Accuracy of Continuous Cardiac Output Measurement With the LiDCOplus System During Intra-Aortic Counterpulsation After Cardiac Surgery

Accuracy of Continuous Cardiac Output Measurement With the LiDCOplus System During Intra-Aortic Counterpulsation After Cardiac Surgery Johannes Menger, MD,* Bruno Mora, MD, PhD,* Keso Skhirtladze, MD,* Arabella Fischer, MD,* Hendrik Jan Ankersmit, MD, MBA,† and Martin Dworschak, MD, MBA* Objective: To evaluate the effect of intra-aortic counterpulsation on precision, accuracy, and concordance of continuous pulse contour cardiac output determined using LiDCOplus (LiDCO Group, London). Design: Prospective trial. Setting: University hospital critical care unit. Participants: Patients with intra-aortic balloon pump support in the 1:1 mode after elective or urgent cardiac surgery. Interventions: Lithium dilution calibrated pulse contour cardiac output was compared with pulmonary artery bolus thermodilution cardiac output during hemodynamically stable conditions in the course of standardized postoperative management. Measurements and Main Results: Fifty-one paired measurements demonstrated good correlation between the 2 methods (r ¼ 0.88, p o 0.001). Mean bias was –0.14 ⫾ 0.81 L/min, limits of agreement 1.48 to –1.77 L/min, and percentage error 28%. Concordance between the 2 techniques regarding directional changes 4⫾10% cardiac output was

100% (p ¼ 0.008). Trending ability was moderate when paired cardiac output changes were assessed using linear regression, 4-quadrant table, and polar plots. When changes o⫾10% of the reference cardiac output were excluded, 90% of the data pairs still lay within the 301 radial limits. Optimal timing of the balloon pump was indispensable for proper determination of pulse contour cardiac output. Conclusions: Because of the LiDCOplus-specific algorithm in determining stroke volume from the arterial pulse waveform, which differs from other devices, accuracy and precision of continuous pulse contour cardiac output only are affected insignificantly by intra-aortic counterpulsation. The authors nevertheless caution that the device should be recalibrated after major hemodynamic alterations or otherwise inexplicable changes of the pulse contour cardiac output to improve trending. & 2016 Elsevier Inc. All rights reserved.

H

occasionally has been associated with significant complications.9,10 To circumvent the risks of pulmonary artery catheter insertion and the cumbersome and time-consuming bolus injections that merely provide intermittent CO measures, semi-invasive continuous CO measurement devices have been introduced into clinical practice. These increasingly are considered as viable alternatives.11 The LiDCOplus (LiDCO Group, London) is such a device that allows beat-to-beat CO measurement and only requires a peripheral venous and an arterial cannula. LiDCOplus has a relatively rapid response time and incorporates 2 different techniques, namely continuous pulse-power analysis (PulseCO) and lithium dilution (LiDCO), to determine CO. Most of the time, lithium dilution merely serves the purpose of calibrating the pulse contour method. Values are reported to be stable for at least 4 to 8 hours, after which recalibration with lithium dilution is advised.12 PulseCO of the LiDCOplus has been validated after vascular, urologic, and thoracic surgery13 and demonstrates good agreement with pulmonary artery ThD-CO in swine14 and in patients with impaired left ventricular systolic function after cardiac surgery.15 LiDCOplus, which could be particularly helpful in critically ill patients requiring mechanical circulatory support, however, has not been validated yet in patients with an IABP, in whom the pulse contour is changed markedly by intra-aortic counterpulsation. To date it is unclear how the altered arterial waveform morphology in IABP patients affects the quality of data of the continuous pulse-wave-derived CO measurements of the LiDCOplus. In contrast, investigations with the FloTrac (Vigileo 1.07; Edwards Lifesciences, Irvine, CA) and the PiCCO (Pulsion Medical Systems, Munich, Germany) devices have indicated that continuous CO measurements using pulse-wave analysis becomes unreliable in the presence of an IABP.16,17 However, the specific algorithms used by these 2 devices to

