Effect of food structure with a focus on fat distribution

The first study that assessed the impact of food structure on postprandial hypertriglyceridaemia was that of Drouin-Chartier et al. ^{(Reference Drouin-Chartier, Tremblay and Maltais-Giguère32)}. The authors showed that dairy fat from a soft cream cheese induced a higher TG response in the early postprandial phase compared with butter or with firm cheddar cheese. However, the measured absolute TG response for the intervention period (6 h) was similar for all products. Although the authors did not measure fat droplet size in their products, they suggested that TG could increase faster with the soft cream owing to the smaller fat droplet size of this product compared with butter or cheddar. However, this may not be the only reason, as discussed later. Hansson et al. ^{(Reference Hansson, Holven and Øyri33)} then reported different TG responses of four different dairy meals (butter, cheese, whipped cream and sour cream) containing the same fat content. The authors reported that the intake of sour cream induced a significantly higher serum TG response than all the other dairy meals. Like Drouin-Chartier et al.,^{(Reference Drouin-Chartier, Tremblay and Maltais-Giguère32)}, the authors attributed these differences to fat droplet sizes without reporting the related values. A limiting factor in this second study was that the different products were given with toast, which probably influenced the TG postprandial response. In both studies, the results could partially be influenced by food consistency (liquid, semi-liquid or solid), as it seems that liquid foods trigger higher TG response of foods with similar composition^{(Reference Dias, Zhu and Thompson11)}. Food structure, which also encompasses food consistency, affects gastric emptying^{(Reference Goyal, Guo and Mashimo34)} and kinetics of lipid digestion, as proven by several in vitro studies^{(Reference Mulet-Cabero, Rigby and Brodkorb35,Reference Salt, Mandalari and Parker36)} . However, these in vitro data sometimes conflict with the conclusion of in vivo studies. For example, Mulet-Cabero et al. ^{(Reference Mulet-Cabero, Rigby and Brodkorb35)} suggested that lipid digestion occurs faster for semi-solid foods than for liquid foods, because of a predicted faster gastric emptying of lipids. These data disagree with the in vivo TG responses to solid and liquid dairy matrices reported by Drouin-Chartier^{(Reference Drouin-Chartier, Tremblay and Maltais-Giguère32)}, Hansson et al. ^{(Reference Hansson, Holven and Øyri33)} and Diaz et al. ^{(Reference Dias, Zhu and Thompson11)}. The comparison and validation of gastric behaviour of such in vitro and in vivo studies is difficult because the in vivo studies scarcely address the structural changes of food in the stomach and the composition of each gastric emptying. However, Kjølbæk et al. ^{(Reference Kjølbæk, Schmidt and Rouy37)} showed a higher TG response for a gel-based casein product than for a casein drink, supporting the hypothesis of Mulet-Cabero^{(Reference Mulet-Cabero, Rigby and Brodkorb35)}. Kjølbæk et al. ^{(Reference Kjølbæk, Schmidt and Rouy37)} suggested that these results could be due to the structure of each product, as the only difference between the two products was that the gel-based product was the result of the acid gelation of the drink. Differences in TG response could also be linked to the fact that casein loses its micellar structure at acid pH, which could affect the lipid digestion rate. However, in the same study, Kjølbæk et al. ^{(Reference Kjølbæk, Schmidt and Rouy37)} reported a similar TG response of two cheddar cheeses with the same nutritional profile but different structures (one solid, one semi-solid), in disagreement with the hypothesis of Mulet-Cabero^{(Reference Mulet-Cabero, Rigby and Brodkorb35)}. It seems that not only macrostructure of foods but also the fine structure of ingredients can affect the postprandial triglyceridaemia response via different digestion rates of lipids. This was illustrated recently in a study that showed a strong influence of the amylose content of starches on lipid digestion of starch-based gels: the higher the amylose content, the higher the lipid digestion per gram of oil^{(Reference Qazi, Ye and Acevedo-Fani38)}. Other in vitro studies have suggested that food processing can also affect lipid digestion^{(Reference Lopez, Cauty and Guyomarc’h39–Reference Qazi, Ye and Acevedo-Fani41)} via structural modification of macro- and micronutrients and, therefore, potential TG response, although validation with in vivo models is needed before health/nutrition-related conclusions can be made, as results may differ^{(Reference Berton-Carabin and Schroën42)}. More recently, Gleize et al. ^{(Reference Gleize, Hiolle and Meunier43)} conducted a study on twelve healthy subjects who received four starchy foods with similar compositions but different structures (custard, pudding, sponge cake and biscuit)^{(Reference Gleize, Hiolle and Meunier43)}. Results showed that custard TG response was significantly higher than pudding and biscuit TG responses. This could be linked to the lipid droplet size/structure of the products: biscuits displayed big fat “flakes” (18 000 μm) while custard was composed of very small droplets (30 μm)^{(Reference Gleize, Hiolle and Meunier43)}. Overall, the available studies suggest that solid food structure limits the TG response, perhaps because solid food structure may modify the accessibility of the lipid droplets to enzymes. Results of Salt et al. ^{(Reference Salt, Mandalari and Parker36)}, who addressed both in vitro lipid digestion of muffins and free oil, support this conclusion. Gleize et al. ^{(Reference Gleize, Hiolle and Meunier43)} also showed that, despite TG postprandial concentrations being different for the four tested products, plasma TG peaked at 3 h for all products. This suggests that, although the food matrix could impact the accessibility of lipids to enzymes, it neither delayed fatty acid absorption by enterocytes nor their secretion into chylomicrons.

