Indocyanine green clearance test combined with MELD score in predicting the short-term prognosis of patients with acute liver failure

Hong-Ling Feng, Qian Li, Lin Wang, Gui-Yu Yuan and Wu-Kui Cao

Tianjin, China

Indocyanine green clearance test combined with MELD score in predicting the short-term prognosis of patients with acute liver failure

Hong-Ling Feng, Qian Li, Lin Wang, Gui-Yu Yuan and Wu-Kui Cao

Tianjin, China

BACKGROUND:Acute liver failure (ALF) is an acute severe deterioration of liver function with high mortality. Early and accurate prognostic assessment of patients with ALF is critically important. Although the model for end-stage liver disease (MELD) scores and King's College Hospital (KCH) criteria are well-accepted as predictive tools, their accuracy is unsatisfactory. The indocyanine green (ICG) clearance test (ICGR15, ICG retention rate at the 15 minutes) is a sensitive indicator of liver function. In this study, we investigated the efficacy of the ICGR15 for the short-term prognosis in patients with ALF. We compared the predictive value of ICGR15 with the MELD scores and KCH criteria.

METHODS:Sixty-nine patients who had been diagnosed with ALF were recruited retrospectively. ICGR15 had been performed by ICG pulse spectrophotometry and relevant clinical and laboratory indices were analyzed within 24 hours of diagnosis. In addition, the MELD scores and KCH criteria were calculated.RESULTS:The three-month mortality of all patients was 47.83%. Age, serum total bilirubin and creatinine concentrations, international normalized ratio for prothrombin time, ICGR15, MELD scores and KCH criteria differed significantly between surviving and deceased patients. A positive correlation was observed between ICGR15 and MELD scores (r=0.328,P=0.006). The ICGR15-MELD model, Logit(P)=0.096×ICGR15+0.174 × MELD score–9.346, was constructed by logistic regression analysis. The area under the receiver operating characteristic curve was 0.855. When set the cut-off point to -0.4684, the sensitivity was 87.90% and specificity, 72.20%. The area under the receiver operating characteristic curve of the ICGR15-MELD model (0.855) was significantly higher than that of the ICGR15 (0.793), MELD scores (0.776) and KCH criteria (0.659).Based on this cut-off value, the patients were divided into two groups. The mortality was 74.36% in the first group (ICGR15-MELD≥-0.4686) and 13.33% in the second group (ICGR15-MELD＜-0.4686), with a significant difference between the two groups (χ2=25.307,P=0.000).

CONCLUSION:The ICGR15-MELD model is superior to the ICGR15, MELD scores, and KCH criteria in predicting the shortterm prognosis of patients with ALF.

(Hepatobiliary Pancreat Dis Int 2014;13:271-275)

acute liver failure;

indocyanine green clearance test;

model for end-stage liver disease;

prognosis

Introduction

Acute liver failure (ALF) is a rapidly progressive disease with extremely high mortality.[1-4]Accurate assessment of the severity of disease is very important when making treatment decisions, such as medication versus liver transplantation, in patients with ALF. Although the model for end-stage liver disease (MELD) scores and King's College Hospital (KCH) criteria are well-accepted as predictive tools,[5-8]their predictive accuracy is unsatisfactory.[9]Therefore, there is an urgent need to develop a convenient, objective, and quantitative method to predict the short-term prognosis in patients with ALF.

Indocyanine green (ICG) clearance test, which accurately evaluates liver function and the number of functioning hepatocytes in real-time, is minimally invasive and can be performed at the bedside, and is regarded as the most reliable method for liver function assessment.[10-12]This test also quantitatively estimates the hepatic functional reserve, thus providing crucial information during the perioperative stage of liver surgery.[13-16]However, there are few reports concerning the clinical significance of the ICG clearance test inALF. In this study, we used the ICG retention rate at the 15 minutes (ICGR15) to assess the short-term (threemonth) prognosis in patients with ALF. To analyze its clinical relevance, we compared the predictive value of ICGR15 with MELD scores and KCH criteria.

