In an experiment with carrots, intensive baking made almost nine times as much carotenoid accessible during simulated digestion as in raw carrots. No vitamins were added. Cooking changed what the digestive system could reach.
Meanwhile, boiling can remove water-soluble vitamins from vegetables. So does cooking preserve nutrients or destroy them? Both can happen to the same meal.
The catch is that three familiar questions are not equivalent: how much of a compound remains in the food, how much can be released during digestion, and whether eating it improves health. Each requires a different experiment. Confusing them has produced decades of confidently contradictory kitchen advice.
✦ Key takeaways
- Cooking changes both nutrient content and nutrient accessibility. A laboratory loss is not automatically a loss to the body.
- Steaming and microwaving often preserve water-soluble nutrients better than boiling, but the food and the exact technique matter.
- Heat can improve the availability of some carotenoids. Small human studies also show that adding oil to cooked tomatoes changes lycopene uptake.
- Food safety and the overall diet outrank nutrient optimisation. There is no universal “healthiest cooking method”.
The four questions hidden inside “healthier cooking”
Four outcomes that sound similar often describe different things.
What survives? A laboratory can measure vitamin C, folate or a plant compound before and after heating. Temperature, cooking time, water contact and oxygen all affect the result. Crucially, an apparent increase in nutrient concentration per 100 g may partly reflect water evaporation. A serious comparison accounts for the changed weight of the cooked food, a measure sometimes called true retention.[1]
What becomes accessible? Our digestive enzymes do not encounter isolated chemicals. They encounter the walls of plant cells, fibrous networks and emulsions. Softening or breaking those structures can make some compounds easier to release during digestion. A laboratory simulation of digestion measures bioaccessibility; it does not necessarily measure absorption into a human bloodstream.[2]
What becomes safer? Adequate cooking can reduce harmful microbes and inactivate naturally occurring toxins. For certain foods, the healthiest raw-versus-cooked comparison is not close at all.
What will someone actually eat? Perfect nutrient retention is of little value if the meal goes uneaten. Taste, cost and habits ultimately matter alongside chemistry.
The carrot experiment: more available nutrients after heating
Carrots make an unusually good demonstration because their orange colour comes partly from carotenoids, including beta-carotene, which the body can convert to vitamin A. These compounds are embedded in plant tissue. The visible colour does not tell us how much will be released in the intestine.
In a 2024 experimental study, researchers cooked carrots by boiling, baking, steaming or microwaving and then used a simulated digestion procedure to assess accessible carotenoids. Under the most intensive treatments tested, total carotenoid bioaccessibility rose by roughly 6.0-fold with microwaving and 8.9-fold with baking relative to raw carrots.[2]
This was simulated digestion, not an absorption trial in people. Digestion, added fat and serving size could change the actual effect.
Carrot results and tomato results measured different endpoints. They cannot be put on one health-benefit scale.
Total carotenoid bioaccessible content, relative to raw carrots. Most intensive treatments tested.
These values come from simulated digestion, not measured absorption in people.
A five-day experiment: 470 g tomatoes daily, cooked with or without 25 ml extra-virgin olive oil.
Change in plasma trans-lycopene in the 11-person oil group; no statistically significant change in the 12-person no-oil group. A short-term biomarker, not demonstrated disease prevention.
Raw carrots remain an excellent food. The experiment challenges the assumption that the highest raw nutrient content always means the greatest nutrient availability.
Why boiling loses nutrients, and why soup is different
A saucepan offers two routes by which water-soluble nutrients can disappear from the vegetables on your plate. Some compounds degrade during heating; others move into the cooking water. When you pour that water down the sink, you may discard material that was never chemically destroyed.
In a 2018 laboratory comparison of selected vegetables, boiling generally produced lower vitamin C retention than microwaving, although outcomes varied considerably by vegetable and treatment. Some measured retention values even exceeded 100%, illustrating how extraction, measurement variation and changes in food structure complicate the interpretation of laboratory numbers.[1] A separate six-vegetable study also reported lower vitamin C losses with microwaving than boiling under its specific test conditions.[3]
The result depends on the vegetable, cooking time, water volume and whether changes in food weight are accounted for.