EMODYNAMIC MONITORING IS a central component in the care of critically ill patients.1 Early goal-directed therapy that incorporates optimization of cardiac output (CO) to assist in guiding therapeutic decision making requires a measured CO that is reliable and accurate.2 Preferably, the CO measurement should be available continuously to rapidly detect hemodynamic changes. Patients with compromised ventricular pump function who most likely benefit from determination of CO, however, occasionally require mechanical assist devices, such as the intraaortic balloon pump (IABP) for hemodynamic stabilization. Although there is controversy regarding the need and benefits of the IABP in cardiac surgery, this device still is in widespread clinical use.3–6 Determination of intermittent bolus thermodilution CO (ThDCO) using a pulmonary artery catheter is considered to be the current clinical reference technique for CO monitoring against which others often are compared.7 However, this method has an inherent error rate of 13%, even with triple measurements,8 and

From the *Division of Cardiothoracic and Vascular Anesthesia and Intensive Care Medicine; and †Department of Surgery; and Christian Doppler Laboratory for Cardiac and Thoracic Diseases, General Hospital Vienna, Medical University of Vienna, Vienna, Austria. Address reprint requests to Martin Dworschak, MD, MBA, Division of Cardiothoracic and Vascular Anesthesia and Intensive Care Medicine, General Hospital Vienna, Medical University of Vienna, Währinger Gürtel 18-20, 1090 Vienna, Austria. E-mail: martin.dworschak@ meduniwien.ac.at © 2016 Elsevier Inc. All rights reserved. 1053-0770/2601-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2015.09.022 592

KEY WORDS: pulse contour analysis, cardiac output, monitoring, intra-aortic counterpulsation, cardiac surgery

Journal of Cardiothoracic and Vascular Anesthesia, Vol 30, No 3 (June), 2016: pp 592–598

PULSE CONTOUR CARDIAC OUTPUT AND INTRA-AORTIC COUNTERPULSATION

calculate CO from the pulse-contour trace are different from that used by LiDCOplus. The objective of this study was to compare ThD-CO and PulseCO measurements of the LiDCOplus system during intraaortic counterpulsation after cardiac surgery. The authors hypothesized that intra-aortic counterpulsation would affect accuracy, precision, and trending ability of PulseCO measurements. METHODS

After receiving approval from the local ethics committee, the authors conducted the investigation at their cardiothoracic intensive care unit. Thirteen patients requiring perioperative implantation of an IABP catheter with a 40-mL balloon (Teleflex, Morrisville, NC) because of severely impaired left ventricular function and a 1:1 assist mode were enrolled in this study. The authors only included adult patients undergoing elective or urgent cardiac surgery. Because IABP implantation is an emergency intervention and the patients usually are sedated or not conscious before the decision for IABP insertion is made, consent was obtained after recovery in survivors or from relatives of nonsurvivors. According to institutional protocol, these emergency patients are managed routinely with a Swan-Ganz catheter. The IABP was triggered either by electrocardiogram or pressure, depending on whichever achieved the best augmentation of diastolic pressure. In addition, left ventricular function was assessed using preoperative transthoracic and/or intraoperative transesophageal echocardiography. Exclusion criteria included oral lithium therapy, pregnancy or women of childbearing age, weight o40 kg, extracardiac or intracardiac shunt, tricuspid or aortic valve regurgitation, ventricular arrhythmias, new-onset atrial fibrillation or atrial fibrillation with uncontrolled ventricular response (ie, heart rate 4100 bpm), and/or known contraindications for pulmonary artery catheter insertion (eg, tricuspid valve stenosis, pulmonary valve stenosis). Parallel measurements were all made during periods of hemodynamic stability. All patients were ventilated mechanically (PEEP 5 cmH2O, tidal volume 5-7 mL/kg peak pressure o25 cmH2O). Measurement of ThD-CO As previously mentioned, a 20-G radial artery catheter, a central venous catheter, and a pulmonary artery catheter (OTD catheter; Edwards Lifesciences) were used as routine monitoring in all patients. The pulmonary artery catheter was connected to a Vigilance monitor (Edwards Lifesciences) for determination of bolus ThD-CO. ThD-CO measurements were made by injecting 10 mL of saline at room temperature through the central venous port. It was calculated as the mean of at least 3 separate measurements randomly conducted during the respiratory cycle. Measurements were repeated when discrepancies 415% between 2 measurements were encountered. Measurement of PulseCO A LiDCOplus monitoring set was connected to the standard intensive care unit hemodynamic monitor (Infinity;

