Fruit phenolic compounds: qualifying fruits

Fruits, like other plant foods, are natural vehicles of phenolic compounds; however, they differ remarkably in content, composition and bioavailability⁽Reference Crozier, Jaganath and Clifford^26–Reference Szajdek and Borowska²⁸⁾. As such, there has been tremendous effort to provide a common system for qualifying and ranking fruits and linking this system to expected biological activity in human subjects. However, so far, there is no unifying scheme that relates varieties of fruit to specific health benefits, specifically due to the incomplete nature of the available data.

Fruits may be classified based on the quantity of dominant phenolic compounds present in them; for instance, quercetin in apples and anthocyanins in strawberries. Fruits may also be classified by their antioxidant capacity using methods such as the Folin–Ciocalteu assay (expressed as total phenols content) or the Oxygen Radical Absorbance Capacity assay. An alternative system is based on the quantity and quality of antioxidants derived from phenols present in fruit, which is expressed as an index (PAOXI)⁽Reference Vinson, Dabbagh and Serry²⁹⁾. Free and total phenol quality is measured by the ability of the fruit to inhibit the oxidation of LDL and VLDL induced by cupric ion. Using this indexing system, Vinson et al. ⁽Reference Vinson, So and Zubik³⁰⁾ evaluated the phenols present in twenty commonly consumed fruits in the American diet. They reported that among the twenty fruits tested, berries are the best source of polyphenol antioxidants, and therefore have the greatest likelihood of delivering an effect in vivo, since concentrations of fruit phenolic compounds in plasma are generally very low (1 μm or less). They also reported that fruits had significantly higher PAOXI values compared with vegetables. Whether higher PAOXI values of fruits translates to greater protection from chronic diseases compared with vegetables is uncertain.

Fruit phenolic compounds: present intake

Intake of phenolic compounds can range several fold depending on individual dietary patterns. Estimates of intake also rely on methods of assessment. Various estimates suggest American's consume 20 mg of flavonoids a day⁽Reference Sampson, Rimm and Hollman¹⁶⁾, based predominately on flavonol and flavone intake, and up to approximately 190 mg daily⁽Reference Chun, Chung and Song³¹⁾, based on intake of flavan-3-ols, flavanones, flavonols, anthocyanidins, flavones and isoflavones. In a dataset of ninety-two students using 7 d food records, Cao et al. (2009)⁽Reference Cao, Zhang and Chen³²⁾ reported mean flavonol (quercetin and kaempferol) intake of 28·55 mg/d. Primary food sources included tea⁽Reference Sampson, Rimm and Hollman^16, Reference Chun, Chung and Song³¹⁾, onions⁽Reference Sampson, Rimm and Hollman¹⁶⁾, apples⁽Reference Sampson, Rimm and Hollman^16, Reference Cao, Zhang and Chen³²⁾, broccoli⁽Reference Sampson, Rimm and Hollman¹⁶⁾, citrus fruit juices⁽Reference Chun, Chung and Song³¹⁾, wine⁽Reference Chun, Chung and Song³¹⁾, potato⁽Reference Cao, Zhang and Chen³²⁾, celery⁽Reference Cao, Zhang and Chen³²⁾ and citrus fruits⁽Reference Sampson, Rimm and Hollman¹⁶⁾. Flavonols and flavones from other identified fruits/fruit sources, namely blueberries and strawberries, have been reported low, < 1 %⁽Reference Sampson, Rimm and Hollman¹⁶⁾. In a seven countries report by Hertog et al. ⁽Reference Hertog, Kromhout and Aravanis³³⁾, flavonoid intake ranged from 2·6 mg/d in West Finland to 68·2 mg/d in Ushibuka, Japan⁽Reference Hertog, Kromhout and Aravanis³³⁾. Consistent with the US reports, quercetin contributes most of the flavonoid intake, contributing >50 % in most countries and up to 100 % in West Finland⁽Reference Hertog, Kromhout and Aravanis³³⁾. In our own estimates of dietary flavonoids intake using the United States Department of Agriculture Flavonoid database 2·1⁽³⁴⁾ and using dietary food intake records from a recently completed study in our lab (unpublished results), we found that the diets of pre- and post-menopausal women in the greater Chicago, IL, area contained anywhere from < 1 mg/d to approximately 400 mg/d; or 0–292 mg/4184 kJ. Flavan-3-ol from tea intake contributed greatly to the higher dietary flavonoid intakes. Dietary flavonoids intake from meal plans taken from the dietary approaches to stop hypertension diet studies⁽Reference Most^{12 )}were estimated at approximately 4·6 mg/4184 kJ for the control diet and approximately 13·7 mg/1000 kcal for the dietary approaches to stop hypertension diet. These estimates were calculated using the 2003 version of the United States Department of Agriculture Flavonoids database (version 2.0), of which only approximately 60 % of foods were found in the database. Additionally, the dietary approaches to stop hypertension diet was designed to target the action of non-flavonoid dietary constituents (K, Ca and Mg), but emphasised fruit and vegetable intake. Hence, dietary flavonoids intake, including classes of flavonoids, can vary immensely depending on food choice. An important additional consideration is timing of intake of dietary flavonoid and non-flavonoid phenolic compounds. Minimising the consequences of energy-dense meals may be an important feature of dietary phenolic antioxidant compounds on human health.

