Plethysmography Variability Index as an Indicator of Adequacy of Preload

BACKGROUND

Optimal intraoperative volume management is a mainstay during general anesthesia. Adequate intraoperative fluid administration enhances preload and supports optimal cardiac output. Intraoperative optimization of cardiac output using volume expansion decreases postoperative morbidity and duration of hospital stay.1 Optimization of intraoperative cardiac output depends on the ability to assess its preload dependence and fluid responsiveness. Static indicators of fluid responsiveness such as central venous pressure (CVP), pulmonary capillary wedge pressure (PCWP), or left ventricular end diastolic area index are invasive or are not easily available and have been shown to be poor predictors of fluid responsiveness.2-9

Dynamic indicators, such as respiratory variations in arterial pulse pressure (PPV)3 or stroke volume variations (SVV)5, relying on the respiratory variations in stroke volume or its surrogates in mechanically ventilated patients have been shown to be superior to static indicators for prediction of fluid responsiveness. 8,10-12 Both techniques rely on pulse contour or pulse power analysis derived from an arterial line and are invasive. Respiratory variations of preload can also be assessed in pulse Doppler aortic flow velocity2,11 or by esophageal Doppler.13

Recently, it has been shown that respiratory variations in the pulse oximeter plethysmography waveform amplitude (∆POP) are sensitive to changes in preload14 and can predict fluid responsiveness in mechanically ventilated patients.15 Pulse oximeter plethysmography waveform amplitudes are strongly related to the respiratory variations in arterial pulse pressure (PPV) 16,17 and are sensitive to changes in ventricular preload.14

Recently published studies have shown promising results regarding the ability of ∆POP to predict fluid responsiveness in the intensive care unit15 and in the operating room.18

Pulse oximeter plethysmography waveform amplitude variations cannot be easily calculated from a bedside monitor. Thus a plethysmography variability index (PVI, Masimo Corp.) has been created. PVI is a new algorithm that automatically calculates ∆POP. PVI is a measure of the dynamic change in perfusion index (PI) that occurs during a specific period of time, sufficiently long enough to include at least one complete respiratory cycle. For PI calculation, the infrared pulsatile signal is indexed against the non-pulsatile infrared signal and expressed as a percentage (PI- (AC/DC×100) reflecting the amplitude of the pulse oximeter waveform. PVI calculation is accomplished by measuring changes in PI over time as PVI = (Pimax-Pimin/Pimax) - 100.

In patients undergoing robotic procedures needing a pneumoperitoneum and steep Trendelenburg position, it is unknown if cardiac output decreases predictably during pneumoperitoneum and to what extent the steep Trendelenburg position offsets the decrease in cardiac output. In addition, pulse pressure variation (PPV) has been shown to reliably predict the response to fluid administration in an animal model in pneumoperitoneum.20 However, no human data are available in this setting.

Thus, the aim of our study is to test the hypothesis that stroke volume will drop significantly after initiation of pneumoperitoneum and that it will increase after placement in Trendelenburg position in patients undergoing robotic gynecological procedures. Secondly, the investigators will test the hypothesis that PVI changes correlate with changes in stroke volume and pulse pressure variation (PPV) and can predict an increase in stroke volume after a fluid bolus. The third hypothesis is that delta PVI is independent of the vaso-status of precapillaries at the measured site.

METHODS

The investigators propose a prospective observational trial with approval of the Human Studies Committees at the University of Louisville and University Hospital Research Integrity Office. All subjects or their legal representatives must give informed, written consent before being enrolled in the study. The study will be conducted under Good Clinical Practice Guidelines.

Inclusion criteria The investigators plan to enroll 40 female subjects scheduled for elective robotic gynecological surgery under general anesthesia.

Exclusion criteria

Patients with atrial fibrillation or other significant arrhythmia (Lown grade 3 or greater) and patients with aortic regurgitation will be excluded.

Protocol

Anesthesia will be induced with propofol (2 mg/kg), rocuronium (0.6 mg/kg) and fentanyl (2 µg/kg) to facilitate intubation. A #8 endotracheal tube will be inserted in the trachea and secured at 22 cm. Anesthesia will be maintained with sevoflurane to maintain the BIS value between 40 and 60. Fentanyl will be given as an infusion at a rate of 2 µg/kg/h. The ventilator will be set at pressure-controlled mode and patients will be ventilated with an airway pressure corresponding to a tidal volume of 8 mL/kg up to a maximum airway pressure of 35 cmH20. Inspired oxygen will be kept near 50% but will be increased if necessary to maintain a hemoglobin oxygen saturation >95%. PEEP will be set at 5 cmH20. Respiratory rate will be 12 breaths per minute.

Each patient will receive an arterial catheter (radial artery) with a transducer and continuous real-time cardiovascular monitor (LiDCO rapid, LiDCO Ltd, London, UK) attached to it. A pulse oximeter (LNOP Adt, Masimo Corp, Irvine, CA) will be attached to the index finger of the patient's same hand and connected to a bedside monitor (Radical 7, Masimo Corp, Irvine, CA); the oximeter will be wrapped to prevent interference from outside light. PVI calculates the respiratory variations in the plethysmography waveform amplitude (Perfusion index [PI]) using the maximum and minimum PI values over a given time period. Arterial waveform will be recorded continuously.

