Bioavailability of epicatechin and effects on nitric oxide metabolites of an apple flavanol-rich extract supplemented beverage compared to a whole apple puree: a randomized, placebo-controlled, crossover trial
Wendy J. Hollands,
Institute of Food Research, Norwich Research Park, Norwich, UK
Flavanol-rich foods are known to exert beneficial effects on cardiovascular health. The biological effects depend on bioavailability of flavanols which may be influenced by food matrix and dose ingested. We compared the bioavailability and dose-response of epicatechin from whole apple and an epicatechin-rich extract, and the effects on plasma and urinary nitric oxide (NO) metabolites.
Methods and results
In a randomized, placebo-controlled, crossover trial, subjects consumed drinks containing 70 and 140 mg epicatechin from an apple extract and an apple puree containing 70 mg epicatechin. Blood and urine samples were collected for 24 h post ingestion. Maximum plasma concentration, AUC(0–24 h), absorption and urinary excretion were all significantly higher after ingestion of both epicatechin drinks compared with apple puree (p < 0.05). Time to maximum plasma concentration was significantly later for the puree compared with the drinks (p < 0.01). Epicatechin bioavailability was >2-fold higher after ingestion of the 140 mg epicatechin drink compared to the 70 mg epicatechin drink (p < 0.05). Excretion of NO metabolites was higher for all test products compared with placebo, which was significant for the high dose drink (p = 0.016).
Oral bioavailability of apple epicatechin increases at higher doses, is reduced by whole apple matrix and has the potential to increase NO bioavailability.
Flavanols are the major flavonoids consumed by humans with their contribution to mean total flavonoid intakes estimated at >80% . Flavanols range from simple monomers such as (+)-catechin and (–)-epicatechin through to more complex oligo-/polymeric flavanols collectively known as proanthocyanidins. Meta-analyses of data obtained from human intervention trials have demonstrated that flavanol-rich foods such as chocolate and cocoa exert beneficial effects on cardiovascular health through their potential to decrease blood pressure and improve endothelial function (as measured by an increase in flow-mediated dilatation (FMD)) [2, 3]. FMD is almost exclusively nitric oxide (NO) dependant and epicatechin-rich cocoa and pure epicatechin have been shown to enhance the bioavailability and bioactivity of NO [4-7]. Furthermore, the maximum effect on endothelial function coincides with peak plasma levels of epicatechin metabolites . It would appear therefore, that epicatechin is an important bioactive component of flavanol-rich foods.
The physiological effects of flavanols, however, are dependent upon their bioavailability (i.e. the proportion of oral dose reaching the site of action). Monomeric flavanols are predominantly absorbed from the small intestine into the gut epithelial cells where they are efficiently metabolised by phase-2 conjugating enzymes to form O-glucuronidated and O-sulfated derivatives of epicatechin and methylepicatechins . After transport to the liver, further methylation and glucuronidation as well as sulfation may occur. Peak plasma concentrations of monomeric flavanols in humans after ingestion of chocolate and cocoa are reported across a wide range (∼0.1 to 12.7 μM), with estimates of total urinary excretion accounting for approximately 10 to 25% of intake [8-19] depending on the dose and the food source/matrix.
There are a number of factors that may affect how well an individual absorbs and metabolises flavanols. The dose consumed and the presence of other polyphenols and macro and micronutrients within the food matrix are important considerations. Concurrent carbohydrate consumption for example, has been shown to significantly increase the uptake of flavanols , whilst the general consensus for milk protein appears to be that it has little or no effect on the bioavailability of flavanols [13, 14, 20, 21]. Very little data exists on the effects of dietary fibre on flavanol absorption but consumption of wheat fibre does not seem to affect the absorption of other flavonoids such as isoflavones .
