Ghee: three short trials, and two different fats under one name

The highest graded evidence about ghee is a 2026 systematic review of 7 randomised trials of animal-based foods in adults aged 18 to 65 with fatty liver disease, which linked ghee intake with improved total cholesterol and alkaline phosphatase. Inside that review ghee rests on a single trial, no pooled estimate is given, and the effects across the different foods were heterogeneous, so the grade is stronger than the finding. The trial that tested ghee most carefully points the other way. In a small crossover trial, 30 healthy men and women living at home ate ghee for 4 weeks in place of an equal amount of olive oil, matched for energy. Fasting apolipoprotein B rose in that small trial by 0.09 g/l (95% CI 0.02 to 0.17) and non-HDL cholesterol by 0.53 mmol/l (95% CI 0.01 to 1.05), against olive oil rather than against no added fat. LDL cholesterol rose 0.29 mmol/l (95% CI -0.05 to 0.63) and did not reach significance, which is a different statement from doing nothing. Two Indian trials fed ghee at 10 percent of energy for 8 weeks against mustard oil. In 63 healthy, young, physically active volunteers total cholesterol rose by 4 weeks and stayed up at 8 weeks. HDL cholesterol rose with it, so the ratio of the two did not change, and neither the effect sizes nor the confidence intervals appear in the only text available for that trial. In 24 young people on a lactovegetarian diet no lipid measure moved at any time point. Even so, 2 of the 11 in the ghee arm gained more than 20 percent in total cholesterol against 1 of 13 controls, and an arm of 11 people is too few to read an absent effect as an absence. One label question belongs at the top rather than in the footnotes. In South Asia the word ghee covers two different fats, and vanaspati ghee is partially hydrogenated vegetable oil that a review of the food supply and food regulations of Pakistan names as one of the major sources of industrial trans fat in Pakistan, next to bakery shortening and hard margarines. That review looked at product labels and food regulations, and it says nothing about the trans fat in dairy ghee, which is a separate question nobody in this cohort answered.

How much is actually in it

• Ghee was the richest source of retinol among South Asian foods analysed for a UK dietary assessment, at 968 micrograms per 100 g. — ethnic foods eaten by South Asians in the UK, analysed with food composition data Evidence D source
  Ghee 968 micrograms retinol per 100 g, the highest retinol value among the foods listed

What a real portion looks like

• Among 7049 Indian factory workers and their families, daily ghee consumption was most common in the urban group, intermediate in migrants and lowest in the rural group. — 7049 participants (41.6% female) from factory sites in Bangalore, Lucknow, Nagpur and Hyderabad: urban, migrant and rural groups (n = 7049) Evidence C source
  Proportion consuming ghee daily highest urban, intermediate migrant, lowest rural, P at or below 0.01; no grams per day reported

What cooking and storage change

• In a Pakistani case-control study, people who reused ghee for cooking had 2.45 times the odds of colorectal cancer, which is an odds ratio and not a risk. — colorectal cancer cases and one matched control per case, Khyber Pakhtunkhwa, Pakistan; questionnaire on diet, addiction and habits Evidence C source
  Reusing ghee OR 2.45, P = 0.0001; reusing oil OR 4.30; fast food OR 3.0; no confidence intervals reported in the abstract

