A recent study correlates use of alternative medicine modalities in general and herbal remedies specifically with a lower chance of successful in vitro fertilization (IVF). The study, while not definitive, raises several important points.

The study was carried out by Jacky Boivin, a psychologist at Cardiff University, and found that women undergoing IVF in conjunction with a CAM modality (about half of which used herbal remedies) had on average 2.4 cycles of IVF treatment and a success rate of 45.2%. Women using only standard IVF techniques underwent 1.9 cycles of treatment with a success rate of 66.4 percent. These differences were statistically significant, and in absolute terms are very clinically significant.

The weakness of the study is that it was not a randomized prospective study. This means that the results are a correlation only and cannot be used to conclude causation. For example it is possible that women under more stress or who had previous failures might have turned in higher numbers to CAM therapies out of desperation. The study does try to address the other confounding factors, such as socioeconomic status, but these sorts of fixes only mitigate the possible effect of confounding factors.

As an aside, the ultimate way to control for possible confounding factors is through randomization, which can only be done in controlled prospective trials. What this means is that subjects are randomly assigned to receive either the study treatment or a control treatment. With large enough numbers and proper randomization, the effects of any confounding factors should average out between the study and control groups.

When it is not feasible to randomize (for example with studies on the correlation of smoking and cancer) you can never completely eliminate possible unknown factors. However, if you do multiple studies, looking at the correlation from many different angles and controlling for any plausible factors that are thought of, you can still achieve a very high probability that a correlation is real and reflects a certain causation. We can reliably say that smoking increases the risk of certain kinds of lung cancer, even though a randomized trial was never done.

What this latest study tells us is that we should not assume that CAM modalities are harmless. Often the use of CAM treatments is justified by the sentiment that it can’t hurt, so why not give it a try on the chance (however small) that it may be of benefit. Well, anything that can theoretically help can also hurt, and therefore anything that can’t hurt can’t help either.

Harm from CAM comes from many sources – there is financial harm from wasted expense, psychological harm from false hope, and potential medical harm from diverting time and attention from more plausible and proven therapies. There is also societal harm from diverting both scarce health care and medical research resources from the probable to the highly improbable.

And there is also direct medical harm from many therapies and interventions. Not all CAM therapies are biologically inactive, like true homeopathic remedies or manipulating non-existent energy fields. Acupuncture involves sticking real needles through the skin which carries a finite risk of infection. Chiropractic neck manipulation can cause damage to arteries in the neck leading to possible stroke.  Chelation therapy can cause dangerous electrolyte imbalance or kidney damage.

An of course, as I have pointed out numerous times before, herbs are drugs and carry all the potential risks of drugs that are actually called drugs. It is very plausible that some herbal remedies given to enhance fertility – apart from being useless – may interfere with researched drugs used in the IVF process. The herbs may also directly affect hormonal levels in a way that decreased the chance of IVF success.

This study is just the latest in a series of findings that commonly used over the counter herbal “remedies” can interfere through drug-drug interactions with prescription drugs. St. John’s Wart is known to interfere with anti-HIV drugs, for example.

This latest study is interesting and concerning, if not definitive, and should spawn follow up research to both replicate the findings and narrow down the factors that are likely to have a negative effect on IVF success. Such preliminary data is properly used to highlight a potential problem and serve and the starting point of more definitive research.

But we can also use this data to inform a risk/benefit assessment – the kind of assessment that is at the core of clinically decision-making. Given that the CAM treatments in question have no proven benefit and are highly implausible, in my opinion that are starting with an unfavorable risk vs benefit ratio. The addition of even preliminary evidence of possible harm further shifts this ratio toward the negative. In mainstream medicine it is not considered proper or ethical to give treatments that may have harm and probably have no benefit.

My biggest complaint about the bizzaro world of CAM is that this basic principle of scientific ethical medicine has been turned upside down. Proponents of CAM feel they do not have to provide any scientific evidence of benefit (or even plausibility), and shift the burden to critics to prove harm or risk (and try to set the bar as high as possible).