593

Dräger, Lübeck, Germany) with input from the pressure transducer connected to the radial artery for continuous determination of PulseCO. A single initial calibration of the device was performed in the intensive care unit during IABP support using 0.3 mmol of lithium chloride (2 mL) to measure LiDCO according to the manufacturer’s instructions. No drugs containing quaternary ammonium ions had been administered within 5 hours before commencement of the trial to prevent interferences. After calibration, measurements of CO were made with both techniques (ie, ThD-CO and PulseCO), after hemodynamics had changed but conditions were stable again. Before measurements were taken, IABP support was checked for proper synchronization. PulseCO was calculated as the mean of at least 3 visually observed CO values that preceded the corresponding ThDCO measurements. As with ThD-CO measurements, up to 2 additional measurements were taken when discrepancies 415% between 2 values were encountered. Obvious outliers, mostly due to single ventricular premature beats, were not counted. The dosage of intravenous drugs and fluids was not altered at the time when a measurement was taken. Statistical Analysis Statistical analysis was performed using R for Windows 3.2.0 (R Foundation for Statistical Computing, Vienna, Austria), including the MethComp package 1.22.2 (Bendix Carstensen, Department of Biostatistics, University of Copenhagen, Denmark). Results are presented as mean ⫾ standard deviation (SD) unless otherwise stated. A p value of o0.05 was chosen to indicate statistical significance. Pearson’s correlation coefficient and linear regression were used to compare LiDCO and PulseCO with ThD-CO. Furthermore, Bland-Altman analysis18 was used to assess accuracy and precision. The authors adjusted for the effects of replicate measurements within each subject in the Bland-Altman analysis using the method published by Carstensen et al.19 A mean percentage error r30% was defined to indicate clinically useful reliability of the LiDCOplus-CO.20 Change in CO (ΔCO) between 2 sequential measurements was used to assess trending. The following 3 statistical approaches were used to analyze the trending ability: (1) The authors evaluated the correlation between sequential ΔCO measured with the 2 techniques using linear regression and determination of the Pearson’s correlation coefficient. (2) The authors analyzed concordance of major (ie, ThDCO 4⫾10%) sequential CO changes with the help of a 4-quadrant table. Concordance of relative directional changes was calculated using the Fisher’s exact test. The authors only included major CO changes because minor CO changes tend to be randomly distributed over a 4-quadrant table and do not reflect the trending ability.21 (3) According to Critchley et al,21 data also were analyzed using a polar plot, which uses the angle and length of the ΔCO vector (polar coordinates) to demonstrate the predictability between different values of CO. It determines the tracking ability of continuous CO monitors and displays information that is lost in concordance analysis.

594

MENGER ET AL

Table 1. Patient Characteristics and Intraoperative Data

Age (yr) Ejection fraction (%)

Sex (male) Type of surgery CABG CABG þ MVR CABG þ AVR AVR Pre-existing conditions COPD Diabetes

Mean

SD

69 30

10 4.6

N

Proportion

12

92%

8 2 2 1

62% 15% 15% 8%

4 3

31% 23%

Abbreviations: AVR, aortic valve replacement; CABG, coronary artery bypass graft; COPD, chronic obstructive pulmonary disease; MVR, mitral valve reconstruction; SD, standard deviation.

