Garlic and Allicin in the Body: Chemistry, Pressure, Lipids, and Safety

Crush a clove of garlic, and the kitchen fills with a scent that was almost non-existent in the whole bulb. This small event captures a scientific problem well: garlic changes even before it is consumed. Allicin, powder, oil, and aged extract share a common origin but enter research studies with different compositions. Therefore, any discussion regarding health benefits begins with the question of exactly which garlic is on the plate or in the capsule.

This familiar ingredient has an unusually long clinical biography: from brief experiments with immune cells to long-term observation of stomach cancer. Within it, one finds encouraging blood pressure changes, modest mean lipid shifts, and well-organized negative trials. To understand their practical significance, one must trace the path from the damaged clove to the measured outcome without skipping over digestion, the form of the preparation, and the characteristics of the participants. Such an approach allows for maintaining interest in garlic while simultaneously discarding promises to "cleanse the vessels" or replace medical treatment for infections.

Chemistry of the clove: why allicin is not equal to garlic

Garlic (Allium sativum) is a source of many compounds, among which organic sulfur is particularly interesting. In intact tissue, the precursor alliin and the enzyme alliinase are spatially separated; cutting or crushing allows them to meet. The enzyme triggers the conversion of alliin, and from the intermediate products, allicin, which belongs to the thiosulfinates, is formed. Consequently, allicin appears as a result of tissue damage, rather than being stored in the whole clove as a constant reserve of ready-made active substance.

This structure explains why methods of preparing raw materials matter. A whole clove, fresh paste, and dried powder may contain different combinations of precursor, enzyme, and reaction products. For powder, it is not only the mass that is significant but also the preservation of the enzyme system: the presence of alliin does not yet guarantee the formation of the expected amount of allicin. It is precisely this problem that Lawson and Hunsaker investigated, comparing food forms with supplements, rather than regarding any garlic capsule as a chemical twin of the fresh plant.

Allicin also does not remain the final point of these transformations. Under appropriate conditions, it yields other sulfur-containing products, including diallyl disulfide and diallyl trisulfide; their properties differ. Polysulfides, for example, are studied as potential sources of hydrogen sulfide in biological reactions, whereas allicin is known for its high reactivity toward thiol groups. Therefore, the phrase "allicin's action" is often too narrow to explain the effects of a complex extract.

A thiol group is a part of a molecule containing sulfur and hydrogen, denoted as –SH. Such groups exist in the amino acid cysteine within proteins and in glutathione, which participates in maintaining cellular redox balance. In experiments on bacteria, allicin altered the cysteine sites of proteins, reduced the available pool of glutathione, and disrupted the function of certain enzymes. This is a specific mechanism of chemical stress, not proof of the selective destruction of pathogens inside a human.

Aged extract, usually designated as AGE, represents a different technological route. Prolonged processing changes the profile of compounds, and in studies, the preparation can be standardized by S-allylcysteine, abbreviated as SAC. For example, in the AGE at Heart study, the daily dose contained 1.2 g of extract and 1.2 mg of SAC: these are different values describing the mixture and its component. Specifying only a large "equivalent of fresh garlic" does not provide the same information about the preparation.

The word "oil" is also ambiguous. Steam distillation and infusing garlic in edible oil are different processes, so the resulting products cannot be compared under a single name. The European monograph describes herbal preparations separately with specific characteristics of raw materials and extraction; the history of use applies to these preparations. For a specialist, the technological description is just as important as the number of milligrams, because without it, it is impossible to understand what material the participant actually received.

This chemistry implies a practical way to read a label: first identify the form, then find the content of the components and the size of a daily dose. The labels "garlic," "allicin," and "aged extract" do not denote the same dose across different packaging. Even the same mass of two preparations does not ensure the same effect if the enzymatic activity and technology differ. Here, it is useful to stop searching for the "strongest" garlic and instead look for an exact match to the product for which data has been published.

