key: cord-0743924-xx8s0jea authors: Shah, Sachin A; Sander, Stephen; White, C Michael; Rinaldi, Mike; Coleman, Craig I title: Evaluation of echinacea for the prevention and treatment of the common cold: a meta-analysis date: 2007-06-25 journal: Lancet Infect Dis DOI: 10.1016/s1473-3099(07)70160-3 sha: 483e7cca05d9d7f861d2b6eddca54b686858122d doc_id: 743924 cord_uid: xx8s0jea Echinacea is one of the most commonly used herbal products, but controversy exists about its benefit in the prevention and treatment of the common cold. Thus, we did a meta-analysis evaluating the effect of echinacea on the incidence and duration of the common cold. 14 unique studies were included in the meta-analysis. Incidence of the common cold was reported as an odds ratio (OR) with 95% CI, and duration of the common cold was reported as the weighted mean difference (WMD) with 95% CI. Weighted averages and mean differences were calculated by a random-effects model (DerSimonian-Laird methodology). Heterogeneity was assessed by the Q statistic and review of L'Abbé plots, and publication bias was assessed through the Egger weighted regression statistic and visual inspection of funnel plots. Echinacea decreased the odds of developing the common cold by 58% (OR 0·42; 95% CI 0·25–0·71; Q statistic p<0·001) and the duration of a cold by 1·4 days (WMD −1·44, −2·24 to −0·64; p=0·01). Similarly, significant reductions were maintained in subgroup analyses limited to Echinaguard/Echinacin use, concomitant supplement use, method of cold exposure, Jadad scores less than 3, or use of a fixed-effects model. Published evidence supports echinacea's benefit in decreasing the incidence and duration of the common cold. According to the National Institute of Allergy and Infectious Diseases, the US population has 1 billion colds annually. Adults have between two and four colds per year, whereas children have between six and ten colds. 1 Although rhinovirus and coronavirus are the most common viruses precipitating cold symptoms, approximately 200 other viruses are also known to cause the common cold. In the USA, about 40% of lost work time and 30% of time lost from school are attributed to symptoms caused by the common cold. 2 The common cold is also associated with a large fi nancial burden on society, with about US$1·5 billion spent annually for physicians' visits and another $2 billion spent on nonprescription cough and cold treatments. 3 In 2002, approximately 20% of the adult US population used nutraceuticals (herbal products, functional foods, animal based supplements). Echinacea, a collection of nine related plant species indigenous to North America, was the most common nutraceutical used and was consumed by over 40·3% of these people. 4,5 Echinacea angustifolia, Echinacea pallida, and Echinacea purpurea (fi gure 1) are the most common species recognised for their medicinal value. 6 The mechanism of action underlying the proposed immunostimulatory eff ects of echinacea remains unclear. Some evidence suggests that upregulation of tumour necrosis factor-α mRNA, which is stimulated by agonistic activity of the cannabinoid receptor (CB2) by alkamides present in echinacea, has a role. 7 The German Commission E, WHO, and the Canadian Natural Health Products Directorate have advocated echinacea use for the common cold. [8] [9] [10] [11] [12] However, there is controversy about the effi cacy of echinacea for the prevention or treatment of the common cold with some studies showing benefi t and others showing a null eff ect. Meta-analysis can be useful in situations such as this, since it can show what the preponderance of evidence in the published work suggests. A past systematic review by Melchart and colleagues 13 concluded that echinacea preparations from the aerial part of the plant were eff ective for the treatment of colds but the evidence for the prevention of a cold was lacking. 13 It is important to note, however, that this review excluded studies using an experimental rhinovirus infection and echinacea preparations with supplements, and it did not include a more recent study by Turner and colleagues. 14 We therefore did a meta-analysis evaluating the eff ect of echinacea on the incidence and duration of the common cold in randomised placebo-controlled studies. A literature search using the terms "Echinacea" and "Purple conefl ower" (limited to human beings and clinical trials) was done by three independent reviewers (SAS, CMW, and CIC) using Medline (1966 to April, Trials were included for analysis if they met the following inclusion criteria: randomised placebo-controlled trials evaluating echinacea-containing products in the prevention and/or treatment of the common cold with adequately reported data on either cold incidence or duration. In cases where a study evaluated the eff ects of diff erent echinacea species or formulations compared with placebo, when possible the data from the echinacea arms were pooled and compared with the placebo arm. 14, 15 When data were reported separately for bacterial and viral infections, only the latter was extracted for inclusion in the analysis. 