As mentioned above, it is widely accepted that the TG postprandial response is influenced not only by the type/amount of fat but also by droplet size distribution^{(Reference Acevedo-Fani and Singh28)}, which is known to change during the digestion process^{(Reference Armand, Borel and Pasquier44)} and impact lipid digestion, as reviewed elsewhere^{(Reference Golding and Wooster30,Reference Singh, Ye and Horne45)} . However, information regarding the specific effect of droplet size on postprandial hypertriglyceridaemia is still scarce. Vors et al. ^{(Reference Vors, Pineau and Gabert46)} evaluated the effect of droplet size of milk and reported higher postprandial chylomicrons, apolipoprotein B-48 and TG concentrations with smaller droplet sizes. However, emulsions were administered together with other food matrices, entangling the extrapolation of the results to droplet size effect only. Similar observations with the same limitations were reported by Tan et al. ^{(Reference Tan, Sun and Goh12)}, Laugerette et al. ^{(Reference Laugerette, Vors and Géloën47)} and Garaiova et al. ^{(Reference Garaiova, Guschina and Plummer48)}. In a recent study, Howard et al. ^{(Reference Howard, Attenbourgh and O’Mahoney49)} tried to overcome this gap by evaluating the effect of two standardised emulsified high-fat meals only differing in fat droplet size on postprandial hypertriglyceridaemia. No other food was given together with the emulsions. They reported a droplet-size-dependent increase of TG concentrations over time, with the increase being more important with fine emulsions than with coarse emulsions. This was in agreement with the observations of the above-mentioned studies that administered emulsions with other foods^{(Reference Tan, Sun and Goh12,Reference Drouin-Chartier, Tremblay and Maltais-Giguère32,Reference Vors, Pineau and Gabert46–Reference Garaiova, Guschina and Plummer48)} . Howard et al. ^{(Reference Howard, Attenbourgh and O’Mahoney49)} showed that a high-fat meal with a smaller lipid droplet size induces a sustained pro-vascular inflammatory and pro-thrombotic milieu, whereas a large lipid droplet size attenuates the rise in vascular inflammatory and thrombotic parameters similarly to a meal with negligible fat content. As this is the first human study evaluating the isolated effect of fat droplet size on vascular inflammation, the results should be considered carefully. Another study showed that emulsion droplet size affected gastrointestinal hormone release and that evenly dispersed, stable, small-emulsion droplets within the stomach lead to prolonged gastric distension and accelerated fat sensing. This prolonged feelings of satiation, which can be seen as something beneficial^{(Reference Steingoetter, Buetikofer and Curcic50)}.

