Validation of the Bispectral Index Monitor During Living Donor Liver Transplantation

Forty two American Society of Anesthesiologists physical class III-IV patients aged 20-50 years with severe liver dysfunction (Child-Pugh Class C) due to liver cirrhosis undergoing living donor liver transplantation in the Mansoura University Liver Transplantation Program from 2007 to 2010 were included in this prospective, observational study after obtaining approval of the local ethical committee and an informed written consent from all participants. The etiology of liver cirrhosis was hepatitis C, Budd Chiari syndrome, hepatocellular carcinoma and autoimmune in all patients.

Anesthetic technique was standardized for all patients. All patients received preoperative intravenous 40 mg of pantoprazole. Patients were monitored with three leads electrocardiography, pulse oximeter, and noninvasive blood pressure. BIS recording electrodes (Aspect Medical Systems. Inc., One Upland Road, Norwood, MA 02062 USA) were applied to the forehead of each patient according to the manufacturer recommendations. The BIS monitor was masked with opaque sheet and BIS data was recorded prior to induction of general anesthesia then continued throughout the procedure. All information obtained from the BIS monitor was continuously downloaded to a computer for offline analysis. The patient's management was not guided by the changes in BIS readings. Before induction of anesthesia, all patients were premedicated with 1-2 mg of intravenous midazolam.

Independent anesthetists, who were not involved in the collection and analysis of the patient's data, provided the perioperative anesthetic management. After preoxygenation, anesthesia was induced with lidocaine 0.75 mg/kg, propofol 1.5-2.0 mg/kg and fentanyl 2µg/kg. Rocuronium 1-1.2 mg/kg was given to suppress the first response on the train-of-four (TOF) stimulations of the ulnar nerve. After tracheal intubation, the lungs were ventilated with an inspired fraction of oxygen (FiO2) of 0.4 to maintain an arterial carbon dioxide tension at 35-45 mm Hg and monitoring of end-tidal carbon dioxide and sevoflurane concentration, direct arterial pressure monitoring through a radial artery catheter and a tympanic membrane temperature were implemented.

A 7.5 Fr continuous thermodilution fiberoptic pulmonary artery catheter (CCO/SvO2) Edwards Life Science, Irvine, CA, USA) was floated through the right internal jugular vein to the right pulmonary artery using waveform and fluoroscopic guidance to measure cardiac output using Angstrom AS5 Monitor (Datex - Ohmeda AS5, Microvitec Display LTD, Bolling Road, Bradford, UK). The final position was confirmed by fluoroscopy and when PAOP less than pulmonary artery diastolic pressure (PADP).

Anesthesia was maintained with 0.5-1.5 minimum alveolar concentration (MAC) of sevoflurane and continuous intravenous infusion of fentanyl (1-3 µg/kg/h) titrated to maintain the mean arterial blood pressure (MAP) and heart rate (HR) within 20% of their baseline values. Rocuronium 0.1-0.3 mg/kg/h was used to maintain suppression of the second twitch in the TOF.

Normothermia was maintained using intravenous infusions of warm fluid and blood products and water-filled thermal mattress according to the authors' protocol. Fluid boluses were given in 250 ml aliquots of either 5% hydroxyethyl Starch 130/0.4 (Voluven® 6%, Fresenius Kabi, Bad Hombourg, Germany) as needed to maintain the CVP and/or PAOP between 5 and 7 mm Hg and the stroke volume (SV) within 20% of its baseline value. Albumin 5% was given as needed to treat hypoalbuminemia. Transfusion of salvaged blood from the cell saver and packed red blood cells was administered as clinically indicated to maintain a hemoglobin level ≥7 g/dL. Intraoperative changes in electrolyte levels and acid-base balance were regularly monitored and treated as appropriate. Hemodynamic control was standardized according to the authors' center protocol. Hypotension (MAP decreased \< 20% from the mean baseline and SVR \< 600 dyne.sec-1.cm-5) was treated with boluses of fluids, ephedrine 5 mg, or epinephrine 5 µg, as needed. Norepinephrine infusion was used for persistent hypotension with low SVR. Dobutamine or epinephrine infusion was administered if the MAP was ≥ 70 mm Hg, CVP and/or PAOP ≥ 7 mm Hg and the cardiac output was \< 4.0 L. min-1. Hypertension (MAP increased > 20% from the mean baseline) was treated with deepening of anesthesia, bolus doses of nitroglycerin 0.05 mg or labetalol 20 mg. Tachycardia (HR > 20% from the baseline values) was treated with boluses of esmolol 20 mg.