13 The following methodological features, most relevant to the control of bias, were assessed: randomisation, random allocation concealment, masking of treatment allocation, blinding, and withdrawals. Jadad scores were calculated to aid in the identifi cation of reports with overall weaker study methodologies. 17 All studies were reviewed and evaluated by three independent reviewers (SAS, CMW, and CIC) with disagreement resolved by consensus. All data were independently abstracted by three investigators (SAS, CMW, and CIC) through the use of a standardised data abstraction tool. The following information was sought from each article: author identifi cation, year of publication, geographical location of the study, study funding source, type of study design (prospective or retrospective, randomised or observational, presence and type of control, blinded or open-label), study population, sample size, duration of patient followup, echinacea product used (specifi c species, dose, preparation type, and branded or unbranded), presence or absence of concomitantly administered supplement, mode of virus exposure (natural or inoculation), and defi nition for incidence or duration of cold (when reported). Incidence of the common cold was treated as a dichotomous variable and reported as an odds ratio with its 95% CI using a DerSimonian and Laird random-eff ects model. Calculation of odds ratios can be problematic when there is an absence of events in one of the comparator groups. For these studies, a nominal value (0·5 colds) was added in each 2×2 cell to enable calculation of an odds ratio. 18, 19 Risk diff erence was also calculated to aid in the assessment of not only statistical but clinical signifi cance as well. Duration of illness was treated as a continuous variable and the weighted mean diff erence (WMD) was calculated as the diff erence between the mean days of the common cold in the echinacea and control groups. Again, a DerSimonian and Laird random-eff ects model was used in calculating the weighted mean diff erence and its 95% CI. Only one study 20 provided 95% CIs for continuous data, for this study the standard deviation of the mean was calculated from the 95% CI using standard methods. 19 Statistical heterogeneity was addressed using the Q statistic (p<0·1 considered signifi cant). Heterogeneity was also assessed through visual inspection of L'Abbé plots. All statistical analyses were done using StatsDirect Version 2.4.6 (StatsDirect Ltd, Cheshire, UK). Numerous subgroup analyses to assess sources of clinical heterogeneity were done. The concomitant administration of additional nutraceuticals with echinacea could potentially result in synergistic, additive, or inhibitory interactions. Therefore, echinacea's effi cacy was assessed both in the presence and absence of other nutraceuticals. Because of the lack of regulation of herbal products, concern has arisen regarding the content of Analyses included in study Incidence in echinacea group* Incidence in control group* Turner et al (2005) 14 Incidence 21 Furthermore, only nine (43%) of 21 standardised preparations met the quality standards as described on the label. 21 Five studies in this meta-analysis included an E purpurea product extracted in 22% alcohol (Echinaguard and Echinacin, Madaus AG, Cologne, Germany). [22] [23] [24] [25] [26] [27] As such, an analysis of benefi t using these two products was done in a subgroup. Finally, studies included in this meta-analysis examined patients who were exposed to (or contracted) a cold either naturally 13, 15, 16, [23] [24] [25] [26] [27] [28] [29] [30] or through investigator inoculation. 14, 22, 31 Since the eff ect this might have had on the effi cacy of echinacea was not known, separate analyses were done to evaluate studies using natural and investigator-inoculated virus. Studies of poorer methodological quality, such as unblinded or open-labelled trials might exhibit exaggerated treatment eff ects. Excluding them might result in increased internal validity but could reduce external validity of the analysis. Additionally, the selection of a random versus fi xed-eff ect model in meta-analyses is controversial. The use of a random-eff ect model in the calculation of confi dence intervals results in wider intervals and thus a more conservative estimate of treatment eff ect compared with a fi xed-eff ect model. To reconcile these issues, sensitivity analysis was done, in which the meta-analysis was reanalysed excluding studies of weaker methodology (Jadad score less than 3) and using a fi xed-eff ects model (Mantel-Haenszel methodology). Egger weighted regression statistics and a visual inspection of funnel plots were used to assess for the presence of publication bias. Study identifi cation, inclusion, and exclusion are shown in fi gure 2. Our initial search strategy yielded 738 citations. Turner (2005) 14 Cohen (2004) 28 Sperber (2004) 22 Schulten (2001) 23 Turner (2000) 31 Grimm (1999) 24 Berg (1998) 25 Melchart (1998) 15 Hoheisel (1997) 