Apples are a rich dietary source of flavanols (mainly epicatechin and epicatechin-rich procyanidins), which have been reported to occur at concentrations between 0.1 and 45 mg per 100 g fresh fruit . Apples also contain other polyphenols such as various glycosides of quercetin and phloretin, chlorogenic acids (caffeoylquinic acids), as well as a mixture of macronutrients such as protein, lipids and carbohydrates and also fibre (largely pectin). In addition to the phenolics in apples, other components of apples have also been reported to provide health benefits. For example, apple pectin has been reported to modify intestinal structure, prolong intestinal transit time, and affect the microbiota (prebiotic effects). The flavanols of apples are very similar to those present in cocoa and dark chocolate, and apples represent an important source of flavanols in human diets, and potential sources of extractable flavanols for inclusion in functional foods.
The purpose of this study was to investigate the acute effects of a whole apple matrix and epicatechin dose of an apple flavanol-rich extract on epicatechin bioavailability, and also on plasma and urinary NO metabolites as a biomarker of endothelial function.
2 Materials and methods
2.1 Chemicals and reagents
Ammonium acetate, catechin, epicatechin, DMF, acetic acid, formic acid, trichloroacetic acid, hydrochloric acid (HCL), sodium phosphate dibasic and the enzymes β-glucuronidase and sulfatase (Helix pomatia Types H-5 and H-1, respectively) were purchased from Sigma-Aldrich (Poole, UK). Taxifolin was obtained from Extrasynthese (Genay, France). Sodium dihydrogen orthophosphate (BDH, Analar) was purchased from VWR international (Leicester, UK) and HPLC-grade methanol from Fisher Scientific (Loughborough, UK). Standards 3′-methylepicatechin, 4′-methyl epicatechin and epicatechin-3-sulfate for epicatechin metabolite identification were synthesised in-house and have been described by the authors elsewhere . Freeze-dried apple granules and apple flavanol-rich extract (EvesseTM EPC); puree and beverage interventions respectively, were provided by Coressence Ltd.
2.2 Subjects and study design
Fourteen men and women aged between 45 and 70 years were recruited to participate in this study. All study participants were assessed for eligibility on the basis of a health questionnaire and the results of clinical laboratory tests. The following exclusion criteria applied: smokers; medical conditions such as asthma, heart disease, gastro-intestinal disease, diabetes, cancer; regular prescribed medication; dietary supplements judged to affect the study outcome; alcohol consumption > approximately 20 g alcohol/day; BMI < 19.5 or >35; clinical results at screening judged to affect study outcome or be indicative of a health problem. The study was conducted at the Institute of Food Research and all participants gave written informed consent prior to participation. The study protocol was approved by the Human Research Governance Committee of the Institute of Food Research (IFR) and the Hertfordshire Research Ethics Committee (Ref: 09/H0311/86). The trial is registered at http://clinicaltrials.gov (Identifier: NCT01097226).
The study was a double blind (except for the apple puree treatment), placebo-controlled, four-phase crossover trial, investigating the bioavailability of epicatechin derived from apple. Participants were randomized to receive a bolus dose of each of the following: (i) 300 g apple puree providing 70 mg epicatechin; (ii) 300 g flavoured water containing an apple flavanol-rich extract providing 70 mg epicatechin; (iii) 300 g flavoured water containing an apple flavanol-rich extract providing 140 mg epicatechin; (iv) 300 g flavoured water as a placebo control.
The beverages were made up immediately prior to consumption by adding the appropriate mass of apple flavanol-rich extract powder required to deliver a 70 or 140 mg dose of epicatechin to a commercially available orange flavoured powder (Kool-Aid, Kraft Foods, USA), before the addition of water. The apple puree was made up as a single batch at the start of the study by re-hydrating freeze-dried whole apple granules with boiling water. The resultant apple puree was sub-divided into smaller aliquots and stored at −20°C until consumption. Immediately prior to consumption, an aliquot of apple puree was defrosted and accurately weighed to provide the required dose of epicatechin.