Preparation and absorption: what remains after the kitchen and digestion

Heating interferes with several processes at once: it can change enzyme activity and substances that have already formed. Therefore, the sequence of actions—crushing first, then heating, or heating whole—has chemical significance. In the Cavagnaro study, whole and crushed cloves were compared after different processing modes, and their extracts were then added to human platelets. Participants did not eat these dishes: the suppression of cell aggregation was measured in vitro.

When boiled or heated in an oven at 200 °C for up to three minutes, laboratory activity was preserved. After six minutes, whole cloves lost it, while pre-crushed ones retained a portion; processing for longer than ten minutes eliminated activity under these conditions. The result shows the dependence of a specific measurement on preparation, but it does not establish that the same time threshold applies to any skillet, portion, or disease. Different heating rates further complicate the transfer of laboratory protocols to a home recipe.

Popular advice to let crushed garlic sit for a strictly defined time often sounds like a medical instruction. Chemically, the pause allows the reaction to continue before heating, but the platelet study did not compare heart attack rates following different kitchen pauses. To derive a therapeutic recommendation, one would have to link preparation to the actual exposure of the substance and then to a clinical outcome. In a home kitchen, it is more reasonable to choose for taste and tolerability than to perceive a stopwatch as a means of preventing thrombosis.

The next stage begins in the digestive tract. Environmental conditions and surrounding food are capable of changing the formation and subsequent transformations of compounds, so the stated yield of allicin in a laboratory system is not yet equal to the effect after ingestion. For some tablets, the coating is important as it changes the location and time of disintegration. To evaluate a product, one needs to know whether the researchers measured only its chemical potential or tracked metabolism in humans.

Lawson and Hunsaker conducted a series of bioassays in 13 healthy volunteers, studying 13 supplements and nine food forms. Over 32 hours, they measured allyl methyl sulfide—a metabolic marker—in exhaled air, comparing it with the response to crushed raw garlic. The assessments depended on the product and food conditions; this was a study of metabolic response without a placebo or blinding. The measured metabolite helps to trace transformations, but it does not in itself report the therapeutic concentration of allicin in the vascular wall.

Particularly revealing is a separate human trial on platelet function. In a crossover study, 18 healthy volunteers ate tzatziki with 4.2 g of raw garlic or a dish without it, and their blood was examined after five hours. No changes in platelet function were detected in the tests used. This small acute experiment does not establish the safety of any long-term high doses, but it shows why adding garlic extract to cells and an ordinary food portion can yield different results.

It follows that preparation cannot be evaluated on a single scale of "benefit preserved or lost." It changes the material, and the body additionally transforms what arrived with the food; a clinical study then answers yet another, separate question. For the reader, this frees garlic from the role of an mandatory raw medicine: the prepared product remains a culinary ingredient. For the researcher, the same sequence requires detailed descriptions of the form and mode so that the next experiment truly tests a comparable intervention.

Arterial pressure: from hydrogen sulfide to clinical effect

Arterial blood pressure depends on cardiac performance and the resistance of the vascular bed, as well as on the volume of circulating fluid and regulatory systems. The systolic value refers to the heart's contraction phase, while the diastolic value refers to the period between contractions. A change in vascular muscle tone can influence pressure, but it does not explain all causes of hypertension. Therefore, the potential vasodilatory effect of garlic compounds should be considered as one mechanism among many, rather than a universal correction for the disease.

Hydrogen sulfide, H₂S, is known as a gas with a pungent odor, yet in biology, it also participates in cellular signaling. In the work of Benavides, human erythrocytes converted certain garlic polysulfides into H₂S with the participation of thiol compounds. The researchers also observed the release of H₂S and the relaxation of isolated rat aortic rings. This combination of experiments involved human cells and animal tissue, rather than human blood pressure measurements taken after a garlic-containing meal.

The hypothesized chain includes the influence of H₂S on potassium channels of the vascular muscle. A change in potassium movement can shift the membrane's electrical state and reduce the conditions for contraction, allowing the vessel to relax. In the described experiment, H₂S production depended on reduced thiols, including glutathione, which links the effect to the specific chemistry of the cellular environment. This is a plausible physiological hypothesis; it does not establish which of the AGE components provides the blood pressure reduction in patients.

