| 1. OBJECTIVETo investigate the dynamic changes of liver stiffness measurement (LSM) by transient elastography (TE) in patients with chronic hepatitis B (CHB) acute exacerbation and determine LSM cutoffs stratified by alanine aminotransferase (ALT) or bilirubin levels to improve diagnostic accuracy of TE for liver cirrhosis diagnosis in patients with compensated chronic hepatitis B.2. METHOD2.1. Dynamic changes of LSM in acute exacerbation of CHB A total of 89 consecutive treatment naive patients with CHB acute exacerbationwere included in the study between March 2009 and April 2010. A diagnosis of CHB was based on persistently positive hepatitis B surface antigens for more than 6 months or liver biopsy proved fibrosis≥METAVIR F3. Acute exacerbation of CHB was defined as ALT more than 5 times the upper limit of normal range (×ULN). Hepatitis B virus DNA higher than 1000 copies/mL was detected in all patients by real-time polymerase chain reaction. Patients with other hepatotropic virus co-infection and a history of autoimmune liver disease, metabolic liver disease, or drug induced liver disease were excluded. Abdominal ultrasonography was performed to exclude fatty liver disease and hepatocellular carcinoma. Ultrasonic images of nodular liver surface and splenomegaly were considered as an ultrasonic image of liver cirrhosis.The initial level was defined as the peak level of ALT. Dynamic changes of LSM and biochemical tests were monitored. LSM was performed every 7-10 days during hospitalization and every 1～3 months for following-up outpatients. Hematology and liver function tests were performed on the same day of as LSM. Liver biopsy was performed as needed.LSM was completed by Fibroscan(?) (ECHOSENS, France) with the probe placed in the usual liver biopsy position. Details of the procedure of LSM have been previously described.The median value, expressed in kilopascal (kPa) was kept as representative of the LSM. Only an LSM with at least ten successful shots, a success rate of more than 60% and an inter-quartile range/median ratio (IQR/M) lower than 30% was considered reliable. Cut-off 18.2 kPa for patients with elevated ALT and 12.0 kPa for normal ALT were selected as the cirrhosis cutoff values in the current study, as suggested by Marcellin et al. and Chan et al.respectively.All statistical analyses were performed with the Statistical Package for Social Science (version 13.0, SPSS Inc, Chicago, IL). Continuous variables were expressed in mean±standard deviation or median (range) as appropriate. Continuous variables comparison was completed by Mann-Whitney test or Kruskal-Wallis test as appropriate. Pearson correlation coefficient or partial correlation coefficient was used to evaluate correlations between LSM, ALT and bilirubin. The significance level was set at 0.05, and all P values were two tailed.2.2. LSM cutoffs for cirrhosis diagnosis stratified by ALT or bilirubin levels in compensated chronic hepatitis B 389 consecutive treatment naive patients with CHB who underwent liver biopsy between July 2007 and July 2010 were prospectively included. A diagnosis of chronic hepatitis B was based on persistently positive hepatitis B surface antigens for more than 6 months or liver biopsy proved CHB. Patients with other hepatotropic virus co-infection and a history of autoimmune liver disease, alcoholic fatty liver disease, non alcoholic steatohepatitis, liver tumor or other hepatobiliary diseases were excluded.All the patients underwent percutaneous liver biopsy, ultrasound examination and routine laboratory tests. Routine laboratory tests were performed within three days of LSM, including hematology, ALT, aspartate aminotransferase (AST), albumin, bilirubin and prothrombin time. LSM were performed by three operators within one week of the liver biopsy. The