Comparison of cardiac output determined by an ultrasound velocity dilution cardiac output method and by the lithium dilution cardiac output method in juvenile horses with experimentally induced hypovolemia Andre C. Shih, DVM; Patricia Queiroz, DVM, MS; Alessio Vigani, DVM, PhD; Anderson Da Cunha, DVM, MS; Romain Pariaut, DVM; Carolina Ricco, DVM, MS; Jennifer Bornkamp, DVM; Fernando Garcia-Pereira, DVM, MS; Carsten Bandt, DVM
Objective—To assess the accuracy of an ultrasound velocity dilution cardiac output (UDCO) method, compared with that of the lithium dilution cardiac output (LiDCO) method, for determination of cardiac output (CO) in juvenile horses with experimentally induced hypovolemia. Animals—12 anesthetized 2- to 6-month-old horses. Procedures—For each anesthetized horse, CO was determined by the LiDCO and UDCO methods prior to any intervention (baseline state), after withdrawal of approximately 40% of the horse’s blood volume (low CO state), after maintenance of hypovolemia and infusion of norepinephrine until mean arterial blood pressure was equal to baseline value (high CO state), and after further infusion of norepinephrine and back-transfusion of withdrawn blood (posttransfusion state). For each of the 4 hemodynamic situations, CO and calculated cardiac index (CI) values were obtained by each method in duplicate (8 pairs of measurements/horse); mean values for each horse and overall mean values across all horses were calculated. Agreement between CI determined by each method (96 paired values) was assessed by Bland-Altman analysis. Results—For the UDCO method–derived CI measurements among the 12 horses, mean ± SD bias was –4 ± 11.3 mL/kg/min (95% limits of agreement, –26.1 to 18.2 mL/kg/min) and mean relative bias was –10.4 ± 21.5% (95% limits of agreement, –52.6% to 31.8%). Conclusions and Clinical Relevance—Results indicated that, compared with the LiDCO method, the UDCO method has acceptable clinical usefulness for determination of CO in foals. (Am J Vet Res 2014;75:565–571)
C
ardiac output can be defined as the volume of blood moved by the heart per unit of time.1,2 Measurement of CO also allows for calculation of oxygen delivery, oxygen consumption, and systemic vascular re-
Received September 4, 2013. Accepted December 19, 2013. From the Departments of Large Animal Clinical Sciences (Shih, Bornkamp, Garcia-Pereira) and Small Animal Clinical Sciences (Vigani, Bandt), College of Veterinary Medicine, University of Florida, Gainesville, FL 32606; the Department of Veterinary Clinical Sciences, School of Veterinary Medicine, Louisiana State University, Baton Rouge, LA 70803 (Queiroz, Da Cunha, Pariaut); and the Department of Small Animal Clinical Sciences, Virginia Maryland Regional College of Veterinary Medicine, Virginia Tech, Blacksburg, VA 24601 (Ricco). Supported in part by an Intramural Research Grant from the School of Veterinary Medicine, Louisiana State University. Presented at the 2012 American College of Veterinary Anesthesia Symposium, Nashville, Tenn, September 2012. The authors thank SanJay Jadhao and John Ladner for technical assistance. Address correspondence to Dr. Shih (
[email protected]). AJVR, Vol 75, No. 6, June 2014
13-09-0268r.indd 565
CI CO LiDCO MAP UDCO TDCO
ABBREVIATIONS
Cardiac index Cardiac output Lithium dilution cardiac output Mean arterial blood pressure Ultrasound velocity dilution cardiac output Thermodilution cardiac output
sistance (Appendix).3 Monitoring of CO can be used to guide treatment, improve clinical outcome in critically ill patients, and monitor patients during anesthesia and has prognostic value.4 Changes in CO are not detected by means of routine hemodynamic monitored variables (eg, arterial blood pressure, central venous pressure, heart rate, and urine output). Furthermore, studies4,5 have revealed that those hemodynamic monitored variables were the worst predictors of survival. On the other hand, hemodynamic variables such as CI, systemic vascular resistance, and oxygen consumption have 565
5/21/2014 10:33:06 AM
good specificity and sensitivity with regard to predicting survival and aid in early goal-directed treatment. 