Human exposure to kava differs significantly in terms of the dosing range, the dosing frequency and, more importantly, the methods of preparation and thus the chemical composition of kava products. These variations are expected to result in different safety profiles. Historically, there have been limited safety disputes surrounding traditional kava use [ 15 ]. Hepatotoxic risk has become a central topic since the early 2000s, due to a number of hepatotoxicity cases reported in Western countries [ 35 165 ]. These adverse events led to a ban on kava use as an herbal anxiolytic drug in Germany between 2002 and 2014 [ 27 ]. The U.S. Food and Drug Administration (FDA) also issued an advisory in March 2002. A number of causes have been hypothesized, including the potential use of low-quality cultivars, the adulteration of non-root parts of the plants, the improper handling and storage of kava materials, the potential for drug–herb interactions, kava overdosing, and others. However, none of these proposals have been validated because the product(s) associated with the purported cases have never been accurately documented. Kava’s hepatotoxic potential has therefore also been proposed as an idiosyncratic phenomenon [ 15 ]. With these challenges in mind, this section mainly focuses on kava’s safety data in relation to human exposure. A few lab animal results published during the past 15 years are discussed to cross-inform potential mechanisms of such toxicity issues. Results using biochemical and in vitro models are not reiterated herein because they have been covered in a number of reviews, including two recent ones [ 68 157 ].
6.1. Kava Safety in Lab Animals
108,115,154,115, Our group evaluated the potential benefits and risks of kava use since 2008. The safety of an ethanolic extract of kava from Gaia Herbs was monitored in several animal models with kava’s cancer prevention potential as the primary research focus [ 107 166 ]. In all of these studies, ethanol and water solvent residues in the kava product were removed. The abundance of the six major kavalactones accounts for about 50% of the dry mass of this ethanolic extract [ 33 ]. Kava was administered to mice or rats in their diet at a level of 1.25–10 mg/g with an exposure period between 14–30 weeks [ 107 166 ]. Kava’s safety profile was monitored by bodyweight, liver weight, relative liver weight, serum ALT, AST, GGT or liver pathology with no signs of adverse effects. Similar results were observed when kava was evaluated in C57BL/6 female mice for 14 weeks at a dose of 500 mg/kg of bodyweight via daily oral gavage, gauged by serum ALT and AST [ 154 ]. Kava’s hepatotoxic risk has also been speculated to be associated with potential drug–herb interactions [ 157 167 ]. Kava and its ingredients have been reported to modulate drug-metabolizing enzymes, particularly CYP enzymes, making drug–herb interactions mechanistically plausible. Since no signs of liver toxicity were observed in three different strains of rodents with kava dosages significantly higher than human exposure, we also explored the potential interaction of kava with the liver toxin drug acetaminophen (Tylenol) as a model compound. Surprisingly, even three days of kava pre-exposure greatly enhanced the hepatotoxicity of acetaminophen in C57BL/6 mice [ 154 ]. Through fractionation, this model identified flavokavains A and B as the potential ingredients for enhancing liver toxicity of acetaminophen. Nonetheless, the clinical relevance of these lab animal results remains an open question because acetaminophen exposure was documented in only one of the kava hepatotoxicity cases. It also remains to be determined whether kava enhances the hepatotoxic risk of other liver toxins, such as alcohol consumption, which may be more clinically relevant based on recent trial results from Sarris et al. [ 98 ].
The most comprehensive lab animal safety evaluation of kava was performed by the U.S. National Toxicology Program (NTP) [ 168 ]. Both genders of F344/N rats and B6C3F1 mice were evaluated with 5-day weekly kava gavage exposure for 2 weeks, 3 months, and 2 years. In these studies, “Kava Kava Extract” was obtained from Cosmopolitan Trading Co (Seattle, WA, USA). Interestingly, the product was described as a medium yellow powder at a density of 0.46 g/mL. No detailed preparation procedures could be found in the report about this kava product. Based on the description of the product, the “Kava Kava Extract” was likely ground kava powder instead of an extract form, because ethanolic extracts, like the one from Gaia Herbs, are an oil form of high viscosity and the acetonic extract is expected to have similar physical properties. If the product used in these NTP studies was indeed kava root powder, the results need to be interpreted again with caution because the powder, administered whole and directly to the rodents, is substantially different from the traditional and anxiolytic kava forms that humans use. The powder would contain all chemicals in kava, particularly those hydrophobic compounds, such as flavokavains A and B, that are of low abundance in the traditional form of kava. The pharmacokinetics in the powder matrix are also expected to be substantially different. Nonetheless, in the 2-week study, kava was given at dosages of 0, 0.125, 0.25, 0.5, 1.0, or 2.0 g/kg of bodyweight for both mice and rats. Increased liver weight was observed when the kava dose was 0.5 g/kg of bodyweight or higher. For the 3-month study, kava dosages were the same as the 2-week study and a liver weight increase was observed when the kava dose was 0.25 g/kg or higher. In the 2-year study, kava doses were 0, 0.1, 0.3, or 1.0 g/kg of bodyweight. Changes in liver weight were observed at a dose of 0.3 g/kg or higher. Pathological lesions in livers were only observed when the kava dose was 1.0 g/kg in rats. In short, the dosages of kava affecting liver weight and pathology in these studies are unlikely to be physiologically relevant as a life-long consumption of 4–5 g of neat kava root daily would be required for a human of 75 kg bodyweight. A few follow-up studies further analyzed the liver tissues from these studies, including immunohistochemical analyses of several CYP enzymes [ 169 ] and global gene expression [ 170 171 ]. No significant changes were observed when the kava dose was 0.3 g/kg or lower.
Based on these long-term exposure results, kava appeared to be well tolerated in rodents, particularly at a dosage not significantly higher than the human equivalent exposure. However, the products used in these studies are not well characterized, including our own studies, which need to improve in the future. In addition, there have been no lab animal safety studies with aqueous kava preparations, which more closely mimic the traditional kava formula.