2.3 Study procedure
Each treatment arm comprised a 2-day period of intervention separated by a washout period of ≥7 days. All four arms were identical in nature, except for the test product consumed. During each period of intervention, participants were instructed to avoid consuming flavanol-rich foods such as chocolate and chocolate products, berry fruits (and derivatives such as jams and juices) and some beverages including alcohol, tea, coffee and red wine. To aid compliance, a list of prohibited foods was given. On day 2 of the intervention period, participants attended the nutrition unit at IFR following a 10-h overnight fast. An intravenous catheter was inserted and a baseline (0 h) blood sample obtained. A baseline (0 h) spot urine sample was also obtained. Participants consumed one of the four test products after which whole blood samples were collected at the following time points: 30 and 60 min and 1.5, 2, 2.5, 3, 4, 6, 8 and 24 h. Urine was collected between 0–2, 2–4, 4–6, 6–8 and 8–24 h post ingestion of the test product. For standardisation purposes, all participants were asked to consume the test product within a 5-min period and refrain from eating and drinking for the next 2 h. A standard breakfast and low flavanol lunch were served 2 and 5 h after test product consumption, respectively. Participants were free to drink water ad libitum after the 2-h time point.
After collection, blood samples were immediately centrifuged (2000 × g; 10 min) and plasma aliquoted into tubes containing a storage solution (0.4 M NaH2PO4 buffer containing 20% ascorbic acid and 0.1% EDTA; pH 3.6). Urine collections were weighed and sub-samples acidified to pH ∼4.5 with 1 M HCL. Plasma and urine samples were immediately frozen and subsequently stored at −80 and −40°C, respectively. Plasma and urine used for NO analysis were frozen and stored without the addition of storage solution or HCL.
2.4 Extraction and analysis of the apple test products
Samples of apple flavanol-rich powder (200 mg) or apple puree (2 g) were weighed into 50 and 20 mL volumetric flasks respectively and partially filled with 70% methanol. Flasks were placed in an ultrasonic bath (10 min) to disperse the samples before incubating in an oven for 10 min at 60°C. Post incubation, samples were placed in an ultrasonic bath for a further 10 min, before cooling and filling flasks to volume with 70% methanol. All samples were centrifuged prior to analysis. Samples of a baking chocolate reference material (NIST SRM 2384) that had previously been defatted using repeated hexane extraction were extracted and analysed alongside test samples.
For determination of flavanol content, samples were analysed by normal-phase HPLC (Agilent HP1100) using a Luna Silica (2) column (250 × 4.6 mm) with a 5 μM particle size. The mobile phase (A, dichloromethane; B, methanol; C, 50% v/v acetic acid) was pumped through the column at a flow rate of 1.0 mL/min. Post column, the eluent passed through a fluorescence detector using wavelengths of 276 nm excitation and 316 nm emission. Epicatechin and catechin content were calculated relative to a response factor derived from an authentic epicatechin standard curve over the range of 0–200 μg/mL. The polymeric fraction (procyanidins) was calculated against the response factor obtained for epicatechin and additional relative fluorescence response data as described by Prior and Gu . The estimate of the concentration of epicatechin in the baking chocolate reference material analysed alongside samples were within the reference range, confirming the validity of the analyses. To determine the content of chlorogenic acid, quercetin glycosides and phloridzin, samples were analysed by reverse-phase HPLC using a Luna (2) column (250 × 4.6 mm) with a 5 μM particle size. The mobile phase (A, 0.1% TFA in water and B, 0.1% TFA in ACN) was pumped through the column at a flow rate of 1.0 mL/min. Post column, the eluent passed through UV diode array detector monitoring over the range 200–700 nm. Chlorogenic acid content of the apple products was calculated relative to a response factor derived from an authentic chlorogenic acid standard curve. The content of flavonoid glycosides (quercetin glycosides and phloridzin) were calculated relative to a response factor obtained for an authentic quercetin-3-rhamnoside standard curve.
