Abstract
Background: Small-scale, short-term intervention studies have suggested that plasma alkylresorcinol (AR) concentrations may be biomarkers of whole grain (WG) wheat and rye intakes.
Objective: The objective was to determine whether plasma AR concentrations reflect self-reported WG food intake in a 16-wk WG intervention study and to establish which phenotypic characteristics influence plasma AR concentrations.
Design: In a randomized parallel-group dietary intervention study, 316 overweight and obese participants with a WG intake of <30 g/d were recruited and randomly assigned to 1 of 3 groups: control (no dietary change), intervention 1 (60 g WG/d for 16 wk), or intervention 2 (60 g WG/d for 8 wk followed by 120 g WG/d for 8 wk). Fasting blood samples were collected at baseline, 8 wk, and 16 wk for the measurement of plasma lipids and ARs.
Results: Plasma samples from 266 study completers were analyzed. Total plasma AR concentrations increased with the WG intervention and could be used to distinguish between control subjects and those who consumed 60 or 120 g WG, but not between those who consumed 60 and 120 g WG. Plasma AR concentrations were higher in men, were positively associated with plasma triglyceride concentrations, and were negatively associated with nonesterified fatty acids.
Conclusions: Plasma AR concentrations were correlated with WG intake and could be used to distinguish between low- and high-WG consumers. Sex and plasma lipid concentrations independently influenced plasma AR concentrations, although plasma triglycerides may explain higher concentrations in men. This trial is registered as ISRCT no. 83078872.
INTRODUCTION
ARs4 are phenolic lipids found in high concentrations in the bran fraction of wheat and rye, in low concentrations in barley, and in only minute quantities in the endosperm (white flour) fraction of these cereals (1, 2). ARs are mainly consumed via products containing WG wheat or rye or bran from these cereals. Because they are absorbed and can be measured in plasma (3, 4), they have been proposed as biomarkers of WG wheat and rye intakes (5, 6). As biomarkers of WG wheat and rye intakes, they may be able to categorize subjects as low or high consumers of WG foods, particularly in epidemiologic studies (7) and as a biomarker of compliance with a WG diet in intervention studies (5, 8, 9).
Small cross-sectional studies (n < 60) examining fasting plasma AR associations have found positive correlations between plasma AR concentrations and AR intake (5), WG wheat and rye intakes (10–12), cereal fiber intake (13, 14), and overall WG cereal intake (10). One larger study found an association with WG wheat and rye intakes but used nonfasting samples (11). This study showed that rye bread, cereal fiber, and total fiber intakes were associated with plasma AR concentrations in postmenopausal Danish women, but these dietary factors explained only 8–12% of the variation in AR concentrations observed (11). Thus, there is a need to have a greater understanding of nondietary factors that may have an effect on plasma AR concentrations.
Most studies of plasma AR concentrations have been performed in Nordic populations that generally have a higher rye-based WG-intake than most other countries (15). The UK population has previously been reported to have low population-level intakes of WG foods (16–18), and most WG consumption is of products based on WG wheat rather than rye.
The WHOLEheart Study, a UK community-based WG intervention study (19) with both parallel intervention and repeated-measures components provided an opportunity to 1) evaluate plasma ARs as biomarkers in a long-term intervention study in subjects with a low habitual intake of WG with wheat as the main source of WG, and 2) evaluate potential determinants of fasting plasma AR concentrations.
SUBJECTS AND METHODS
Chemicals
Pentadecylresorcinol (15:0), heptadecylresorcinol (17:0), nonadecylresorcinol (19:0), eicosylresorcinol (20:0), heneicosylresorcinol (21:0), tricosylresorcinol (23:0), and pentacosylresorcinol (25:0) (all >95% purity by nuclear magnetic resonance) were purchased from Reseachem AG. The 20:0 AR was used as an internal standard because it is not found in nature. N-Methyl-N-trifluoroacetamide with 1% trimethylchlorosilane for derivatization was from Pierce Chemical Company. Oasis MAX solid phase extraction columns were from Waters SA. All solvents used were from Merck and were of chromatography grade.
Study design
The study design and primary results of the WHOLEheart Study were published elsewhere (19). Briefly, 316 overweight but otherwise healthy volunteers with a low estimated WG intake (<30 g/d), as determined with an FFQ, were randomly assigned to 1 of 3 groups (control, intervention group 1, and intervention group 2). Intervention group 1 was instructed to incorporate three 20-g servings/d of WG foods (provided free to the subjects) in their normal diet for 16 wk. Intervention group 2 was given the same instructions, but asked to increase their WG intake from three to six 20-g servings/d between weeks 8 and 16. The control group was asked to maintain their normal diet throughout the 16 wk and was not informed that the study concerned WG foods. Subjects could choose freely from prepackaged WG foods listed in Table 1 to reach their target WG intake. Fasting blood samples were taken at 0, 8, and 16 wk of the study. Subjects were recruited from the Newcastle upon Tyne and Cambridge areas of England. The WHOLEheart Study was approved by the Newcastle and North Tyneside Local Research Ethics Committee 1 and by the Newcastle upon Tyne Hospitals NHS Foundation Trust Research and Development Department and registered at isrtcn.org (ISRCT no. 83078872).
