Vitamin C and cancer revisited

ResearchBlogging.orgIt’s been quite a while since I wrote about this topic, but, quite frankly, I didn’t think anything new was likely to come up that would interest me sufficiently to take it on again. I was almost right; it’s been well over two years since the last time I discussed the issue of whether or not vitamin C has any role in treating cancer.

When last I left the topic, two studies had been released that were being widely cited as a “vindication” of Linus Pauling. As you may recall, Pauling was the Nobel Laureate who succumbed to what’s sometimes called the “Nobel disease” in that he turned into a major crank late in his life. The main manifestation of his crank tendencies was his major role in the founding of the orthomolecular medicine movement. So-called “orthomolecular medicine” generally involves treating all disease with megadoses of various vitamins and other nutrients, chief among which was vitamin C, also known as ascorbic acid. Thanks largely to Pauling’s influence and the power and influence that come with his having been a two-time Nobel Laureate, the concept that megadoses of vitamin C could be used to treat diseases mild and serious, ranging from the common cold to cancer, has hung around far longer than it would likely have if it were judged on a strictly scientific basis. There are endless studies, some of which seem to show a benefit in preclinical models but only at enormous concentrations, and there are clinical studies that fall into two categories: flawed studies that seem to show a benefit (one of which Pauling himself inflicted on the medical literature) or better-designed studies that show no effect. All in all, I remain rather puzzled at the continued interest in vitamin C as some sort of cancer cure, given that the largest effects ever reported were in fact quite modest and given how huge the doses necessary to achieve even those modest responses in animal models had to be. The whole thing strikes me as the proverbial long run for a short slide. But apparently there are a lot of people who like getting their butts dirty sliding into base, because the attraction to vitamin C still lives on. Never mind that there are charges that Pauling could not abide negative data to the point of suppressing them or that there were also data that suggested that vitamin C may may stimulate the growth of some tumors in mice. To many in the world of “complementary and alternative medicine” (CAM), vitamin C at large doses remains a “natural” cure for what ails you.

Most recently, I deconstructed two studies. One study was of the dreaded “best case” series, a type of study I absolutely detest because it’s an excuse for purveyors of dubious therapies to cherry pick their “best cases,” even if it’s a handful out of hundreds or thousands, and this one was no different. Worse, even as “best cases,” these cases were not all that impressive as evidence for a therapeutic effect due to vitamin C. The other study, by Qi Chen, Mark Levine, and colleagues at the National Institute of Diabetes and Digestive and Kidney Diseases at the NIH was published in PNAS. Sadly, it wasn’t all that striking either, being only an in vitro study looking at the effect of very high concentrations of vitamin C on various tumor cell lines in cell culture. All I could do was to wonder was whether this study was evidence for clinically useful anti-tumor activity due to ascorbic acid or just another example demonstrating that it’s possible to kill tumor cells in vitro with almost anything if you jack the concentration high enough. I will admit that it was interesting that the vitamin C-mediated tumor cell killing depended upon the generation of H2O2 (peroxide) and could be abrogated by inhibitors of H2O2 generation. In any case, Levine’s conclusion was that vitamin C/ascorbate is a prodrug that results in the selective generation of H2O2 in tumor cells but not in whole blood. All in all it was a mildly interesting result, but hardly a stunning vindication of Linus Pauling.

Now, over two years later, Levine et al are back, this time with another study published in PNAS entitled Pharmacologic doses of ascorbate act as a prooxidant and decrease growth of aggressive tumor xenografts in mice. This time, they’ve done animal studies. Once again, these studies show mildly positive results. Once again, the whole endeavor strikes me as a long run for a short slide. Once again, I remain puzzled as to why there is such intense interest in what, even if it “works,” strikes me as a cumbersome and only mildly effective therapy.

