Echinacea products (ECHINACIN and others) are recommended inter alia for the supportive treatment of recurrent infections of the airways. To date, due to the absence of proven benefit, we have found no indication for their use (a-t 2003; 34: 15).

However, in a recent meta-analysis (2) of the preventive or therapeutic use of echinacea for colds which has also attracted the interest of the press (1), the authors conclude that echinacea exhibits impressive efficacy: Taken prophylactically echinacea extract is claimed to reduce the risk of colds by almost 60%, while used therapeutically it is claimed to reduce the duration of pre-existing colds by a mean 1.4 days. Thus, while drawing on a similar database to that used in a Cochrane review (3), which describes for echinacea unproven benefits in respect of prevention and possible, but not uniformly demonstrated, efficacy in respect of treatment, the meta-analysis reaches more favourable conclusions. The caution expressed by the authors of the present meta-analysis, who call for large-scale studies before a general recommendation can be made, is thoroughly justified on closer scrutiny of the database.

According to the Jadad score* used in the meta-analysis, the studies included are judged to be of moderate to good quality. However, studies with glaring methodological flaws, such as unclear or incorrect randomisation, unvalidated procedures for measuring endpoints and evaluation of multiple end-points without adjustment, also receive this "seal of quality". In one study, although only employees of Madaus, the ECHINACIN producer, were enrolled, blinding was not controlled (4). An additional source of irritation is the fact that the authors of the meta-analysis "weight" the studies according to a principle that is not described in detail. Due to this process a large, qualitatively good negative study (5) exerts almost no influence on the results, whereas small studies with glaring flaws which describe significant benefits are highly weighted (6, 7). An Israeli study (8) in 430 children is most highly weighted with respect to evaluations of both prevention and treatment, although 25% of the study participants withdrew prematurely from the study and were not included in the evaluation. In addition, the statistical procedure was evidently changed in the course of the study. Further, a combination of echinacea, propolis ("bee resin") and vitamin C was tested in this study and, consequently, the reported effects of treatment cannot be safely attributed only to echinacea.

To make matters worse, the results of a further study (9) are wrongly presented: instead of showing findings numerically unfavourable to echinacea (longer duration of illness compared with placebo), the data are turned upside down and included in the calculation as advantage of echinacea. Therefore, the overall meta-analysis is entirely inconclusive and readers could have been spared it.

In the absence of proven efficacy, potential severe side effects assume great significance: pronounced immunological reactions, most conspicuous following parenteral administration (now withdrawn), can also be observed when echinacea is given per os (a-t 1999; No. 3: 40).

	Due to serious methodological errors, the new meta-analysis of the prophylactic and therapeutic efficacy of echinacea products (ECHINACIN and others) lacks significance.
	The efficacy of echinacea remains unproven. For this reason and because of possible severe side effects, we advise against its use.

(R = randomised study, M = meta-analysis)

1

Süddeutsche Zeitung, 28 June 2007, page 16

2

SHAH, S.A. et al.: Lancet Infect. Dis. 2007; 7: 473-80

3

LINDE, K. et al.: Echinacea for preventing and treating the common cold. The Cochrane Database of Systematic Reviews 2007, Issue 2; Status Nov. 2005

4

SCHULTEN, B. et al.: Arzneimittelforschung 2001; 51: 563-8

5

TAYLOR, J.A. et al.: JAMA 2003; 290: 2824-30

6

BRÄUNIG, B., KNICK, E.: Naturheilpraxis 1993; 1: 72-5

7

LINDENMUTH, G.F., LINDENMUTH, E.B.: J. Altern. Complement. Med. 2000; 6: 327-34

8

COHEN, H.A. et al.: Arch. Pediatr. Adolesc. Med. 2004; 158: 217-21

9

BARRETT, B.P. et al.: Ann. Intern. Med. 2002; 137: 939-46

*

Jadad score = scale for methodical evaluation of study quality, ranging from 0 (worst) to 5 (best), in which randomisation, blinding and description of withdrawals and dropouts are the only variables considered.