Honey is nectar that bees have worked over. Nectar is typically 70 to 80% water, and bees fan it and add enzymes until the finished honey holds about 18% water. A bee’s honey stomach holds only about 40 mg of liquid, and filling it can take visits to more than a thousand flowers.
Nutritionally, honey is sugar. A tablespoon (21 g) has 64 calories and 17.2 g of sugars, mostly fructose and glucose, and no vitamin or mineral reaches 5% of a daily value. Our nutrient density score for honey is 0 because it counts all of honey’s sugar as added sugar. That fits how WHO treats it, as free sugars, and how the FDA lists it, under added sugars.
Honey also has a hard age rule: never give it to a baby under 12 months. It can carry spores of the bacterium behind infant botulism, and a 1979 California study found them in 9 of 90 honey samples. For older children, honey is one of the better-studied cough remedies, and manuka honey is used on wounds as a medical dressing, not as food. The results below are often mixed.
Health benefits of honey
The top findings from 27 studies, strongest evidence first
What the stars mean
- Strong. Several randomized trials, pooled in a meta-analysis, point the same way, or regulators (FDA or EFSA) have approved the claim.
- Moderate. Good randomized trials support it, but there are fewer of them, or their results vary.
- Limited. Small, short or single trials, or mixed results. Promising, but it needs confirmation.
Population and lab studies get no stars: they show a link or a mechanism, not proof.
How we rate evidence →- ★★Probably eased children’s cough, not adults’ bronchitisAcross 6 trials (899 children), honey probably cut cough frequency by 1.05 points against no treatment and 1.62 against placebo; most had one night’s use. In 194 adults with acute bronchitis, honey did not shorten the cough compared with usual care.
- ★★Mixed results on blood sugar and cholesterolOne pooling of 18 trials found honey lowered fasting glucose by 0.20 mmol/L and LDL cholesterol by 0.16 mmol/L, mostly with low certainty. A pooling of 69 trials found 10 g a day may lower HbA1c but may worsen blood pressure, triglycerides and fasting glucose.
- ★★Faster burn healing with honey dressingsIn a Cochrane review, honey dressings (on the skin, not eaten) healed partial-thickness burns 4.68 days faster than conventional dressings, based on 2 trials. For leg and diabetic foot ulcers the effects were unclear, and manuka honey dressings did not significantly help leg ulcer healing in 368 people.
- ★★Less severe mouth sores during radiotherapyIn 7 trials of 412 head and neck cancer patients, honey made severe mouth sores much less likely (relative risk 0.22). It did not lower how often mouth sores occurred overall, and the authors call the evidence limited.
- ★★No less pain after tonsil removal in childrenIn 400 children after tonsil removal, 5 mL of Marri or manuka honey did not reduce pain compared with standard care or placebo. In practice children took only 2 to 3 doses a day.
All 27 studies, strongest evidence first
In a Cochrane review, honey probably eased children’s cough better than no treatment or placebo, but the authors found no strong evidence for or against it
Review of 6 randomized trials (899 children aged 12 months to 18 years) of honey for acute cough. On a 7-point scale, honey cut cough frequency by 1.05 points against no treatment and 1.62 against placebo (moderate-certainty evidence). It looked similar to dextromethorphan (low certainty), and beyond 3 days it had no advantage over placebo. Most children were treated for only one night.
Honey eased cough and cold symptoms better than usual care in a meta-analysis of 14 studies
Meta-analysis of 14 studies of upper respiratory tract infections. Compared with usual care, honey improved cough frequency (a standardized difference of 0.36) and cough severity (0.44). The two studies against placebo showed no clear difference in combined symptoms. Risk of bias was moderate, the authors call for better placebo-controlled trials, and one author keeps bees as a hobby.
Honey lowered fasting glucose by 0.20 mmol/L and LDL cholesterol by 0.16 mmol/L in a meta-analysis of 18 trials
Meta-analysis of 18 controlled trials (1,105 people) that lasted at least 1 week. Triglycerides fell 0.13 mmol/L and HDL rose 0.07 mmol/L. The certainty of the evidence was low for most results and high only for HDL. The clearest effects were seen with robinia, clover and raw honey.