The phenolic composition of the low-dose flavanol-rich apple extract was: (-)-epicatechin 70 mg; (+)-catechin 1.8 mg; procyanidins (flavanol-3-ol oligo-and polymers) 62.8 mg; total quercetin glycosides 7.1 mg; phloridzin 3.2 mg and chlorogenic acid 7.4 mg. The phenolic composition of the apple puree was: (-)-epicatechin 70 mg; (+)-catechin 6.1 mg; procyanidins 200 mg; total quercetin glycosides 5.7 mg; phloridzin 28.5 mg and chlorogenic acid 102.6 mg.
Aliquots of apple puree (re-hydrated freeze-dried apple) analysed at the start (freshly made puree) and toward the end of the intervention period (frozen puree) showed no significant differences in epicatechin content, indicating that the polyphenols were not lost during puree preparation and were stable in the frozen puree throughout the intervention period. The epicatechin content of the beverages (made up from the flavanol-rich powder on the day of consumption) was also measured and shown to be stable over a 30-min period followed by mixing with water and orange flavouring, which demonstrates the epicatechin was not lost during drink preparation and was stable until consumed by the volunteer (within 10–15 min of preparation).
2.5 Extraction and analysis of epicatechin in plasma and urine
Urine samples (200 μL) were mixed with phosphate buffer (200 μL; pH 6.8), sulfatase (80 μL; 80 U), β-glucuronidase (80 μL; 800 U) and taxifolin (10 μL; 10 μg/mL) prior to incubation for 2 h at 37°C. Post incubation, DMF (570 μL) and formic acid (40 μL) were added. Similarly, plasma samples (200 μL) were mixed with phosphate buffer (200 μL; pH 5.0), sulfatase (20 μL; 20 U), β-glucuronidase (20 μL; 200 U) and taxifolin (10 μL; 1 μg/mL) prior to incubation for 2 h at 37°C. Post incubation, DMF (100 μL) and 50% aqueous trichloroacetic acid (20 μL) were added. All plasma and urine samples were then centrifuged (13 000 RPM; −3°C) for 15 min prior to LC-MS analysis as described by the authors elsewhere 
Quantification of epicatechin was achieved using a response factor obtained from an epicatechin standard curve that was generated from a series of samples of blank plasma or urine that had been spiked with known quantities of epicatechin and then processed in exactly the same manner as the unknown human plasma and urine samples and run alongside those samples. Standard curves were linear up to 3.5 and 69 μM in plasma and urine, respectively. Regression coefficient values were >0.99. The LODs for epicatechin were 70 and 35 nM for urine and plasma respectively, and the intra-day and inter-day variance was less than 10 and 20%, for urine and plasma, respectively. Reference urine and plasma samples with a known epicatechin concentration were extracted and analysed alongside every study sample batch and epicatechin concentration remained within 20% of the mean.
2.6 Quantification of NO metabolites in plasma and urine
Nitrate and nitrite determinations were performed using a Sievers 280i NO analyser comprising of a purge vessel attachment, hot water bath to control the temperature of the reaction vessel and a chill bath to control the condenser. The 280i NOA reduces nitrite, nitrate and nitrosothiols to NO in the purge vessel which is then quantified according to the chemiluminescence signal released transiently within the instrument. Plasma and urine samples were diluted with water (1:1 and 1:25; plasma and urine, respectively) before analysis. Nitrite was determined after adding a sample of plasma or urine to a mild reducing solution of sodium iodide in acetic acid at room temperature. Nitrates were determined by adding a sample of plasma or urine to vanadium (III) chloride in HCL at 90°C. Samples (50 μL) were analysed in triplicate. The LOD determined for a 50 μL sample was 80 μmols/dm−3. Intra-day and inter-day variability were 2.5 and 12.4%, respectively.
2.7 Statistical analysis
Statistical analyses were performed using the R data analysis software (R Foundation for Statistical Computing, Vienna, Austria; http://www.r-project.org) Pharmacokinetic parameters, maximum plasma concentration (Cmax), time to maximum plasma concentration (Tmax), plasma elimination half-life (T1/2) and area under the plasma concentration time curve (AUC 0–24h) were calculated from the measured plasma data. For AUC, the trapezoidal method was applied. Epicatechin absorption was estimated by fitting a single compartment model to the plasma data using total body water as the volume of distribution. Paired t-tests were used to examine differences between variables and when the data was not normally distributed, a non-parametric equivalent was used. Differences were considered significant when p < 0.05.
