to the editor: In Dr. Cupp's article, “Herbal Remedies: Adverse Effects and Drug Interactions,”¹ ginseng is implicated as a cause of decreased response to warfarin. Danshen, another widely used herbal medication, has the opposite effect when used with warfarin: it potentiates the anticoagulant action of warfarin.^2,3

Danshen is a popular herbal medication often used to treat various medical complaints, particularly those associated with coronary artery disease, in the Chinese community.^3,4 Thus, physicians should be alert to the possibility of this drug interaction when anticoagulation control becomes difficult or deviates from the optimum, and when no other causes are apparent.^3,5

to the editor: Complementary or alternative medicine has become increasingly popular among many patients, as evidenced by the fact that approximately one in three Americans uses herbal products as part of a therapeutic regimen.¹ Findings from a 1997 survey revealed that total patient visits to alternative medicine practitioners exceeded visits to primary care physicians during that year; the out-of-pocket expenses paid to alternative medicine practitioners surpassed the out-of-pocket expenditures for all hospitalizations in the United States.¹ Burgeoning interest in botanical medicine has led to greater scientific scrutiny concerning product safety and efficacy, and the adverse effect potential of drugs and herbs, as discussed in Dr. Cupp's article.²

Although touted as “natural,” herbal medicines should not be considered inactive or benign. For example, many products contain several different compounds with variable concentrations depending on plant genetics, growth and harvest conditions, processing techniques and plant parts used. The potential for contamination and adulteration, which occurs more frequently with imported medicinals or raw herbs, must also be considered. The association between dosing and toxicity is especially problematic in pediatric patients.^3,4 Unfortunately, prospective studies of the long-term efficacy of alternative medicine are limited. However, the U.S. Food and Drug Administration has identified 250 herbs as “generally recognized as safe” by using information derived from their historical use and lack of reported side effects.⁵

It is crucial that primary care physicians ask patients about their use of botanical remedies, because most patients will not disclose such information. In some cases, having such information may be lifesaving, as evidenced by the following case. A young immigrant woman presented with a history suggestive of a viral syndrome. However, her vital signs were suspicious: blood pressure, 100/80 mm Hg; pulse rate, 47 beats per minute; respiratory rate, 16 breaths per minute; and temperature, 38.9°C (102°F). The patient was sent home with the reassurance that this was a benign illness, only to return several hours later with a deterioration of her condition. During her second visit, it was disclosed that she had been taking oleander leaves in tea in order to relieve her menstrual pain, which began at the same time as her upper respiratory complaints. The lesson from this case was clear: an appropriate history, especially in patients from different cultures, should include an inquiry about the use of home remedies.

The oleander plant is a beautiful, ubiquitous plant found in the western United States, which can cause bradycardia and heart block similar to that seen with digoxin toxicity. In this situation, the patient's slow heart rate should have raised concerns about another diagnosis, because most febrile illnesses cause tachycardia instead of low heart rate. Simply questioning the patient about what treatments had been used at home may have uncovered the reason for the abnormal cardiac response. The increased number of patients born outside of the United States who are using the U.S. health care system has brought an increased risk of adverse effects from traditional or indigenous therapies. In response, family physicians should learn about the most common herbal products and the toxic reactions they can cause.

to the editor: Dr. Cupp's article begins the important task of addressing herb/drug interactions,¹ but her article may be misleading. For example, Dr. Cupp states that “herbal products are not tested with the scientific rigor required of conventional drugs . . .” and “herbal health products aren't tested to be sure they're safe, so they may cause problems (patient information sheet).”¹ However, she fails to address the many recent reviews and meta-analyses^2–5 that assess double-blind, placebo-controlled trials of botanical medicines that point to convincing evidence of efficacy. Also, a complete discussion of the subject should mention the many economic⁴ and political reasons for the historical lack of scientific studies on herbal medicines in the United States. Many of the science and efficacy reviews have taken place in other countries.⁶

We agree that patients and practitioners need to be educated about the safe use of herbal medicines. Special emphasis should be placed on plants that have well-documented adverse effects or contraindications, such as Ephedra species when taken in high dosages or in certain medical conditions. It is also the responsibility of scientists and specialists in botanical medicine to avoid confusion such as that created by Dr. Cupp's warnings (see patient information sheet) about the dangers of plant pollen adulterants (has any such problem ever been documented?) or by her pointing out that herbs “aren't natural to the human body,” a statement that was not well defined and doesn't help patients learn how to stratify risk or make well-informed decisions about natural and synthetic substances.

Furthermore, it would be helpful to patients and practitioners if the herbal side effect profiles were qualified and put into the context of their pharmaceutical counterparts. Most herbal products are extremely safe in comparison with their pharmaceutical counterparts, and adverse outcomes, such as the list for St. John's wort, are rare. While the annual mortality rate attributable to conventional medicine numbers in the tens of thousands, herbal medicines cause a dozen or so deaths each year. With similar numbers of people at risk, there can be little doubt about which type of treatment poses a higher risk. We need this kind of objective information to help make intelligent decisions about current use of botanical medicines and to illustrate the need for future research.

in reply: Drs. Kiefer and Barrett are correct that meta-analyses and clinical studies of some herbal products exist. Unfortunately, such information is unavailable for the vast majority of herbal products and therapeutic claims. The U.S. Food and Drug Administration requires studies that are appropriately designed to assess drug safety and efficacy. Usually, double-blind, placebo-controlled studies involving thousands of patients are required. Inclusion and exclusion criteria must be delineated, and statistical methods used to assess the data must be appropriate. Meta-analyses and reviews of studies that may not have been appropriately designed are not substitutes for this process. In addition, the results of studies using one herbal preparation may not be applicable to other herbal products that were prepared using a different method and by a different manufacturer.

Drs. Kiefer and Barrett state that “herbal medicines cause a dozen or so deaths each year,” but do not reference this statement. This lack of data underscores the importance of clinicians reporting problems associated with herbal products via the U.S. Food and Drug Administration's MedWatch program. Previously undocumented adverse effects and drug interactions continue to be reported in association with herbal products. For example, although St. John's wort has been used for many years worldwide, its ability to induce theophylline metabolism was just recently recognized.¹

Drs. Kiefer and Barrett are not comfortable with my statement that herbs are not natural to the human body. My point was that “natural” is not synonymous with “safe.” Just because the chemicals in herbs are manufactured by plants, it does not make them safer than chemicals manufactured in the laboratory. Patients have the right to know this.

Patients also have the right to know exactly what is in the products that they purchase. Drs. Kiefer and Barrett ask if there is documentation for the statement concerning pollen contamination. I do not have documentation for this because this information was added later by the editors. However, the warning does not state that this has occurred, only that it might occur. Furthermore, contamination of herbal products by heavy metals,² benzodiazepines,³ steroids, NSAIDs, hepatotoxic herbs⁴ and digitalis⁵ is well documented.

I appreciate the additional information provided by Drs. Cheng and Bullock regarding oleander and danshen, and caution readers that my article is not meant to be a comprehensive, all-inclusive review of all of the adverse effects and interactions that have been reported with herbal remedies.