A direct clinical answer was sought in the double-blind AGE at Heart trial. A total of 104 people with inadequately controlled hypertension were randomized, most of whom were already taking antihypertensive medications; 95 completed the participation, and the main analysis included 88. Over the course of 12 weeks, 1.2 g of AGE per day, containing 1.2 mg of SAC, was studied against a placebo. The intergroup difference in systolic pressure was −5.0 ± 2.1 mmHg, where the value after the ± sign is the standard error; no significant difference in diastolic pressure was obtained.

For interpretation, it is important to distinguish between the overall result and a subgroup of "responders." If people are selected specifically because their blood pressure decreased, the average amount of decrease will naturally become larger and will no longer describe the entire initial group. Exclusions after randomization can also alter the comparability of groups. The trial was funded by the manufacturer Wakunaga: this does not invalidate the measurements, but together with the small sample size and attrition, it explains the need for independent replication.

Summarizing several trials provides a broader, yet still heterogeneous picture. The 2020 Ried meta-analysis included 12 studies and 553 participants with hypertension; the average reduction in systolic pressure relative to control was about 8.3 mmHg. For the diastolic indicator, eight studies with 374 participants yielded about 5.5 mmHg. These estimates refer to a selected hypertensive population and various preparations, rather than to the preventive consumption of garlic by people with normal blood pressure.

A broader meta-analysis by Behrouz and colleagues, published in the 2026 issue of Nutrition Reviews, combined 108 trials of various cardiovascular risk factors with 7,137 participants. For blood pressure, the combined difference was more modest: −3.71 mmHg systolic and −1.97 mmHg diastolic, with 95% confidence intervals of −5.07 to −2.36 and −2.86 to −1.08, respectively. Not all 7,137 people necessarily participated in every individual outcome analysis. The differences between the reviews show how much the average effect depends on the included populations and trials.

In practical work, this is an argument for discussing the supplement as a possible addition for a specific task, while pre-evaluating medications and the risk of bleeding. The result should be judged by repeated blood pressure measurements in comparable conditions, since a single reading does not describe a stable response. Furthermore, the given millimeters of mercury are not percentages of prevented strokes: such events were not the basis for calculating the effect here. Prescribed therapy should not be recalculated based on the average value from a garlic meta-analysis.

Lipids and glucose: why the average result does not become a promise

Cholesterol is transported through the blood within lipoproteins because it is poorly soluble in an aqueous environment. LDL delivers it to tissues, and the liver removes a significant portion of these particles via specific receptors; the content of LDL-cholesterol reflects one segment of a complex metabolism. It is influenced by the formation and utilization of particles, diet, and treatment. Therefore, the explanation involving "dissolving fat with garlic" is physiologically incorrect: the clinical question is whether the concentration of a specific lipoprotein parameter changes.

One of the most informative answers was provided by an experiment by Gardner and colleagues. It included 192 adults with moderately elevated LDL, with a baseline of 130–190 mg/dL, and assigned them to raw garlic, a powdered supplement, AGE, and a placebo. The intervention lasted six months, six days a week; the regimens were selected to correspond to approximately a 4 g clove, while controlling the chemical characteristics of the materials. Comparing several forms at once allowed for testing not only the hypothesis of benefit but also the assumption that the failures of previous trials were explained solely by the type of supplement.

No garlic form provided a statistically or clinically significant improvement in LDL relative to the placebo. No convincing advantage was found for other studied lipid parameters either. The negative result is especially useful for a person expecting that a daily clove will necessarily fix an analysis: such an expectation did not withstand a direct test. At the same time, the study describes its preparations, doses, and participants, and not all possible products labeled "garlic." Gardner Results.

In pooled data, a small positive average effect is encountered. A 2013 meta-analysis by Ried, Toben, and Fakler included 39 trials and, for people with elevated total cholesterol taking it for longer than two months, reported a reduction in total cholesterol of approximately 17 mg/dL and LDL of 9 mg/dL. A broader analysis by Behrouz gave −5.90 mg/dL for LDL, with a 95% interval from −10.51 to −1.29. These are concentration changes that cannot be converted into a measured frequency of prevented heart attacks specifically through the use of garlic.