Methods

Patients

This retrospective study included 69 patients with ALF from October 2010 to August 2012 in our hospital. They were 59 males and 10 females aged from 17 to 76 years (45.93 ±14.25). The causes of ALF were hepatitis B in 52 patients, hepatitis C in 1, hepatitis E in 4, hepatitis B and E in 1, autoimmune liver disease in 1, alcoholic hepatitis in 2, drug-induced hepatitis in 1, and unknown reason in 7. The diagnosis of ALF was made according to the American Association for the Study of Liver Diseases Position Paper; The Management of Acute Liver Failure: Update 2011.[17]Three months after the diagnosis, the patients were followed up by telephone. Thirty-three patients died within 3 months and 36 survived beyond this time.

Treatment

All patients had received basic treatment of ALF with nutritional support, hepatocyte growth factor, and glutathione for hepatocyte proliferation and restoration of liver function. They also received other therapies including prevention of infection, correction of anomalies in the coagulation system, and improvement of cerebral edema. In addition, plasmapheresis was performed when necessary (2600-3000 mL each time, PlasmaFlux P2 dry).

Outcomes

During the first 24 hours after the diagnosis of ALF was made, the following data were collected: time of onset of hepatic encephalopathy and hepatic, renal, and coagulation function. Death or survival of patients at three months was recorded.

Prediction models

MELD scores were calculated as follows: MELD=3.8 × Ln(TBil[mg/dL])+11.2×Ln(INR)+9.6×Ln(Cr[mg/dL])+6.4× cause, where TBil is serum total bilirubin concentration, INR is international normalized ratio for prothrombin time, and Cr is serum creatinine concentration. When ALF was due to cholestasis or alcohol, the cause was counted for 0; other causes counted for 1. The final results were truncated as integers.[18]

The KCH criteria included INR ＞6.5 or any three of the following criteria: 1) age ＜10 or ＞40 years; 2) non-A/non-B hepatitis, drug-induced hepatitis, and halothane hepatitis; 3) time from jaundice to hepatic encephalopathy ＞7 days; 4) TBil ＞300 μmol/L (17 mg/ dL); and 5) INR ＞3.5, which suggests a poor prognosis.[19]

ICG clearance test

ICG (0.5 mg/kg; Jishi Pharmaceutical, Shenyang, China) was injected intravenously in 5-10 seconds through the median cubital vein. ICGR15 was determined by pulse spectrophotometry (DDG-3300K, Japan).

Statistical analysis

Data were analyzed using SPSS 20.0 (IBM, USA). Measurement data were analyzed with Student'sttest and enumeration data with the Chi-square test. Correlation analysis was performed with Pearson's product-moment correlation coefficient test. APvalue less than 0.05 was considered statistically significant. Logistic regression analysis was used to establish the prognosis model. Area under the receiver operating characteristic (AUROC) curve was used to assess the ability of the newly developed prognosis model, ICGR15, MELD scores, and KCH criteria to predict the shortterm prognosis of ALF patients.

Results

Comparison between deceased and surviving patients

At three months after the diagnosis of ALF, 33 (47.83%) patients died and 36 (52.17%) survived, of whom 20 (60.61%) and 28 (77.78%), respectively, had received plasmapheresis. There was no significant difference between deceased and surviving patients with regard to plasmapheresis treatment (P＞0.05). No patient in this study had received a liver transplantation.

There were significant differences between deceased and surviving patients in age, TBil, Cr, INR, ICGR15, and MELD scores (allP＜0.05). Significantly fewer surviving (6, 16.67%) than deceased (16, 48.48%) patients had met the KCH criteria (P=0.005) (Table 1).

Correlation analysis between ICGR15 and MELD scores

The correlation between ICGR15 and MELD scores in ALF patients was statistically significant (r=0.328,P=0.006).

Establishment of the new prognosis model

ICGR15 and MELD scores were subjected to logisticregression analysis, the forward logistic regression method being used to establish the ICGR15-MELD model. The prognosis (Logit(P)) equations are as follows:

Logit(P)=Ln[P/(1–P)];P=1/(1+e-Logit(P))

Logit(P)=0.096×ICGR15+0.174×MELD score–9.346

Patients withP＞0.5 were judged as deaths and those withP＜0.5 as survivors; or if Logit(P)＞0, the patients were judged as deaths, while if Logit(P)＜0, the patients were judged as survivors. The AUROC curve was 0.855 when this equation was used in the judgment, with a sensitivity of 87.90% and a specificity of 72.20%. This model was statistically significant (χ2=31.021,P=0.000). Goodness of fit test was performed with the Hosmer-Lemeshow model (χ2=1.928,P=0.983). The regression coefficients of the ICGR15-MELD model were statistically significant (Table 2).

