Chapter  20:  Garlic: Effects on Cardiovascular Risks and Disease, Protective Effects Against Cancer, and Clinical Adverse Effects: Evidence Report/Technology Assessment Number 20

Objectives

This evidence report summarizes the effects of garlic on cardiovascular-related factors and disease, associations between garlic and cancer, and possible adverse effects of garlic.

Search Strategy

English and non-English citations were identified through February 2000 from 11 electronic databases, references of pertinent articles and reviews, manufacturers, and technical experts.

Selection Criteria

We limited review of cardiovascular-related effects to randomized controlled trials in humans that lasted at least 4 weeks and compared garlic with placebo, no garlic, or another active agent. Review of associations with cancer was limited to controlled studies that compared any precancerous or cancerous lesions in humans consuming varying amounts of garlic. All types of studies in humans were used to assess adverse clinical effects.

Data Collection and Analysis

Two physicians abstracted data from cardiovascular and cancer studies; one physician abstracted data about adverse effects. Lipid outcomes from cardiovascular trials were examined quantitatively using standardized and unstandardized mean differences (adjusted for baseline differences).

Main Results

• Thirty-seven randomized trials, all but one in adults, consistently showed that compared with placebo, various garlic preparations led to small, statistically significant reductions in total cholesterol at 1 month (range of average pooled reductions 1.2 to 17.3 milligrams per deciliter [mg/dL]) and 3 months (range of average pooled reductions 12.4 to 25.4 mg/dL). Eight trials with outcomes at 6 months showed no significant reductions of garlic compared with placebo. Changes in low-density lipoprotein (LDL) levels and triglycerides mirrored total cholesterol results; no significant changes in high-density lipoprotein (HDL) levels were found.

• Twenty-seven small, randomized, placebo-controlled trials, all but one in adults, reported mixed but never large effects of various garlic preparations on blood pressure outcomes.

• Twelve small, randomized trials suggested that various garlic preparations had no clinically significant effect on glucose in persons with or without diabetes. Two small short trials reported no statistically significant effects of garlic compared with placebo on serum insulin or C peptide levels.

• Ten small trials, all but one in adults and of short duration, showed the effects of various garlic preparations on platelet aggregation and mixed effects on plasma viscosity and fibrinolytic activity.

• There were insufficient data to confirm or refute garlic's effects on clinical outcomes such as myocardial infarction and claudication.

• Scant data, primarily from case-control studies, suggest, but do not prove, that dietary garlic consumption is associated with decreased odds of laryngeal, gastric, colorectal, and endometrial cancer and adenomatous colorectal polyps.

• Adverse effects of oral ingestion of garlic are "smelly" breath and body odor. Other possible, but not proven, adverse effects include flatulence, esophageal and abdominal pain, small intestinal obstruction, dermatitis, rhinitis, asthma, and bleeding.

Conclusions

Trials show several promising, modest, short-term effects of garlic supplementation on lipid and antithrombotic factors. Effects on clinical outcomes are not established, and effects on glucose and blood pressure are none to minimal. High dietary intake of garlic may be associated with decreased odds of multiple cancers. Ability to interpret existing data is substantively limited by marked variability in types of garlic preparations that have been studied and inadequate definition of active constituents in the various preparations.

This document is in the public domain and may be used and reprinted without permission except those copyrighted materials noted for which further reproduction is prohibited without the specific permission of copyright holders.

Suggested Citation:
Mulrow C, Lawrence V, Ackermann R, et al. Garlic: effects on cardiovascular risks and disease, protective effects against cancer, and clinical adverse effects. Evidence Report/Technology Assessment No. 20 (Contract 290-97-0012 to the San Antonio Evidence-based Practice Center based at The University of Texas Health Science Center at San Antonio and the Veterans Evidence-based Research, Dissemination, and Implementation Center, a Veterans Affairs Health Services Research and Development Center of Excellence). AHRQ Publication No. 01-E023. Rockville, MD: Agency for Healthcare Research and Quality. October 2000.

Overview

This evidence report is a systematic review that summarizes clinical studies of garlic in humans. It addresses three areas: (1) effects on cardiovascular-related disease and factors such as lipids, blood pressure, glucose, atherosclerosis, and thrombosis; (2) any protective associations with cancer; and (3) clinical adverse effects. The report was requested by the National Center for Complementary and Alternative Medicine, a component of the National Institutes of Health, and sponsored by the Agency for Healthcare Research and Quality. The following are the rationale for this report: (1) availability of multiple clinical studies with promising but conflicting results, and (2) high consumer usage of garlic as a health supplement. The report is intended primarily for agencies interested in funding clinical garlic studies, clinicians, and researchers, and secondarily for consumers.

Reporting the Evidence

The report addresses the following topics:

• Whether oral ingestion of garlic (fresh, cooked, or supplements) compared with no garlic, other oral supplements, or drugs lowers lipids, blood pressure, glucose, and cardiovascular morbidity and mortality.

• Whether garlic increases insulin sensitivity and antithrombotic activity.

• Associations between garlic and precancerous lesions, cancer, or cancer-related morbidity and mortality

• Types and frequency of adverse effects of oral, topical, and inhaled garlic dust.

• Interactions between garlic and commonly used medications.

Methodology

Findings

Limitations

Notable limitations in summarizing findings from garlic research include the substantial variability in types of garlic and garlic preparations that have been studied and an inadequate definition of the active, biologically available constituents in the various preparations. In addition, many trials that evaluated the effects of garlic on cardiovascular-related endpoints are limited by short durations; inadequate randomization and blinding procedures; lack of clear specification of contents of garlic preparations-including their constituents and dissolution properties; lack of intention-to-treat analyses; and incomplete reporting of data. The meta-analysis we performed is limited by some missing data at different time points and by the need to impute variability data from some trials.

We found few studies assessing associations between garlic consumption and cancer. Some pertinent studies may have been missed because they addressed associations with multiple foods and either did not report or analyze findings specific to garlic. Studies that were found sometimes failed to distinguish the type of garlic exposure (raw, cooked, or specific supplement), used subject recall to assess different frequencies of use over varying time periods, and adjusted for various potential confounders in different ways. Although we believe that we found most reported adverse-effect literature regarding garlic, adverse effects in general are frequently underreported or reported in ways that do not allow causality and frequency to be determined.

Future Research

Scope and Objectives

This evidence report was requested by the National Center for Complementary and Alternative Medicine, a component of the National Institutes of Health. It was contracted by the Agency for Healthcare Research and Quality, part of the U.S. Department of Health and Human Services. The following are the rationale for this report addressing garlic: (1) availability of multiple clinical studies with promising but conflicting results and (2) high consumer usage of garlic as a health supplement. This report is intended primarily for agencies interested in funding clinical garlic studies, clinicians, and researchers, and secondarily for consumers. This chapter highlights the rich history of garlic, the complexity of its chemistry, recent research, the variety of available commercial preparations, and challenges in conducting garlic research in humans. The evidence report is a systematic review that summarizes the studies in humans that address the following three areas:

• Effects of garlic on cardiovascular-related factors and disease. Results of randomized trials lasting at least 4 weeks comparing garlic preparations with placebo or other agents are presented. Effects on the following outcomes are addressed: clinical cardiovascular disease, lipids, atherosclerosis, blood pressure, glucose, insulin sensitivity, and antithrombotic activity.

• Associations of garlic in preventing cancer. Results of case-control and cohort studies that compare the occurrence of various precancerous conditions and cancer in persons consuming or avoiding garlic are described.

• Clinical adverse effects of garlic ingestion or contact. Various reported adverse effects, including dermatological, gastrointestinal, and hematological, are summarized.

Origins and Folklore of Garlic

Garlic, scientifically known as Allium sativum L., is closely related to other "smelly" bulbs such as onions, leeks, and chives.¹ There is an old story originally told by the ancient Islamic religious leader Mahomet that garlic sprang from the left footprint, and onion from the right, of Satan as he walked out of the Garden of Eden.² Another story, translated from the 5 ^th century AD Bower Manuscript , says that garlic grew from the drops of blood that fell to the earth when Janar-Dana decapitated Asura, the ruler of the Old-Indian gods.³ Historians believe that garlic was native to Siberia and was spread to other geographic regions more than 5,000 years ago by Siberian nomadic tribes.^{4 5}

Throughout recorded history, garlic has played rich and diverse commercial, culinary, literary, and mythical roles. Ancient Egyptian and Middle Eastern cultures used garlic as a currency. Around 2,500 BC, 15 pounds of garlic could purchase a healthy slave.⁶ The culinary delights of garlic over manna were extolled by followers of Moses in the Old Testament (Numbers 11:5,6). Virgil, in his Second Idyll , described how Thestylis used the juices of wild thyme and garlic as a prophylactic against snakebites, while Homer wrote that garlic helped Ulysses escape from being changed by Circe into a pig like his companions.^{3 4}

Balkans rubbed garlic on doorknobs and window frames to discourage vampires from haunting them, and they hung garlic on doors and windows so that nobody would take milk from their cows and the families would be safe from witches.^{3 6} In other parts of Europe, fisherman and seamen wore cloves of garlic to protect themselves from evil spirits and diseases, and mothers hung garlic on children to protect them from the demons of sickness.³ Other religions and cultures shunned garlic and believed it to be evil. Sects of Christianity, Hinduism, Islam, and Zen Buddhism have considered garlic "unclean."^{2 7} Nepalese shamans of the Brahan and Cherti castes were forbidden to eat garlic because it could cause suffering in the afterworld.² In ancient Greece, people who had eaten garlic were forbidden from entering the Temple of Cybele because its smell was considered offensive.²

