Assessment of indocyanine green clearance test combined with the model for end-stage liver disease in the prediction of short-term prognosis for liver failure

Hong-ling FENG,Qian LI,Lin WANG,Wan-you YU,Gui-yu YUAN,Wu-kui CAO,Ji-ming YANG
DOI: https://doi.org/10.3760/cma.j.issn.1000-6680.2013.10.004
2013-01-01
Abstract:Objective To investigate the efficacy of the indocyanine green (ICG) clearance test (ICGR15) combined with the model for end-stage liver disease (MELD) for assessing the short-term prognosis of patients with liver failure.Methods Eighty patients with liver failure were analyzed retrospectively.ICGR15 and relevant clinical data within 24 hours of diagnosis were analyzed.Meanwhile,the MELD score and King's College Hospital (KCH) were evaluated.All findings were tested for correlation with 3-month mortality.Quantitative data were analyzed with analysis of variance and Student's t-test.Count data were analyzed with chi-square test.Correlation analysis was performed with Pearson's coefficient test.Results Among 80 patients with liver failure,39 patients survived and 41 died.The mortality rate of all patients was 51.2%.The serum total bilirubin,creainine concentrations,ICGR15,MELD scores and patient number in accordance with KCH criteria of surviving patients were (288.0±109.1) μmol/L,(63.3±24.4) μmol/L,(48.1±10.2)%,20.6±4.4,and 6 cases,respectively,which were lower than those in dead patients [(340.7 ± 108.2) μmol/L,(98.8 ± 59.1) μmol/L,(60.2 ± 10.6) %,26.9 ± 7.1 and 19 cases,respectively] (P =0.033,P= 0.001,P= 0.000,P= 0.000 and P =0.003,respectively).There was no significant difference of ICGR15 among four types of liver failure.A positive correlation was observed between ICGR15 and MELD score (r=0.289,P=0.009).The ICGR15-MELD model was created by subjecting ICGR15 and MELD scores to Logistic regression analysis.The following ICGR15-MELD model,Logit (P) =0.105 × ICGR15 + 0.178 × MELD score-9.734,was constructed by Logistic regression analysis.The area under the receiver operating characteristic (ROC) curve was 0.860 and the cut offpoint of 0.3 had sensitivity of 85.40% and specificity of 74.40%.The area under the curve of the ICGR15-MELD model was significantly higher than those of ICGR15 (0.791),MELD score (0.770) and KCH criteria (0.655).Conclusions ICGR15 and MELD scores perform better than the KCH criteria in predicting the prognosis of liver failure.The ICGR15-MELD model is superior to ICGR15,MELD score,and KCH in predicting the short term prognosis of patients with liver failure.
What problem does this paper attempt to address?