Fatal fulminant hepatic failure induced by a natural therapy containing kava

In reply: Thomsen et al raise several points regarding our recent report of a death following ingestion of a preparation containing kava.1

The first is that the results of their analysis of the herbal preparation differ from those of the Therapeutic Goods Administration (TGA). The TGA found Passiflora incarnata in the tablets, whereas Thomsen and colleagues did not. The explanation for this lies in the fact that different batches of the product were analysed by the two laboratories. The analysis we reported relates to batch 13861, which was the source of the patient’s tablets. Thomsen and colleagues have analysed material from a different batch that may have contained different components.

Thomsen et al also raise the possibility that Teucrium, a known adulterant of Scutellaria, may have been present in the preparation. With respect to this issue, the TGA tested for both T. chamaedrys and T. canadensis and found none present. They also tested for the presence of the chemicals teucreoside and verbascoside, both of which are found in many Teucrium species. Neither was present. The raw materials used to make the tablets were also analysed, and no evidence of Teucrium substitution or contamination could be found.

The next point raised is whether the association between kava use and hepatotoxicity is causal, and they quote an analysis published in the journal Phytomedicine, in which the authors claimed that kava was the probable cause in only two of 68 reviewed cases. In marked contrast, the Journal of Hepatology, in July 2003, published an analysis of 36 cases of hepatitis associated with the use of kava.2 The review concluded that kava was the certain or probable cause of the hepatitis in 24 of the 36 cases. Worryingly, eight of these patients required liver transplantation.

Thomsen and colleagues then point to a lack of evidence of hepatotoxicity in studies in rats administered kavalactones.3 We believe it is inappropriate and misleading to suggest that a study of this kind could rule out the possibility that kava preparations cause life-threatening but uncommon idiosyncratic hepatotoxic reactions in humans.

Finally, Thomsen et al wonder whether the patient’s illness may have been due to chromium toxicity. However, the illness was not suggestive of chromium poisoning.4 There is no reason to believe that chromium was present at other than normal background levels.