Fruit phenolic compounds: Mechanisms of action for health benefit

Fruit phenolic compounds may provide benefit to human subjects via several mechanisms⁽Reference Nijveldt, van Nood and van Hoorn³⁵⁾. The best-described and most well-known mechanism is through their antioxidant properties and modulation of biological oxidative stress to prevent damage to cellular lipids, proteins and DNA. Directly, they may scavenge superoxide and other reactive oxygen species (ROS) such as hydroxyl and peroxy radicals. For example, from a thermodynamic standpoint, the chemical nature of flavonols lends them favourable to effectively scavenge superoxide anions, singlet oxygen and lipid peroxy radicals⁽Reference Galleano, Verstraeten and Oteiza³⁶⁾. Indirectly, some flavonoids may spare/recycle endogenous antioxidant (e.g. glutathione, urate, vitamin E and vitamin C) and selected flavonoids, such as catechins and quercetin, may chelate redox active metals, such as Fe and Cu, thereby removing a causal factor in free radical generation⁽Reference Galleano, Verstraeten and Oteiza^36–Reference Ferrali, Signorini and Caciotti³⁸⁾. Flavonoid–metal complexes have also been reported to mimic superoxide dismutase activity providing another mechanism of ‘antioxidant defence’ benefit⁽Reference Kostyuk, Potapovich and Strigunova³⁹⁾. However, the biological relevance of these described health-promoting mechanisms of flavonoid, such as free radical scavenging and metal chelation is questioned due to a kinetically unfavourable situation with respect to endogenous compounds such as glutathione, urate, vitamin C and a panoply of endogenous metal chelators that exist in plasma and tissues at much higher concentrations than flavonoid compounds (μm-mm v. nm-lower μm, respectively)⁽Reference Galleano, Verstraeten and Oteiza³⁶⁾. Similarly, questions have been raised about fruit-derived phenolics on the sparing or recycling of certain endogenous antioxidant defence compounds. Fruits deliver fructose along with phenolic compounds and some of the actions often ascribed to phenolics may be due to fructose. Fructose has been known for years to increase plasma urate levels through fructokinase-mediated metabolism to fructose 1-phosphate⁽Reference Maenpaa, Raivio and Kekomaki^40, Reference Heuckenkamp and Zollner⁴¹⁾. An increase in urate concentrations commensurate with increased plasma antioxidant capacity, which has been reported after strawberry, spinach and apple consumption⁽Reference Cao, Russell and Lischner^42, Reference Lotito and Frei⁴³⁾. Further, Lottito & Frei⁽Reference Lotito and Frei^{43 )}reported parallel increases in urate and plasma antioxidant capacity after human volunteers consumed red delicious apples or fructose solution, each providing approximately 64 g fructose. Future studies on the antioxidant effects of flavonoid-rich foods should consider the effects of other non-phenolic contributors, such as the content of fructose, sucrose and sorbitol.

Indirect mechanisms of flavonoids involving lipid–flavonoid or protein–flavonoid interactions may be considered from a different perspective. Interactions requiring higher molecular specificity can withstand lower availability of substrate. Hence, alterations in membrane and protein function that involve lipid–flavonoid or protein–flavonoid interaction can happen at very low flavonoid concentrations and have significant biological effects at flavonoid concentrations attainable through diet. Stimulating endogenous antioxidant defence systems (e.g. superoxide dismutase and glutathione peroxidase) and/or inhibiting enzymes that generate large amounts of ROS such as xanthine oxidase and NAD(P)H oxidase are examples of ‘indirect’ antioxidant effects derived from protein–flavonoid interactions. Inhibiting absorption of already oxidised products, such as lipid hydroxyperoxides⁽Reference Ursini, Zamburlini and Cazzolato^8, Reference Ursini and Sevanian⁴⁴⁾, may be yet another mechanism of benefit by some phenolic compounds.

