiographic views over three consecutive heart cycles and the mean values were calculated. The pacing enabled measurements at heart rate 100 bpm, thereby compensating for individual variability and hypothermia induced heart rate reduction. ECG, two-dimensional (2D) and Doppler echocardiography were recorded and analysed offline (EchoPac version 202, GE Healthcare, Horten, Norway).
Calculations of the electrical events
The ECG lead II was used for electrical measurements. The electrical systole (QT interval) was measured from onset of QRS to end of T wave (T end) in ECG. The QT interval was heart rate corrected according to the Bazett’s formula (QTc) .T end was determined by using the manual tangent method and defined as the intersection of the isoelectric line with the tangent to the steepest downslope of the T wave. T peak was defined as the maximum positive or negative deflection of the T wave from the isoelectric line [28, 29] . Electrical dispersion was measured as variation in T peak to T end duration (TpTe).
Calculation of the mechanical events
From the echocardiographic apical 4- chamber long axis view, mitral and aortic valve (AV) opening (O) and closing (C) were recorded. Isovolumic contraction and isovolumic relaxation time were measured in long axis view from mitral valve closing to aortic valve opening and aortic valve closing to mitral valve opening respectively.
Ejection time was measured as the duration of the pulsed wave Doppler signal in left ventricular outflow tract (ETPW). The interval from onset of QRS to the Doppler registered AVC represented the duration of the mechanical systole (QAVC). EMW was calculated by subtracting the electrical systole from the duration of the mechanical systole in the same heartbeat, QAVC-QT interval, (Fig 1).
Global cardiac function
Left ventricular volumes were measured from 2D echocardiographic images (apical 4- and 2- chamber views), and ejection fraction was calculated by the modified Simpson biplane method. Systolic velocity was recorded from mitral ring tissue Doppler velocity images in apical 4-chamber views. Cardiac output was measured by pulmonary artery catheter thermodilution and stroke volume calculated. Peak systolic left ventricular pressure was recorded by the micromanometer catheter placed in the left ventricle.
Regional strain and mechanical dispersion
Longitudinal strain was obtained by speckle tracking echocardiography from the basal and mid segments in three apical left ventricular projections (4-chamber, 2-chamber and long axis, framerate 63 ± 15 ms). The endocardial border was traced manually, and region of interest was adjusted to fit the myocardial thickness. Segments that failed to track were manually adjusted, and if subsequent failing the segments were excluded. The echocardiographic probe was directly placed on the heart with a subsequent loss of the apical segments, whereof peak strain was derived from 12 segments. Time to peak strain in each segment was defined as the time from Q onset on ECG to peak longitudinal strain in three cycles and averaged. Mechanical dispersion was defined as the standard deviation of time to peak negative strain  in the 12 left ventricular segments. Septal and lateral strains were measured from basal segments in 4-chamber view.
Doppler registered QT and QAVC were re-measured by the same observer to assess intraobserver repeatability. A second observer measured the same variables in seven random selected animals to assess interobserver reproducibility.
Statistical analyses were performed in SPSSv.25 software (SPSS, Inc., Chicago, IL, USA). Parametric data are presented as mean ± standard deviation. Data from normothermia and hypothermia at spontaneous and heart rate 100 bpm were compared by Student´s t-test. P-value, p < 0.05 was considered statistically significant. Intra- and interobserver variations were analysed by using intraclass correlation coefficient concerning single measures .