Let’s look at the study itself. There’s one interesting thing that stands out right away when I look at it, and that’s that the manuscript was received for review on May 1 and accepted for publication on June 6. That’s incredibly fast, and it makes me think that this article was communicated, rather than submitted to peer review. (The manuscript lists the article as having been “edited” by Academy member Bruce Ames.) One thing that most people don’t understand about PNAS is that any member of the National Academy of Sciences can submit a manuscript and essentially choose its reviewers, whether that manuscript is by the member or “communicated” to the journal for another investigator. Indeed, getting a paper published in PNAS is quite easy for an Academy member and incredibly difficult for a non-member who does not have the connections that allow him to line up an Academy member willing to act as referee and thus is forced to submit his manuscript directly to the journal. On the other hand, PNAS used to be used more often that many would like to admit as a dumping ground by Academy members for their lesser work because of how easily they could publish in it. (Indeed, Linus Pauling first published his truly execrable retrospective and unrandomized “study” purporting to show that high dose vitamin C prolonged the lives of patients with advanced cancer in PNAS in the 1970s.) In recent years, fortunately, the journal has tightened up editorial standards and peer review considerably. I only mention all of this because it’s been a bit of a pet peeve of mine about PNAS, and, although the quality of published manuscripts in PNAS is generally high, stinkers do sometimes sneak their way in there, thanks to the lax peer review standards for Academy members, some of whom are occasionally tempted to abuse the privilege of being able to publish essentially anything they want or by the occasional Academy member who turns crank, as Linus Pauling did.

But I digress. Back to the meat of the science, so to speak.

Oddities specific to submitting manuscripts to PNAS aside, how good is the study? Answer: It’s OK but it’s nothing special. In fact, it’s not particularly “meaty,” with only four figures, one of which is essentially a rehash and extension of data from Levine’s previous PNAS paper. Indeed, if this paper were not about a controversial and questionable therapy that had been thought to have been found wanting and instead been about a new cancer chemotherapeutic, my guess is that this result would have been viewed as mildly promising but nothing earthshaking. However, it’s not about some new chemotherapy agent; it’s about vitamin C. Consequently, it got a fair amount of attention in the press last week. Basically, this study reports the results of in vivo experiments in mouse tumor models using high dose vitamin C to treat tumors. In the introduction, Levine argues that the reason previous studies using high dose vitamin C to treat cancer failed to show an antitumor effect was because the drug was administered orally. The body, he points out, keeps a pretty tight control over the allowable blood concentration of vitamin C, and it’s not possible to boost that concentration above a certain level with oral dosing alone. If the blood level of ascorbate could be boosted far beyond the limit that can be achieved with oral dosing, Levine hypothesized, it might reach concentrations in the millimolar range, which is the concentration range in which ascorbate has been shown to kill tumor cells in vitro.

The important thing to remember here, though, is that at these doses ascorbate is not acting as a nutrient. It is not a “nutritional” treatment for cancer, as some CAM advocates claim. Rather, it is acting as a drug. Indeed, these are doses that are many, many times greater than what is required for good nutrition; they are true megadoses. To illustrate the point, I note that the doses used in mice by Levine et al were on the order of up to 4 g/kg twice daily. Translating that dose to the proverbial 70 kg man would mean a dose of up to 560 g per day, or nearly a pound and a quarter, administered intravenously. Of course, mice are not humans, and, as it turns out, it doesn’t take doses quite that high to reach blood concentrations in humans as high as those achieved in the mice using 4 g/kg ascorbate (more on that later). Even so, we’re still talking huge doses of ascorbate just to get its blood and tissue concentrations into a theoretical potentially therapeutic range.

After in essence recapping results presented in Figure 1 of their previous study for this study again and adding more cell lines to the panel to show that high concentrations of ascorbate are selectively toxic to tumor cells compared to normal cells, Levine gets to the meat of his results by testing his vitamin C regimen in three different mouse models of cancer in athymic nude mice: Ovcar5, a human ovarian cancer cell line; Pan02, a mouse pancreatic cancer cell line; and 9L, a rat glioblastoma cell line. Why he chose these particular cell lines is not really explained clearly. Personally, if I were doing this experiment I would have used nothing but human tumor cell lines. Be that as it may, tumors were allowed to grow to a volume of 50 ± 10 mm3, at which point treatment with either intraperitoneal injections of ascorbate or saline placebo were commenced. The results are below:


As you can see, the growth of the tumors is delayed, but there doesn’t appear to be any tumor shrinkage. This is actually not an uncommon result for a potential cancer chemotherapeutic drug (and, make no mistake, in this study ascorbate is being used as a chemotherapeutic agent); so I wouldn’t hold it against this study. Against these three cell lines, at least, ascorbate at these doses does appear to exhibit some in vivo antitumor activity. It’s modest and it requires massive doses producing high concentrations of ascorbate (millimolar range) in the blood to achieve, but it’s clearly there. At the end of the study, treated tumors were anywhere from 40-53% smaller than the untreated controls, and in the 9L cell line there was reportedly a major decrease in the frequency of detectable metastases.

