Putting Ginkgo to the Test


	Summer 2001

	From The Editor Jet Lag May Harm Memory Putting Ginkgo to the Test Depression and Memory Patient Profile Memory Tip

Putting Ginkgo to the Test
We don't know yet whether ginkgo biloba prevents or treats Alzheimer's disease. Why not?

Get the printable version.

View the large type version.

In the premiere issue of Memory Loss & the Brain, we took a critical look at research on the alleged memory-enhancing effects of ginkgo biloba. Extracts made from the leaves of the ginkgo biloba tree are also under investigation as a treatment for Alzheimer's disease -a debilitating illness marked by impairments in memory, thought, and emotion. But in dozens of published studies, including a major trial in the prestigious Journal of the American Medical Association (JAMA), the evidence doesn't support ginkgo as a treatment for Alzheimer's disease or for boosting memory power in otherwise healthy people.

Why the big black hole of uncertainty? It all comes down to the quality of the studies conducted so far. By understanding what makes a drug study tick, you can get some insight into what counts as convincing evidence in medical research-and why ginkgo biloba hasn't yet moved from the dietary supplement racks to the prescription pads of U.S. doctors.

What's a good study?

Scientists conduct small pilot studies, involving less than a dozen people, to investigate interesting research questions. If results show that the research is on the right track, the researchers may undertake larger "clinical trials" to evaluate whether a new drug or treatment is safe and effective. At a minimum, these clinical trials must have three features:

1. Steps to minimize bias
These features of the study help to minimize influences on the test results not directly related to the chemical actions of the drug itself:

Placebo control: In the best clinical trials, subjects are divided into at least two groups. One receives the test drug; the other receives an inactive placebo pill. It's done this way because people tend to report improvement in their symptoms following any kind of treatment-even an inactive compound, such as a sugar pill). The placebo group provides a way to distinguish actual drug effects from the "placebo effect." (In some studies, the group receiving the test drug is compared against one or more groups receiving currently-marketed drugs, instead of or in addition to, comparing against the placebo group.)

Randomization: This means that researchers assign people in the study to the drug treatment or placebo group at random, decreasing the chance that people will cluster in patterns that may influence the results. For example, assigning the most healthy individuals to the treatment group could make the drug look more effective than it really is.

Double blinding. This insures that neither the patients nor the researchers know who received the drug until the final results are tallied. Otherwise, the knowledge of who did and didn't get the drug could influence the scientists or the study participants.

2. Careful selection of subjects
In a well-done clinical trial, all the participants are screened carefully for the disease that the drug is supposed to treat. This sounds obvious, but a lot of health problems have similar symptoms. For example, both early Alzheimer's disease and depression can both cause memory impairments.

3. Measuring what matters
In any drug trial, researchers have to choose an appropriate way to measure the effect of the drug. In trials for Alzheimer's drugs, for instance, participants undergo standard tests to determine if the drug reduces memory impairment and other symptoms. The best clinical trials also check for "clinical significance," or whether the treatment made a meaningful difference in patients' lives. If patients taking a new Alzheimer's drug show a small improvement in a laboratory test of memory, this is all well and good. But does this improved test score translate into a decrease in forgetfulness or confusion in everyday life-a meaningful benefit? Typically, a doctor determines this during individual meetings with patients. Information from family members is also included in many studies.

Past imperfect

Not many studies of ginkgo have met these standards. A few years ago, several researchers in Oregon reviewed all the studies they could find of ginkgo biloba as a treatment for dementia, focusing on those studies involving people with Alzheimer's disease (the leading cause of dementia). Out of 50 studies, only four met the minimum standards for a clinical trial of a drug for Alzheimer's.

Together, the four studies included 212 subjects treated with ginkgo and 212 subjects given a placebo. The Oregon researchers concluded that daily treatment with 120 to 240 mg of ginkgo for three to six months brought about a small improvement in test scores, but it did not appear that the change was clinically significant.

There were other problems, too. Past studies of ginkgo and dementia (Alzheimer's and other types) were often conducted by a single scientist, or team of scientists, on patients recruited in one location. These "single-site" studies cannot be trusted as much as those conducted at different locations around the country, explains Lon Schneider, MD, a professor of psychiatry, neurology, and gerontology at the University of Southern California in Los Angeles. Single-site studies "are hugely subject to bias, because at one site the investigator is holding all the cards," Schneider says. "These are never as compelling as multicenter studies." In multi-center trials, any errors or bias contributed by a particular research group has less of an impact.

Another weakness has been the reliance on small groups of research subjects. In the clinical trials to test current Alzheimer's drugs, hundreds of research subjects were involved. When a study involving a dozen or two subjects claims to show that ginkgo helps people with dementia, people like Schneider get suspicious. Was the improvement in one or two people misreported or mismeasured, which then bumped up the average? "You can show improvement with 20 people only when there is an exaggeratedly large and unrealistic effect," he adds. More typically, hundreds of participants are required just to distinguish drug effects from those of the placebo.

