Onion rings: nobody has studied them, so this is what fried food data says

In a randomised crossover meal, 12 obese insulin-resistant women ate vegetables deep-fried in extra-virgin olive oil. Against a meal of identical composition built with 25 g of uncooked oil, the fried version lowered their C-peptide area under the curve at 120 minutes (P = 0.0020). In the five lean women who ate the same two meals there was no difference. That is seventeen women, one sitting, postprandial markers only, and what a lower C-peptide curve after a single meal means for health is not something the trial measured. Extra-virgin olive oil is also not the fat a commercial fryer holds, so this tells you nothing about a basket of rings. We found no study that names onion rings at all. What is on this page was gathered under the terms the research actually uses: fried food, deep frying, acrylamide. Pooling observational studies published up to June 2020, the highest fried food intake carried a relative risk of 1.16 (95% CI 1.07 to 1.25) for overweight or obesity in adults. Hypertension came out at 1.20 (95% CI 1.05 to 1.38), and type 2 diabetes showed no clear link (RR 1.07, 0.90 to 1.27). Heterogeneity in that pooling ran as high as 92 percent and the authors ask for caution themselves. These are associations between an eating pattern and an outcome, not a measured effect of one food.

The numbers, per 100 g

What your body absorbs

• In 12 obese insulin-resistant women, a meal whose vegetables were deep-fried in extra-virgin olive oil cut the C-peptide area under the curve at 120 minutes against the same meal built with 25 g of uncooked oil (P = 0.0020). — 12 obese insulin-resistant women (BMI 32.8 kg/m2) and five lean women (BMI 22.2), randomised to two meals of equal composition (n = 17) Evidence B source
  C-peptide AUC lower after the fried meal at 120 min (P = 0.0020), 150 min and 180 min (P = 0.0039); no difference in lean women

What cooking and storage change

• Dipping battered and breaded patties in a 10% whey protein isolate solution at pH 2 before deep frying gave the largest lipid reduction, and the only changes a trained panel could perceive were colour, hardness and crunchiness. — battered and breaded chicken patties, crackermeal or Japanese breadcrumb coating, judged by a trained descriptive sensory panel on 16 attributes (n = not stated in abstract) Evidence B source
  Largest lipid reduction at 10% whey protein isolate, pH 2, with darker and less yellow crust and no perceivable change in hardness or crunchiness

What it does to your health

• A meta-analysis of 18 case-control studies found the highest fried food intake carried an odds ratio of 1.52 (95% CI 1.23 to 1.87) for gastric cancer, across 5,739 patients and 70,933 controls. — gastric cancer patients (n = 5739) and healthy adults (control, n = 70,933) (n = 5739 cases, 70933 controls) Evidence C source
  OR 1.52 (95% CI 1.23-1.87), I2 = 76%, p = 0.0001; odds, not risk
• Pooling observational studies, the highest fried food intake carried a relative risk of 1.16 (95% CI 1.07 to 1.25) for overweight or obesity and 1.20 (95% CI 1.05 to 1.38) for hypertension in adults, with no clear link to type 2 diabetes. — adults in cohort, cross-sectional and case-control studies published up to 17 June 2020 (n = not stated in abstract) Evidence C source
  RR 1.16 (1.07-1.25) overweight/obesity; RR 1.20 (1.05-1.38) hypertension; RR 1.07 (0.90-1.27) T2DM

Safety, limits and interactions

• Across five studies and 53,870 mother and newborn pairs, the highest quartile of the acrylamide haemoglobin adduct went with an odds ratio of 1.20 (95% CI 1.08 to 1.33) for small for gestational age and a birth weight 131 g lower (95% CI 204 to 58 g lower). — 53,870 mother-newborn pairs in five eligible studies (n = 53870) Evidence C source
  OR 1.20 (95% CI 1.08-1.33) for SGA; birth weight -131 g (95% CI -204 to -58); head circumference -0.31 cm (95% CI -0.58 to -0.04). Odds, not risk
• EFSA holds that acrylamide in food potentially increases cancer risk for consumers in all age groups, and that on a body weight basis children are the most exposed age group. — consumers of all age groups in the European Union Evidence E source
  No numeric risk estimate given by the agency; the position rests on animal studies and human exposure data

What people believe that the data does not support

• Across 16 studies and 1,151,189 adults followed for a median 14.9 years, dietary acrylamide exposure averaging 23 micrograms a day showed no association with any site-specific non-gynecological cancer. — 1,151,189 non-occupationally exposed adults, 48,175 of whom developed cancer, median follow-up 14.9 years (n = 1151189) Evidence C source
  No association at highest versus lowest exposure, no threshold in the dose-response analysis; lung cancer in smokers the single exception

Who this works differently for

• Pooling five cohort studies of 54,728 participants, high gestational acrylamide exposure went with an odds ratio of 1.14 (95% CI 1.08 to 1.21) for the child being overweight or obese at age 3. — five cohort studies including 54 728 participants (n = 54728) Evidence C source
  OR 1.14 (95% CI 1.06-1.23) for small for gestational age; OR 1.14 (95% CI 1.08-1.21) for overweight or obesity at age 3, falling to OR 1.09 (1.02-1.16) at age 8. Odds, not risk