Such coexistence of a positive average and negative large-scale experience does not require choosing only the convenient publication. A meta-analysis aggregates studies with different baseline disorders, preparations, durations, and concomitant therapy; the mean value describes this set, rather than guaranteeing a result for every participant. If the materials differ significantly, statistical aggregation does not create a unified clinical dose. For a specialist, the discrepancy becomes a reason to more carefully compare the patient with the studied population and product, and for a reader — not to buy a supplement based on a single promising number.

Regarding glucose, the baseline situation is also important. A 2024 review combined 22 randomized trials with 1,567 participants and interventions ranging from three weeks to a year, including different garlic forms. The mean difference in fasting glucose was −7.01 mg/dL, and for HbA1c it was −0.66 percentage points, with 95% intervals from −8.53 to −5.49 and from −0.76 to −0.55, respectively. For glucose, the results were highly heterogeneous: I² was about 95%, so a single mean poorly describes the predictability of an individual response.

HbA1c shows the proportion of hemoglobin bound to glucose and reflects a longer period than a single fasting measurement. A decrease of 0.66 percentage points means a change, for example, from 8.00% to 7.34%, if other conditions correspond to this conditional example; it is not a relative decrease of 0.66%. However, the example explains the units rather than predicting a result for a specific patient. The heterogeneous doses and timelines of the review do not allow deriving a universal garlic regimen or a basis to reduce glucose-lowering treatment from it.

The practical question is therefore better formulated through the goal of treatment and a repeatable indicator. If a person takes a supplement for lipids, the lipid profile should be evaluated, not the smell or subjective vitality; if for glucose — take into account medications and the baseline metabolic disorder. Several statistically significant markers cannot be added up into a summary "percentage of health." It is more important to understand whether a small additional effect changes the clinical decision for a given person and justifies the costs and interactions.

Infections and immune cells: three distinct questions about protection

The antimicrobial reputation of garlic has a real laboratory basis, but it should be described accurately. In Müller's work, allicin inhibited the growth of Escherichia coli and chemically modified the bacterium's proteins, causing a disruption in cellular function. Other experiments show that the sensitivity of species and strains varies: some bacteria are capable of withstanding much higher concentrations. These are in vitro studies of microorganisms, useful for finding future therapeutic approaches, but not clinical trials of consumed garlic.

To treat an infection, a substance must reach the target site in an active form, remain there, and be safe enough for human tissues. In a laboratory dish, bacteria are in direct contact with a specific concentration, whereas after ingestion, the substance passes through digestion and numerous reactions. Furthermore, the inhibition of bacterial growth does not answer the question of a viral cold. Therefore, the term "natural antibiotic" masks several skipped stages and hinders the understanding of which specific infection was investigated.

A separate human study by Nantz and colleagues focused on immune cells and well-being. In a double-blind trial, 120 healthy adults received 2.56 g of aged garlic extract (AGE) daily or a placebo for 90 days. After 45 days, the ability of NK cells and γδ-T cells to proliferate—that is, to multiply under assay conditions—was evaluated; for these cellular results, the authors indicated n=56. Some indicators were higher in the extract group; however, this is a description of a specific cellular response, not a universal assessment of the entire immune system.

Illness diaries provided a different picture: the frequency of colds and flu-like conditions did not differ in a statistically significant way. At the same time, the AGE group had 21% fewer reported symptoms and 58% fewer days of work or school missed due to illness. These relative differences relate to symptoms and days, not to the probability of getting sick. The clinical indicators were secondary and dependent on self-reporting; the study was funded by Wakunaga, which is important to consider alongside its scale.

Another frequently cited experiment was conducted by Josling in 2001. A total of 146 volunteers took an allicin-containing supplement (one capsule per day) or a placebo for 12 weeks, recording cold episodes in a diary. The authors reported 24 episodes in the supplement group versus 65 in the placebo group. These are numbers of registered episodes, including potential repeat illnesses, so they cannot be read as "24 sick people versus 65" or used to derive an absolute reduction in individual risk without additional information.