Honey dressings healed partial-thickness burns 4.68 days faster than conventional dressings, but the evidence beyond burns is weak
Cochrane review of 26 trials (3,011 people) of honey put on wounds. This is topical use, not eating honey. The burn result rests on 2 trials (992 people). For venous leg ulcers and diabetic foot ulcers the effects were unclear (low-quality evidence), and the authors call most other evidence low or very low quality. Some authors ran a trial of honey dressings that received a small cash gift from a maker of honey dressings.
A review of 69 trials found that 10 g of honey a day may lower HbA1c but may worsen blood pressure, triglycerides and fasting glucose
Overview of reviews with an updated meta-analysis of 69 randomized trials (3,544 people), including a dose analysis. The authors advise caution with long-term or high-dose honey in people with health concerns. The result differs from the 2023 meta-analysis above, so the effect of honey on blood sugar and blood pressure is not settled.
In adults with acute bronchitis, honey did not shorten the cough compared with usual care
Open randomized trial in 194 adults in primary care in Catalonia that compared honey, dextromethorphan, an ipratropium inhaler and usual care for up to 14 days. The median number of days with moderate or severe cough was 6 with honey and 5 with usual care. The trial missed its planned size because of COVID-19, so it may have been too small to see a modest effect.
Manuka honey dressings did not significantly improve healing of venous leg ulcers at 12 weeks compared with usual care
Open-label randomized trial in 368 people, who got calcium alginate dressings with manuka honey or usual care, with compression bandaging for all. This was topical use. At 12 weeks 55.6% of ulcers had healed with honey and 49.7% with usual care. Honey was probably more costly and linked to more adverse events (relative risk 1.3).
Honey reduced the severity of radiotherapy mouth sores in head and neck cancer but did not lower how often they occurred
Meta-analysis of 7 trials (412 patients) of honey during radiotherapy. Severe oral mucositis was much less likely (relative risk 0.22) and fewer treatments were interrupted, but the overall rate of mucositis did not drop significantly (relative risk 0.69). The authors call the evidence limited and ask for larger trials.
Honey after tonsil removal did not reduce children’s pain compared with standard care or placebo
Double-blind trial in 400 children at 3 centers in Western Australia. Children took 5 mL of Marri honey, manuka honey or placebo six times a day for at least 7 days, but in practice took only 2 to 3 doses a day. Pain scores followed similar paths in all groups. Vomiting was more common in the honey and placebo groups in the early days, and 20% were lost to follow-up or withdrew.
Honey dressings shortened chronic wound healing time by 17.13 days in a meta-analysis of 8 studies, but the evidence was very low quality
Meta-analysis of 8 studies (906 people) of honey dressings on chronic wounds, which is topical use. Healing rate, time to clear bacteria and hospital stay did not differ significantly, and honey may cause more painful discomfort during treatment. The authors cite risk of bias, inconsistency and publication bias.
One dose of buckwheat honey before bed eased children’s night cough more than no treatment in a single trial
Trial in 105 children aged 2 to 18 with colds, given a single dose of buckwheat honey, honey-flavored dextromethorphan or nothing 30 minutes before bed. Parents, who were only partly blinded, rated honey best. Honey beat no treatment for cough frequency, while dextromethorphan did not, and honey and dextromethorphan did not differ significantly.
Two weeks of 50 g of carbohydrate a day from honey, sugar or corn syrup gave similar effects on blood sugar, blood fats and inflammation
Crossover trial in 55 adults, 28 with normal glucose tolerance and 27 with impaired glucose tolerance, comparing honey, sucrose and high-fructose corn syrup. All three raised triglycerides in both groups, and raised blood sugar and inflammation responses in the people with impaired glucose tolerance. Body weight did not change.
Clover honey instead of sugar for 4 weeks: triglycerides stayed raised with sugar but not with honey
Crossover trial in 37 adults, who ate clover honey or sucrose (1.2 g of carbohydrate per kg of body weight a day) in two 4-week periods. Total, LDL and HDL cholesterol did not change on either. People ate less energy, carbohydrate and sugars during the honey period, and body weight rose briefly in the first week on honey.
70 g of honey a day lowered total cholesterol, LDL and triglycerides in young adults, while the same amount of sugar raised them
Double-blind trial in 60 healthy people aged 18 to 30, comparing 70 g of honey with 70 g of sucrose a day. HDL rose with honey and fell with sugar. The abstract gives no figures for the size of the changes, and the authors ask for more trials.