operators were unaware of the clinical details of the patient. Only biopsy samples with a length >15mm and including at least 10 complete portal tracts were considered adequate. Liver fibrosis was semi-quantitatively evaluated by the METAVIR scoring system as follows:F0, no fibrosis; F1, portal fibrosis without septa; F2, portal fibrosis and few septa; F3, numerous septa without cirrhosis; F4, cirrhosis.Statistical analysis was performed using the Statistical Package for Social Science (SPSS version 13.0; SPSS Inc., Chicago, IL, USA). Continuous variables were expressed as mean±standard deviation or median (range) as appropriate. Spearman correlation coefficient was used to evaluate the correlation between LSM and METAVIR fibrosis stage. Pearson correlation coefficient was used to evaluate correlations between LSM, ALT and bilirubin. Categorical variables were compared by chi-squared test or Fisher's exact test. Continuous variables were compared by Student t-test or Mann-Whitney test as appropriate. The overall accuracy of LSM in cirrhosis diagnosis was calculated using the area under receiver operating characteristics curve(AUROC) and its 95% confidence internal(CI). Optimal cutoff values of LSM for confirming and excluding cirrhosis were chosen to obtain positive likelihood ratio(PLR) nearly 10.0 and negative likelihood ratio(NLR) nearly 0.1, respectively. All statistical tests were two-sided. Statistical significance was taken as P<0.05.3. RESULT3.1. Dynamic changes of LSM in acute exacerbation of CHBA total of 282 LSMs were performed. Median follow-up period (range) was 26 (7-217) days, with an average of 55 days. LSM was significantly correlated with ALT (r=0.175, P=0.003) and bilirubin (r=0.483, P<0.001). During the period of follow-up, 54 out of 89 patients had records at ALT levels of 2-5xULN,40 at ALT levels of 1-2×ULN, and 33 at ALT normalization. LSM in patients with ALT normalization was significantly lower than with other ALT levels (Kruskal-Wallis H test:x2=36.607, P<0.001). Likewise, the mean reduction of LSM in patients with ALT normalization was significantly greater than with other ALT levels (x2=14.301, P=0.001).Forty-one patients were detected with baseline bilirubin above 3×ULN, and 30 of them had a continuous drop in their bilirubin levels. During the period of follow-up, 17 out of the 30 patients had records at bilirubin levels of 2-3xULN,17 at bilirubin levels of 1-2×ULN, and 7 at bilirubin normalization. LSM in patients with bilirubin normalization was significantly lower than with other bilirubin levels(Kruskal-Wallis H test:x2=23.99, P<0.001). Parallel with subsequent bilirubin levels, reductions of LSMs were 10.66±14.67kPa,16.83±13.08kPa and 22.11±7.63kPa at corresponding levels of bilirubin 2-3xULN (n=17), 1-2xULN (n=17) and 0-1xULN (n=7). The mean reduction of LSM in patients with bilirubin 0-1xULN was significantly greater than in other groups (Kruskal-Wallis H test:χ2=6.559,P=0.038).Among the 89 patients,33 were observed till ALT normalization whose LSMs dropped accordingly (28.79±15.62 vs.15.56±14.25, Z=-4.194, P<0.001).23 of them were initially diagnosed with cirrhosis by cut-off 18.2kPa. Parallel with ALT normalization,15 patients were still diagnosed with cirrhosis by cut-off 12.0 kPa. Among them,1 patient was histologically proved METAVIR F4,1 patient had typical ultrasonic images of cirrhosis, and 10 patients were characterized with persistent bilirubin elevation. In the other 8 patients with LSMs reduced to non-cirrhosis levels. 