In patients with acute or chronic liver failure, the range of hepatic encephalopathy varies from minimal cerebral functional deficits to coma, and may result in fatal brain damage in severe cases.1,2 Monitors of disturbed cerebral function can include measurements of cerebral perfusion pressure, intracranial pressure, electroencephalography, cerebral blood flow (velocity), and jugular venous oximetry.3
However, monitoring of these variables may be invasive, complicated, or expensive, or require a high degree of professional expertise. Furthermore, such measurements may occasionally be associated with severe bleeding complications, owing to compromised hemostasis in liver failure patients.4 Therefore, a continuous and noninvasive monitoring method that indicates cerebral disturbances arising from hyperemia/ischemia is needed at the bedside of patients with hepatic encephalopathy.5
Near-infrared spectroscopy has been extensively used to investigate the parenchymal and microcirculatory oxygenation of the frontal cerebral cortex in a variety of clinical situations, alerting attending physicians to changes in the critical balance between arterial oxygen delivery and cerebral oxygen consumption.6–9 However, despite the utility of this noninvasive rSO2 method, there are few reports regarding the association between rSO2 values and laboratory variables, such as hyperbilirubinemia, which may cause a falsely low rSO2 reading.10–12 Moreover, the extent to which rSO2 values are affected by abnormal laboratory variables has not yet been fully investigated in end-stage liver disease patients before liver transplantation surgery.11,13
In the present study, we therefore assessed the association between near-infrared spectroscopy rSO2 values and laboratory variables that might affect these measurements. We also sought to determine optimal cutoff values associated with low rSO2 levels in awake end-stage liver disease patients awaiting liver transplantation.
We studied patients with acute or chronic hepatic failure scheduled for liver transplantation between May 2007 and June 2008. Patients were excluded if rSO2 monitoring was unavailable at the time of analysis or had “poor signal quality” on the near-infrared spectroscopy monitor screen. Patients were also excluded if they had sedative medications or preexisting cerebral pathology, such as previous stroke episodes. Patients with West Haven criteria14 grade 0, 1, or 2 hepatic encephalopathy consciousness levels were included. However, patients were excluded if they had a grade 3 or 4 level. After obtaining approval from our IRB and informed consent from all patients, 172 subjects ages 18 to 68 years were assessed in this single-center observational study. In patients undergoing liver transplantation surgery, rSO2 monitoring is one of the routine monitors and our institutional standard. Our IRB waived written patient consent in this observational study; therefore, the rSO2 monitoring procedure was explained orally. The severity of liver disease was evaluated using Child class and Model for End-Stage Liver Disease. In patients with Child class A or B end-stage liver disease, we used laboratory information obtained during routine preoperative evaluation within 1 week (median 2.1 days; range, 0 to 7 days) before the day of surgery, whereas blood samples from patients with Child class C end-stage liver disease were obtained 1 day before or on the morning of surgery for measurements of serum total bilirubin (TB), hemoglobin (Hb), creatinine, albumin, sodium, potassium, magnesium, glucose, and ammonia concentrations, and prothrombin time. Cerebral rSO2 was measured in the operating room before induction of general anesthesia in patients who were breathing room air. None of these end-stage liver disease patients received any sedative or narcotic premedications before baseline rSO2 measurement.
Cerebral rSO2 Measurement
After a patient was placed supine on the operation table, cerebral rSO2 was measured using an INVOS 5100B cerebral oximeter (operational range, 15% to 95% oxygen saturation; Somanetics, Troy, MI). Two disposable cerebral oximetry sensors were placed bilaterally on the right and left sides of the forehead according to the manufacturer's instructions. The 5100B cerebral oximeter generates low-intensity, near-infrared light and directs the light, at wavelengths of 730 and 805 nm, onto the patient's forehead; the light penetrates the skull and passes through the cerebral cortex. The sensor comprises a light-emitting diode and 2 detectors located 3.0 cm and 4.0 cm from the diode, respectively, intended to remove the extracranial contribution of scattered light by the application of a subtraction algorithm. The proximal detector receives a signal from the shallow part of the brain, whereas the distal detector measures the saturation of all adjacent tissues, including the skin, muscle, skull, and brain.
Cerebral rSO2 data were collected automatically from the digital output port of the monitor and were continuously recorded at 5- to 6-second intervals on a computer hard disk, for later off-line analysis. Cerebral rSO2 was defined as the average saturation of both sides of the forehead over a 1-minute period, from 12 rSO2 values recorded at 5- to 6-second intervals for each hemisphere. Measurements commenced approximately 3 minutes after the sensors were applied.
