Editor's Note: This is the second in an occasional series of articles on psychopharmacological treatments. The articles were solicited by George M. Simpson, M.D., editor of the journal's Psychopharmacology column, who has found that some important topics require more extensive discussion than is possible in the relatively short space of a column. The first article in this series reviewed pharmacology of depression in children and adolescents (see the May 2000 issue, pages 627-633). Another will focus on attention-deficit hyperactivity disorder across the life span.
Once considered a counterculture phenomenon, herbal medicines are now a $4 billion industry, and herbal products are readily available in drugstores and supermarkets (1). It has been estimated that one of every three Americans has used herbal remedies (1), with 50 million people using them in a given month (2). Four of the 12 most commonly used herbal medications—St. John's wort, kava, ginkgo, and valerian—are taken for the prevention or treatment of psychiatric symptoms, which is the most rapidly growing segment of the herbal product market (1).
The increasing interest in and use of herbal remedies by patients has created a need for psychiatrists and other mental health professionals to become familiar with the effects of more commonly used herbs, including their risks, side effects, and contraindications. Unfortunately, this topic is not covered in most medical school curricula, psychiatry residency programs, or textbooks of psychiatry and psychopharmacology. However, this situation is beginning to change, as evidenced by a recent issue of JAMA devoted to alternative medicine (November 11, 1998).
Our review is intended as a brief introduction to the subject. Readers are referred to several recent texts and publications for further information (3,4,5,6,7).
St. John's wort (Hypericum perforatum) is an aromatic perennial that is native to Europe but now grows wild in parts of Asia, North America, and South America. Its use can be traced back to the texts of the ancient Greek physicians Hippocrates and Galen. In the past decade it has become the second most commonly used herbal remedy in Germany (3) and is currently used in that country for the treatment of depression four times more often than the most commonly used prescription antidepressant (8). In the United States, St. John's wort is the second most commonly purchased herbal product. An estimated 17 percent of Americans have taken products containing St. John's wort (1).
Much of the popularity of St. John's wort in the United States and other English-speaking countries followed the publication of a review article by Linde and colleagues (9) in the British Medical Journal in 1996. The article received considerable publicity in the lay press (8,10). More than two dozen reports of clinical trials of hypericum extracts have been published; the majority of trials have been conducted with methanol extracts of the herb (3,9,11).
However, if the analysis is restricted to well-controlled trials using standardized dosages and standard outcome measures, nine studies are of interest (3). Five of these were placebo-controlled studies using 900 mg per day of an aqueous methanol extract referred to in the literature as LI 160, with a treatment duration of at least four weeks. The five studies (3) used the Hamilton Rating Scale for Depression (HAMD) as an outcome measure (12). Overall, patients receiving the extract showed a slightly greater improvement in HAMD scores than those receiving placebo. When response was defined as at least a 50 percent decline in HAMD scores, 61 percent of the patients receiving the extract responded, compared with 24 percent receiving a placebo.
Four studies of St. John's wort compared a dosage of 900 mg per day of the LI 160 preparation of hypericum against low dosages of maprotiline, imipramine, and amitriptyline for patients with mild to moderate depression (13,14,15,16). Although no statistically significant differences in response were found, the studies were small, raising the possibility of type II error, and no placebo control groups were included, raising the question of whether the conventional antidepressants were any better than placebo at the dosages used.
Because of the relatively small number of well-controlled studies of St. John's wort, the National Institutes of Health recently funded a multisite study comparing the LI 160 preparation with sertraline and placebo in an eight-week trial. Results should be available in two to three years.
The components of hypericum extracts that may be responsible for any antidepressant actions are unknown. Extracts of St. John's wort contain a large number of compounds (3,17), and although it is often assumed that hypericin is the active ingredient, it has not been proved. The mechanism of action is also unclear. Early studies suggested that extracts of St. John's wort inhibited monoamine oxidase (18), but more recent studies have failed to find such inhibition at the tissue concentrations that occur following typical dosages (19,20). Instead, it appears that the extract may block the reuptake of norepinephrine and serotonin and down-regulate serotonin receptors (21).
