Major hemodynamic fluctuations frequently precipitate during general anesthesia. These acute alterations make the intraoperative management of patients challenging. Sympathoadrenal and renin-angiotensin systems are assumed to be biochemical mediators of such undesirable circulatory effects.1 To decrease associated morbidity and mortality, information regarding the effects of general anesthesia on the cardiovascular system and optimization of intraoperative hemodynamic monitoring is crucial.
In anesthetized small animals, blood pressure monitoring is routinely performed to estimate the cardiovascular status and organ perfusion. Hypotension indicates mean arterial blood pressure (MAP) < 60 to 65 mm Hg.2 Anesthetists use MAP as a major determinant for therapeutic decisions perioperatively. By referring to Ohm's law and fluid flow, it is evident that pressure (ie, MAP) does not equate directly to flow (ie, cardiac output [CO]); hence, changes in MAP can occur due to changes in CO or systemic vascular resistance or both.3,4 If hypotension is treated aggressively by solely focusing on MAP values, a deleterious impact on overall cardiac performance, coronary blood flow, and CO may occur. Cardiac output is affected by stroke volume, heart rate, preload, afterload, and contractility. Therefore, procuring an accurate picture of intraoperative hemodynamics by measuring CO as the primary variable and analyzing it along with secondary variables like MAP and heart rate can boost the quality of patient care and anesthetic management. Pulmonary artery thermodilution (PATD) is an invasive technique for measuring flow and pressures and is considered the “gold standard” for CO monitoring in humans and animals.5,6 However, the utility of PATD in clinical settings is questionable due to the risks of cardiac catheterization, required training and skills, lower time efficiency, and high instrumentation costs.7,8
In the past few years, there has been a surge of alternate minimally invasive devices to replace PATD, such as esophageal Doppler monitor (EDM)9–15 and transesophageal echocardiography (TEE),16–19 in dogs. Even though recent literature14 suggests the potential benefit of a veterinary EDM for closely tracking hemodynamics in a canine model hemorrhagic shock, there are minimal data on its performance during pharmacological interventions. Similarly, limited literature exists to determine whether TEE can guide cardiovascular monitoring during drug-induced acute changes in CO. During pharmacologic manipulation of hemodynamics in healthy, anesthetized Beagle dogs, the objectives of the present study were to (1) evaluate the level of agreement of PATD with EDM and TEE for measuring CO, and (2) investigate the relationship of EDM-derived variables, ie, stroke distance (StrokeD), minute distance (MinuteD), flow time corrected (FTc), peak velocity (PV), and mean acceleration (MA) with respect to CO and MAP values. We hypothesized that (1) EDM and TEE will exhibit an acceptable level of agreement with PATD in this experimental setting, and (2) EDM-derived variables will closely correlate with the acute, drug-specific variations in CO and MAP values.
Methods
Study animals
Six purpose-bred, male adult Beagle dogs (age, 11 to 17 months; mean ± SD weight, 10.8 ± 0.7 kg) participated in this prospective, crossover, randomized experimental study. Virginia Tech University-Institutional Animal Care and Use Committee (protocol number 20-229) approved the experimental procedures and animal use for this study. All dogs were deemed healthy based on a thorough physical exam, CBC, and serum chemistry panel. Based on previous canine studies evaluating CO monitoring techniques,14,20–23 an a priori power analysis confirmed a required sample size of 6 animals to demonstrate a 30% significant difference in CO in response to hemodynamic manipulation, assuming a statistical power of 0.8 and an alpha level of 0.05 (G*Power 3.1, Heinrich-Heine-Universität).
Induction of general anesthesia and instrumentation for standard monitoring
Before the day of the experiment, dogs were fasted for 12 hours with free access to water. On the study day, following cephalic vein catheterization, oxygen was supplemented (4 L/min) for 5 minutes using a fitted facemask connected to a small animal anesthesia machine and rebreathing system. General anesthesia was induced with IV propofol until tracheal intubation was achieved. The dog was transitioned into dorsal recumbency and connected to the rebreathing circuit and a ventilator-integrated anesthesia machine (Aestiva 5/7900; GE-Datex Ohmeda). Anesthetic maintenance was isoflurane in oxygen (1 to 2 L/min) targeting the end-tidal concentration of isoflurane (ETISO) between 1.4 and 1.6% as continuously measured by an infrared gas analyzer included in a multiparameter monitor (S/5; GE-Datex Ohmeda). A lead II ECG, heart rate (HR), end-tidal concentration of carbon dioxide, esophageal temperature, pulse oximetry, and spirometry were also recorded by the same monitor. A forced-air warming device and circulating water blanket helped maintain body temperature between 36.7 and 38 °C throughout the study.
Rocuronium was administered as an IV bolus of 0.4 mg/kg followed by a constant rate infusion rate of 0.4 mg/kg/h to produce and maintain neuromuscular paralysis. The train-of-four supramaximal stimulation (Stimpod 450X; Xavant Technology, SA) of the common peroneal nerve was used to monitor the depth of blockade. The dogs underwent controlled ventilation throughout the anesthetic period with a tidal volume of 12 mL/kg and respiratory rate of 8 to 18 breaths/min to maintain the end-tidal concentration of carbon dioxide between 35 and 45 mm Hg. Arterial catheterization was performed aseptically in the dorsal pedal artery for recording invasive systolic, diastolic, and mean arterial pressure. Another multiparametric monitor (Carescape B850; GE Healthcare) was connected to a heparinized saline-flushed (3 IU/mL) disposable pressure transducer system (Deltran II; Utah Medical Products Inc) that was leveled and zeroed approximately at the level of the right atrium. After clipping and sterile skin preparation, an 18-gauge 3.2-cm IV catheter was inserted in the left jugular vein and was connected to a t-port extension tubing (B Braun Medical Inc) and was designated for the administration of study drugs.
Instrumentation for CO monitoring by intermittent PATD
Using the modified Seldinger technique, a 6-Fr 8.5-cm hemostasis introducer (Fast-Cath; Abbott Cardiovascular) that was aseptically inserted and secured into the right jugular vein was used to advance a 5-Fr 75-cm thermistor-tipped Swan Ganz catheter (132FS; Edwards Lifesciences Corp) until its distal tip was located in the pulmonary artery (Figure 1). The methodology for PATD was exactly replicated from the numerous canine studies.14,20,21 At each time point, data collected from the CO monitor (Carescape B850; GE Healthcare) were equal to the mean of 3 consecutive measurements within a 10% variation. Injections were done manually always by the same researcher to reduce bias, with at least a 3-minute interval between injections.