Over the last several years, polyphenolic compounds have been studied for their action in cellular signalling: modifying pathways of inflammation, insulin action, platelet function and vascular relaxation⁽Reference Nijveldt, van Nood and van Hoorn³⁵⁾. For example, it is known that oxidants increase pro-inflammatory transcription factor NF-κB, whereas antioxidants such as pyrrolidine dithiocarbamate and N-acetyl cysteine inhibit NF-κB activation⁽Reference Schreck, Meier and Mannel⁴⁵⁾. Polyphenolic compounds from red wine have been shown to modulate NF-κB activation, suggesting a redox-mediated action of red wine polyphenols on inflammatory pathways⁽Reference Blanco-Colio, Valderrama and Alvarez-Sala⁴⁶⁾. Extracts of strawberry rich in anthocyanins have been shown to attenuate oxidative stress (H₂O₂)-induced impairment of insulin signalling by restoring insulin receptor substrate-1 activation in human skeletal muscle cells⁽Reference Sandhya, Tadapaneni and Banaszewski⁴⁷⁾, an effect also shown in hepatoma cells with tea-derived ( − )-epigallacatechin-3-gallate⁽Reference Lin and Lin⁴⁸⁾. Nitric oxide (NO)-stimulated endothelium-dependent relaxation by red wine, grape seed extract (GSE) and strawberries is due to activation of the protein kinase B (Akt)/phosphatidylinositol-3 kinase pathway, also well known for its regulation by cellular redox status⁽Reference Ndiaye, Chataigneau and Lobysheva^49–Reference Edirisinghe, Burton-Freeman and Varelis⁵¹⁾. Overall, phenolic compounds found in plant foods, including fruits, have multiple paths for benefiting human health, most notably, through their actions as antioxidants and modifying cellular events. Their specific actions are likely to be dependent on the composition and time course of compound-/metabolites appearing in plasma⁽Reference Crozier, Jaganath and Clifford^26, Reference Manach, Williamson and Morand^52, Reference Hollman, van Trijp and Buysman⁵³⁾. Continued advancement in technology and investment in research will help to elucidate these details.

Postprandial glycaemia and insulinaemia

Postprandial hyperglycaemia is widely recognised as producing oxidative stress and is associated with exacerbating a number of disease and disease-related conditions, such as diabetes and atherosclerotic CVD, metabolic syndrome, hypertension and obesity⁽Reference Laakso^57–Reference Stratton, Adler and Neil⁵⁹⁾ as depicted in Fig. 1. Hyperglycaemia is associated with the generation of ROS and reactive nitrogen species, respectively, arising from several intracellular sources. In normal physiology, antioxidant defence systems balance ROS/reactive nitrogen species. When these systems are overwhelmed, excess ROS/reactive nitrogen species alters redox balance increasing the activation of pro-inflammatory pathways (e.g. NF-κB), impairing cellular signalling (e.g. insulin signalling) and causing irreversible modifications of cellular and non-cellular components⁽Reference Williams, Spencer and Rice-Evans⁶⁷⁾. Hyperglycaemia can promote non-enzymatic glycation of proteins (Fig. 1), such as LDL⁽Reference Nivoit, Wiernsperger and Moulin⁶⁸⁾, and can increase the susceptibility of LDL to oxidation⁽Reference Ceriello, Quagliaro and Catone^69, Reference Diwadkar, Anderson and Bridges⁷⁰⁾. Oxidised LDL is implicated in the initiation, progression and complication of atherosclerotic CVD⁽Reference Steinberg and Lewis^71, Reference Berliner, Novab and Fogelman⁷²⁾. Hence, strategies to modify postprandial glycaemia to favour a balanced system are warranted.

Fig. 1 Postprandial dysmetabolic-induced oxidative stress and possible mechanisms for fruit phenolic action. FFA, free fatty acid; sd LDL, small dense LDL; ox LDL, oxidised LDL; NO, nitric oxide.

The glycaemic response to a meal is driven by preabsorptive and postabsorptive factors, including characteristics of the meal (dose and digestibility of carbohydrate, co-existing dietary factors and food matrix), alimentary metabolism and a variety of associated neuroendocrine responses, most notably the secretion and action of insulin to facilitate glucose clearance from the blood. The ability of fruits, based on their inherent phenolic composition, to favourably augment glucose metabolism would likely be through one of these factors. Interference with glucose absorption and/or modifying insulin-mediated glucose metabolism are two probable mechanisms; the latter having relevance in both glucose and lipid metabolisms and linking the antioxidant properties of fruit phenolics with cellular redox-sensitive pathways.