There is also a simple culinary escape clause: if the cooking liquid becomes part of the meal, some of the dissolved nutrients remain in the dish. A vegetable soup and a plate of boiled vegetables drained in a colander are not the same experiment.
This is why the instruction to “never boil vegetables” is too blunt. Boiling is easy to control, and it can improve texture, digestibility or safety. Steaming is often a useful default when the goal is to limit water-soluble vitamin loss; it is not a moral achievement.
The microwave paradox
Microwaves have acquired a reputation for being unnatural, although the relevant kitchen mechanism is familiar: electromagnetic energy warms parts of the food, and heat changes its chemistry. Water-soluble vitamins do not recognise whether the heat came from a saucepan, oven or microwave.
Because microwaving can cook quickly and with little added water, it sometimes gives good nutrient retention. That is a consequence of time, heat distribution and liquid contact, not a special nutrient-preserving property of the appliance.[1][3]
Microwaves can also heat unevenly. With foods that require a safe internal temperature, check it rather than assuming a short cooking time is sufficient. Optimise the process, not the appliance.
Why tomatoes behave differently with oil
Tomatoes reveal a second confusion. When a compound moves from a vegetable into your bloodstream more readily, the raw food may appear nutritionally inferior even if it had plenty of that compound to begin with.
Lycopene, the red carotenoid in tomatoes, is fat-soluble. Mechanical processing, heating and eating it with fat can change its availability. That does not make raw tomatoes deficient; it makes the food’s physical form relevant.
In a small human experiment, participants ate a daily meal of 470 g of diced tomatoes cooked either with or without 25 ml of extra-virgin olive oil for five days. Among 11 people in the oil group, plasma trans-lycopene rose by 82% and cis-lycopene by 40%. Among 12 people in the no-oil group, trans-lycopene showed no statistically significant change, while cis-lycopene rose by 15%.[4]
Unlike the carrot experiment, this measured a response in people. It still measured a blood biomarker over five days, not long-term health.
Tomatoes cooked with oil and raw tomatoes in a salad are both sensible choices; the preparation changes which qualities you emphasise.
The dark side of browning
Some of the flavours we most admire come from chemical reactions under heat. Browning creates aromas and textures that make food appealing. It also provides a warning against treating temperature as though it had only nutritional benefits or only dangers.
For muscle meats, high-temperature cooking, especially grilling over flame or prolonged contact with a hot surface, can generate heterocyclic amines (HCAs) and polycyclic aromatic hydrocarbons (PAHs). These have mutagenic properties in experiments. Animal studies using high doses have found carcinogenic effects, but epidemiological studies have not established precisely how much risk ordinary dietary exposure to these compounds adds for humans.[5]
The distinction between hazard and individual risk matters. Detecting a potentially harmful chemical does not by itself quantify the probability of disease from a particular dinner. Nor does uncertainty about that risk mean repeated heavy charring is worth pursuing.
For starchy foods such as toast and potatoes, another compound matters: acrylamide, which can form during high-temperature cooking. European food-safety advice favours golden rather than very dark browning of toast and fried potatoes.[6]
And there is an even more immediate danger at the other end of the thermometer: undercooked food. US food-safety guidance specifies a 74°C (165°F) internal temperature for poultry and 71°C (160°F) for ground meat. An appropriate food thermometer is more useful for this purpose than judging colour alone.[7]
Raw or inadequately cooked red kidney beans can contain enough phytohaemagglutinin to cause acute gastrointestinal illness. FDA guidance recommends soaking them for at least five hours, discarding the soaking water and then boiling in fresh water for at least 30 minutes. A slow cooker alone may not get hot enough to inactivate the lectin reliably.[8]
Cooking is therefore not merely a choice about retaining nutrients. In some foods it is an essential safety intervention.
Do you need a special high-temperature cooking oil?
Advice about cooking oil often begins and ends with its smoke point. That is too simple. The visible onset of smoke is influenced by free fatty acids and other features of the oil. It does not provide a single-number description of oxidation, flavour, nutritional quality and performance across repeated cooking cycles.[9]
Consider the more meaningful distinction: what fat does the oil add to the diet, how much do you use, how long and how hot do you heat it, and is it repeatedly reused? For everyday cooking, extra-virgin olive oil and other oils rich in unsaturated fats can be reasonable choices when used appropriately. A high advertised smoke point does not turn an oil into a health treatment, and a lower one does not automatically disqualify it from ordinary sautéing.