Placement of a balloon-tipped, multilumen pulmonary artery (PA) thermodilution catheter (Swan Ganz) through a hemostasis introducer inserted into the right jugular vein using the modified Seldinger technique. The catheter is advanced by pressure-waveform guidance first through the right atrium (RA) and then the right ventricle (RV) until its distal tip is located in the PA. The characteristic pressure waveforms and pressure values of RA, RV, and PA are displayed on the cardiac output (CO) monitor using disposable pressure transducer systems. Once in the PA, the balloon can be inflated to occlude the branch of PA and provide pulmonary artery wedge (PAW) pressure that is equivalent to the pressure of the left atrium. The proximal port is used for injecting chilled 0.9% sodium chloride solution for determination of CO by intermittent pulmonary artery thermodilution. The distal port is used for PA pressure measurements and PAW pressure readings if the balloon is inflated (using the balloon port). The thermistor port is connected to the CO monitor via a cable and allows the display of continuous core body temperature values. A temperature curve is plotted over time by the CO monitor as the cool injectate causes the pulmonary artery temperature to fall. Cardiac output values are calculated using the modified Stewart-Hamilton equation.
Citation: American Journal of Veterinary Research 84, 8; 10.2460/ajvr.23.05.0101
Instrumentation for CO monitoring by TEE
A—Transesophageal echocardiographic image of the left ventricular outflow tract: The zoom function is applied to a mid esophageal long-axis view and a mid systolic frame is selected. The left ventricular outflow tract is then measured using an inner edge-to-inner edge technique at a location just proximal to the aortic valve leaflets. Ao = Aorta. LV = Left ventricle. B—Transesophageal echocardiographic transgastric image: This image plane was used to guide placement of the spectral Doppler gate at the level of the left ventricular outflow tract, in the same location where this structure's diameter was measured. This image plane is used because it allows the most parallel alignment with the aortic flow. C—Transesophageal spectral Doppler of left ventricular outflow tract: on the top right corner, the 2-D transgastric image used for directing the placement of the pulsed-wave Doppler gate (double parallel lines) is visualized. The spectrograms of 3 consecutive beats are visualized in the lower half of the image. Using a software package, the area of each spectrogram is traced and the velocity-time integrals are calculated.
Citation: American Journal of Veterinary Research 84, 8; 10.2460/ajvr.23.05.0101
Instrumentation for EDM, velocity-time waveform, recording of EDM-derived indexes, and estimation of CO
The EDM veterinary model (CardioQ-EDMV+; Deltex Medical) attached to a 4.02-MHz continuous Doppler ultrasound emitting probe (K9P 120 cm; Deltex Medical) was lubricated using aqueous gel and inserted into the oral cavity alongside the existent TEE probe (Supplementary Figure S1). Advancement of the probe into the esophagus continued until the probe tip was in the region of the fifth to sixth thoracic vertebra, paralleling the aortic blood flow. The probe was maneuvered by the operator to adjust the depth and subtle rotational movements were made until the real-time velocity-time, triangular aortic waveform was visualized coinciding with a distinct, maximal pitch Doppler “whip crack” sound associated with peak blood flow from the descending thoracic aorta (Supplementary Figure S2). A detailed description of the working principle of this method, Doppler waveform analysis, and derived variables has been recently reported in anesthetized dogs.14 This waveform was utilized to obtain EDM-derived StrokeD, MinuteD, FTc, PV, and MA.
To ensure the reliability of EDM data, the HR from this monitor was cross-matched with the pulse rate from the pulse-oximetry and arterial waveforms and the HR from the ECG before every recording of EDM-derived indexes. Before data collection at each time point, (1) at least 2 minutes were given after a consistent highest quality signal and waveform were obtained, and (2) for each variable, the average value over a 1-minute cycle was recorded.
Pharmacologic manipulation of hemodynamics and data collection
Once instrumentation was completed, each dog went through 4 treatments causing acute pharmacologic variations in the hemodynamics (Figure 3). These treatments were administered in a randomized order (https://www.randomizer.org/). Baseline data were collected before dobutamine (DOBUbaseline), esmolol (ESMObaseline), phenylephrine (PHENbaseline), and high-dose isoflurane (highISObaseline). Predetermined hemodynamic goals were set to ensure a significant treatment-specific change in the hemodynamics was observed. These treatment goals were as follows: (1) DOBU: dobutamine IV infusion of 3 to 10 μg/kg/min aimed to increase CO monitoring by PATD (PATDCO) by > 40% as compared with DOBUbaseline; (2) ESMO: esmolol IV bolus of 100 μg/kg followed by infusion of 50 to 200 μg/kg/min aimed to decrease PATDCO by > 40% as compared with ESMObaseline; (3) PHEN: phenylephrine IV infusion of 0.2 to 1 μg/kg/min aimed for MAP > 120 mm Hg; and (4) highISO: ETISO > 3% aimed for MAP < 50 mm Hg.
Study timeline and data collection performed in 6 isoflurane-anesthetized Beagle dogs subjected to 4 randomized pharmacological interventions to acutely change the hemodynamics. Data included cardiac output measurements by intermittent pulmonary artery thermodilution (PATDCO), transesophageal echocardiography (TEECO), esophageal Doppler monitor (EDMCO), and other EDM-derived indexes. Baseline data were collected before dobutamine (DOBUbaseline), esmolol (ESMObaseline), phenylephrine (PHENbaseline), and high-dose isoflurane (highISObaseline) treatments. Predetermined hemodynamic goals specific to each treatment were (1) DOBU: dobutamine IV infusion of 3 to 10 μg/kg/min aimed to increase PATDCO by > 40% as compared with DOBUbaseline; (2) ESMO: esmolol IV bolus of 100 μg/kg followed by infusion 50 to 200 μg/kg/min aimed to decrease PATDCO by > 40% as compared with ESMObaseline; (3) PHEN: phenylephrine IV infusion of 0.2 to 1 μg/kg/min aimed for mean arterial pressure > 120 mm Hg; and (4) highISO: end-tidal isoflurane concentration > 3% aimed for mean arterial pressure < 50 mm Hg. Ten minutes were given for hemodynamic stabilization after achieving the goal associated with each treatment and before any data were collected. At least a 30-minute washout period was assigned between treatments. Baseline readings before individual treatments were only taken when PATDCO value was within 10% variation as compared to the previous baseline reading. FTc = Flow time corrected. MA = Mean acceleration. MAP = Mean arterial pressure. MinuteD = Minute distance. PV = Peak velocity. StrokeD = Stroke distance.