70 g of honey a day for up to 30 days lowered triglycerides by 11% and LDL cholesterol by 5.8% in overweight people with normal values
Trial in 55 overweight or obese people, 38 on honey and 17 on 70 g of sucrose a day. Honey also lowered total cholesterol 3%, fasting glucose 4.2%, CRP 3.2% and body weight 1.3%, and raised HDL 3.3%. The groups were uneven in size, and the abstract does not say how blinding was done.
75 g of honey a day for 14 days did not lower LDL cholesterol in people with high cholesterol
Trial in 60 people with high cholesterol, split by sex and statin use, who took a honey solution or a similar sugar solution. Neither changed cholesterol or triglycerides in the whole group. In women, LDL rose with the sugar solution but not with honey.
In type 2 diabetes, 50 g of honey a day for 8 weeks raised HbA1c by 0.17% while the control diet lowered it by 0.22%
Crossover trial in 42 people with type 2 diabetes who finished (53 started), with 8 weeks on each diet and a 4-week washout. The rise on honey was not significant by itself, but the difference between the two conditions was. Fasting glucose did not change, and waist circumference fell with honey.
In type 2 diabetes, 8 weeks of honey lowered cholesterol and body weight but raised HbA1c
Trial in 48 people with type 2 diabetes, who got honey or no honey. Fasting blood sugar did not differ between groups, LDL cholesterol and triglycerides fell and HDL rose. The authors advise cautious use of honey in people with diabetes because of the HbA1c rise. The abstract does not give the dose.
Eating buckwheat honey raised plasma antioxidant capacity in healthy adults compared with corn syrup
Healthy adults ate 1.5 g per kg of body weight of corn syrup or of buckwheat honey with two different levels of phenols (0.79 and 1.71 mg per gram). Plasma phenols, antioxidant capacity and reducing capacity rose after both honey doses. This is a blood test, not a health outcome.
Birch pollen honey before the season cut allergy symptoms by 60% compared with usual medication alone
Randomized pilot study in 44 people with birch pollen allergy who ate birch pollen honey (pollen added to honey) or regular honey daily from November to March, plus 17 on usual medication alone. The two honey groups did not differ significantly, so the study cannot separate the pollen from the honey. The authors call the results preliminary.
A honey and coffee paste cut cough scores from 2.9 to 0.2 in adults with a cough that lingered after a cold
Double-blind trial in 97 adults with a cough of more than 3 weeks after a cold. Scores fell from 3.0 to 2.4 on prednisolone and from 2.8 to 2.7 on guaifenesin. The paste also contained coffee, so the result cannot be credited to honey alone.
Honey 2 to 6 times a week went with lower odds of fatty liver (0.86), while daily honey showed no link
Cross-sectional study of 21,979 adults in Tianjin, China, using a food frequency questionnaire. The odds ratio was 1.10, not significant, for honey at least once a day. A snapshot like this cannot show cause, and the pattern did not rise or fall steadily with intake (P for trend 0.90).
Among 2,943 infant botulism cases from 2007 to 2021, 3.8% of US patients had eaten honey, against 19.7% elsewhere
Worldwide summary of laboratory-confirmed infant botulism from surveillance, hospital inquiries and the medical literature. Patients outside the US were more likely to need a breathing tube (50.3% against 21.2%). The authors say pediatricians should keep telling parents about honey as one risk factor, so honey is one avoidable source, not the only one.
Botulism spores were found in 10% of 90 honey samples, and honey was linked to infant botulism
Investigation of infant botulism cases in California. Botulism bacteria, but no ready-made toxin, were found in 6 honey samples fed to 3 patients. In California 29.2% (12/41) of hospitalized babies had been fed honey before symptoms began, and worldwide 34.7% (28/75). The authors concluded that honey should not be fed to infants.
In the lab, manuka honey disrupted biofilms of three common wound bacteria and kept them from sticking to human skin cells
Test-tube and cell experiments, not people. Manuka honey caused extensive cell death in biofilms of Staphylococcus aureus, Pseudomonas aeruginosa and Streptococcus pyogenes. Doses too low to kill the bacteria still cut their adhesion to wound proteins and to human skin cells.
