Reviewed: Hung S-K, Ernst E. Herbal medicine: an overview of the literature from three decades. J Diet Suppl. 2010;7(3):217-226.

Using focused search terms and limits, the authors identified all published articles on herbal medicine appearing between 1977 and 2007 in any language in the Medline/PubMed database at the US National Library of Medicine. The number of English-language articles on all topics on Medline has dramatically risen during this time, while the number of articles in other languages has fallen. The same was found by these authors to be true for herbal articles. After English, the most articles were in Chinese. The next 3 most common languages were German, Russian, and Japanese, each used in many fewer articles.

Articles published during the sample years 1977, 1987, 1997, and 2007 were further analyzed. For each, the number of herbal medicine articles in any language was compared with the number in each of the 5 most popular languages. Also, for each of these years, the number of systematic reviews (SRs), reviews, clinical controlled trials (CCTs), randomized controlled trials (RCTs), and case reports in any language were calculated.

CCTs and RCTs in any language were identified along with the herbal medicine they investigated in order to determine the herbal medicines most often tested in CCTs and RCTs. Finally, full-text copies of the first 20 English-language RCTs in PubMed results for each sample year were scored for quality on the Jadad scale, the most widely accepted rating scale for determining the quality of the design of clinical trials. For years when there were not 20 RCTs published, the search was extended chronologically until the quota was filled.

Over the 30-year study period, the number of herbal medicine articles published rose from 739 in 1977 to 6,364 in 2007. The largest numbers were reviews and RCTs. CCTs were least frequently reported, rising from 6 in 1977 to 35 in 2007, with only 70 in total for all 4 sample years. English dominated linguistically with over 75% of all articles in the sample years. Chinese, the second most popular language, accounted for 18%.

For each of the 3 decades surveyed, the most frequently studied herbal medicine in CCTs and in RCTs, and the most frequently studied in each of the 5 most popular languages, was determined. Between 1977 and 1987, guar gum (from the bean of Cyamopsis tetragonoloba, Fabaceae) was most studied overall, in 7 of 29 CCTs and 16 of 123 RCTs. Between 1988 and 1997, evening primrose (Oenothera biennis, Onagraceae) oil took the lead in CCTs with 9 of 40, while standardized extract of the leaf of ginkgo (Ginkgo biloba, Ginkgoaceae) led in RCTs with 31 of 512. From 1998 to 2007, ginseng (not further specified, presumably of the genus Panax in the family Araliaceae) had the most CCTs with 11 of 67. Ginkgo again held sway in RCTs with 106 of 2,430. Over the entire 30 years studied, most reports on CCTs in English concerned evening primrose oil and guar gum, with 18 of 337 each, while the most RCTs reported in English were on ginkgo (107/1,873). Ginkgo also predominated in RCTs reported in Chinese (10/1,037) and German (19/89), but Russian-language RCTs most often concerned garlic (Allium sativum, Liliaceae; 3/11), while those in Japanese most often reported on saireito, a traditional kampo and Chinese medicine comprising 12 herbs (2/5). Reports on CCTs in Chinese most often concerned tripterygium (Tripterygium wilfordii, Celastraceae; 2/49). Those in German were most often on ginkgo extract (5/31); those in Russian, on leuzea (Leuzea carthamoides, Asteraceae, syn. Rhaponticum carthamoides); while each of 4 CCTs reported in Japanese concerned a different herbal medicine.

To find 20 RCTs in each of the sample years, the search for 1977 was extended into 1986; for 1987, 1990; and for 1997, 1998. The year 2007 saw 20 or more RCTs published. The shorter successive search periods needed to reach 20 RCTs to confirm the rising number of RCTs throughout the study period. Mean Jadad scores also rose steadily, from 1.9 (standard deviation [SD] 0.85) in 1977 to 2.6 (SD 0.94) in 2007. Between 1977 and 1997, the mean difference reached statistical significance at 1.00 (P=0.014), with a mean score in 1997 of 2.9 (SD 1.07). While similar results have been reported by others investigating, e.g., RCTs in Traditional Chinese Medicine journals or RCTs on most commonly used herbs, the average quality of RCTs on herbal medicines is still rather poor, according to this review. (This appears to conflict with the conclusions of an independent Swiss study in 2007 which found that, in general, the methodological and reporting quality of trials of Western herbal medicine was on average superior to trials of conventional medicine. Although, in both groups a clear majority of studies was of inadequate or uncertain quality.¹) No sample year had an average of 3.0 in Jadad scores. Whether this is better or worse than the average quality of RCTs on conventional medicines is an issue that should not affect the increasing quality of herbal RCTs seen in this study. Use of the herbal medicine-specific Consolidated Standard of Reporting Trials (CONSORT) checklist for reporting RCTs might further improve herbal RCTs as the original CONSORT has been reported to do in conventional medicine studies. (The primary distinction in the CONSORT guidelines for RCTs for conventional medicines and herbal RCTs relates to the recommendation for more fully detailed descriptions of herbal preparations used in herbal RCTs.²)

The main limitation of this article is that, regardless of the comprehensive search terms and limits used, results in Medline depend on how articles are indexed, and there is apparently room for error. For this study, when the search was limited to CCTs, not all articles returned were CCTs. Indexing errors or omissions could have led to errors in all of the data obtained.

—Mariann Garner-Wizard

References

1. Nartey L, Huwiler-Muentener K, Shang A, Liewald K, Jueni P, Egger M. Matched-pair study showed higher quality of placebo-controlled trials in Western phytotherapy than conventional medicine J Clin Epidemiol. 2007; 60(8):787-794.
2. 2. Gagnier JJ, Boon H, Rochon P, Moher D, Barnes J, Bombardier C. Improving the quality of reporting randomized controlled trials evaluating herbal interventions: implementing the CONSORT statement. HerbalGram. 2006; 71:50-56