RESULTS

The mean age of the 13 study patients (male [n ¼ 12]) was 69 ⫾ 10 years. All patients had undergone cardiac surgery with cardiopulmonary bypass and moderate hypothermia. Patient characteristics and intraoperative data are provided in Table 1. All patients had an IABP implanted perioperatively and required full 1:1 support. The mean study period per patient was 2:00 ⫾ 1:20 hours, with a mean time between 2 measurements of 40 ⫾ 32 minutes and a mean number of paired measurements per patient of 4.2 ⫾ 1.0. Continuous infusion of up to 0.65 mg/kg/min of norepinephrine and up to 6.1 mg/kg/min of dobutamine was required in most patients. The dosage of inotropes and vasoconstrictors was not changed during measurement periods. Ten patients were in sinus rhythm, and 3 patients demonstrated intermittent atrial fibrillation. Mean perioperative ejection fraction was 30% ⫾ 5%. Before measurements were taken, the PulseCO was calibrated with LiDCO (n ¼ 13). The relationship with simultaneously determined ThD-CO can be described as: LiDCO ¼ 0.79  ThD-CO þ 0.95 (r ¼ 0.82; p o 0.001; 95%

confidence interval [CI]: 0.49-0.94; Fig 1A). Bland-Altman analysis showed a mean bias of –0.24 ⫾ 0.90 L/min and limits of agreement of 1.55 to –2.03 L/min (Fig 1B). The percentage error (1.96 SD/mean of the reference method) was 30%. After confirmation of optimal timing of the IABP, 51 paired measurements were taken. This number is comparable with other method comparison studies.16,22 Patients in sinus rhythm exhibited identical heart rates independent of the way it was assessed (ie, electrocardiogram, pulse oximetry, or pulse contour). Improper timing of the IABP and (predominantly isolated) extrasystoles led to inaccurate determination of pulse contour stroke volume and/or pulse contour heart rate and occasionally to a severe deviation between PulseCO and ThD-CO. Mean CO was 5.8 ⫾ 1.5 L/min (range: 3.3-9.2 L/min) for ThD-CO and 5.9 ⫾ 1.7 L/min (range: 3.2-10.0 L/min) for PulseCO. The correlation coefficient between ThD-CO and PulseCO was r = 0.88 (p o 0.001; 95% CI, 0.79-0.93) and r2 = 0.77 (Fig 2A). Bland-Altman analysis revealed a mean bias of –0.14 ⫾ 0.81 L/min and limits of agreement (1.96 SD) of 1.48 to –1.77 L/min (Fig 2B). The percentage error (1.96 SD/ mean of the reference method) was 27.6%. When BlandAltman analysis data were corrected for replicate measurements,19 the mean bias became –0.21 ⫾ 0.83 L/min and limits of agreement 1.43 to –1.85 L/min. The percentage error (1.96 SD/mean of the reference method) was 28% To evaluate the trending ability, 38 sequential ΔCO measurements were recorded. The overall ΔCO range measured was –1.7 to 1.3 L/min for ΔThD-CO, with a mean absolute change of 0.4 ⫾ 0.4 L/min; and –1.4 to 2.5 L/min for ΔPulseCO, with a mean absolute change of 0.5 ⫾ 0.6 L/min. The correlation coefficient between ΔThD-CO and ΔPulseCO was r ¼ 0.53 (p o 0.001; 95% CI 0.25-0.72) and r2 ¼ 0.28 (Fig 3). The authors’ observations included 10 paired measurements with major ThD-CO changes in relation to previous measurements. In 7 measurements, ThD-CO increased (median [IQR range]): þ16 (þ16 to þ22 [þ14 to þ28])%). All 7 paired PulseCO values also demonstrated increasing values (þ14 (þ10 to þ34 [þ7 to þ63])%). In 3 measurements, ThD-CO

Fig 1. (A) Relationship between bolus thermodilution cardiac output (ThD-CO) and lithium dilution cardiac output (LiDCO) measured at the initial calibration of the LiDCOplus pulse contour cardiac output (PulseCO) technique. (B) Bland-Altman plot of cardiac output measured using bolus lithium dilution (LiDCO) and intermittent bolus thermodilution (ThD-CO) during calibration of pulse contour cardiac output (PulseCO).