27 Combined (random) 0·1 0·2 Odds ratio (95% CI) Of these, 665 were excluded manually and electronically by limiting our search to human beings and clinical trials. The remaining 73 abstracts were reviewed of which 50 were excluded for evaluating echinacea for outcomes other than cold incidence or duration. Therefore, 23 abstracts remained and underwent full-text article review. Eight of the 23 studies did not report data on either primary endpoint of the analysis (incidence or duration), and one of the 23 used an active control. Therefore, 14 unique studies [14] [15] [16] 20, [22] [23] [24] [25] [27] [28] [29] [30] [31] [32] were therefore included in this meta-analysis, encompassing 1356 study participants for incidence and 1630 participants for duration. Table 1 shows characteristics of the included studies. Seven studies evaluated monotherapy with E purpurea, [22] [23] [24] [25] 27, 30, 32 one study evaluated E angustifolia, 14 one evaluated E pallida, 13,16 one study did not specify which specifi c species of echinacea was studied, 31 and four studies evaluated a combination of diff erent Echinacea species. 15, 20, 28, 29 Two studies evaluated echinacea's eff ect in children. 28, 32 Five studies used either Echinaguard or Echinacin products made by one company, Madaus AG. [22] [23] [24] [25] 27 Virus exposure using rhinovirus inoculation was done in three studies, 14, 22, 31 and four studies evaluated the eff ect of echinacea along with a supplement. 20,28-30 A summary of individual study data on the incidence and duration of colds is provided in table 2. Meta-analysis showed that echinacea decreased the odds of a patient contracting a cold by 58% (odds ratio [OR] 0·42, 95% CI 0·25-0·71; Q statistic p<0·001), corresponding with a risk diff erence of -0·17 (-0·25 to -0·08; number needed to treat 6). Echinacea was also found to decrease the duration of cold by 1·4 days (WMD -1·44, -2·24 to -0·64; p=0·01), as shown in fi gure 3 and fi gure 4. The Q statistic showed signifi cant heterogeneity in both the incidence and duration analyses. However, review of the L'Abbé plot for incidence showed that included studies generally agreed on echinacea's positive eff ect, but not the magnitude of the benefi t (fi gure 5). Some degree of asymmetry was noted upon review of the funnel plots for both the incidence and duration analyses, resulting in our inability to rule out the presence of publication bias in our analyses (fi gure 6). However, when publication bias was assessed using the Egger weighted regression statistic, no signifi cant publication bias was detected for either the incidence or duration analyses (p=0·64 and p=0·79, respectively). Table 3 depicts the results of subgroup and sensitivity analyses. Regardless of whether echinacea was administered in the presence or absence of other supplements or nutraceuticals, substantial reductions in the incidence of the common cold were seen. Whereas the subgroup of those receiving echinacea with a supplement showed a signifi cant eff ect on shorting the duration of cold in its own right (p<0·0001), the subgroup receiving echinacea without a supplement showed only a trend towards benefi t (p=0·27). In the analysis limited to fi ve studies evaluating Echinaguard or Echinacin products, [22] [23] [24] [25] 27 similar signifi cant reductions in patients' odds compared with the overall analysis were observed (p=0·0009). A reduction in the odds of contracting a cold was observed when virus exposure occurred naturally or was investigator inoculated. A decrease in duration of cold was also maintained when only natural virus exposure studies were evaluated. 13, 15, 16, 20, [28] [29] [30] 32 For the endpoint of cold duration, no studies used Echinaguard/Echinacin or evaluated investigational rhinovirus inoculation; thus these analyses could not be undertaken. Finally, neither the use of a fi xed-eff ects model instead of a random-eff ects model nor the exclusion of studies with a Jadad score less than 3 had any eff ect on overall study conclusions. More than 800 products containing echinacea are available, which come in tablet, extract, fresh juice, tincture, and tea formulations. 33 There are three commonly used species of echinacea, diff ering parts of the plant can be used in diff erent products (fl ower, stem, root), and the same plant species may contain diff ering levels of constituent molecules in diff erent parts of the year or geographical location. Although concentration variances exist, all three species of echinacea contain water-soluble polysaccharides, a lipophilic fraction (alkamides, polyacetylenes), caff eoyl conjugates (echinacoside, chicoric acid, caff eic acid) and fl avonoids. 