Chemistry of Garlic

Figure 1. Chemical profile of garlic

   Figure 1. Chemical profile of garlic

Figure 1

Crushing or cutting garlic results in a comingling of precursor compounds alliin and methiin, found in the cytosol, with robust, abundant enzymes known as alliinases, which are concentrated in stem cells. The result is cleavage of alliin and methiin to give 2-propenesulfenic acid and methanesulfenic acid, respectively. Sulfenic acids are highly reactive intermediates, rapidly condensing to give compounds known as thiosulfinates ( RS(O)SR'), which are the actual primary flavor compounds. Four sulfenic acids (RSOH, R = methyl, allyl, 1-propenyl, and propyl) are formed when garlic is cut. These acids combine in all possible permutations to give thiosulfinates, whose specific ratio is responsible for the unique flavor of garlic. The major thiosulfinate from garlic is allicin, CH ₂ =CHCH ₂ S(O)SCH ₂ CH=CH ₂ (RS(O)SR, where R = allyl), formed in less than 10 seconds when alliinase acts on alliin. Thiosulfinates are quite reactive and undergo a variety of further transformations, depending on conditions. Thus, in organic solvents or oils, allicin decomposes into 2-propenesulfenic acid and thioacrolein; the latter compound rapidly undergoes self-condensation giving isomeric dithiins. In a second, more complex process, allicin can form ajoene. This reaction occurs in organic solvents or food oils, but only at trace levels in water. Thiosulfinates react with heated water to give dialk(en)yl polysulfides such as diallyl polysulfides and allyl methyl polysulfides. These polysulfides have a characteristic odor, which is similar to allicin. The above transformations of allicin and other thiosulfinates can occur in the laboratory as well as when cooking with garlic. Frying garlic, which involves temperatures higher than 100 degrees Celsius (°C), denatures alliinases and converts alliin to cysteine and allyl alcohol. Allicin is moderately stable in aqueous solutions, with a half-life of about 30 days at room temperature and about 6 months at 4° C.³ Alliinases also are deactivated at low pH. Their activity is maximal at pH between 5 and 8 and rapidly diminishes at higher or lower pH. Under conditions when allicin and cysteine or allicin and glutathione (the main intracellular thiol of mammalian cells) react, S -allylmercaptocysteine or S -allylmercaptoglutathione is formed, respectively.^{3 11 12}

When fresh garlic is ingested, 2-propenethiol, an odorous compound even at very low concentrations, can be detected on the breath. This compound was identified in human garlic breath by collecting and analyzing breath samples immediately after ingestion of sliced fresh garlic. Within a few minutes, the 2-propenethiol disappears and is replaced by allyl methyl sulfide, the methylated metabolite of 2-propenethiol. Acetone also is detected in human breath after garlic ingestion and may parallel increased lipid metabolism. Allicin, ajoene, dithiins, and certain allyl sulfides have antithrombotic activity. Allicin, ajoene, and allyl sulfides also may have antilipidemic properties, while allyl sulfides have anticancer activity. Organoselenium compounds may follow a similar metabolic pathway to that of sulfur. Thus, γ-glutamyl- Se -methylselenocysteine is presumably hydrolyzed in the gut to Se -methylselenocysteine, which is enzymatically cleaved to methaneselenol, and then methylated to dimethyl selenide, another compound detected in human garlic breath. Dehydrated tablets manufactured from fresh garlic also contain alliin, γ-glutamylcysteine derivatives, alliinases, and all of the compounds found naturally in garlic. When garlic cloves are consumed, N -acetyl- S -allylcysteine is excreted in the urine.

Historical Uses of Garlic

Arthritis, asthma, toothaches, freckles, baldness, athlete's foot, plague, cancer, and cardiovascular disease are just a few of the many maladies that people throughout history have treated with garlic. The teachings of the father of Ayurvedic medicine, Charak (around 3,000 BC), described early uses of garlic to maintain the fluidity of blood and strengthen the heart.¹³ Similar uses were reiterated in the Codex Ebers , an Egyptian medical manuscript dating to approximately 3,500 years ago.^{7 13 14} In the Talmud , an important collection of ancient rabbinic writings, garlic is recommended for treating wounds.^{3 14} Pliny the Elder, the author of what is believed to be the first encyclopedia, Historia Naturalis, wrote that garlic could cure 62 ailments, including consumption, hemorrhoids, dog and snake bites, and tumors.⁷ Hippocrates recommended garlic as a laxative and diuretic.^{3 7} In the mid-1500s, Adam Lonitzer, a German physician author of an herbal reference book, recommended garlic juice externally for killing lice and nits and internally to rid the body of worms and poison.³ By the 1800s, Louis Pasteur had reported that garlic kills bacteria,¹⁴ and in World War II, garlic juice with water was used to disinfect wounds.^{15 16} During the last half of the 20 ^th century, investigations relating to the medicinal qualities of garlic exploded. The next two sections will briefly describe some of the research relating to cardiovascular factors and cancer.

Garlic Research Addressing Cardiovascular-Related Factors and Disease

Animal studies in rats and rabbits show multiple effects of garlic on cardiovascular-related factors, including antithrombotic effects, regression of atheromata, and decreases in total cholesterol, glucose, and blood pressure.^{17 18 19 20 21 22} Although some observational studies and small trials in humans corroborate animal studies,^{20 23 24 25 26 27 28 29} data are sometimes confusing and conflicting. For example, some studies in humans show that short-term administration of garlic for less than 4 months has either no lipid-lowering or transient detrimental effects,^{30 31 32 33} while others suggest that garlic has short-term beneficial effects.^{34 35 36 37 38 39 40} A possible reason for the conflicting data is the variable release of allicin or allicin-derived compounds from the dosage forms used in the studies.⁴¹

Several reviews that were published in the early 1990s summarized randomized controlled trial (RCT) data about garlic and cardiovascular factors.^{13 42 43 44 45 46 47} Most focused on the effects of garlic on lipids, although one summarized the effects on blood pressure.⁴⁶ The reviews included different studies, gave varying attention to specific garlic preparations and doses, and sometimes reached different conclusions. Moreover, multiple new trials are available, and some focus on endpoints, such as antithrombotic activity and progression of atherosclerotic plaques, that were not addressed in prior reviews.

Garlic Research Addressing Cancer

Most, but not all, in vivo and in vitro studies have shown that garlic constituents inhibit tumor cell metabolism, inhibit tumor initiation and promotion, and/or modulate immune responses.^{48 49 50 51} For example, in vitro studies have suggested that constituents such as allicin and ajoene have antimitotic and antimutagenic properties, while other studies have suggested that constituents such as alliin do not inhibit mutagenesis.⁴⁹ Some in vivo studies have shown an inhibition of tumor growth in rats and mice with fresh garlic but not with garlic in which alliinase had been inactivated.^{48 50} Extrapolation of effective doses from in vivo experiments to potentially equivalent doses in humans has been estimated at 25 to 450 mg of garlic cloves per kilogram (kg) of body weight or 2 to 32 g per 70-kg person.^{52 53 54} Recent epidemiological studies in humans have examined whether garlic consumption is related to incidence of breast, colorectal, gastric, lung, and prostatic cancer.^{55 56 57 58 59 60 61} These studies have reported intriguing results; a detailed review of their design, method of ascertaining cancer, and measurement of garlic intake is warranted.

Garlic Preparations

Table 1. Historical examples of different preparations (more...)

Table 1. Historical examples of different preparations of garlic

Garlic enemas in ancient Egypt to promote cleanliness.3 Slivers of garlic in mouth or behind ear for toothaches.6 Wine of garlic or garlic juice to stimulate hair growth in baldness.4 Pounded garlic cloves with honey to treat sore joints and muscles.15 Garlic syrup to cure hoarseness and breathing ailments.1 Garlic showers in Japanese sanatoriums for coughs.3 Bruised garlic mixed with lard and rubbed on chest or placed in shoes for whooping cough.4 6 Minced garlic applied to soles of feet for smallpox or to baby's navel for colds.4 6 Garlic juice diluted with water to treat wounds in World War II.4 15 63

Summary of Challenges in Conducting and Interpreting Garlic Research in Humans

Table 2. Summary points about garlic

Table 2. Summary points about garlic

Garlic has a rich cultural history that may influence current beliefs about appropriate uses and preparations of garlic. Different whole garlic bulbs contain variable concentrations of inert constituents based on strain and growth condition. Crushing or cutting the whole garlic bulb liberates multiple substances with variable half-lives. These substances vary based on initial substrate concentrations and storage conditions and alter with time. End products of processed garlic depend on above as well as the mode of preparation. Major constituents of a particular preparation may contain compounds with concentrations significantly less than that found in whole garlic or new substances entirely absent in the undisturbed bulb. Gastric pH and other dietary factors may affect further the liberation or destruction of compounds. An individual's variation in absorptive potential, first-pass metabolism, and coexistent disease processes may alter true biologic effectiveness of garlic.

Second, true active ingredients and mechanisms of action of garlic constituents are varied and sometimes unclear. For example, whether allicin is truly an active ingredient has been questioned because it has a very transient half-life and limited or no bioavailability that may be due to a significant hepatic first-pass metabolism, although metabolites of allicin do have substantial activity. Proposed mechanisms of action of various garlic constituents include, but are not limited to, the following: decrease in dietary fat absorption, decrease in hepatic cholesterol synthesis, alteration in lipoprotein composition, alteration in total body lipid composition, vasorelaxation, antioxidation, and increased tissue insulin sensitivity.

Third, garlic constituents have varying bioavailability. Whether the in vitro activity of garlic constituents correlates with true physiological effects in humans is difficult to assess. Consuming identical preparations of garlic may result in varying bioavailability depending upon several factors. For example, particular foods that are consumed with garlic may impair absorption, alter gastric pH, or react directly with key constituents of the garlic product. Some substances, including medications, may alter hepatic first-pass metabolism. Particular underlying conditions in individuals such as dyslipidemia, atherosclerosis, or genetic predispositions to cancer also may affect biological efficacy.

Fourth, there is substantial variability between garlic compounds, based on where and how the whole garlic is grown, as well as how the end preparation is prepared and stored. Fresh garlics from around the world vary in amounts of compounds, based on climate and soil conditions. Some purported active ingredients, such as SACs, are not present above trace levels in fresh garlic but are instead produced by aging chopped garlic in dilute alcohol. Other organosulfur compounds are produced through maceration in the presence of edible oils. Another source of variability is the stability of garlic compounds in the different stages of storage between harvest and consumption.

Fifth, quantitative release of allicin (a possible main active compound) from dehydrated preparations needs verification under simulated gastrointestinal conditions or by an in vivo method. This may be critical because allicin release depends on alliinase activity, which can be greatly decreased by improper powder or tablet manufacturing, gastric acid, and neutral pH that is typical of the intestinal tract. Tablets that dissolve quickly after entering the intestinal tract have the greatest amount of alliinase activity. Notably, dissolution properties of preparations are not routinely reported in the trials, but some Kwai ^® preparations that were used in early trials have been shown to release three times as much allicin as the Kwai ^® tablets that were used in the more recent trials.⁴¹

Sixth, studies of garlic intake are somewhat unique regarding the ability to successfully conduct "blinded" studies. One of the characteristics of the "stinking rose" that separates garlic from almost any other dietary component is its strong and unique taste, as well as the related issues of associated body and breath odors. Whether garlic study participants can detect the garlic or placebo group to which they have been assigned raises at least two concerns. The first is the classical concern that provides the rationale for "blinding" study participants to which group they are assigned. When there is a perceived benefit of one assignment over another, research participants might alter their behavior (e.g., diet or exercise) if they think that they have been assigned to the garlic group. If this is the case, it becomes difficult to determine to what extent study results are attributable to the garlic versus the associated behavior changes. A second concern is perhaps less intuitive and more hypothetical. It is possible that the putative health benefits of ingesting garlic are attributable to the presence, concentration, and form of its many sulfur-containing compounds. The sulfur compounds generate the unique taste and odor of garlic. Garlic studies that successfully blind participants to group assignments may be using garlic products that have undetectable taste and body and breath odor as well as low levels of active sulfur compounds present. If benefits were tied to sulfur compounds, such studies would underestimate them.

