By Seed prodded, or there’s less to these studies than meets the eye

OK, I’ve been prodded enough!

Yes, I’ve been aware of the study purporting to present good anecdotal case reports showing that there might be something to the hypothesis that megadoses of vitamin C can cure cancer where other therapies fail. I’ve also been aware of an in vitro study that suggested selective toxicity of vitamin C to tumor cells compared to normal cells. I’ve even been meaning to write about since I first saw it a couple of weeks ago, but the AACR intervened, as did a number of other topics, and, like so many other topics that I want to write about but somehow never find the time, I put this one off again and again–until I had it thrown in my face not just once, not just twice, but three times by fellow ScienceBloggers yesterday.

Dang it all, don’t they know that Monday night is my night to watch 24 (this week more than ever since it was revealed last week that it wasn’t the Vice President behind it all but President Logan himself, a plot twist so awe-inspiringingly unbelievable, given how wimpy Logan has been thus far, that it could only be on 24)? Or that I also have a talk on Wednesday that I have to get ready for? I don’t want to interrupt the bliss of watching Jack Bauer kick terrorist butt. Worse, my tendency to procrastinate until the last possible minute when getting talks together can only be worsened by a blog prodding demanding a response.

Oh, well. The copious commercials left plenty of time for me to comment, and I can always pull a late-nighter tomorrow if I have to. Besides, at least one of the studies is interesting enough that it might make me be a little more open to the possibility that megadoses of vitamin C might have therapeutic value–but only a little until I see more and better data.

In the first article, Lee presents Linus Pauling as an unfairly vilified genius who may now be vindicated, starting out:

In essence, he [Pauling] challenged the established notions of nutrition, particularly regarding vitamin C. Pauling was influenced by the work of biochemist Irwin Stone, who pointed out that the majority of plants and animals produce their own vitamin C in proportion to their bodyweight, with humans, apes and a handful of other (primarily fruit-eating) species being the rare, incapable exceptions due to a purported genetic defect. Pauling reasoned that since vitamin C is present in most animals in concentrations vastly greater than the amount most people consume each day, some common illnesses might be the result of this perpetual vitamin C deficiency.

Consequently, he proposed that “megadoses” of vitamin C could effectively treat several illnesses, most notably cancer and the common cold, and published a few books to popularize these ideas. In 1973, he formed the Linus Pauling Institute of Medicine, where he performed multiple experiments to verify his claims.

The real trouble started when other researchers tried–and failed–to replicate his results.

No, the real trouble started with the poor experimental design of many of Pauling’s own studies (and his shameless overselling of their results). They were not well-designed prospective studies, but rather retrospective studies in which the control groups were not properly matched, as described by Stephen Barrett:

In 1982, William D. DeWys, M.D., chief of the clinical investigations branch of the National Cancer Institute’s cancer therapy program, pointed out that the vitamin C and control groups had not been properly matched. First he observed that no data had been published to demonstrate that the patients had been matched by stage of their disease, functional ability, weight loss, and sites of metastasis, all of which are important judging the stage of the disease. Then he pointed out that Cameron’s patients began getting vitamin C when Cameron judged them “untreatable” and their subsequent survival was compared to that of the control patients from the time they had been labeled “untreatable.”

DeWys reasoned that if the two groups were comparable, the average time from the initial diagnosis to “untreatable” status should be similar for both groups. But they were not. He concluded that many of Cameron’s patients had been labeled untreatable earlier in the course of their disease and would therefore be expected to live longer. DeWys also noted that more than 20% of the patients in the control group had died within a few days of being labeled untreatable, whereas none of Cameron’s patients had died. This, too, suggested that Cameron’s patients had had less advanced disease when they were labeled untreatable.

As experimental clinical protocols go, this study was a mess. Linus Pauling was not a clinician and had no experience in clinical trial design, and it really showed. There’s no standardization, no good matching of controls by age, stage, and performance status. The results of this study are essentially meaningless. If you want a quote from his original paper that shows this better than anything, here it is: “We believe that the ascorbate-treated patients represent a random selection of all the terminal patients in the hospital, even though no formal randomization process was used.” As clinical trials go, even retrospective ones, it’s pretty bad.