34, 35 It is yet to be determined if it is one, a few, or the combined eff ect of many constituents (mainly alkamides, chicoric acid, and polysaccharides) that induce immunostimulation. Despite all of these factors that can infl uence the effi cacy of echinacea and the diff erent doses Review that can be used, the results of our meta-analysis show that echinacea reduces the incidence as well as the duration of the common cold. Since the Echinaguard or Echinacin products both contain the fresh pressed juice of E purpurea in 22% alcohol extract, were manufactured by the same company, and were evaluated in fi ve diff erent studies, [22] [23] [24] [25] [26] [27] we analysed the benefi ts of these products separately and found a reduction in the incidence of cold by 56%. This might be important since the variability in the echinacea product evaluated between these trials would be minimised. Our meta-analysis had only one cold incidence study 28 that used echinacea with other supplements (vitamin C and propolis). This study found an 86% reduction in the incidence of the common cold. As such, we cannot determine if the combination of echinacea with other nutraceuticals yields better results than echinacea alone. Several experimental studies have shown that vitamin C might have eff ects on the immune system. 36 Propolis, a natural resinous product collected by honeybees from various plant sources, has also been used in the prevention of respiratory infections. 37 For cold duration as the outcome, four studies 20,28-30 used echinacea combined with additional supplements (vitamin C, propolis, lemongrass leaf, spearmint, peppermint, thyme, citric acid, rosemary leaf, eucalyptus, and fennel seed) and yielded a 1·3-day shorter duration of cold than placebo. Echinacea used alone, although showing a similar benefi t, did not show a signifi cantly shorter duration of cold than placebo (p=0·27), suggesting that this sub-group analysis was underpowered. Comparing the results on duration of cold in the overall analysis to the subgroup analyses suggests that the benefi t is caused by echinacea rather than the other supplements. We evaluated the method of viral exposure on the outcome of cold induction. If echinacea was given prophylactically in an attempt to reduce the incidence of natural cold induction, the incidence was reduced by 65% versus placebo. When echinacea was given as prophylaxis against cold induction caused by direct rhinovirus inoculation, the incidence was only reduced by 35%. One postulation for the possible reduced benefi ts with direct inoculation is that echinacea works better on preventing the common cold caused by viruses other than rhinovirus. With over 200 viruses capable of causing the common cold, echinacea could have modest eff ect against rhinovirus but marked eff ects against other viruses. Of the direct rhinovirus inoculation trials, the most touted is the study by Turner and colleagues 14 published in 2005. The authors compared patients given E angustifolia equivalent to 900 mg/day with placebo and showed that echinacea did not have "clinically signifi cant eff ects on infection with a rhinovirus or on the clinical illness that results from it". The German Commission E has approved E purpurea at a recommended dose of 900 mg but has not approved E angustifolia. 8 The 1999 WHO monograph recommends E angustifolia at a dose of 3 g, a dose more than three times the dose used by Turner and colleagues. 38 As such, the dose used in this trial may have been too low to be fully eff ective. The previous meta-analysis done by Melchart and colleagues 13 and updated in November, 2005, 39 included 16 trials encompassing 22 analyses and showed a benefi t of echinacea for the treatment but not prevention of a common cold. 13, 39 By comparison, our meta-analysis included 14 trials encompassing 16 analyses. Although our results are in agreement with the previous meta-analysis, our results suggest an additional benefi t of echinacea for use in the prevention, as well as the treatment, of a cold. The meta-analysis by Linde and colleagues 39 assessed a cold severity endpoint, included two unpublished evaluations, and excluded studies that used experimental rhinovirus inoculation or that combined echinacea with other nutraceutical ingredients. In our analysis we chose not to evaluate cold severity because of concerns about the potential heterogeneity of the methods used for cold severity assessment in the studies. We included studies evaluating echinacea with other nutraceuticals in our analysis, as well as studies evaluating direct rhinovirus inoculation, which have the highest internal validity since the virus, the degree of exposure, and the exact time of exposure are all known. We decided to address the eff ect of these potential confounders through the use of subgroup and sensitivity analyses rather than through exclusion, which provides more information from which to make a determination of the effi cacy of echinacea. Furthermore, we included one study that was published after Melchart and colleagues updated their analysis 39 and excluded unpublished studies because data in such studies have not undergone rigorous peer review. There are several limitations to this meta-analysis that must be addressed. First, the studies by Barrett 20 and Turner 14 and their colleagues used alfalfa and a mixture of alcohol beverages, respectively, as their placebo arm. Since alfalfa and alcoholic concoctions may have immunostimulatory benefi ts, their use in the placebo arm is controversial. 