Few studies have examined the relationship between fruit phenolic compounds and postprandial glycaemia. Ten human trials were selected for the review (Table 1). These trials met criteria that included a minimum 2 h postprandial assessment of plasma glucose and included adequate controls or comparative data to evaluate the effect of the studied fruit on glycaemia. Criteria for delivery of fruit with a meal was desired, but was not mandatory given the small number of published trials. Cranberry juice⁽Reference Wilson, Singh and Vorsa^73, Reference Vinson, Bose and Proch⁷⁴⁾, wild blueberry⁽Reference Kay and Holub⁷⁵⁾, strawberry⁽Reference Cao, Russell and Lischner⁴²⁾ and a fruit meal of banana, kiwi and apple⁽Reference Erdmann, Topsch and Lippl⁷⁶⁾ were the subject of five trials. Five additional trials were included that tested red wine or components of red wine with a meal⁽Reference Ceriello, Bortolotti and Motz^54, Reference Natella, Ghiselli and Gudi^77–Reference Pal, Naissides and Mamo⁷⁹⁾ or was a comparator in another study⁽Reference Cao, Russell and Lischner⁴²⁾. Of these ten trials, six trials provided fruit as a beverage or in a beverage formula with meals⁽Reference Ceriello, Bortolotti and Motz^54, Reference Cao, Russell and Lischner^42, Reference Kay and Holub^75, Reference Natella, Ghiselli and Gudi^77–Reference Pal, Naissides and Mamo⁷⁹⁾. As shown in Table 1, no remarkable effects on plasma glucose were reported overall. Only one trial reported a possible action in modifying glucose metabolism⁽Reference Wilson, Singh and Vorsa⁷³⁾. In the trial by Wilson et al. ⁽Reference Wilson, Singh and Vorsa⁷³⁾, incremental area under the glucose curve after normal energy cranberry and energy control drinks was not different; however, the pattern of glycaemia was modified in the cranberry-containing drink, suggesting delayed absorption or altered distribution of glucose to insulin-sensitive tissues. Quercetin, a major phenolic compound in cranberry, has been shown to inhibit gastric uptake of glucose in a porcine model⁽Reference Cermak, Landgraf and Wolffram⁸⁰⁾. Quercetin and myricetin have been demonstrated to inhibit GLUT-4-mediated uptake by adipocytes⁽Reference Strobel, Allard and Perez-Acle⁸¹⁾; whereas insulin-sensitising actions of quercetin in liver and muscle have been reported in fructose-fed rat model of insulin resistance⁽Reference Kannappan and Anuradha⁸²⁾. Based on this limited evidence, no specific claims can be made relative to postprandial glucose control and cranberry or its associated phenolic compounds. Likewise, no evidence supports a glucose-lowering effect of consuming polyphenolic-rich fruits or beverages with meals. One trial reported an increase in postprandial glucose with wild blueberry consumption delivered in a beverage with a high-fat meal⁽Reference Kay and Holub⁷⁵⁾. The elevated glucose may be due to differences in mono-and di-saccharide content of treatments and/or the higher fibre content of the wild blueberry altering glucose absorption and metabolism. Insulin concentrations or other metabolic indices were not reported to help explain the results.

Table 1 Clinical trials examining postprandial glucose and insulin responses after fruit-associated polyphenolic treatments

T2DM, Type 2 diabetes mellitus; RW, red wine; EtOH, ethyl alcohol; OW, overweight or obese; M, male; F, female; post-M, postmenopausal; dyslip, dyslipidaemia; Pbo, placebo; Glu, glucose; Ins, insulin; J, juice; Vit C, vitamin C; Str, strawberry; Ctrl bev, control beverage; WBB, wild blueberry; CB, cranberry; AUC, area under the curve; FD, freeze-dried; carb, carbohydrate; veg, vegetable; LoJ, low kcal juice; NmJ, normal juice; LoHFCS J, low high fructose corn syrup juice; NmHFCS J, normal high fructose corn syrup juice; ↓ , decrease or reduction (P < 0·05); ↑ , increase (P < 0·05); –, neutral; no effect (P>0·05) compared with placebo or baseline (i.e. comparator trials).

* Describes general health, age or weight status. Healthy indicates normolipidaemic, between 20 and 70 years, not overweight or obese and no documented disease or condition.

† Where meal is indicated, fruit or fruit-based beverage was consumed with the meal.

‡ Meal is likely the same as that provided in reference Naissides et al. ⁽Reference Naissides, Mamo and James⁷⁸⁾.

§ Time point that a second meal was introduced. The results represent the glucose and insulin responses to this point.