Replacing olive oil with coconut oil solely because coconut oil tolerates a particular cooking technique would be a questionable nutrition rule: coconut oil is comparatively rich in saturated fat. Likewise, the blanket warning that “margarine is unhealthy” ignores modern unsaturated soft spreads without industrial trans fats.
For more on selecting oil rather than worshipping a smoke-point table, see our guide to olive oil.
Fermented food is not automatically probiotic food
Heat is not the only way to transform food. Microbes can do chemistry on our behalf. In fermentation, bacteria, yeasts or fungi change sugars, proteins and other compounds, producing new flavours, textures and sometimes better keeping qualities.
But fermented and probiotic are not synonyms. A fermented food may have no live organisms when it reaches the table. Baked sourdough bread, for instance, has undergone fermentation even though the baking step kills the fermenting microbes. A true probiotic claim requires evidence of a health benefit from sufficiently characterised live microorganisms in an adequate dose.[10]
Yoghurt and kefir can be useful foods, but their effects depend on the product and the outcome studied. Warming a fermented ingredient may alter or remove live microbial activity without necessarily removing the nutritional value of the underlying food. Our kefir review examines what human trials do and do not show.
It is entirely possible to enjoy fermented foods for their flavour without turning them into a universal prescription for the microbiome.
Why “processed” is the wrong dividing line
To cut a carrot is to process it. To cook beans, pasteurise milk or freeze peas is to process food. The question is what the processing does, not whether it happened.
There is nevertheless an important finding behind concerns about highly processed diets. In a tightly controlled inpatient crossover trial, 20 adults ate an ultra-processed or unprocessed diet for two weeks each, with foods offered freely. Despite matching several presented nutritional characteristics, the participants consumed an average of 508 kcal more per day during the ultra-processed phase.[11]
That is a genuine experimental signal about the two diets tested. It is not evidence that a tin of tomatoes is harmful, that every ingredient classified as ultra-processed causes overeating, or that a home-made meal is always nutritionally better. Eating rate, texture, energy density and the wider food environment may be involved.
A tin of beans that makes a nourishing dinner possible on a busy evening can be more helpful than a basket of “pristine” ingredients that goes unused. Healthy cooking has to survive contact with ordinary life.
A practical food-by-method comparison
This comparison distinguishes the endpoint actually measured in each source: chemical retention, simulated digestion, short-term blood response, or food safety. It is not a ranking of cuisines or a long-term disease-outcome study.
| Food or method | What changes | Useful kitchen decision | Evidence measured |
|---|---|---|---|
| Carrots: raw vs cooked | 6.03× (microwaving) to 8.90× (baking) accessible carotenoid content under intensive experimental conditions. | Eat both forms; those fold changes are not measured human absorption. | In-vitro simulated digestion, not a human trial [2] |
| Vegetables: boiling vs steaming | Water-soluble compounds may move into cooking liquid. | Steam when convenient, or retain broth in soups. | Nutrient-retention assays [1][3] |
| Vegetables: microwaving | Short cooking with little water can reduce some losses. | Useful for quick portions; heat evenly and sufficiently. | Nutrient-retention assays [1][3] |
| Tomatoes with olive oil | Plasma trans-lycopene rose 82% in the oil group. | The experiment used 25 ml oil per 470 g tomatoes for five days, not a few drops. | Human biomarker study; oil group n=11, no-oil group n=12 [4] |
| Meat over direct flame | More intense heating can generate HCAs and PAHs. | Avoid prolonged heavy charring; cook safely. | Chemistry, animal and observational evidence [5] |
| Toast or potatoes: dark browning | Acrylamide can form during high-temperature browning. | Prefer golden rather than dark brown. | Food chemistry and risk assessment [6] |
| Dried kidney beans | Adequate boiling inactivates a harmful lectin. | Never rely on inadequate low-temperature cooking. | Toxicology and food safety [8] |
| Fermented products | Microbial fermentation changes food; live microbes may not survive further processing. | Separate flavour, live cultures and proven probiotic effects. | Microbiology and product-specific trials [10] |
How to make a better kitchen decision
The most useful healthy cooking tips are not thirteen separate commandments. They form a short decision process.
