to the editor: In this article's Strength of Recommendations Taxonomy (SORT) table, three of the recommendations said “results are mixed,” “studies are inconsistent,” or “the evidence is conflicting.” When conflicting results are found in small controlled studies and epidemiologic studies, the most likely explanation is that there is no real effect. Ioannidis has shown why most published research findings are false.¹ Research methodologist Bausell shows in his book Snake Oil Science: The Truth About Complementary and Alternative Medicine that many factors conspire to produce false-positive results, particularly in studies of alternative medicine.²

A rigorously scientific analysis of all the green tea research can be best interpreted as telling us there is no good evidence to support its use for conditions such as cancer, weight loss, or cardiovascular disease. The Natural Medicines Comprehensive Database, a resource for evidence-based, clinical information on natural medicines, rates green tea as only “possibly effective.” This rating is below their ratings of “effective” and “probably effective.”³ It raises a number of concerns about safety and interactions with other drugs, laboratory tests, and diseases. Additional concerns are that the dosage from brewed tea is variable, and commercial green tea extracts are classified as “diet supplements” under the Diet Supplement Health and Education Act and are not regulated for safety, purity, and content. Contamination and inconsistency of such products is common. In my opinion, we should not recommend green tea to our patients unless more convincing evidence is found.

in reply: We appreciate Dr. Hall's thoughtful letter in response to our review of green tea. We agree that the evidence for the health benefits of green tea is mixed; that it is available in several forms, including teas and extracts; that tea strength varies according to brewing technique; and that extracts of green tea are classified as dietary supplements. Potential adverse effects, interactions, and contraindications of various green tea preparations were discussed in our review. The safety of green tea as a beverage is well established and compares favorably with the risks associated with many of the other interventions we use in medicine every day.

Dr. Hall argues that green tea should not be recommended to patients until more data are available. If one takes this approach with all therapies meeting Strength of Recommendations Taxonomy B levels of evidence, few therapies will remain to recommend. We suggest the following approach to the available evidence, which takes into account what is known about safety and effectiveness. Weiger and colleagues¹ and Cohen and colleagues² proposed a “useful criteria for placing individual therapies along a continuum.”¹ Where the evidence clearly supports effectiveness and safety, we can recommend use. Where the evidence is inconclusive but the safety profile is good, we can accept use. Wherever the evidence indicates ineffectiveness or serious risk, we should discourage use.^1,2

We also agree that much published research is questionable. However, we disagree with the assertion that research supporting the use of green tea is more likely to be based on “false-positive results” than research supporting conventional therapies in the larger context of medicine today, where conflicts of interest and industry influence loom large.

Many issues surrounding medical research conspire to lead to false-positive results. Physicians must be vigilant regarding information presented in the medical literature, but evidence suggests that in this regard, complementary and alternative medicine is no more suspect than conventional medicine.

As physicians, many of our decisions will continue to be made in gray areas where we do not have conclusive evidence. For each potential therapy, available evidence should be discussed with patients, and when risk is outweighed by benefit, we can help patients make the most informed decision and monitor for safety and benefit, thus providing responsible patient-oriented care.