The JAMA study: almost but not quite

One recent study of gingko for Alzheimer's avoided some of these pitfalls-the study published in JAMA in 1997. Hundreds of patients with early Alzheimer's took 120 mg of ginkgo per day for a full year, and were tested periodically. The study suggested that ginkgo delayed the progression of the symptoms for up to six months. These widely-publicized findings influenced many people to take ginkgo in hopes of preventing or treating Alzheimer's disease.

But when other researchers took a close look, they found flaws-some fairly serious. For one thing, some of the people in the placebo group did not worsen as much as they should have without treatment. "It was not as it ought to be, and it leaves you wondering if these are truly patients with dementia," observes Paul Solomon, Ph.D., a professor of psychology at Williams College in Massachusetts. Also, the change in mental skills detected by the study was small-about 25 percent of what would be expected in a patient treated with existing Alzheimer's drugs, Solomon says.

Clinical significance

The real kicker was the fact that researchers could not tell the difference between treated subjects and those on placebo. In other words, the study could not prove that the effect of ginkgo on these people was clinically significant. Based on information provided by caregivers and family members, the researchers did document a difference, although it was relatively small: for every seven people treated with ginkgo, caregivers and family members detected improvement in one person on ginkgo.

"The outcome measure that is required in the United States is that observers who are blinded can notice a difference in the patients, and no one could notice any difference," Solomon notes. "If that study was submitted to the FDA to have a drug approved, it would not be acceptable."

The next wave

Despite all the flaws of past studies, the evidence was good enough to launch two new and improved studies of ginkgo for dementia in the United States. In 1999, the National Institutes of Health (NIH) in Bethesda, Maryland, started a $15 million multicenter study to determine if treatment with ginkgo either prevents or delays the onset of Alzheimer's in older people at risk of the disease. The six-year study will involve 2,000 people recruited at four locations around the country.

Another large study of ginkgo for dementia concluded this year. The study is led by Schneider and financed by Dr. Willmar Schwabe Pharmaceuticals, a company based in Germany. Schwabe is also supplying the daily doses of ginkgo 240 mg/day) for the NIH study. The Schwabe trial is a "regulatory quality" study on the level of the trials on which the FDA approved current Alzheimer's drugs, Schneider says. It involved about 500 people with Alzheimer's disease at dozens of different centers who were given 120 or 240 mg of either ginkgo or a placebo and followed for six months. The study concluded earlier this year, and the first results may be announced this summer.

Leaping over the counter

Today, U.S. manufacturers of over-the-counter ginkgo supplements can make general claims that the herb "helps memory" or "improves concentration." The brass ring for a pharmaceutical giant like Schwabe would be to get approval by the FDA to make so-called "health claims," such as "Delays progression of Alzheimer's disease."

The NIH and Schwabe studies may not provide the final word on ginkgo and dementia. Clinical trials are so complex and difficult, it's unusual for a single study, however well-conducted, to settle a medical controversy like the one swirling around ginkgo biloba. Often it takes many years for evidence to pile high enough for experts to reach a consensus. In the meantime, Alzheimer's treatment will continue to rely on what works: skilled and compassionate caregivers and the handful of medications that won a place in physician's formulary by playing by the rules of evidence.

Further Reading:

The Physician's Desk Reference for Nutritional Supplements, by Sheldon Saul Hendler Ph.D., M.D., and David Rorvik. (Montvale, NJ: The Medical Economics Company, Inc., 2001, 700 pp., $59.95). This frank and science-based book reviews the scientific evidence for ginkgo biloba and other herbal remedies.

The US governments Food and Drug Administration (FDA) maintains a website with information on drugs and supplements, as well as information about what claims a specific product may make: http://www.fda.gov.

To learn more about proven therapies for Alzheimer's Disease as well as research on new ones, contact the Alzheimer's Disease Education and Referral (ADEAR) center. It can answer questions about the disease, provide free brochures, and offer referrals to patients and caregivers. Tel: 800-438-4380, or visit the website: http://www.alzheimers.org.

The study by Oregon researchers that reviewed existing gingko studies:

B. Oken and others, "The efficacy of ginkgo biloba on cognitive function in Alzheimer disease," in Archives of Neurology, November 1998, vo. 55, no. 11, pp. 1409-1415.

A few of the many conflicting studies testing the effects of ginkgo on Alzheimer's disease and dementia, including the JAMA article cited in the text:

"A placebo-controlled, double-blind, randomized trial of an extract of ginkgo biloba for dementia," by the North American EGb Study Group, in Journal of the American Medical Association, October 1997, vol. 278, no. 16, pp. 1327-1332.

"The efficacy of ginkgo for elderly people with dementia and age-associated memory impairment: new results of a randomized clinical trial," by M. van Dongen and others, in Journal of the American Geriatric Society, October 2000, vol. 48, no. 10, pp. 1183-1194.

"A 26-week analysis of a double-blind, placebo-controlled trial of the ginkgo biloba extract EGb 761 in dementia," by P. Le Bars and others, in Dementia and Geriatric Cognitive Disorders, July-August 2000, vol. 11, no. 4, pp. 230-237.