Citation: American Journal of Veterinary Research 84, 8; 10.2460/ajvr.23.05.0101
Ten minutes were given for hemodynamic stabilization after every treatment goal was achieved and before any data were collected. At least a 30-minute washout period was assigned between treatments to prevent residual cardiovascular effects of the previous treatment from impacting the hemodynamic data of the next treatment. Baseline readings before individual treatments were only taken when PATDCO value was within 10% variation as compared to the previous baseline reading. Three researchers were assigned to collect data from a specific CO-monitoring technique and were blinded to the data recorded by the others. One researcher was designated to administer the treatments. The order for collecting CO data from the 2 test methods in the study (EDMCO and TEECO) was randomized (https://www.randomizer.org/); however, the PATDCO measurements were always performed after acquiring the data from other CO techniques. This was done to prevent the fluid volume from multiple 0.9% saline injections for PATDCO potentially impacting the data from the test methods. The TEE and EDM probes were adjacently placed next to each other in the dog's esophagus, and at least 2 minutes were given after obtaining a consistent quality signal and waveform (for EDM) or clear crisp images (for TEE) before data were acquired. The PATDCO, TEECO, and EDMCO data were obtained at the end of the expiration phase in the respiratory cycle.
Anesthetic recovery
After final data collection, rocuronium was discontinued. The train-of-four ratio ≥ 0.9, consistent spontaneous ventilation, and generation of > 3 cm H2O negative pressure during spontaneous inspiration were considered indicative of recovery from rocuronium-induced neuromuscular block. The Swan Ganz jugular and arterial catheters were removed and the isoflurane vaporizer was turned off. After extubation, 0.3 mg/kg, IV methadone was administered and dogs were moved to individual kennels. They were further monitored using HR, oscillometric blood pressure, respiratory rate and pattern, body temperature, demeanor, appetite, pain assessment, and catheter sites periodically for the next 96 h.
Statistical analysis
To analyze the normality of the physiological effects of different drugs (eg, DOBU and ESMO) across a range of variables (eg, CO and MAP), the Shapiro-Wilk and D'Agostino-Pearson tests were used, and the data were presented as mean ± SD for normally distributed variables. A repeated measures ANOVA was used to compare mean differences across variables among the data acquisition time points within each dog as well as between dogs during the 4 treatments. A pairwise t test for parametric data or a Wilcoxon signed rank test for nonparametric data were conducted for paired sample analysis. The statistical significance for all analyses was set at P < .05.
To assess the correlation between TEECO and EDMCO with PATDCO, a least squares regression analysis was conducted. The bias for the 2 methods was calculated as PATDCO – TEECO and PATDCO – EDMCO, and the normality of the bias for both methods was assessed. Although the bias was normally distributed, the relative bias expressed as a percentage instead of mean bias was analyzed to account for the large physiographical range of CO across different treatments.27 A positive relative bias (%) represented underprediction, and a negative bias represented overprediction. The limits of agreement (LOA) were calculated as relative bias ± 1.96 X SD for a 95% confidence interval. A <30% overall relative bias was deemed acceptable, and the percentage of individual observations with relative bias > 30% were also evaluated for both TEE and EDM. Lin's concordance correlation (ρc) was used to evaluate the reproducibility of the measurements.28
A Bland-Altman analysis for nonuniform differences (to account for proportioning effect) was performed to demonstrate agreement between PATDCO and EDMCO values, and PATDCO and TEECO values,29,30 while a polar plot analysis was used to evaluate trending ability and agreement between the PATD and the 2 test methods.31,32 Statistical software SAS Version 9.4 (SAS Institute Inc) was used for statistical tests, the Bland-Altman plots were created using Excel, Microsoft Corp, and the polar plots were generated using Polar Plot 3 analysis add-in (accessed March 10, 2023; https://andypope.info/charts/polarplot3.html).
Results
The anesthetic induction, maintenance, and recovery were uneventful in all dogs. In each dog, instrumentation for PATD and placement of the thermodilution catheter was successfully performed and no complications were reported. There were no missing data disclosed in the study for PATD, TEE, and EDM. The dogs were normothermic (37.8 ± 0.4 °C) and normocapnic (38 ± 3 mm Hg) throughout the anesthetic period. There were no differences in the total anesthesia time (P = 0.88), time spent during the 4 treatments, ie, DOBU (P = 0.38), ESMO (P = 0.69), PHEN (P = 0.47), and highISO (P = 0.29), among all the dogs. No difference was noted in the time from obtaining a baseline reading to achieving the hemodynamic goal (P > .05) between treatments.
Hemodynamic goals achieved with the treatments across study dogs
The mean ± SD of pertinent standard monitoring and hemodynamic data across the 4 pharmacologic interventions is presented (Table 1). During DOBU, dobutamine dosed at 7 ± 1.5 μg/kg/min was necessary to cause a 45.5 ± 4.5% significant increase (P < .001) in PATDCO values. During ESMO, esmolol infusion of 150 ± 25 μg/kg/min significantly decreased (P < .001) PATDCO values by 44.8 ± 6.3%. During PHEN, a phenylephrine dosage of 0.5 ± 0.2 μg/kg/min was needed to achieve a MAP of 133 ± 8 mm Hg, while ETISO 3.6 ± 0.2% lowered the MAP values to 43 ± 4 mm Hg during highISO treatment. The percent variation for PATDCO values between DOBUbaseline, ESMObaseline, PHENbaseline, and highISObaseline readings was 8.6 ± 3.1%.
