Garlic

What it is

Garlic (Allium sativum) is a common cooking bulb that also contains bioactive compounds studied for heart and metabolic effects [1]. Trials have tested garlic in several forms, including aged garlic extract [2] and a supplement with 180 mg of allicin [3].

How it might work

Proposed mechanisms, not yet established: garlic has antiplatelet properties, meaning it may make blood clot less readily, which is one reason it raises bleeding concerns [4]. Aged garlic extract is proposed to increase the activity of natural killer and gamma-delta T cells, which may account for part of the milder illness seen in one trial [2].

What the research says

Blood pressure has the most consistent results. A 2016 meta-analysis of 20 trials with 970 people found systolic blood pressure about 5.1 mm Hg lower on garlic than on placebo, and 8.7 mm Hg lower in people whose blood pressure was already high [5]. A 2026 meta-analysis of 108 trials with 7,137 adults found systolic about 3.7 mm Hg lower and diastolic about 2.0 mm Hg lower [6]. A 2025 meta-analysis of 10 trials in people with hypertension found similar drops [7]. A 2015 meta-analysis of 17 trials found the systolic drop only in people with high blood pressure, not in people with normal readings [8].

Garlic also modestly lowers blood fats. The 2026 meta-analysis found total cholesterol about 10 mg/dL lower and LDL about 6 mg/dL lower, with HDL about 2 mg/dL higher [6]. A 2024 meta-analysis found total cholesterol 14 mg/dL lower and LDL 8 mg/dL lower [9]. Triglyceride results disagree: the 2026 analysis found a small drop [6], while the 2024 analysis found no substantial change [9]. An umbrella review of nine meta-analyses found total cholesterol fell in eight of them, while HDL, LDL and triglyceride results were inconsistent [10].

Fasting blood sugar also fell in several meta-analyses. A 2024 analysis found fasting glucose 7.0 mg/dL lower and HbA1c lower by 0.66 [9]. A 2019 analysis of 33 studies in people with diabetes found fasting blood sugar 10.9 mg/dL lower on garlic than on placebo [11]. A 2015 analysis of 7 trials with 513 people also found lower fasting glucose [12].

For colds, one trial of 120 healthy adults found that aged garlic extract (2.56 g a day for 90 days) did not significantly change how many colds and flu illnesses people had, but the illnesses were milder: 21% fewer symptoms and 58% fewer work or school days missed [2]. The same trial found that natural killer and gamma-delta T cells multiplied better after 45 days [2]. A Cochrane review found only one trial of garlic for colds that met its criteria: 146 people taking a 180 mg allicin supplement for 12 weeks had 24 colds, against 65 on placebo, but the review concluded the evidence is insufficient [3].

In an observational analysis of 21 studies with more than 540,000 people, those who ate the most Allium vegetables (onion, garlic and relatives) had lower odds of stomach cancer than those who ate the least (odds ratio 0.54) [13]. Most were case-control studies, which compare people with and without cancer, so this is linked to lower risk rather than proof that garlic lowers it [13].

Doses used in studies

These are doses from trials, not recommendations. Aged garlic extract: 2.56 g a day for 90 days [2]; 5 mL twice a day for 12 weeks, alongside warfarin [4]. Allicin-standardised supplement: 180 mg once a day for 12 weeks [3].

Safety and interactions

Common side effects: garlic breath and body odour are the best-documented complaints [14][15]. Gastrointestinal discomfort and bad breath were the most common side effects in a 2025 meta-analysis, with adverse reactions slightly more frequent on garlic but not significantly so [7]. Allergic contact dermatitis, urticaria and anaphylaxis have been reported in case reports [14].

Bleeding and blood thinners: garlic has antiplatelet properties [4]. In a 12-week trial of aged garlic extract with warfarin, 48 people completed the study and none had increased bleeding [4]. Case reports describe excessive surgical bleeding in people eating about 12 g of garlic a day [16] and in a man taking garlic and fish oil supplements before bypass surgery [17]. A review concludes garlic is strongly associated with surgical bleeding independent of anticoagulants [18]. Garlic may enhance the effect of warfarin and fluindione [14]. Anaesthesia guidance recommends stopping herbal medicines one to two weeks before elective surgery [19].

Other medicines: garlic may reduce the effectiveness of the HIV drug saquinavir [14].

Upper limits: no upper intake limit was found in the studies read. A case of bleeding at about 12 g a day [16] is not a safe limit.

Pregnancy and breastfeeding: no trials of garlic supplements in pregnancy were found. Garlic flavour passes into breast milk, and infants can detect it within hours of a single maternal meal [20]. Garlic eaten by nursing mothers has been reported to change infant behaviour during breastfeeding [14].

Kidney, liver and bleeding disorders: no studies in people with kidney or liver disease, or with bleeding disorders, were found in the searches. The bleeding evidence above is the closest guide for people with bleeding problems.

Caveats and open questions

Most trials are small and short, and the blood pressure results vary: diastolic results in one 2015 meta-analysis had high heterogeneity, meaning the trials disagreed [8]. An earlier review of trials found cholesterol reductions at 1 and 3 months but not at 6 months, and found no effect on glucose; it described the trials as of marginal quality and noted that the active compounds in preparations were not well defined [15]. A Cochrane review found that claims about colds rely largely on poor-quality evidence [3]. One Cochrane author received support for training in review methods, not for that review [3].

The forms differ (raw garlic, aged extract and allicin-standardised supplements), so results do not transfer neatly from one product to another. The meta-analyses read here pooled blood pressure, blood fats and blood sugar; their authors call for long-term trials on heart disease and death [5]. The cold and immune trials read here ran for 90 days or less [2][3].