I have to wonder how this got published, but then I saw that it was originally published in the Proceedings of the National Academy of Sciences, a journal that I’ve never seen a clinical trial published in before, as it’s not a clinical journal. Not surprsingly, given his Nobel Prizes, Linus Pauling was a member of the National Academy of Sciences. What is not really known much outside the scientific community is that members of the NAS can contribute papers to this journal as they see fit and in essence pick their reviewers. Also, one way that non-members can have papers published in PNAS is if a member of the Academy agrees to submit their manuscript for them, and, in fact, members are supposed to take the responsibility for having such papers reviewed before “communicating them” to PNAS. Thus, in essence a member of the Academy can get nearly anything he or she wishes published in PNAS, whether written by him or herself or a friend. Normally, that’s not such a big problem for quality, because getting into the Academy is so incredibly difficult and only the most prestigious scientists are invited to join. Consequently, PNAS is still a highly prestigious journal with a high impact factor, and most of its papers are of high quality. Scientists know, however, that sometimes Academy members will use it as a journal of last resort to publish some of their leftover findings. They also know that on occasion, when rare members fall for dubious science, as Pauling did, they can “communicate” their questionable findings and get them published in PNAS unless they’re so outrageously ridiculous that even the deferential editorial board can’t stomach publishing them. All they have to do is to find a couple of sympathetic colleagues to review their manuscripts and then submit them. What keeps the overall quality of most of the journal’s articles high is primarily the desire of members of the Academy not to sully their names by communicating papers that they consider to be poor quality science.

I doubt that the concept that vitamin C might cure cancer and other diseases would ever have been taken as seriously as it was for as long as it has been if it hadn’t been championed by a scientific figure as towering and impressive as Linus Pauling, two-time Nobel Laureate. In his zeal, Pauling popularized his ideas not primarily by publishing in scientific journals, but mostly by writing books, giving talks, and forming his own insitute to do experiments designed to prove his ideas. There’s nothing inherently wrong with forming one’s own institute to study a question one is passionately interested in, but much of his work didn’t pass scientific muster through peer review, and the results of his clinical trial could not be replicated. Moreover, it was more than just cancer. According to Linus Pauling, megadoses of vitamin C could cure the common cold and a variety of other illnesses, contentions that did not stand up to scrutiny, and he also believed that vitamin C could cure vascular disease. (Whenever one treatment is represented as a cure for multiple unrelated diseases, my skeptical antennae start twitching furiously.) Also, his institute published a number of publications that made dubious claims, including a claim that “no progress” had been made in cancer in 20 years, and Pauling himself is known to have strongly defended certain quacks. (Ignore the irrelevant rants against Pauling’s left-wing politics in the cited article.)

Another sure sign of a zealot, Pauling couldn’t tolerate data that contradicted his belief in vitamin C. Indeed, when data from the experiments of a colleague at his institute, Arthur Robinson, suggested that vitamin C at the doses advocated by Pauling might actually increase the rate of tumor growth in an experimental model in mice:

Robinson, a former student and long-time associate of Pauling, helped found the institute and became its first president. According to an investigative report by James Lowell, Ph.D., in Nutrition Forum newsletter, Robinson’s own research led him to conclude in 1978 that the high doses (5-10 grams per day) of vitamin C being recommended by Pauling might actually promote some types of cancer in mice [18]. Robinson told Lowell, for example, that animals fed quantities equivalent to Pauling’s recommendations contracted skin cancer almost twice as frequently as the control group and that only doses of vitamin C that were nearly lethal had any protective effect. Shortly after reporting this to Pauling, Robinson was asked to resign from the institute, his experimental animals were killed, his scientific data were impounded, and some of the previous research results were destroyed. Pauling also declared publicly that Robinson’s research was “amateurish” and inadequate. Robinson responded by suing the Institute and its trustees. In 1983, the suit was settled out of court for $575,000. In an interview quoted in Nature, Pauling said that the settlement “represented no more than compensation for loss of office and the cost of Robinson’s legal fees.” However, the court-approved agreement states that $425,000 of the settlement was for slander and libel.