Mean ± SD of standard data, intermittent pulmonary artery thermodilution cardiac output (PATDCO) measurements, and esophageal Doppler monitor (EDM)-derived indexes recorded in 6 healthy, mechanically ventilated, isoflurane-anesthetized Beagle dogs subjected to 4 randomized pharmacological interventions to acutely change the hemodynamics.
| Variable | DOBUbaseline | DOBU | ESMObaseline | ESMO | PHENbaseline | PHEN | highISObaseline | highISO |
|---|---|---|---|---|---|---|---|---|
| PATDCO (L/min) | 1.77 ± 0.35 | 2.58 ± 0.24* | 1.85 ± 0.11 | 1.06 ± 0.10† | 1.82 ± 0.23 | 0.92 ± 0.12‡ | 1.92 ± 0.16 | 0.73 ± 0.10§ |
| HR (beats/min) | 95 ± 12 | 107 ± 16 | 100 ± 9 | 95 ± 13 | 98 ± 11 | 66 ± 10‡ | 115 ± 6 | 93 ± 8§ |
| MAP (mm Hg) | 72 ± 9 | 87 ± 7* | 76 ± 8 | 64 ± 5† | 69 ± 6 | 133 ± 9‡ | 78 ± 4 | 43 ± 4§ |
| ETISO (%) | 1.6 ± 0.1 | 1.5 ± 0.1 | 1.5 ± 0.0 | 1.5 ± 0.1 | 1.6 ± 0.1 | 1.5 ± 0.1 | 1.5 ± 0.1 | 3.6 ± 0.2§ |
| EDM indexes | ||||||||
| StrokeD (cm) | 15.9 ± 0.4 | 19.4 ± 0.7* | 17.2 ± 0.8 | 11.5 ± 0.6† | 16.7 ± 0.9 | 9.3 ± 0.9‡ | 15.1 ± 0.3 | 12.9 ± 0.6§ |
| MinuteD (cm/min) | 1,526 ± 89 | 2,095 ± 101* | 1,737 ± 65 | 1,096 ± 99† | 1,636 ± 72 | 606 ± 125‡ | 1,751± 93 | 1,212 ± 82§ |
| FTc (ms) | 329 ± 12 | 372 ± 18* | 332 ± 9 | 351 ± 13 | 344 ± 10 | 258 ± 15‡ | 315 ± 11 | 385 ± 10§ |
| PV (cm/s) | 104 ± 11 | 124 ± 8* | 109 ± 6 | 81 ± 10† | 100 ± 9 | 77 ± 13‡ | 110 ± 7 | 85 ± 5§ |
| MA (m/s2) | 11.8 ± 2.1 | 14.4 ± 2.6* | 10.9 ± 3.8 | 8.2 ± 2.9† | 11.2 ± 2.1 | 8.8 ± 1.9‡ | 12.3 ± 3.3 | 9.4 ± 1.9§ |
Baseline data were collected before dobutamine (DOBUbaseline), esmolol (ESMObaseline), phenylephrine (PHENbaseline), and high-dose isoflurane (highISObaseline) treatments. Predetermined hemodynamic goals specific to each treatment were (1) DOBU: dobutamine IV infusion of 3 to 10 μg/kg/min aimed to increase PATDCO by > 40% as compared with DOBUbaseline; (2) ESMO: esmolol IV bolus of 100 μg/kg followed by infusion of 50 to 200 μg/kg/min aimed to decrease PATDCO by > 40% as compared with ESMObaseline; (3) PHEN: phenylephrine IV infusion of 0.2 to 1 μg/kg/min aimed for mean arterial pressure > 120 mm Hg; and (4) highISO: end-tidal isoflurane concentration > 3% aimed for mean arterial pressure < 50 mm Hg.
ETISO = End-tidal isoflurane concentration. FTc = Flow time corrected. HR = Heart rate. MA = Mean acceleration. MAP = Mean arterial pressure. MinuteD = Minute distance. PATDCO = Intermittent pulmonary artery thermodilution cardiac output. PV = Peak velocity. StrokeD = Stroke distance.
Comparisons between PATDCO, TEECO, and EDMCO measurements
For each dog, a pair of CO measurements was acquired using TEE, EDM, and PATD across 4 treatments (DOBU, ESMO, PHEN, and highISO) along with baseline data collected before each treatment. Thus, 48 paired observations for a total of 6 dogs were obtained. The mean ± SD relative bias evaluated with respect to PATDCO for TEECO (0.35 ± 25.2%) had a substantially smaller magnitude compared to the EDMCO (−27.2 ± 22.5%) which indicated an overall better performance by TEECO. The LOA were reported to be between −71.4% and 17% for EDMCO and between −49.1% and 49.8% for TEECO. When the relative bias was evaluated individually for each observation, less than 15% of observations were reported to be outside the 30% acceptable limit for TEECO. On the other hand, approximately 40% of observations were reported to be outside the 30% acceptable limit of relative bias for EDMCO. The percent error calculated for EDMCO was significantly higher at 44.1% versus 27.6% for TEECO.
Although r2 was greater than 0.90 for TEECO and EDMCO, the ρc value of 0.82 for TEECO and 0.66 for EDMCO indicated the difference in performance between the 2 methods. The residuals for the scatter plots were normally distributed for TEECO. The scatter plot for TEECO (Figure 4) did not indicate a consistent bias in estimation highlighting a good fit with PATDCO for a range of CO values. In addition, the slope about Y = X close to 1 indicated TEE's promising potential. The residuals for the scatter plots were normally distributed for EDMCO. The scatter plot for EDMCO (Figure 5) indicated a consistent positive bias highlighting an overestimation by EDMCO compared to PATDCO for a range of CO values. In addition, the slope about Y = X of 1.34 indicated that EDM showed a lower accuracy and estimation of PATDCO values.
A—Scatter plot representing TEECO and PATDCO in 6 healthy, anesthetized Beagle dogs across 4 treatments (dobutamine, esmolol, phenylephrine, and high-dose isoflurane) along with baseline data collected before each treatment, thus yielding 48 paired observations (circles). Regression analysis about the line Y = X (dashed line) resulted in a good fit (solid line) as shown by the slope 0.97 and r2 0.95. B—Bland-Altman analysis for nonuniform differences for TEECO as compared to PATDCO in 6 healthy, anesthetized Beagle dogs across 4 treatments (dobutamine, esmolol, phenylephrine, and high-dose isoflurane) along with baseline data collected before each treatment, thus yielding 48 paired observations (circles). Each circle represents an individual comparison of the difference with the mean and the central line represents the nonuniform mean bias of the difference. As displayed, the solid lines indicate the mean (green) and upper (blue) and lower (red) limits of agreement, and the dashed lines indicate the 95% confidence intervals around these values. C—Polar plot representing changes in TEECO as compared to PATDCO in 6 healthy, anesthetized Beagle dogs across 4 treatments (dobutamine, esmolol, phenylephrine, and high-dose isoflurane) along with baseline data collected before each treatment, thus yielding 48 paired observations (circles). Dotted lines represent 10% agreement boundaries (ie, 10% = 0.153 L/min as mean PATDCO = 1.53 L/min). The distance from the center represents the absolute values of the mean change in CO ([ΔPATDCO + ΔTEECO]/2), and the angle from the horizontal 0° radial axis represents the disagreement. The polar plot analysis revealed a good trending pattern for TEECO across the wide range of CO values as < 30% of the data points were located outside the limits of good agreement.