Continuous variables were compared using independent t tests or Mann–Whitney U test. The relationships between mean rSO2 and laboratory variables were assessed by Pearson or Spearman correlation analyses as appropriate. P values of <0.05 were considered statistically significant. Multivariate analysis by forward logistic regression was performed to identify independent factors observing rSO2 <50%.
Receiver operating characteristic (ROC) curves were constructed to determine the optimal cutoff values of laboratory variables for rSO2 <50%, with the optimal cutoff defined as that with the highest product of sensitivity and specificity. Statistical analyses were performed using SPSS software version 12.0 (SPSS Inc., Chicago, IL) and MEDCALC (MedCalc software, Mariakerke, Belgium).
Table 1 shows the clinical characteristics of the subjects and their measured laboratory values. Of our 172 patients, 129 (75%) had rSO2 ≥50%, and 43 (25%) had rSO2 <50%, in whom 162 patients (94%) belonged to West Haven criteria grade 0 and 1 hepatic encephalopathy. One hundred thirty-four patients (78%) had hepatitis B virus–related cirrhosis, and others had alcoholic cirrhosis (n = 12), fulminant hepatic failure (n = 8), hepatitis C virus–related cirrhosis (n = 6), cryptogenic cirrhosis (n = 5), autoimmune hepatitis (n = 3), Budd–Chiari syndrome (n = 2), and primary biliary cirrhosis (n = 2). There was no significant difference between left and right forehead rSO2 values (54.9% ± 13.9% vs. 55.9% ± 13.7%; 95% confidence interval [CI] for the difference = 0.75 to 0.99), and they were highly correlated (r = 0.94, P < 0.001). We observed significant correlations between rSO2 level and serum concentrations of TB, Hb, creatinine, sodium, and magnesium, prothrombin time (P < 0.001 each), and heart rate (P = 0.002, Table 1). Only variables with P < 0.05 were then entered into the final model of multivariable analysis. Multiple logistic regression analysis showed that the significant and independent parameters of observing rSO2 <50% were increased TB concentration (median 2.9 mg/dL; range 0.4 to 66 mg/dL; odds ratio = 1.31; 95% CI = 1.18 to 1.45) and low Hb (median 10.6 g/dL; range 5.3 to 15.7 g/dL; odds ratio = 0.21; 95% CI = 0.11 to 0.43) (Nagelkerke R2 = 0.82, Table 2).
Scattergrams showing the relationships between rSO2 and TB or Hb concentration are shown in Figure 1. We observed a significant but only moderate negative correlation between rSO2 and TB concentration (r = −0.76, P < 0.001) and a significant but only modest positive correlation between rSO2 and Hb concentration (r = 0.52, P < 0.001).
We found that the optimal cutoff points for observing rSO2 <50% were TB >7.2 mg/dL (sensitivity 89%, specificity 90%) and Hb <9.6 g/dL (sensitivity 70%, specificity 82%) (Fig. 2). The area under the ROC curve for TB was significantly superior to that for Hb (0.93 vs. 0.81; P = 0.002) (Fig. 3).
We found that high TB level and low Hb concentration were both significantly and independently associated with rSO2 <50% in end-stage liver disease patients with West Haven criteria grade 0, 1, and 2 hepatic encephalopathy, and that rSO2 appeared to be related to Hb concentration even in patients with high TB concentration. The optimal cutoff points for observing an rSO2 <50% were TB >7.2 mg/dL and Hb <9.6 g/dL.
Light of near-infrared wavelengths readily penetrates skin and bone, but very few biological substances absorb such radiation. Hb and cytochrome aa3 are the only substances (chromophores) that show detectable changes in near-infrared absorption in response to hypoxia or ischemia in normal patients.15 However, patients with severe jaundice may have high enough serum bilirubin concentrations to interfere with near-infrared spectroscopy.11,13 We observed a negative correlation between rSO2 level and TB concentration (r = −0.76, P < 0.001), similar to that observed in a previous study of interference by icterus on cerebral oximetry (r = −0.72).11 The systematic influence of bilirubin on rSO2 measurements is most likely due to its competitive absorption of light, because bilirubin does not influence either cerebral blood flow or oxygen uptake.11 Differences in the impact of bilirubin on oxygen readings of cerebral oximetry in comparison with pulse oximeter oxygen saturation [SpO2] and mixed venous oxygen saturation [SvO2] have shown that TB affects rSO2 but not SpO2 or SvO2.11,16 Tissue pigmentation deposits (e.g., biliverdin) and hyperbilirubinemia (free or conjugated) may also interfere with measurements of rSO2.11,17 The design of this study, however, could not distinguish among these differences, indicating that additional studies are needed to address these issues.