PULSE CONTOUR CARDIAC OUTPUT AND INTRA-AORTIC COUNTERPULSATION

595

Fig 2. (A) Relationship between pulmonary artery thermodilution cardiac output (ThD-CO) and pulse contour cardiac output determined using the LiDCOplus device (PulseCO) for all measurements (n ¼ 51), including regression line PulseCO ¼ (0.99  ThD-CO) þ 0.23. (B) BlandAltman plot for repeated measurements showing mean bias and limits of agreement (1.96 SD). ThD-CO, pulmonary artery thermodilution cardiac output; PulseCO, pulse contour cardiac output determined using the LiDCOplus device.

decreased (–20 (–22 to –19 [–24 to –18])%). Corresponding PulseCO values also showed decreasing values (–10 (–20 to – 5.7 [–30 to –1])%). PulseCO thus showed 100% concordance in tracking directional ThD-CO changes 4⫾10% (p ¼ 0.008; Table 2). The analysis in the polar plot21 was limited by the few data points but depicted acceptable trending ability; 100% of all pairs (n ¼ 38) lay within 1 L/min, or 17% limits of agreement. Inside 0.5 L/min, or 9% limits of agreement, were 87% (n ¼ 33) of all data pairs (Fig 4). To calculate the limits of agreement the authors used the mean ThD-CO (ie, 5.8 L/ min). When minor changes of the reference CO (ie, o⫾10% ThD-CO) were excluded, the 95% radial limits of agreement were 271 and the mean polar angle was 171 ⫾ 121. DISCUSSION

This was the first study to compare ThD-CO determined using a pulmonary artery catheter and lithium dilution calibrated continuous PulseCO using the LiDCOplus pulse-power algorithm in patients requiring 1:1 IABP support immediately after cardiac surgery. The accuracy of LiDCOplus in continuously measuring CO has been confirmed before in a similar environment (ie, in patients after cardiac surgery who were administered inotropes and vasopressors)15 and in various other clinical settings.13 Furthermore, a recent review that included 5 studies comparing PulseCO of LiDCOplus with pulmonary artery ThD-CO also demonstrated a pooled weighted percentage error of 27%.11 Interestingly, bias, limits of agreement, and percentage error in this trial were almost analogous to values that have been observed when PulseCO was evaluated in cardiac surgery patients with diminished left ventricular pump function whose treatment did not include an IABP. 15 As previously mentioned, intermittent ThD-CO measurement with a Swan Ganz catheter is the clinical gold standard for measuring CO. Previous trials showed that other dilution techniques, such as indocyanine green dye dilution23 and transpulmonary thermodilution,17 are comparable to pulmonary artery ThD-CO during IABP support. The IABP, therefore, did not seem to impair the results of noncontinuous dilution techniques to determine CO. LiDCOplus uses lithium dilution

for calibration of continuous PulseCO. The specific accuracy of the lithium dilution technique has not been tested yet during IABP support but in multiple other settings. Linton et al,24 for example, compared pulmonary artery ThD-CO and LiDCO in postoperative intensive care patients. They showed excellent correlation and adequate agreement (r2 ¼ 0.94, bias ¼ – 0.25 ⫾ 0.46 L/min) between LiDCO and pulmonary artery ThD-CO. Although the number of measurements was limited (n ¼ 13), the results of this study suggested that, in this setting, lithium dilution correlated well with pulmonary artery ThD-CO (r ¼ 0.82) and appeared to be as reliable as other dilution techniques. On the other hand, inconclusive observations concerning the reliability of different systems using pulse-contour methods to measure CO during IABP support have been reported. During IABP use the pulse-pressure curve is distorted by pre-set volume changes in the aorta. As the balloon of the IABP inflates with appearance of the dicrotic notch, a second augmented diastolic pressure peak appears after the unassisted peak. The balloon deflates just at the end of diastole and thereby creates suction power, which decreases the aortic end-diastolic pressure below the patient’s baseline value. Deflation is maintained until the onset of the next systolic left ventricular ejection causing the first unasisted peak in the pulse pressure trace.25 Lorsomradee et al16 compared uncalibrated arterial pulsecontour CO (Vigileo 1.07; Edwards Lifesciences) and pulmonary artery ThD-CO after cardiac surgery. In the IABP subgroup (n ¼ 12), they found only poor agreement between the 2 techniques. The Vigileo frequently displayed error messages during different periods and failed to show CO values for several minutes in 80% of the patients with IABP support. Janda et al17 studied the impact of IABP on CO measurements in swine using transpulmonary thermodilution and pulsecontour analysis with the PiCCO system (Pulsion Medical Systems). The authors found close agreement between transpulmonary and pulmonary artery ThD-CO (r ¼ 0.94) and no interference with the thermodilution measurements by the IABP, but the pulse-contour function of the PiCCO system was not working. Scolletta et al22 reported that another novel uncalibrated pulse-contour method (MostCare; Vytech Health, Padova,