Citation: American Journal of Veterinary Research 84, 8; 10.2460/ajvr.23.05.0101
A—Scatter plot representing the CO measurements using EDMCO and PATDCO in 6 healthy, anesthetized Beagle dogs across 4 treatments (dobutamine, esmolol, phenylephrine, and high-dose isoflurane) along with baseline data collected before each treatment, thus yielding 48 paired observations (circles). Regression analysis about the line Y = X (dashed line) resulted in a good fit (solid line) as shown by the slope 1.34 and r2 0.97. B—Bland–Altman analysis for nonuniform differences for EDMCO as compared to PATDCO in 6 healthy, anesthetized Beagle dogs across 4 treatments (dobutamine, esmolol, phenylephrine, and high-dose isoflurane) along with baseline data collected before each treatment, thus yielding 48 paired observations (circles). Each circle represents an individual comparison of the difference with the mean and the central line represents the nonuniform mean bias of the difference. As displayed, the solid lines indicate the mean (green) and upper (blue) and lower (red) limits of agreement, and the dashed lines indicate the 95% confidence intervals around these values. C—Polar plot representing changes in EDMCO as compared to PATDCO in 6 healthy, anesthetized Beagle dogs across 4 treatments (dobutamine, esmolol, phenylephrine, and high-dose isoflurane) along with baseline data collected before each treatment, thus yielding 48 paired observations (circles). Dotted lines represent 10% agreement boundaries (ie, 10% = 0.153 L/min as mean PATDCO = 1.53 L/min). The distance from the center represents the absolute values of the mean change in CO ([ΔPATDCO + ΔEDMCO]/2), and the angle from the horizontal 0° radial axis represents the disagreement. The polar plot analysis revealed a good trending pattern for EDMCO across the wide range of CO values as < 30% of the data points were located outside the limits of good agreement.
Citation: American Journal of Veterinary Research 84, 8; 10.2460/ajvr.23.05.0101
The Bland-Altman analysis for TEECO (Figure 4) showed a good agreement and only a weak trend (slope = −0.04; intercept = 0.08; P < .001) between the bias and the average CO data. The results highlighted a consistent performance of TEE for a range of CO values as only a couple of observations were outside the LOA, and the magnitude of the bias was very small across a range of CO values. The Bland-Altman analysis for EDMCO (Figure 5) showed a strong negative trend (slope = −0.42; intercept = 0.23; P < .001) between the bias and the average CO data indicating a consistent overestimation across a broad range of CO values. The mean bias was also reported to be higher for higher values of CO which indicated that as the CO increased, the difference between EDMCO and PATDCO also increased. Although only a few observations were outside the LOA, the relatively high magnitude of bias across a range of CO values indicated a satisfactory agreement between EDM and PATD.
The polar plot analysis highlighted a good trending ability for both TEECO (Figure 4) and EDMCO (Figure 5) for a range of CO values as < 30% of the data points were located outside the limits of good agreement (ie, 10% mean CO = 0.153 L/min as mean PATDCO = 1.53 L/min).
Effect of treatments on EDM-derived indexes
There was no statistical difference observed in StrokeD, MinuteD, FTc, PV, and MA values between DOBUbaseline, ESMObaseline, PHENbaseline, and highISObaseline readings in all dogs. During DOBU, StrokeD, MinuteD, FTc, PV and MA values significantly increased as compared to DOBUbaseline (P < .001). During ESMO, StrokeD, MinuteD, PV, and MA values were significantly decreased (P < .001), whereas, FTc was unchanged as compared to ESMObaseline. Once the hemodynamic goal for PHEN and highISO was achieved, it was observed that StrokeD, MinuteD, PV, and MA significantly declined as opposed to PHENbaseline and highISObaseline measurements (P < .001). However, FTc was significantly lower during PHEN (P = .01) but higher (P = .02) during highISO.
Discussion
The present study observed a good agreement between TEECO and PATDCO, but EDMCO consistently overestimated PATDCO values. The value for ρc was 0.82 for TEECO and 0.66 for EDMCO, while the percent error for TEECO and EDMCO was ± 27.6% and ± 44.1%, respectively, indicating that EDM's error was not within the acceptable range (< 30%) published as a standard.27 However, from a clinician's standpoint, simply focusing on the absolute accuracy of a CO method may not always be the priority, rather, reliable trending patterns may also give crucial information,33 as seen in the present study.
Perioperative TEE is a valued diagnostic tool that provides the following: (1) reasons for hemodynamic instability and improves patient outcomes34,35; (2) rapid, real-time, continuous assessment of cardiac structures and their function; (3) information on anatomical or regional abnormalities, intracardiac neoplasia, thrombi formation, shunt direction, and magnitude; and (4) estimation of CO, maximal velocity in LVOT, left ventricular ejection fraction, LVOT VTI, and ventricular end-diastolic dimensions. Intracardiac air is also easily confirmed and managed by using this technique.36–38 Despite the credibility and reliability of TEE, it is very operator dependent, and for this reason, it is mandatory that a trained, skilled echocardiographer performs it, who has in-depth knowledge regarding the indications, potential complications, correct image acquisition, and echocardiographic data analysis.26 Moreover, a comprehensive understanding of ultrasound physics and TEE equipment controls is crucial for the optimization of the image quality, reduction in artifacts, and avoiding image misinterpretation. Reported TEE-associated complications in humans have an overall complication rate of 0.18% to 4% with < 0.02% mortality rate.25,39 In dogs, mucosal damage of the lower esophageal sphincter, displacement of the aorta and patent ductus arteriosus, and pin-point mucosal erosions at the heart base are reported, but the incidence is low and the severity of these complications is mild.25,40 Modern TEE probes possess a safety mechanism where the transducer stops functioning if the transducer tip temperature reaches 41 to 42 °C, thereby preventing thermal injury.25
In another study with healthy Beagles,16 after analyzing 30 paired measurements of PATDCO and TEECO, LOA and percent error were 0.03 ± 0.26 L/min and ± 12.3% for TEECO. The TEE provided better agreement with PATD versus transthoracic bioimpedance and partial carbon dioxide rebreathing during sevoflurane anesthesia in dogs. The comparison of transgastric view of LVOT-derived CO values with PATDCO in healthy, anesthetized dogs showed a strong positive correlation between TEECO and PATDCO (P < .0001), and the bias (LOA) was 0.14 ± 0.29 L/min (−0.44 to 0.72 L/min) with a percent error of ± 12.32%.17 This value was higher but within acceptable limits in our study (± 27.6%), and mean ± SD relative bias (LOA) was 0.35 ± 25.2% (−49.1% to 49.8%). Another study utilized longitudinal cranial-esophageal aorta long-axis-view to measure LVOTArea and calculate CO during postural changes in anesthetized dogs. The TEE successfully detected changes in preload, afterload, and HR.18 Evidence in mechanically ventilated canine patients suggests TEE also has benefit in estimating CO for further determining whether pulse pressure variation, aortic flow velocity-time integral variation and peak aortic velocity variation possess clinical value for evaluating fluid responsiveness.19,41