Adding methylglyoxal to ordinary honey matched the biofilm-killing power of manuka honey in the lab
Test-tube experiment on 5 strains of Staphylococcus aureus. Honey solutions with at least 0.53 mg/mL of methylglyoxal killed biofilms. The manuka honey tested held 790 mg/kg, and the authors conclude methylglyoxal is only partly responsible for manuka honey’s effect.
In rats, a fructose-rich honey diet raised insulin resistance but not liver enzymes over 120 days
Animal study in 44 rats fed a normal diet, a high-fructose diet, high-fructose fruits or honey. Liver enzymes rose only on the high-fructose diet, and the liver tissue showed no fatty changes in any group. This is not proof for people.
Who should be careful with honey
Most people can eat honey in small amounts. Take a little care if one of these applies:
- Babies under 12 months. The CDC and the NHS both say not to give honey to a child under 1 year old, because it can carry spores that grow in a baby’s gut and make a toxin. Honey is one avoidable source of the spores, not the only one.
- Diabetes or a sugar limit. Honey raises blood sugar like other sugars. In two trials in type 2 diabetes, 8 weeks of honey raised HbA1c, so count it as carbohydrate and ask your care team how much fits.
- Irritable bowel syndrome. Monash University lists honey as an excess-fructose, high-FODMAP food. If fructose upsets your gut, test small amounts.
How to choose, store and eat honey
- Raw or processed: in a 2023 meta-analysis, raw, clover and robinia honey showed the clearest drops in fasting glucose and total cholesterol, but the certainty was low.
- In place of sugar: in a 2-week crossover trial, 50 g of carbohydrate a day from honey, sugar or corn syrup had similar effects on blood sugar and triglycerides.
- For a child’s cough: only over 12 months old, and in the Cochrane review honey had no edge over placebo beyond 3 days.
Buy with some care, because honey is a known target for fraud: in the EU’s 2021 to 2022 “From the Hives” action, 147 of 320 imported consignments (46%) were suspected of containing added sugar syrups, though the Commission says the methods are not yet fully validated and one simple test cannot prove adulteration. Keep honey at room temperature on the counter or a pantry shelf, since the fridge speeds up crystallization, which does not mean the honey has spoiled. To make crystallized honey runny again, stand the jar in warm water and stir.
Sources
- Botulism prevention. CDC.
- Foods to avoid giving babies and young children. NHS.
- WHO calls on countries to reduce sugars intake among adults and children. World Health Organization. 2015.
- Added sugars on the Nutrition Facts label. US Food and Drug Administration.
- Johnson RK, Appel LJ, Brands M, et al. Dietary sugars intake and cardiovascular health: a scientific statement from the American Heart Association. Circulation. 2009.
- Label reading and FODMAPs. Monash FODMAP blog.
- Food fraud: how genuine is your honey?. European Commission Joint Research Centre. 2023.
- How should I store honey?. Ohio State University Extension, citing National Honey Board guidelines. 2016.
- Tanzi MG, Gabay MP. Association between honey consumption and infant botulism. Pharmacotherapy. 2002.
History of honey
People ate honey long before they kept bees. A rock painting in a cave in Valencia, Spain, dated to at least 8,000 years ago, shows two honey hunters gathering comb from a wild nest. Bee products were already in use by the first farmers: beeswax fingerprints turned up in pottery from Neolithic sites across Europe, the Near East and North Africa, going back to the seventh millennium BC.
The oldest known honey remains come from clay vessels in a tomb in Georgia, dated to between 4,700 and 5,500 years ago, where several kinds of honey were buried with a person. In Egypt, bee images date to the Old Kingdom, about 2400 BC, and the earliest written records of beekeeping come from there. Honey sweetened cakes, and bees were kept at temples for offerings.
Beekeeping was serious business in ancient Greece. In 594 BCE the lawmaker Solon required hives to stand 300 feet (90 meters) from those of another beekeeper.
Honey has also been a weapon. In the Third Mithridatic War, forces from the Kingdom of Pontus are said to have wrecked a Roman column by leaving “mad honey” in its path. The honey held grayanotoxins, which come from rhododendron nectar and cause dizziness, low blood pressure and slow heartbeat. Poisoning is rarely fatal, and the same “mad honey” problem still shows up around the Black Sea.