The postprandial insulin response is critical in the management of postprandial glucose metabolism as well as lipid and lipoprotein metabolisms. Insulin signalling is sensitive to cellular redox balance⁽Reference Bashan, Kovsan and Kachko⁸³⁾, suggesting that fruits rich in antioxidant compounds could beneficially impact nutrient metabolism through improved or enhanced insulin signalling in insulin-sensitive tissues. Five of the ten trials identified reported changes in postprandial insulin concentrations. Four reported no differences among treatments⁽Reference Ceriello, Bortolotti and Motz^54, Reference Wilson, Singh and Vorsa^73, Reference Erdmann, Topsch and Lippl^76, Reference Pal, Naissides and Mamo⁷⁹⁾ and one⁽Reference Naissides, Mamo and James⁷⁸⁾ reported elevated postprandial insulin concentrations after red wine consumption compared with dealcoholised red wine or water with a standardised meal of 75 g cheese and 50 g fat-free crackers in postmenopausal women. Recent work in our lab suggests that a strawberry-based drink delivering approximately 126 mg total flavonoids (approximately 82 mg anthocyanins) consumed with a moderate fat/carbohydrate meal reduces the postprandial insulin response in overweight men and women compared with when the same meal is consumed with a placebo (non-flavonoid containing) drink⁽Reference Edirisinghe, Krishnankutty and Cappozzo⁸⁴⁾. No differences in postprandial glucose concentrations were observed between treatment conditions. These data suggest reduced insulin requirement to achieve glucose homoeostasis, supporting improved postprandial insulin sensitivity, an effect that appears to be mediated through redox-sensitive insulin signalling pathways⁽Reference Kannappan and Anuradha^82, Reference Edirisinghe, Krishnankutty and Cappozzo⁸⁴⁾. More studies are needed to elucidate the potential effects of fruit phenolics on insulin action in human subjects.

Postprandial lipaemia

Elevated postprandial TAG are associated with increased risk for the development and progression of atherosclerotic CVD⁽Reference Patsch, Miesenbock and Hopferwieser⁸⁵⁾ and can be more discriminating than fasting concentrations in assessing high-risk atherosclerotic conditions⁽Reference Bansal, Buring and Rifai⁸⁶⁾. Fasting and postprandial hyperlipidaemia is apparent in individuals with obesity, diabetes mellitus and metabolic syndrome⁽Reference van Oostrom, van Wijk and Cabezas^9, Reference Patsch, Miesenbock and Hopferwieser^85, Reference Alipour, Elte and van Zaanen⁸⁷⁾. These conditions are also characterised by elevated NEFA, low-grade inflammation and impaired insulin sensitivity⁽Reference Alipour, Elte and van Zaanen⁴⁾, all of which promote and exacerbate disturbances in lipid metabolism. It is important to note that inflammatory mediators alone can trigger insulin resistance in cells, in the absence of obesity and other factors⁽Reference Shoelson, Lee and Goldfine⁸⁸⁾. For example, ingestion of a high-fat or glucose challenge meal provokes an acute inflammatory response⁽Reference Bashan, Kovsan and Kachko^83, Reference Hyson, Paglieroni and Wun^89–Reference Mohanty, Hamouda and Garg⁹²⁾. Excessive intake of energy-dense, nutrient-poor foods provides potent substrate for inducing postprandial inflammation. Several lines of evidence, including work in our laboratory, suggest that the acute response arises from induction of oxidative stress and stimulated inflammatory pathways leading to reduced insulin sensitivity⁽Reference Edirisinghe, Krishnankutty and Cappozzo^84, Reference Shoelson, Lee and Goldfine^88–Reference Mohanty, Hamouda and Garg⁹²⁾ (Fig. 1). Under these conditions, lipaemia is magnified and the relative composition of all lipoproteins changes supporting an atherogenic state of the plasma. VLDL remnants are cholesterol ester enriched and in exchange, HDL become TAG rich and cholesterol depleted, which after further modification leads to the dissociation of the structurally important protein apo A-I, resulting in clearance and reduced HDL. Through a similar process, LDL become small and more susceptible to oxidation⁽Reference Alipour, Elte and van Zaanen^4, Reference Alipour, Elte and van Zaanen⁸⁷⁾. Oxidised LDL are proposed to play a causative role in early atherogenesis through their ability to up-regulate scavenger receptors on activated monocytes, transforming them into macrophages, eventually leading to foam cell formation⁽Reference Fuhrman and Aviram⁹³⁾.

