Curcumin is a yellow–orange pigment obtained from the plant Curcuma longa. The powdered rhizome of this plant, called turmeric, is a common ingredient in curry powders and has a long history of use in traditional Asian medicine for a wide variety of disorders. In the last decade a large number of reports have been published on the beneficial effects of curcumin, and it has repeatedly been claimed that this natural product is efficient and safe for the prevention and treatment of several diseases including cancer.1–3 It is not surprising, therefore, that curcumin is currently sold as a dietary supplement and that numerous clinical trials are ongoing or recruiting participants to evaluate curcumin activity. But there is accumulating evidence that curcumin may not be so effective and safe. Because such evidence is not generally acknowledged, the purpose of this letter is to briefly review the negative properties of curcumin so that they can be balanced against its beneficial effects.

Most of the evidence that supports the therapeutic potential of curcumin is mainly based on in vitro studies in which curcumin was tested at concentrations in the micromolar range. Several reports have demonstrated, however, that the plasma concentrations of curcumin in people taking relatively high oral doses of this compound are very low, typically in the nanomolar range (reviewed in Ref. 4). For instance, a recent study examined the pharmacokinetics of a curcumin preparation in 12 healthy human volunteers 0.25–72 hr after an oral dose of 10 or 12 g. Using a high-performance liquid chromatography assay with a limit of detection of 50 ng mL⁻¹, only 1 subject had detectable free curcumin at any of the time points assayed.5 The fact that curcumin also undergoes extensive metabolism in intestine and liver6, 7 means that high concentrations of curcumin cannot be achieved and maintained in plasma and tissues after oral ingestion. This is a major obstacle for the clinical development of this agent and suggests that the therapeutic potential of oral curcumin is limited. The low clinical efficiency of curcumin in the treatment of several chronic diseases such as Alzheimer's disease and cardiovascular diseases has been discussed recently.8

As far as cancer is concerned, in vitro studies have demonstrated that cancer cells do not die unless they are exposed to curcumin concentrations of 5–50 μM for several hours.4, 9, 10 Because of its poor bioavailability, these concentrations are not achieved outside the gastrointestinal tract when curcumin is taken orally. Because of its extensive metabolism in intestine and liver, these concentrations cannot be maintained for several hours in the gastrointestinal tract. This suggests that the chemotherapeutic potential of oral curcumin is limited even for the treatment of cancers of the gastrointestinal tract. Accordingly, when 15 patients with advanced colorectal cancer were treated with curcumin at daily doses of 3.6 g for up to 4 months, no partial responses to treatment or decreases in tumor markers were observed.11

A search of the website www.clinicaltrials.gov in July 2009 showed 34 clinical trials using curcumin in a wide variety of diseases, particularly in cancer. In some of these trials, patients with several types of cancer are receiving or will receive curcumin through the oral route. For instance, in an ongoing Phase II clinical trial (NCT00094445), participants with pancreatic cancer are receiving 8 g of curcumin by mouth every day for several 8-week-periods. As discussed before, the plasma concentrations of curcumin in people taking relatively high oral doses of curcumin are very low, typically in the nanomolar range. This means that the oral administration of curcumin does not lead to cytotoxic concentrations outside the gastrointestinal tract. If one assumes that tumor cell death is necessary to achieve an efficient therapeutic response, one should not expect a very positive outcome from this trial. A Phase II Trial is also recruiting participants to test if a daily oral dose of 8 g of curcumin can improve the efficacy of the standard chemotherapy gemcitabine in patients with locally advanced or metastatic adenocarcinoma of the pancreas (NCT00192842). The rationale for this trial is based on in vitro and in vivo data that suggest that noncytotoxic concentrations of curcumin may sensitize cancer cells to the effects of anticancer drugs such as gemcitabine.4, 12 Although a daily dose of 1 g kg⁻¹ of curcumin increased the antitumor effects of gemcitabine in an orthotopic model of pancreatic cancer,12 this dose of curcumin (e.g. 70 g in a 70-kg person) is almost 10 times higher than that used in the clinical trial testing the combination of curcumin and gemcitabine (8 g). This makes the outcome of this trial uncertain, as curcumin can either increase or reduce the efficiency of chemotherapy depending on the concentration at which it is used.4, 13