The danger to babies was found much later. Infant botulism was first recognized in the United States in the late 1970s. In 1979 a California team found its bacteria in honey fed to patients and concluded that honey should not be given to infants, advice that the CDC has since adopted.
Sources
- Honey. Wikipedia.
- Roffet-Salque M, Regert M, Evershed RP, et al. Widespread exploitation of the honeybee by early Neolithic farmers. Nature. 2015.
- Turner MD. Mad honey and the Poisoner King: a case of mass grayanotoxin poisoning in the Roman military. Cureus. 2023.
- Jansen SA, Kleerekooper I, Hofman ZL, et al. Grayanotoxin poisoning: “mad honey disease” and beyond. Cardiovasc Toxicol. 2012.
- Arnon SS, Midura TF, Damus K, et al. Honey and other environmental risk factors for infant botulism. J Pediatr. 1979.
- Tanzi MG, Gabay MP. Association between honey consumption and infant botulism. Pharmacotherapy. 2002.
Frequently asked questions about honey
Is honey healthier than sugar?
Only slightly, if at all. In a 2-week crossover trial of 55 people, 50 g of carbohydrate a day from honey, sugar or corn syrup had similar effects on blood sugar, triglycerides and inflammation, and a 2023 meta-analysis of 18 trials found small drops in fasting glucose (0.20 mmol/L) and LDL (0.16 mmol/L) with low certainty. A tablespoon (21 g) holds 64 calories and 17.2 g of sugars, and WHO and the FDA both count honey as free or added sugar.
Why can’t babies have honey?
Honey can carry spores of the bacterium that causes infant botulism, which grow and make a toxin in a baby’s gut. The CDC and the NHS both say no honey under 1 year old. A 1979 study found these bacteria in 10% of 90 honey samples, and 29.2% of hospitalized California babies had been fed honey. Honey is one avoidable source, not the only one: 3.8% of US patients between 2007 and 2021 had eaten honey.
Does honey help a cough?
For children over 12 months, probably a little. A Cochrane review of 6 trials (899 children) found honey probably eased cough better than no treatment or placebo, with moderate certainty, but no strong evidence for or against it, and no edge over placebo beyond 3 days. In 194 adults with acute bronchitis, honey did not shorten the cough. Never give it to a baby.
Is raw honey better than regular honey?
The evidence is thin. In a 2023 meta-analysis, raw, clover and robinia honey showed the clearest drops in fasting glucose and total cholesterol, but the certainty was low. No study here shows that raw honey is safer or more nutritious. Raw or not, honey is not safe for babies under 12 months.
How much honey can you eat a day?
No study here sets an ideal amount, so treat honey as added sugar. WHO advises less than 10% of calories from free sugars and says below 5%, about 25 g or 6 teaspoons, brings extra benefit. One tablespoon of honey has 17.2 g of sugars, which is about two-thirds of that lower limit.
Is honey okay for people with diabetes?
It can fit in small amounts, but honey raises blood sugar like other sugars, so count it as carbohydrate. In a crossover trial of 42 people with type 2 diabetes, 50 g a day for 8 weeks raised HbA1c by 0.17%, while the control diet lowered it by 0.22%. A 2023 meta-analysis found lower fasting glucose overall, so the evidence is mixed. Ask your care team how much fits.
Does manuka honey heal wounds?
Medical dressings with honey may help some wounds, but the evidence is weak. A Cochrane review of 26 trials found honey dressings healed partial-thickness burns 4.68 days faster than conventional dressings, but effects on leg ulcers and diabetic foot ulcers were unclear. In a trial of 368 people with venous leg ulcers, 55.6% healed with manuka honey dressings and 49.7% with usual care, a difference that was not significant. Lab studies show manuka honey disrupts bacterial biofilms.
Is store-bought honey fake?
Some of it may be diluted, but suspicion is not proof. In the EU’s “From the Hives” action, 147 of 320 imported honey consignments (46%) were suspected of containing added sugar syrups, now mostly made from rice, wheat or sugar beet. The Commission says the methods are not yet fully validated, so a single test cannot confirm fraud.