Another aspect of postprandial lipaemia is endothelial dysfunction (Fig. 1). Many endothelial-derived factors, of which NO plays a central role, regulate vasomotion, permeability, proliferation and smooth muscle cell migration. Elevated TAG and NEFA are associated with reduced NO, a result that appears related to oxidative stress generation⁽Reference van Oostrom, van Wijk and Cabezas^9, Reference Ceriello, Taboga and Tonutti⁵⁶⁾. The oxidative imbalance results in reduced endothelial NO synthase activity and peroxynitrile generation and activation/up-regulation of NF-κB and NF-κB-dependent factors, such as monocyte chemoattractant protein-1, cellular adhesion molecules (e.g. intercellular adhesion molecule and vascular adhesion molecule), cytokines, growth factors and metalloproteinases, as well as pro-thrombotic factors including increased plasminogen activator inhibitor-1 and tissue factor to name a few⁽Reference Kris-Etherton, Lefevre and Beecher⁹⁴⁾. Overall, the lipaemia-induced oxidative stress promotes a pro-inflammatory, pro-thrombotic unstable atherogenic environment. In principle, preventing or decreasing the postprandial-induced oxidative stress or the associated inflammatory response should interrupt the metabolic–oxidative–inflammatory circle providing balance to the system and resistance to pathology. This may be one mechanism by which eating fruits and vegetables regularly favour reduced disease risk.

Focused on fruits and their inherent phenolic quality to impart an action on postprandial lipaemia or associated events, fifteen trials using the postprandial testing paradigm were reviewed and are listed in Table 2. Changes in postprandial TAG were reported in twelve trials, of which one of the twelve trials⁽Reference Pal, Naissides and Mamo⁷⁹⁾ reported changes in plasma TAG and chylomicron (B-48) concentrations postmeal and another⁽Reference Blanco-Colio, Valderrama and Alvarez-Sala⁴⁶⁾ reported changes in TAG, chylomicron and VLDL concentrations postmeal. Overall, seven trials reported no specific advantage on postprandial TAG management when consuming wild blueberry⁽Reference Kay and Holub^75, Reference Mazza, Kay and Cottrell⁹⁵⁾, cultivated blueberry⁽Reference Serafini, Testa and Villaño⁹⁶⁾or red wine^(54,77,79) or GSE⁽Reference Sugiyama, Akazome and Shoji⁹⁸⁾ with a high/moderate fat meal (or whole milk⁽Reference Serafini, Testa and Villaño⁹⁶⁾). Serafini et al. ⁽Reference Serafini, Testa and Villaño⁹⁶⁾ reported a 6 % increase in total cholesterol at 5 h after consumption of 200 g blueberry with whole milk compared with the same dose of fruit with water; however, three other trials showed no changes in postprandial total cholesterol, LDL or HDL⁽Reference Kay and Holub^75, Reference Natella, Ghiselli and Gudi^77, Reference Natella, Belelli and Gentili⁹⁷⁾. Three trials reported decreased postprandial TAG after consuming strawberry⁽Reference Burton-Freeman, Linares and Hyson¹⁰⁾, apple procyanidins (600 mg capsule)⁽Reference Sugiyama, Akazome and Shoji⁹⁸⁾ and cranberry juice⁽Reference Vinson, Bose and Proch⁷⁴⁾ with moderate fat, high fat or no fat meals, respectively. The TAG-lowering effect of fruits or their active constituents may have been by inhibition of pancreatic lipase activity⁽Reference Sugiyama, Akazome and Shoji⁹⁸⁾, reduced dietary TAG absorption⁽Reference Pal, Naissides and Mamo⁷⁹⁾ and/or by enhanced clearance, possibly through antioxidant-mediated enhanced postprandial insulin sensitivity⁽Reference Décordé, Teissedre and Sutra⁹⁹⁾. Two trials reported elevated TAG when drinking red wine with a fat-enriched meal⁽Reference Naissides, Mamo and James^78, Reference Blanco-Colio, Valderrama and Alvarez-Sala⁴⁶⁾. Increased postprandial TAG was evident after consuming red wine with alcohol compared with dealcoholised red wine or water⁽Reference Naissides, Mamo and James⁷⁸⁾ or after consuming a low-alcohol red wine or no wine⁽Reference Blanco-Colio, Valderrama and Alvarez-Sala⁴⁶⁾, respectively, despite differences in trial population (postmenopausal dyslipidaemic women v. healthy young men and women). Elevation of VLDL indicates alcohol-induced stimulated hepatic lipogenesis with red wine⁽Reference Blanco-Colio, Valderrama and Alvarez-Sala⁴⁶⁾; however, when alcohol (vodka) was controlled and the ratio of sugar to red wine alcohol varied (increased) at an equivalent energy level, VLDL was not increased compared with no wine with the fatty meal. These data suggest a threshold effect of alcohol plus sugar in stimulating hepatic lipogenesis or modifying clearance of VLDL. In either case, excess or prolonged circulation of TAG-rich lipoproteins, like VLDL, can be unfavourable, as discussed earlier. Interestingly, despite elevated TAG and VLDL after red wine consumption with a fat-enriched breakfast, Blanco-Colio et al. ⁽Reference Blanco-Colio, Valderrama and Alvarez-Sala⁴⁶⁾ reported reduced NF-κB activation in cultured mononuclear cells in a time-dependent manner that was not evident in the no-wine or vodka conditions. Hence, components of red wine, presumably phenolic compounds, provided a relative protection from the potentially deleterious effects of the associated postprandial hyperlipidaemia.