Seventh, even sustained beneficial effects on intermediate physiological factors such as lipids and blood pressure do not necessarily portend clinically important benefits. At present, the research evidence from human studies in the cardiovascular area focuses almost exclusively on intermediate physiological factors rather than clinical outcomes. The very multiplicity of potential physiological effects of garlic underscores the need for studies with broadly defined clinically important outcomes.

Expert Input

We owe a major debt of gratitude to the following groups of multidisciplinary experts from around the world who assisted in preparing this report: 10 national advisory panel members and 5 technical experts who helped define the scope and shape the content, 20 peer reviewers representing a variety of backgrounds and viewpoints, 4 scientific authors who provided additional data from their studies, and 15 staff members of the San Antonio Evidence-based Practice Center and the San Antonio Veterans Evidence-based Research, Dissemination, and Implementation Center, a Veterans Affairs Health Service Research and Development Center of Excellence. Their names are listed in Appendix B. Acknowledgments.

Questions Addressed in Evidence Report

Table 3. Questions and selection criteria grouped by (more...)

Table 3. Questions and selection criteria grouped by search strategies

Questions about cardiovascular risk factors and disease	Selection criteria
1 In adults or children with or without dyslipidemia, does oral ingestion of garlic (fresh, cooked, or supplements) compared with no garlic, other oral supplements, or drugs lower plasma lipid levels? 2 In adults or children with or without hypertension, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs lower blood pressure? 3 In adults or children with or without diabetes, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs increase insulin sensitivity? 4 In adults or children with or without diabetes, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs lower plasma glucose levels or glycosylated hemoglobin? 5 In adults or children, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs increase antithrombotic activity?	Study type: Randomized controlled trials or systematic reviews of randomized controlled trials greater than 4 weeks duration. Participants: Humans. Intervention group: Fresh, cooked, or supplement of garlic. Control group: Placebo, usual care, or other active agent. Outcomes (physiological): Lipids, blood pressure, insulin sensitivity, glucose or glycosylated hemoglobin, or antithrombotic activity.
Questions about cancer	Selection Criteria
6 In adults at average or high risk for cardiovascular disease, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs decrease cardiovascular morbidity or mortality? 7 Do different preparations of garlic vary in effectiveness regarding any of the above outcomes?	Outcomes (clinical): Cardiovascular morbidity or mortality such as stroke, myocardial infarction, angina incidence, or severity; peripheral vascular disease; or numbers of cardiac procedures.
Questions about cancer	Selection Criteria
1 In adults without cancer risk factors, precancerous conditions, or malignancy, does oral ingestion of garlic compared with no garlic, placebo, or other oral supplements reduce the risk of developing cancer? 2 In adults with cancer risk factors but no precancerous conditions or malignancy, does oral ingestion of garlic compared with no garlic, placebo, or other oral supplements reduce the risk of developing cancer? 3 In adults, does oral ingestion of garlic compared with no garlic, placebo, or other oral supplements reduce the risk of developing precancerous lesions? 4 In adults with cancer, does oral ingestion of garlic compared with no garlic, placebo, or other oral supplements reduce morbidity or mortality of cancer? 5 Is garlic more effective against some types of risk factors, precancerous conditions, or cancer than others?	Study type: Cohort or case-control studies with greater than 50 participants (also randomized controlled trials eligible for Question No. 4). Participants: Humans. Case group: Participants with cancer or precancerous lesions. Control group: Participants without cancer. Exposure: Fresh or prepared garlic in diet or as supplement. Outcomes (clinical): Precancerous lesions, cancer, morbidity, and mortality.
Questions about adverse effects	Selection criteria
1 What are the symptomatic and serious adverse effects of various garlic preparations, and what is their frequency? 2 Are there interactions between garlic and commonly used medications for dyslipidemia, diabetes, or thrombogenic disease (e.g., statins, sulfonylureas, or antithrombotic agents)?	Study type: Case reports, case series, cohort, or surveillance studies or randomized controlled trials. Participants: Humans exposed to garlic. Control group: Not required. Outcomes: Any reported adverse effect.

Literature Search and Selection Methods

Sources and Search Methods

Table 4. Electronic sources searched

Table 4. Electronic sources searched

Electronic database	Description
AMED (Alternative and Allied Medicine Database) Searched from 1985 to October 1998	This database contains 100,000 references from 400 journals on alternative and complementary medicine going back to 1985.
CINAHL (Cumulative Index to Nursing and Allied Health Literature) Searched from 1984 to July 1999	This database includes citations from more than 500 biomedical and popular sources, including National League of Nursing and American Nurses Association publications, covers publications from 1982 to the present, and is considered the premier nursing database.
CISCOM (Centralised Information Service for Complementary Medicine) Searched 1968 to July 1999	This database contains more than 34,000 references and combines data from MEDLINE, AMED, and other specialist European databases.
Cochrane Library (http://www.cochrane.org) DARE (Database of Reviews of Effectiveness) (http://www.york.ac.uk/inst/crd/) The Cochrane Controlled Trials Registry Searched Issue 2, 1999	These databases contain references of randomized controlled trials and systematic reviews identified from electronic bibliographic sources and hand-searching of multiple journals and symposia or meeting proceedings.
Dissertation Abstracts Searched from 1961 to Dec 1998	This library indexes doctoral dissertations and masters' abstracts from more than 1,000 institutions.
EMBASE Searched from 1988 to July 1999	This database contains biomedical and pharmaceutical citations and is considered the premier biomedical database in Europe.
MEDLINE (PubMed) Searched from 1966 to July 1999PubMed searched July 1999 to February 2000	These databases (MEDLINE and PubMed) index almost 4,000 international biomedical journals from 1966 to the present, include references from Index Medicus, International Nursing Index, and Index to Dental Literature, and are considered the premiere biomedical databases in the United States.
MICROMEDEX contains: DRUGDEX Product Index (drug ingredients) POISONDEX/IDENTIDEX (toxicology) DRUG-REAX (interactive drug interactions) Searched July 1999	This database provides a major source for drug, poison, and acute care information.
NAPRALERT (Natural Products Alert) Searched 1650 to September 1999	This database contains records from 1650 to the present on natural products, including the pharmacology, biological activity, taxonomic distribution, ethnomedicine, and chemistry of plant, microbial, and animal extracts.
PHYTODOK (German database) Searched 1995 to July 1999	This database contains 8,800 references from approximately 300 journals worldwide on toxicology, pharmacology, and therapeutic uses for natural compounds and on isolation of natural compounds from plant material.
Science Citation Index Searched from January 1990 to March 1999	This index covers 4,400 scientific and 1,400 social science journals worldwide, together with selected coverage of related material.

English and non-English citations were identified through February 2000 from 11 electronic databases (Table 4); references of pertinent articles and reviews; manufacturers; and technical experts. An update search that was limited to PubMed was conducted in February 2000. Database searching used maximally sensitive strategies that identified all papers on garlic and on herbal treatments for cardiovascular disease or cancer. The International Bibliographic Information on Dietary Supplements (IBIDS) database was not searched because it did not contain information on botanicals.

Titles, abstracts, and keyword lists of the 11 electronic databases listed in Table 4 were searched using the following terms, which include Latin names for garlic and names of garlic extracts and constituents (".tw." indicates text word searches, "/" indicates keyword searches, and "$" indicates truncated words that facilitate searching all formations and suffixes of particular terms):

Sources and Search Methods

2-propenesulfenic acid.tw.	alliinase.tw.	dipropyl disul$ide.tw.
aglio.tw.	diallyl sulfide.tw.	dipropyl sul$ide.tw.
ajo.tw.	allium sativum.tw.	garlic extract/
ajoene.tw.	allyl mercaptan.tw.	garlic oil/
alisat.tw.	allyl mercaptan.tw.	garlic.tw.
allicin.tw.	diallyl disulphide.tw.	garlic/
kwai.tw.	diallyl sulfide.tw.	knoblauch$.
kyolic.tw.	diallyl sulphide.tw.	thiosulfinates.tw.
s-allyl cysteine.tw.	s-allylcysteine.tw.	vinyl-dithiin$.tw.
s-allyl$cysteine.tw.	thioallyl derivative$.tw.	vinyl$dithiin$.tw.
		vinyldithiin$.tw.

Selection Processes

Figure 2. Flow diagram of selection process

   Figure 2. Flow diagram of selection process

At least two independent reviewers scanned the titles and abstracts of all records identified from the search, using selection criteria given in Table 3. Selection criteria that were specified for each formulated question included the types of participants, interventions, control groups, outcomes, and study designs that were deemed appropriate. Cardiovascular-related trials were arbitrarily limited to those that were at least 4 weeks in duration, because the national advisory panel thought that several weeks of garlic administration might be necessary to demonstrate effects on factors such as glucose, blood pressure, and lipids. Figure 2 schematically presents the selection process. Of 1,798 records, reviewers excluded 1,480 with certainty when screening titles and abstracts. Most of these were in vitro studies, involved animals, or did not meet design inclusion criteria. When screening the full text of the remaining 317 records, 178 were excluded. Of the 139 records meeting selection criteria, 49 were records of 45 randomized trials (some trials were reported in multiple publications or records), 17 were records of 13 cancer survey, case-control, or cohort studies, and 73 were reports of adverse effects or results of skin patch tests with garlic. Of note, of the 45 randomized trials meeting eligibility requirements, 34 were cited in MEDLINE, 2 additional trials were identified from EMBASE only, 1 was identified from PHYTODOK, and 8 were identified either by experts or from scanning symposium proceedings.

Data Abstraction Process

Two independent physicians abstracted data from trials that were identified in the efficacy searches. They were not blinded either to study title or to author names. Items that were related to the quality of assessed studies included adequacy of randomization (method and concealment of assignment); whether the trial was single or double blind; whether the intervention and control groups were adequately matched to maintain blinding; cointerventions such as diet, exercise, and cardiovascular medications; and the number of dropouts. Disagreements in abstractions were uncommon (less than 1 percent of items) and were resolved by consensus. No formal reliability testing was done. All abstracted outcome data were verified by a third person with expertise in quantitative data. Abstractions were filed electronically to enable easy updating.

