Address for correspondence: Karen Brown, 502 RKCSB, Department of Cancer Studies and Molecular Medicine, University of Leicester, Leicester, United Kingdom. email@example.com
An expanding body of preclinical evidence suggests resveratrol has the potential to impact a variety of human diseases. To translate encouraging experimental findings into human benefits, information is first needed on the safety, pharmacokinetics, pharmacodynamics, and, ultimately, clinical efficacy of resveratrol. Published clinical trials have largely focused on characterizing the pharmacokinetics and metabolism of resveratrol. Recent studies have also evaluated safety and potential mechanisms of activity following multiple dosing, and have found resveratrol to be safe and reasonably well-tolerated at doses of up to 5 g/day. However, the occurrence of mild to moderate side effects is likely to limit the doses employed in future trials to significantly less than this amount. This review describes the available clinical data, outlines how it supports the continuing development of resveratrol, and suggests what additional information is needed to increase the chances of success in future clinical trials.
An increasing body of evidence from preclinical studies suggests resveratrol has the potential to impact a variety of human diseases. To translate encouraging experimental findings into human benefit, information is first needed on the safety, pharmacokinetics, pharmacodynamics, and ultimately, clinical efficacy of resveratrol. Valuable data are accumulating on the safety and tolerability of resveratrol in humans, while the clinical pharmacokinetic and metabolism profiles are becoming reasonably well-defined. However, there are currently no published demonstrations of therapeutic or protective effects of resveratrol in appropriately designed clinical trials. Furthermore, there is presently very little evidence of pharmacological activity in terms of molecular or biochemical changes in humans that might be useful as surrogate efficacy biomarkers; this is an area that urgently needs addressing. A review of the database (http://clinicaltrials.gov/) reveals a total of 16 studies involving resveratrol that are either active or recruiting, plus six more that have recently been completed.1 These include trials aimed at investigating the potential role of resveratrol in the management of type 2 diabetes, obesity, Alzheimer's disease, and cancer. The following paper reviews the available clinical data, outlines how it supports the continuing development of resveratrol, and suggests what additional information is needed.
Resveratrol pharmacokinetics and metabolism in humans
A number of clinical investigations have assessed the pharmacokinetics and metabolism of resveratrol in humans following oral ingestion of either the single synthetic agent or as a constituent of a particular food or drink.2–7 These studies, which are summarized in Table 1, have employed a range of doses and administration schedules, as well as a variety of formulations. An understanding of the clinical pharmacokinetics and metabolism of resveratrol is essential to enable appropriate doses to be chosen for larger efficacy trials, for defining relevant concentrations that should be used in preclinical mechanistic studies, and for identifying metabolites that may contribute to or diminish the activity of the parent compound.
Table 1. Summary of published clinical trials involving resveratrol. The figures in parentheses (column 1) refer to the number of participants in each study
Form of resveratrol
Healthy males (12)
Delivered in white wine, white grape juice or vegetable juice
25 mg/70 kg
Resveratrol absorption was similar in all three matrices
Among the first of these pharmacokinetic investigations was that by Goldberg et al., which involved oral administration of resveratrol (25 mg/70 kg) to healthy male subjects in three different matrices, white wine, white grape juice, and vegetable juice, to examine the influence on absorption.8 Blood samples were taken up to 4 h after ingestion, and the highest recorded serum resveratrol/metabolite levels were consistently achieved after 30 min. Concentrations then declined rapidly, reaching baseline levels within 4 h. Free resveratrol accounted for only a small fraction of the total dose in plasma (1.7–1.9%), with glucuronide and sulfate conjugates dominating the profile in both plasma and urine. The total absorption and peak concentrations of resveratrol related species were broadly equivalent in aqueous and alcoholic matrices. However, the low levels of free resveratrol attained (<40 nmol/L) led the authors to conclude that circulating concentrations generated through consumption of dietary sources would be inadequate to elicit biological effects, based on that needed for activity in cultured cells in vitro (5–100 μM).8
A study comparing resveratrol pharmacokinetics after administration of grape juice, which predominantly contains glucosides, or as the pure free aglycone dissolved in a small amount of whisky and diluted with water,9 essentially reinforced the findings of the Goldberg investigation. In addition, the results suggested that glycoside forms of resveratrol are absorbed to a lesser extent than the aglycone.8
In a more comprehensive study, the absorption, bioavailability, and metabolic fate of resveratrol was traced through the use of radioisotope labeling.10 Following oral ingestion of a dietary relevant amount (25 mg) of [14C]-resveratrol by six healthy volunteers, at least 70% of the dose was absorbed, which is unusually high for a dietary polyphenol.10 However, despite relatively efficient absorption, the bioavailability of unchanged resveratrol was very low, due to rapid and extensive metabolism. Peak plasma levels of radioactivity equated to ∼2 μM total [14C]-labeled species, but the vast majority of this was due to the presence of phase II metabolites; only trace amounts of unchanged resveratrol (<5 ng/mL) could be detected. A secondary peak, observed at 6 h after oral dosing was indicative of enterohepatic recirculation of conjugated metabolites by reabsorption after intestinal hydrolysis.
