In this large population-based study on cooking methods, raw, boiling, pan-frying, and toasting showed beneficial profiles for inflammatory markers, while frying and stewing showed detrimental inflammatory profiles. Raw, boiling, and pan-frying showed beneficial differences in renal biomarkers; no other cooking methods influenced renal function. Boiling was associated with lower TSH levels. Pan-frying and toasting were associated with a higher serum vitamin D, while frying was associated with a lower level of this vitamin. Therefore, raw, boiling, pan-frying, and toasting are associated with healthy profiles. However, frying and, to a less extent, stewing showed unhealthier profiles. Cooking methods that included no added fats where healthier than those including added fats heated at high temperatures or during longer periods of time.
Our results, obtained in a sample of older adults, are an example of novel exposures influence that have hardly been studied in the past and that may also be used as a healthy ageing strategy21. Thus, although some cooking methods are considered healthy, and in many cases older adults have gradually adopted them, comprehensive information that provides guidance on food preparation has never been provided in literature before.
Several mechanisms could alter food while it is being processed, including thermal procedures and the addition of certain components such as fats. For instance, mechanical processing or soaking could also modify nutrient bioavailability22. Then, there is a wide range of processes involved: from liquid evaporation and its substitution with fats23,24, variations on antioxidant activity25,26, formation of advanced glycation end products27,28, to altered glycemic and insulinemic responses29,30. In addition, some specific cooking methods are considered harmful. During deep-frying, polycyclic aromatic hydrocarbons and heterocyclic amines are produced, which are genotoxic agents, and have been associated with cancer incidence31. Also, saturated fatty acids formation could increase during frying32 that, depending on its food source, are related varied biological effects.
Our findings support that, among the elderly, raw food consumption is associated with beneficial marker profiles (such as lower levels of hs-CRP, neutrophils, GDF-15, IL-6, urinary albumin, and serum uric acid). Two cross-sectional studies assessed the consumption of raw food-only diets. The first concluded that such a diet was associated with weight reduction, but also with underweight33. In the second one, an exclusive raw food diet was associated with lower LDL-cholesterol and HDL-cholesterol levels, in addition to vitamin B-12 deficiency34. Hence, along with general recommendations for a healthy diet (quality food choices and adequate daily consumption), it seems very reasonable to favor raw food consumption for this age group.
Regarding inflammatory markers, raw, boiling, and toasting food consumption were associated with lower hs-CRP levels. A possible explanation could rely on the fact that these cooking methods usually do not imply the addition of unhealthy fat sources35. In an analysis of the National Health and Nutrition Examination Survey (NHANES) conducted with 4900 adults, it was observed that saturated fat consumption was modestly associated with elevated hs-CRP36. In addition, raw food has high fiber and antioxidants contents, which are also associated with lower levels of hs-CRP37. On the contrary, frying and stewing were associated with higher hs-CRP levels in this study. The addition of harmful fats or a longer cooking time might be associated with higher levels of this inflammatory marker. Regarding this latter, recent studies have found that a longer cooking time—such in stewing, increases fat content of food38,39 that could lead to deleterious effects on health40.
Concerning other inflammatory markers, the association of raw food consumption (mainly fruit and vegetables) with lower counts of white blood cells and neutrophils was also examined in other studies. An observational study with 51 vegan participants found a reduction in CD4, CD8 and natural-killer cells41. Likewise, Mediterranean diet adherence, rich in raw food, was associated with lower leukocyte count levels42. This is important since neutrophils regulate inflammatory function and it has been suggested that lower neutrophil levels are associated with lower atherothrombotic processes and cardiovascular diseases43,44.
In our study, raw food consumption was also associated with lower GDF-15 levels. Among older adults, the importance of GDF-15 relies on its role as biomarker for cardiac fibrosis45 as well as a marker for chronic disease burden46. An essential cofactor of this biomarker is selenium—mostly found in raw consumed food46, and it has been described an inverse relationship between them47. On the other hand, although cooking methods were not evaluated, a 6-month clinical trial found no association between GDF-15 levels and plant-based or protein-based diets48.
Raw and toasting food consumption showed decreasing differences in IL-6 levels. These cooking methods involve dressing oils and hidden fats from fruits, nuts and seeds (usually consumed toasted)49 which in turn promote carotenoid absorption. Carotenoid levels are inversely related to serum IL-6 due to modifications in systemic inflammatory responses50,51. Similarly, in a cohort with 3075 older adults an association was found between lower IL-6 levels and a dietary pattern that included high consumption of fruit, vegetables, whole grains, and non-fried poultry52. Lower levels of IL-6 might be beneficial as an inverse relationship with coronary heart disease has been observed53.
Concerning renal function, there was a lower mean urinary albumin with raw, boiling, and pan-frying consumption in this population. A high renal excretion of albumin could reflect greater animal protein and fat intake, and, generally, these cooking methods do not involve the addition of fats. Therefore, these could be beneficial for both overweight and high blood pressure prevention54, and might also be recommended in patients with hypertension or decline in renal function. On the other hand, lower uric acid levels occurred with raw food consumption. This is in line with the literature as a lower purine content in raw food could be responsible for a lower uric acid formation55,56.
In relation to hormones and other nutritional biomarkers, our results suggest that boiling is associated with lower TSH levels and it could potentially prevent subclinical hypothyroidism. Though, further research is still needed. A study demonstrated that boiled cauliflower had lower levels of progoitrin, a precursor of goitrin57. However, other boiled cyanogenic plants (cabbage, radish, cassava) have shown high anti-thyroid peroxidase levels that could exert opposite effects58. Additionally, beneficial PD in serum vitamin D was observed with pan-frying and toasting. Conversely, the exposure to high temperatures has been shown to cause adverse effects in vitamin D3 content in food, but knowledge regarding this matter is limited, and the stability of vitamin D3 in added vegetable oils during cooking has not yet been determined59,60. Of note is that for pan-frying and toasting, food is exposed to heat for shorter periods of time.
The main strength of this study is the use of a dietary history that allowed to collect detailed information on food consumption and cooking methods. Its validity and reproducibility has been shown in Spanish population. Standardized procedures have been followed during analytical processing of biological samples in a central laboratory. Likewise, several confounding factors were taken into account (including socio-demographic determinants and lifestyle), although some degree of residual confounding cannot be ruled out.
This study has several limitations that have to be considered. First, since this study has a cross-sectional design, the results cannot establish causality. Second, as when a specific food is studied, it is not possible to distinguish the effect of cooking from the consumption of the food itself. Similarly, we cannot distinguish the effect of a specific cooking method from the effect of added fats when occurring together. Finally, because of administrative reasons, some measurements were only conducted in the subsample recruited from January, 2017 to the end of the study resulting in a reduction in sample size, but without deriving in selection bias.