There may be situations where having a small lipid droplet size to provide a fast delivery of lipids/FSM to the body may be beneficial to health, in case of specific supplementations for subgroups of the population such as elderly. Based on the protective effect of FSM, and considering that emulsions with small droplet sizes enhance FSM bioaccessibility/bioavailability^{(Reference Gleize, Hiolle and Meunier43,Reference Tan and McClements51)} , FSM may help to tackle the potential negative effect of fine emulsions on postprandial hypertriglyceridaemia (see below).

Effects of emulsifiers

Polysorbates and carboxymethylcelluloses, two emulsifiers commonly used by the food industry and research community, were shown to have a detrimental effect on health^{(Reference Bancil, Sandall and Rossi52–Reference Viennois and Chassaing54)}, although their effect on postprandial hypertriglyceridaemia has not yet been deeply considered. In an in vivo study in rats, Nassara et al. ^{(Reference Nassra, Bourgeois and Subirade55)} compared the effect of water-in-oil emulsions on postprandial hypertriglyceridaemia. The different emulsions presented similar droplet sizes (9–10 µm) but were stabilised either with synthetic emulsifiers (polyoxyethylene sorbitan monooleate, also known as Tween80, or sodium stearoyl-2-lactylate) or protein-based emulsifiers (sodium caseinate or β-lactoglobulin). The authors concluded that the use of proteins to emulsify lipids has the potential to decrease lipid postprandial response compared with synthetic emulsifiers. The effect of casein and whey proteins – two protein-based emulsifiers – on postprandial hypertriglyceridaemia has also been evaluated by Mariotti et al. ^{(Reference Mariotti, Valette and Lopez56)}. The authors concluded that the type of milk protein did not affect postprandial plasma glucose, amino acids, insulin or NEFA, but reported that caseins markedly reduced postprandial TG and plasma chylomicrons. No significant differences between the meals regarding postprandial oxidative stress, endothelial function or low-grade inflammation were found. In a different study, Keogh et al. ^{(Reference Keogh, Wooster and Golding57)} evaluated the effect of emulsions stabilised with egg lecithin, sodium sterol lactylate, sodium caseinate/monoglyceride and Tween80 on postprandial hypertriglyceridaemia. The interpretation of their results is difficult, as the authors did not report emulsion droplet size and, therefore, it is unclear if reported differences between emulsions are related to droplet size or to the used emulsifier. Authors reported in vitro fatty acid release rate, and emulsions with the highest digestibility rate were not always the emulsions with the greatest impact on TG postprandial concentrations^{(Reference Keogh, Wooster and Golding57)}. This suggests that both droplet size and emulsifiers are important parameters to take into consideration when evaluating the impact of food emulsions on postprandial hypertriglyceridaemia.

Effects of FSM-rich foods

A meta-analysis encompassing the effect of a wide range of foods/food components (sugars, fibre-rich foods, alcohol, fats, proteins, alcohol and more), some of them containing FSM, on postprandial TG response can be found in the literature^{(Reference Lee, Low and Chen60)}, although the specific effect of FSM was not assessed. Some studies indicate a relationship between FSM and TG postprandial response. Available information is summarised in Table 1. For instance, Gomez-Marin et al. ^{(Reference Gomez-Marin, Gomez-Delgado and Lopez-Moreno65)} showed that long-term consumption of a Mediterranean diet, known to be rich in FSM, improves to a greater extent the postprandial TG levels in individuals with type 2 diabetes than just following a low-fat diet. Here, we addressed the linkage between FSM-rich foods, notably tomato meals^{(Reference Arranz, Martínez-Huélamo and Vallverdu-Queralt66,Reference Burton-Freeman, Talbot and Park67)} , legumes^{(Reference Bourdon, Olson and Backus68–Reference Tan, Siow and Peh72)}, nuts^{(Reference Berry, Tydeman and Lewis73–Reference Salas-Salvadó, Guasch-Ferré and Bulló79)}, orange juice^{(Reference Berryman, Grieger and West74,Reference Cerletti, Gianfagna and Tamburrelli80–Reference Sahar, Ur Rahman and Ishaq83)} or certain diets^{(Reference Gomez-Marin, Gomez-Delgado and Lopez-Moreno65,Reference Muñoz-Perez, Gonzalez-Correa and Astudillo-Muñoz84)} (Table 2) and postprandial hypertriglyceridaemia.

