| BackgroundAcute-on-chronic liver failure (ACLF), with a case fatality rate of 60% to 70%, is the most common type of liver failure and mainly caused by infecting with hepatitis B virus (HBV). Estimating severity of HBV-related ACLF accurately and assessing disease progression promptly have been always highly concerned. Since Model for End-stage Liver Disease (MELD) founded in 2000, it has been widely used in assessing the severity of liver failure and predicting prognosis of patients. From then on, Ruf et al. had found that both hyponatremia and MELD Scoring System were risk factors affecting mortality, if using the two assessment methods in combination would be helpful for better prognosis. Therefore, the MEID-Na scoring system had been derived. Biggins et al. had found that MELD-Na scoring system has better prediction capacity upon survival rate than single MELD scoring system by studying a total of 753 patients with end-stage liver disease from several liver transplantation centers. Tang Changhua et al. of China also verified that MELD scoring system will have higher reliability on the assessment and prediction of ACLF conditions when used in combination with serum sodium factor. Fifty years ago Child-Turcotte-Pugh score was raised by Child and Turcotte and then revised by Pugh, which became a model for assessing liver reserve function and prognosis of serious liver diseases. MEID-Na scoring system derived from End-stage Liver Disease Model with characteristic of objectivity and detailed as well as Child-Turcotte-Pugh (CTP) score with easy operation are the most widely used model systems for liver failure assessment in clinic. Both systems are based upon liver failures caused by HCV infection and alcoholic liver impairment. On the contrary, there are a few studies focusing on HBV-related ACLF. Whether the two models are appropriate for the clinical Chinese patients requires further study.ObjectivesTo study predictive value of MELD-Na and CTP Scoring system by evaluating short-term prognosis of HBV-related ACLF, and to offer clues and basis for clinical options of appropriate therapeutic regimen and rapid prognosis assessment.Methods1. Baseline DataAll cases are inpatients of HBV-related ACLF who were treated in The Eighth People’s Hospital of Guangzhou and Nanfang Hospital of Southern Medical University from January 2010 to December 2012. All patients were HBsAg positive and met diagnostic criteria for ACLF in Diagnostic and treatment guidelines for liver failure (2012 version). Patients with the following conditions were excluded:ones combined with hepatitis virus infection and HIV infection, ones with cholestatic, alcoholic and autoimmune liver disease, ones with drug intoxication induced liver failure, ones with liver tumor or liver transplantation, and ones with kidney underlying disease induced renal insufficiency or with long-term anticoagulant therapy. All patients received similar anti-virus, liver protection, symptomatic treatment and other comprehensive treatment.2. Observation Indicators and Follow-Up End-PointAll clinical information of timing nodes with highest serum total bilirubin during the hospitalization and test indicators were collected. Clinical information covered ascitic volumes and hepatic encephalopathy stages. The definition of ascitic volumes taken from International Ascites Club (IAC) was listed as follow:small volume indicates ascites which can only be detected by ultrasonic test; medium volume indicates ascites which can induce moderate symmetric abdominal distention; and the large volume indicates tension ascites which would cause apparent abdominal distention. Test indicators cover serum total bilirubin, albumin, creatinine, prothrombin time (PT), PT international normalized ratio (INR) and serum sodium concentration. Liver failure stages referred to Diagnostic and treatment guidelines for liver failure (2012 version). Laboratory test index of the two hospitals are in the same reference value range. Within 90-day (since hospitalization) survival or death points are taken as follow-up endpoint. The judgments of end point are on the basis of following conditions:cases of whose condition worsening, giving up emergency treatment and being discharged or dead are grouped into death cases; while ones who revisited the hospital after 90 days are grouped into survival cases; for ones who didn’t seek medical consultation after discharge, survival/death status must be confirmed with the patient or his/her family by telephone follow-up..3. Score CalculationAs all subjects of this study are HBV relative failure patients, the MELD scores equals to 3.8×loge (serum bilirubin umol/L×0.058) plus 11.2xloge (PT