Spontaneous reports of assumed herbal hepatotoxicity by black cohosh: is the liver-unspecific Naranjo scale precise enough to ascertain causality?

Purpose Causality assessment of cases with herbal hepatotoxicity represents a major regulatory challenge and included, in the past, the application of a diagnostic algorithm consisting of causality evaluation methods with either liver-specific or liver-unspecific characteristics. To evaluate various causality assessing methods in cases with suspected herbal hepatotoxicity, two different scales were now used for reasons of comparison. Methods We used the liver-specific scale of the updated Council for International Organizations of Medical Sciences (CIOMS) as well as the Naranjo scale that is not organ specific and therefore not liver specific. Both scales were applied to 22 cases of spontaneous reports with initially assumed herbal hepatotoxicity caused by black cohosh, used for menopausal symptoms. Results The analysis shows that causality was either unlikely (n = 6) or excluded (n = 16), using the updated CIOMS scale. There were various confounding variables: pre-existing liver diseases (n = 6) including genuine autoimmune hepatitis or alcoholic or cardiac hepatopathy; hepatotoxicity induced by interferon or fluoxetine (n = 2); marginally increased serum activities of alanine aminotransferase (n = 2) or gamma-glutamyltranspeptidase (n = 2) of unassessable causality; a mixed group consisting of unassessable cases (n = 6) and cases with questionable, poorly documented hepato-biliary diseases (n = 3); and rosuvastin-induced rhabdomyolysis (n = 1). These confounding factors were not recognized by the Naranjo scale. Conclusions Structured hepatotoxicity-specific causality assessment methods such as the updated CIOMS scale are the preferred tools for causality assessment of assumed herbal hepatotoxicity and should replace the liver-unspecific Naranjo scale. Applying the updated CIOMS scale to cases with initially assumed hepatotoxicity by BC, causality was now found either unlikely or excluded. Copyright (C) 2011 John Wiley & Sons, Ltd.