In the same study, urinary analysis revealed that most of the radioactivity following oral dosing was recovered in the urine (53–85%) with the proportion excreted in the feces highly variable (0.3–38%).10 Administration of a higher unlabeled oral dose (100 mg) to one subject allowed characterization of metabolites eliminated via the urine by high performance liquid chromatography-mass spectrometry (HPLC-MS) and demonstrated the involvement of three transformation pathways. Five major metabolites were identified: two isomeric resveratrol monoglucuronides, a dihydroresveratrol monoglucuronide, a resveratrol monosulfate, and a dihydroresveratrol sulfate. Intestinal microflora were suggested as the potential source of hydrogenated resveratrol metabolites formed by saturation of the aliphatic double bond. Sulfate metabolites excreted in the urine accounted for approximately 24% of the dose administered, and this conjugation reaction was proposed as the rate limiting step governing the poor bioavailability of resveratrol, whereas glucuronides contributed about half this much. The remainder seemed to consist largely of early eluting, polar derivatives of unknown structure.10
A recent phase I dose escalation study evaluated the safety and pharmacokinetics of resveratrol administered as a single dose (0.5, 1.0, 2.5, or 5.0 g) in healthy volunteers.11 Plasma and urine concentrations of resveratrol and metabolites were determined by HPLC with UV detection and structurally identified using HPLC-tandem mass spectrometry (LC-MS/MS). In addition, fecal samples were collected for analysis. At the highest dose, the average peak plasma concentration (Cmax) of resveratrol was 539 ± 384 ng/mL (mean ± SD, n= 10), which equates to ∼2.4 μM. In comparison, the Cmax levels of two monoglucuronides (resveratrol-3-O-glucuronide and resveratrol-4′-O-glucuronide) plus resveratrol-3-O-sulfate were three- to eight-fold higher (Fig. 1). Furthermore, the area under the plasma concentration versus time curve (AUC) for these metabolites was up to 23-fold greater than for resveratrol, indicating considerably higher systemic exposure to the conjugates. Additional metabolites detected were resveratrol-4′-O-sulfate and resveratrol disulfate, although these were relatively low in abundance. Renal excretion of resveratrol and its metabolites was rapid, with 77% of all urinary agent-derived species excreted within 4 h of consuming the 0.5 g dose. The appearance of a secondary resveratrol peak in the plasma, together with the detection of a large amount of parent compound in the feces compared to metabolites, is again consistent with enterohepatic circulation suggested in the earlier [14C]-resveratrol study by Walle et al.10
In a follow up study, resveratrol was administered at the same doses for 29 days to assess pharmacokinetics following repeated dosing.12 Resveratrol and metabolite levels were measured by HPLC-UV in plasma pre-dose and at 0.25, 1.0, 1.5, 5, 12, and 24 h post-dose on a day between the 21st and the 28th day of dosing. The mean average plasma concentration (Cav) and Cmax values of parent resveratrol across the four groups increased with dose and ranged from 0.04–0.55 μM and 0.19–4.24 μM, respectively. Using this particular regimen, similar plasma concentrations of resveratrol and its main metabolites were attained following multiple or single dosing;11,12 significant accumulation with repeated ingestion was only evident at the highest dose. Resveratrol-3-O-sulfate, resveratrol-4′-O-glucuronide, and resveratrol-3-O-glucuronide were the major plasma metabolites, with Cmax values between 2.4- and 13-fold greater than resveratrol itself. AUC values for these metabolites also exceeded those of the parent, in the case of resveratrol-3-O-sulfate by up to 20-fold. The apparent total body clearances and volumes of distribution were calculated following urinary analysis and were consistent with the rapid metabolism and low bioavailability of resveratrol.