Analysis of EDMCO and PATDCO in the present study highlighted consistent overestimation, mean ± SD relative bias (LOA) of −27.2 ± 22.5% (−71.4% to 17%) but had an overall good trending pattern. Previously, in anesthetized healthy, normotensive dogs, EDMCO values were consistently lower than PATDCO values with a ρc value of 0.73 and percent error of ± 38.9% in a pool of 18 paired observations. The average difference (bias) between EDMCO and PATDCO was –0.94 ± 0.44 L/minute.9 Studies44,45 in human medicine have reported conflicting data regarding the relationship between EDMCO and PATDCO, hinting at good42,43 and poor agreement. When VTI and PV measurements from EDM human model were compared with transthoracic echocardiographic measurements in anesthetized dogs that were free of cardiac disease, EDM displayed clinically acceptable agreement.12
The esophagus is a conducive minimally invasive location where the ultrasound transducer uses piezoelectric crystals to alternately transmit and receive ultrasound signals close to the descending aorta. The accuracy of the descending aortic blood velocity measurement is majorly influenced by the alignment between the Doppler ultrasound beam and the direction of blood flow, known as the angle of insonation.46,47 If errors occur in measuring blood velocity by EDM, they could be focused on the alignment of the Doppler beam in EDM technology as compared to TEE. When the angle is wider between the Doppler beam and blood flow, the higher is the inaccuracy in blood velocity measurements due to an erroneous cosine in the Doppler equation, as well as the deviation of the actual angle from the assumed angle of insonation.46–48 Additionally, the velocity-time waveforms obtained by TEE and EDM are similar, where VTI (with TEE) corresponds to StrokeD (with EDM). The TEE offers clear visual confirmation of the Doppler beam paralleling with the structures of interest (ie, aorta) and providing very good alignment. In contrary with EDM, perfect alignment is assumed when a clear and sharp aortic waveform with the highest peak velocity that corresponds to a maximal pitch sound is procured. Moreover, due to the location of the tip of EDM probe, flow traces are not derived from the LVOT but rather from a portion of the descending aorta. Considering the LVOTArea and HR values were identical between the equations for TEECO and EDMCO, the source of error can be centered on the site of measurement of EDM traces and any deviation from a parallel alignment (cosine of 0° = 1) that will result in inaccurate estimation of real velocities. Additionally, the cost for TEE standard views and basic models is roughly 3 to 5 times the cost of EDM, which may be an important consideration for clinicians interested in advanced hemodynamic monitoring in the perioperative setting.
By displaying a velocity-time waveform (Figure 3) in real time, EDM's preferred indexes for monitoring preload are StrokeD, MinuteD, and FTc. In healthy and endotoxemic dogs, these 3 variables have shown promising results by closely tracking CO changes and guiding hemodynamic optimization.11,14 We attribute the increase in StrokeD, MinuteD, and FTc during DOBU to the β1-adrenergic receptor agonism resulting in augmentation of myocardial contractility, ventricular filling, and output. Considering PV and MA are contractility markers, if the ventricle is stimulated by an inotrope, the amplitude of the waveform and speed of blood on the upstroke will increase, similar to what was observed in our study dogs. Conversely, in a hypodynamic state (ESMO and highISO), the waveform will appear dampened with decreased amplitude yielding lower PV and MA values, and StrokeD and MinuteD may decline due to decreased left ventricular function. Caution should be exercised in using FTc solely to diagnose hypovolemia, as it also shares an inverse relationship with systemic vascular resistance. When afterload was boosted during PHEN, a narrow waveform with decreased amplitude was noted, which yielded lower FTc, PV, and MA values. Interestingly, during highISO, due to the volatile anesthetic-induced peripheral vasodilation especially with ETISO > 3%, FTc was higher, but PV and MA declined possibly due to the negative inotropy induced by inhalants. Clinicians must be cognizant of the changes in 1 variable that can accompany compensatory variations in other variables. Hence, no single variable derived by EDM can be relied on to gain information on preload, afterload, or contractility.
The present study has multiple limitations. The sample size is small, which causes the receiver operating characteristic curves to not generate. Hence, the predictive values, cutoffs, sensitivity, and specificity for the EDM-derived variables are not reported. The fewer time points for data collection restrict the number of paired observations for robust comparisons between test methods and PATD. Since the primary focus of this study was on hemodynamics, factors like body temperature, end-tidal carbon dioxide concentration, sympathetic stimulation, and anesthetic depth were controlled to isolate the impact of treatments on the hemodynamic variables. This study utilizes healthy Beagle dogs as its study population. Hence, evaluating TEECO and EDMCO measurements and EDM-derived indexes in anesthetized canine patients with systemic diseases is warranted.
Our study observed a good agreement between TEECO and PATDCO, and a consistent overestimation of PATDCO by EDMCO in healthy, anesthetized Beagle dogs undergoing acute pharmacologic manipulation of hemodynamics. The value for ρc was 0.82 for TEECO and 0.66 for EDMCO, and the percent error was ± 27.6% and ± 44.1% for TEECO and EDMCO, respectively. Both techniques displayed a good trending pattern in this CO comparative study. The EDM-derived indexes, ie, StrokeD, MinuteD, FTc, PV, and MA, correlated with changes in CO, contractility, and afterload induced by the treatments. These variables may have the potential in guiding therapeutic decisions on the use of inotropes, vasopressors, and fluid therapy during hemodynamic instability, and their further clinical investigation is imperative.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org.
Acknowledgments
We thank Jordi Mauragis for assistance with creating graphical illustrations, Andy Mears for providing expertise during the CardioQ-EDMV+ Esophageal Doppler monitor (Deltex Medical, UK) training, and Mindy Quigley for helping with the organization of the manuscript.
Disclosures
The authors have nothing to disclose. No artificial intelligence-assisted technologies were used in the generation of this manuscript.
Funding
This study was a part of a hemodynamic research project funded by Virginia-Maryland College of Veterinary Medicine, Veterinary Memorial Fund (grant No. 446641).
