We read with great interest the article by Mannaerts et al.,1 who demonstrated the inhibitory effect of valproic acid (VPA) on hepatic stellate cell activation and consequent liver fibrosis in a chronically injured mouse liver. The authors' initial concept is rather surprising because VPA, one of the most widely prescribed anticonvulsants, has been considered a hepatotoxic agent.2 The successful results of this study may make clinicians expect implications for therapeutic options for liver fibrosis. Although they have presented hopeful information, we as pathologists insist that one must carefully ponder the risk-benefit balance.
There have been many reports warning of the hepatotoxicity of VPA, including microvesicular steatosis, steatohepatitis, hepatocellular carcinomas, and even fulminant hepatitis.3, 4 We have also experienced an autopsy case of severe steatohepatitis related to chronic VPA administration. The patient, a 22-year-old man with an approximately 14-year medical history of VPA administration under a diagnosis of Lennox-Gastaut syndrome, died of fatal aspiration pneumonia. Most autopsy findings were concordant with clinical diagnoses, except for a markedly fatty liver. The liver was tender and yellowish and weighed 955 g. A cut surface of the liver showed a peculiar nodular appearance (Fig. 1A). Histological examination revealed a hallmark of incomplete septal cirrhosis associated with steatohepatitis (Fig. 1B). Because he was a nonobese nondrinker and was negative for any hepatitis viruses, it was concluded that VPA was the most likely cause of steatohepatitis in this case. The liver disorder had been subclinical (serum aminotransferases had been almost normal until just before death); therefore, an accurate diagnosis could not have been made while he was alive. However, the liver dysfunction might have contributed to the development and exacerbation of his pneumonia.
Our concern is that the authors did not mention predictable adverse findings. They certainly demonstrated the absence of a VPA effect on the serum aminotransferase levels of the examined mice, but they did not report histopathological findings other than liver fibrosis shown by Sirius red–stained sections. Were steatotic liver disorders not seen in the liver tissues of the mice as expected? The authors have a responsibility to pay maximum attention to the hepatotoxic effects of VPA before the clinical use of VPA or its derivatives is begun. According to our experience, VPA can induce even liver fibrosis via undetectable persistent inflammation and steatosis. We believe that they need to check this point. As an anticonvulsant, VPA is being used worldwide currently because its beneficial effect with respect to the prevention of seizure is considered to be greater than the risk of hepatic injury. Does VPA administration provide sufficient benefits to liver fibrosis patients?