An important aspect of resveratrol pharmacokinetics is assessing whether it reaches the proposed sites of action after oral ingestion in humans and determining the concentrations attained in target tissues. Although this has been studied in detail in rodents,13,14 it clearly presents challenges in humans. A recent clinical trial conducted in colorectal cancer patients has begun to address these issues.15 Recruited individuals (10 per group) consumed either one or two 500 mg resveratrol caplets daily for eight consecutive days prior to resection. During surgery, samples of tumor and adjacent sections of apparently normal colon tissue were obtained for analysis. In addition to resveratrol, six metabolites were identified in the tissue by HPLC-UV and/or LC-MS/MS: resveratrol-3-O-glucuronide, resveratrol-4′-O-glucuronide, resveratrol-3-O-sulfate, resveratrol-4′-O-sulfate, resveratrol disulfate, and resveratrol sulfate glucuronide. In contrast to the pattern in plasma of healthy volunteers12 where the sulfate glucuronide is a minor component or more commonly absent, it was a predominant metabolite in tissue from 14 out of 20 patients.15 There was substantial variation in the tissue concentrations measured, both in different samples from the same patient and among individuals. When all ten patients on each dose were considered together, the highest mean concentration of resveratrol was found in normal tissue localized proximal to the tumor on the right side, where it reached ∼19 and 674 nmol/g, after the 0.5 and 1.0 g interventions, respectively. The corresponding values in tumor tissue were lower at ∼8 and 94 nmol/g (or μmol/L assuming 1 mL has a mass of 1 g). Importantly, the levels of free resveratrol detected in many cases exceeded the concentrations that have been widely reported to have activity in numerous preclinical systems.16 Maximal mean tissue concentrations determined for resveratrol metabolites (in nmol resveratrol equivalents/g) were 86 for resveratrol-3-O-glucuronide at the 0.5 g dose level and 67 for resveratrol-3-O-sulfate in patients on 1.0 g resveratrol, both observed in normal right-sided colorectal tissue proximal to the tumor.15 Interestingly, levels of resveratrol and its metabolites were consistently higher in tissues originating in the right side of the colon compared to the left. This difference was best exemplified by one patient who had two tumors removed; the levels of resveratrol-derived species were considerably higher in the cecal (right-sided) tumor than the one excised from the sigmoid colon (left-sided). This phenomenon may be due to a higher resveratrol content of fecal matter passing through the right side of the colon relative to the left. In addition, during this passage the luminal contents become more solid, which may hinder the absorption of resveratrol.
Although high levels of resveratrol are achievable in the colon, the profile in other internal tissues might be expected to more closely reflect the situation in plasma. Therefore, whether resveratrol is active in these organs may be more dependent on the intrinsic activity of the metabolites present and/or their ability to regenerate the parent compound. Emerging data on the properties of resveratrol metabolites suggests that the 3- and 4′-O-sulfates engage several mechanisms consistent with anticancer activity.17,18 Although reports to date indicate that resveratrol glucuronides are inactive in vitro, they may provide a pool for local or systemic regeneration of resveratrol in vivo through β-glucuronidase mediated hydrolysis.19 In evaluating the potential health benefits of resveratrol it is therefore of key importance to determine the distribution and metabolite profile in other human target tissues and advance understanding of the inherent activity and pharmacokinetics of resveratrol metabolites.