One physician abstracted data about adverse effects. Items that were abstracted included study design (case report, case series, case control, cohort, and controlled trial) and type of specific adverse effect. Several explicit criteria that were aimed at assessing drug adverse effect causality were assessed, such as appropriate temporal relationship, lack of apparent alternative causes, known toxic concentrations of the drug at the time of the appearance of the symptom, disappearance of the symptom with drug discontinuation, dose-response relationship, and reappearance of the symptom if the drug was readministered.

Unpublished Data

We found two randomized trials that were only published in abstract form.^{46 64} We obtained the full report of one of these trials.⁶⁴ The one for which we could not obtain a full report is not included in this review.⁶⁵ It was a crossover trial with 16 participants that compared standardized dehydrated garlic (Kwai^® ) with placebo.⁴⁶ No data prior to the crossover were given in the abstract. Several published studies met selection criteria but did not report critical design features or outcome data. Authors were contacted and requested to provide information regarding randomization procedures and lipid outcomes. Three of the 11 requested authors provided unpublished raw data for lipid outcomes.

Data Analysis Process

Data were synthesized descriptively, emphasizing methodological characteristics of the studies such as populations enrolled, definitions of selection and outcome criteria, sample sizes, adequacy of randomization process, interventions and comparisons, cointerventions, biases in outcome assessment or intervention administration, and study designs. Relationships among clinical outcomes, participant characteristics, and methodological characteristics were examined in evidence tables and graphical summaries such as forest plots.

Primary outcomes in studies were measured with continuous rather than categorical variables. Two methods were used to estimate "effect size" measures for each study. First, we used the standardized mean differences between treatment and comparison group scores. Hedges' g was used to compute the standardized mean difference for each trial:

where, for a given trial are the mean clinical outcome scores for the treatment group and comparison group, respectively, and s_pooled is the pooled standard deviation for the difference between the two means.⁶⁶ These estimates were adjusted for between-group differences at baseline and for small sample bias.⁶⁶ Adjusting for baseline differences was accomplished by calculating an effect size at baseline; by definition, it should be zero if study groups were well matched. When a nonzero effect size at baseline was found, outcome effect sizes were adjusted by subtracting the baseline effect size. Second, we used the unstandardized mean differences between treatment and comparison group scores and then adjusted for baseline differences using meta-regression models.

Published reports seldom provided estimates of s_pooled. One of three strategies was used to estimate s_pooled when the authors did not directly provide it. First, the individual group variances were used to estimate s_pooled . If these data were not reported, the pooled variance was back-calculated from either the test statistic or the p-value for differences at followup.⁶⁷ If neither was possible, a mean variance that was derived from studies of similar size was used. Studies in which the pooled variance was calculated using either of the two latter methods were flagged in the event the magnitude of the effect size resulted in the study being identified as a potential outlier in analyzing heterogeneity.

Placebo-controlled randomized trials with lipid outcomes were quantitatively pooled using a random effects estimator.^{66 67} We tried to identify outliers using a standard heterogeneity chi-square test, funnel plot, and Galbraith plot. Studies were considered outliers if the probability for the chi-square value was less than 0.1 and/or the study fell outside of the funnel or Galbraith plot. (A Galbraith plot is a graphical method used to aid in assessing heterogeneity and is particularly useful when the number of studies is small.⁶⁸ The position of each study along the two axes indicates the weight allocated in the meta-analysis. The vertical axis [a Z statistic equal to the effect size divided by its standard error] gives the contribution of each study to the Q [heterogeneity] statistic. Points outside the confidence bounds are those studies that have major contributions to heterogeneity; in the absence of heterogeneity, all points would be expected to be within the confidence bounds.)

Standardized mean difference effect sizes were converted to clinical laboratory units to aid in interpreting effect size standard deviation units. As noted above, the effect size statistic is calculated by dividing the difference between group means by the pooled standard deviation of the two groups. Because both numerator and denominator are expressed in original units (e.g., milligrams per deciliter [mg/dL]), the units cancel out and the effect size is "unitless." Effect sizes can be back-converted to a value with the original unit, for example, mg/dL, by multiplying the effect size value by a standard deviation value. The statistical significance of the values (effect sizes or converted values) do not change. However, the magnitude of the "converted effect" will vary up or down depending on the magnitude of the standard deviation used. Because these converted clinical units are based on a common standard deviation across all studies, individual study values do not always agree with author-reported results.

Lacking population standard deviation values, we chose to use the "average" standard deviation value for the pooled studies within each group for conversions. Two "averages" were examined: a weighted pooled standard deviation across studies (weighted by sample size) and the median pooled standard deviation. When the two values were substantially different (representing skewness), the median value was chosen. When the values were similar, the weighted pooled standard deviation value was used. The actual weighted average standard deviations that were used in conversions of lipid analyses were total cholesterol: 40.8 mg/dL; low-density lipoprotein level (LDL): 29.1 mg/dL; high-density lipoprotein level (HDL): 11.4 mg/dL; and triglycerides: 85.9 mg/dL.

The following are the rationale for pooling lipid studies: (1) multiple small- to moderate-size studies were available; (2) similar control groups were used; (3) lipid outcomes were measured using similar parameters (e.g., serum total cholesterol) at similar followup times (i.e., 4 to 6 weeks, 8 to 12 weeks, and 20 to 24 weeks); and (4) actual numeric results were often reported. No quantitative summary analysis of blood pressure, glucose, or thrombotic outcomes was performed. Although several studies measured blood pressure, few had a priori hypotheses about blood pressure, and numeric results commonly were not reported, raising the possibility of publication bias in the studies that did report numeric outcomes. Few studies reported glucose or thrombotic outcomes; those that did reported these in several different manners (e.g., fibrinolytic activity and platelet adhesiveness).

Sensitivity and Subgroup Analysis

Different preparations of garlic, including combination preparations, were used in the trials. Subgroup analyses were conducted for trials that used similar dried standardized preparations of garlic and enrolled participants with hypercholesterolemia. Analyses with and without the studies that evaluated the combination preparation of garlic and ginkgo or garlic and hawthorn compared with placebo were conducted because ginkgo is not known to affect lipid parameters. The study that evaluated a garlic and fish oil combination was not pooled with other studies because of possible independent effects of fish oil on lipid parameters. Subgroup analysis based on "doses" of garlic supplements was not conducted because of limited variability of dosing among trials.

Presentation of Results

For this report, results of the standardized mean difference analyses are presented with overall results converted back to original "mg/dL" units; the results and conclusions of the standardized mean difference analyses do not differ substantially from the unstandardized mean difference analyses. Results for total cholesterol, HDL, LDL, and triglyceride levels are all presented in mg/dL. To convert cholesterol, LDL, and HDL values from mg/dL to millimoles per deciliter (mmol/dL), divide by 38.7. To convert triglyceride values from mg/dL to mmol/dL, divide by 88.2. Results were presented in clinical laboratory units to aid interpretation. The magnitude of values depends entirely on the value of the standard deviation. For this report, we chose to use a weighted average standard deviation across all studies at baseline (see the preceding "Data Analysis Process" section). Different values for the common standard deviation would change the magnitude of clinical laboratory unit results (would not change the magnitude of the effect size units), but not the statistical significance.

Results: Cardiovascular Risks and Disease

Overview of the Evidence

There were 45 RCTs lasting at least 4 weeks that tested the effects of garlic on cardiovascular-related endpoints (Evidence Table 1). All but three RCTs reported intermediate physiological outcomes rather than clinical outcomes such as cardiovascular events or mortality.^{69 70 71} Reported trial protocols included lipid measurements (n=44), blood pressure measurements (n=30), serum glucose concentration (n=12), antithrombotic activity (n=10), walking distance (n=2), myocardial infarctions and deaths (n=1), and regression of atherosclerotic plaques (n=1). Placebo comparisons were used in 39 trials, a "no garlic" comparison in 2, an active antilipidemic or antihypertensive agent in 3, and a head-to-head comparison of two garlic preparations in 1.

Trials were conducted in Germany (n=13), North America (n=14), India (n=5), United Kingdom (n=4), Thailand (n=2), Poland (n=2), Switzerland (n=1), Italy (n=1), Europe (n=1), and Australia (n=2). Industry sponsorship and free provision of commercially developed garlic products were stated in 35 trials and were unclear in 3. Seven trials stated no commercial sponsorship.^{34 35 40 71 72 73 74} Four trials were published in symposia or meeting proceedings,^{39 71 75 76} one in a book,⁷⁷ one as a thesis, ⁷⁸ and the remainder in journals. We did not review the peer-review processes that were used by the journals.

Table 5. Characteristics of trials that evaluated standardized (more...)

Table 5. Characteristics of trials that evaluated standardized dehydrated garlic powder preparations