INR) plus 9.6xloge (serum creatinie umol/L×0.011) plus 6.4. MELD-Na scores equals to MELD scores plus 1.59×(135-Na+). If levels of serum Na+was greater than or equal to 135mmol/L,135mmol/L would be taken for the calculation; if levels of serum Na+ was less than or equal to 120mmol/L,120mmol/L would be taken for the calculation; if levels of serum Na+ was between 120mmol/L to 135mmol/L, exact values was taken. According to scoring criterion, Child-Turcotte-Pugh (CTP) scores were calculated by five indicators including hepatic encephalopathy stages, ascites, TBil, albumin and prolongation of PT.4. Statistical MethodsMeasurement data was expressed as mean±SD (x±s) and Mest was adopted. Enumeration data was expressed as ratio (%) and compared by x2-est. Spearman rank sum test was used for correlation analysis of liver failure stages and MELD-Na and CTP scores. If the correlation coefficient rs above 0.75 was considered as high correlation, between 0.45 and 0.75 was considered as medium correlation, and under 0.45 was considered as low correlation. Receiver operating characteristic (ROC) curve and area under the curve (AUC) were used for MELD-Na and CTP scores assessing HBV ACLF patients’short-term prognosis; when the AUC was above 0.7, scores was considered with clinical application value; when it was above 0.8, it would be considered with good predictive/judgment accuracy; Z-test is adopted for the AUC comparison; The best Cut-off value and Youden index are determined according to the sensitivity and specificity of the ROC curve. All statistical analyses were analyzed with SPSS 13.0 statistical software. A P<0.05 level was used to determine a significant difference.Results1. Baseline DataA total of 339 patients with HBV-related ACLF were collected (232 cases from The Eighth People’s Hospital of Guangzhou and 107 cases from Nanfang Hospital of Southern Medical University); the patient group comprised 299(88.2%) males and 40(11.8%) females, aged 43.16±12.15 years (range 18-82 years). Averages of measurements:Serum TBil of 493.10±170.73umol/L, PT INR of 3.08±1.69, serum creatinine of 104.64±89.57umol/L, serum sodium of 134.93±5.93 mmol/L, PT of 31.32±5.29s, albumin of 38.13±30.05g/L, MELD-Na scores of 33.70±11.24 and Child-Turcotte-Pugh scores of 11.70±1.61.2. Correlation between Liver Failure Stages and MELD-Na and CTP Score2.1 Liver Failure Stages and MELD-Na and CTP ScoreFor patients in ACLF-early stage (n=87), the MELD-Na score was 26.38±7.00 while CTP scores was 11.76±1.51. For ones in ACLF-middle stage (n=74), MELD-Na and CTP scores were 32.32±11.29 and 11.77±1.35 respectively. For ones in ACLF-advanced stage (n=178), MELD-Na and CTP scores were 37.86±10.97 and 11.65±1.75 respectively. MELD-Na scores in advanced patients was obvious higher than that in the middle stage patients (P<0.001) and early stage patients (P<0.001). And scores in middle stage patients were apparently higher than that in early stage ones (P<0.001). But for CTP scores, comparisons between patients in early, middle or advanced liver failure stage had no statistical significance.2.2 Correlation between Liver Failure Stages and MELD-Na and CTP ScoresMELD-Na scores were positively related with liver failure stages with a Spearman correlation coefficient (rs) of 0.485, which indicated moderate correlation; and there was a statistical difference (P<0.001). CTP scores were also positively related with liver failure stage with an rs of 0.306, indicating low correlation; and there was a statistical difference (P<0.001).3. Correlativity between Liver Failure Stage, MELD-Na Scores, CTP Scores and Prognosis3.1 Correlativity between Liver Failure Stages and PrognosisIn all 339 cases,191 (56.3%) are grouped into the survival group while 148 (43.7%) are grouped into the death group. In the early liver failure stage group,70 (80.5%) of 87 patients survived and 17 (19.5%) were dead. In the middle liver failure stage group,41 (55.4%) survived and 33 (44.6%) did not. In the liver failure advanced group,80 patients (44.9%) survived and 98 (55.1%) died. Differences of the short-term mortality between patients in three stages were significant (χ2=30.001, P<0.001).3.2 Correlativity between MELD-Na, CTP Scores and PrognosisThe MELD-Na scores (40.44±11.56) of the death group was obvious higher than that of the survival group (28.49±7.66) with statistical significance (t=-11.426, P<0.001); Difference of CTP scores between the two groups (11.76±1.42 vs. 11.66±1.74) was without statistical significance (t=-0.551, P=0.582).For different prognosis of MELD-Na and CTP scoring system, inter-range differences were significant. MELD-Na scoring system group:there are 84 patients with scores