Performance of unidimensional transient elastography in staging non-alcoholic steatohepatitis.

BACKGROUND/AIMS:

Transient elastography (TE) is a noninvasive method for predicting liver fibrosis, mainly validated in patients with viral hepatitis. Information is still limited concerning its performance in non-alcoholic steatohepatitis (NASH) patients. We aimed to assess the value of TE in the prediction of fibrosis stage in NASH as well as the factors determining the discordance between the TE-predicted and the biopsy-proven fibrosis stage in these patients.

METHODS:

Liver biopsy and TE were performed on 72 consecutive NASH patients. Fibrosis, lobular inflammation, ballooning and steatosis were evaluated (Brunt system).

RESULTS:

Liver stiffness (LS) values ranged from 2.80 to 16.90 kPa. In the univariate analysis, LS was correlated with fibrosis (r=0.661; p<0.0001), steatosis (r=0.435, p<0.0001), ballooning (r=0.385; p=0.001) and lobular inflammation (r=0.364; p=0.002). In multivariate analysis, only fibrosis significantly correlated with LS (p<0.0001). The median (and range) LS values (kPa) according to the fibrosis stages were: 4.90 (2.80-7.30) for F0; 6.15 (4.80-12.50) for F1; 6.90 (3.30-16.90) for F2 and 14.00 (10.70-14.10) for F3, with significant difference between stages, except for F1-F2 (p=0.249). Cut off values were calculated for predicting each fibrosis stage: 5.3 kPa (AUROC=0.879) for F1; 6.8 kPa (AUROC=0.789) for F2; and 10.4 kPa (AUROC=0.978) for F3. Patients with false-positive results had a significantly higher ALT level than those with concordant results (p=0.039).

CONCLUSION:

In NASH patients, TE allows a reliable assessment and prediction of liver fibrosis, especially in advanced stages. Steatosis, ballooning and inflammation do not influence liver stiffness.