Table 1.Comparison of clinical characteristics between patients with different outcomes

Table 2.The regression coefficients and indices of the ICGR15-MELD model

Table 3.Prognostic values of the various models

Prognostic values of the various models

The AUROCs curve of the various models are in the following order: ICGR15-MELD model＞ICGR15＞MELD scores＞KCH criteria (Table 3, Fig.). According to the results, the cut-off value of the ICGR15-MELD model was -0.4686. Based on this cut-off value, the data were analyzed in two groups. The first group (ICGR15-MELD≥-0.4686) included 39 patients, of whom 29 (74.36%) died. The second group (ICGR15-MELD＜-0.4686) included 30 patients, of whom 4 (13.33%) died, with a significant difference between the two groups (χ2=25.307,P=0.000) (Table 4).

Fig.AUROC curves of the ICGR15-MELD model, ICGR15, MELD scores and KCH criteria.

Table 4.Comparison of mortality based on the cut-off value of the ICGR15-MELD model

Discussion

ALF is a severe condition with a high mortality. Its prognostic factors have been extensively studied and various scoring systems are suggested such as the KCH criteria and MELD scores. The causes of ALF vary between countries. In the developed countries,acetaminophen-induced hepatic toxicity is the primary cause,[20]whereas in China the major cause is viral hepatitis.[21]Therefore, the predictive value of established scoring systems in Chinese patients remains controversial.

The KCH criteria have been widely used to assess the severity of ALF and the timing of liver transplantation.[6-8,19]In this study, the performance of these criteria was very poor (AUROC of 0.659), suggesting that they may not be appropriate for Chinese patients. Our finding is consistent with a previous report.[22]Recently, much evidence has shown that the fulfillment of the KCH criteria is highly predictive of poor outcomes in patients with non-acetaminophen ALF. However, the KCH criteria does not accurately predict survival.[7]In this study, 17 of the 47 patients who did not meet the KCH criteria died within three months, suggesting that these criteria are too stringent for Chinese patients with ALF.

In 2000, Malinchoc et al[23]introduced the MELD for evaluating the survival rate of patients with liver cirrhosis treated with transjugular intrahepatic portosystemic shunt procedures. The United Network of Organ Sharing adopted the MELD in 2002 as sorting criteria for recipients awaiting liver transplantation.[24]The use of the MELD has since extended to patients with liver cirrhosis, liver cancer, and liver failure.[5,6,8,18]In our study, the MELD scores of the deceased patients were significantly higher than those of the survivors, with an AUROC of 0.776, demonstrating the strong performance of this model in prediction of short-term survival in Chinese patients with ALF. There were significant differences between deceased patients and survivors in some components of the MELD model that significantly affected the final scores, such as TBil, Cr, and INR. However, cholestatic and alcoholic causes score 0 and other causes score 1 in the MELD scores, which is an unimportant consideration in Chinese ALF patients who predominantly have viral hepatitis. In addition, low blood volume and excessive diuretics increase Cr, decrease absorption of vitamin K; cholestasis elevates INR; malnutrition and infection increase TBil. Such extrahepatic factors may influence the performance of the MELD model in the prediction of liver diseases.[18]Therefore, the MELD model also has limitations in the prediction of ALF prognosis.

Because ICGR15 reflects three phases of ICG metabolism in the liver, namely uptake, processing, and excretion, it reflects the integrity and function of hepatocytes. In recent years, the development of pulse spectrophotometry has facilitated the clinical use of ICGR15 because this test can now be performed at the bedside, in real-time, and is minimally invasive.[11,12]Merle et al[25]found that ICG clearance test predicts the prognosis of ALF patients with an AUROC of 0.90, a sensitivity of 85.7%, and a specificity of 88.9%. In our study, ICGR15 was significantly higher in deceased than in surviving patients and performed significantly better than the MELD model and KCH criteria (AUROC of 0.793, 0.776, and 0.659, respectively), suggesting that ICGR15 is preferable for predicting ALF prognosis. However, because ICGR15 is sensitive to hepatic blood flow and bile secretion, it needs some modification by other prediction models or variables.