Several strategies have been proposed to overcome the low oral bioavailability of curcumin.4, 14–18 One of these strategies has entered clinical trials and consists of using the black pepper alkaloid piperine (bioperine) to increase the bioavailability of curcumin.14 This strategy, however, should be used cautiously, as piperine is a potent inhibitor of drug metabolism and may cause toxicity in people taking specific drugs.8, 19, 20 In addition, it is important to note that any strategy that increases the levels of curcumin in tissues will not only increase the effectiveness of curcumin, but also its toxicity.

A relatively high number of reports suggests that curcumin may cause toxicity under specific conditions. In 1976 Goodpasture and Arrighi found that turmeric caused a dose- and time-dependent induction of chromosome aberrations in several mammalian cell lines; these alterations were observed at concentrations of 10 μg mL⁻¹.21 Accumulating data have demonstrated since then that curcumin can induce DNA damage and chromosomal alterations both in vitro and in vivo at concentrations similar to those reported to exert beneficial effect.22–30 For instance, curcumin concentrations of 2.5 and 5 μg mL⁻¹ were shown to induce DNA damage to both the mitochondrial and nuclear genomes in cells.28 These reports raise concern about curcumin safety, as the induction of DNA alterations is a common event in carcinogenesis.

In 1993 the National Toxicology Program (USA) published an extensive report on the toxic and carcinogenic properties of an organic extract of turmeric, called turmeric oleoresin.31 This extract is commonly added to food items and contains a percentage of curcumin (79–85%) similar to that of commercial grade curcumin. Rats and mice were fed diets containing several concentrations of turmeric oleoresin for 3 months and 2 years, and the possible toxic and carcinogenic effects were evaluated. In the 2-year feeding studies, turmeric oleoresin ingestion was associated with increased incidences of ulcers, hyperplasia, and inflammation of the forestomach, cecum and colon in male rats and of the cecum in female rats. In female mice, ingestion of diets containing turmeric oleoresin was associated with an increased incidence of thyroid gland follicular cell hyperplasia. The report also concluded that there was equivocal evidence of carcinogenic activity in female rats, female mice, and male mice. These conclusions were based on increased incidences of clitoral gland adenomas in female rats, hepatocellular adenomas in female mice, and carcinomas of the small intestine and hepatocellular adenomas in male mice. The increased incidence of carcinomas of the small intestine was observed in mice taking average daily doses of curcumin of ∼0.2 mg kg⁻¹ body weight.31 A recent report has also shown that curcumin can promote lung cancer in mice.32

These negative effects of curcumin may be mediated by several possible mechanisms. Evidence suggests that reactive oxygen species (ROS) such as superoxide anion and hydrogen peroxide may play an important role in carcinogenesis.33–36 This evidence is based on the facts that (i) ROS can induce cell malignant transformation,37–40 (ii) cancer cells commonly have increased levels of ROS,41–43 and (iii) the malignant phenotype of cancer cells can be reversed by reducing the cellular levels of ROS.40, 44–48 Experimental studies have demonstrated that, although low concentrations of curcumin induce antioxidant effects, higher concentrations of this compound increase the cellular levels of ROS.4, 9, 23, 28, 49–53 The presence of 2 α,β-unsaturated ketones in the chemical structure of curcumin may also mediate some of its negative properties. These chemical groups are known to react covalently with exposed thiol groups of cysteine residues of proteins through a reaction termed Michael addition. This reaction may explain, for instance, why curcumin generates ROS by irreversibly modifying the antioxidant enzyme thioredoxin reductase,49 why curcumin induces topoisomerase II-mediated DNA damage,10, 54, 55 and why curcumin inactivates the tumor suppressor protein p53.56, 57