The rSO2 may be affected by Hb concentration,18–20 differential optical path length,17 extracranial blood contamination,21,22 skull thickness, and area of the cerebrospinal fluid layer.20 Optical path length, which measures how far near-infrared light can travel through tissue, is important in estimating rSO2. There is a negative correlation between optical path length and Hb concentration.23 When the total amount of absorbing Hb within the brain is low, the number of detected photons may be increased; these photons may travel great distances through the brain, resulting in an increased optical path length.24 The modified Lambert–Beer formula to calculate rSO2 includes optical path length as a constant. As the optical path length changes with changes in Hb concentration, Hb can cause measurement error in rSO2. In the present study, we found that high Hb concentrations were correlated with high rSO2 values, indicating that Hb concentration is also an important determinant of rSO2 values even in patients with hyperbilirubinemia. However, reduced Hb level may have a real impact on rSO2 and may not be due simply to interference, because cerebral oxygen delivery depends on Hb, cardiac output, and arterial oxygen saturation. Thus, low rSO2 may reflect a true reduction in cerebral oxygenation, although hyperbilirubinemia may cause a false reduction in measured rSO2 due to absorption of relevant wavelengths. However, the effects of Hb concentration on optical path length in highly icteric patients have not been determined. Further studies are required to address this issue in patients with hyperbilirubinemia.
Our study had several limitations. First, laboratory information and rSO2 values were not obtained simultaneously in Child class A and B patients. Thus, laboratory values may have changed in the interval from laboratory sampling to measurement of rSO2. We expect, however, that any such differences would have been small because these patients (35% of our study population) were in a chronic steady state for at least 1 week before surgery. Indeed, our pilot study in Child class A and B patients revealed that 95% CIs for the difference of TB and Hb values obtained from blood samples at median 3-day intervals (range of 1 to 7 days) were 0.14 to 0.18 mg/dL (2.0 ± 1.2 mg/dL vs. 2.1 ± 1.2 mg/dL, P = 0.823) and 0.24 to 0.32 g/dL (11.9 ± 2.2 g/dL vs. 11.9 ± 2.1 g/dL, P = 0.983), respectively (data not shown in the RESULTS section). Furthermore, the correlation coefficient between rSO2 and TB measured in the present study was similar to that of a previous study (−0.76 vs. −0.72).11 Second, our study population did not represent all end-stage liver disease patients for whom liver transplantation was planned. We had to exclude patients with West Haven criteria grade 3 or 4 hepatic encephalopathy to allow discrimination between patients suffering from true cerebral disturbances and those with low rSO2 values arising from abnormal laboratory measurements. Moreover, we did not monitor patients with low rSO2 to detect changes in cerebral oxygen delivery. Another limitation was that we did not evaluate the effects of other possible factors—such as skull thickness, cerebrospinal fluid volume, and extent of extracranial components—on rSO2 values.15,18,20 Specifically, we only included mean arterial blood pressure and pulse pressure, and did not measure cerebral blood flow which may affect baseline rSO2 values.
In conclusion, we have shown here that a bedside monitor of rSO2 may give false low values when TB is elevated above normal or when Hb concentrations are lower than normal in end-stage liver disease patients with West Haven criteria grade 0, 1, or 2 hepatic encephalopathy. The results of this study identify patients in whom a low rSO2 may be an artifact rather than a warning of cerebral ischemia. However, because we did not measure cerebral oxygenation directly, further study may be required to ascertain our findings.
Name: Jun-Gol Song, MD, PhD.
Contribution: This author participated in the conduct of the study and manuscript preparation.
Name: Sung-Moon Jeong, MD, PhD.
Contribution: This author participated in the conduct of the study.
Name: Won-Jung Shin, MD.
Contribution: This author participated in the conduct of the study.
Name: In-Gu Jun, MD.
Contribution: This author participated in data analysis.
Name: Kyoon Shin, MD.
Contribution: This author participated in data analysis.
Name: In-Young Huh, MD, PhD.
Contribution: This author participated in data analysis.
Name: Young-Kug Kim, MD, PhD.
Contribution: This author participated in study design.
Name: Gyu-Sam Hwang, MD, PhD.
Contribution: This author participated in study design and manuscript preparation.