All operations were performed by the same surgeons using the piggyback technique. After portal vein clamping, infusion rates of fentanyl and rocuronium were reduced by approximately 50%. Fentanyl and rocuronium infusions were discontinued after peritoneal closure. After completion of skin closure, sevoflurane was discontinued. The patients were transferred to the ICU immediately after surgery, while intubated and postoperative analgesia was achieved with rescue boluses of fentanyl 0.5 µg/kg when needed.

Extubation criteria included cooperative, alertness, train-of-four ratio ≥ 0.9, spontaneous breathing with tidal volume > 5 mL/kg, respiratory rate > 10 and \< 28 breaths/min, maximum inspiratory pressure ≤ -20 cm H2O, stable hemodynamics, minimal bleeding, core temperature > 35.5°C, urine output > 0.5 mL/Kg/h, arterial carbon dioxide tension ≤ 45 mm Hg, arterial oxygen tension > 100 mm Hg and FiO2 \< 50%.

On the second postoperative day, the patients were asked about perioperative awareness and recall by asking three simple questions using standard interview "What was the last thing you remembered happening before you went to sleep? What is the first thing you remember happening on waking? Did you dream or have any other experiences whilst you were asleep?" Another investigator, who was not involved in the patient's management, collected the saved patients' data. Preoperative MELD score, durations of warm ischemia and anhepatic phase, intraoperative blood loss and fentanyl consumption were collected. Preoperatively (baseline) absolute values and intraoperative average values for BIS, Et-Sevo, heart rate (HR), MAP, CVP, MPAP, CO, PAOP, SVR, PVR, core temperature, pH, PaCO2, PaO2, SaO2 SvO2, and PvO2 were recorded during hepatic dissection, anhepatic, and neohepatic phases and during surgical closure. Postoperative total bilirubin, albumin level, liver enzymes (AST, ALT, GGT, and alkaline phosphatase), PT, factors V and VII, time to extubation, ICU and hospital length of stays, neurological complication, encephalopathy and 3 months-mortality were recorded.

Data were tested for normality using the Kolmogorov-Smirnov test. Repeated-measures analysis of variance was used for analysis of serial changes in the patients' data at different times. Fisher exact test was used for categorical data. Postulated independent predictors of BIS readings included age, gender, MELD score, durations of surgical phases and warm ischemia, Et-Sevo, average of intraoperative hemodynamic, temperature, oxygenation and acid base variables, intraoperative blood loss and fentanyl consumption, were examined in a stepwise manner into a multiple regression model, with entry and retention set at a significance level of P \< 0.05 and removal set at P ≥ 0.1. Moreover, multivariate logistic regression was done to identify the correlations between BIS values and blood loss, time to tracheal extubation, postoperative liver function and coagulation variables, ICU length of stay, and 3 months-mortality.

To evaluate the accuracy of BIS in predicting survival rate after liver transplantation, receiver operating characteristic (ROC) curves were generated to describe the performance characteristics of the BIS readings during different phases of surgery in predicting the survival outcome. An ROC area of 1.0 is characteristic of an ideal model, whereas an area of 0.5 indicates a model of no diagnostic value. Cutoff values for mortality with the best diagnostic sensitivity and specificity were derived from the ROC curve analyses. Unweighted accuracy ((sensitivity + specificity) / 2) was also reported.

The studied patients were further divided according to their outcome variables into survivors (n = 32) and non-survivors (n = 10) subgroups. Independent Student-t- and Mann Whitney tests were used as appropriate. Data were expressed as mean (standard deviation), number (%), or median [range]. A value of P \< 0.05 was considered to represent statistical significance.