What it does to your health

• In 30 healthy adults, 4 weeks of ghee instead of an equal amount of olive oil raised fasting apolipoprotein B by 0.09 g/l (95% CI 0.02 to 0.17) and non-HDL cholesterol by 0.53 mmol/l (95% CI 0.01 to 1.05). — 30 healthy men and women, free-living outpatients, two 4-week periods with a 2-week wash-out (n = 30) Evidence B source
  Apo B +0.09 g/l (95% CI 0.02, 0.17), P = 0.018; non-HDL cholesterol +0.53 mmol/l (95% CI 0.01, 1.05), P = 0.046; LDL cholesterol +0.29 mmol/l (95% CI -0.05, 0.63), P = 0.092, not significant
• In 63 healthy, young, physically active adults, ghee supplying 10% of energy for 8 weeks raised total cholesterol and HDL cholesterol together, so the total to HDL ratio did not change. — 63 healthy, young, physically active adults (52 male, 11 female); ghee group 30, mustard oil control group 33; 8 weeks (n = 63) Evidence B source
  Significant rise in serum total cholesterol at 4 weeks that persisted at 8 weeks, with a similar rise in HDL cholesterol, so the total cholesterol to HDL ratio showed no significant change
• In 24 healthy young vegetarians, ghee supplying 10% of energy for 8 weeks changed no serum lipid measure, although 2 of 11 people in the ghee group had a rise in total cholesterol above 20%. — 24 healthy young volunteers on a lactovegetarian diet; ghee group 11, mustard oil control group 13; 8 weeks (n = 24) Evidence B source
  No significant change in serum lipid profile at any time point; 2 of 11 in the ghee group and 1 of 13 controls had more than 20% rise in total cholesterol at 8 weeks
• A 2026 systematic review of 7 randomised trials of animal-based foods in adults with fatty liver disease linked ghee intake with improved total cholesterol and alkaline phosphatase, and that link rests on 1 trial. — adults 18 to 65 years with metabolic dysfunction-associated steatotic liver disease; 7 randomised trials of animal-based foods, of which 1 used ghee (n = 7 trials) Evidence A source
  Ghee intake linked with improved total cholesterol and alkaline phosphatase; no pooled estimate given, effects across foods were heterogeneous
• In 500 Pakistani adults, 250 with coronary artery disease and 250 controls, eating ghee showed no significant association with the disease while chicken, beef, eggs and junk food did. — 500 adults in Lahore, Pakistan: 250 coronary artery disease patients (mean age 57.8 years) and 250 age and gender matched healthy controls (n = 500) Evidence C source
  No significant association between ghee consumption and coronary artery disease; significant associations for chicken, beef, eggs and junk food
• In 210 Bangladeshi adults, ghee and butter oil carried lower odds of coronary artery disease than junk food, which had an odds ratio of 5.49 (95% CI 2.25 to 13.38). — 210 adults: 105 hospitalised coronary artery disease patients and 105 healthy urban controls, Dhaka, Bangladesh (n = 210) Evidence C source
  Junk food OR 5.49 (95% CI 2.25 to 13.38), chicken OR 4.54, beef OR 2.68, eggs OR 2.38, fish OR 2.81; ghee and butter oil grouped together had lower odds with no separate value reported

Safety, limits and interactions

• No agency sets a limit for ghee itself, so the number that bounds it is the WHO ceiling of 10% of total energy intake for saturated fat, a strong recommendation for adults and children. — adults and children, general population Evidence E source
  Ceiling of 10% of total energy intake for saturated fatty acids, strong recommendation; a further reduction below 10% is a conditional recommendation

What people believe that the data does not support

• The word ghee covers two different fats on labels in South Asia, and vanaspati ghee is partially hydrogenated vegetable oil that a review of the food supply and food regulations of Pakistan names as a major source of industrial trans fat there. — food supply of Pakistan, industrially produced foods and product labels Evidence E source
  Vanaspati ghee, bakery shortening, hard margarines and fat spreads identified as the major sources of trans fatty acids; existing national limits judged insufficient

Who this works differently for

• Two children who were force-fed ghee as a cultural practice developed lipoid pneumonia with destruction of the alveolar septa, and their chest infection did not respond to antimicrobial therapy. — 2 children, Saudi Arabia, after forced feeding with animal fat (ghee) (n = 2) Evidence D source
  Collapse and consolidation of the right middle and left lower lobes on X-ray and CT; thickening and destruction of alveolar septa on electron microscopy