596

MENGER ET AL

Fig 3. Relationship between sequential change in pulmonary artery thermodilution cardiac output (ΔThD-CO) and sequential change in pulse-contour cardiac output determined using the LiDCOplus device (ΔPulseCO) for all measurements (n ¼ 38), including regression line ΔPulseCO ¼ (0.70  ΔThD-CO) þ 0.05.

Table 2. Four-Quadrant Table of Paired Measurements From Major (4⫾10%) Changes in ThD-CO in Relation to the Previous Measurement and Corresponding PulseCO Measures ΔPulseCO 40%

ΔPulseCO o0%

0 —

ΔThD-CO 4–10% ΔThD-CO (%)

7 þ16 (þ16 to þ22 [þ14 to þ28]) þ14 (þ10 to þ34 [þ7 to þ63]) 0 —

ΔPulseCO (%)

—

ΔThD-CO 4þ10% ΔThD-CO (%) ΔPulseCO (%)

— 3 –20 (–22 to –19 [–24 to –18]) –10 (–20 to –5.7 [–30 to –1]) p ¼ 0.008

NOTE. Values are either median (interquartile range [range]) or numbers. Abbreviations: ΔPulseCO, change in pulse-power cardiac output; ΔThD-CO, change in thermodilution cardiac output.

Italy) was reliable (percentage error: 24%) in patients with IABP support (n ¼ 15) after isolated CABG surgery. These results, however, were in stark contrast with those obtained by Paarmann et al26 who compared CO values obtained using MostCare with pulmonary artery ThD-CO in patients after various cardiac surgery procedures (n ¼ 23, excluding IABP). They found a huge percentage error of 87% that caused them to advise against the use of MostCare to determine CO in cardiac surgery patients—even with an unaltered pulse-pressure curve. Reasons for the poor performance of the PiCCO and Vigileo systems could be inborn algorithms that are used to calculate stroke volume. The PiCCO system, for example, uses only the systolic part of the pulse-pressure wave.27 It therefore needs to detect the dicrotic notch to determine stroke volume.17 On the other hand, the Vigileo monitor identifies falsely high pulse rates because each upslope of the pulse waveform is counted as a single heartbeat.28 These differences in signal processing may explain why the PiCCO and Vigileo devices commonly are not

PULSE CONTOUR CARDIAC OUTPUT AND INTRA-AORTIC COUNTERPULSATION

135

180

mean Δ CO (L/min)

90

45

0

0

0.5 1 225

1.5

315

2 270 Fig 4. Polar plot to show trending ability. The distance from the center of the plot represents the mean change in cardiac output (ΔCO) and the angle with the horizontal axis (0-degree radial) represents agreement. The closer the agreement between CO measurements, the closer data pairs will lie along the vertical radial axis.

capable to accurately interpret the pulse-pressure curve being altered by intra-aortic counterpulsation. This drawback seems to make both of them unsuitable for continuous CO measurement in IABP-supported patients.16,17 Contrary to the previously mentioned devices, LiDCOplus uses pulse-power analyses to analyze the arterial pulse-pressure curve. This nonmorphology-based algorithm uses the whole pulse-contour curve (ie, the systolic and the diastolic parts), which seems to be advantageous in patients with pulse-pressure waves being distorted by an IABP. The underlying physical law is conservation of power. Net power changes after a pulse consequently should reflect the balance between input (ie, stroke volume) and output (ie, blood being pushed into the peripheral vasculature). To achieve a linear relationship between net power and net flow, the system requires corrections for vascular compliance and calibration.25,26 The observations made during this study demonstrated that in patients receiving IABP support, it was extremely crucial that IABP timing was optimal so that the device could detect the full

