The risks of CAM: How much do we know?

Working in pharmacies where supplements are sold alongside traditional (over-the-counter) medications, I’m regularly astonished at the different perceptions consumers can have about the relative efficacy and safety of different types of products. Once, speaking with a customer about a medical condition she wanted to treat, I indicated that there were no effective non-prescription therapies — she needed to see a physician for access to an effective treatment by prescription — and I gestured behind the counter. “Back there?!” she pointed. “That’s where you keep the stuff that kills people! I want something natural!” Suggesting that my patients with heart disease or HIV had a somewhat different perspective, I tried (unsuccessfully) to talk her out of a questionable-looking supplement (Hint: avoid anything from a company with a P.O. box as a mailing address.) This appeal to nature, combined with a perception that natural products are safe, and conventional drugs are unsafe, is pervasive.

One of the responses regularly tossed at health professionals who argue for higher regulatory standards for supplements, CAM, and other natural health products is that conventional medicine kills, and prescription drugs are its preferred weapon. In contrast, they argue, supplements are safe. Any credible health professional will agree that conventional medicine can cause harms. David Gorski has examined Science-Based Medicine and patient safety before. The number of preventable adverse events from medical treatments is far too high. And even the idiosyncratic events — freak accidents, basically — mean we must always consider the rare but possible harms of the therapies we use. But as Harriet Hall, another SBM blogger, has pointed out, we cannot look at drug harms in isolation: the variable perpetually ignored by critics of mainstream medicine is the benefits associated with those drugs, which can make the risks seem quite acceptable. Both expected benefits and possible harms need to be incorporated into our decision-making process.

When it comes to drug (or supplement) safety monitoring, causality assessment is the evaluation of the odds that a particular treatment is associated with a particular adverse event. A single adverse event may have characteristics that suggest the relationship to a treatment is highly probable. But because correlation may well not equal causation, data are usually rolled up an analyzed in aggregate, by national and world-wide agencies to help identify the rare but serious adverse events. What this means is, if data are not collected, we have an incomplete picture of a product’s adverse event profile. Yet when it comes to supplements, good data on safety and efficacy are often lacking.

How Safe is the ‘Alternative’?

One of the recurrent themes of science-based medicine is that any medical intervention that can plausibly cause physiological benefit can also plausibly cause physiological harm.
— Steven Novella, Vitamin E and Stroke, Science-Based Medicine Blog

Despite the widespread perception that supplements are safe, some do have undesirable biological and pharmacological activity. There is no inherent safety to any product, natural or not. There is both anecdotal and more systematically-captured information to suggest harms can be associated with supplements. So how do we gather good data to inform decision making?

The primary challenge we face when evaluating the risk profile of supplements is the lack of good quality clinical trial evidence. When these data do exist, studies are typically of short duration with small sample sizes. Without clinical trials that systematically measure and report adverse events, there’s no harms profile, and the “Side Effects” section of the monograph stays nice and short. In contrast to supplements, new drugs undergo extensive clinical trials and adverse events are actively monitored, collated, and eventually are summarized in the product monograph. It should surprise no-one that the list of adverse events is longer, when you actually go out an systematically look for them.

As regulators worldwide generally don’t require any meaningful pre-sale testing of supplements for safety or efficacy, there’s typically little data available when any product becomes marketed. (Some regulators will approved products based on a history of “traditional use”, which offers little insight into a products safety profile, as that data have never been systematically collected.) Instead, regulators worldwide (e.g., Canada and the USA) rely almost entirely passive surveillance to monitor the safety of marketed products.

Passive surveillance depends on the spontaneous reporting of adverse events from health professionals, alternative medicine purveyors, manufacturers, and even consumers directly. Passive surveillance is the key method of collecting safety data on drugs and supplements once they have been licensed for sale.

However, passive surveillance is an imperfect means of collecting safety data. It is much more effective at identifying rare and infrequent adverse events (e.g., liver toxicity) than incremental increases in the risks of common medical conditions (e.g., heart attacks). While the true denominator cannot be know with certainty, it’s generally accepted that only a fraction of adverse events are identified, and only a fraction of these are actually reported. Despite these limitations, the Institute for Safe Medication Practices reports that in 2009, the FDA received 19,551 reports of patient deaths associated with drug therapy. Is this the smoking gun that Big Pharma is killing us? Not quite. Causation has not been established, and even then, we need to consider harms in the context of the benefits a therapy may offer. The system isn’t perfect, but it is producing data that may be helpful for decision making, and understanding the evolving risk profile of licensed products.