4 Cardiac output is routinely measured in humans undergoing anesthesia in critical care units. Horses have intrinsic characteristics (body size, temperament, and unique anatomic features) that make routine use of CO in this species a challenge. Cardiac output measurement in veterinary patients is not routinely performed because of the invasiveness of currently validated procedures.3,6 The clinical gold standard of CO measurement is the TDCO method, which requires the use of a pulmonary artery catheter. Use of pulmonary artery catheters is associated with major risks including development of arrhythmias, cardiac endothelial damage, infection, knotting of the catheter in the pulmonary artery, pulmonary thromboembolism, and pulmonary artery rupture.2,3 Consequently, there has been an interest in less invasive yet accurate methods of CO measurement including the LiDCO method, esophageal Doppler ultrasonography, thoracic electrical bioimpedance assessment, and arterial pulse analysis.7–9 The values of CO provided by the LiDCO technique have excellent agreement with values of CO provided by the TDCO technique in horses, dogs, and cats.2,10,11 The LiDCO method is widely applied in veterinary medicine and is the most common technique used to study CO in anesthetized foals.9–11 The UDCO method is a novel technique that eliminates some of the drawbacks of the TDCO and LiDCO methods. The UDCO method is minimally invasive, requiring placement of only peripheral venous and arterial catheters; also, it uses a physiologic noncumulative indicator, saline (0.9% NaCl) solution.12,13 Although numerous repeated LiDCO measurements can result in considerable blood loss, there is minimal blood loss associated with numerous repeated UDCO measurements. The UDCO measurements are performed by injecting a bolus of isotonic saline solution (0.5 to 2 mL/kg) into the venous circulation, which results in a transient hemodilution and change in the velocity of the blood detected by the ultrasound sensors. To evaluate a new CO monitoring technique, a comparative study should be conducted in the species of interest during various clinical scenarios and over a wide physiologic range of CO of interest to demonstrate that the measurements obtained are in agreement with those obtained with validated techniques. Use of the LiDCO method has been validated and compared with the TDCO method in neonatal foals and other species.14,15 The UDCO technique was recently evaluated in anesthetized euvolemic neonatal foals (1 to 3 days of age).6 There is, however, significant variation in CO values among horses of differing ages. Foals have high heart rates and CI values because of their high metabolic rates and immature sympathetic nervous system.16 Mean ± SD CI in 2-hour-old foals is 155.2 ± 8.1 mL/ kg/min, and in 14-day-old foals, the value is 222 ± 21.6 mL/kg/min; in adult horses, CI is typically 72 mL/kg/ min.16–18 The purpose of the study reported here was to assess the accuracy of a UDCO method, compared with that of the LiDCO method, for determination of CO (and CI calculated on the basis of the CO values) in juvenile horses with experimentally induced hypo566
13-09-0268r.indd 566
volemia. Our hypothesis was that values of CI derived by use of the UDCO technique would have good agreement with values determined by use of the LiDCO technique in juvenile horses with hypovolemia induced via controlled hemorrhage. Materials and Methods Animals—Data were obtained from 12 juvenile horses (2 to 6 months of age). Body weights ranged between 70 and 189 kg (mean ± SD, 103.4 ± 45.3 kg). Horses were determined to be healthy on the basis of findings of a physical examination and CBC; no biochemical analyses were performed. The study was approved by an institutional animal care and use committee. Instrumentation of horses and study design—For each horse, anesthesia was induced with xylazine hydrochloride (0.5 mg/kg, IV) and ketamine hydrochloride (2.2 mg/kg, IV). Following orotracheal intubation, anesthesia was maintained with isofluranea and 100% FIO2 carrier gas. The mean ± SD end-tidal concentration of isoflurane was kept constant at 1.8 ± 0.1% for the duration of each experiment. Mechanical ventilation was set at 8 to 10 breaths/min and a tidal volume of 10 to 15 mL/kg; the ventilator was not turned off during CO measurements. A 7F, 30-cm-long, doublelumen catheterb was placed in the right jugular vein, and the tip was advanced to the level of the right atrium. A 10-gauge, 15-cm long catheter was placed in the left jugular vein for blood sample withdrawal. Right atrium catheter position was determined by waveform tracing. A venous blood pressure monitor transducer was placed and zeroed at the level of the right atrium (point of the horse’s shoulder). A 20-gauge, 1.88-inch catheter was placed in the right metatarsal artery.c The blood pressure transducer was calibrated with a mercury manometer, and end-tidal gas concentration was calibrated with a manufactured known-concentration gas cylinder. d LiDCO method measurement of CO—Hemoglobin and sodium concentrations required for CO measurement were determined before each data set with a bench-top blood gas analyzer.e Lithium dilution CO was determined by injecting a bolus of lithium chloridef (0.003 mmol/kg) at the time that blood was being withdrawn at a constant rate (4 L/min) across a lithium sensor.g Total blood withdrawn during each LiDCO measurement was approximately 20 mL. Cardiac output was calculated on the basis of the lithium dose administered and the area under the concentration-time curve prior to recirculation.g UDCO method measurement of CO—For each horse, an arteriovenous extracorporeal arteriovenous loop was made by attaching loop tubing between the metatarsal arterial and one of the central venous catheters.12,h Two ultrasound velocity sensorsi were attached to this circuit; the venous flow sensor was placed upstream from the venous catheter, and the arterial flow sensor was placed downstream from the arterial catheter.13 The same extracorporeal arteriovenous loop was primed with saline solution containing heparin and connected between the metatarsal arterial catheter AJVR, Vol 75, No. 6, June 2014
