Curcumin has broad spectrum chemopreventive activity in preclinical studies and appears to be safe in animal and human studies [3, 12, 13]. For chemopreventive interventions to be successful, they must be provided in doses that are effective, but be nearly free of toxicities. Several studies have demonstrated minimal toxicity with moderate doses of curcumin given in various formulations. Soni and Kuttan  administered 500 mg capsules of a 98% pure curcumin formulation to 10 volunteers daily for 7 days and reported no clinical toxicity. Two trials evaluating the efficacy of turmeric or curcumin for the treatment of arthritis or postoperative inflammation found that doses of 1,200 to 2,100 mg of curcumin per day for 2–6 weeks were without adverse effects [15, 16]. Cheng et al  reported no treatment-related toxicity up to 8,000 mg/day using a 99.3% pure curcumin tablet (500 mg of curcumin in each tablet). However, doses up to 12,000 mg were unacceptable to patients due to the bulky volume of the tablets. In the present study, we recognized minimal toxicity up to 12,000 mg in a single dose of standardized powder extract (C
Complex™, Sabinsa Corporation) obtained from Alleppey finger turmeric (Table 1). Thus, to our knowledge, no maximum tolerated dose has been identified in humans, although successful administration may be affected by differences in drug formulation.
Low serum levels of curcumin after single dose administration is consistent with the previously reported pharmacokinetic studies in animals and humans [7–9, 17–20]. Wahlstrom et al  administered 1 g/kg to Sprague-Dawley rats and found 65–85% of the dose excreted unchanged in the feces with negligible amounts in the urine. Studies with deuterium and tritium-labeled curcumin given orally to rats resulted in most of the detectable radioactivity in the feces [18, 20]. Negligible amounts of radioactivity were detectable in plasma when smaller doses of [3H]-curcumin were administered . Sharma et al  administered Curcuma extract 440 to 2,200 mg/day (36 to 180 mg of curcumin) for up to 29 days to patients with advanced colorectal cancer and failed to detect curcumin or its metabolites in blood or urine. Serum concentrations may be dependent on the dose that is administered. Cheng et al  determined the concentration of curcumin by HPLC in 25 subjects with high risk cancer lesions. The peak serum concentration after 4,000, 6,000, and 8,000 mg were 0.51 ± 0.11 μM, 0.64 ± 0.06 μM, and 1.77 ± 1.87 μM, respectively, but doses below 4,000 mg were barely detectable. These findings are consistent with the present study, in which low levels of curcumin were measured only at doses ≥8,000 mg (Table 2).
Low serum levels of curcumin may be due in part to the extensive intestinal and hepatic metabolic biotransformation. Preclinical work has demonstrated that avid sulfation, glucuronidation, and reduction of curcumin occurs in the gastrointestinal tracts of rats and humans [19, 21–23]. A challenge for future chemopreventive strategies lies in the conflicting evidence of the biological effects of the resultant metabolites. For example, lymphocytic glutathione S-transferase activity, a potential surrogate biomarker of curcumin activity, was significantly decreased in patients taking 440 mg/day despite the lack of measurable serum curcumin . This finding suggests that curcuminoid metabolites, which may not have been detected, resulted in a systemic biological effect. In contrast, experiments by Ireson et al  suggest that the metabolism of curcumin generates species with reduced ability to inhibit COX-2 expression compared to the parent compound. Given these conflicting data, the poor bioavailability of curcumin, and the fact that the gastrointestinal tract is exposed to the greatest concentration of unmetabolized curcumin, colorectal cancer chemoprevention is the most attractive area for future efforts.
To further elucidate the cause of low serum curcumin levels after large dose consumption in the present study, the curcumin capsules were analyzed. Each capsule was found to be 75% curcumin, 23% demethoxycurcumin and 2% bisdemethoxycurcumin, which is consistent with other commercial curcuminoid products. However, the curcumin activity as measured by HPLC area count for a weighed portion of capsular content was 66% less than an equivalent weighed portion of a previously supplied, pure curcumin powder (unpublished data). Thus, curcumin bioavailability in humans and biologic activity might differ based upon the formulation used.