EASL-ALEH 2015 Algorithm for the Use of Transient Elastography in Treatment-naive Patients With Hepatitis B

An Independent Validation

Jean Nana; Kristina Skaare; Jean Luc Bosson; Vincent Leroy; Tarik Asselah; Michael Adler; Nathalie Sturm; Jean-Pierre Zarski


J Viral Hepat. 2021;28(8):1169-1176. 

In This Article


In this study, we were able to validate the EASL-ALEH 2015 algorithm based on LSM and ALT serum levels in treatment-naive patients with chronic Hepatitis B in a large population of well-characterized HBV patients as compared to liver biopsy. Since they were published in 2015, this algorithm has not been independently validated. Fibrosis stage was assessed using the METAVIR scoring system (F0–F4). In addition, TE and various blood markers were performed at the same time.

First, we have validated performance of TE for diagnosing significant fibrosis and cirrhosis. The optimal cut-off values in our study were 6.2 kPa for ≥F2, 8.3 kPa for ≥F3 and 10 kPa for F4. These results are similar to those published by Goyal et al. (6 kPa for ≥F2 and 11 kPa for F4)[24] and also observed by different authors.[25–32] As already observed, performance of TE was better for detecting cirrhosis (F = 4) with an AUROC of 0.92 at a cut-off of 10 kPa. These results are consistent with previous studies.[25–30] The cut-off was lower for significant fibrosis and cirrhosis than that published for chronic hepatitis C.[31,32] This tendency of a lower cut-off in CHB may be explained by two unique features. First, Sturm et al.[33] concluded that the total amount of liver fibrosis reflected by the fibrosis area was significantly lower in patients with CHB, because the fibrous septa might be thinner in these patients than in those with CHC with the same histological stage (F4). Secondly, because CHB tends to progress to cirrhosis with larger nodules (macronodular cirrhosis) than CHC, the TE pulse is more likely to pass through the normal liver parenchyma between fibrotic bands in patients with CHB than in those with CHC. In the present study, cut-off values for diagnosis of cirrhosis at which there is very high positive predictive value (PPV) for cirrhosis were different from those previously published.[34,35] These differences could be due to the study population which comprised patients with chronic liver diseases of various aetiologies. Secondly, in the two previously published studies, cut-off values were chosen to have a positive predictive value of more than 90%, which favours specificity. In this study, cut-off values were chosen to maximize the sum of sensitivity and specificity.

We stratified the 413 patients with available TE and ALT data into three categories: viral load lower than 20,000 IU/ml and ALT normal (n = 40), viral load >20,000 IU/ml and ALT > 2 times the upper limit of the normal value (n = 118) and others patients (n = 255). Numerous studies have demonstrated the influence of necrotic and inflammatory activity on TE, with increased LSM being associated with elevated ALT levels.[36,37] In the present study, as expected, patients with elevated ALT levels showed higher LSM values. However, performances of TE were not compromised with these ALT elevations. In others patients, those represent majority of HBV-infected patients in whom liver biopsy was performed, and performances of TE to stage significant fibrosis were good.

To validate the EASL-ALEH 2015 algorithm, we evaluated the prevalence of significant fibrosis within two groups of patients according to ALT levels and LSM. In the first group (patients with normal ALT levels), the prevalence of significant lesions was 9%, 38% and 67%, respectively, when LSM was <6kPa, between 6 and 9 kPa or more than 9 kPa. For patients with normal ALT levels and LSM between 6 and 9 kPa, the indication of liver biopsy should be discussed. We propose to take into account HBV-DNA level. For patients with normal ALT levels and LSM more than 9 kPa, EASL 2015 algorithm proposed to consider treatment screening for varices and hepatocellular carcinoma (HCC). This approach permits a good selection of the majority of patients (67% in our study), for whom antiviral treatment is indicated, despite a subsequent risk of overestimation of extensive fibrosis/cirrhosis. We propose to also add HBV-DNA serum level. Liu C et al have suggested that HBV-DNA level decreased with the increase in fibrosis or the presence of cirrhosis on liver tissues.[38] Indeed, Praneenararat S, et al showed that HBV-DNA level >5.5 log IU/ml was able to predict significant liver fibrosis in HBeAg-negative CHB patients.[39]

For patients with elevated ALT level but lower than 5 times ULN, the prevalence of significant fibrosis was 15%, 52% and 85%, respectively, when LSM was <6kPa, between 6 and 12 kPa or more than 12 kPa. EASL 2015 algorithm suggested to exclude other causes of elevated ALT levels and consider to regularly repeat TE when LSM was <6kPa, to perform liver biopsy when LSM was between 6 and 12 kPa and to consider treatment, screening for varices and HCC when LSM was more than 12 kPa. In our study, performances of TE were excellent in the subgroup of patients with ALT levels >N but <5N (n = 306). The percentages of well-classified patients were 73% for F ≥ 2, 82% for F ≥ 3 and 91% for F ≥ 4 patients. Clearly, our results showed that we can use LSM alone to diagnose significant fibrosis and cirrhosis in this subgroup of patients as well as in patients with ALT levels up to five times the ULN as observed in our study.

Significant discrepancies (≥2 points) between liver histology and LSM were observed in 13 patients. In cases of discrepancies between LSM and the liver fibrosis stage, LB specimens were re-analysed by the senior pathologist (N.S.) to clarify reasons for misclassification. Increased LSM in HBe antigen–negative patients was associated with extensive necrosis probably due to a peak of activity at the time of biopsy. However, 5 patients with LSM<6 KPa and ALT>ULN but ≤5ULN had significant fibrosis. All were F = 3 and had very high HBV-DNA serum level maybe explaining discrepancies observed between LSM and liver biopsy. In these patients, mean HBV-DNA serum level was 6.42 log UI/ml. Praneenararat et al.[39] have shown that HBV-DNA serum levels >5.5 log IU/mL could predict on multivariate analysis significant liver fibrosis (OR = 28.012, [95% CI: 1.631–481.240], p = .022). The workforce is certainly small but their study highlights the importance of taking into account the level of viral load in patients for whom a liver biopsy is recommended. Several patients with significant fibrosis that should be treated are missed even by a rather low LSM cut-off at 6kPa.

The main limitations of our study were the small number of included patients with F3 and F4 fibrosis stages and the absence of follow-up. However, the distribution of fibrosis stages in our population corresponded to that usually observed in France and in Europe. In addition, we think that in patients with LSM between 6 and 9kPa, we should look for co-morbidities, especially metabolic syndrome and steatosis and repeat TE at regular intervals to minimize errors. The effects of non-alcoholic fatty liver disease on TE in patients with chronic hepatitis B were not investigated in our study. Some biopsies included in the study are below the recommended quality criteria biopsies; 27.8% had <9 portal tracts.