Side effects of hypericum extracts are mild and uncommon. In a study of more than 3,000 patients taking hypericum, only 2.4 percent reported side effects, primarily allergic reactions and gastrointestinal upset (22). Other reported side effects include dry mouth, sedation, and headache (16). Photosensitivity appears to be a risk at dosages higher than those typically used to treat depression (23). There do not appear to be significant adverse effects on cardiac conduction (24). The use of hypericum extracts is contraindicated in pregnancy and lactation due to inadequate safety data (25).
Because the mechanism of hypericum extracts may involve serotonin reuptake blockade, extracts should not be combined with monoamine oxidase inhibitors or selective serotonin reuptake inhibitors, as the combination could cause a serotonin syndrome (26). In addition, an interaction between hypericum extracts and olanzapine has been reported, with a patient having a 300 percent increase in olanzapine levels after starting St. John's wort (27). The latter interaction may have been due to an effect of some component of the extract inhibiting CYP 1A2 and thus interfering with the metabolism of olanzapine.
The kava shrub (Piper methysticum) is native to Polynesia and the Pacific Islands, and it has traditionally been taken by Pacific Islanders as a beverage mixed with water and coconut milk (3). Kava is becoming a popular herbal product in the United States, with sales increasing rapidly (1). Most medicinal forms are either ethanol-water or acetone-water extracts (3).
Kava is one of the few herbal remedies in which the pharmacologically active ingredient is known. Meyer (28) proved that the effects of kava are due to the kavapyrones, which in animal models act as muscle relaxants and anticonvulsants, protect against strychnine poisoning, and reduce limbic system excitability. Exactly how the kavapyrones produce these effects is unclear, as they have a variety of actions involving inhibition of voltage-dependent sodium channels, increasing GABAA receptor densities, blocking norepinephrine reuptake, and suppressing the release of glutamate (29,30,31,32).
Several double-blind, placebo-controlled trials have been conducted using a standardized extract containing 70 percent kavapyrones given in a dosage of 210 mg a day (3). In two of these studies, a significant difference in scores on the Hamilton Anxiety Rating Scale (HAMA) (33) was seen after only one week of treatment; in the third and largest study, a significant difference in HAMA scores was seen after eight weeks, with the difference continuing until the study terminated at week 25 (3). Several double-blind studies have also demonstrated that DL-kawain, one of the kavapyrones, in dosages of 200 to 600 mg a day is more effective than placebo as measured by reduced HAMA scores (3).
The major criticism of all of these studies has been their ill-defined patient populations; in addition, many had small sample sizes and were of short duration. Although it appears that kava extracts have anxiolytic properties, it remains to be seen for which patients they may be most useful and how they compare with conventional anxiolytics in efficacy.
When kava has been taken in dosages ranging from 100 to 210 mg of kavapyrones daily, it has been associated with few adverse effects. In studies comparing kava and oxazepam, kava appears not to adversely affect cognitive function, mental acuity, or coordination (34,35), although slight morning tiredness and reduced reactivity while driving have been reported (4), as has ataxia (36). In rare cases, kava may lead to allergic reactions, yellowing or scaling of the skin, gastrointestinal complaints, pupil dilation, and blurred vision (4,34). Heavy use by Australian aborigines and Pacific Islanders has been associated with hepatotoxicity, hematuria, macrocytic anemia, ataxia, increased patellar reflexes, weight loss, hair loss, and rash (3,37). However, aborigines in one study ingested 50,700 mg a day (37); a typical therapeutic dosage is 210 mg. Kava is contraindicated during pregnancy and lactation (4).
Kava may potentiate the effects of alcohol, benzodiazepines, and other sedative-hypnotic agents through additive effects (4,38). Patients taking benzodiazepines should be advised not to take kava.
Ginkgo trees (Ginkgo biloba) are native to East Asia and are grown ornamentally in Europe and North America. Used in China for more than 2,000 years as a tea for treatment of asthma, ginkgo is now the most commonly sold herbal product in Germany and one of the top three herbals in the United States, where it is taken primarily to prevent or treat memory problems (1,3). U.S. sales increased markedly after a report in JAMA of a clinical trial of the use of kava for patients with dementia (39).