What is still unknown

• Only five in vivo studies have ever tested biomarkers of fried food intake, so how much fried food a person actually eats still cannot be measured objectively. — four human intervention studies and one animal study, mostly on fried potato and meat (n = 5 studies) Evidence D source
  Short-chain fatty acids, amino acids and urinary acrolein and acrylamide derivatives appear as candidate markers, none validated

The bottom line

Stomach cancer is the association people repeat, and it has an exact shape. A meta-analysis of 18 case-control studies, 5,739 patients against 70,933 controls, put the highest fried food intake at an odds ratio of 1.52 (95% CI 1.23 to 1.87). Those are odds, not risk, and they come from studies that asked people to recall diets they had already eaten, with heterogeneity of 76 percent and a different definition of fried food in each of the 18. Acrylamide, which forms in the browned crust, is the other worry, and the adult human data runs the other way. Sixteen studies and 1,151,189 adults were followed for a median of 14.9 years, and dietary acrylamide averaging 23 micrograms a day showed no association with any site-specific non-gynecological cancer, with no threshold in the dose-response analysis. Intake there was estimated from food frequency questionnaires, which measure it poorly, so the null is weak reassurance and not proof of safety. Pregnancy is where the acrylamide numbers stop being null. Across five studies and 53,870 mother and newborn pairs, the highest quartile of the acrylamide haemoglobin adduct went with an odds ratio of 1.20 (95% CI 1.08 to 1.33) for a baby small for gestational age. Birth weight in that quartile was 131 g lower (95% CI 204 to 58 g lower). The exposure is observational, it rests on five studies, and the adduct counts acrylamide from every source, smoking included. Five cohorts of 54,728 participants linked high exposure in pregnancy to the child being overweight or obese at age 3 (OR 1.14, 95% CI 1.08 to 1.21), easing to OR 1.09 (1.02 to 1.16) by age 8. Those are odds, not risk, and exposure was estimated from diet or from adducts, and that second finding sits on this page as a caution for pregnancy as much as a fact about children. EFSA holds that acrylamide in food potentially increases cancer risk for consumers in all age groups, and that children are the most exposed on a body weight basis. The agency attaches no effect size to that, and it is a precautionary position rather than a measured one. We have no card here on allergy, none on any medicine, and none on a portion size that changes anything. The frying itself can be altered. Dipping battered and breaded chicken patties in a 10 percent whey protein isolate solution at pH 2 before deep frying produced the largest lipid reduction in that trial. A trained panel judging 16 attributes could perceive only a darker, less yellow crust, with nothing it could feel in hardness or crunchiness. The patties were chicken, the size of the lipid reduction is not given in the abstract, and it reaches a battered, breaded, deep-fried onion only by analogy. The honest limit is measurement. Only five in vivo studies have ever tested biomarkers of fried food intake, four in humans and one in animals, mostly on fried potato and meat, and none of the candidate markers is validated. So everything above rests on people writing down what they think they ate.

Sources

1. onri-r5-07 · randomised controlled trial
   This study demonstrated that in obese, insulin-resistant women, food fried in extra-virgin olive oil significantly reduced both insulin and C-peptide responses after a meal.
   J Med Food, 2011 · checked 2026-09-24
2. onri-r5-06 · randomised controlled food process trial
   Thus, JBP that showed the highest lipid reduction (10% WPI at pH 2) were observed to be darker, less yellow, but did not produce any perceivable changes in hardness or crunchiness
   J Food Sci, 2009 · checked 2026-09-24
3. onri-r5-01 · meta-analysis of case-control studies
   There was a significant positive association between gastric cancer risk and fried food intake (OR = 1.52, 95% CI = 1.23-1.87, I2 = 76%, p = 0.0001).
   Nutrients, 2023 · checked 2026-09-24
4. onri-r5-02 · meta-analysis of observational studies
   In comparing the highest to lowest fried-food intake, the pooled RRs (95% CIs) were 1.16 (1.07-1.25...Our meta-analysis indicates fried-food consumption is associated with increased risk of overweight/obesity and hypertension but not T2DM in adults
   Crit Rev Food Sci Nutr, 2022 · checked 2026-09-24
5. onri-r5-04 · meta-analysis of observational studies
   Higher risk of SGA and lower birth weight and head circumference were observed in the highest quartile of AA-Hb
   Environ Res, 2022 · checked 2026-09-24
6. onri-r5-08 · agency position
   acrylamide in food potentially increases the risk of developing cancer for consumers in all age groups. While this applies to all consumers, on a body weight basis, children are the most exposed age group.
   EFSA, Acrylamide topic page, accessed 2026-09-24 · checked 2026-09-24
7. onri-r5-03 · dose-response meta-analysis of observational studies
   Pooled analysis showed no association between the highest vs. lowest dietary acrylamide exposure and each site-specific cancer investigated, with no evidence of thresholds in the dose-response meta-analysis.
   Front Nutr, 2022 · checked 2026-09-24
8. onri-r5-05 · meta-analysis of cohort studies
   In addition, pooled ORs suggested that children had a high risk of developing overweight/obesity in the future in maternal high AA exposure group
   Public Health Nutr, 2020 · checked 2026-09-24
9. onri-r5-09 · systematic review of in vivo studies
   In conclusion, the number of published papers in this research subject shows that it remains largely uninvestigated.
   Food Funct, 2025 · checked 2026-09-24