A 2014 Cochrane review found only one trial that met its inclusion criteria—the work of Josling. Despite a positive signal, the authors deemed the clinical data insufficient to draw a conclusion about the efficacy of preventing or treating the common cold. This is an example of how a review evaluates reproducibility and limitations rather than simply repeating the most impressive number. The composition and regimen of the single capsule studied also do not translate into a dose of raw garlic for the winter season.

For practical purposes, it is useful to separate three tasks: changing a cellular laboratory characteristic, reducing the severity of symptoms, and preventing disease. Each requires its own outcome and appropriate design; success in the first does not guarantee the other two. If a cold has already begun, the cited trials of long-term prophylactic intake do not establish the efficacy of an emergency increased dose. Garlic may remain part of a habitual diet, but the decision regarding assistance during an infection should be based on diagnosis and symptoms, not on the result of a dish with bacteria.

Gastric cancer: what long-term observation really showed

The history of gastric cancer requires a longer timeline than studies on blood pressure or the common cold. Chronic Helicobacter pylori infection can sustain mucosal inflammation, which, in some people, is followed by atrophy, intestinal metaplasia, and further changes. Atrophy means the loss of normal glands; metaplasia is the appearance of a different type of tissue; neither is equivalent to cancer that has already developed. This sequence helps explain why researchers evaluate the condition of the mucosa, new cases of the disease, and mortality separately.

The Shandong trial was conducted among residents of a Chinese region with a high risk of gastric cancer. The analysis included 3,365 people aged 35–64; a factorial design separately allocated interventions against H. pylori, a vitamin supplement, and a garlic preparation or their corresponding placebos. This allowed for the study of multiple exposures in one trial while maintaining their distinction. An early report after the intervention period found no favorable effect of the garlic supplement on the prevalence of precancerous changes or cancer incidence.

The garlic regimen included 800 mg of aged extract and 4 mg of steam-distilled oil per day for 7.3 years. Follow-up continued for a total of 22.3 years, meaning the supplement was not taken during the entire period. The combination of two preparations also makes it impossible to attribute the result solely to allicin, SAC, or a common dish. When reading a long-term report, one must keep the duration of the intervention, the duration of follow-up, and the technological composition of the product separate.

For gastric cancer incidence in the long-term analysis, an odds ratio of 0.81 was obtained with a 95% confidence interval of 0.57–1.13. Because the interval includes unity, a statistically convincing reduction in the number of new cases for the garlic intervention was not established. Furthermore, the odds ratio is not an absolute probability difference and does not mean that cancer was successfully prevented in 19 out of every hundred participants. This is the result of a specific comparative analysis for which both the point estimate and its uncertainty are important.

For gastric cancer mortality, the authors reported a favorable result: the hazard ratio was 0.66, with an interval of 0.43–1.00. The upper bound is rounded to unity, so the signal should be described as borderline, maintaining the authors' assessment of significance. Such an indicator compares the intensity of events over time and does not denote an absolute reduction in mortality by 34 percentage points. The main point is that the mortality result cannot be substituted for the missing convincing result of prevention of new cases.

Biologically, incidence and mortality may diverge because between the onset of a tumor and death lie stage, diagnosis, treatment, and clinical features. But the divergence itself does not prove which of these processes the garlic preparation altered. To explain the mechanism, additional measurements and reproduction of the result would be needed, rather than just a plausible story about cellular protection. The study leaves an interesting question, yet it does not provide grounds for prescribing multi-year supplements to any person concerned about oncological risk.

The applied conclusion shifts attention back to confirmed causes and individual risk. The NCI describes data according to which the elimination of H. pylori reduces the risk of gastric cancer in certain populations; the decision on screening and treatment depends on clinical indications. The garlic part of the Shandong trial was not a replacement for antibiotic intervention and does not prove independent eradication of the infection. For a patient, knowledge about the infection and the state of the mucosa is more useful than recalculating capsules from a high-risk region into daily cloves.

