The Journal of Informed Pharmacotherapy 2000;2:209-211.
Download a PDF version of this article
Reviewer: Kari L. Olson, Pharm D.
Reviewer's email address: email@example.com
Reviewer's profession/specialty: Research Fellow, EPICORE Centre and Division of Cardiology, University of Alberta
Stevinson C, Pittler MH, Ernst E. Garlic for treating hypercholesterolemia. A meta-analysis of randomized clinical trials. Ann Intern Med 2000;133:420-9. PubMed Cit
Two previous meta-analyses have been published on the effect of garlic on serum total cholesterol levels. The last meta-analysis was published in 1994 and several independent studies have conducted since this time. The objective of the current article was to conduct a an updated meta-analysis of all studies meeting the inclusion criteria and to thus reassess the effect of garlic (as compared to placebo) on total serum cholesterol in patients with elevated (i.e. greater than 5.17 mmol/L) levels. No efforts were made by the investigators to distinguish between studies that focused on patients with coronary heart disease (CHD) versus those without.
1. Did the overview address a focused clinical question?
Yes. The quality of the methods utilized by each study identified was assessed using a scale which quantified the likelihood of bias of the trial based upon the description of the randomization, blinding, and patient withdrawals. The article did not indicate the score (if any) that was needed to be attained in order for a study to be included in the analysis.
Yes. The meta-analysis used the chi-square test although the specific type of test used was not identified. This test for heterogeneity was significant indicating statistical heterogeneity amongst the trials. The heterogeneity was due to one study in particular. When study was excluded, homogeneity between the remaining 12 studies was found. No further tests for heterogeneity were undertaken.
1. What are the overall results of the review?
Thirty-nine studies were identified, however only 13 met the inclusion criteria for meta-analysis. There were 796 patients enrolled in the 13 studies, 10 trials reported benefits with garlic, while 3 found no difference as compared to placebo. Overall, the meta-analysis revealed that garlic significantly (p<0.01) reduced total cholesterol levels. Treatment with garlic resulted in a 5.8% reduction in total cholesterol levels from baseline. The authors did not report the reduction in cholesterol levels that occurred with placebo. The authors also evaluated the effect of garlic on total cholesterol levels using the 12 homogenous studies and found a smaller (4.3%) reduction in cholesterol levels with garlic. Gastrointestinal symptoms and garlic breath were the most frequently reported adverse effects. It is unclear if there were significant differences in the incidence of adverse effects between the garlic and placebo groups.
2. How precise were the results?Among the 13 studies included, treatment with garlic resulted in a 5.8% reduction in total cholesterol levels from baseline. The weighted mean difference was - 0.41 mmol/L (95% CI, -0.66 to -0.15 mmol/L). Among the 12 homogenous studies, the reduction in total cholesterol levels with garlic was 4.3%, equivalent to a mean difference of -0.30 mmol/L (95% CI, - 0.48 to - 0.11 mmol/L).
3. How much does allowance for uncertainty change the results?
Two sensitivity analyses were performed. The first analysis utilized 5 methodologically similar studies (i.e. same garlic preparation, same dose and also controlled for dietary factors). No significant difference in the reduction of total cholesterol with garlic (as compared to placebo) was found. The second analysis included 6 of diet-controlled studies and, once again, the authors found no significant difference between garlic and placebo in the reduction of cholesterol.
2. Were all clinically important outcomes considered?
Cardiovascular disease, of which CHD is the most common, is the leading cause of death in Canada. Elevated cholesterol is an independent risk factor for heart disease. Numerous studies have shown that the incidence of coronary heart disease and elevated cholesterol to be continuous and graded. Numerous large, randomized, controlled trials have shown that reducing cholesterol levels with HMG-CoA reductase inhibitors, in patients with and without CHD, and with elevated and "average" cholesterol levels reduces mortality and morbidity. In addition to knowing which patients derive the greatest benefit from therapy, the duration of therapy required to obtain benefits, the side effects from these medications are known, and for the most part, minimal. In order to be considered an "alternative" to standard therapies, studies evaluating other forms of lipid-lowering interventions, such as garlic, need to evaluate clinically relevant endpoints, as have the HMG-CoA reductase trials. In the absence of these studies, recommending these "alternative" interventions is difficult. Numerous unanswered questions exist from the current meta-analysis on the use of garlic in reducing total cholesterol levels. As previously mentioned, there are no data on the effect of garlic on the incidence of coronary heart disease, reductions in fatal or non-fatal myocardial infarctions, hospital admissions, or total mortality. Furthermore, it is unclear which types of patients derive the greatest benefit. The meta-analysis included patients with and without CHD and children. The form of garlic administered to patients enrolled in the studies varied from essential oil to standardized powder, or spray-dried powder. Whether there are differences in endpoints based upon the form of garlic used is unknown. Furthermore, the dose of garlic required is unknown. Patients inquiring about the use of garlic for reducing cholesterol should be informed of the minimal reduction in cholesterol levels observed in clinical trials and the aforementioned pitfalls of the current literature. Future randomized, controlled studies of garlic, should evaluate specific groups of patients, specific garlic preparations and doses, and evaluate clinically important endpoints.
Copyright © 2000 by the Journal of Informed Pharmacotherapy. All rights reserved.