40 Although we agree that this may be a potential confounder in our analyses, it should be noted that if these agents do in fact have benefi cial properties that reduce the incidence and/or duration of the common cold, then this would result in an underestimation of echinacea's benefi t. Second, although the Egger statistic shows absence of publication bias, our funnel plot shows asymmetry, suggesting that the potential for publication bias cannot be eliminated. Publication bias arises when trials with negative outcomes have a lower propensity to be published. Third, heterogeneity was present in our meta-analysis; however, the L'Abbé plot shows that the heterogeneity is a result of studies' disagreement in the magnitude, but not the direction, of echinacea's benefi t. Furthermore, after doing various subgroup analyses to assess the eff ect of clinical heterogeneity, echinacea maintained signifi cant eff ects on the reduction of cold incidence and duration. Finally, this analysis focuses on the effi cacy but not the safety of echinacea. Although adverse events with echinacea are not commonly reported, gastrointestinal upset and rash have been reported. 6 Much more work needs to be done to elucidate the safety of prolonged therapy since its eff ect on the rate-corrected QT interval, blood pressure, and other safety parameters is not well known. Of note, echinacea is a human cytochrome P450 3A4 enzyme inhibitor so the potential for drug interactions also needs to be assessed. 41 An analysis of the current evidence in the literature suggests that echinacea has a benefi t in decreasing the incidence and duration of the common cold; however, large-scale randomised prospective studies controlling for variables such as species, quality of preparation and dose of echinacea, method of cold induction, and objectivity of study endpoints evaluated are needed before echinacea for the prevention or treatment of the common cold can become standard practice. We declare that we have no confl icts of interest. These are described in detail in the Methods section on page 473. The complete German Commission E monographs: therapeutic guide to herbal medicines In: WHO monographs on selected medicinal plants. Geneva: World Health Organization Echinacea angustifolia: natural health products directorate Echinacea purpurea: natural health products directorate Echinacea pallida: natural health products directorate Echinacea for preventing and treating the common cold An evaluation of Echinacea angustifolia in experimental rhinovirus infections Echinacea root extract for the prevention of upper respiratory tract infections: A double-blind, placebo-controlled randomized trial Therapeutic experiences with Echinacea pallida in infl uenzal infections Assessing the quality of reports of randomized clinical trials: is blinding necessary? Much ado about nothing: a comparison of the performance of meta-analytical methods with rare events Cochrane handbook for systematic reviews of interventions 4.2.6 Treatment of the common cold with unrefi ned echinacea. A randomized, double-blind, placebo controlled trial Echinacea and truth in labeling Echinacea purpurea for prevention of experimental rhinovirus colds Effi cacy of Echinacea purpurea in patients with a common cold: a placebocontrolled, randomised, double-blind clinical trial A randomized controlled trial of the eff ect of fl uid extract of Echinacea purpurea on the incidence and severity of colds and respiratory infections Infl uence of echinacin (EC31) treatment on the exercise-induced immune response in athletes The ABC clinical guide to herbs Echinaguard treatment shortens the course of the common cold: a double-blind, placebo-controlled clinical trial Eff ectiveness of an herbal preparation containing echinacea, propolis and vitamin C in preventing respiratory tract infections in children: a randomized, double-blind, placebo-controlled, multicenter study The effi cacy of echinacea compound herbal tea preparation on the severity and duration of upper respiratory and fl u symptoms: a randomized, double-blind placebo-controlled study Effi cacy in the treatment of the common cold of a preparation containing an echinacea extract Ineff ectiveness of echinacea for prevention of experimental rhinovirus colds Effi cacy and safety of echinacea in treating upper respiratory tract infections in children: a randomized controlled trial Use of echinacea in upper respiratory tract infection Physicians' desk reference for herbal medicines Applied therapeutics: the clinical use of drugs Vitamin C supplementation and respiratory infections: a systematic review Review of the biological properties and toxicity of bee propolis Echinacea angustifolia in rhinovirus infections Echinacea for preventing and treating the common cold Echinacea for the common cold An in vitro evaluation of human cytochrome P450 3A4 inhibition by selected commercial herbal extracts and tinctures