In 1994, Robinson and two colleagues summarized the results of four mouse studies he had carried out while working at the Pauling Institute [20]. Nearly all of the mice developed skin cancers (squamous cell carcinomas) following exposure to ultraviolet radiation. Altogether, 1,846 hairless mice received a total of 38 different diets. The researchers found that (a) the rate of onset and severity of tumors could be varied as much as 20-fold by just modifying dietary balance; (b) diets with the worst balance of nutrients had the greatest inhibitory effect on cancer growth; and (c) no cures or remissions were observed (although the researchers were not looking for this). In 1999, Robinson commented:

The results of these experiments caused an argument between Linus and me, which ended our 16-year period of work together. He was not willing to accept the experimentally proved fact that vitamin C in ordinary doses accelerated the growth rate of squamous cell carcinoma in these mice.

At the time, Linus was promoting his claim that “75% of all cancer can be prevented and cured by vitamin C alone.” This claim proved to be without experimental foundation and not true. . . . Vitamin C increased the rate of growth of cancer at human equivalents of 1 to 5 grams per day, but suppressed the cancer growth rate at doses on the order of 100 grams per day (near the lethal dose), as do other measures of malnutrition [21].

Of course, this could just be sour grapes after a falling out among old friends, or it could be only one side of a more nuanced story. It could also be that the acceleration of growth reported by Robinson may not be the behavior of all tumors or even most tumors in response to high dose vitamin C. However, if Robinson’s results as he reported later were correct, in the case of at least one tumor vitamin C appeared not to inhibit growth but rather to accelerate it. Pauling’s apparent willingness to suppress data that did not fit with his vitamin C hypothesis and to fire the man whose experiments yielded such data must be acknowledged, particularly given the libel settlement that resulted from legal action by Robinson. Lee’s representation of Pauling as scientist unfairly vilified for his ideas that were outside the mainstream may have an element of truth to it, but just an element. The whole story is not so flattering to Pauling, sadly. Pauling deserved more than a little of the vilification heaped upon him for his vitamin C advocacy.

But what about the recent PNAS paper and the Canadian Medical Association Journal study from a couple of weeks ago? Do they rehabilitate Pauling and suggest that maybe he was right about vitamin C after all, as Lee seems to be implying? The PNAS study is simply an in vitro study using five tumor cell lines that shows selective toxicity to tumor cells as compared to normal cells at concentrations that are achievable with intravenous doses of vitamin C, with vitamin C being around four- to twenty-fold more toxic to the tumor cells tested. This cell killing depended upon hydrogen peroxide generation, as it was abrogated when cells were pretreated with peroxide scavengers. It’s a pretty straightforward paper. My only quibbles with it are that they only used monocytes, lymphocytes, fibroblasts, and breast epithelial cells as the normal cells and that they only exposed the cells to the ascorbate for one hour, even though the paper itself states that it can take several hours for a large intravenous dose of vitamin C to be eliminated from the body. The paper represents interesting in vitro observations, but, as they say, you can kill any cells in vitro. In vitro experiments can generate hypotheses, but they are inherently highly artificial systems. As clinicians, we want to know if this works in vivo. Many are the chemotherapeutic agents that appeared to kill cancer cells selectively in vitro but failed in animal experiments. Another problem is that the concentrations of ascorbate used are very high, in the 2 to 20 mM range (hence the need for very high intravenous doses of vitamin C to achieve such levels in the blood).

That’s why the CMAJ paper seems on first glance to be of more interest. Based on the concept that only intravenous dosing can produce high enough blood levels of ascorbate and that that is the reason previous trials failed to show a benefit from high dose vitamin C, three cancer cases are reported, all of whom received high dose vitamin C and all of whom showed much longer than expected survival. The problem is, there’s a lot less there than has been reported.