Table 1.

Summary of relevant studies regarding the effect of fat-soluble micronutrients from supplements on postprandial hypertriglyceridaemia and/or inflammation

Effects of legumes

The results of Nilsson et al. ^{(Reference Nilsson, Johansson and Ekström69)} suggested that consumption of 101 g (weight before cooking) of legumes (beans) as a starter for dinner can regulate the inflammation response of the first meal during the following day in comparison with the consumption of 89 g of white bread. Indeed, they reported a significant decrease in IL-6 and interleukin 18 (IL-18), as well as in blood glucose and insulin at postprandial state (between 30 and 180 or 30 and 120 min for glucose and insulin, respectively). Although the FSM content of the legumes used in this study was not addressed, it is known that beans do contain FSM (e.g. α-, β- and γ-tocopherols: 4–17 mg/100 g^{(Reference Campos-Vega, Loarca-Piña and Oomah89,Reference Ryan, Galvin and O’Connor90)} ). We hypothesise that not only fibre but also FSM could play a role in regulating this inflammation response. Other studies did not report a significant decrease in commonly used biomarkers, such as interleukins or TG^{(Reference Bourdon, Olson and Backus68,Reference Reverri, Randolph and Steinberg71,Reference Tan, Siow and Peh72)} , or reported a decrease only for certain bean varieties^{(Reference Olmedilla-Alonso, Pedrosa and Cuadrado70)}. However, studies with a lack of decrease in hypertriglyceridaemia biomarkers did show a protective effect against postprandial metabolic stress (decrease of postprandial insulin/enhanced cholecystokinin response associated with reductions of plasma glucose and insulin concentrations in diabetic patients) (see Table 2 for detailed information)^{(Reference Bourdon, Olson and Backus68,Reference Olmedilla-Alonso, Pedrosa and Cuadrado70–Reference Tan, Siow and Peh72)} . As previously stated, legume effects can largely be attributed to their high content in fibre, and the beneficial effect of FSM from legumes is a hypothesis that should be validated. However, Reverri et al. ^{(Reference Reverri, Randolph and Steinberg71)} showed that, compared with fibre-matched or micronutrient-matched meals, the bean meals were more efficient in modulating insulin response, likely highlighting a synergic effect between fibre and micronutrients, supporting our hypothesis.

Table 2.

Summary of relevant studies regarding the effect of foods containing fat-soluble micronutrients on postprandial hypertriglyceridaemia and/or inflammation

* Although FSM content was not reported, legumes are known to contain vitamin E (α-, β- and γ-tocopherols: 4–17 mg/100 g dry weight^{(Reference Campos-Vega, Loarca-Piña and Oomah89,Reference Ryan, Galvin and O’Connor90)} ).

† Although FSM content was not reported, walnuts are known to contain vitamin E and carotenoids (vitamin E: 43 mg/100 g dry weight, carotene: 0.03 mg/100 g dry weight^{(Reference Ni, Zhang and Chen117)}).

‡ Although FSM content was not reported, almonds are known to contain vitamin E (mainly α-tocopherol: 20–80 mg/100 g dry weight^{(Reference Barreca, Nabavi and Sureda92,Reference Yada, Lapsley and Huang93)} ). All values are for orientation.

It is important to note that observations of these studies cannot be attributed only to the effect of fat-soluble micronutrients, as these foods/meals contain other compounds with a potential beneficial effect on the measured biomarkers (e.g. fibre).