Two of the fourteen participants randomized to treatment did not complete: one completed only one of the four treatment arms (placebo drink) and cited work as the reason for withdrawal, the other completed two of the four treatment arms (low- and high-dose epicatechin drink) but was withdrawn by the researcher after difficulties in siting an intravenous catheter on the third phase of the trial.
Subject characteristics for the 13 participants included in the analysis are shown in Table 1.
Table 1. Subject characteristics at start of trial
Male (n = 6)
Female (n = 7)
Data are mean ± SD.
58.2 ± 5.9
55.1 ± 6.2
56.7 ± 6.4
173.6 ± 3.3
166.2 ± 4.8
170.0 ± 5.3
79.5 ± 10.6
69.5 ± 7.4
74.5 ± 9.2
26.3 ± 3.2
25.1 ± 2.5
25.7 ± 2.7
3.2 Identification of epicatechin and epicatechin metabolites in plasma and urine
In a recent paper, it was reported that epicatechin sulfates and methylepicatechin sulfates were very poor substrates for commercially available aryl sulfatase enzymes, and that peaks of sulfated epicatechin and methylepicatechin remained after the enzyme hydrolysis step . Here, six peaks generating ion currents in the multiple reaction monitoring MS analyses were identified in hydrolysed post-intervention plasma and urine samples. Two of the peaks produced m/z ratios of 289/109 and corresponded with the retention times of authentic catechin and epicatechin reference standards. Two further peaks were shown to correspond with 3′-methylepicatechin and 4′-methylepicatechin, based on having the same retention times as the authentic standards and generating the expected mass transition (m/z = 303/137). Substantial ion current peaks were observed for the transitions m/z = 369/289 and 383/303 and were identified as epicatechin sulfates and methylepicatechin sulfates, respectively. Epicatechin glucuronides were not detected confirming that they had been effectively hydrolysed. Estimates of total epicatechin were derived from the combined sum of epicatechin, 3′methylepicatechin, 4′methylepicatechin, epicatechin sulfates and methyl-epicatechin sulfates quantified against an authentic epicatechin standard.
For the majority of subjects, no epicatechin was detectable in the baseline urine or plasma samples, indicating excellent compliance with the low-flavanol diet. Epicatechin appeared in plasma within 30 min of ingestion of all three treatments with only small amounts still detectable 24 h later.
3.3 Effects of matrix on epicatechin bioavailability
The pharmacokinetic data for plasma epicatechin are presented in Table 2 and Fig. 1. The mean Cmax for plasma epicatechin after ingestion of apple puree containing 70 mg epicatechin was 2.1 ± 0.8 μmol/L which was significantly lower than the equivalent dose fed as a flavoured beverage (3.5 ± 1.4 μmol/L; p = 0.00049 for difference). Similarly AUC0→24h was significantly reduced after apple puree ingestion compared with the same dose beverage (9.5 ± 4.3 μmol.h/L compared with 15.8 ± 5.5 μmol.h/L; puree and drink respectively; p < 0.0001). There was no significant difference in the half-life of epicatechin elimination in plasma between the two matrices. However, there was a significant difference in the time taken to reach maximum plasma concentration between the two matrices with the Cmax being reached significantly later for the apple puree compared to the equivalent dose fed as a flavoured drink (1.7 ± 0.7 h compared with 0.9 ± 0.3 h, puree and drink respectively; p = 0.0056).
Table 2. Pharmacokinetic parameters of epicatechin after ingestion of a drink containing an apple flavanol-rich extract and whole apple puree
Puree (241 μmol)
Drink (241 μmol)
Drink (483 μmol)
Data are presented as mean ± SD.