597

pulse-pressure wave (ie, the merged systolic and diastolic part) of a single left ventricular ejection. This is essential to calculate stroke volume, and then, by multiplying the derived stroke volume with the number of pulse-pressure waves per minute (ie, heart rate), eventually determine CO. Thereby, tracking CO changes with PulseCO becomes technically feasible. Because the LiDCOplus device always allows recalibration of PulseCO when readings do not match the clinical situation, the authors decided to analyze only major changes of the reference method to assess the trending ability of substantial CO changes. Concordance and polar plot analysis as proposed by Critchley et al29 were modulated accordingly. The results with the LiDCOplus system indicated good agreement between ThD-CO and PulseCO, provided IABP timing was optimal. Under this circumstance, the LiDCOplus device rarely showed error messages or implausible values when hemodynamics were stable. Furthermore, the ability of the device to adequately track clinically relevant directional changes of CO was acceptable. Nevertheless, there are some limitations that need to be mentioned in this context. Because measurements were conducted during full IABP support, the authors can merely report on LiDCOplus characteristics during the 1:1 assist mode. Accuracy during weaning from IABP (ie, in the 2:1 or 3:1 mode) appeared to decrease, although the authors did not evaluate this systematically. The specific limitations of the LiDCOplus technology have been mentioned previously. There also were limitations regarding the methodology of the trial presented here. First, because assessments of ThD-CO and PulseCO were not performed simultaneously, but rather with a small time delay, slight hemodynamic alterations that may have occurred during this interval could have influenced the results. However, it can be assumed that these alterations should be small because the authors measured only during stable hemodynamics that were free from arrhythmias. Second, although the authors found acceptable tracking capability for major CO changes, recalibration of PulseCO has been recommended after gross hemodynamic perturbances.30,31 In patients with IABP support, the authors propose recalibration after major, persistent alterations of PulseCO or inexplicable PulseCO readings, especially before decisive therapeutic actions are taken. Third, the authors did not evaluate each technique (PulseCO/LiDCO) separately but only continuous PulseCO after initial lithium calibration to restrict exposure to lithium.

REFERENCES 1. Pinsky MR, Payen D: Functional hemodynamic monitoring. Crit Care 9:566-572, 2005 2. Rivers E, Nguyen B, Havstad S, et al: Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 345: 1368-1377, 2001 3. Unverzagt S, Buerke M, de Waha A, et al: Intra-aortic balloon pump counterpulsation (IABP) for myocardial infarction complicated by cardiogenic shock. Cochrane Database Syst Rev CD 007398, 2015 4. Theologou T, Bashir M, Rengarajan A, et al: Preoperative intra aortic balloon pumps in patients undergoing coronary artery bypass grafting. Cochrane Database Syst Rev CD 004472, 2011

5. Mannacio V, Di Tommaso L, De Amicis V, et al: Preoperative intraaortic balloon pump for off-pump coronary arterial revascularization. Ann Thorac Surg 93:804-809, 2012 6. Gargiulo G, Stabile E, Perrino C, et al: Contemporary use and results of intra-aortic balloon pump counterpulsation. Minerva Cardioangiol 2014 Nov 25 [Epub ahead of print]. 7. Ganz W, Donoso R, Marcus HS, et al: A new technique for measurement of cardiac output by thermodilution in man. Am J Cardiol 27:392-396, 1971 8. Stetz C, Miller R, Kelly G, et al: Reliability of the thermodilution method in the determination of cardiac output in clinical practice. Am Rev Respir Dis 126:1001-1004, 1982