Effect of supplements

Extensive evidence has demonstrated that antioxidants such as vitamin E or carotenoids (but also vitamin C and polyphenols) have protective effects in preventing cardiovascular diseases^{(Reference Wang, Chun and Song94)}. Some authors have reported specific beneficial effects of FSM (Table 1) and foods/meals containing FSM (Table 2) during the postprandial state. For example, a study by Deplanque et al. ^{(Reference Deplanque, Muscente-Paque and Chappuis61)} in a randomised, double-blind, parallel-group, placebo-controlled study, with 146 healthy normal-weight subjects reported that consumption of a mix of phytonutrients for a period of 2 weeks (a mixture of lycopene and phytosterols: 15 mg of each), phytoene and phytofluene (4 mg total) and β-carotene (0·5 mg)) was enough to increase plasma levels of lycopene, phytofluene and phytoene after a high-fat meal (postprandial state, 8 h after food intake) and decreased oxidised LDL concentration in comparison with the control group. These authors did not observe significant differences in glucose, insulin and TG levels during the postprandial state (8 h after food intake, except for insulin that was measured 2 h after food intake), although values were lower for the supplemented group.

Some information about the effect of vitamin E in postprandial hypertriglyceridaemia can also be found in the literature. Bae et al. ^{(Reference Bae, Schwemmer and Lee62)} reported that vitamin E supplementation of a high-fat meal did not reduce serum TG concentration, although they observed protective effects of vitamin E on endothelial function when measuring brachial artery vasodilation during the postprandial period (changed from 13·3% to 6·6% (p ≤ 0·05), 7·1% (p ≤ 0·05) or 13·2% at 2, 4 or 6 h after eating a high-fat meal). In later studies, Neri et al. ^{(Reference Neri, Calvagno and Mauceri63)} and Hejazi et al. ^{(Reference Hejazi, Dabbaghmanesh and Mazloom64)} reported that serum TG profile at postprandial state was not significantly modified after vitamin E supplementation; even though Neri et al. ^{(Reference Neri, Calvagno and Mauceri63)} supplementation also contained vitamin C. However, Hejazi et al. ^{(Reference Hejazi, Dabbaghmanesh and Mazloom64)} observed a decrease in malondialdehyde levels (a lipid peroxidation marker) in the postprandial state (2 h after food intake) in comparison with the placebo group. Neri et al. ^{(Reference Neri, Calvagno and Mauceri63)} also reported that, even though no effect on lipid profile was found, vitamin E treatment in healthy subjects and subjects with impaired glucose tolerance induced a significant postprandial improvement in redox balance parameters, serum CRP and VCAM-1 levels. No other study focusing on the effect of single FSM is available.

Overall, published data suggest that the presence of FSM in foods tends to decrease postprandial hypertriglyceridaemia and/or inflammation. However, available studies are still limited and therefore should be considered carefully. More research is needed to fully understand the mechanisms underlying FSM effects in this context.

miRNAs

miRNAs are a class of gene expression regulators^{(Reference Catalanotto, Cogoni and Zardo97)} that can potentially be interesting biomarkers to evaluate postprandial hypertriglyceridaemia after a high-fat meal. Indeed, monocytes can interact with postprandial TRL before they migrate to the endothelium^{(Reference Peng and Wu27,Reference Varela, Ortega and Bermudez98)} , affecting the transcription of genes involved in lipid homoeostasis and inflammation. Measurement of miRNAs can be used for early detection of asymptomatic and symptomatic diseases^{(Reference Navickas, Gal and Laucevičius99 Reference Karkeni, Bonnet and Marcotorchino,100)}, including cardiovascular ones. Certain miRNAs likely modulate cholesterol efflux and reverse cholesterol transport in mammals, as reviewed by Dávalos and Fernández-Hernando^{(Reference Dávalos and Fernández-Hernando101)}. In a study in rainbow trout, Zhu et al. ^{(Reference Zhu, Corraze and Plagnes-Juan102)} evaluated the different expression of miRNAs known to be involved in cholesterol metabolism when feeding the animals with either a vegetable- or marine-based diet, to address whether miRNAs can be used as biomarkers to determine postprandial hypertriglyceridaemia. These authors reported that a plant-based diet significantly reduced the expression of miR-223, probably a consequence of the absence of dietary cholesterol, in agreement with data obtained in mice^{(Reference Vickers, Landstreet and Levin103)} or in humans^{(Reference Talbot, Mensink and Smolders104)}. However, authors also observed discrepancies for other miRNAs, suggesting that there is a limitation regarding the extrapolation of results between different species^{(Reference Zhu, Corraze and Plagnes-Juan102)}.

