--- topic: "Cooking Oil Smoke Points" title: "Smoke Points and Cooking Oils" order: 1 description: "Smoke point is a weak guide to safe frying. Stability, polyunsaturated fat, reuse and ventilation matter more. What olive, coconut and seed oils do under heat." question: "Which oils are safe to cook with at high heat?" answer: "Oils low in polyunsaturated fat, such as olive and coconut, likely hold up best; stability and not reusing oil matter more than smoke point." actions: - "Choose a heat-stable oil such as olive oil: olive oils lasted 24–27 hours of simulated frying before the legal breakdown limit, against 15 hours for a vegetable-oil blend [1]." - "Never heat oil until it smokes; use an extractor fan or open a window: women who cooked until oil smoked, without an extractor, had much higher lung-cancer odds [2][3]." - "Fry only occasionally, and discard oil that is dark, foamy or smells acrid instead of reusing it again and again [4][5]." cautions: - "With heart disease, type 2 diabetes or high blood pressure, keep fried food occasional: higher intake was linked to about 28% more major cardiovascular events in an observational meta-analysis [6]." - "Anyone who cooks daily over hot oil should ventilate well: cooking-oil fumes carry aldehydes, and sunflower oil gave off the most in kitchen tests [7][3]." confidence: 3 confidenceNote: "Lab tests agree on which oils break down least under heat, but almost no human trials show that the difference changes health." evidenceSummary: rctCount: 1 observationalCount: 2 metaAnalysisCount: 2 totalParticipants: "40,000+ in the largest cohort; two meta-analyses add more" replicated: true newestStudyYear: 2021 references: - id: "1" pmid: "20678538" title: "Olive oil stability under deep-frying conditions" authors: "Casal S, Malheiro R, Sendas A, et al." year: 2010 journal: "Food Chem Toxicol" - id: "2" pmid: "10645816" title: "Chinese food cooking and lung cancer in women nonsmokers" authors: "Ko YC, Cheng LS, Lee CH, et al." year: 2000 journal: "Am J Epidemiol" - id: "3" pmid: "29164315" title: "The risk of lung cancer among cooking adults: a meta-analysis of 23 observational studies" authors: "Jia PL, Zhang C, Yu JJ, et al." year: 2018 journal: "J Cancer Res Clin Oncol" - id: "4" pmid: "10091835" title: "Impaired endothelial function following a meal rich in used cooking fat" authors: "Williams MJ, Sutherland WH, McCormick MP, et al." year: 1999 journal: "J Am Coll Cardiol" - id: "5" pmid: "26148922" title: "Possible adverse effects of frying with vegetable oils" authors: "Dobarganes C, Márquez-Ruiz G" year: 2015 journal: "Br J Nutr" - id: "6" pmid: "33468573" title: "Fried-food consumption and risk of cardiovascular disease and all-cause mortality: a meta-analysis of observational studies" authors: "Qin P, Zhang M, Han M, et al." year: 2021 journal: "Heart" - id: "7" pmid: "27780622" title: "Effects of cooking method, cooking oil, and food type on aldehyde emissions in cooking oil fumes" authors: "Peng CY, Lan CH, Lin PC, et al." year: 2017 journal: "J Hazard Mater" - id: "8" pmid: "30263791" title: "Physicochemical properties and oxidative stability of frying oils during repeated frying of potato chips" authors: "Yu KS, Cho H, Hwang KT" year: 2018 journal: "Food Sci Biotechnol" - id: "9" pmid: "18203133" title: "Acrolein: sources, metabolism, and biomolecular interactions relevant to human health and disease" authors: "Stevens JF, Maier CS" year: 2008 journal: "Mol Nutr Food Res" - id: "10" pmid: "17995742" title: "Chemistry of deep-fat frying oils" authors: "Choe E, Min DB" year: 2007 journal: "J Food Sci" - id: "11" pmid: "32244669" title: "Potential Adverse Public Health Effects Afforded by the Ingestion of Dietary Lipid Oxidation Product Toxins: Significance of Fried Food Sources" authors: "Grootveld M, Percival BC, Leenders J, et al." year: 2020 journal: "Nutrients" - id: "12" pmid: "22275385" title: "Consumption of fried foods and risk of coronary heart disease: Spanish cohort of the European Prospective Investigation into Cancer and Nutrition study" authors: "Guallar-Castillón P, Rodríguez-Artalejo F, Lopez-Garcia E, et al." year: 2012 journal: "BMJ" --- Smoke point tells you when an oil starts to smoke, not how well it copes with heat. What matters more is how much polyunsaturated fat the oil contains: in frying tests, olive oil and refined coconut oil broke down less than soybean or sunflower oil [1][8][7]. Keep the heat moderate, ventilate, and don't reuse frying oil many times. ## How it works **What a smoke point is.** It is the temperature at which an oil starts to give off visible smoke. Part of that smoke is acrolein, a reactive compound formed when the glycerol in fats breaks down with heat and when polyunsaturated fats oxidize [9]. Smoking oil is a sign you have overheated it. But oils begin to degrade well before they smoke. **Stability is what counts.** Heating in air triggers oxidation, hydrolysis and polymerization. The more unsaturated fatty acids an oil has, the faster it oxidizes, and high temperature, many fryings, free fatty acids and metals all speed the process [10]. Oxidising polyunsaturated fats form aldehydes, some of them toxic, which pass into the fried food [11]. **What frying tests show.