Resveratrol safety and tolerability
Only a small number of clinical trials using resveratrol as a single-agent formulated as a medicinal product have formally addressed and reported on safety and tolerability.6,11,12,20,21 Only one of these studies included a placebo control group,21 so it is difficult to completely ascribe the adverse effects experienced to ingestion of resveratrol; however, a consistent pattern is evident that allows some conclusions to be drawn. The dose escalation study reported by Brown et al. describes toxicity data from 44 healthy volunteers (10–12 per group) who consumed resveratrol for 29 days at a daily dose of 0.5, 1.0, 2.5, or 5.0 g.12 Resveratrol was deemed safe, as borne out by the lack of serious adverse events detected by clinical, biochemical, or hematological indices, during both the intervention and 2 week follow-up phase. Overall, 28 participants reported at least one adverse effect that was considered possibly due to resveratrol, with the majority occurring in people taking the two highest doses. The most common toxicity was gastrointestinal, particularly diarrhea, nausea, and abdominal pain, all of which occurred only in individuals taking in excess of 1 g resveratrol per day. Typically, the onset of gastrointestinal symptoms was ∼1 h after caplet ingestion with improvement during the course of the day. Approximately, 90% of all events reported were graded as mild, according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE). In addition, there were four cases of moderate severity diarrhea. Based on the overall findings from this study, the authors recommended that daily doses of resveratrol for subsequent clinical evaluation should not exceed 1 g.12
Results from a similarly designed 4-week study by Chow et al. reinforce these data in that 1 g resveratrol taken once daily was generally well-tolerated in healthy participants.20 All reported adverse events were CTC grade 1 or 2 with many being mild and transient; the frequency of side effects experienced were consistent with those observed in the trial described by Brown et al.12 and shorter-term studies involving fractionated daily doses of resveratrol.6,21,22
Pharmacodynamic effects of resveratrol in humans
Considering the wealth of preclinical data on the potential mechanisms of action engaged by resveratrol there is relatively little published evidence of pharmacodynamic effects in humans. In the phase I study described by Brown et al., ingestion of resveratrol for four weeks by healthy volunteers caused a small but significant decrease in circulating insulin-like growth factor-1 (IGF-1) and IGF binding protein 3 (IGFBP-3), compared to pre-dosing values.12 The IGF signaling system influences malignant development; IGFs, which are mitogenic and antiapoptotic,23,24 can affect cell differentiation, neoplastic transformation, and metastasis.24–26 The IGF system is regulated by IGF binding proteins, particularly IGFBP-3, which binds IGFs in the extracellular milieu, reducing circulating levels and the potential for interaction with IGF receptors. The potential importance of IGF-1 in cancer is illustrated by several studies suggesting a direct relationship between levels of IGF-1 and risk of colorectal, prostate, breast, or lung cancer.27 The role of IGFBP-3 is less clear. Conventionally, this peptide is thought to have a protective effect due to sequestration of IGF-1; however, there are also data indicating that higher concentrations are associated with an increased rather than reduced risk of premenopausal breast cancer.28 Furthermore, although high levels have been inversely correlated with lung cancer risk in a meta-regression analysis, there was no such association with other cancers such as colon and prostate.28 In the volunteer study, when results from all trial participants were combined, consumption of resveratrol reduced IGF-1 and IGFBP-3 levels weakly compared to baseline, albeit significantly.12 However, when each dose group was considered separately, those taking 2.5 g resveratrol experienced the most prominent and consistent reductions in plasma IGF-1, whereas no significant changes were apparent with lower or higher doses (0.5, 1.0, 5.0 g). Mean IGFBP-3 concentrations in individuals on 1.0 or 2.5 g resveratrol were also significantly reduced but the other doses had no effect. The ability of resveratrol to decrease circulating IGF-1 and IGFBP-3 in humans may constitute an anti-carcinogenic mechanism. However, further studies are required to confirm or refute this association and, if appropriate, elucidate why the relationship appears to be nonlinear; an understanding of the underlying mechanisms would strengthen the potential value of IGF-1 and/or IGFBP-3 as biomarkers to monitor resveratrol efficacy.