Despite their widespread use, surveillance of supplements reveals generally few adverse events. It could be that many products lack meaningful pharmacologic activity — either positive or negative. An adverse event is probably less likely in drugs without any meaningful beneficial effects. But beyond efficacy, there are several other factors, unique to supplement use, that may contribute to even less frequent adverse event reporting. These were identified in a series of qualitative semi-structured interviews conducted by Rishma Walji in 2010, and summarized in the paper Consumers of Natural Health Products: natural born pharmacovigilantes?:

  • Consumers believe supplements are safe, because they are “natural.” Therefore they’re less likely to associate adverse events with supplement use.
  • Supplements don’t require a prescription. No health professional may be involved in monitoring response to therapy, who might be able to identify adverse events as part of a monitoring plan.
  • Because consumers may not involve health professionals in their decisions to take a supplement, they may feel obliged to deal with it on their own, and not involve health professionals. Perceptions about how their health professional will react to their use of supplements may come into play.
  • Consumers may not know how to manage and adverse event, and who to report an adverse event to. They didn’t get it from a doctor, so they don’t report the outcome to a doctor.
  • Even if they do contemplate reporting an adverse event, the process can be complex, especially for consumers.
  • Users may have a distrust of “conventional” medicine which drove their initial use of the product. When adverse events were identified, there were reluctant to consult “conventional” health care providers for assistance.
  • Some surveyed stated a fear of losing access to supplements, driving a reluctance to report adverse events.

From my own practice, I’d add few additional factors to the list. Supplements are commonly used for short durations, and for self-limiting conditions, in comparison to prescription drugs that may be used chronically. Prescription drug manufacturers are required by law to report to regulators any actual or potential adverse events that they learn of. This requirement was recently implemented in the USA for supplement manufacturers — however, contact information may not yet appear on supplement labels.

Disappointingly, consumers surveyed reported a preference to discuss adverse with health food store personnel, family or friends, which appears to be a factor of both comfort and convenience. Enablers to reporting adverse events included knowledge and importance of adverse event reporting, and personal comfort with their health care provider.

Real World Harms

People often ask “what’s the harm” of supplements and other forms of alternative medicine. This question can be answered, if incompletely. Considering the numerous case reports of harm helps us understand patterns, look at causality, and guide the safe use of these products. A recent case series of adverse events supports some of the observations that Walji made in her paper, and expands on the possible consequence of CAM use. The paper, Adverse events associated with the use of complementary and alternative medicine in children, is from Lim et al., and appears in the Archives of Disease in Childhood.

Monthly surveillance reports of CAM-associated adverse events in Australian children were collated over a three year period. This active surveillance program (pediatricians are asked monthly to report any rare disorders that have been observed) followed adverse event reports with detailed questionnaires, evaluating causality, severity, and whether the adverse event was potentially related to a failure to use “conventional” therapy. The program identified 39 adverse events, with 64% considered severe, life threatening, or fatal. Thirty cases (77%) were considered to be related to CAM, and 17 cases (44%) were felt to be due to a failure to use conventional medicine. Four deaths were documented, all attributed to a failure to use conventional medicine in favour of CAM. One of the deaths, from seizure, was attributed to a decision to use CAM rather than drugs, with the aim of avoiding the side effects of drugs.

The authors make the following observations on CAM use and adverse events:

  • Reporting can be complicated because information on the product may not be available.
  • Products can be contaminated or adulterated with therapeutic, undisclosed drugs, complicating evaluations of causality.
  • Because CAM products are perceived as safe, they may not be stored properly in the home.
  • CAM use may be accompanied by dietary advice that is dangerous, and not based on good evidence. Two of the fatalities associated with CAM occurred in children on marked dietary restrictions prescribed by a CAM practitioner.

It’s a small case series, and we know neither the actual numerator or denominator. Still the severity of some of the reactions is concerning, and the author’s comments point to additional factors that may influence the incidence of adverse events reported.

Information is Beautiful … But Incomplete

As a pharmacist, I constantly need to contextualize data directly for patients, translating trial results into more understandable figures like numbers needed to treat and numbers needed to harm. I’m always searching for better ways of communicating the understanding of risks and benefit. Information is Beautiful is one of my favourite blogs, as the authors highlight innovative and effective ways of graphically presenting data. So I was initially excited about their attempt to summarize the data for supplements — click on the image to view the data on the IiB website:

Clicking each bubble will link you to the supporting clinical data. But there’s a crucial element missing in their presentation: harms. You cannot evaluate the value of a treatment based on efficacy alone. For example, selenium is ranked very highly for colorectal cancer prevention based on this study. Not only is there evidence to the contrary, selenium consumption is associated with a significant increase in the risk of type 2 diabetes. The risks actually seem more compelling and relevant than the potential benefits. Or look at red yeast rice, another supplement rated highly. While RYR does seem effective, is less effective and more dangerous than drugs derived from it. I like the IiB visual approach, but it fails in communicating meaningful data on risk and benefit, both of which are essential for decision-making.

