We read with interest the article by Wieckowska et al. reporting the potential usefulness of plasma keratin-18 fragments as a noninvasive diagnostic means of differentiating nonalcoholic steatohepatitis (NASH) from simple steatosis.1 We would like to draw attention to similar studies on serum keratin-18 in the differentiation of steatohepatitis from simple steatosis in alcoholics. A keratin-18 fragment, the so-called tissue polypeptide specific antigen (TPS), began to be widely used as a tumor marker for epithelial malignancies many years ago (reviewed in Barak et al.2). Later, it was recognized that serum TPS levels may increase in patients with nonmalignant liver disease.3 This represents an obvious drawback to TPS as a tumor marker but could provide some diagnostic value as a marker of liver disease. We observed that serum TPS levels are markedly increased in patients with alcoholic liver disease.4–6 In these patients, serum TPS levels can differentiate steatohepatitis (alcoholic hepatitis) from simple steatosis more accurately than standard liver tests.4 A TPS cut-off level of 235 U/l showed a 91% sensitivity and 100% specificity in the diagnosis of alcoholic steatohepatitis.4 Thus, the diagnostic accuracy of TPS in the differentiation of steatohepatitis from simple steatosis in alcoholics seems similar to that reported by Wieckowska et al. in the differentiation of NASH from simple steatosis by means of keratin-18 fragments.1 It should be noted, however, that the TPS immunoassay and the immunoassay employed by Wieckowska et al. use different detector antibodies and therefore measure different keratin-18 epitopes. The TPS assay detects the M3 epitope (keratin-18 residues 322-340),7 whereas the method used by Wieckowska et al. detects the M30 neoepitope (keratin-18 Asp396-NE) that is generated during caspase-3–related apoptosis.1 The latter may therefore be a specific marker of cell apoptosis, whereas increased serum TPS levels could result, theoretically, from necrosis or apoptosis. In fact, TPS levels have been proposed as an extracellular marker of apoptosis in vitro.8 The TPS assay has the additional advantage of automated determination as a routine procedure in many centers.
The elevation of serum keratin-18 or its fragments in steatohepatitis is not surprising. Keratin-18 is a key protein of the hepatocyte cytoskeleton which plays an important role in liver health and disease.9 We observed that serum TPS levels are strongly correlated with the number of hepatocytes containing Mallory's hyaline degeneration,4 a peculiar histological finding of both alcoholic and nonalcoholic steatohepatitis which is mainly composed of aggregates of modified keratin-18 and keratin-8 in addition to nonkeratin components. Serum TPS levels are also correlated with abnormal hepatocyte keratin-8/keratin-18 expression in alcoholic liver disease.5, 6 In summary, these findings on the behavior of serum keratin-18 in alcoholic patients are consistent with those of Wieckowska et al. in nonalcoholic patients.1 Determination of serum keratin-18 (or its fragments) may therefore be useful as a noninvasive method to differentiate simple steatosis from steatohepatitis in corresponding settings. A word of caution should be added, because elevated levels of serum keratin-18 is not specific to steatohepatitis. Serum TPS is now routinely determined in patients with liver disease from our department, and preliminary analysis of these data indicates that elevated serum concentrations of keratin-18 (TPS) can be found in virtually all liver diseases, particularly in cases of acute hepatitis of a variety of origins.10 Further studies are warranted to elucidate the diagnostic utility of serum keratin-18 and its fragments as peculiar markers of liver disease.