What is still unknown

• Everything known about ghee and blood lipids comes from short trials in young healthy people, and the authors of the largest of them, 63 volunteers over 8 weeks, say the result cannot be carried over to older or hyperlipidaemic people. — 63 healthy, young, physically active adults, 8 weeks; no trial found in people with high cholesterol, in older people or over longer than 8 weeks (n = 63) Evidence B source
  Authors state more studies are needed in older subjects, hyperlipidaemic subjects and people with less healthy lifestyles before extrapolating

The bottom line

There is no USDA record for ghee at all, so the composition block on this page is empty and the one number we have for what is in it comes from elsewhere. Ghee held the highest retinol value, 968 micrograms per 100 g, among South Asian foods analysed for a UK dietary assessment, a figure taken from food composition tables rather than from measuring jars on a shelf. That is per 100 g of fat, which is nobody’s portion. Frequency is the only intake figure we found. Among 7049 factory workers and their families at four Indian sites, daily ghee eating was most common in the urban group, intermediate in migrants and lowest in the rural group (P at or below 0.01), by food frequency questionnaire. It recorded how often and never how much. Reheating is where the one unpleasant number sits. In a Pakistani case-control study, people who reused ghee for cooking had 2.45 times the odds of colorectal cancer (P = 0.0001), an odds ratio rather than a risk, reported with no confidence interval and no stated sample size. Reusing frying fat travelled together with reused oil (OR 4.30) and fast food (OR 3.0), so ghee was never isolated from them. Two case-control studies asked instead about heart disease, and neither produced a significant risk estimate for ghee itself. Among 500 adults in Lahore, 250 with coronary artery disease and 250 matched controls recruited by convenience with hospital attendants as controls, ghee showed no significant association with the disease while chicken, beef, eggs and junk food did. Among 210 adults in Dhaka, 105 hospitalised patients against 105 community controls, junk food carried an odds ratio of 5.49 (95% CI 2.25 to 13.38) while ghee was pooled with butter oil in the low-odds group and given no value of its own. Neither study measured amounts, and a missing association in one hospital sample is not a clean bill. Nobody regulates ghee by name. No agency sets a limit for it, the FDA claim rules, the EU claims register and the NIH office of supplements hold nothing under the word, and the number that bounds it is the WHO ceiling of 10 percent of total energy from saturated fat. That is a strong recommendation for adults and children, graded by WHO itself on moderate certainty evidence, and it is a statement about saturated fat rather than about ghee. The only human harm report we found is narrow, and worth naming for what it is. Two children force-fed ghee as a cultural practice developed lipoid pneumonia with thickening and destruction of the alveolar septa, and a chest infection that did not respond to antimicrobials. These are two case reports, so no frequency can be read from them, and the harm came from forcing fat into a child’s airway rather than from ghee in food. We found no study of ghee with any medicine, and nothing on pregnancy or breastfeeding, so if you take something where dietary fat matters, that question is not answered here. What is missing is most of the question. Everything known about ghee and blood lipids comes from short trials in young healthy people, and the authors of the largest of them, 63 volunteers over 8 weeks, write that the result cannot be carried across to older people, to people with high cholesterol or to people with less healthy lifestyles. We found no trial longer than 8 weeks, no trial in anyone with raised cholesterol, no study of how ghee changes the absorption of any nutrient, and no ghee entry in the USDA retention tables, so cooking losses are also unmeasured. If you cook with ghee, the two things this page supports are reading the label to learn which fat is in the jar, and counting it inside your saturated fat rather than outside it.