Study	Country of Study	Size (n=)	Mean Age	Male %	Participant Characteristics	Inclusion Criteria	Recruitment Setting	Garlic Preparation	Daily Dosage
Adler (1997)79	Canada	23	46	100	No diabetics or cardiac disease No lipid or antihypertensive drug	TC > 200 mg/dL	Unclear	Dehydrated tablet: Kwai®	900 mg
Auer (1990)39	Germany	47	58	45	45% with TC > 250 8% Diabetic	DBP 95 to 104 mm Hg	11 General practices	Dehydrated tablet: Kwai®	600 mg
Holzgartner (1992)85	Germany	98	57	39	40% Smokers 35% Hypertensive 2% Diabetic	TC or TG > 250 mg/dL Hyperlipoproteinemia types IIa, IIb, or IV	5 General practices	Dehydrated tablet: Kwai®	900 mg
Isaacsohn (1998)86	United States	50	58	54	2% Coronary disease 2% Smokers 22% Hypertensive	LDL > 160 mg/dL and TG < 350 mg/dL	Specialty clinic	Dehydrated tablet: Kwai®	900 mg
Jain (1993)24	United States	42	52	45	No diabetics 26% Smokers	TC > 220 mg/dL	Community volunteers	Dehydrated tablet: Kwai®	900 mg
Kandziora (1988)37	Germany	40	Range 43-65	Not given	Base: TC 280 mg/dL Base: TG 209 mg/dL No diabetics	DBP 95 to 104 mm Hg	Unclear	Dehydrated tablet: Kwai®	600 mg
Kandziora (1988)38	Germany	40	56	83	Base: TC 292 mg/dL Base: TG 207 mg/dL No diabetics 70% Smokers	DBP 95 to 104 mm Hg	Unclear	Dehydrated tablet: Kwai®	600 mg
Kannar (1998)78	Australia	90	54	54	No diabetics or cardiac disease No lipid or antihypertensive drug	TC > 250 mg/dL	Community volunteers	Dehydrated enteric-coated tablet: noncommercial	880 mg
Kiesewetter (1991)87	Germany	60	24	30	13% Smokers	Elevated spontaneous platelet activity	Unclear	Dehydrated tablet: Kwai®	800 mg
Kiesewetter (1993)69	Germany	80	60	54	75% Smokers 75% Elevated TC 60% "Obese" 40% Diabetic	Lower extremity peripheral vascular disease	Unclear	Dehydrated tablet: Kwai®	800 mg
Koscielny (1999)88	Germany	280	60	71	65% Elevated TC 27% Smokers 38% Hypertensive 6% Diabetic	Asymptomatic advanced atherosclerotic plaques in femoral or carotid artery	Unclear	Dehydrated tablet: Kwai®	900 mg
Lash (1998)76	United States	35	45	49	All postrenal transplant	TC > 240 mg/dL and LDL > 160 mg/dL	Unclear	Dehydrated tablet: Pure-Gar®	1,360 mg
Mader (1990)89	Germany	261	59	44	47% Hypertensive 7% Diabetic 12% Coronary disease	TC and/or TG > 200 mg/dL	Community clinic	Dehydrated tablet: Kwai®	800 mg
Mansell (1996)84	United Kingdom	60	63	76	All noninsulin dependent diabetes	Unclear	Unclear	Dehydrated tablet: Kwai®	900 mg
McCrindle (1998)90	Canada	31	14	52	No significant alcohol or tobacco use	TC > 185 mg/dL, family history of hyperlipidemia and early coronary artery disease	Specialty clinic	Dehydrated tablet: Kwai®	900 mg
Melvin (1996)82	Canada	34	41 to 77	47	All hyperlipidemic	Off all drug therapy "Elevated cholesterol"	Unclear	Dehydrated tablet: Kwai®	900 mg
Neil (1996)47	United Kingdom	115	53	61	Mean BMI > 27 15% Smokers	TC 230 to 325 mg/dL and LDL > 130 mg/dL	Community clinic	Dehydrated tablet: Kwai®	900 mg
Rotzsch (1992)91	Germany	24	37	42	29% Hypertensive	HDL < 10 mg/dL (men); HDL < 15 mg/dL (women)	Unclear	Dehydrated tablet: Kwai®	900 mg
de Santos (1993)81	United Kingdom	60	52	33	No lipid or antihypertensive drug	TC > 250 mg/dL	Community clinic	Dehydrated tablet: Kwai®	900 mg
de Santos (1995)83	United Kingdom	80	56	35	No lipid or antihypertensive drug	Mild hypercholesteremia	1 general practice	Dehydrated tablet: Kwai®	600 mg
Saradeth (1994)92	Europe	72	39	29	No fibrinolytic or anticoagulant meds	Age 18 to 50	Community volunteers	Dehydrated tablet: Kwai®	600 mg
Simons (1995)93	Australia	31	54	52	No diabetics or cardiac disease No lipid or antihypertensive drug	TC 230 to 300 mg/dL; TG < 260 mg/dL	Community volunteers	Dehydrated tablet: Kwai®	900 mg
Superko (2000)80	United States	50	53	Unclear	No lipid drugs	Moderate hypercholesteremia	Unclear	Dehydrated tablet: Kwai®	900 mg
Vorberg (1990)75	Germany	40	50	43	No diabetics	TC 230 to 350 mg/dL	Community clinic	Dehydrated tablet: Kwai®	900 mg

Note: To convert cholesterol, LDL, and HDL values from mg/dL to mmol/dL, divide by 38.7. To convert triglyceride values from mg/dL to mmol/dL, divide by 88.2.

Abbreviations used: MI = body mass index; DBP = diastolic blood pressure; HDL = high density lipoprotein; LDL = low density lipoprotein; mg/dL = milligrams per deciliter; mm Hg = millimeters mercury; TC = total cholesterol; TG = triglycerides.

Twenty-two studies used dehydrated garlic preparations that were standardized to an alliin content of 0.52 percent of the total tablet mass (Table 5).^{24 37 38 39 47 69 75 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93} Such standardization is reportedly equivalent to a potential 0.6 percent allicin release. One study standardized to a guaranteed minimum release of 3,000 per million (0.3 percent) allicin.⁷⁶ Another study standardized to produce 2.4 mg allicin per tablet.⁷⁸

Twenty-two trials used nonstandardized preparations or preparations in which standardization was unclear (Evidence Table 2).^{30 34 35 36 40 64 70 71 72 73 74 77 83 94 95 96 97 98 99 100} Three trials used garlic ether extracts;^{34 71 72} four used "aged garlic extract^TM"^{30 77 94} one used cold-pressed garlic oil;³⁶ two used raw or fried garlic;^{35 74} one used "a garlic oil;"⁸³ one used a steam-distilled, enteric-coated, beta-cyclodextrin-bound garlic oil preparation;⁹⁵ two used "dried garlic tablets;"⁹⁶ two used a "spray-dried garlic" preparation;^{40 73} and five used combination products of garlic with other substances such as fish oil, ginkgo, soya, and hawthorn.^{70 97 98 99 100} Although some authors equated different garlic preparations to equivalencies of whole or "fresh garlic," processes such as alcohol extraction, steam distillation, and aging may liberate constituents not present in fresh or whole garlic, making such comparisons meaningless. Thus, dosages in Evidence Table 2 are given only in terms of individual preparations that were used in each study.

Clinical Outcomes

Antilipidemic Effects

Of the 45 randomized trials, 44 evaluated the effectiveness of garlic preparations on various serum lipid endpoints. Five evaluated combination preparations of garlic with either fish oil, hawthorn, or ginkgo biloba.^{70 97 98 99 100} Placebos were used as comparisons except in the following instances: a "no garlic" control,³⁵ an antilipidemic agent,⁸⁵ an antihypertensive agent,³⁸ Kwai^® plus an antihypertensive agent versus an antihypertensive agent only,³⁷ and a head-to-head comparison of two different garlic preparations.⁸³

Ten studies specified low-fat, low-cholesterol, high-fiber diets for participants.^{47 80} One study was conducted under "dietary extremes during Christmas holiday season" with, presumably, high fat and caloric intakes.⁹⁹ Another gave participants 100 g of butter to test effects of garlic on postprandial lipid influences.⁹¹ Other studies defined no specific dietary measures and generally allowed "usual diets."

Because of significant variability in lipid measurement, strict standards of measurement have been suggested. Ideally, weight and diet should be stable for more than 2 weeks, repeat measurements should be used to establish a baseline mean, and some measurements, such as triglycerides, should be obtained after at least a 12-hour fast. LDL should be calculated when possible using the Friedewald equation. Trials of garlic used various lipid measurement protocols; a few followed the suggested stringent guidelines.^{47 64 77 78 80 86}

Trial Results

One multicenter trial involving 98 adults with hyperlipidemia found no differences in lipid outcomes at 3 months between persons given an antilipidemic agent (bezafibrate) and persons given a standardized dehydrated garlic preparation (Kwai^®).⁸⁵ The trial that compared a garlic fish oil combination with placebo reported decreases in total cholesterol with active therapy that were of greater magnitude than reductions of most other trials and that were statistically significant.⁹⁷ The single head-to-head comparison of a standardized dehydrated preparation (Kwai^®) with garlic oil (Hoefels Original Garlic Oil^®) was an open trial that did not attempt blinding.⁸³ This trial found no significant differences between the two preparations in total cholesterol or HDL measurements at 1, 2, 3, and 4 months. A statistically significant reduction in low density lipoprotein cholesterol (LDL-C) favoring Kwai^®over Original Garlic Oil was found at 4-month, but not 2-month, followup observations.

Figure 3. Galbraith plot of studies with total cholesterol (more...)

   Figure 3. Galbraith plot of studies with total cholesterol results at 4 to 6 weeks

Figure 4. Funnel plot of studies with total cholesterol (more...)

   Figure 4. Funnel plot of studies with total cholesterol results at 4 to 6 weeks.

Figure 5. Galbraith plot of studies with total cholesterol (more...)

   Figure 5. Galbraith plot of studies with total cholesterol results at 8 to 12 weeks

Figure 6. Funnel plot of studies with total cholesterol (more...)

   Figure 6. Funnel plot of studies with total cholesterol results at 8 to 12 weeks

Figure 7. Galbraith plot of studies with total cholesterol (more...)

   Figure 7. Galbraith plot of studies with total cholesterol results at 20 to 24 weeks

Figure 8. Funnel plot of studies with total cholesterol (more...)

   Figure 8. Funnel plot of studies with total cholesterol results at 20 to 24 weeks

Meta-analyses of other trials that were placebo controlled and that reported total cholesterol outcomes at 4 to 6 weeks, 8 to 12 weeks, and 20 to 24 weeks were conducted. (Of note, the study by Kandziora that compared garlic plus an antihypertensive agent with antihypertensive therapy alone is included in some of the analyses.) Prior to conducting meta-analyses, examination of Galbraith and funnel plots as well as statistical testing for heterogeneity identified some trials with outlier results (Figures 3-8). Review of the design and protocol characteristics of such trials did not elucidate obvious systematic differences that could explain their outlier status. Quantitative analyses were performed with and without outliers. Results are described below for analyses without the outlier studies. (Including the outliers yielded slightly different effect sizes that were in the same direction and usually of the same statistical significance as the results that excluded the outliers.)

Figure 9. Forest plot of studies with total cholesterol (more...)

   Figure 9. Forest plot of studies with total cholesterol results at 4 to 6 weeks

Figure 9 shows results of meta-analyses for trials reporting total cholesterol outcomes at 4 to 6 weeks. Combining all studies, regardless of preparation type, statistically showed that garlic preparations compared with placebo significantly reduced total cholesterol levels on average by 7.2 mg/dL (95 percent confidence interval [CI] 1.2 to 13.2 mg/dL). Combining only studies that used standardized dehydrated garlic preparations showed average reductions of 10.2 mg/dL (95 percent CI 3.1 to 17.3 mg/dL). A few studies that reported outcomes at 4 to 6 weeks could not be included in the figure or meta-analyses. Two studies reported that garlic significantly decreased cholesterol.^{36 82}

Figure 10. Forest plot of studies with total cholesterol (more...)