Cmax = maximum peak plasma concentration; Tmax = time to reach peak
plasma concentration; AUC = area under the plasma concentration time curve; T1/2 = elimination half-life; absorption predicted using one compartment model fitted to plasma data.
a)–f)Values with same superscript are significantly different from each other, p < 0.05 (paired t-test).
Further evidence that bioavailability is affected by matrix was observed after fitting the plasma data to a one-compartment model, using total body water as the volume of distribution, to predict epicatechin absorption. After ingestion of apple puree, the predicted epicatechin absorption was 44.3 ± 15.4% which was significantly lower than the same-dose beverage (59.2 ± 21.8%; p = 0.00488).
The rate of urinary excretion (μmol epicatechin/h) was greatest across the 0–2 h time point for both the low-dose drink and apple puree treatments (Fig. 2). The absolute quantity of epicatechin excretion over the entire 24-h collection period after ingestion of apple puree was 69.0 ± 19.1 μmol and accounted for ∼ 29% of intake. These were significantly lower than that excreted after ingestion of the low-dose drink (88.5 ± 21.9 μmol) accounting for ∼ 37% of intake (p = 0.0016 for difference) (Table 2).
3.4 Effects of dose on epicatechin bioavailability
There was a significant effect of dose on AUC0→24h (44.87 ± 18.9 μmol.h/L compared with 15.8 ± 5.5 μmol.h/L; high- and low-dose drink respectively; p < 0.0001) and plasma Cmax (8.9 ± 3.7 μmol/L compared with 3.5 ± 1.4 μmol/L; high- and low-dose drink respectively; p < 0.0001). However, there were no significant differences between the two beverage doses on the time taken to reach C-max or the half-life of epicatechin elimination from plasma (see table 2).
It is apparent that from the parameters estimating bioavailability from the plasma data (Cmax and AUC), that double the dose of epicatechin in the form of a flavoured drink results in more than a twofold increase in bioavailability. After fitting the plasma data to a one-compartment model, predicted epicatechin absorption was significantly higher after ingestion of the high dose drink compared with the low-dose drink (82.5 ± 31.0 and 59.2 ± 21.8, respectively; p = 0.00028).
The absolute quantity of epicatechin excretion over the entire 24-h collection period after ingestion of the high-dose drink was 201.3 ± 41.7 μmol which was significantly higher than subjects fed the low-dose drink (p < 0.0001 for difference). The differences in urinary yield (% of dose) between the drink doses reached borderline significance (p = 0.0733) (see Table 2).
3.5 Effects of epicatechin on NO metabolites in plasma and urine
Plasma nitrate concentrations increased 30 min after ingestion of all three treatments and returned to baseline levels within 1–1.5 h after which it remained fairly constant throughout the timed collection period (data not shown). However, there were no significant differences in plasma concentration of nitrates between the placebo and treatments 30 min after ingestion when differences were most apparent. Plasma nitrite remained unchanged regardless of treatment (data not shown).
Urinary excretion of nitrates increased to a peak at 2 h after ingestion of all three treatments. We found no significant differences between any of the treatments and placebo at this time point. However, at the 24-h time point, urinary nitrate excretion was significantly higher after ingestion of the high-dose drink compared with placebo (1.46 versus 1.15 μmols nitrate/mg creatinine respectively, p = 0.0161), but was not significantly changed for the low-dose drink or puree (see Fig. 3).
In this four-phase, randomized, crossover, single-dose pharmacokinetics, trial we have shown that (i) the bioavailability of apple epicatechin was negatively affected by the whole apple matrix, (ii) the dose normalized bioavailability of epicatechin from a flavanol-rich apple extract increased at higher doses and (iii) urinary excretion of NO increased after consumption of all epicatechin-containing test products, and was significant for the high-dose epicatechin drink.
