To the Editor:—
We read the recent editorial by Dyer and James,1referencing the paper by Langesæter et al .2with great interest. Langesæter et al .2have successfully demonstrated that a minimally invasive technology that measures cardiac output (LiDCOplus , LiDCO Ltd., Cambridge, United Kingdom) can be used for maternal hemodynamic monitoring. Other arterial pressure waveform–based systems that provide beat-by-beat assessment of cardiac output and stroke volume include the PiCCOplus (Pulsion Medical Systems, Munich, Germany) and the Vigileo monitor/FloTrac sensor (Edwards Lifesciences LLC, Irvine, CA). Dyer and James1note in their editorial that these less invasive methods of hemodynamic monitoring are attractive to the obstetric anesthesiologist. However, the editorial1view that the Vigileo monitor is unsuitable for the study of rapid maternal hemodynamic changes is not a logical conclusion.
Bland-Altman analysis is routinely used to assess precision and bias when comparing two different measurement techniques.3–9As acknowledged in the editorial, the acceptance of a new technique of cardiac output measurement should rely on limits of agreement of up to ± 30% between the minimally invasive techniques and the existing “gold-standard” (i.e., the thermodilution pulmonary artery catheter).5There are currently several studies in the critical care and cardiac anesthesia literature that have evaluated precision and bias (using Bland-Altman analyses) in the measurements of cardiac output with the Vigileo monitor/FloTrac sensor.3,4,7–10Based on these studies, statistically and clinically acceptable precision and bias has been shown to exist in the measurement of cardiac output by the Vigileo monitor. The novel arterial pressure cardiac output algorithm used by the Vigileo monitor provides cardiac output assessments that agree satisfactorily for clinical purposes with intermittent and continuous thermodilution techniques in postoperative cardiac surgical patients4,7–9and in the critically ill population.3,10Fluid and pharmacologic therapy algorithms based on cardiac output and stroke volume variation measurements from the Vigileo monitor are now being studied (with the hope of improving outcome) in patients who are critically ill and/or undergoing major surgery.3Further, such studies may be needed in the noncritically ill spontaneously breathing patient populations, including pregnant women undergoing cesarean sections.
We will now examine the specific situation of the measurement of cardiac output after the administration of phenylephrine during cardiac surgery (based on which the editorial concludes that the Vigileo monitor “may not be suitable”).9The basic physiologic principle underpinning the Vigileo device is that left ventricular stroke volume and arterial pulsatility are proportional to each other, and the proportionality constant, κ, is a number that describes the resistance and compliance of the arterial tree: Stroke volume ≈ pulsatility x κ.6Pulsatility is calculated by using the SD of the arterial pressure waveform over a 20-s period analyzed at a 100-Hz frequency. The κ value is calculated every minute by the latest operating system and is based on patient weight, height, age, mean arterial pressure, skewness, and kurtosis of the pressure wave.11It has been implied by Lorsomradee et al .9that immediately after a phenylephrine bolus (or after sternotomy), the Vigileo/FloTrac system overestimated the cardiac output relative to the continuous cardiac output measurement by the thermodilution pulmonary artery catheter. Phenylephrine (especially as a bolus dose) will abruptly affect the pulsatility of the arterial system. This acute arteriolar constriction will cause the left ventricular stroke volume to be ejected into a more “pressurized” central arterial tree, simulating an increase in pulsatility. The measurements obtained by the Vigileo monitor will initially interpret this increase in pulsatility as an increase in left-sided stroke volume and cardiac output. The right-sided cardiac output being measured by the thermodilution pulmonary artery catheter will not reflect this sudden change. The current software version on the Vigileo system, however, recalibrates itself constantly, and over the next 1 to 2 min the FloTrac sensor will “relearn” this new increased vascular tone, recalculate κ, and then report updated stroke volume and cardiac output values.11The converse may be expected acutely when a vasodilator (such as nitroprusside) is administered. The vascular tone and therefore pulsatility will acutely decrease, and the initial Vigileo cardiac output readings will underestimate thermodilution measurements. Understanding the mechanisms of cardiac output measurement by the Vigileo monitor allows us to anticipate such measurement errors. Every device has its limitations. A bolus of thermally active injectate (e.g., red blood cells/fluids being administered rapidly) will produce errors in the measurement of cardiac output by the thermodilution pulmonary artery catheter. Similarly, acute changes in arterial pulsatility will produce errors in the arterial waveform based measurements systems. Since we do not need to externally calibrate the Vigileo system for measurement (the system regularly autocalibrates), we are essentially exchanging ease of use for clinically acceptable precision and bias. Waiting a few minutes for autorecalibration and the recalculation of κ is appropriate before relying on Vigileo-reported measurements.
Dyer and James1also state that central venous pressures and pulmonary capillary wedge pressures are unlikely to predict the response to fluid administration, and that pulse pressure variation and stroke volume variation (SVV) may be better indicators of fluid responsiveness. Systolic blood pressure variation, pulse pressure variation, and SVV are all clinically reliable measures of fluid responsiveness in nonobstetric populations.12–14However, these cardiorespiratory interactions have mostly been studied in mechanically ventilated, critically ill patients. Robust data in spontaneously breathing subjects are not widely available. Pulse pressure variation or SVV changes may be harder to interpret during spontaneous ventilation, given the inherent variations in respiration. The LiDCOplus system displays a continuous measure of SVV, and it would be interesting to know if this data were acquired by Langesæter et al . Since subjects in this study2adhered to a relatively strict fluid administration protocol (with no prehydration and moderate cohydration) and a standardized vasopressor administration protocol, we could gain useful information from any SVV data that might be available.
*Tufts University School of Medicine, Baystate Medical Center, Springfield, Massachusetts. neil.roy.connelly@bhs.org