   Figure 10. Forest plot of studies with total cholesterol results at 8 to 12 weeks

Figure 10 gives total cholesterol results at 8 to 12 weeks. Combining all studies, regardless of preparation type, statistically showed that garlic preparations compared with placebo significantly reduced total cholesterol levels on average by 17.1 mg/dL (95 percent CI 12.4 to 21.9 mg/dL). Combining only studies that used standardized dehydrated garlic preparations showed average reductions of 19.2 mg/dL (95 percent CI 13.0 to 25.4 mg/dL). Of note, one 12-week study that is not shown in the figure stated that "garlic tablets (Kwai^®) led to a marginal fall in total cholesterol at 6 weeks that was not sustained at 12 weeks."⁸⁴

Figure 11. Forest plot of studies with total cholesterol (more...)

   Figure 11. Forest plot of studies with total cholesterol results at 20 to 24 weeks

Figure 11 shows results of the eight trials that reported total cholesterol outcomes at 20 to 24 weeks. Average reductions with garlic preparations compared with placebo at this time point were not significant (1.2 mg/dL, 95 percent CI --8.2 to 10.7 mg/dL). Limiting this analysis to the three 6-month studies that compared standardized dehydrated garlic preparations with placebo also was not significant (2.8 mg/dL, 95 percent CI -8.7 to 14.4 mg/dL) (Figure 11). Removing Koscielny's study from these analyses because of its questioned authenticity yields a larger but still nonsignificant effect. The statistical significance of these analyses should be interpreted cautiously because of few studies and low power. One additional trial that was not shown in the figure reported that total cholesterol decreased significantly after 3 years of followup in postinfarction patients who were given 6 to 10 g of garlic ether extract compared with patients given placebo (reported total cholesterol changes: 6.85 to 6.21 mmol with garlic and "no change" with placebo).⁷¹

Figure 12. Forest plot of pooled results of total cholesterol (more...)

   Figure 12. Forest plot of pooled results of total cholesterol findings at different time periods

Figure 12 shows the total cholesterol summary results for the different time periods. Observed reductions at 4 to 6 weeks appear smaller than observed reductions at 8 to 12 weeks, but not smaller than reductions at 20 to 24 weeks. It is not clear whether these observations are due to systematic differences among studies that have longer and shorter followup durations or are due to time-dependent effects of garlic. For example, there were few long-term studies; longer studies had higher dropout rates, and longer studies may have used preparations of Kwai^® that had lower allicin-release properties. Observations depicted in Figure 12 also suggest, but do not prove, that standardized dehydrated preparations may result in greater short-term (4- to 12-week) total cholesterol reductions than other preparations.

LDL reductions at 8 to 12 weeks for "all studies regardless of preparation type compared with placebo (n=13)" and for "dehydrated preparation studies only compared with placebo (n=10)" were 6.2 mg/dL (95 percent CI 0.8 to 11.7 mg/dL) and 6.7 mg/dL (95 percent CI 0 to 13.5 mg/dL), respectively. Triglyceride reductions at 8 to 12 weeks for "all studies regardless of preparation type compared with placebo (n=17)" and for "dehydrated preparation studies only compared with placebo (n=13)" were 19.1 mg/dL (95 percent CI 7.6 to 30.4 mg/dL) and 21.1 mg/dL (95 percent CI 8.3 to 34.0 mg/dL), respectively. None of the analyses regarding HDL values were statistically significant, possibly because few studies reported HDL. At 8 to 12 weeks, the average HDL reduction for "all garlic preparations compared with placebo (n=14)" was 0.9 mg/dL (95 percent CI --1.0 to 2.8 mg/dL), and for "standardized dehydrated preparations only compared with placebo (n=10)" was 0.2 mg/dL (95 percent CI --2.1 to 2.4 mg/dL). LDL, triglyceride, and HDL analyses are given in Appendix A.

Sensitivity analyses that were limited to trials that specifically included only participants with hyperlipidemia did not vary significantly from those presented above. (Few trials had normal mean values of total cholesterol at baseline; thus, few were excluded in this sensitivity analyses.) Finally, sensitivity analyses that excluded studies that were not designed as double blind were not different than the analyses presented above.

Antihypertensive Effects

Thirty trials reported protocols that included blood pressure measurement.^{24 36 37 38 39 64 69 70 71 75 77 78 79 80 81 83 84 85 86 87 88 89 90 92 93 94 96 98 100} Two did not comment on blood pressure results.^{88 89} Baseline blood pressures and outcomes from the 27 trials that reported results are listed in Evidence Table 3. Nineteen of the trials tested standardized dehydrated garlic preparations.^{24 39 64 69 75 79 80 81 83 85 86 87 88 89 90 92 93 93} All except four used placebo controls: Holzgartner compared garlic with an antilipidemic agent,⁸⁵ Kandziora compared garlic with an antihypertensive agent (a reserpine-diuretic combination),³⁸ Kandziora used an antihypertensive cointervention in both the garlic and control group with no placebo addition in the control group,³⁷ and de Santos compared the dehydrated preparation with a garlic oil preparation.⁸³

Participants in the studies were sometimes prescribed antilipidemic diets,^{78 80 81 85 86 90 94} or physical activity.⁶⁹ Both trials with antihypertensive regimens specifically excluded participants from taking other antihypertensive medications.^{37 38} Seven trials precluded use of antihypertensive agents.^{64 70 78 79 81 83 87}

Initial blood pressures of participants were variable. In 14 trials, participants were normotensive. Six trials included normotensive and hypertensive participants; mean baseline blood pressures ranged from systolic pressures of 140 to 165 mmHg and/or diastolic pressures of 85 to 100 mmHg.^{36 75 81 83 85 98} Three studies specifically evaluated hypertensive patients, and one studied patients with atherosclerotic peripheral vascular disease.^{37 38 39 70} Average baseline blood pressures in these four trials ranged from systolic pressures of 170 to 180 millimeters of mercury (mm Hg) and diastolic pressures of 99 to 102 mm Hg.

Trial Results

Several trials reported significant reductions in blood pressure in participants given garlic at 10-- to 13--week followup times (within-group comparisons). However, our comparisons between groups showed that only three trials demonstrated statistically significant reductions in blood pressure with garlic compared with placebo.^{39 75 79} Three showed small statistically significant reductions in diastolic blood pressure that ranged from 2 to 7 percent. ^{39 75 79} One showed small statistically significant reductions in systolic blood pressure of approximately 3 percent.⁷⁹ Other trials showed no statistically significant differences in blood pressure between participants assigned to garlic and those assigned to placebo.^{37 38 64 69 70 78 80 81 84 85 90 92 94 96 98} Of note, these data were not pooled because half of the studies did not present numerical data that could be used in a quantitative analysis, multiple different methods of blood pressure measurement were used, and few studies had a priori hypotheses related to blood pressure.

Three trials, which did not specifically study hypertensive participants, reported blood pressure outcomes at 5 to 6 months.^{77 81 94} One of these trials found that a standardized, dehydrated tablet preparation (Kwai^®) reduced systolic (by 22 mm Hg) and diastolic blood pressures (by 11 mm Hg) significantly more than placebo.⁸¹ Another reported that "aged garlic extract^TM" significantly reduced systolic and diastolic blood pressures compared with placebo, but our analyses using adjusted values derived from the reported baseline figures and followup values from the crossover point did not confirm statistically significant diastolic differences in favor of garlic.⁹⁴ A third trial reported that "aged garlic extract^TM" did not significantly reduce systolic or diastolic blood pressures more than placebo.⁷⁷ The 3-year trial in postinfarction patients reported that blood pressure levels at followup were "significantly lower" in participants given garlic compared with those given placebo, but actual values were not given.⁷¹

Analyses adjusting for baseline differences of the trial that compared different preparations of garlic found that the standardized dehydrated preparation (Kwai^®) led to statistically significant greater reductions in systolic and diastolic blood pressures compared with the oil preparation (Hoefels Original Garlic Oil^®) at 1-, 2-, 3-, and 4-month followup periods.⁸³ There were no statistically significant differences in blood pressure between groups at 4- and 12-week followup points in the study with hypertensive subjects that compared dehydrated garlic with an antihypertensive agent.³⁸ The trial with hypertensive subjects that compared garlic with no garlic in patients who were given an antihypertensive agent also found no significant differences in blood pressure between groups.³⁷

Antiglycemic Effects

Twelve trials assessed the effect of garlic supplementation on serum glucose concentration (Evidence Table 4).^{24 37 38 39 70 72 73 84 87 88 96} Two of the trials considered glucose a primary outcome;^{73 84} one included glycosylated hemoglobin outcomes;⁸⁴ one addressed the responsiveness of serum insulin concentration to oral glucose tolerance testing in patients treated with garlic;⁷³ and another included serum insulin and C peptide results.⁸⁴

Participant characteristics in the trials with glucose outcomes were quite variable. Two trials studied adults with diabetes.^{73 84} Three followed participants with a known atherosclerotic disease process such as coronary or peripheral arterial occlusive disease.^{70 72 88} Three had participants with hypertension,^{37 38 39} and two had participants with high serum lipids.^{37 38} No trials assessed changes in diet or activity level between the groups. Four reported equivalent body masses of patients at baseline, but only one reported that there were no changes in body mass between the groups at the end of the intervention.²⁴

Trial Results

One study reported a statistically significant benefit in serum glucose lowering from garlic supplementation that was not found in the control group.⁸⁷ This trial studied 60 patients with baseline elevated spontaneous platelet aggregation values and measured serum glucose as a secondary endpoint over 4 weeks of study. Baseline serum glucose was 89.0 mg/dL. No other studies reported significant effects of garlic on glucose. No significant changes in postprandial insulin levels, glycosylated hemoglobin, or C peptide were reported.^{73 84}

Antithrombotic Effects

Ten trials assessed the effectiveness of oral garlic supplementation on potential prothrombotic risk factors (Evidence Table 5).^{36 69 70 72 74 87 88 90 96 102} Four used standardized dehydrated garlic preparations. ^{69 87 88 90} All except one, which used no treatment in the control group,⁷⁴ compared garlic supplementation with placebo controls. Results were reported in all but one of the publications,⁸⁸ amounting to a cumulative database of only 409 patients.