5 patients were tested without typical ultrasonic images of cirrhosis, and the other 3 were histologically proved METAVIR F3. Following the continuous reduction in ALT, four patterns of LSM changes were observed:descendent pattern, wave pattern, biphasic pattern and assurgent pattern. In patients without clinical, imageological or histological signs of cirrhosis, LSM decreased to lower than cirrhosis cut-off 12.0kPa parallel with ALT normalization and bilirubin decline. However, in patients with liver biopsy proven or clinical, ultrasonic cirrhosis, LSM varied and remained at levels higher than the cirrhosis cut-off. In patients with cholestasis, LSM rose to and remained at a higher level regardless of ALT normalization.3.2. LSM cutoffs for cirrhosis diagnosis stratified by ALT or bilirubin levels in compensated chronic hepatitis BAmong 389 patients with CHB underwent liver biopsy,61 patients were excluded due to inadequate liver biopsy sample size (n=56) or unreliable LSM (n=5). 13 hepatic decompensated patients due to bilirubin higher than 51μmol/L were also excluded. At last,315(81.0%) patients were included in present study. Among them, 74 (23.5%) patients were classified as fibrosis F4, and 70(22.2%),99(31.4%), 65(20.6%),7(2.2%) patients were fibrosis F3, F2, F1, F0, respectively. LS was significantly correlated with liver fibrosis stage(r=0.745, P<0.001). The AUROC was 0.888(95%CI:0.848-0.928) for cirrhosis (METAVIR F4) diagnosis. With excluding cut-off 10.3 kPa and confirming cut-off 20.5 kPa,218(69.2%) patients were discriminated with or without cirrhosis with accurate rate of 92.2%.LSM was correlated with ALT (r=0.216, P<0.001). Cut-offs for excluding and confirming cirrhosis was similar in patients with normal ALT, ALT 1-2xULN and ALT 2-5xULN, but far different from that with ALT≥5xULN. The prevalence of cirrhosis is similar in patients with ALT<5xULN and ALT≥5xULN (22.8% vs.26.9%,χ2=0.491, P=0.484), while LSMs were lower in patients with ALT<5xULN [9.1(3.2-48.0) vs.10.3(4.8-54.5), Z=-2.893, P=0.004]. Patients with high ALT tend to have higher LSM than those with low ALT levels at the same fibrosis stage. ALT was correlated with LSM (r=0.270, P=0.020) in patients with METAVIR F4, and LSMs were higher in patients with ALT≥5xULN (Z=-2.413, P=0.016). ALT was not significantly correlated with LSM in patients with METAVIR F3, F2 and F1. Further cut-offs was determined:10.3kPa for excluding cirrhosis, and 16.9kPa for confirming cirrhosis in patients with ALT<5xULN; excluding cut-off 13.0 kPa and confirming cut-off 24.5 kPa for patients with ALT≥5xULN. Based on these cutoffs,72.7% of all patients [72.4% of patients with ALT<5xULN and 76.1% (51/67) of patients with ALT≥5xULN. The accurate rates were 92.1% and 92.2%, respectively] could be exempted from biopsy.LSM was correlated with bilirubin with Pearson's coefficient 0.355(P<0.001). LSM [median (range)] in patient with elevated bilirubin was higher than that with normal bilirubin [14.1(4.5-54.5) vs.8.3 (3.2-33.3), Z=-3.979, P<0.001]. Bilirubin was significantly correlated with LSM in patients with METAVIR F2, F3 and F4. AUROC for cirrhosis diagnosis in patients with normal bilirubin was superior to that with elevated bilirubin [0.896(95%CI 0.844-0.949) vs.0.844(95%CI 0.766-0.922)]. Cut-offs in patients with normal bilirubin were determined as 10.5kPa for excluding cirrhosis and 16.8kPa for confirming cirrhosis. With these two cutoffs,78.3% of patients with normal bilirubin could be exempted from biopsy with an accuracy of 92.4%. However, in patients with elevated bilirubin, based on excluding cut-off 9.1kPa (NLR 0.068) and confirming cut-off 24.5kPa (PLR 8.4), only 50.5% of patients could be exempted from biopsy with accurate rate of 91.5%.4. CONCLUTIONLSMs in patients with CHB acute exacerbation were significantly correlated with serum ALT and bilirubin levels. LSM dropped parallel with decreasing ALT and bilirubin levels. In patients with CHB, LSM correlated well with METAVIR fibrosis stage. ATL cutoffs stratified by ALT 5xULN and bilirubin normalization would... |
PDF Full Text Request