5/21/2014 10:33:06 AM
and the distal portion of the lumen of one of the central venous catheters. During UDCO measurements, blood was continuously circulating through the loop by use of the manufacturer’s roller pump.j After UDCO measurement, the loop was flushed with saline solution containing heparin and the pump was stopped. To obtain measurements, a roller pump circulated blood through the arteriovenous loop from the artery to the vein at a rate of 12 mL/min for 6 minutes. During this process, isotonic saline (0.9% NaCl) solution (0.5 mL/ kg; prewarmed to body temperature in a bag warmerj [constant 37°C]) was injected into the venous side of the arteriovenous loop. The arterial sensor measured the change in concentration of saline solution in the circulating blood (dilution), and the UDCO determination was the difference between ultrasound velocities in blood and saline solution. Ultrasound velocity in blood is 1,560 to 1,585 m/s and is mainly determined by the presence of blood cells and plasma protein.19 Ultrasound velocity in warm saline solution is lower than that in blood (1,533 m/s). Injection of a saline solution bolus (0.5 mL/kg) generates transient hemodilution and consequently decreases the ultrasound velocity, as measured by the arterial flow probe. Saline solution was injected through the right atrium catheter just proximal to the venous flow sensor. Cardiac output was calculated with the monitor computer from the ultrasonic speed dilution curve and was based on a derivation of the Stewart-Hamilton principle: CO = ([UVblood – UVsaline] X Volinj)/f UVa(t)dt
where UVblood – UVsaline is the difference between the ultrasound velocities in blood and saline solution, Volinj is the volume of isotonic saline solution injected (mL), and f UVa(t)dt is the function of the change in ultrasound velocity in arterial blood as measured by the arterial sensor over time. Cardiac output is the product of the volume of isotonic saline solution injected and the decrease in ultrasound velocity as the saline solution passes through the ultrasonic sensor.12–23,i,k Experimental design—After instrumentation of each horse, values of CO were determined by use of the LiDCO and UDCO methods (baseline state). Measurements were obtained for 3 additional hemodynamic situations (designated as low CO state, high CO state, and posttransfusion state). For each of the 4 hemodynamic situations, the horse was maintained at a stable hemodynamic plane for a minimum of 10 minutes prior to data collection. Paired LiDCO and UDCO measurements were obtained in duplicate for each hemodynamic situation (8 paired measurements/horse); a value of CI was calculated on the basis of each CO measurement (8 paired measurements/horse). Cardiac index was obtained by dividing CO by body weight. A mean of the duplicate measurements of CO and a mean of the duplicate determinations of CI were calculated for each horse. For data analysis, an overall mean CO and CI derived by each method across all 12 horses were calculated for each hemodynamic situation. All 96 pairs of CI values were used in the Bland-Altman analyses. AJVR, Vol 75, No. 6, June 2014
13-09-0268r.indd 567
For each horse, experimental manipulations were undertaken to generate the hemodynamic situations other than baseline state. A hypovolemic, hypotensive, low CO state was achieved by withdrawing blood until the MAP was stable at 40 mm Hg for a period of 10 minutes. The amount of blood withdrawn was equivalent to a mean volume of 30.7 ± 8.2 mL/kg or approximately 40% of blood volume. Total blood volume was calculated as 90 mL/kg.24 Withdrawn blood was collected from the central venous catheter and stored in citrate-phosphate-dextrose-adenine solutionl at 37°C in an incubator. A hypovolemic, normotensive, high CO state was then achieved by maintaining hypovolemia and titrating a norepinephrine infusion until MAP was equal to the baseline value. Next, a euvolemic, normotensive, high CO state was achieved by continuing the norepinephrine infusion and transfusing all of the blood volume withdrawn to induce hypotension back to the horse; this was the posttransfusion state. The order in which the CO states were induced or methods of CO assessment were undertaken was