Gingko biloba is a plant that has long been used as a Chinese herbal remedy in a standard extract. Since the deregulation of the supplement industry in the US in 1994 gingko has become one of the more popular herbal supplements sold in the US (annual sales for 1997 were estimated at $240 million, and have only increased since). The primary claim for gingko is that it improves memory and cognitive function, and may even stave off the onset of Alzheimer’s dementia. Other claims include that it is useful for headache, depression, and tinnitus (ringing in the ears).

The problem with the current regulation is that it essentially allows companies to make claims without having to provide evidence to back up those claims; therefore, the marketing is ahead of the research. This has been true for gingko as well – it has been marketed as a memory aid for years and only later has the research started coming in.

The latest issue of Neurology online contains a double-blind placebo-controlled study of gingko biloba extract in a healthy elderly population (118 subjects) that was observed for 42 months. The study is essentially negative – not supporting the use of gingko in preventing memory decline in the elderly – but there are some complexities to the interpretation of the study.

The primary outcome was an intention to treat analysis – that means the net effect was measured without taking into account whether or not the subject actually took the treatment or placebo. This type of analysis is useful if the question is whether or not it is useful to prescribe a treatment. If a treatment works, but no one can take it because it is too complex, inconvenient, or has too many side effects, then it is not a very practical treatment. But if your question is whether or not a specific treatment works (regardless of whether it is practical) then you have to account for compliance.

In this study the primary outcome was intention to treat and it was completely negative – no benefit from gingko in preventing progression to dementia. The authors did a secondary analysis, however, where they accounted for compliance and with this analysis two effects emerged: a mild protective effect from gingko, and an increase in strokes and transient ischemic attacks (TIAs). How should we interpret this?

First – primary outcomes in a study always have more weight, for that is how the study was designed and the question that was asked prior to the data being collected. Secondary analysis is always suspect – it usually involves additional statistical complexity, may involve an analysis of only a subset of the subjects, and also may involve a re-analysis of data after it is collected, which opens the door for bias and further statistical anomalies.

Also, this study was designed as a pilot study, which means it was low power (118 subjects is a very small number for such a study). This means that the results are not very reliable but were intended to help design a larger more definitive trial. For example, the authors used the data from this trial to calculate that a follow-up trial would need at least 300 subjects to have an 80% chance of detecting an effect (depending on certain variables it may also require as many as 2800 subjects). This study also indicates that any follow-up study should closely track strokes and TIAs as a possible adverse side effect.

My opinion is that the bottom line from this study is that it does not tell us anything definitive, but that it argues against any clinically significant effect from gingko, but of course does not rule out a small effect. The increased risk of stroke and TIA is likely not real, but it could be. Gingko is known to be a mild blood thinner and should not be taken with aspirin or warfarin (Coumadin), which are also blood thinners. There is concern that gingko increases the risk of bleeding, but usually (if anything) a blood thinning effect is associated with a decreased risk of non-bleeding strokes. So either this side effect was spurious or the effects of gingko on blood clotting and platelet function are more complex than we currently know.

To put this one small study into the broader context of the clinical literature, a recent Cochrane review of clinical trials of gingko and memory concluded:

The evidence that Ginkgo has predictable and clinically significant benefit for people with dementia or cognitive impairment is inconsistent and unconvincing.

That basically means that the evidence suggests gingko does not work for memory, despite the hundreds of millions of dollars of sales based upon that claim. This new study does not alter this conclusion.

The other context that is critical to thinking about gingko and its clinical utility is that all of this discussion of clinical trials, possible medical benefits, and concerns over side effects is identical to the type of discussion we would have about any drug. There is a good reason for that – Gingko biloba extract is a drug or combination of drugs and it should be treated as such. The medical and research community treats it exactly like a drug – with pharmacological effects that can be studied using standard techniques, active ingredients, drug-drug interactions, and potential clinical effects and side effects. Why then is it regulated as if it were a dietary supplement?

Our experience with gingko also shows that centuries of traditional use is not a good predictor of clinical efficacy. Such a history therefore does not obviate the need to conduct proper clinical trials. Gingko may, in fact, be causing more harm than good (or harm without any good).

The only rational conclusion is that drugs should be regulated as drugs.