1. Ferenci P, Herneth A, Puspok A, Steindl P. Neuropharmacologic modulation of hepatic encephalopathy: experimental and clinical data. Adv Exp Med Biol 1994;368:103–12
2. Jones EA. Pathogenesis of hepatic encephalopathy. Clin Liver Dis 2000;4:467–85
3. Aggarwal S, Obrist W, Yonas H, Kramer D, Kang Y, Scott V, Planinsic R. Cerebral hemodynamic and metabolic profiles in fulminant hepatic failure: relationship to outcome. Liver Transpl 2005;11:1353–60
4. Blei AT, Olafsson S, Webster S, Levy R. Complications of intracranial pressure monitoring in fulminant hepatic failure. Lancet 1993;341:157–8
5. Nielsen HB, Tofteng F, Wang LP, Larsen FS. Cerebral oxygenation determined by near-infrared spectrophotometry in patients with fulminant hepatic failure. J Hepatol 2003;38:188–92
6. Calderon-Arnulphi M, Alaraj A, Amin-Hanjani S, Mantulin WW, Polzonetti CM, Gratton E, Charbel FT. Detection of cerebral ischemia in neurovascular surgery using quantitative frequency-domain near-infrared spectroscopy. J Neurosurg 2007;106:283–90
7. Samra SK, Dy EA, Welch K, Dorje P, Zelenock GB, Stanley JC. Evaluation of a cerebral oximeter as a monitor of cerebral ischemia during carotid endarterectomy. Anesthesiology 2000;93:964–70
8. McCusker K, Chalafant A, de Foe G, Gunaydin S, Vijay V. Influence of hematocrit and pump prime on cerebral oxygen saturation in on-pump coronary revascularization. Perfusion 2006;21:149–55
9. Nishiyama T. Recent advance in patient monitoring. Korean J Anesthesiol 2010;59:144–59
10. Plachky J, Hofer S, Volkmann M, Martin E, Bardenheuer HJ, Weigand MA. Regional cerebral oxygen saturation is a sensitive marker of cerebral hypoperfusion during orthotopic liver transplantation. Anesth Analg 2004;99:344–9
11. Madsen PL, Skak C, Rasmussen A, Secher NH. Interference of cerebral near-infrared oximetry in patients with icterus. Anesth Analg 2000;90:489–93
12. Skak C, Rasmussen A, Kirkegaard P, Secher NH. Cerebral oxygen saturation and blood flow during liver transplantation. Anesth Analg 1997;84:730–3
13. Nielsen HB, Larsen FS. Cerebral oximetry and hyperbilirubinemia. Anesth Analg 2001;93:245–6
14. Ferenci P, Lockwood A, Mullen K, Tarter R, Weissenborn K, Blei AT. Hepatic encephalopathy—definition, nomenclature, diagnosis, and quantification: final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, 1998. Hepatology 2002;35:716–21
15. Smythe PR, Samra SK. Monitors of cerebral oxygenation. Anesthesiol Clin North America 2002;20:293–313
16. Veyckemans F, Baele P, Guillaume JE, Willems E, Robert A, Clerbaux T. Hyperbilirubinemia does not interfere with hemoglobin saturation measured by pulse oximetry. Anesthesiology 1989;70:118–22
17. Owen-Reece H, Smith M, Elwell CE, Goldstone JC. Near infrared spectroscopy. Br J Anaesth 1999;82:418–26
18. Kishi K, Kawaguchi M, Yoshitani K, Nagahata T, Furuya H. Influence of patient variables and sensor location on regional cerebral oxygen saturation measured by INVOS 4100 near-infrared spectrophotometers. J Neurosurg Anesthesiol 2003;15: 302–6
19. Yoshitani K, Kawaguchi M, Iwata M, Sasaoka N, Inoue S, Kurumatani N, Furuya H. Comparison of changes in jugular venous bulb oxygen saturation and cerebral oxygen saturation during variations of haemoglobin concentration under propofol and sevoflurane anaesthesia. Br J Anaesth 2005;94:341–6
20. Yoshitani K, Kawaguchi M, Miura N, Okuno T, Kanoda T, Ohnishi Y, Kuro M. Effects of hemoglobin concentration, skull thickness, and the area of the cerebrospinal fluid layer on near-infrared spectroscopy measurements. Anesthesiology 2007;106:458–62
21. Germon TJ, Kane NM, Manara AR, Nelson RJ. Near-infrared spectroscopy in adults: effects of extracranial ischaemia and intracranial hypoxia on estimation of cerebral oxygenation. Br J Anaesth 1994;73:503–6
22. Germon TJ, Young AE, Manara AR, Nelson RJ. Extracerebral absorption of near infrared light influences the detection of increased cerebral oxygenation monitored by near infrared spectroscopy. J Neurol Neurosurg Psychiatry 1995;58:477–9
23. Kurth CD, Uher B. Cerebral hemoglobin and optical pathlength influence near-infrared spectroscopy measurement of cerebral oxygen saturation. Anesth Analg 1997;84:1297–305
© 2011 International Anesthesia Research Society
24. Duncan A, Meek JH, Clemence M, Elwell CE, Fallon P, Tyszczuk L, Cope M, Delpy DT. Measurement of cranial optical path length as a function of age using phase resolved near infrared spectroscopy. Pediatr Res 1996;39:889–94