Spontaneous platelet aggregation

Five trials lasting at least 4 weeks measured the effects of garlic consumption on platelet aggregation; four trials provided results.^{36 69 70 87 102} In two of these trials, Kiesewetter reported statistically significant reductions in platelet aggregation with a standardized dehydrated garlic preparation (Kwai^®) compared with placebo. One of Kiesewetter's trials involved persons with known elevations of spontaneous platelet aggregation and a followup period of 4 weeks.⁸⁷ The other trial involved patients with lower extremity peripheral arterial occlusive disease who were followed for 12 weeks.⁶⁹ Another study in healthy Canadian volunteers reported a 16.4 percent reduction in platelet aggregation after 4 weeks of daily supplementation with cold pressed garlic compared with a 5.6 percent increase in the placebo group.³⁶ Czerny's trial involved patients with hyperlipidemia, peripheral vascular disease, and claudication who were followed for 16 weeks.⁷⁰ Outcomes were measured as the increase in added adenosine diphosphate (ADP) required to aggregate platelets. A 46 percent reduction in aggregation in the treatment group after 16 weeks compared with a 7 percent increase in the placebo group was found. A 6-month trial assessed platelet aggregation in a subgroup of 15 participants from a larger trial of 52 participants.¹⁰² "Aged garlic extract^TM" compared with placebo did not have statistically significant effects on ADP-induced aggregation, but it did significantly decrease epinephrine-induced platelet aggregation.

Fibrinolytic activity

Three trials assessed changes in fibrinolytic activity. Two were performed in Indian patients with preexisting coronary artery disease.^{72 74} In one, Bordia reported that garlic oil ethyl acetate extract compared with placebo significantly increased fibrinolytic activity at 6- and 12-week followup points.⁷² In the other, Chutani reported that both raw and fried garlic preparations given in 3 g of butter compared with no garlic resulted in increased fibrinolytic activity at 4 weeks.⁷⁴ The third trial assessed effects of a nonstandardized dehydrated tablet in German patients with type IIa, IIb, or IV hyperlipoproteinemia.⁹⁶ This crossover trial with therapy periods of 6 weeks reported "no significant change" in either the garlic or placebo groups.

Serum fibrinogen levels

None of four trials reported any significant change in this endpoint. ^{72 87 90 96}

Plasma viscosity

Changes in plasma viscosity following supplementation with a standardized dehydrated tablet were assessed in three trials.^{69 87 88} One did not report viscosity results.⁸⁸ The remaining two trials were performed on German participants by the same investigators. One 4-week trial reported that Kwai^® was associated with a statistically significant reduction in viscosity that was not reproduced in the placebo group, but estimates of between-group differences were not statistically significant.⁸⁷ The second trial reported that Kwai^® statistically reduced viscosity significantly more than placebo at 12 weeks.

Effects of Garlic on Morbidity and Mortality

Of three trials with clinical endpoints, two assessed improvement in pain-free walking distances.^{69 70} Kiesewetter evaluated 80 patients with cardiovascular risk factors and stage II peripheral vascular occlusive disease of the lower extremities (femoral, popliteal, or tibial occlusive disease more than 60 percent).⁶⁹ All participants received 90 minutes of supervised physical therapy per week and were randomized to standardized dehydrated garlic or placebo. They were followed for 12 weeks. Treatment and placebo groups were comparable in their consumption of antilipidemic, antihypertensive, antidiabetic, and other cardiac drugs. Outcomes were assessed in 64 patients who completed the study. Attrition was 20 percent in each group; 10 percent of the treatment group dropped out due to intolerance of garlic odor. Pain-free walking from baseline increased by approximately 40 meters in the treatment group compared with approximately 30 meters in the placebo group. Although the authors reported these as significant differences, our analyses showed that the differences were not statistically significant. The authors apparently compared shorter walking distances obtained during a run-in period with distances obtained at 12 weeks, while we compared longer distances obtained at baseline when randomization occurred with those obtained at followup.

Czerny assessed pain-free treadmill walking distance in 100 patients with "arteriosclerotic dependent illnesses such as phase II intermittent claudication."⁷⁰ Participants were randomized to 4 months of supplementation with a garlic oil macerate/soya lecithin/hawthorn oil/wheat germ oil combination or placebo. No attrition was reported in either group. Maximum walking distance increased significantly in the treatment group (114 percent) compared with the placebo group (17 percent) (p<0.05). Whether soya lecithin, hawthorn oil, or wheat germ oil have independent effects on walking distance could not be evaluated.

Bordia addressed reinfarction rates in 432 patients who had sustained prior myocardial infarctions.⁷¹ Recruitment setting, method of diagnosis of infarction, and time between prior infarct and randomization were not clear. Despite contacting this author repeatedly, we were unable to verify the randomization process and any blinding methods. Baseline health characteristics were reportedly equivalent between groups, and all participants received "standard medical therapy" for postinfarction patients. Compliance, adverse effects, and dropouts were not reported. At the 3-year followup, the author reported 11 deaths and 15 reinfarctions in the 222 participants randomized to garlic (0.1 gram per kilogram per day [g/kg/d]) and 20 deaths and 22 reinfarctions in the 210 placebo recipients. Although the author reported significant differences, when data between groups were analyzed with a chi-square test, there were no statistically significant differences in total mortality (p=0.07) or myocardial infarction (p=0.13).

Atherosclerotic Plaque Regression

Regression of atherosclerotic plaque volume and intimal-medial plaque thickness as estimated by B-mode ultrasound of carotid and femoral arteries was assessed in one trial.⁸⁸ Two hundred eighty patients with atherosclerotic risk factors and advanced, but not flow-limiting, stenoses were treated with standardized dehydrated garlic tablets (Kwai^®) or placebo over 48 months. Baseline risk factors, plaque localization, average plaque volume, and gender ratio were matched at baseline. However, women assigned to placebo were younger than those assigned to garlic: 39 percent in the placebo group were 45 to 55 years old compared with only 9 percent in the garlic group. Results were given for the 152 (91 placebo and 61 garlic) persons who completed all 4 years of the trial (46 percent attrition). Mean plaque volume decreased 2.6 percent with garlic and increased 15.6 percent with placebo (p<0.0001 for difference between groups). The progression of plaque volume in women in the placebo group (53 percent increase over 4 years) was significantly greater than any other group (p<0.0001). The difference between garlic- and placebo-treated men over the course of the study was 4.4 percent favoring garlic treatment (p<0.0001). (Of note, the authenticity of this trial has been challenged because of questions about the legitimacy of the ultrasound figures that accompanied the publication and because of the high unequal dropout rates that were not accounted for in the analysis.^{101 103} As of December 1999, these allegations were still under investigation.)

Results: Associations With Cancer

One survey (analyzed as a case-control study), 10 case-control studies, and 2 case-cohort studies examined whether garlic consumption is associated with decreased incidence of precancerous lesions or cancer.^{55 56 59 60 61 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118} One case-cohort study¹⁰⁹ examined associations with multiple cancers, including breast,¹⁰⁶ colorectal,¹⁰⁸ gastric,¹¹² and lung cancer.¹¹³ Several other case-control studies evaluated associations of cancer with Allium vegetables such as onions and leeks or with mixtures of ingredients, but did not separately assess garlic associations. Such studies were not included in this report.^{58 119 120 121 122 123 124 125 126 127 128} In addition, a Chinese survey that did not state the numbers of cases and controls was not included.¹²⁹ This survey reported that death from gastric cancer was lower in a county where garlic consumption was high compared with a neighboring county where garlic consumption was low. (Of note, we know of a relevant recently completed case-control study from Israel, but were unable to obtain analysis from this study as of December 30, 1999. In addition, we found an ongoing, but not yet reported, large trial in China that is evaluating effects of two garlic preparations on gastric precancerous lesions.¹³⁰ )

Results: Adverse Effects of Garlic

Garlic and Cardiovascular Disease Summary Points

1. In adults or children with or without dyslipidemia, does oral ingestion of garlic (fresh, cooked, or supplements) compared with no garlic, other oral supplements, or drugs lower plasma lipid levels?

Thirty-seven randomized trials, all but one in adults, consistently showed that compared with placebo, various garlic preparations led to small reductions in total cholesterol at 1 month (range of average pooled reductions 1.2 to 17.3 mg/dL) and 3 months (range of average pooled reductions 12.4 to 25.4 mg/dL). Eight placebo-controlled trials reported outcomes at 6 months; pooled analyses showed no significant reductions of total cholesterol with garlic compared with placebo. It is not clear whether statistically significant positive short-term effects, but negative longer term effects, are due to systematic differences in studies that have longer and shorter followup durations, fewer longer term studies, or time-dependent effects of garlic. Statistically significant reductions in LDL (range 0 to 13.5 mg/dL) and in triglycerides (range 7.6 to 34.0 mg/dL) also were found in pooled analyses at 3 months. No significant changes in HDL were seen in pooled analyses at 1 and 3 months. One multicenter trial involving 98 adults with hyperlipidemia found no differences in lipid outcomes at 3 months between persons given an antilipidemic agent and persons given a standardized dehydrated garlic preparation (Kwai^® ). Most trials had significant methodological flaws such as unclear randomization processes and no intention-to-treat analyses, which limited the ability to make firm conclusions.

2. In adults or children with or without hypertension, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs lower blood pressure?

Twenty-seven small, randomized trials, all but one in adults and of short duration, reported mixed, but never large, effects of garlic on blood pressure outcomes. Most studies did not find significant differences in persons randomized to garlic compared with those randomized to placebo. The one small trial (n=40) that directly compared a standardized dehydrated garlic preparation to an active antihypertensive agent found no differences in blood pressure between groups. Because of unclear randomization processes, lack of intention-to-treat analyses, missing data, and variability in blood pressure measurement techniques, no firm conclusions can be made from these trials.

3. In adults or children with or without diabetes, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs increase insulin sensitivity?

Two small, short trials, both in adults, reported no statistically significant effects of garlic compared with placebo on serum insulin or C peptide levels.

4. In adults or children with or without diabetes, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs lower plasma glucose levels or glycosylated hemoglobin?

Twelve small, randomized trials, all in adults and of short duration, suggested that garlic has no significant effect on glucose in persons with or without diabetes. Because of the small number of trials, the short followup, unclear randomization processes, no intention-to-treat analyses, and missing data, no firm conclusions can be made.

5. In adults or children with or without diabetes, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs increase antithrombotic activity?

Ten small, randomized trials, all but one in adults and of short duration, showed promising effects of garlic on platelet aggregation and mixed effects on plasma viscosity and fibrinolytic activity. Because of few participants in the trials, short followup durations, unclear randomization processes, no intention-to-treat analyses, missing data, and variability in techniques used to assess outcomes, no firm conclusions can be made.

6. In adults at average or high risk for cardiovascular disease, does oral ingestion of garlic compared with no garlic, other oral supplements, or drugs decrease cardiovascular morbidity or mortality?