not randomized. For each horse, measurements were obtained for the baseline state, low CO state, high CO state, and finally posttransfusion state. On each occasion that paired CO measurements were obtained, the LiDCO measurement was performed first, ensuring that the saline solution used in the UDCO method would not interfere with determining the LiDCO measurement. Blood withdrawn for LiDCO measurement was not returned to the horse. At the end of the experiment, each horse recovered from anesthesia and experimental manipulations without complications. Heart rate, systolic arterial blood pressure, MAP, diastolic arterial blood pressure, oxygen saturation (as measured by pulse oximetry), and end-tidal concentrations of carbon dioxide and isoflurane were monitoredd continuously throughout each phase of the experiment. However, for purposes of analysis, a value of each variable was obtained at the time of CO measurement by either method during each hemodynamic situation. A mean of the duplicate measurements of each variable was calculated for each horse in each hemodynamic situation. For data analysis, an overall mean of each variable across all 12 horses was calculated for each hemodynamic situation. Data analysis—Normality of the data and equality of variances were assessed with the Kolmogorov-Smirnov and Levene tests, respectively. Agreement between CI values derived by use of the LiDCO and UDCO methods was determined by means of the Bland and Altman method.25,26 For each of the 96 pairs of values, bias was calculated as follows: (CILiDCO – CIUDCO), where CILiDCO and CIUDCO are the paired CI values obtained at a given time measured by LiDCO and UDCO, respectively. For each of the 96 pairs of values, bias was also expressed as a percentage of the mean CI (relative bias). Relative bias was calculated by use of the following equation27: (CILiDCO – CIUDCO)/(0.5 X [CILiDCO + CIUDCO]) X 100
The difference between the mean values determined by the 2 methods was plotted against the mean 567
5/21/2014 10:33:06 AM
of the mean values determined by the 2 methods. A general linear regression model with repeated observations was applied to assess the effect of the method of CO measurement and CO state (baseline, low CO, high CO, and posttransfusion) on CI bias. Positive bias or relative bias reflected underestimation of LiDCO method–derived CO by the UDCO method, whereas negative bias or relative bias reflected overestimation of the LiDCO method–derived CO by the UDCO method. The limits of agreement for bias or relative bias were reported as ± (1.96 X SD). A regression analysis was performed to compare body weight and bias or relative bias. The level of CI was compared with a 1-way ANOVA on ranks. The concordance correlation coefficient (ρc) integrates a measurement of precision and accuracy as follows: ρc = ρ X Cb, where ρ is the Pearson correlation coefficient that measures how far each observation deviates from the line of best fit (ie, precision) and Cb is a bias correction factor that measures how far the line of best fit deviates from the 45° line through the origin (ie, accuracy).27,28 A ρc of 1 indicates perfect concordance. For all analyses, values of P < 0.05 were considered significant. Results For each horse, a pair of CO measurements was obtained by use of the LiDCO and UDCO methods in duplicate during each of the 4 hemodynamic situ-
ations (ie, baseline state, low CO state, high CO state, and posttransfusion state). Eight pairs of measurements were obtained for each horse; thus, overall 96 pairs of measurements were obtained from all 12 horses. Each CO value was subsequently used to calculate a CI value. After completion of the experiments, all horses recovered uneventfully and no long-term adverse events or complications were attributed to the study procedures. Mean ± SD anesthesia time was 160 ± 42 minutes, and duration of controlled hemorrhage (ie, duration of the hypovolemic and high CO states) was 59 ± 12 minutes. Compared with findings at the time of baseline measurements, there were no significant changes in oxygen saturation, heart rate, or end-tidal concentrations of carbon dioxide and isoflurane during the other 3 hemodynamic situations (Table 1). Systolic arterial blood pressure and MAP were decreased significantly (P = 0.002) in the low CO state, compared with findings in the baseline state. In the high CO state, MAP increased significantly (P = 0.004), compared with findings in the low CO state. Values of CI derived by use of the LiDCO method ranged between 8 and 152 mL/kg/min. Mean ± SD CIs derived by use of the LiDCO method in the baseline, low CO, high CO, and posttransfusion states were 69.0 ± 23.3 mL/kg/min, 23.2 ± 11.5 mL/kg/min, 40.1 ± 14.5 mL/kg/min, and 87.7 ± 28.8 mL/kg/min, respectively (Table 2). Values of CI derived by use of UDCO ranged between 15 and 182 mL/kg/min.