Intraoperative fluid management will be performed with plasmolyte 2 mL/kg/h. Fluid boluses will be given as guided by SVV. Hetastarch will be given as a bolus in 250 mL increments to a maximum of 50 mL/kg. The fluid bolus will be given over 5 minutes. A bladder (foley) catheter will be inserted after induction of anesthesia. This is part of conventional care and will allow for more accurate fluid volume tracking.

Once the patient is in Trendelenburg position and a measurement yields that SVV or PVI stays above 10%, fluid boluses of 250 mL of hetastarch will be given over 5 minutes until SVV or PVI falls below 10%. The fluid bolus will also be repeated if there is a rise in stroke volume of 10%.

Anesthesia will continue until surgery is completed. Paralysis will be continued throughout the anesthesia period with the goal to maintain a train of four (TOF) at 1 to 2 twitches. At the end of surgery the patient's trachea will be extubated according to routine extubation criteria.

During abdominal CO2 insufflation the intra-abdominal pressure will be kept at 15 cmH2O. The Trendelenburg position will be 30º, which corresponds to the maximum inclination of the OR table.

Measurements

Hemodynamic data (blood pressure, cardiac index, pulse pressure variation) will be measured by pulse power analysis. PVI will be measured by pulse oximetry. Hemodynamic data (MAP, SVV, PPV, CO, PI and PVI) will be recorded at the following intervals: after induction of anesthesia (Baseline), after initiation of pneumoperitoneum, after positioning the patient in Trendelenburg position, and every 15 minutes thereafter. At the end of surgery measurements will be recorded after reversal of Trendelenburg position and after deflation of the peritoneum.

Forearm - fingertip temperature gradients will be measured by Mon-a-thermometer (Mallinckrodt Medical, Inc. St. Louis, MO) and disposable Mon-a-therm thermocouples. Vasoconstriction will be identified when forearm - fingertip temperature gradient is ≥ 00C and PI <1.

Intra-abdominal pressures will be recorded during each measurement of hemodynamic parameters. Trendelenburg position will be measured with an angle ruler.

Data Analysis Sample size analysis: Preliminary data have shown a likelihood of 75% to see an increase of stroke volume after positioning the patient in Trendelenburg position. Taking into consideration the annual patient population of 200 patients and assuming a confidence interval of 65-80% at the 0.05 significance level a total of 35 patients will be needed. Thus, the investigators plan to study 40 patients. An interim analysis will be performed after 20 patients.

The investigators will use paired t-tests and repeated measures ANOVA for all numerical measurements, when comparing data (BP, HR, and SV), before and after interventions (baseline to insufflation, insufflation to Trendelenburg position, and stroke volume before and after fluid bolus). Chi-square tests will be used for categorical data such as change in SVV and PPV. Logistic regression, Bland-Altmann plot and ROC curves will be used to predict the agreement and precision of SVV and PVI changes. Results will be considered statistically significant when P < 0.05.

Significance

The change of hemodynamics during the non-physiologic (iatrogenic) events of abdominal insufflation and placing the patient in steep Trendelenburg position has not been elucidated. Shedding light on hemodynamic derangements may help the clinician to optimize stroke volume and cardiac output early in the process of robotic surgery. The evaluation of a non-invasive device (perfusion variability index) to determine feasibility may allow measuring dynamic hemodynamic parameters non-invasively in the future.

This observational trial does not impose any additional risks on the patient. Arterial cannulation for invasive blood pressure measurements is conventional care during robotic surgery; the applied devices (LiDCO, pulsoximetry and skin thermometers) are non-invasive devices.

Gender and Minority Inclusion

There is no data indicating that ethnicity affects hemodynamic changes during robotic surgery. The investigators accordingly will accept patients from all minority groups. Based on the population of greater Louisville, the investigators expect that 85% of the subjects will be Caucasian. The investigators will include female patients only in the study, as the investigators will only enroll patients for gynecological pathologies.

Potential Risks

The proposed study will be conducted with approval of the Human Studies Committees at the University of Louisville. The investigators must receive written, informed consent from participating patients or authorized representatives. The principal investigator and all collaborators have completed an IRB-certified human subject protection training course.

The investigators will only enroll patients scheduled for robotic surgery undergoing steep Trendelenburg position. The risks of steep Trendelenburg position are swelling of the upper body, specifically eyes, and the larynx and pharyngeal structures. Swelling may cause the risk of delayed extubation or, very seldom, postanesthesia re-intubation. Increased intracerebral pressure may cause headache.

Insufflation of the abdomen may cause subcutaneous emphysema and shoulder pain due to irritation of the phrenic nerve.

Arterial cannulation may rarely cause hematoma at the wrist and arterial thrombosis and infection of the cannulation site.

All potential complications will be explained to the patient twice. First, the patient will have the information presented to them as part of the conventional care consents for surgery and anesthesia. Second, when investigators have a consenting discussion with the potential subjects about the research protocol.