Standardized commercial preparations usually contain the active components flavone glycosides (24 percent) and terpenoids (6 percent). The extracts most commonly used therapeutically are designated as EGb 761 and LI 1370 (3).
A review by Kleijnen and Knipschild (40) of more than 40 controlled trials of gingko showed that all but one found clinically significant improvement in symptoms such as memory loss, concentration difficulties, fatigue, anxiety, and depressed mood. However, most of the studies had poorly defined patient populations and small sample sizes and used nonstandard outcome measures.
A more recent 52-week, randomized, double-blind, placebo-controlled, multicenter study of more than 300 patients with Alzheimer's disease or vascular dementia used the EGb 761 extract at a dosage of 120 mg a day (39). The group taking ginkgo extract showed significantly less decline on two of the three standardized rating scales. The outcome measure was improvement of 4 points or more on the cognitive subscale of the Alzheimer Disease Assessment Scale (41), which is roughly equivalent to a six-month reversal of symptoms. On the basis of this measure, 27 percent of the EGb group improved, compared with 14 percent of the placebo group. Although these effects are modest, they may be valuable to patients and families and are not unlike results from studies of cholinesterase inhibitors. More study is clearly warranted to determine which patients might benefit from this treatment.
Ginkgo extracts contain a large number of substances that have been found to have a variety of pharmacological effects (3,40). The ginkgo flavonoids are thought to be antioxidants, and the ginkgolides, especially ginkgolide B, inhibit platelet-activating factor. There is also evidence that ginkgo extracts can improve vascular perfusion by modulating vessel wall tone.
Side effects from ginkgo are relatively uncommon but include headache, gastrointestinal upset, and allergic skin reactions (40). Rarely, ginkgo preparations have been associated with cerebral hemorrhage (42). The safety of ginkgo in pregnancy or lactation has not been established.
Because ginkgolide B is a potent inhibitor of platelet-activating factor, ginkgo extracts have the potential to interact with platelet-antiaggregating and antithrombolytic therapies (4,42). In addition, they should be used with caution by patients who consume alcohol or who have other risk factors for hemorrhagic stroke (42).
Approximately 250 different species of valerian exist. The one used most commonly for medicinal purposes (Valeriana officinalis) is a perennial that is native to Europe and Asia. Valerian root is frequently made into a tea by adding 3 to 5 g of dried valerian root to hot water and straining after ten to 15 minutes (4). In addition, a variety of extracts and tinctures have been prepared, with considerable differences in composition between the aqueous and ethanol extracts (3). Preparations available in the United States are often mixtures that include other ingredients such as passion flower.
Several studies of the effects of valerian extracts on sleep have been conducted. In healthy human subjects, 400 to 900 mg of valerian extract decreased sleep latency and nocturnal awakenings and improved subjective sleep quality (43,44,45). However, placebo effects were marked in some studies, and in some cases the beneficial effects of valerian were not seen until after two to four weeks of therapy.
Valerian extracts contain more than 100 different constituents. Which of them are responsible for the pharmacological actions is not known with certainty, as the whole valerian extract has been demonstrated to have central nervous system actions not attributable to valeric acids, valepotriates, or volatile oils (3). In laboratory animals, valerenic acids have sedative and anticonvulsant effects, and valerian extracts have been demonstrated to have a variety of effects on GABA-ergic neurons, including increased release of GABA, decreased GABA reuptake, and decreased GABA degradation (3).
Adverse effects of valerian preparations are rare but may include gastrointestinal upset, contact allergies, headache, restless sleep, and mydriasis (4). Valerian appears to be relatively safe in overdose (46), with the major effect being central nervous system depression (47). Currently, little is known about its safety in pregnancy and lactation.
The major drug interactions of valerian are with other sedative-hypnotics. The sedative effects of valerian may potentiate the effects of other central nervous system depressants.