Whilst there is an abundance of literature data concerned with the oral bioavailability of flavonoids from different sources, and a number of reports describing the effects of food macro-components such as protein, fat and fibre on bioavailability, very little is known of how an apple matrix affects the absorption of apple flavanols. There are a number of possible reasons why the bioavailability of apple epicatechin is reduced in the presence of the whole apple matrix, including: (i) Reduced bioaccessibility due to binding of epicatechin to apple fibre or entrapment of epicatechin within intact cells; (ii) the increased viscosity of the bolus caused by the apple pectin (iii) reduced rates of epicatechin uptake by enterocytes due to competition for enterocyte phase-2 metabolising enzymes from other polyphenols (e.g. quercetin, phloretin, chlorogenic acid) that serves to increase the intracellular epicatechin concentration and reduce the lumen-cellular epicatechin concentration gradient that drives passive diffusion; (iv) decreased rates of basolateral (serosal) efflux of epicatechin conjugates (=reduced absorption) due to competition from other polyphenols for enterocyte efflux transporters; (v) increased rates of apical (luminal) efflux of epicatechin conjugates (=reduced absorption) due to activation of apical membrane efflux transporters.
In addition to being a rich source of dietary polyphenols, apples contain a mixture of other nutritional components including protein, carbohydrates and fibre (largely pectin). These non-phenolic components may contribute to the health benefits of consuming apples (e.g. by improving gut transit times or changing the composition of the gut flora), but they may also influence the bioavailability of the phenolics. It was previously reported that concurrent ingestion of sugars or a carbohydrate-rich food (bread) enhanced cocoa flavanol absorption ; the authors proposed that the observed effects on flavanol absorption might be mediated by a carbohydrate-specific effect on gut physiology or a yet unidentified sugar-flavanol transporter. In contrast, no effect of lipid or protein-rich meals was observed on the absorption of cocoa flavanols . Pectin may modify intestinal structure  and delay gastric transit times [26, 27]. For subjects in our study, the mean time taken to reach maximum plasma concentration of epicatechin was significantly longer after ingestion of the whole apple puree compared with the equivalent dose of epicatechin fed as an apple-extract supplemented beverage (lacking apple macro-components). Given the propensity for pectin to interfere with absorption, and taking into account that the pectin content in the apple puree fed to subjects in our study was relatively high (approx. 17g/100g), it would be reasonable to assume that the pectin component of the whole apple product was, at least to some extent, responsible for the observed increased Tmax and decrease in epicatechin bioavailability. Few studies to date have directly investigated the effects of pectin on flavonoid bioavailability but Nishijima et al. have reported that chronic consumption of apple pectin enhances the intestinal absorption of the flavonol quercetin in rats . However, since there was no effect on the Tmax of plasma quercetin, the authors concluded that this effect was likely to be the result of biological changes induced by the pectin (i.e. improved nutrient transport efficiency, and/or changes in the absorptive capacity of the small intestine) rather than physico-chemical interactions between quercetin and the pectin.
The whole apple puree used in the study described here contained significant quantities of quercetin and phloretin glycosides, and chlorogenic acid whereas the flavanol-rich extract contained substantially lower quantities of phloretin glycosides and chlorogenic acid, but similar quantities of quercetin glycosides (see section 'Extraction and analysis of the apple test products'). Once the quercetin gluco- and galacto-sides and the phloretin glucosides have been hydrolysed by luminal lactase phloridzin hydrolase  and/or cytosylic β-glucosidase , and the chlorogenic acid (caffeoyl-quinate esters) by intestinal esterases , they would diffuse into the small intestinal enterocytes and become conjugated with sulfate and/or glucuronic acid. It is therefore likely that quercetin, phloretin and caffeic acid present in the whole apple competed with epicatechin for enterocyte phase-2 conjugating enzymes and efflux transporters. The expected effect would be to reduce the rate of absorption of epicatechin in the puree compared to the extract supplemented beverage, which is consistent with what we observed in this study (∼40% reduction in AUC0–24h after ingestion of apple puree compared with the same dose beverage). Previously, Silberberg et al.  reported that after supplementing rats an equal dose of quercetin and catechin in combination or alone, a significant decrease in plasma concentrations of both quercetin and catechin was observed when they were ingested together, compared with plasma concentrations obtained when the flavonoids were ingested separately.
