Value Of MELD To Assess The Short-term Prognosis Of Severe Hepatitis By Artificial Liver Treatment

The outcome of artificial liver treatment depends on the severity of the disease, andthe cost of the treatment is expensive. So how to select the appropriate artificial livertreatment objects is important. Model for End-Stage the Liver Disease (MELD) hasbeen widely confirmed a good predictive value for short-term or long-term survival ofend-stage liver disease. We applied the scoring system in438patients with liver failurefrom our hospital in past seven years. And we study the patients’ MELD scores by ROC(Receiver Operating Characteristic Curve) to assess the MELD’ ability for predictingthe short-term prognosis of severe hepatitis after artificial liver treatment. We studythe two groups of patients’s clinical data by case-control analysis, and try to choose thebest artificial liver treatment groups.Objectives：To assess the value of the model for end-stage liver disease(MELD) onpredicting the short-term prognosis of severe hepatitis by artificial liver treatment,andlook forward to providing the best value range of MELD scores for reasonably choosingartificial liver treatment in rescuing patients from dying.Methods: The clinical information of438cases with liver failure was collected, datefrom2003.1to2009.12. Depending on whether they were treated by artificial liversupport, the patients with liver failure were divided into treatment group (treatmentgroup) and medical therapy group (control group). Their survival times were observedwithin3months. On basis of the receiver operating characteristic curve (ROC) of thetwo groups, we can evaluate MELD scores’ accuracy in predicting mortality. Eachcut-off value of ROC curve was calculated. According to the value of cut-off, the veryscope of the MELD scores adapting to artificial liver support was be predicted. The treatment group’s mortality was compared with the control group’s to search for otherscopes of the MELD value, in which the mortalities of the tow group are different.Results:29.28points was the treatment group’s MELD mean, and the control group’MELD mean was28.95points. ROC curve area for the treatment group was0.797, and31is the cut-off value. The control group ROC curve area is0.825,26is the cut-offvalue. Between26points and31points, mortalities of the corresponding two groupshad a statistically significant difference (P<0.01); Between31points and41points, thecorresponding mortalities of the two groups were statistical different (P<0.05); WhenMELD scores were less than26points or larger than41points, the mortality betweenthe two groups were not statistically different.Conclusions: MELD scoring system can predict short-term prognosis of severe hepatitis patientstreated by artificial liver treatment and medical therapy. The MELD score was higher, the short-termprognosis was worse. Artificial liver treatment can improve short-term survival in some patients.Patients whose MELD scores between26and41should try to get artificial liver support treatment.However, when the score was larger than41, liver transplant may be the only means to save the lifewith liver failure.