598

9. Connors AF, Speroff T, Dawson NV, et al: The effectiveness of right heart catheterization in the initial care of critically III patients. JAMA 276:889-897, 1996 10. Hadian M, Pinsky MR: Evidence-based review of the use of the pulmonary artery catheter: impact data and complications. Crit Care 10 (Suppl 3):S8, 2006 11. Schlöglhofer T, Gilly H, Schima H: Semi-invasive measurement of cardiac output based on pulse contour: a review and analysis. Can J Anesth 61:452-479, 2014 12. Hamilton TT, Huber LM, Jessen ME: PulseCO: a less-invasive method to monitor cardiac output from arterial pressure after cardiac surgery. Ann Thorac Surg. 74:S1408-S1412, 2002 13. Pittman J, Bar-Yosef S, SumPing J, et al: Continuous cardiac output monitoring with pulse contour analysis: a comparison with lithium indicator dilution cardiac output measurement. Crit Care Med 33:2015-2021, 2005 14. Kurita T, Morita K, Kato S, et al: Comparison of the accuracy of the lithium dilution technique with the thermodilution technique for measurement of cardiac output. Br J Anaesth 79:770-775, 1997 15. Mora B, Ince I, Birkenberg B, et al: Validation of cardiac output measurement with the LiDCOTM pulse contour system in patients with impaired left ventricular function after cardiac surgery. Anaesthesia 66: 675-681, 2011 16. Lorsomradee S, Lorsomradee S, Cromheecke S, et al: Uncalibrated arterial pulse contour analysis versus continuous thermodilution technique: effects of alterations in arterial waveform. J Cardiothorac Vasc Anesth 21:636-643, 2007 17. Janda M, Scheeren TW, Bajorat J, et al: The impact of intraaortic balloon pumping on cardiac output determination by pulmonary arterial and transpulmonary thermodilution in pigs. J Cardiothorac Vasc Anesth 20:320-324, 2006 18. Bland JM, Altman DG: Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1:307-310, 1986 19. Carstensen B, Simpson J, Gurrin LC: Statistical models for assessing agreement in method comparison studies with replicate measurements. Int J Biostat 4:Article 16, 2008

MENGER ET AL

20. Critchley LA, Critchley JA: A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. J Cin Monit Comput 15:85-91, 1999 21. 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 111:1180-1192, 2010 22. Scolletta S, Franchi F, Taccone FS, et al: An uncalibrated pulse contour method to measure cardiac output during aortic counterpulsation. Anesth Analg 113:1389-1395, 2011 23. Herzlinger GA: Absolute determination of cardiac output in intra-aortic balloon pumped patients using the radial arterial pressure trace. Circulation 53:417-421, 1976 24. Linton R, Band D, O’Brien T, et al: Lithium dilution cardiac output measurement: a comparison with thermodilution. Crit Care Med 25:1796-1800, 1997 25. Trost JC, Hillis LD: Intra-aortic balloon counterpulsation. Am J Cardiol 97:1391-1398, 2006 26. Paarmann H, Groesdonk HV, Sedemund-Adib B, et al: Lack of agreement between pulmonary arterial thermodilution cardiac output and the pressure recording analytical method in postoperative cardiac surgery patients. Br J Anaesth 106:475-481, 2011 27. De Wilde R, Schreuder J, Van Den Berg P, et al: An evaluation of cardiac output by five arterial pulse contour techniques during cardiac surgery. Anaesthesia 62:760-768, 2007 28. Maus TM, Lee DE: Arterial pressure–based cardiac output assessment. J Cardiothorac Vasc Anesth 22:468-473, 2008 29. Critchley LA, Yang XX, Lee A: Assessment of trending ability of cardiac output monitors by polar plot methodology. J Cardiothorac Vasc Anesth 25:536-546, 2011 30. Cooper ES, Muir WW: Continuous cardiac output monitoring via arterial pressure waveform analysis following severe hemorrhagic shock in dogs. Crit Care Med 35:1724-1729, 2007 31. Yamashita K, Nishiyama T, Yokoyama T, et al: Cardiothoracic anesthesia, respiration and airway cardiac output by PulseCO is not interchangeable with thermodilution in patients undergoing OPCAB. Can J Anesth 52:530-534, 2005