Table 1—Mean ± SD heart rate, systolic arterial blood pressure (SAP), diastolic arterial blood pressure (DAP), MAP, end-tidal concentrations of isoflurane and carbon dioxide (ET-Iso and ET-CO2, respectively), and oxygen saturation (as measured by pulse oximetry) at the times that CO measurements were obtained by use of the LiDCO and UDCO methods in 12 anesthetized juvenile horses during 4 hemodynamic situations. Hemodynamic situation Variables Heart rate (beats/min) SAP (mm Hg) DAP (mm Hg) MAP (mm Hg) ET-Iso (mm Hg) ET-CO2 (mm Hg) Oxygen saturation (%)
Baseline state
Low CO state
High CO state
Posttransfusion state
53 ± 8 102 ± 10 58 ± 8 74 ± 13 1.8 ± 0.1 41 ± 3 99 ± 1
53 ± 13 71 ± 8* 28 ± 3* 37 ± 5* 1.8 ± 0.3 39 ± 11 98 ± 1
54 ± 14 98 ± 17 64 ± 12 73 ± 8† 1.8 ± 0.1 40 ± 13 99 ± 1
55 ± 7 110 ± 18 59 ± 9 79 ± 10 1.8 ± 0.2 41 ± 5 98 ± 1
For each anesthetized horse, CO was determined by the LiDCO and UDCO methods prior to any intervention (baseline state); after withdrawal of approximately 40% of the horse’s blood volume (low CO state); after maintenance of hypovolemia and infusion of norepinephrine until MAP was equal to baseline value (high CO state); and after further infusion of norepinephrine and back-transfusion of withdrawn blood (posttransfusion state). For each of the 4 hemodynamic situations, CO values were obtained by each method in duplicate (8 pairs of measurements/horse). *Within a variable, value differs significantly (P < 0.05) from baseline state value. †Within a variable, value differs significantly (P < 0.05) from the low CO state value. Table 2—Mean ± SD CO (L/min) and calculated CI (mL/kg/min) derived by use of the LiDCO and UDCO methods in 12 anesthetized juvenile horses during 4 hemodynamic situations. Hemodynamic situation Variable CO CO CI CI
Method
Baseline state
Low CO state
High CO state
Posttransfusion state
LiDCO UDCO LiDCO UDCO
5.4 ± 1.3 5.6 ± 1.2 69 ± 23.3 72.3 ± 23
1.8 ± 0.8 * 2.3 ± 0.9* 23.2 ± 10.2* 28.6 ± 11.5*
3.1 ± 0.8 3.5 ± 1.3 40.1 ± 14.5 44.0 ± 17.6
7.1 ± 2.3* 7.4 ± 2.3*† 87.7 ± 28.8* 91 ± 31.3*†
For each of the 4 hemodynamic situations, CO and calculated (CI) values were obtained by each method in duplicate (8 pairs of measurements/horse); overall, 96 measurements obtained by each method were available for analysis. Cardiac index was obtained by dividing CO by body weight. See Table 1 for key. 568
13-09-0268r.indd 568
AJVR, Vol 75, No. 6, June 2014
5/21/2014 10:33:06 AM
Mean CIs derived by use of the UDCO method in the baseline, low CO, high CO, and posttransfusion states were 72.3 ± 23 mL/kg/min, 28.6 ± 11.3 mL/ kg/min, 44 ± 17.3 mL/kg/min, and 91 ± 31.3 mL/kg/ min, respectively. This yielded a mean difference be-
tween the 2 methods of –4 mL/kg/min. The mean total blood volume withdrawn during the low CO state was 30.7 ± 8.2 mL/kg; the duration of the controlled hemorrhage procedure during the low CO state was approximately ≤ 15 minutes. With regard to CI, Bland-Altman analysis revealed no significant effect of hemodynamic situation on bias or relative bias (P = 0.8 and P = 0.7, respectively). The mean ± SD bias of CI derived by use of the UDCO method was –4 ± 11.3 mL/kg/min (limits of agreement, –26.1 to 18.2 mL/kg/min; Figure 1). The mean ± SD relative bias of CI derived by use of the UDCO method was –10.4 ± 21.5% (limits of agreement, –52.6% to 31.8%). Fourteen of the 96 (14.6%) observations had an absolute value of relative bias > 30%. Concordance correlation coefficient (ρc) between CI values derived by use of the LiDCO and UDCO methods was 0.89, and r2 was 0.8 (P < 0.001). No linear correlation was observed between body weight and bias or relative bias. Discussion
Figure 1—Bland-Altman analysis plot of bias (A) and relative bias (B) for CI values derived from paired CO measurements obtained by use of the LiDCO and UDCO methods in 12 anesthetized juvenile horses. For each anesthetized horse, CO was determined by the LiDCO and UDCO methods prior to any intervention (baseline state [white circles]), after withdrawal of approximately 40% of the horse’s blood volume (low CO state [black circles]), after maintenance of hypovolemia and infusion of norepinephrine until MAP was equal to baseline value (high CO state [white triangles]), and after further infusion of norepinephrine and back-transfusion of withdrawn blood (posttransfusion state [black triangles]). For each of the 4 hemodynamic situations, CO and