The first patient, for instance, had renal cell carcinoma with putative lung metastases and received 65 g vitamin C intravenously twice a week for 10 months. The lung metastases were never biopsy-proven to be renal cell carcinoma, leaving some doubt about them. However, let’s accept that they probably were legitimate lung metastases, even though none of them were ever biopsied. (We don’t always biopsy metastases to prove that they are metastases, particularly if the biopsy would be unduly risky, although in general ideally we do like to biopsy suspected metastases before starting therapy.) The problem is that renal cell carcinoma is a tumor type that is one of the more common tumor types to be reported to undergo unexplained spontaneous remissions. It’s pretty uncommon, but it has definitely been documented to happen. Also, the patient received a number of other “alternative therapies,” including thymus protein extract, N-acetylcysteine, niacinamide, and whole thyroid extract. If the authors believe that alternative therapies like high dose vitamin C can cure certain cancers, why did they automatically assume that it had to be the vitamin C and not the other alternative therapies used or some com? Why couldn’t it have just as plausibly been one of the other treatments or some combination of them? The fate of this patient was ultimately not good: “Unfortunately, the woman — a long-time smoker — was diagnosed four years later with small-cell lung cancer. She again opted for vitamin C treatment, but the second cancer did not respond.” Was there vitamin C resistance in the second tumor as opposed to the first, or did the first tumor just go into spontaneous remission, or was it something else? It’s impossible to say much of anything from this case.

The second case presented was of a man who had a primary bladder tumor with multiple satellite tumors. These were all resected with apparently clear margins transurethrally using a cystoscope. Rather than additional conventional therapy, the man chose intravenous vitamin C and was reported alive and well nine years later. Normally, this would be a pretty poor prognosis tumor without more therapy, but, as the accompanying commentary points out, long term survival has been reported with surgical therapy alone. Indeed, this case reminds me a lot of breast cancer patients who undergo excision of their primary tumor, forgo adjuvant therapy for alternative therapy, and then attribute their survival not to the conventional therapy (surgery) but rather to the alternative therapy. In addition, this patient also took a variety of other alternative therapies, leading to the question again: Why did the authors assume that it was the vitamin C? They may be correct, but there are too many confounding factors. It could just as plausibly been the surgery.

Finally, the third patient had a large paraspinous mass that turned out to be diffuse B-cell lymphoma. She agreed to radiation therapy but refused chemotherapy, opting instead for intravenous vitamin C. She, too, used a variety of other alternative therapies, including beta-carotene, bioflavonoids, chondroitin sulfate, coenzyme Q10, dehydroepiandrosterone, a multiple vitamin supplement, N-acetylcysteine, a botanical supplement, and bismuth tablets. It’s utterly ridiculous to include this case as “evidence” for teh efficacy of vitamin C. Besides the confounding factor of using multiple alternative therapies, the patient reported only had, as far as can be told from the case report, a Stage I diffuse B cell lymphoma. Radiation therapy alone used to be a common treatment for such tumors and is well known to be able to produce long term survival in early stage lymphoma. To try to show that this was something other than Stage I disease, the authors make much of the observation that the patient subsequently developed enlarged lymph nodes in numerous other nodal basins over ten years. However, nowhere do they report that any of these enlarged lymph nodes were were ever biopsied to prove that they were recurrent lymphoma, and the only pathology result reported is from the biopsy of the paraspinous mass. One thing oncologists that I’ve worked with damned near always insist upon before diagnosing a recurrence of lymphoma is proof by tissue diagnosis, and it is very puzzling why such a simple procedure was not done sometime over the ten year history reported. It would have been very easy to do biopsies of any of these superficial lymph nodes. In fact, looking at the authors’ descriptions, only one of these nodes would I have even considered particularly suspicious (a 3 cm supraclavicular node). The majority of the nodes reported were merely described as merely “palpable” or as measuring 1.0 or 1.5 cm.

Do you now see why I say that there’s a lot less to this paper than meets the eye?It’s possible that high dose vitamin C might have had something to do with the long term survival of these patients, but these case reports certainly do not provide particularly compelling evidence to suspect that it did. Worse, despite claims to the contrary, high dose vitamin C is not completely benign. High IV doses can precipitate renal failure, and acute hemolytic anemia in patients with glucose-6-phosphate dehydrogenase deficiency. It would therefore be wrong to subject people to the risks of this therapy without some better indication that it might actually have significant antitumor effect. As the commentary points out, even if these three cases do represent legitimate tumor responses to vitamin C, we have no idea what the denominator is. These cases were collected over many years, and it’s impossible to know how many thousands of patients underwent similar vitamin C regimens with no response at all. In fact, I find it hard not to point out that these are the best cases the authors could come up with over many years, and yet they are only marginally more convincing than the anecdotes that alties like to spread in that these cases at least had definitive pathologic tissue diagnoses reported. Finally, the hype around vitamin C without convincing evidence for its efficacy is leading to real harm, such as parents of a girl with a potentially curable cancer choosing intravenous vitamin C rather than chemotherapy.