Table 2 Clinical trials examining postprandial TAG and LDL oxidation responses after fruit-associated polyphenolic treatments

LDLox, LDL oxidation; AC, antioxidant capacity; M, male; F, female; RW, red wine; EtOH, ethyl alcohol; TRAP, total radical-trapping antioxidant parameter; post-M, postmenopausal; dyslip, dyslipidaemia; OW, overweight or obese; de-alco, de-alcoholised; CM, chylomicrons; Str, strawberry; Pbo, placebo; FD, freeze-dried; WBB, wild blueberry; ORAC, oxygen radical absorbance capacity; TAS, total antioxidant status; TEAC, trolox equivalent antioxidant capacity; GSE, grape seed extract; AP, apple polyphenol extract; CB, cranberry; vit C, vitamin C; BB, blueberry; J, juice; RW, red wine, WW, white wine; FRAP, ferric-reducing antioxidant power; TBAR, thiobarbituric acid reactive substances; ↓ , decrease or reduction (P < 0·05); ↑ , increase (P < 0·05); –, neutral, no effect (P>0·05) compared with placebo or baseline (i.e. comparator trials).

* Describes general health, age or weight status. Healthy indicates normolipidaemic, between >19 years, not overweight or obese and no documented disease or condition.

† Where meal is indicated, fruit or fruit-based beverage was consumed 5 d with the meal.

‡ Meal is likely the same as that provided in reference Naissides et al. ⁽Reference Naissides, Mamo and James⁷⁸⁾.

§ Time point that a second meal was introduced. The results represent the glucose and insulin responses to this point.

Excessive metabolic substrate and dysregulated lipid (and carbohydrate) metabolism exposes cells to an overload of nutrients, generating a relative imbalance of ROS/reactive nitrogen species to antioxidant defences triggering a number of endogenous cellular responses that support and perpetuate the imbalanced state. The antioxidant properties of phenolic compounds in fruits and their contribution to the total antioxidant defence system are likely the underlying mechanism by which fruit consumption impart their greatest benefit. In the case of lipid metabolism, the protection of LDL from oxidation is a critical matter. Elevated plasma concentrations of oxidised LDL are apparent in patients with CHD and diabetes and predict future risk of CHD events in apparently healthy men⁽Reference Imazu, Ono and Tadehara^100–Reference Meisinger, Baumert and Khuseyinova¹⁰³⁾. Several in vitro and ex vivo investigations (after feeding trials of days to weeks with fruit, fruit juice and wine) have demonstrated a relationship between increasing plasma antioxidant status and decreasing the susceptibility of LDL to oxidation⁽Reference Miyagi, Miwa and Inoue^104–Reference Aviram, Dornfeld and Resenblat¹⁰⁹⁾. In contrast, few investigations have examined acute protection of LDL, such as in a postprandial challenge paradigm. Postprandial LDL are more susceptible to oxidation than fasting LDL⁽Reference Lechleitner, Hoppichler and Föger¹¹⁰⁾ and postprandial LDL have been shown to accumulate more readily in activated macrophages than LDL derived from fasting samples⁽Reference Steinberg and Lewis^71, Reference Lechleitner, Hoppichler and Föger¹¹⁰⁾. The mechanism by which LDL become oxidised are not completely clear. Absorption of dietary-derived lipid hydroperoxides that incorporate into lipoproteins and act as initiators for further lipoprotein oxidative modification is one possibility⁽Reference Ursini, Zamburlini and Cazzolato⁸⁾. The typical intake of lipid hydroperoxides in a fat-rich western diet has been estimated as 1·5 mmol/d⁽Reference Wolff and Nourooz-Zadeh¹¹¹⁾. Consumption of polyphenols with a high-fat meal has been shown to decrease plasma lipid hydroperoxides. Natella et al. ⁽Reference Natella, Belelli and Gentili⁹⁷⁾ showed in eight healthy men who consumed a standard ‘Milanese’ style meat and potatoes meal with (active) and without (placebo) 300 mg GSE that plasma lipid hydroperoxides were 1·5-fold higher in the placebo meal 1 h after the meal compared with the GSE active meal. Correspondingly, plasma antioxidant capacity was higher in the active GSE meal and resistance to LDL oxidation was enhanced, although statistical significance was not achieved between placebo and GSE active meals, a factor possibly due to timing of assessment and small sample size.