Several other lines of evidence raise concern about curcumin safety. Curcumin was recently found to be an active iron chelator in vivo and to induce a state of overt iron deficiency anemia in mice fed with diets poor in iron.58 This suggests that curcumin has the potential to affect systemic iron metabolism, particularly in people with suboptimal iron status.58, 59 Curcumin has also been shown to inhibit the activity of the drug-metabolizing enzymes cytochrome P450, glutathione-S-transferase, and UDP-glucuronosyltransferase.8, 60–62 The inhibition of these enzymes in people taking curcumin may lead to an undesired increase in the plasma concentrations of some drugs and cause toxicity.8

Although this letter is focused on the negative properties of curcumin, it is important to note that there is evidence suggesting that curcumin has beneficial properties and could be developed into a drug for the prevention and treatment of diseases such as cancer.1–4 There are indeed numerous reports showing beneficial effects of curcumin in colon cancer, yet a potential role for curcumin in colon cancer has been shown in few clinical studies only.63In vivo data suggest that curcumin may also exert beneficial effects in organs outside the gastrointestinal tract;4 however, since curcumin does not reach these organs at high concentrations, it is not clear at this point whether these beneficial effects are due to low levels of curcumin, to metabolites of curcumin or to an unknown indirect effect. To develop curcumin into a preventive or therapeutic drug, it is important to consider the dose levels which elicit desirable/undesirable effects. Based on human studies, it is generally recognized that curcumin does not cause significant short-term toxicity at doses up to 8 g day⁻¹; this dose of curcumin is considered suitable for trials in which curcumin is administered for short periods of time. This dose of curcumin is not completely harmless, however, as human studies have shown that curcumin at doses ranging from 0.9 to 3.6 g day⁻¹ for 1–4 months originates some adverse effects including nausea and diarrhea and causes an increase in serum alkaline phosphatase and lactate dehydrogenase.11 Curcumin could be used at doses higher than 8 g day⁻¹ in situations in which no effective therapies exist (e.g. advanced pancreatic cancer), as toxicity is acceptable in these situations. But no human studies have been conducted to date to test the dose levels which cause long-term toxicity. This information would be valuable in the design of chemoprevention trials, in which chemopreventive agents are administered for long periods of time and toxicity is not always acceptable. Epidemiological data suggest that the lower incidence of gastrointestinal cancers in India may be due to diets rich in curcumin, and it has been estimated that the doses of curcumin in people consuming high amounts of turmeric in their diet are ∼0.15 g day⁻¹.63 In the absence of long-term toxicity studies in humans, this dose may be considered suitable when curcumin is used for long periods of time. This dose of curcumin is similar to that recommended by the World Heath Organization,31 but ∼10 times lower than that generally recommended by suppliers of curcumin supplements. The lack of long-term toxicity studies in humans should not only be considered by health professionals, but also by people taking supplements of curcumin (with and without piperine) that are readily available in the market.

It is unfortunate that curcumin is regarded in the scientific literature as efficient and safe when its efficiency and safety have not yet been proven. The fact that curcumin is a common dietary constituent is not enough to prove its safety, as other common dietary constituents have shown toxicity when used as dietary supplements.64 The fact that no major toxicity has been found in short-term studies in humans is not a proof of curcumin safety either. For a drug to be safe, it must also be devoid of long-term toxicity, and the most complete long-term toxicity study conducted to date raises concern about curcumin safety.31 In addition, the effectiveness of a drug is usually established by randomized, placebo-controlled, double-blind clinical trials, and no such trials have shown curcumin to be effective so far.8 Finally, the fact that the number of studies showing positive effects of curcumin is much higher than that showing negative effects does not necessarily mean that the benefit-risk ratio of curcumin is shifted towards the benefit; it may just indicate that there are more researchers evaluating the beneficial effects of curcumin than evaluating its toxicity. It is the opinion of the authors that future research is needed to establish the benefit-risk profile of curcumin. In the meantime, we believe that it is important to acknowledge the negative properties of curcumin so that the use of curcumin is not based on unbalanced information.