There are insufficient data to confirm or refute effects of garlic on clinical outcomes such as myocardial infarction and claudication. One randomized trial with 492 participants found no statistically significant decreases in numbers of myocardial infarctions and deaths when placebo was compared with 6 to 10 g of garlic ether extract after 3 years. Numbers of participants who adhered to prescribed garlic regimens or who failed to complete the study followup were not described. This trial was not published in peer-review literature. Details confirming its randomization process and followup were not obtained despite requests to the author.

Two trials in patients with atherosclerotic lower extremity disease found that garlic increased pain-free walking distance at 12 to 16 weeks compared with placebo. In the first trial, 80 participants were randomized to standardized dehydrated garlic or placebo, but only 64 completed the trial. Among those, pain-free walking distance increased by approximately 40 meters with garlic compared with approximately 30 meters with placebo. In the second trial, 100 participants were randomized to placebo or a garlic oil macerate/soya lecithin/hawthorn oil/wheat germ oil combination. At 4 months, maximum walking distance increased significantly more in the garlic group (114 percent) than in the placebo group (17 percent) (p<0.05).

7. Do different preparations of garlic vary in effectiveness regarding any of the above outcomes?

RCTs do not establish whether garlic effectiveness varies across preparations or dosages. Limited data, not derived from head-to-head comparisons, suggest, but do not prove, that standardized dehydrated preparations may have greater short-term (1- to 3-month) lipid-lowering effects than other preparations.

Garlic and Cancer Summary Points

1. In adults without cancer risk factors, precancerous conditions, or malignancy, does garlic ingestion compared with no garlic, placebo, or other oral supplements reduce the risk of developing cancer?

Scant data, primarily from case-control studies, suggest, but do not prove, that dietary garlic consumption is associated with decreased odds of laryngeal, gastric, colorectal, and endometrial cancer, and adenomatous colorectal polyps. Single case-control studies suggest, but do not prove, that dietary garlic consumption is not associated with breast or prostate cancer. Data regarding esophageal cancer are conflicting. No conclusions regarding associations between the use of particular garlic supplements and precancerous lesions or cancer can be made. Preliminary 3-year evidence from a large cohort study suggests that consumption of "any" garlic supplement does not reduce the risk of breast, lung, colon, or gastric cancer. This study has not reported associations relevant to the consumption of fresh or raw garlic, and its data about supplements are limited because information is not available about different types and brands of garlic supplementations.

2. In adults with cancer risk factors, but no precancerous conditions or malignancy, does garlic ingestion compared with no garlic, placebo, or other oral supplements reduce the risk of developing cancer?

No controlled studies in humans specific to this question were found.

3. In adults, does garlic ingestion compared with no garlic, placebo, or other oral supplements reduce the risk of developing precancerous lesions?

There are very limited data regarding the associations of garlic with precancerous lesions. One small survey that was limited by low response rates showed no statistically significant associations between the consumption of raw garlic and precancerous gastric lesions. One large case-control study suggested that increased garlic consumption was associated with lower odds of colorectal polyps. There is an ongoing, but not yet reported, large trial in China that is evaluating the effects of two garlic preparations on gastric precancerous lesions.

4. In adults with cancer, does garlic ingestion compared with no garlic, placebo, or other oral supplements reduce morbidity or mortality of cancer?

No controlled studies in humans were found that inform this question.

5. Is garlic more effective against some types of risk factors, precancerous conditions, or cancer than others?

Evidence regarding different precancerous lesions and cancers is summarized under the first and third questions above. Associations with cancer have been studied using different definitions of garlic exposure, different study populations, and different study designs. These issues preclude sound judgments regarding whether garlic has greater protective associations with some cancers than with others.

Adverse Effects Summary Points

1. What are the symptomatic and serious adverse effects of different garlic preparations, and what is their frequency?

Adverse effects of oral ingestion of garlic are "smelly" breath and body odor. Other possible, but not proven, adverse effects include flatulence, esophageal and abdominal pain, small intestinal obstruction, contact dermatitis, rhinitis, asthma, bleeding and hematoma, and myocardial infarction. The frequency of adverse effects and whether they vary by particular preparations are not established. Adverse effects of inhaled garlic dust include allergic reactions such as asthma, rhinitis, urticaria, angioedema, and anaphylaxis. Adverse effects of topical exposure to raw garlic include contact dermatitis, skin blisters, and ulcero-necrotic lesions. Frequency of reactions to inhaled garlic dust or topical exposures of garlic is not established.

2. Are there interactions between garlic and commonly used medications for dyslipidemia, diabetes, or thrombogenic disease (e.g., statins, sulfonylureas, antithrombotic agents)?

We found no evidence in humans to inform this question except for reports of two patients taking warfarin who experienced increases in their International Normalized Ratio (INR) when taking garlic pearls or tablets.

Limitations

Many randomized trials that have evaluated effects of garlic on cardiovascular-related endpoints are limited by short durations, inadequate randomization and blinding procedures, lack of clear specification of contents of garlic preparations (including their constituents and dissolution properties), lack of intention-to-treat analyses, and incomplete reporting of data. The meta-analysis of lipid data that we performed is limited by some missing data at different time points and by the need to impute variability data from some trials.

Few studies assessed associations between garlic consumption and cancer, and some studies may have been missed because they addressed associations with multiple foods. Findings specific to garlic may not be analyzed or reported and, even if reported, may be presented only in an appendix or as a single line in the text. Studies that were found sometimes failed to distinguish the type of garlic exposure (raw, cooked, or specific supplement), used subject recall to assess different frequencies of use over varying time periods, and adjusted for various potential confounders in different ways. Although we believe we found most reported adverse-effect literature regarding garlic, adverse effects in general are frequently underreported or reported in manners that do not allow determination of causality and frequency.

Equivalency of Garlic Preparations

Equivalency of different garlic preparations is unclear; studies that estimate equivalency for different constituents are warranted. This is an essential step before conducting more large trials with cardiovascular outcomes. The exact or standardized content of preparations and the amount of release of the main constituents under simulated gastrointestinal dissolution conditions should be specified.

Proper Placebos

Garlic supplements, including ones labeled as odor free, may produce easily detectable breath or body odor that may impair the ability to conduct double-blind trials. Placebos that are designed to simulate this odor should be designed and tested.

Cardiovascular-Related Effects

Studies in humans with lipid and antithrombotic outcomes are promising, but they are limited by unclear randomization procedures, short durations, and unclear ability to blind active treatment administration and outcome assessments. More studies with similar design features are not needed. Rather, a few well-designed trials of longer duration that are powered to assess morbidity and mortality outcomes, as well as lipid and thrombotic outcomes, are needed. Such trials should attempt and assess adequacy of blinding.

It is unclear whether particular preparations, constituents, and/or dosages of garlic are superior to others in affecting physiological outcomes. Comparative studies of different preparations, constituents, and dosages with both intermediate physiological and clinical outcome measures are warranted. Comparative studies with alternative agents, such as 3-hydroxy-3-methyl-glutaryl (HMG) co-reductase inhibitors (statins), also are warranted.

Cancer-Related Associations

Scant data, primarily from case-control studies, suggested, but did not prove, that garlic consumption is associated with decreased odds of laryngeal, gastric, colorectal, and endometrial cancer and adenomatous colorectal polyps. More case-control and cohort studies that evaluate such associations, as well as more studies evaluating associations with other cancer types, are needed. Future studies should aim for more precision in assessing types, formulations, and constituents of garlic that are consumed. The frequency and the duration of consumption also warrant more precise quantification. Sampling techniques that allow greater levels of garlic consumption to be represented should be emphasized as some existing studies are limited by population samples that consume very small quantities of garlic. Consideration should be given to mounting more trials, such as the ongoing Chinese trial, that evaluate the protective effects of different garlic preparations in persons with very high risk of cancer or precancerous lesions. Future systematic reviews in this area should search more broadly for diet-related population studies and aim to place findings that are specific to garlic in a broader context that takes into account findings regarding other Allium-containing vegetables as well as other foods.

Adverse Effects

The frequency and severity of adverse effects that are related to garlic should be quantified. Whether adverse effects are specific to particular preparations, constituents, and doses of garlic should be elucidated. Whether certain adverse effects are unique to particular types of garlic exposure (e.g., inhaled, oral, or topical) should be clarified.

The most serious potential adverse effects of garlic that have been cited are complications related to bleeding. Whether particular preparations and constituents of garlic affect physiological parameters related to bleeding such as platelet adhesiveness, prothrombin time, and partial thromboplastin time, as well as whether particular preparations lead to clinically significant bleeding, warrants more study.

Interactions

As the most promising results of garlic supplements relate to potential decreases in lipids and many persons are already taking medications for dyslipidemia, interactions between garlic supplements and these medications should be evaluated. Whether there are particular subsets of persons, such as those who are taking warfarin or other antithrombotic agents who are particularly susceptible to bleeding related to garlic, also warrants study.

Appendix A. Supplemental Figures

Appendix B. Contributors

We owe a major debt of gratitude to the following groups of multidisciplinary experts from around the world who assisted in preparing this report: 10 national advisory panel members and 5 technical experts who helped define the scope and shape the content, 20 peer reviewers representing a variety of backgrounds and viewpoints, and 4 scientific authors who provided additional data from their studies.

Appendix C. Acronyms and Abbreviations

ADP:adenosine diphosphate

ASA:acetyl-salicilic acid

BMI:body mass index

BP:blood pressure

°C:degrees Celsius

CI:95% confidence interval

Cr:crossover

CS:change score

D:diastolic measure

DBP:diastolic blood pressure

FA:fibrinolytic activity

g:gram

g/kg/d:grams per kilogram per day

HDL:high-density lipoprotein levels

IBIDS:International Bibliographic Information on Dietary Supplements

IgE:Immunoglobulin E

INR:International Normalized Ratio

kg:kilogram

LDL:low-density lipoprotein levels

M:mean measure

mcM:micromoles

mg:milligram

mg/dL:milligrams per deciliter

mg/g:milligrams per gram

mg/kg:milligrams per kilogram

ml:milliliter

mm Hg:millimeters Mercury

mmol/dL:millimoles per deciliter

mPas:millipascals

NCEP:National Cholesterol Education Program

OR:odds ratio

PA:platelet aggregation

PV:plasma viscosity

RCT:randomized controlled trial

RR:risk ratio

S:systolic measure

SAC:S-allylcysteine

SDstandard deviation

Sd:standing

SE:standard error of the mean

SF:serum fibrinogen

SH:serum homocysteine

Sp:suprine

St:sitting

TC:total cholesterol

TG:triglycerides

Uc:position unclear