The potential use of miRNAs as a biomarker to determine hypertriglyceridaemia/inflammation has been suggested by several studies^{(Reference Lopez, Bermudez and Montserrat-de la Paz7,Reference Ortega, Cardona-Alvarado and Mercader105)} , with different focuses on how to use miRNAs as biomarkers. miRNAs can potentially be used to evaluate differences between healthy and metabolically disturbed/intervened individuals at a fasting state (lipaemia)^{(Reference Ortega, Cardona-Alvarado and Mercader105,Reference Ortega, Mercader and Catalán106)} or to evaluate the differences in the same individuals but at a different state (fasting versus postprandial state)^{(Reference Lopez, Bermudez and Montserrat-de la Paz7,Reference Quintanilha, Pinto Ferreira and Ferreira9,Reference Daimiel, Micó and Valls107,Reference Mantilla-Escalante, de las Hazas and Gil-Zamorano108)} . Table 3 summarises the main results of these studies. The circulating miRNA profile of adults with obesity performed by Ortega et al. ^{(Reference Ortega, Mercader and Catalán106)} suggested that the miRNA profile in plasma was different when comparing obese with non-obese subjects at fasting state. According to these results ^{Reference Ortega, Mercader and Catalán} , patients with morbid obesity showed a significant increase in the levels of the miRNAs miR-140-5p, miR-142-3p and miR-222 and a significant decrease in the levels of miR-532-5p, miR-125b, miR-130b, miR-221, miR-15a, miR-423-5p and miR-520c-3p, compared with healthy individuals at fasting state. In a later study by the same research group^{(Reference Ortega, Cardona-Alvarado and Mercader105)}, a comparison of miRNA levels of sedentary individuals before and after 8 weeks of a normocaloric diet enriched with 30 g of nuts was performed. As in their previous study, ^{Reference Ortega, Cardona-Alvarado and Mercader} authors observed differences between miRNA profiles before and after the diet intervention. Interestingly, they reported a decrease in mi-R221, which is somehow contradictory with the results of their previous study (mi-R221 concentration being lower in obese than in healthy individuals), which highlights the different regulation of this miRNA in plasma in healthy versus obese/metabolically compromised subjects.

Table 3.

Relevant human studies regarding the expression of miRNAs related to postprandial hypertriglyceridaemia/inflammation

Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)} also evaluated differential miRNA expression between fasting and postprandial states (2 h after food consumption) in peripheral blood mononuclear cells (PBMCs) of healthy humans. Compared with miRNA expression signature at fast (and when following the inclusion criteria of >1·2-fold under- or overexpression and statistical difference (p < 0·05)), the authors reported nine down-regulated and nine up-regulated miRNAs in the postprandial state (Table 3). miRNAs hsa-miR-223-3p and hsa-miR-223-5p – the above-mentioned miRNAs involved in cholesterol metabolism in different animal models^{(Reference Zhu, Corraze and Plagnes-Juan102–Reference Talbot, Mensink and Smolders104)} – were also detected in this study (see Table S1 of Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)}) but did not meet the criteria to be selected as overexpressed miRNAs. In a later study, Mantilla-Escalante et al. ^{(Reference Mantilla-Escalante, de las Hazas and Gil-Zamorano108)} evaluated the whole mouse miRNome and tried to establish a relationship between the expression of mouse and human miRNAs in the postprandial state. Not all the selected miRNAs from the mouse model (Table 3) showed the same trend in human subjects, but some of these miRNAs (hsa-miR 206, hsa-miR 409-3p and hsa-miR 27b-5p) were up-regulated both in mice and in humans. Interestingly, Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)} also reported an increase of hsa-miR 206 and hsa-miR 27b-5p in the postprandial state, although the increase was not sufficient to meet their inclusion criteria. Contrary to the results of Mantilla-Escalante et al. ^{(Reference Mantilla-Escalante, de las Hazas and Gil-Zamorano108)}, Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)} found a lower expression of mRNAs hsa-miR 409-3p in the postprandial state. Regarding the other miRNAs selected by Mantilla-Escalante et al. ^{(Reference Mantilla-Escalante, de las Hazas and Gil-Zamorano108)}, Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)} reported lower levels of hsa-miR-10a-3p and hsa-miR-340-3p, and higher levels of hsa-miR-543 and hsa-miR-125a-3p when comparing fasting with postprandial state, although, again, the difference remained insufficient to meet their inclusion criteria. More recently, Quintanilha et al. ^{(Reference Quintanilha, Pinto Ferreira and Ferreira9)} studied the effect of high-fat high-saturated meal ingestion on plasma miRNA expression during the postprandial period in healthy women (Table 3). When comparing their results with the results of Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)}, we observed a shared trend between results (only hsa-miR 200c-3p and hsa-miR 92b-3p showed different tendencies), although in the study of Lopez et al. (2018) none of the miRNAs identified by Quintanilha et al. ^{(Reference Quintanilha, Pinto Ferreira and Ferreira9)} meet the inclusion criteria, which could be influenced by the sex difference between studies. Finally, Daimiel et al. ^{(Reference Daimiel, Micó and Valls107)} evaluated the effect of olive oil consumption on miRNAs during postprandial state, focusing on fifty-three cardiovascular-related miRNAs. They found that, regardless the polyphenol level in the olive oil, has-let-7e-5p was down-regulated in the postprandial period (Table 3), in agreement with the trend presented by Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)}. Conversely, observations for miR-328a-3p were contradictory, as it was decreased in one study^{(Reference Daimiel, Micó and Valls107)} and increased in another^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)}. Moreover, Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)} observed almost no differences between fasting and postprandial state for miR-20a-5p.

These differences between studies highlight the need for more and larger human in vivo studies to be able to define a reliable range of miRNA biomarkers suitable to assess human postprandial hypertriglyceridaemia. One of the limitations of the available studies is that they tend to focus (or only show the results) on certain miRNAs instead of reporting the whole miRNA human profile, which restricts the comparison with the results of other studies, unless compared studies evaluate the same miRNAs. Only Lopez et al. ^{(Reference Lopez, Bermudez and Montserrat-de la Paz7)} reported all miRNAs. Since the expression pattern of miRNAs regulates gene expression and, in some cases, can be associated with different human pathologies, the identification of problematic miRNAs can be of great use for the research community both to try to regulate their expression and to explore their use as relevant biomarkers.

Circulating microvesicles

Endothelial microvesicles can be released into the blood stream from the endothelium as a response to activation, injury or apoptosis of endothelial cells^{(Reference Dignat-George and Boulanger109)}. Associations between high levels of these microvesicles, endothelial dysfunction and obesity have been reported^{(Reference Esposito, Ciotola and Schisano110–Reference Helal, Defoort and Robert112)}. Their potential use as biomarkers to determine the effect of fat consumption in the postprandial state, both in healthy and in non-healthy individuals, has been already evaluated by some studies^{(Reference Araujo, Silva and Guerra8,Reference Bulut, Jelich and Dacanay-Schwarz113)} that reported a significant increase after fat consumption. However, the levels of microvesicles seem to be also affected by blood flow^{(Reference Jenkins, Padilla and Boyle114,Reference Silva, Sales and Araujo115)} . This could affect the interpretation of the results, although Araujo et al. ^{(Reference Araujo, Silva and Guerra8)} showed that the circulation of endothelial microvesicles increases in a similar way in individuals with postprandial lipaemia and those with postprandial lipaemia plus disturbed blood flow. According to a recent study by Kumar et al. ^{(Reference Kumar, Sundaram and Mu116)}, it is not only the blood concentration of endothelial microvesicles that changes after fat consumption, but also their fat composition. High-fat diet dramatically changes the lipid profile of intestinal epithelial exosomes from predominantly phosphatidylethanolamine to phosphatidylcholine, which results in the inhibition of the insulin response and therefore contributes to insulin resistance^{(Reference Kumar, Sundaram and Mu116)}. The use of microvesicles as biomarkers has not been fully validated yet, but the available results support their high potential.