Circulating LDL is subject to oxidation in a pro-oxidative environment (postprandial state) when defences are inadequate. Acute protection may require direct association of antioxidant compounds with lipoproteins and/or activation of or preservation of the HDL-associated paraoxonase 1⁽Reference Aviram, Kaplan and Rosenblat¹¹²⁾. α-Tocopherol is a well-known lipid-soluble antioxidant present in LDL. Previous work has shown that dietary phenolic compounds can bind human LDL⁽Reference Lamuela-Raventos, Covas and Frito¹¹³⁾. Phenolic compounds that bind LDL are likely to provide a relative protection of LDL from ROS through their peroxyl-scavenging activity or by structural modification resulting in in vivo resistance to oxidation⁽Reference Witzum^114, Reference Tikkanen, Wahala and Ojala¹¹⁵⁾. Bound polyphenols may also spare losses of endogenous antioxidants, like α-tocopherol⁽Reference Natella, Ghiselli and Gudi⁷⁷⁾.

Present therapeutic strategies in CVD target improving serum cholesterol, generally LDL cholesterol, as assessed by fasting total and LDL concentrations. Because oxidised LDL is an early initiator and propagator of atherosclerosis, there may be merit in strategies that target protection of LDL from oxidation during the postabsorptive period. Trials with vitamin supplements such as vitamin E, vitamin C and β-carotene have been disappointing, possibly because they can act as pro-oxidants as well as antioxidants⁽Reference Aviram, Kaplan and Rosenblat¹¹²⁾. Long-term supplementation with polyphenolic compounds or polyphenolic-rich fruits on the susceptibility of fasting LDL has been disparate. Timing of intake (with or without a meal) may be one explanation for whether a favourable outcome will be achieved. Table 2 represents the few studies identified that measured oxidised LDL in plasma/serum after a meal that was accompanied by a phenolic-rich treatment or placebo. In relatively healthy individuals consuming a moderate-/high-fat meal, strawberry⁽Reference Burton-Freeman, Linares and Hyson¹⁰⁾ and red wine⁽Reference Natella, Ghiselli and Gudi⁷⁷⁾ blocked the postprandial rise in oxidised LDL or susceptibility of LDL to oxidation, respectively. GSE (300 mg capsule) with a high-fat meal in healthy men enhanced LDL protection, but this was not statistically different from the meal with placebo capsule⁽Reference Natella, Belelli and Gentili⁹⁷⁾. After a very low-fat meal (two plain bagels)⁽Reference Caccetta, Croft and Beilin⁶⁴⁾ or no meal challenge⁽Reference Ziegler, Kostner and Thallinger¹¹⁶⁾, changes in LDL oxidation were neutral, suggesting that the benefit of polyphenol consumption with a meal is to attenuate or block meal-induced oxidative stress when the stress is sufficient to result in oxidised LDL. In type 2 diabetic subjects⁽Reference Ceriello, Bortolotti and Motz⁵⁴⁾, red wine attenuated meal-induced decrease in plasma antioxidant capacity (as measured by total radical-trapping antioxidant parameter) and partially protected LDL from the effect of the meal. Two investigations using a berry mix (240 mg; 80 mg each bilberries, lingonberries, black currant)⁽Reference Berliner, Navab and Fogelman¹⁰¹⁾or cactus fruit (500 g)⁽Reference Tesoriere, Allegra and Butera¹¹⁷⁾ compared postprandial LDL oxidation susceptibility with baseline (fasting) susceptibility measurements (no placebo or other comparator) after consuming fruit alone. Consumption of berries resulted in improved antioxidant capacity with neutral effect on LDL oxidative susceptibility. Cactus fruit, rich in peroxyl-scavenging betalains resulted in decreased susceptibility of LDL to oxidation at 3 and 5 h postprandially. The effect appeared dependent on betalain accumulation in LDL and sparing effects of lipid-soluble antioxidants. As with much of the postprandial effects of fruits and their inherent phytochemicals, additional research is required to understand betalains and indicaxanthins in health protection.

The acute effect of increasing antioxidant capacity of plasma is fairly consistent; however, changes in a biological marker of disease risk are not always indicated. This may be due to a variety of reasons ranging from composition and dose requirements to elicit an effect to choice of marker being measured. In addition, polyphenolic compounds likely act beyond a classic ‘antioxidant’ role to impart benefit, and hence changes in antioxidant capacity may not necessarily predict a biological response. Likewise, the interrelationship between acute and long-term effects of polyphenolic compounds on disease risk, as studied in various acute and chronic feeding paradigms, requires continued attention to better understand the translation and predictability of our models to describe biological impact and mechanism of action in a preventative or potentially therapeutic role in certain populations.