Multivariate analysis has been extensively used in recent years. The KCH criteria are limited in prediction capability, complicated to calculate, and based on subjective variables. In addition, the KCH criteria and MELD model have several components in common, including the cause of ALF, TBil, and INR. Therefore, we excluded the KCH criteria from the logistic regression analysis. In this study, we found a positive correlation between the ICGR15 and MELD scores. ICGR15 reflects the hepatic functional reserve and MELD scores mirror pathophysiological changes in severely impaired liver function, such as hyperbilirubinemia, renal failure, and coagulation dysfunction. Therefore, a combination of ICGR15 and MELD scores can improve the accuracy of prediction of ALF prognosis. We created the ICGR15-MELD model by subjecting ICGR15 and MELD scores to logistic regression analysis, and found that this model is better to predict ALF prognosis than ICGR15 or the MELD model alone, with an AUROC of 0.855. If the cut-off value is set to -0.4686, the sensitivity is 87.90%, and the specificity, 72.20%. The patients were divided into two groups according to the cut-off value. The mortality was 74.36% in the first group (ICGR15-MELD ≥-0.4686) and 13.33% in the second group (ICGR15-MELD＜-0.4686), with a significant difference between the two groups. For those whose score are higher than -0.4686 should arrange liver transplantation. The ICGR15-MELD model, comprising objective and quantitative variables and reflecting the severity of the illness, is preferable for clinical practice.

We conclude that the combination of ICGR15 and MELD scores are better than the KCH criteria or ICGR15 or MELD scores alone in predicting ALF prognosis. However, ALF is diverse in onset, etiology, clinical type, disease course, complications, and treatment methods. Therefore, a prospective, multicenter, large study with the same standards in diagnosis and data collection criteria is necessary to further validate our conclusion.

Acknowledgment:We thank Professor Wei Hou (Tianjin Second People's Hospital) to help to improve the literary beauty of thelanguage.

Contributors:FHL proposed the study. FHL and LQ wrote the first draft and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. LQ is the guarantor.

Funding:The study was supported by a grant from the Foundation of the Ministry of Health, China (2008ZX1005).

Ethical approval:Not needed.

Competing interest:No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

1 Wojcicki M, Lubikowski J, Wrzesinski M, Post M, Jarosz K, Hydzik P, et al. Outcome of emergency liver transplantation including mortality on the waiting list: a single-center experience. Transplant Proc 2007;39:2781-2784.

2 Wei G, Bergquist A, Broomé U, Lindgren S, Wallerstedt S, Almer S, et al. Acute liver failure in Sweden: etiology and outcome. J Intern Med 2007;262:393-401.

3 Liou IW, Larson AM. Role of liver transplantation in acute liver failure. Semin Liver Dis 2008;28:201-209.

4 Taylor RM, Tujios S, Jinjuvadia K, Davern T, Shaikh OS, Han S, et al. Short and long-term outcomes in patients with acute liver failure due to ischemic hepatitis. Dig Dis Sci 2012;57: 777-785.

5 Katoonizadeh A, Decaestecker J, Wilmer A, Aerts R, Verslype C, Vansteenbergen W, et al. MELD score to predict outcome in adult patients with non-acetaminophen-induced acute liver failure. Liver Int 2007;27:329-334.

6 Dhiman RK, Jain S, Maheshwari U, Bhalla A, Sharma N, Ahluwalia J, et al. Early indicators of prognosis in fulminant hepatic failure: an assessment of the Model for End-Stage Liver Disease (MELD) and King's College Hospital criteria. Liver Transpl 2007;13:814-821.

7 Simpson KJ, Bates CM, Henderson NC, Wigmore SJ, Garden OJ, Lee A, et al. The utilization of liver transplantation in the management of acute liver failure: comparison between acetaminophen and non-acetaminophen etiologies. Liver Transpl 2009;15:600-609.

8 Du WB, Pan XP, Li LJ. Prognostic models for acute liver failure. Hepatobiliary Pancreat Dis Int 2010;9:122-128.

9 Polson J. Assessment of prognosis in acute liver failure. Semin Liver Dis 2008;28:218-225.

10 Wheeler HO, Cranston WI, Meltzer JI. Hepatic uptake and biliary excretion of indocyanine green in the dog. Proc Soc Exp Biol Med 1958;99:11-14.