Supporting Science-Based Decision Making

Whether one is considering a “conventional” drug or a supplement, the questions we should be asking are the same:

  • What benefits can I expect given my own condition? What data exist to support this? How much more effective is this treatment than a placebo?
  • What harms with this treatment have been documented? Has the product been studied extensively enough to understand the potential risk profile?
  • In how many patients has this product been studied? How closely do they match my situation?
  • What alternatives exist, and what are their risk-benefit profiles? What are the risks of no treatment at all?
  • What are the economic consequences of the treatment strategy I decide upon? How much “bang for the buck” should I expect?


Any active treatment can have harms, yet robust safety data can be hard to come by for supplements and natural health products, painting a skewed picture of their risks and benefits. Our passive system of collecting safety information isn’t likely to remedy this situation. While regulators may have created different standards for supplements and CAM, permitting sale in the absence of good safety and efficacy data, our decision-making should make no such exceptions. Evaluating what we know about efficacy and safety, and what we don’t know, and the limitations of both, are critical to science-based decision making.


Lim A, Cranswick N, & South M (2010). Adverse events associated with the use of complementary and alternative medicine in children. Archives of disease in childhood PMID: 21178176

Walji, R., Boon, H., Barnes, J., Austin, Z., Welsh, S., & Baker, G. (2010). Consumers of natural health products: natural-born pharmacovigilantes? BMC Complementary and Alternative Medicine, 10 (1) DOI: 10.1186/1472-6882-10-8

13 thoughts on “The risks of CAM: How much do we know?

  1. Pingback: The Risks of CAM | PEI Curmudgeon's Blog

  2. “The primary challenge we face when evaluating … (is)… the lack of good quality clinical trial evidence. When these data do exist, studies are typically of short duration with small sample sizes.”

    We can say this of any number of pharmaceutical drugs, and in fact, judges have said it for us: Vioxx, Ezetimibe, Vytorin, Prepulsid, the Statins Zyprexa. The list goes on. Today, Feb 19, dabigatran is being spammed to us in dozens of mainstream media articles, in spite of the fact the non-industry funded drugs research body Therapeutics Initiative says there are serious adverse effect concerns. Like myocardial infarction. Therapeutics Initiative further tell us they have asked for all the relevant data and pharma has refused.

    How many other drugs now withdrawn were pushed onto the consumer like this, heavily marketed and lightly researched, with the consumer not told of risks, and not aware of what information is being withheld?

    I’m not a supplement user (or drugs if I can help it and I can) but I do have to admit a yawn when pharmacists get worried about supplements. Bring your advocacy back where the money is Scott, right under your nose.

    You like links? I got ’em.

    • I look at the examples of drug withdrawals you have listed as only reinforcing my point: the need for both robust, large clinical trials as well as good post-marketing surveillance. Recall that all of those withdrawn products underwent Phase I, II and III trials that underwent significant scrutiny by regulators world wide. It’s the clinical trial data, and in some cases, Phase IV (real-world) evaluations, as well as mandatory pharmacovigilance, that gave a better understanding of actual risk-benefit perspective of these products. The criticism of these products that you’ve correctly identified is only possible because of a rigorous scientific process that compels manufacturers to complete clinical trials.

      In contrast, with CAM products, we have nothing even approaching the minimum requirement for prescription drugs. Not only does the absence of data mean that claims of efficacy are usually unfounded, but there’s also no persuasive evidence that the products are safe either. An absence of efficacy and safety data does not mean these products are safe or effective. For example, I’ve blogged before about a product that failed as a prescription drug and has simply been repackaged and rebranded as a “natural health product” and put on the market. Why? Because the regulatory bar is lower.

      If you take a look at some of the articles here you’ll see that I’m advocating for an equivalent safety standard for all products, instead of a double standard simply because a product is considered “natural” or a “supplement”.

  3. Vioxx pharma withheld data. Baycol pharma withheld data. Vytorin was stopped because forced to stop. Trials were rigged, end-points moved.

    If you ignore the bleeding obvious and go where it’s neat and easy, it says a lot about you. And pharmacy.