calculated CI values were obtained by each method in duplicate (8 pairs of measurements/horse); overall, 96 measurements obtained by each method were available for analysis. Cardiac index was obtained by dividing CO by body weight. Bias was calculated as follows: CILiDCO – CIUDCO, where CILiDCO and CIUDCO are a pair of CI values measured by the LiDCO and UDCO methods, respectively. Relative bias (ie, bias expressed as a percentage of the mean CI) was calculated by use of the following equation: (CILiDCO – CIUDCO)/(0.5 X [CILiDCO + CIUDCO]) X 100. Mean bias and mean relative bias (solid horizontal line) and upper and lower 95% limits of agreement (defined as the mean ± 1.96 X SD [dashed lines]) are indicated. AJVR, Vol 75, No. 6, June 2014
13-09-0268r.indd 569
Hypovolemic shock is a common type of shock in humans and other animals and is characterized by a decrease in CO and an increase in systemic vascular resistance. Horses can develop hypovolemic shock as a result of hemorrhage, loss of circulating volume into the gastrointestinal tract (third space effect), or leakage of vascular fluid into the interstitium and abdominal space as a result of capillary damage. Hypovolemia remains a major cause of morbidity and death in juvenile horses, which may be a consequence of clinicians’ inability to evaluate the hemodynamic state of patients. Cardiac output is influenced by the volemic status, cardiac contractility, and heart rate and by the resistance of the vascular bed (ie, systemic vascular resistance) against which the heart is pumping. For a CO monitor to be useful, it has to provide accurate measurements over a wide range of clinically important scenarios including normovolemia, hypovolemia, and hyperdynamic conditions. Results of the present study indicated that the UDCO method is a feasible alternative to the LIDCO method for CO measurement in juvenile horses undergoing severe hypovolemic shock. In the present study, we were able to evaluate the bias, precision, and accuracy of a novel method of CO measurement (UDCO) against an accepted and clinically relevant method (LiDCO) over a range of physiologic values in juvenile horses. The experimental procedures 569
5/21/2014 10:33:07 AM
were undertaken to manipulate total blood volume as well as cardiac contractility and systemic vascular resistance of the horses. The data obtained indicated satisfactory agreement between the CO values obtained by the LiDCO and UDCO methods. This agreement remained acceptable after acute hemorrhage even when systemic vascular resistance was pharmacologically altered with norepinephrine administration. Other authors have reported that CO determined by the UDCO method compared favorably with CO determined by pulmonary thermodilution and aortic flow probe methods in mice, rats, and humans.12,13 Bias and precision estimates are derived from comparison of the difference between mean values determined by 2 measurement techniques against the mean for each paired reading.27,28 Cardiac output measurement techniques have a variable degree of inherent error ranging from 10% to 20% (plus or minus) as a result of cyclic changes caused by respiration. The current recommendations for accepting a new technique are based on limits of agreement of –30% to 30% between the new and the reference technique.27,28 In the present study, 80.5% of the 96 UDCO method–derived CO measurements was within limits of agreement of –30% to 30%, suggesting the potential applicability of this new method, as compared with the LiDCO method. A negative bias identified in the present study indicated that the use of the UDCO method tended to slightly overestimate CO, compared with values determined by the LiDCO method. The advantages of the UDCO method for assessment of CO include the use of a safe and inexpensive indicator (isotonic saline solution) and absence of blood loss. The recommended bolus of isotonic saline solution is 0.5 to 2 mL/kg. At conservative doses, repeated bolus administration is unlikely to cause fluid overload or any long-term adverse effects. The UDCO method requires the bolus injection to be completed in 2 to 5 seconds. This may be a limitation of its use in larger animals (eg, adult horses) as the volume of saline solution increases. The placement of short largebore central catheters is essential to ensure smooth and rapid administration of the bolus injection. In addition, handling of the saline