The rest of the study involved mainly measuring the levels of H2O2 in the tissues of treated animals. Amazingly, peak levels of as high as 30 mM were obtained, which led the authors to reference a recently published phase I clinical trial of high dose intravenous vitamin C in advanced malignancy that shows that it is possible to approach such high concentrations of ascorbate in humans with doses of 1.5 g/kg. That’s still a lot but more achievable than 4 g/kg. Even so, it’s still problematic, as the infusion of high volumes of high osmolarity solution did cause problems in this trial. Nonetheless, if appropriate screening for conditions that might predispose to adverse events, even at the highest dose range, the results of the phase I clinical trial suggest that high dose intravenous vitamin C is fairly well tolerated.

Based on the animal studies and the safety data from the phase I trial, Levine concludes that high dose intravenous ascorbate represents a promising new therapy that should be further investigated. If his were the only evidence I saw being presented, I’d agree that it’s mildly to moderately promising as an anticancer therapy. However, I know the history of the use of vitamin C in cancer. More importantly, I know that the phase I clinical trial above did not identify a single patient with an objective response to therapy. True, the lack of even one objective response in a phase I trial does not necessarily mean that the drug doesn’t work, but it does dampen considerably optimism that it will work as a single agent. Taking the animal studies and the phase I trial together, I view the results with mild optimism that high dose vitamin C might have efficacy in human cancer. However, putting these results in context with what has been reported before over 30 years decreases even that mild enthusiasm, especially when coupled with my observation of just how modest the responses were in the mice and by just how much ascorbate was required to result in even a mild antitumor effect. Indeed, the authors of the phase I trial write:

In summary, this study shows that 1.5 g/kg ascorbic acid infused >90-120 min three times weekly is essentially free of risk and important side-effects when simple precautions are taken. In people with normal renal function, this dose achieves a plasma ascorbic acid concentration >10 mmol/l for several hours. No patient experienced an objective anticancer response, although two patients at the 0.6-g/kg dose received greater than six cycles of ascorbic acid with stable disease. Even though only six patients received the recommended phase II dose, our results suggest that the likelihood of an objective anticancer response to i.v. ascorbic acid alone is slight in unselected patients with multiply treated advanced cancer.

All of this brings up the question that just can’t be avoided: Does this study vindicate Linus Pauling and all the alternative practitioners of high dose vitamin C therapy for malignancy, such as the ones who treated Katie Wernecke for lymphoma? It depends on what you mean by “vindicate,” but my answer is: Not really, except perhaps in the weakest of ways, and here’s why. Remember, Linus Pauling didn’t claim that vitamin C in combination with other chemotherapeutic drugs might provide an additional incremental benefit to patients. That’s far too modest a claim. He said that “75% of all cancer can be prevented and cured by vitamin C alone.” Similarly, boosters of vitamin C as a cancer therapy don’t claim that it provides a small additional benefit when administered with chemotherapy, either. They claim that high dose vitamin C cures cancer at a high rate. Even the most optimistic appraisal of Levine’s results of Levine’s mouses study or the phase I clinical trial do not support any such claims.

The bottom line from my perspective: Yes, vitamin C probably has some antitumor activity for some tumors, but the word “underwhelming” comes to mind. Naturally, the authors of the phase I trial suggest using vitamin C in combination with chemotherapy, which is a perfectly reasonable approach for drugs that don’t show any objective evidence of activity as a single agent. However, this will not be easy, given that the sheer mass and volume of ascorbate that must be administered could easily interfere with other chemotherapy. One thing is fairly clear from the data known thus far: If high dose intravenous ascorbate has antitumor activity in humans, it is probably modest at best, and it definitely requires very high doses to achieve. High dose ascorbate might ultimately find its way into the armamentarium of science-based oncology, but it’s not likely to become a mainstay of treatment for any malignancy. It’s just too wimpy.


1. Chen, Q., Espey, M.G., Sun, A.Y., Pooput, C., Kirk, K.L., Krishna, M.C., Khosh, D.B., Drisko, J., Levine, M. (2008). From the Cover: Pharmacologic doses of ascorbate act as a prooxidant and decrease growth of aggressive tumor xenografts in mice. Proceedings of the National Academy of Sciences, 105(32), 11105-11109. DOI: 10.1073/pnas.0804226105

2. Hoffer, L.J., Levine, M., Assouline, S., Melnychuk, D., Padayatty, S.J., Rosadiuk, K., Rousseau, C., Robitaille, L., Miller, W.H. (2008). Phase I clinical trial of i.v. ascorbic acid in advanced malignancy. Annals of Oncology DOI: 10.1093/annonc/mdn377