Several other potential markers of activity were investigated in blood samples from the volunteers. Ingestion of resveratrol for 29 days failed to significantly affect circulating levels of prostaglandin E-2, reflecting perturbation of the arachidonic acid cascade, or influence leukocyte levels of the malondialdehyde-DNA adduct M1dG, which is a measure of DNA oxidation.12 With respect to activity in target tissues, resveratrol ingestion for 8 days reduced the proliferation of colorectal epithelial cells in cancer tissue compared to pre-dose levels, as assessed by Ki-67 positivity detected by immunohistochemistry.15 The effect was small and was only significant when data were combined from patients in both dose groups (0.5 and 1.0 g/day); however, it suggests resveratrol has the potential to favorably alter cell proliferation in humans.
The ability of resveratrol to protect against DNA damage and resulting initiating mutations has been linked to suppression of carcinogen metabolic activation and/or increased detoxification through modulation of enzymes involved in phase I or II transformations. However, it is possible that such changes may also affect the efficacy or toxicity of concomitant medications since the same enzymes are responsible for drug metabolism. To ascertain whether such interactions might be a potential problem for resveratrol, the effect of once daily dosing on the activity of drug and carcinogen metabolizing enzymes, phase I cytochrome P450s plus the conjugating enzymes glutathione S-transferase (GST) and UDP-glucuronosyl transferase (UGT) 1A1, was investigated in healthy volunteers.20 Subjects received resveratrol (1 g daily) for 4 weeks and when pre and post intervention activities were compared resveratrol was found to significantly inhibit the phenotypic indices of plasma CYP3A4, 2D6 and 2C9, while inducing 1A2. In addition, in subjects with low baseline values, intervention was associated with an induction of GST-π protein expression and UGT1A1 activity. Inhibition of CYP3A4 and 2C9 in particular, which metabolize a broad range of drugs, could lead to elevated plasma concentrations and increased likelihood of toxicity. The authors therefore acknowledged that although the enzyme modulation observed could be one of the mechanisms through which resveratrol inhibits carcinogenesis, doses of less than 1 g should be evaluated in future trials to minimize adverse drug interactions.20
A further demonstration of the capability of resveratrol to alter physiological parameters in humans was shown in a randomized double-blind placebo-controlled crossover study.29 Single oral doses of resveratrol (250 or 500 mg) were associated with a dose-dependent pattern of higher cerebral blood flow in the prefrontal cortex during task performance, as indexed by total concentrations of hemoglobin. This is the first indication that resveratrol can alter cerebral blood flow variables in humans and supports additional research on its effects on brain function.
It is extremely encouraging for the future clinical development of resveratrol that doses of up to 5 g/day, taken for a month, are safe and reasonably well-tolerated. However, it seems likely that the occurrence of dose-related mild to moderate side effects, coupled with the ability of resveratrol to alter the activity of drug metabolizing enzymes, will limit the dosage employed in future studies to <1.0 g/day. In humans, resveratrol is efficiently absorbed after oral administration; however, rapid phase II metabolism drastically limits its plasma bioavailability, and this may also prove to be the case for internal tissues. In the absence of distribution data for specific human tissues, the levels attained in plasma should guide the choice of concentrations used in preclinical in vitro studies to more accurately reflect the clinical scenario. Consequently, concentrations employed should not surpass ∼1 μM in order to mimic the resveratrol levels attainable through intake of 1 g/day or less.12 The presence of high concentrations of parent resveratrol in colorectal tissues, in excess of that required for activity in vitro, supports the colon as a target organ. The efficacy of resveratrol in other tissues may be largely dependent on whether its metabolites have significant activity or are able to regenerate resveratrol either locally or systemically. Alternatively, concentrations of resveratrol below the range typically studied in vitro, at least in cancer research, may be sufficient for activity. There is some evidence of biological activity at extremely low concentrations of resveratrol as well as biphasic dose–response relationships; these effects should be investigated further and the underlying mechanisms elucidated in order to help identify the optimum doses to be used in the next phase of clinical trials aimed at evaluating efficacy.30