Herbal teas have long been consumed for their presumed sedative-hypnotic effects. Besides valerian, common ingredients in such teas are hops, lemon balm, chamomile, and passion flower, although a variety of other herbs have also been used (3). The best studied of these is chamomile. Although chamomile has been widely used as a folk remedy (3,6), there does not appear to be experimental confirmation of its effectiveness as a sleep aid. However, receptor-binding studies have found components of chamomile extract to bind to GABA receptors (6). Far less is known of the effectiveness of the other herbs.
The herbal preparations described here and conventional medications are regulated in very different ways. In the United States, herbal preparations are regulated as dietary supplements under the Dietary Supplement Health and Education Act of 1994 (5). This situation differs considerably from that in Germany, where herbal preparations are regulated more like drugs (3,5). For this reason few controlled studies of the efficacy and safety of herbal preparations are published in English-language journals, making it difficult for most American psychiatrists to gain access to information about these compounds. Furthermore, compared with conventional medications, much less is known about the safety and efficacy of these compounds.
Another important implication of the less rigorous U.S. regulation is that herbal preparations are not standardized in the same way that pharmaceutical products are. Products may vary greatly in their composition depending on variation in the raw plant material due to genetic (varietal) factors, climate, growing season, soil, rainfall, and other growing conditions; method of preparation; and the type of solvent used in the extraction process (3). Thus tea brewed from a dried herb will differ in composition from an alcohol extract, and both will differ from a volatile oil distilled from the crude plant material. Because these herbal products contain a variety of compounds, and because the active compounds are often unknown, it is difficult to standardize preparations. In addition, adulteration and substitution may occur when the plant material is expensive.
For all of these reasons, we cannot conclude, for example, that a U.S. product that is described as containing a certain amount of ginkgo extract would be the same as a product such as EGb 761 made in Germany, where standardization is more exact (5). Thus it is extremely difficult to generalize the results of European studies in which well-standardized preparations are used to the results that might be obtained using American preparations purchased in a drugstore, health food store, or supermarket. This difficulty is a further concern when investigators try to evaluate the safety and efficacy of these products.
This problem is illustrated in a report published by the Los Angeles Times in which samples of ten different brands of St. John's wort were sent to an independent laboratory for analysis (48). The amount of hypericin, a component that is often used for standardization purposes (3), varied from 20 to 140 percent of the amount claimed on the label, with half of the brands containing less than 80 percent of the labeled amount, and two brands containing more than 120 percent of the labeled amount. Thus one cannot generalize from the results obtained with one preparation to those that might be expected with a different preparation, even if the labels indicate that they are similar.
It is becoming increasingly important for physicians to be familiar with the herbal remedies commonly used in the patient populations they serve. Our patients are frequently exposed to sketchy information about these products through the media or marketing campaigns, and they will quite rightly expect their physicians to be able to answer their questions about indications, risks, interactions, and side effects of herbal products.
In addition, increasing numbers of clients express a preference for the use of remedies they perceive to be "natural," and physicians familiar with the available research will be better equipped to make informed and safe decisions about whether it is appropriate to recommend herbal remedies in selected cases.
Although evidence of the efficacy of certain herbal preparations in the treatment of psychiatric conditions is growing, translating the results of efficacy studies into effective treatments for patients is hampered by the chemical complexity of the products and the lack of standardization of commonly available preparations, not to mention the paucity of well-controlled studies of safety and efficacy. In particular, well-controlled studies comparing herbal remedies with conventional medications are few. As a result, it is premature for psychiatrists to recommend herbal remedies over established conventional treatments.
Dr. Beaubrun is a resident in general psychiatry at the Martin Luther King, Jr.-Drew Medical Center in Los Angeles. Dr. Gray is professor and chairman of the department of psychiatry and human behavior at the Charles R. Drew University of Medicine and Science and director of the Augustus F. Hawkins Comprehensive Community Mental Health Center at the King-Drew Medical Center, 1720 East 120th Street, Los Angeles, California 90059 (e-mail, firstname.lastname@example.org). Send correspondence to Dr. Gray.