In our study, we observed ∼ 2.5-fold increase in plasma C-max, AUC, and absolute urinary excretion of epicatechin after ingestion of the beverage providing a 140-mg dose of epicatechin compared with the 70-mg dose, demonstrating that there is not a simple linear relationship between the oral dose and the appearance in plasma or excretion in urine. Indeed, our data indicate that the absorption and excretion of epicatechin increases at higher doses on a dose-normalised basis. There are few published reports of studies comparing the bioavailability for different doses of flavanols in humans. But, in a previously reported study in which subjects consumed 1.5, 3.0 and 4.5 g of decaffeinated green tea solids on three different occasions, 2.7 to 3.4-fold increases in plasma Cmax values were observed between the 1.5 and 3.0 g doses , which is in keeping with our observation. However, no further increase in plasma Cmax was observed for the 4.5 g dose indicating that there may be a saturation effect at higher doses. It is possible to rationalise the observed increased bioavailability of epicatechin at higher doses. Higher doses of epicatechin may cause the cells to become overloaded with hydrophilic conjugates. Their efflux from the cells is an active process and although unequivocal evidence of the role of individual transporters in flavonoid conjugate efflux is lacking, it does exist for some efflux transporters (e.g. breast cancer-resistant protein) . The expression of these efflux transporters is polarised – that is, transporters are expressed on either the apical (luminal) or basolateral (serosal) membrane, but not both. Because the efflux process depends on active transport via membrane transporters, it is possible that when the epicatechin dose is increased, apical efflux of epicatechin conjugates becomes saturated to some extent. A consequence of saturation of apical efflux would be an increase in intracellular concentrations of epicatechin conjugates, which would in turn drive an increase in the rate of basolateral efflux. This would manifest as a non-linear super-absorption response to increased doses of epicatechin, which is consistent with what we have reported here.
Understanding the absorption and metabolism of flavanols is an important consideration in determining the biological significance. Flavanol-rich foods and beverages, as well as pure epicatechin, have been shown to beneficially affect endothelial function [2-4]. FMD is a clinically valid biomarker for endothelial function and is almost exclusively NO dependent. Epicatechin-rich cocoa and pure epicatechin have been shown to enhance the bioavailability and bioactivity of NO [4-7]. Moreover, the maximum effect on endothelial function coincides with peak plasma levels of epicatechin metabolites . It would appear therefore, that epicatechin is an important bioactive component of flavanol-rich foods. Contrary to others [4, 34] we were not able to demonstrate a significant effect of a flavanol-rich food on plasma NO metabolite levels. However, we did observe and increase in 24-h urinary excretion of nitrates for all three test products compared with placebo, which was significant for the high-dose beverage. This study was not powered to assess the effects of the treatments on NO metabolites as a biomarker of endothelial function, which likely explains why we did not observe a statistically significant increase in plasma NO metabolites following the consumption of the flavanol-rich foods/beverages.
In conclusion, our data show that both whole apple and a flavanol-rich apple extract are good sources of bioavailable (–)-epicatechin. But, epicatechin consumed from whole apple is less bioavailable than when delivered in the form of an epicatechin-rich apple extract incorporated in a water-based beverage, supporting the notion that the apple matrix reduces epicatechin bioavailability. The data reported here contributes to knowledge of the factors affecting epicatechin bioavailability and will be of benefit in future in the design of foods and beverages that are sources of bioavailable flavonoids including epicatechin.
The authors thank the research nurses at the human nutrition unit for the excellent care of our dedicated volunteers. We also thank Shikha Saha for technical input with mass spectrometry. This research was funded by Danisco A/S.
Potential conflict of interest statement: O.H. and K.T. are employed by DuPont (formerly Danisco) and R.W. is a director and employee of Coressence Ltd. P.A.K. is an independent member of the Coressence Ltd. Scientific Advisory Board.