solution must be minimized because changes in temperature and presence of microbubbles can influence the UDCO readings.6 A disadvantage of the UDCO method is potential for platelet damage and microthrombi formation from blood flow through the extracorporeal loop attached to a peristaltic pump. For animals with severe hemorrhagic shock, this can be a major risk because the patient is hypoperfused and in an activated hypercoagulable state. In the present study, we did not detect formation of clots at the catheter sites or in the extracorporeal tubing used in the study horses. Further studies are needed to determine whether there is an increase in morbidity rate associated with this technique in clinically compromised patients. The limitations of the present study included a small sample size (12 horses). All comparisons were made during 1 anesthetic event, and the order of the hemodynamic states was not randomized, thereby increasing the risk of temporal bias. We do believe that the total 570
13-09-0268r.indd 570
blood volume increased with time as blood was shifted from the interstitial space toward the central compartment and with possible splenic contraction. This was reflected by the fact that mean CO and CI were greatest in the posttransfusion state, compared with findings in the other 3 hemodynamic situations. However, UDCO and LiDCO measurements of CO were obtained almost simultaneously and any temporal effect on the measurements should have been minimal. A Bland-Altman analysis does not compensate for the proportional relationship between errors and the magnitude of CO over a wide physiologic range.27,28 Thus, as CO increases, the error can increase. It is recommended to calculate a percentage error for each hemodynamic situation instead of a single percentage error calculated for the average of all the CO states, as performed in the present study. Another limitation of the study was that CO was determined by the LiDCO method and not by the gold-standard TDCO method. Use of the TDCO method to obtain CO measurements would have required the placement of a pulmonary artery catheter, which is an invasive procedure associated with risk; moreover, a catheter of suitable size for use in juvenile horses was not available at the time of the study. The measurement of CO by means of the LiDCO and TDCO methods has been compared, and there is good agreement of CO values between the 2 techniques.3 A further limitation of the present study was that the CO measurements were only obtained in duplicate. The percentage error with the TDCO method has been reported to be 22% with 1 CO measurement but can decrease to 13% if measurements are performed in triplicate.28 Both the UDCO and LiDCO methods are minimally invasive techniques for measurement of CO that require the placement of a large-bore venous catheter and a peripheral arterial catheter. Both methods require a roller pump to move the indicator at a constant velocity, and both CO methods require some degree of experience to use. Results of the present study indicated that use of the UDCO method in horses weighing 71 to 180 kg was a feasible CO monitoring technique. A previous study6 also validated the use of UDCO for measurement of CO in neonatal foals (weight range, 35 to 48 kg). Further studies are needed to evaluate the usefulness of the UDCO method for accurately monitoring hemodynamic changes in horses of different ages and body sizes with and without ongoing disease processes. a. b. c. d. e. f. g. h.
i. j. k. l.
IsoFlo, Abbott Laboratories, North Chicago, Ill. Mila International Inc, Denver, Colo. Terumo Surflow 20G Terumo Med Corp, Summerset, NJ. A/S5 Gas Calibration Canister. Datex-Ohmeda Division, Helsinki, Finland. ABL 725 Radiometer Inc, Bronshoj, Denmark. Lithium Chloride, LiDCO Limited, London, England. CM 31-01, LiDCO Ltd, London, England. Gleed RD, Kislukhin V, Krivitski N, et al. The effect of injection temperature on cardiac output determination by ultrasound dilution in an anesthetized sheep (abstr), in Proceedings, 7th World Cong Vet Anaesthesiol 2000;15. Ultrasound COStatus flow probe, Transonic System Inc. Ithaca, NY. CoStatus Roller pump, Transonic Systems Inc, Ithaca, NY. HFW1000, Transonic Systems Inc, Ithaca, NY. CPDA-1 blood collection bag, Terumo Corp Teruflex, Tokyo, Japan. AJVR, Vol 75, No. 6, June 2014
5/21/2014 10:33:07 AM
References 1.
2. 3. 4. 5. 6.
7. 8. 9.
10. 11. 12.
13.