** - Olive oils lasted 24 to 27 hours of simulated home frying before total polar compounds (a measure of breakdown) hit the 25% legal limit. A vegetable-oil blend lasted 15 hours [1]. - In 80 repeated fryings of potato chips, soybean oil built up the most oxidation products of four fats tested (refined coconut oil, soybean oil, pure olive oil, shortening), and each oil's aldehydes reflected its fatty-acid makeup [8]. - In kitchen fume tests, sunflower oil released the most aldehydes whatever the method. Palm and rapeseed oil released less. Deep frying released more than pan frying, and pan frying more than stir-frying [7]. So oils rich in monounsaturated fat (olive) or saturated fat (coconut, palm) are more stable than oils rich in linoleic acid (sunflower, soybean). Ghee and tallow follow the same logic, but we found no head-to-head tests of them. Typical smoke points differ between refined and unrefined versions; see [olive oil](/olive-oil/health-benefits), [avocado oil](/avocado-oil/cooking-and-health-benefits) and [seed oils](/seed-oils/healthy-alternatives). **Does it matter for health?** After a meal rich in fat that had been used for deep frying in a fast-food restaurant, ten men's arteries widened less than after the same meal made with unused fat [4]. A Spanish cohort of 40,757 adults found no link between fried food and heart disease, whether the frying oil was olive or sunflower [12], yet a meta-analysis found higher fried-food intake went with 28% more major cardiovascular events [6]. Fumes matter too: among women who cooked, poor ventilation and waiting for oil to smoke went with more lung cancer [2][3]. **Practical rules.** Match the oil to the job. Heat it until it shimmers, not until it smokes. Ventilate. Use fresh oil, and don't keep reusing it. ## The evidence in detail ### Head-to-head frying studies In a Portuguese study, oil was sampled every three hours of simulated domestic frying until total polar compounds reached the 25% legal limit. All olive oils lasted 24 to 27 hours, against 15 for a commercial vegetable-oil blend. The extra-virgin oil showed the least oxidation and hydrolysis and the most minor antioxidants, and the vegetable oil, despite more vitamin E, oxidized far faster [1]. A Korean study fried potato chips 80 times in refined coconut oil, refined soybean oil, pure olive oil and vegetable shortening. Polyunsaturated fatty acids fell in every oil. Soybean oil had the highest conjugated dienes (21.8 mmol/L) and p-anisidine value (47.7), while total polar compounds ended between 8.1% and 9.5% in all four, under the rejection limit [8]. A Taiwanese team compared four oils (palm, rapeseed, sunflower, soybean), three methods and two foods in an ordinary kitchen: deep frying gave the most aldehydes, stir-frying the fewest, and sunflower oil the most of any oil [7]. These are chemistry measurements from simulations, not measurements of harm. ### What eating or breathing the breakdown products does Ten men had flow-mediated brachial artery dilation measured before and four hours after three meals. After the meal made with used restaurant frying fat, dilation fell from 5.9% to 0.8% (P = 0.0003); it did not change after the same meal with unused fat or after a low-fat meal [4]. The study was tiny, single-meal, and did not standardize the oil. A 2020 review argues that polyunsaturated oils heated for frying generate cytotoxic aldehydes and epoxy-fatty acids, then concedes the "urgent and unmet" need for epidemiological trials linking intake of these compounds to disease [11]. A 2015 review likewise puts most suspicion on abused frying oils, not on frying in general [5]. For fumes, a Taiwanese case-control study of 131 nonsmoking women with lung cancer, 252 hospital controls and 262 community controls found adjusted odds ratios of 2.0 to 2.6 for women who waited for oil to emit fumes before cooking and 3.2 to 12.2 for those without a fume extractor [2]. A meta-analysis of 23 observational studies (9,411 cases) found cooking women had about double the odds of lung cancer (OR 1.98 in nonsmokers), poor ventilation raised odds (OR 1.20), and stir-frying was linked to more risk (OR 1.89) than deep frying (OR 1.41, not significant). Heterogeneity was high [3]. The stir-frying finding runs opposite to the aldehyde-emission ranking above, so emissions do not map neatly onto disease, and most of this research concerns hot wok cooking in Asia. ### Fried foods and heart disease In 40,757 Spanish adults followed for a median 11 years, there were 606 coronary events. The hazard ratio for the highest versus lowest quarter of fried-food intake was 1.08 (0.82 to 1.43), with no difference between olive and sunflower oil users, and none for all-cause mortality [12]. A meta-analysis pooled 17 studies: highest versus lowest intake went with relative risks of 1.28 (1.15 to 1.43) for major cardiovascular events, 1.22 (1.07 to 1.40) for coronary heart disease and 1.37 (1.07 to 1.75) for heart failure, but 1.03 (0.96 to 1.12) for all-cause mortality. Heterogeneity was high (I² 82%), and the authors flag recall and misclassification bias [6]. Fried food may be a marker for fast food, chips and calories rather than for the oil. ### Where the evidence conflicts or is missing We found no trial comparing health outcomes of different frying oils. Peer-reviewed head-to-head tests exist for olive, coconut, palm, rapeseed, soybean and sunflower oils, but not for ghee, tallow or avocado oil in the papers we reviewed. Published smoke points vary by refining and test method, so treat them as rough guides. ### Bottom line Confidence is 3 of 5. The ranking of oils by stability is consistent across several laboratory and kitchen studies, and good practice (moderate heat, ventilation, fresh oil) is sensible. Whether the choice of oil changes disease risk is unknown.