    • I’m confused, insurgent. Just how do mistakes by Big Pharma (and they do exist, as you’ve demonstrated, and no one disputes these issues) have anything to do with the effectiveness or potential harmfulness of CAM? As Scott said, these incidents are unfortunate and a rightful indictment of profit-driven industry running drug research and development. But it shows that with continued independent research, these issues are eventually sorted out. They should be sorted out *before* the drugs are widely sold.

      CAM is also a profit-driven industry with even less oversight than Big Pharma. This can lead to major problems in both quality and efficacy, but so far CAM practitioners seem happy to promote and sell their remedies with little to no supporting evidence and despite (in some cases) poor quality control of their remedies, leading to illness or other harm (e.g., Zicam). CAM remedies aren’t free and are sometimes promoted on TV, so where is your outrage when those don’t work or when there is media coverage of harm?

      The answer to “CAM has issues” isn’t “Well, Big Pharma also has issues”. What does that solve? Now we have two industries full of issues. The answer is “let’s fix both and/or demand that they be fixed by parties in a position to do so”. This can only be done with rigorous and objective investigation of efficacy.

    • I think these are great examples that further reinforce an argument for rigorous oversight. All of these cases were identified because of the appearance of data and information. And in some cases, they came out despite deliberate measures to block their release. It’s cases like these that build a stronger case for disclosure and transparency in the approval process. The commercialization of drugs is in the hands of for-profit companies (just like the manufacturers of CAM products). Companies can exercise bad judgment in their actions – and sometimes behave even worse, as your examples point out. But the response to bad behavior from pharmaceutical companies shouldn’t be to throw our hands up and turn to allowing CAM manufacturers to make safety and efficacy statements about products without requiring any good evidence of either.

  4. Insurgent — if I concede that the the FDA / Health Canada approvals process has fallen down in some cases, will you concede that that fact doesn’t make any CAM remedy effective or safe? Because right now, your response to Scott’s concerns about CAM essentially amounts to “Pharma’s worse.” We can debate whether that’s true or not, but either way, it doesn’t address the problems with CAM — it’s just a deflective tactic.

  5. There was no “independent” research, not at the beginning, and certainly, no “continued”. It borders on criminal. In fact, courts are deciding that it IS.

    As for the difference between a pharma drug induced disability or death and a CAM one, none for the end point, but a lot. Many consumers today are finding doctor dismissal letters in their mail boxes when they refuses the blockbuster shibboleths. I’ve got one here for statins, and I’m not alone. Standard of care, whether it is really care, or just higher stockholders share. Seniors in nursing homes are not being forced to take CAM, children in school who are anxious about exams are being put on Zyprexa at the school’s “either or”. Further more, we are not paying for CAM studies, because as has been noted, there really aren’t any. But we sure are forking out for pharma studies. And please don’t bleat costs of R&D, because I’ve got studies that blow that all to smithereens. Here in Canada, we pay coming and going, in our taxes, in the high costs of me-too drugs shilled to patients (it should be RD&M) and it’s not for “CAM”. Which by the way does not mean unproven. It means “complementary” and alternative. Ie) a doctor treating HCV will use milk thistle too. (And probably get very promising results esp if he drops the interferon).

  6. “*…) will you concede that that fact doesn’t make any CAM remedy effective or safe?”

    I can’t concede something I’ve not asserted nor defended.

    I think it’s a cop-out, and rather lame, to run after CAM when there’s much bigger game to be had.

  7. JAMA February 18, 1911
    JAMA. 2011;305(7):722. doi: 10.1001/jama.2011.38

    “Most physicians of modern times would be quite sure, if they thought about the matter at all, that the pharmaceutic problems which prove so troublesome in our time represent abuses that have crept into medicine because of peculiar developments in the last few generations. Certainly very few would think that these problems-proprietary remedies and their abuse, substitution, counterfeiting of drugs, the use of long, imposing Greek names to make new compounds popular and impressive-were all features of a previous phase of medical history. In this matter the story of medicine at Rome during the early empire is particularly interesting…”

  8. Insurgent, yes, some drugs are bad. It seems as though you are trying to jump to the non sequitur conclusion: “You should not criticize CAM, because there are major problems with industry.”

    Why must it be either or? Go after pharma, and go after CAM. These are *all* large, profit-driven companies who certainly have motives to withhold information.

    One of Scott’s points is that, in terms of CAM, we simply have no way of knowing what, if any harms are happening out there in the community, because there is no post-marketing surveillance system. Therefore, we need a consistent approach to any substances offered in the marketplace which claim to have pharmacological effects.

Comments are closed.