Valverde A, Giguere S, Sanchez LC, et al. Effects of dobutamine, norepinephrine, and vasopressin on cardiovascular function in anesthetized neonatal foals with induced hypotension. Am J Vet Res 2006;67:1730–1737. Corley KT, Donaldson LL, Furr MO. Comparison of lithium dilution and thermodilution cardiac output measurements in anaesthetised neonatal foals. Equine Vet J 2002;34:598–601. Corley KT, Donaldson LL, Durando MM, et al. Cardiac output technologies with special reference to the horse. J Vet Intern Med 2003;17:262–272. Mellema M. Cardiac output, wedge pressure and oxygen delivery. Vet Clin North Am Small Anim Pract 2001;31:1175–205. Mellema, M. Cardiac output monitoring. In: Silverstein D, Hopper K, eds. Small animal critical care medicine. St Louis: Saunders Elsevier, 2009. Shih A, Giguere S, Sanchez LC, et al. Determination of cardiac output in neonatal foals by ultrasound velocity dilution and its comparison to the lithium dilution method. J Vet Emerg Crit Care (San Antonio) 2009;19:438–443. Hoffman GM, Ghanayem NS, Tweddell JS. Noninvasive assessment of cardiac output. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2005;12–21. Connors AF Jr, Speroff T, Dawson NV, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. SUPPORT Investigators. JAMA 1996;276:889–897. Giguere S, Bucki E, Adin DB, et al. Cardiac output measurement by partial carbon dioxide rebreathing, 2-dimensional echocardiography, and lithium-dilution method in anesthetized neonatal foals. J Vet Intern Med 2005;19:737–743. Mason DJ, O’Grady M, Woods JP, et al. Assessment of lithium dilution cardiac output as a technique for measurement of cardiac output in dogs. Am J Vet Res 2001;62:1255–1261. Beaulieu KE, Kerr CL, McDonell WN. Evaluation of a lithium dilution cardiac output technique as a method for measurement of cardiac output in anesthetized cats. Am J Vet Res 2005;66:1639–1645. Eremenko A, Balykov I, Chaus N, et al. Use of an extracorporeal arteriovenous tubing loop to measure cardiac output in intensive care unit patients by ultrasound velocity dilution. Asaio J 1998;44:M462–M464. Kitzler TM, Sergeyeva O, Morris A, et al. Noninvasive measurement of cardiac output in hemodialysis patients by task force monitor: a comparison with the Transonic System. Asaio J 2007;53:561–565.
14. Corley KT, Donaldson LL, Furr MM. Comparison of lithium dilution and thermodilution cardiac output measurements in anaesthetised neonatal foals. Equine Vet J 2002;34:598–601. 15. Mason DJ, O’Grady M, Woods JP, McDonell W. Assessment of lithium dilution as a technique for measurement of cardiac output in dogs. Am J Vet Res 2001;62:1255–1261. 16. Evans DL. Neonatal foal. Vet Clin North Am Equine Pract 1985;1:513–531. 17. Thomas WP, Madigan JE, Backus KQ J Reprod Fertil Suppl 1987;35:27–31 18. Bonagura JD, Muir WW. The cardiovascular system. In: Muir WW, Hubbell JA, eds. Equine anesthesia: monitoring and emergency therapy. St Louis: Mosby Elsevier, 1991:39–104. 19. Veal N, Moal F, Wang J, et al. New method of cardiac output measurement using ultrasound velocity dilution in rats. J Appl Physiol 2001;91:1274–1282. 20. Linton RA, Jonas MM, Tibby SM, et al. Cardiac output measured by lithium dilution and transpulmonary thermodilution in patients in a paediatric intensive care unit. Intensive Care Med 2000;26:1507–1511. 21. Beaulieu JM, Sotnikova TD, Yao WD, et al. Lithium antagonizes dopamine-dependent behaviors mediated by an AKT/glycogen synthase kinase 3 signaling cascade. Proc Natl Acad Sci U S A 2004;101:5099–5104. 22. Melchior R, Darling E, Terry B, et al. A novel method of measuring cardiac output in infants following extracorporeal procedures: preliminary validation in a swine model. Perfusion 2005;20:323–327. 23. Krivitski NM. Theory and validation of access flow measurement by dilution technique during hemodialysis. Kidney Int 1995;48:244–250. 24. Finsterer U, Prucksunand P, Brechtelsbauer H. Critical evaluation of methods for determination of blood volume in the dog. Pflugers Arch 1973;341:63–72. 25. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–310. 26. 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;15:85–91. 27. Lin LI-K. A concordance correlation coefficient to evaluate reproducibility Biometrics 1989;45:255–268. 28. Lin LI-K. A note on the concordance correlation coefficient. Biometrics 2000;56:324–325.
Appendix Equations commonly used to determine CO and other variables. CO = Heart rate X stroke volume Systemic vascular resistance = (MAP–central venous pressure) X 80/CO Oxygen delivery = CO X arterial oxygen concentration Oxygen consumption = arterial oxygen concentration – venous oxygen concentration X CO
AJVR, Vol 75, No. 6, June 2014
13-09-0268r.indd 571
571
5/21/2014 10:33:07 AM