Now that I’ve shown why these studies aren’t what they’re being hyped to be, I will say that the in vitro study is sufficiently interesting that it is probably worth doing some animal experiments with various mouse tumor models to see if there is any hint of efficacy, to see vitamin C at doses high enough to produce blood levels as high as the in vitro study show any antitumor effect. It may even be worth doing a Phase I trial to determine the maximum tolerated dose in cancer patients and whether there is any objective evidence of a response, but I wouldn’t advance to that stage until after at least some animal work.

So, what’s the bottom line? Is it Lee’s view, as stated below:

The frightening implication of all this is, if it somehow happened to one of the greatest figures of 20th century science, it can certainly happen to anyone who speaks out about the credibility of unorthodox ideas. For instance, without the benefit of these recent studies, a science journalist writing about Pauling’s vitamin C theories without calling them baloney and flapdoodle would likely have been a target for professional derision and ridicule – and perhaps still will be.

But perhaps instead the ultimate vindication of a great mind’s last discovery is at hand. And hopefully, Pauling’s tragic downfall and unfolding redemption will provide a valuable lesson in science’s fallibility for those who truly seek objectivity.

Possibly, but most likely not. Pauling’s work on vitamin C was too riddled with poor methodology, poor experimental design, and blatant bias. Even if this more recent work pans out, it won’t completely rehabilitate his image, particularly given that numerous quacks have appropriated his name and his concept of “orthomolecular medicine” to sell all sorts of dubious “cures,” and the Journal of Orthomolecular Medicine still publishes all sorts of questionable science, such as treating HIV infection with vitamin C. Nor is this particularly likely to be an example of “today’s “debunked” ideas might ending up being tomorrow’s ‘new’ exciting discoveries.”

Tara, characteristically, gets it much closer to correct:

And this brings me back to Pauling. Indeed, his work in this field is often synonymous with quackery, or a tale of caution when it comes to knowing where your expertise lies–and where it doesn’t. Does that mean he could still be right–and megadoses of vitamin C could be beneficial? Sure. It’s also unfortunate that snake oil salesman such as Matthias Rath, who was affiliated with the Linus Pauling Institute, continue to trade on Dr. Pauling’s name. So while all this is unfortunate, what I see here is a bit more like the mythology of Barry Marshall’s ostracism–complete now with the potential happy ending of Pauling’s “vindication.” Again–call me closed-minded, but I’ll wait for the evidence.

Don’t worry, Tara, you’re not at all close-minded, just properly skeptical. It is those who have so credulously trumpeted these two studies as proof that Pauling was right who are so open-minded that their brains fall out. It’ll take a whole lot more evidence than these two papers even to begin to vindicate even partially Pauling’s concept. Based on the in vitro study more than the case reports, I’m provisionally willing to reconsider the concept of using vitamin C to treat cancer one last time, after having pretty much dismissed it totally in the past, based on the three Mayo Clinic studies and other evidence. For now, I’m willing to make two predictions: First, the likelihood that any of this will pan out to result in a treatment for cancer more effective than the present chemotherapeutics that we have is probably pretty low. Second, if my first prediction turns out to be incorrect and Phase I, Phase II, and Phase III studies ultimately show vitamin C to be highly effective against specific cancers, we skeptical “conventional doctors” reviled by alties (including me) will embrace vitamin C, coopt it, incorporate it into our anticancer armamentariums, and study it futher to understand better its mechanisms, limitations, side effect profile, and risk-benefit ratio. Indeed, personally, I’d even raise a glass to the old guy, congratulating him for getting the basic concept right even if he botched the execution…

That’s how medical science works when it works as it should.