Sources

1. ghr5-11 · food composition analysis
   The top five ethnic foods containing highest levels of selected nutrients were lamb balti (3mg/100g iron), lamb kebab (3.2mg/100g zinc), mixed dhal (62μg/100g folate), fish curry (1.4μg/100g vitamin D), ghee (968μg/100g retinol) and toor dhal (9.1g/100g dietary fibre).
   Food Chem, 2013 · checked 2026-09-24
2. ghr5-12 · cross-sectional survey
   The proportion of people consuming tandoori roti, dal with vegetables, potato and ghee on a daily basis was highest in the urban sample, intermediate in the migrant group and lowest in the rural group (P ≤ 0·01).
   Public Health Nutr, 2010 · checked 2026-09-24
3. ghr5-05 · case-control study
   Our study found that certain dietary factors, such as consuming fast food (OR: 3.0; P = 0.0001) and reusing ghee (OR: 2.45; P = 0.0001) and oil (OR: 4.30; P = 0.0001), increase the risk of CRC.
   Nutrition, 2024 · checked 2026-09-24
4. ghr5-01 · randomised controlled trial
   Thirty healthy men and women were studied in a free-living outpatient regimen...Compared with the olive oil, the diet with ghee increased fasting plasma apo-B (apo B) (0·09, 95 % CI 0·02, 0·17 g/l, P = 0·018), non-HDL-cholesterol (non-HDL-cholesterol) (0·53, 95 % CI 0·01, 1·05 mmol/l, P = 0·046)
   Br J Nutr, 2022 · checked 2026-09-24
5. ghr5-02 · randomised controlled trial
   The present study was conducted on sixty-three healthy, young, physically active adult volunteers (52 male, 11 female)...The serum total cholesterol level showed a significant rise in the experimental group at 4 wk; the rise persisted at 8 wk. A similar rise was also seen in HDL cholesterol. Hence the total cholesterol/HDL cholesterol ratio did not show any significant change.
   Indian J Physiol Pharmacol, 2005 · checked 2026-09-24
6. ghr5-03 · randomised controlled trial
   In neither group was there any significant change in the serum lipid profile at any point in time. At 8 wk, 2 volunteers in the experimental group, and 1 volunteer in the control group had more than 20% rise in serum total cholesterol as compared to their 0 wk values
   Indian J Physiol Pharmacol, 2002 · checked 2026-09-24
7. ghr5-04 · systematic review of randomised controlled trials
   A total of seven RCTs demonstrated food-specific and heterogeneous effects...Ghee intake was linked with improved total cholesterol and liver enzyme (ALP).
   Nutrients, 2026 · checked 2026-09-24
8. ghr5-06 · case-control study
   A total of 500 subjects were selected, of which 250 were coronary artery disease patients and 250 were healthy controls...A key finding is that consumption of chicken, beef, eggs and junk food are associated with a high risk of CAD whereas consumption of ghee is not associated with the risk of CAD.
   Int J Environ Res Public Health, 2022 · checked 2026-09-24
9. ghr5-07 · case-control study
   this case-control study was performed among 210 subjects...However, fat-free milk, ghee/butter oil, curd/yogurt, and fruits had lower ORs revealing no or less risks for CAD.
   J Health Popul Nutr, 2021 · checked 2026-09-24
10. ghr5-09 · WHO guideline
    WHO recommends that adults and children reduce saturated fatty acid intake to 10% of total energy intake (strong recommendation). 2. WHO suggests further reducing saturated fatty acid intake to less than 10% of total energy intake (conditional recommendation).
    Saturated fatty acid and trans-fatty acid intake for adults and children: WHO guideline, 2023 · checked 2026-09-24
11. ghr5-10 · review of product labels and food regulations
    Vanaspati ghee (partially hydrogenated vegetable oil), bakery shortening, hard margarines, and fat spreads are identified as the major sources of TFAs.
    J Clin Hypertens (Greenwich), 2020 · checked 2026-09-24
12. ghr5-08 · case reports
    We report the ultrastructural findings and observations in animal fat associated lipoid pneumonia from two children following a cultural practice of forced feeding with animal fat (ghee). Clinical findings showed an acute or chronic chest infection which failed to respond to anti-microbial therapy.
    East Afr Med J, 2000 · checked 2026-09-24
13. ghr5-13 · randomised controlled trial
    However, more studies are needed on older subjects, hyperlipidaemic subjects, and on subjects following less healthy lifestyles before the results of this study can be extrapolated to the general population.
    Indian J Physiol Pharmacol, 2005 · checked 2026-09-24