Safety: bleeding, medications, and home use

Tolerance begins with common complaints: odor, abdominal pain, flatulence, and nausea. NCCIH also points out the possibility of allergic reactions, so pronounced symptoms should not be automatically explained as "getting used to" a beneficial product. A food serving and a concentrated supplement create different conditions of use, and increasing the dose for the sake of an effect may increase discomfort. If garlic causes unpleasant sensations, there is no physiological obligation to eat it raw or endure irritation for the sake of an unproven benefit.

The most significant systemic risk is a potential increase in bleeding, especially when taken alongside anticoagulants or aspirin. Platelets are involved in the primary closure of vascular injuries, so the reduction of their activity is not equivalent to the colloquial term "blood thinning" and may have an undesirable side effect. However, a small pilot study with AGE in patients on warfarin did not find an increase in bleeding: 52 people were randomized, and 48 completed 12 weeks of taking 5 ml twice a day. This is the result of a specific product under observation, which is insufficient to exclude rare events or resolve all possible combinations.

Before surgery, it is necessary to inform the surgeon and anesthesiologist about the supplement, its composition, and the regular consumption of large amounts of garlic. The adopted EMA monograph contains a warning to avoid consuming garlic seven days before a surgical intervention due to the postoperative risk of bleeding. This is a regulatory precaution for the described herbal products; a specific plan should be coordinated with the medical team, taking into account the procedure and medications. Discontinuing a prescribed anticoagulant on one's own instead of discussing the supplement creates a different danger and does not follow from garlic research.

Black tarry stools, blood in the urine, vomiting blood, and prolonged, unstoppable bleeding require urgent medical attention. During anticoagulant therapy, pronounced unexplained bruising and increased menstrual bleeding are also alarming; such symptoms should not be dismissed as the effect of a "natural remedy." These signs describe potential bleeding in general, not a specific diagnosis of a garlic-related complication. When seeking medical help, it is important to state all medications and supplements so that the combination of effects is evaluated, not just the last product taken.

Interactions may also manifest as a decrease in drug concentration. In ten healthy volunteers receiving 1200 mg of Fortovase saquinavir three times a day during study periods, garlic tablets taken twice a day for days 5–25 reduced the mean area under the concentration-time curve by 51% and the maximum concentration by 54%. This is a relative change in drug exposure, not the proportion of patients with treatment failure. The study concerned a specific regimen; the EMA has separately contraindicated the combination of the described garlic products with saquinavir/ritonavir, so the issue of compatibility is decided based on the drug, not on the general "herbal" label.

Pregnancy and breastfeeding require a separate assessment: the NCCIH warns that amounts higher than those found in food may be unsafe. Data on adults with hypertension do not answer questions about fetal development, safety for an infant, or pediatric dosage. Even good tolerability in a small group does not establish the safety of rare outcomes. Therefore, research-level grams of extract should not be turned into self-prescription for a pregnant or breastfeeding woman or a child.

External use of raw garlic has its own history of damage, including chemical burns. A 2021 systematic review of published cases gathered 39 patients from 32 articles; second-degree burns were frequently described, and necrosis occurred in some instances. A case series does not allow for the calculation of the complication rate among all users, but it confirms its reality. Garlic paste should not be applied under a bandage or used on mucous membranes, including the vagina: the assumed antimicrobial activity does not make such an application a safe treatment.

Finally, garlic in homemade oil requires food safety, not pharmacological, caution. The NCHFP warns of the risk of botulism at room temperature and recommends preparing the mixture fresh, storing it at 40 °F or below, approximately 4.4 °C, for no more than four days, or freezing it. Garlic's antimicrobial reputation is no substitute for these conditions, and a commercial product should be evaluated according to its own instructions. In everyday life, garlic can well remain a pleasure of eating; a supplement, however, deserves a specific purpose, a comparable preparation, and a discussion of medications, as it is these details that determine the ratio of potential benefit and risk.