11 Haruna M, Kumon K, Yahagi N, Watanabe Y, Ishida Y, Kobayashi N, et al. Blood volume measurement at the bedside using ICG pulse spectrophotometry. Anesthesiology 1998;89: 1322-1328.

12 Nishioka M, Ishikawa M, Hanaki N, Kashiwagi Y, Miki H, Miyake H, et al. Perioperative hemodynamic study of patients undergoing abdominal surgery using pulse dye densitometry. Hepatogastroenterology 2006;53:874-878.

13 Du ZG, Li B, Wei YG, Yin J, Feng X, Chen X. A new scoring system for assessment of liver function after successful hepatectomy in patients with hepatocellular carcinoma. Hepatobiliary Pancreat Dis Int 2011;10:265-269.

14 Sheng QS, Lang R, He Q, Yang YJ, Zhao DF, Chen DZ. Indocyanine green clearance test and model for end-stage liver disease score of patients with liver cirrhosis. Hepatobiliary Pancreat Dis Int 2009;8:46-49.

15 Cheung TT, Chan SC, Chok KS, Chan AC, Yu WC, Poon RT, et al. Rapid measurement of indocyanine green retention by pulse spectrophotometry: a validation study in 70 patients with Child-Pugh A cirrhosis before hepatectomy for hepatocellular carcinoma. Hepatobiliary Pancreat Dis Int 2012;11:267-271.

16 Li H, Li B, Wei Y. Potential factors dedicated to postoperative liver dysfunction in patients with normal preoperative ICG-15 clearance rate. Dig Dis Sci 2013;58:1163-1164.

17 Lee WM, Larson AM, Stravitz RT. AASLD Position Paper; The management of acute liver failure: update 2011. Hepatology 2011;1-22.

18 Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM, Kosberg CL, et al. A model to predict survival in patients with end-stage liver disease. Hepatology 2001;33: 464-470.

19 O'Grady JG, Alexander GJ, Hayllar KM, Williams R. Early indicators of prognosis in fulminant hepatic failure. Gastroenterology 1989;97:439-445.

20 O'Grady JG. Broadening the view of acetaminophen hepatotoxicity. Hepatology 2005;42:1252-1254.

21 Liver Failure and Artificial Liver Group, Chinese Society of Infectious Diseases, Chinese Medical Association; Severe Liver Diseases and Artificial Liver Group, Chinese Society of Hepatology, Chinese Medical Association. Diagnostic and treatment guidelines for liver failure. Zhonghua Gan Zang Bing Za Zhi 2006;14:643-646.

22 Yantorno SE, Kremers WK, Ruf AE, Trentadue JJ, Podestá LG, Villamil FG. MELD is superior to King's college and Clichy's criteria to assess prognosis in fulminant hepatic failure. Liver Transpl 2007;13:822-828.

23 Malinchoc M, Kamath PS, Gordon FD, Peine CJ, Rank J, ter Borg PC. A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts. Hepatology 2000;31:864-871.

24 Freeman RB Jr, Wiesner RH, Harper A, McDiarmid SV, Lake J, Edwards E, et al. The new liver allocation system: moving toward evidence-based transplantation policy. Liver Transpl 2002;8:851-858.

25 Merle U, Sieg O, Stremmel W, Encke J, Eisenbach C. Sensitivity and specificity of plasma disappearance rate of indocyanine green as a prognostic indicator in acute liver failure. BMC Gastroenterol 2009;9:91.

Received March 27, 2013

Accepted after revision November 11, 2013

Author Affiliations: Intensive Care Unit, Tianjin Second People's Hospital, Tianjin 300192, China (Feng HL, Li Q, Wang L, Yuan GY and Cao WK)

Qian Li, MD, Intensive Care Unit, Tianjin Second People's Hospital, 75 Sudi Road, Nankai District, Tianjin 300192, China (Tel: 86-22-27468550; Email: crbicu@163.com)

© 2014, Hepatobiliary Pancreat Dis Int. All rights reserved.

10.1016/S1499-3872(14)60040-0

Published online March 27, 2014.