References
- 1.↑
Hoar PF, Stone JG, Faltas AN, Bendixen HH, Head RJ, Berkowitz BA. Hemodynamic and adrenergic responses to anesthesia and operation for myocardial revascularization. J Thorac Cardiovasc Surg. 1980;80(2):242–248. doi:10.1016/S0022-5223(19)37798-0
- 2.↑
Skelding A, Valverde A. Review of non-invasive blood pressure measurement in animals: part 2–evaluation of the performance of non-invasive devices. Can Vet J. 2020;61(5):481–498.
- 3.↑
Magder S. The meaning of blood pressure. Crit Care. 2018;22(1):257. doi:10.1186/s13054-018-2171-1
- 4.↑
Yamada T, Vacas S, Gricourt Y, Cannesson M. Improving perioperative outcomes through minimally invasive and non-invasive hemodynamic monitoring techniques. Front Med (Lausanne). 2018; 5:144. doi:10.3389/fmed.2018.00144
- 5.↑
Swan HJ, Ganz W, Forrester J, Marcus H, Diamond G, Chonette D. Catheterization of the heart in man with use of a flow-directed balloon-tipped catheter. N Engl J Med. 1970;283(9):447–451. doi:10.1056/NEJM197008272830902
- 6.↑
Marshall K, Thomovsky E, Johnson P, Brooks A. A review of available techniques for cardiac output monitoring. Top Companion Anim Med. 2016;31(3):100–108. doi:10.1053/j.tcam.2016.08.006
- 7.↑
Evans DC, Doraiswamy VA, Prosciak MP, et al. Complications associated with pulmonary artery catheters: a comprehensive clinical review. Scand J Surg. 2009;98(4):199–208. doi:10.1177/145749690909800402
- 8.↑
Kobe J, Mishra N, Arya VK, Al-Moustadi W, Nates W, Kumar B. Cardiac output monitoring: technology and choice. Ann Card Anaesth. 2019;22(1):6–17. doi:10.4103/aca.ACA_41_18
- 9.↑
Canfrán S, Cediel R, Sández I, Caro-Vadillo A, Gómez de Segura IA. Evaluation of an oesophageal Doppler device for monitoring cardiac output in anaesthetised healthy normotensive dogs. J Small Anim Pract. 2015;56(7):450–455. doi:10.1111/jsap.12362
- 10.
Sandez I, Soto M, Torralbo D, Rioja E. Effect of different analgesic techniques on hemodynamic variables recorded with an esophageal Doppler monitor during ovariohysterectomy in dogs. Can Vet J. 2018;59(4):419–424.
- 11.↑
Muehlestein MB, Steblaj B, Joerger FB, Briganti A, Kutter APN. Evaluation of the ability of haemodynamic variables obtained with minimally invasive techniques to assess fluid responsiveness in endotoxaemic Beagles. Vet Anaesth Analg. 2021;48(5):645–653. doi:10.1016/j.vaa.2021.02.008
- 12.↑
Sandez I, Verdier N, Redondo JI, et al. Agreement between transthoracic echocardiography and esophageal Doppler on aortic flow variables in anesthetized mechanically ventilated dogs. Can Vet J. 2022;63(7):722–726.
- 13.
Henze IS, Hilpert L, Kutter APN. Development and comparison of an esophageal Doppler monitoring-based treatment algorithm with a heart rate and blood pressure-based treatment algorithm for goal-directed fluid therapy in anesthetized dogs: a pilot study. Front Vet Sci. 2022;9:1008240. doi:10.3389/fvets.2022.1008240
- 14.↑
Paranjape VV, Henao-Guerrero N, Menciotti G, Saksena S. Esophageal Doppler-derived indices and arterial load variables provide useful hemodynamic information during assessment of fluid responsiveness in anesthetized dogs undergoing acute changes in blood volume. Am J Vet Res. 2023;84(3): ajvr.22.11.0198. doi:10.2460/ajvr.22.11.0198
- 15.↑
de Figueiredo LF, Cruz RJ Jr, Silva E, Silva M. Cardiac output determination during experimental hemorrhage and resuscitation using a transesophageal Doppler monitor. Artif Organs. 2004;28(4):338–342. doi:10.1111/j.1525-1594.2004. 47351.x
- 16.↑
Yamashita K, Ueyama Y, Miyoshi K, et al. Minimally invasive determination of cardiac output by transthoracic bioimpedance, partial carbon dioxide rebreathing, and transesophageal Doppler echocardiography in beagle dogs. J Vet Med Sci. 2007;69(1):43–47. doi:10.1292/jvms.69.43
- 17.↑
Mantovani MM, Fantoni DT, Gimenes AM, et al. Clinical monitoring of cardiac output assessed by transoesophageal echocardiography in anaesthetised dogs: a comparison with the thermodilution technique. BMC Vet Res. 2017;13(1):325. doi:10.1186/s12917-017-1227-9
- 18.↑
Goya S, Wada T, Shimada K, et al. Effects of postural change on transesophageal echocardiography views and parameters in healthy dogs. J Vet Med Sci. 2017;79(2):380–386. doi:10.1292/jvms.16-0323
- 19.↑
Gonçalves LA, Otsuki DA, Pereira MA, Nagashima JK, Ambrosio AM, Fantoni DT. Comparison of pulse pressure variation versus echocardiography-derived stroke volume variation for prediction of fluid responsiveness in mechanically ventilated anesthetized dogs. Vet Anaesth Analg. 2020;47(1):28–37. doi:10.1016/j.vaa.2019.08.047
- 20.↑
Paranjape VV, Henao-Guerrero N, Menciotti G, Saksena S. Volumetric evaluation of fluid responsiveness using a modified passive leg raise maneuver during experimental induction and correction of hypovolemia in anesthetized dogs. Vet Anaesth Analg. 2023;50(3):211–219. doi:10.1016/j.vaa.2023.02.009
- 21.↑
Paranjape VV, Henao-Guerrero N, Menciotti G, Saksena S, Agostinho M. Agreement between electrical cardiometry and pulmonary artery thermodilution for measuring cardiac output in isoflurane-anesthetized dogs. Animals. 2023;13(8):1420. doi:10.3390/ani13081420
- 22.
