icon-folder.gif   Hepatitis Reports for NATAP  
    Back grey_arrow_rt.gif
Prediction of histology in hepatitis C
  Hepatology April 2003, Volume 37, Number 4. Claudio Puoti, Department of Gastroenterology, Genzano Hospital, Rome, Italy
To the editor:
We read with interest the article by Pradat et al.1 on the predictive value of alanine aminotransferase (ALT) levels for histologic findings in chronic hepatitis C. However, we have some concerns about the methodology in this report.
Prada study: Predictive value of ALT levels for histologic findings in chronic hepatitis C: A European collaborative study The aim of this retrospective study was to determine the predictive value of alanine aminotransferase (ALT) levels for histologic findings in patients with chronic hepatitis C virus (HCV) infection. Data on 864 HCV RNA-positive patients were collected. ALT values were obtained at the time of biopsy (before treatment), and normal ALT values were defined as normal values obtained at serial evaluations during a 6-month period. Histologic results were scored using the METAVIR system. Among all patients, 99% of those with elevated ALT levels had a score of at least F1 (positive predictive value [PPV], 99%) and 88% had a score greater than A1F1. Among patients with persistently normal ALT values, 65% had a score of at least F1 (negative predictive value [NPV], 35%) and 26% had a score greater than A1F1. The receiver operating characteristics analysis indicates that the ALT threshold for the best compromise sensitivity-specificity was about 2.25 times the upper limit of normal (ULN). In conclusion, almost all HCV RNA-positive patients with elevated ALT levels have some degree of fibrosis. However, an important proportion of patients with persistently normal ALT levels also show some histologic signs of fibrosis; the degree of fibrosis is usually mild but is sometimes more marked, and in rare cases cirrhosis may be present. In this subset of patients, the indication of liver biopsy and the potential benefit of therapy need to be further evaluated. These results suggest the need to revisit the algorithm for liver biopsy practice. (HEPATOLOGY 2002;36:973-977.)
In this retrospective study, "persistently" normal ALT levels (PNAL) were defined as 2 to 4 sequential normal values obtained during a 6-month period. However, it has been shown that this observation period is too short and thus inadequate to discriminate between hepatitis C virus (HCV) carriers with PNAL and patients with transient biochemical remission. Two recent prospective studies found that many subjects referred as HCV carriers with PNAL on the basis of this definition did suffer from ALT flares during follow-up. In our study, 21% of the patients showed ALT flare-ups during a 34-month follow-up period (range, 24-48 months), and Martinot-Peignoux et al. found ALT increases even after 60 months of follow-up. Therefore, it is possible that many subjects that Pradat et al. considered as having PNAL had only transient ALT normality, thus deriving a higher-than-expected prevalence of severe liver damage in this group. As the investigators correctly stated, persistently normal ALT level is really a function of the stringency of follow-up and therefore depends on the frequency of ALT determinations. Were these subjects still followed-up after the initial 6-month period? If so, we question how many people showed ALT flares during the follow-up, and if these subjects at the time of analysis were included in group 1 (PNAL) or group 2 (abnormal ALT). Furthermore, we wonder if there were differences in terms of histologic, virologic, and demographic features between subjects with PNAL and those with ALT flares during the follow-up period.
Further, no information is given regarding virologic features in these patients. Although viral load does not differ between subjects with PNAL or abnormal ALT nor does it predict the severity of liver damage, several studies found a higher frequency of non-1 HCV genotype in subjects with PNAL. This finding could have potential implications for antiviral treatment, because it is known that non-1 HCV genotypes are associated with higher response to interferon treatment. We ask whether HCV genotype distribution and viral load were different between the two groups of patients, and whether correlations were found between virologic features and the presence or absence of PNAL. Furthermore, it could be very interesting to know whether genotype distribution was significantly different between the 6 centers.
Finally, Pradat et al. did not make clear the details of patient recruitment, and no information is provided as to how HCV carriers with PNAL were identified. Were these subjects incidentally detected after a chance finding of a positive anti-HCV test, or were they already followed-up because of previous blood exposure (e.g., blood tranfusion, intravenous drug addiction, etc.)? It could be very informative to compare histologic findings according to the presence or absence of risk factors for HCV exposure, because, to our knowledge, no such data exist in the literature.