lower than or equal to 25 in which 14(16.67%) died; 71 ones with scores above 25 but lower than or equal to 30 in which 16 (22.54%) died; 56 ones with scores above 30 but lower than or equal to 35 in which 19 (33.93%) died; 41 patients with scores were higher than 35 but lower than 40 in which the death rate was 58.54%(n=24); death rate of score range between 40 and 45 was 90.63%(29/32) and that of score range above 45 was 83.64%(46/55); the x2 value for inter-range comparison of MELD-Na scoring system is 108.037 (P<0.001). CTP scoring groups: there are 23 patients whose scores were 9 or below with no death case; 66 patients with 10 scores out of which 7 (10.61%) died; 69 ones with 11 scores and 19 patients died(27.45%); in whose scores were 12, the death rate is 55.88%(38/68); and in 68 patients with 13 CTP scores, the death rate was 63.24%(n=43); and for ones whose scores are 14 or the above, the death rate was 41/45 (91.11%); the x2 value for inter-range comparison of CTP scores is 110.346 (P<0.001)。In conclusion, the short-term mortality was elevated along with MELD-Na and CTP scores increasing.4. The Predictive Value of MELD-Na and CTP Assessment upon the Short-Term Prognosis of Acute-on-chronic Hepatitis B Liver FailureThe score AUC of MELD-Na was 0.813(95%CI:0.765-0.860; cut-off value: 34.28; sensitivity:0.716; specificity:0.832; Youden index:0.548). The score AUC of CTP score was 0.823(95%CI:0.780-0.867; cut-off value:11.5; sensitivity:0.824; specificity:0.691; Youden index:0.515). Both MELD-Na and CTP scoring system had favorable survival predictive capacity for HBV-related ACLF short-term prognosis. There was overlapping parts of the 95% CI in the two scoring models. Z-test was adopted for comparing the predictive capacities between the two scoring systems (Z=0.229, P=0.819); the difference had no statistical significance.According to Cut-off value of the MELD-Na scoring system(=34.28),339 patients were divided into MELD-Na scores was 34 or less group and MELD-Na scores above 34 group, case fatality rate of the two groups were 21%(41/195) and 74.3%(107/144) respectively, difference between them was statistically significant (x2=95.594, P< 0.001).According to Cut-off value of the CTP score scoring system (=11.5), patients were divided into CTP lower than 12 group and CTP scores higher than or equal to 12 group, and case fatality rate of the two groups were 16.5%(26/158) and 67.4%(122/181) respectively, the difference was statistically significant (x2=89.020,P<0.001).Conclusions1. In differentiating the early, middle and late stage of liver failure, MELD-Na scoring system can make obvious distinguish, accurately reflecting the severity of liver failure conditions to some extent, while the CTP scores showed no statistical significance upon different liver failure stages.2. MELD-Na and CTP scoring system were positive related to liver failure stages, while the correlation of the later is not close. MELD-Na Score held higher correlation with different liver failure stages.3. The 90-days mortality of HBV-related ACLF patients is 43.7%, while the mortality of advanced liver failure patients would be up to 5.1%.4. When the patients are divided into the survival group and death group, the MELD-Na scores of the later is significantly higher than that of the former, which indicated that MELD-Na scoring system can better reflect the severity of the conditions in patients with liver failure, while the CTP scores of the two groups are without statistical significance.5. It was showed that the mortality is elevated with increasing MELD-Na and CTP scores. Therefore, both MELD-Na and CTP scores can predict the short-term prognosis of liver failure patients.6. MELD-Na and CTP scoring system have good prediction accuracy and authenticity for short-term prognosis prediction for patients with liver failure. Both MELD-Na and CTP scores are preferable models for the short-term prognosis prediction in HBV-related ACLF patients.7. According to the Cut-off values, case fatality rate in the group of MELD-Na scores higher than 34 was 74.3%,while that in the group of CTP scores higher than or equal to 12 was 67.4%, indicating patients in this condition have poor prognosis.Compared with CTP scoring system, MELD-Na scoring system is relatively better on the short-term prognosis prediction in HBV-related ACLF patients. Clinically, the two models can be functional complementation by MELD-Na playing the major role. Only closely combined the two models with clinical practice, we can work for precise assessment of short-term prognosis in HBV-related ACLF patients, then to carry out corresponding treatments as soon as possible. |
PDF Full Text Request