Paranjape VV, Shih AC, Garcia-Pereira FL. Use of a modified passive leg-raising maneuver to predict fluid responsiveness during experimental induction and correction of hypovolemia in healthy isoflurane-anesthetized pigs. Am J Vet Res. 2019;80(1):24–32. doi:10.2460/ajvr.80.1.24
- 23.↑
Paranjape VV, Shih AC, Garcia-Pereira FL, Saksena S. Transpulmonary ultrasound dilution is an acceptable technique for cardiac output measurement in anesthetized pigs. Am J Vet Res. 2022;83(6):ajvr.21.11.0189. doi:10.2460/ajvr.21.11.0189
- 24.↑
Puchalski MD, Lui GK, Miller-Hance WC, et al. Guidelines for performing a comprehensive transesophageal echocardiographic: examination in children and all patients with congenital heart disease: recommendations from the American Society of Echocardiography. J Am Soc Echocardiogr. 2019;32(2):173–215. doi:10.1016/j.echo.2018.08.016
- 25.↑
Domenech O, Oliveira P. Transoesophageal echocardiography in the dog. Vet J. 2013;198(2):329–338. doi:10.1016/j.tvjl.2013.08.014
- 26.↑
Porter TR, Shillcutt SK, Adams MS, et al. Guidelines for the use of echocardiography as a monitor for therapeutic intervention in adults: a report from the American Society of Echocardiography. J Am Soc Echocardiogr. 2015;28(1):40–56. doi:10.1016/j.echo.2014.09.009
- 27.↑
Critchley LA, Critchley JA. A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. J Clin Monit Comput. 1999;15(2):85–91. doi:10.1023/a:1009982611386
- 28.↑
Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics. 1989;45(1):255–268. doi:10.2307/2532051
- 29.↑
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307–310. doi:10.1016/S0140-6736(86)90837-8
- 30.↑
Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res. 1999;8(2):135–160. doi:10.1177/096228029900800204
- 31.↑
Critchley LA, Lee A, Ho AM. A critical review of the ability of continuous cardiac output monitors to measure trends in cardiac output. Anesth Analg. 2010;111(5):1180–1192. doi:10.1213/ANE.0b013e3181f08a5b
- 32.↑
Critchley LA, Yang XX, Lee A. Assessment of trending ability of cardiac output monitors by polar plot methodology. J Cardiothorac Vasc Anesth. 2011;25(3):536–546. doi:10.1053/j.jvca.2011.01.003
- 33.↑
Peyton PJ, Chong SW. Minimally invasive measurement of cardiac output during surgery and critical care: a meta-analysis of accuracy and precision [Erratum in Anesthesiology 2012;116(4):973]. Anesthesiology 2010;113(5):1220–1235. doi:10.1097/ALN.0b013e3181ee3130
- 34.↑
Ramachandran G, Sundar AS, Venugopal V, Shah HD, Dogra N. Recent advances in cardiac anaesthesia. Indian J Anaesth 2023;67(1):78–84. doi:10.4103/ija.ija_972_22
- 35.↑
Efrimescu CI, Moorthy A, Griffin M. Rescue transesophageal echocardiography: a narrative review of current knowledge and practice. J Cardiothorac Vasc Anesth 2023;37(4):584–600. doi:10.1053/j.jvca.2022.12.031
- 36.↑
Gouveia V, Marcelino P, Reuter DA. The role of transesophageal echocardiography in the intraoperative period. Curr Cardiol Rev 2011;7(3):184–196. doi:10.2174/157340311798220511
- 37.
Smith WB, Robinson AR, Janelle GM. Expanding role of perioperative transesophageal echocardiography in the general anesthesia practice and residency training in the USA. Curr Opin Anaesthesiol 2015;28(1):95–100. doi:10.1097/ACO.0000000000000146
- 38.↑
Akiyama K, Arisawa S, Ide M, Iwaya M, Naito Y. Intraoperative cardiac assessment with transesophageal echocardiography for decision-making in cardiac anesthesia. Gen Thorac Cardiovasc Surg 2013;61(6):320–329. doi:10.1007/s11748-013-0208-6
- 39.↑
Hilberath JN, Oakes DA, Shernan SK, Bulwer BE, D'Ambra MN, Eltzschig HK. Safety of transesophageal echocardiography. J Am Soc Echocardiogr 2010;23(11):1115–1221. doi:10.1016/j.echo.2010.08.013
- 40.↑
Stoner CH, Saunders AB, Heseltine JC, Cook AK, Lidbury JA. Prospective evaluation of complications associated with transesophageal echocardiography in dogs with congenital heart disease. J Vet Intern Med 2022;36(2):406–416. doi:10.1111/jvim.16356
- 41.↑
Fantoni DT, Ida KK, Gimenes AM, et al. Pulse pressure variation as a guide for volume expansion in dogs undergoing orthopedic surgery. Vet Anaesth Analg 2017;44(4):710–718. doi:10.1016/j.vaa.2016.11.011
- 42.↑
Chew HC, Devanand A, Phua GC, Loo CM. Oesophageal Doppler ultrasound in the assessment of haemodynamic status of patients admitted to the medical intensive care unit with septic shock. Ann Acad Med Singap 2009;38(8):699–703.
- 43.↑
Schubert S, Schmitz T, Weiss M, et al. Continuous, non-invasive techniques to determine cardiac output in children after cardiac surgery: evaluation of transesophageal Doppler and electric velocimetry. J Clin Monit Comput 2008;22(4):299–307. doi:10.1007/s10877-008-9133-0
- 44.↑
Kim K, Kwok I, Chang H, Han T. Comparison of cardiac outputs of major burn patients undergoing extensive early escharectomy: esophageal Doppler monitor versus thermodilution pulmonary artery catheter. J Trauma 2004;57(5):1013–1017. doi:10.1097/01.ta.0000105925.51950.08
- 45.↑
Phan TD, Kluger R, Wan C, Wong D, Padayachee A. A comparison of three minimally invasive cardiac output devices with thermodilution in elective cardiac surgery. Anaesth Intensive Care 2011;39(6):1014–1021. doi:10.1177/0310057X1103900606
- 46.↑
Schober P, Loer SA, Schwarte LA. Perioperative hemodynamic monitoring with transesophageal Doppler technology. Anesth Analg 2009;109(2):340–353. doi:10.1213/ane.0b013e3181aa0af3
- 47.↑
Singer M. Oesophageal Doppler. Curr Opin Crit Care 2009;15(3):244–248. doi:10.1097/MCC.0b013e32832b7083
- 48.↑
Cholley BP, Singer M. Esophageal Doppler: noninvasive cardiac output monitor. Echocardiography 2003;20(8):763–769. doi:10.1111/j.0742-2822.2003.03033.x