First, make the food safe. Properly cook poultry, ground meat and foods containing naturally occurring toxins. Where temperatures matter, measure them.
Second, decide what you are trying to preserve or obtain. For a delicate vegetable rich in water-soluble vitamins, brief steaming may make sense. For carrots or tomatoes, some cooking can improve accessibility of specific compounds. For beans, adequate cooking is non-negotiable.
Third, account for what happens to the water and the oil. Nutrients in a discarded pot of water and nutrients in a soup are different cases. A small amount of oil can change carotenoid uptake, but pouring in more oil is not automatically a nutritional improvement.
Fourth, avoid needless extremes. Repeated heavy charring is not required for good flavour. Neither is an obsession with preserving every milligram of vitamin C.
Finally, ask whether you will eat the meal again. A useful cooking habit is one that produces enjoyable food, fits real life and supports a varied dietary pattern. The World Health Organization’s framework emphasises adequacy, balance, moderation, diversity and safety, not a single ideal pan or appliance.[12]
What the studies cannot tell us
Most experiments about cooking methods measure food chemistry, simulated digestion or short-term changes in blood markers. They rarely randomise people to different cooking techniques for years while holding their entire diets and lifestyles constant. The endpoint is often a vitamin concentration, not fewer heart attacks.
Nor is there one outcome to optimise. Vitamin C retention, carotenoid accessibility, food safety, energy intake and taste can pull in different directions. No experiment supplies a universal weighting for them.
There is something liberating about that conclusion. Healthy cooking is not the pursuit of a flawless technique. It is choosing sensible trade-offs, informed by measurements rather than kitchen folklore, and making food that belongs in a sustainable diet.
Sources and how to interpret them
- Effects of different cooking methods on the content of vitamins and true retention in selected vegetables. Food Science and Biotechnology, 2018. Laboratory food-composition study, not a health-outcomes trial.
- Towards more sustainable cooking practices to increase the bioaccessibility of colourless and provitamin A carotenoids in cooked carrots, 2024. Simulated gastrointestinal digestion; fold changes do not equal measured human absorption.
- Effect of cooking methods on the nutritional quality of selected vegetables. Heliyon, 2023. Six-vegetable laboratory comparison; results are specific to its preparation protocols.
- Increases in plasma lycopene concentration after consumption of tomatoes cooked with olive oil, 2005. Short human intervention; biomarker, not clinical outcome.
- Chemicals in Meat Cooked at High Temperatures and Cancer Risk. US National Cancer Institute. Explains HCA/PAH formation and uncertainty in human risk.
- Acrylamide in Food. European Food Safety Authority. Food-safety recommendations on excessive browning.
- Safe Minimum Internal Temperature Chart. USDA FSIS. Official food safety guidance.
- Bad Bug Book, second edition: phytohaemagglutinin in beans. US FDA. Toxicology and preparation safety.
- Deciphering the Complexity of Smoke Point in Virgin Olive Oils to Develop Simple Predictive Models, 2025. Oil-composition study; smoke point is not a clinical endpoint.
- The ISAPP consensus statement on fermented foods. Nature Reviews Gastroenterology & Hepatology, 2021. Definitions, microbiology and limits of probiotic claims.
- Ultra-processed diets cause excess calorie intake and weight gain: an inpatient randomised controlled trial. Cell Metabolism, 2019. Twenty adults; two weeks per diet.
- Healthy diet. WHO, 2026. Population-level dietary framework, not evidence for one cooking appliance.
Editorial method. The article distinguishes food-chemistry assays, simulated digestion, short-term human biomarkers, controlled dietary experiments and safety guidance. These endpoints must not be conflated. The article is educational, not individual medical or dietetic advice.
Publication history. First published in July 2023; comprehensively rewritten in October 2026. An earlier version carried the byline of Mikhail Faranov, MD. The present version is attributed to Ilya Nikitin, Ph.D.
Ease your mind about the ingredients in your meals. When you’re the chef, you have full visibility and control over the amount of oil, sugar, and other components you add. This transparency allows you to be responsible for what goes into your food. However, it’s important to be aware of the nutritional content of the dishes you prepare. If a recipe doesn’t align with your dietary preferences or health goals, feel free to remove it from your collection and substitute it with something that better suits your needs.