TABLE 1.
WG and AR contents of foods provided to WHOLEheart Study participants1
| Portion size |
WG (weight/portion) |
|||||
| Product name | WG | Total AR content | Male | Female | Male | Female |
| g/100 g | μg/g | g | g | g | g | |
| Whole-wheat bread | 54.8 | 488 | 37 | 37 | 20.0 | 20.0 |
| WG breakfast cereal with dried fruit2 | 57.6 | 349 | 67 | 59 | 38.6 | 34.0 |
| Multigrain cereal3 | 64.3 | 92 | 43 | 35 | 27.6 | 22.5 |
| Porridge oats 1 | 11.24 | 0 | 2544 | 2164 | 28.4 | 24.2 |
| Brown basmati rice | 344 | 0 | 2204 | 1814 | 74.8 | 61.5 |
| Whole-wheat pasta | 30.94 | 172 | 2574 | 2004 | 79.4 | 61.0 |
| Wheat breakfast biscuits5 | 80.8 | 389 | 42 | 36 | 33.8 | 29.0 |
| Porridge oats 2 | 52.8 | 0 | 38 | 38 | 20.0 | 20.0 |
| Oat snack bar6 | 60 | 5 | 38 | 38 | 22.8 | 22.8 |
| WG chips7 | 69 | 87 | 29 | 29 | 19.8 | 19.8 |
The WG contents were provided by the food manufacturers. AR, alkylresorcinol; WG, whole grain.
Nestlé Fruitful Shredded Wheat (Cereal Partners UK Ltd).
Nestlé Cheerios (Cereal Partners UK Ltd), made with WG corn, oats, barley, wheat, and rice.
Value represents cooked weight, whereas all other WG contents are per 100 g dry weight.
Weetabix (Weetabix Ltd).
Quaker Oat Bar (Pepsico Ltd).
Sun-Chips (Pepsico Ltd), made with WG corn, wheat, and oats.
Analytic methods
AR in food samples used during the intervention were extracted by using hot 1-propanol (1) and analyzed by using HLPC with Coularray electrochemical detection (2). The variation with replicate analyses was <4%. Plasma ARs were analyzed after extraction and purification by mixed-mode, anion-exchange, solid-phase extraction with gas chromatography–mass spectrometry (4, 20) with the addition of a deproteinization step with 0.5 mL 50% ethanol before extraction, because we found that this slightly improves AR peak areas. Concentrations of the individual AR homologs 17:0, 19:0, 21:0, 23:0, and 25:0 were quantified to determine the total plasma AR concentration. All samples were analyzed in duplicate and reanalyzed if the percentage difference between duplicates was >15%. Intra- and interbatch CVs for total ARs were <15% and <17% on the basis of reference plasma samples. Plasma total cholesterol, HDL cholesterol, triglycerides, NEFAs, glucose, insulin, and C-reactive protein were analyzed by automated enzymatic methods outlined by Brownlee et al (19).
Determination of WG and alkylresorcinol intake
WG, total cereal, and WG wheat intakes were determined by a 149-question semiquantitative FFQ based on the Cambridge HNR-MRC (Human Nutrition Research unit of the Medical Research Council) version of the European Prospective Investigation into Cancer and Nutrition FFQ. Questions covered food intake over the past 7 d. The FFQ was adapted to include questions on WG foods provided in the study, and those available on the UK market (19) and validated against 4-d food diaries (21). Data were converted into grams of WG per day by using the American Association of Cereal Chemists definition of WG (22) and previously published information on the WG content of foods available in the United Kingdom (16). AR intake from foods or food groups containing WG or refined wheat or rye was estimated from analysis of the WG foods provided by using data available for specific foods (2) or estimates based on the average known flour content in each food or food group (15).
Statistical analyses
Determinants of plasma AR concentrations were analyzed by using normalization for both concentration (nmol/L) and total lipids (nmol/mmol plasma lipids), because it was previously reported that, because ARs are transported in lipoproteins, correction for total lipids may remove some variation between subjects (23). Because the data were skewed, log transformation was applied, and all estimates are reported as geometric means and their 95% CIs.
Differences due to the intervention were analyzed by using a linear mixed model with intervention and identified covariates for plasma AR concentration as fixed effects. An interaction between intervention and time was included in the model to assess the differences over the 3 time points. A random subject effect (intercept and slope) was included in the model. Sidák's adjustment for multiple comparisons was used to correct for the multiplicity of comparisons at the different time points.
Potential determinants of plasma AR were analyzed by using a linear mixed model with potential intervention and physiologic predictors of AR concentration as fixed effects. The model included subject number as a random effect.
Spearman's rank correlation coefficients were calculated for the associations between total plasma AR and total WG intake, WG wheat intake, total AR intake, age, BMI, and clinical chemistry measures (triglyceride, NEFA, glucose, insulin, C-reactive protein, and fasting plasma total, LDL-, and HDL-cholesterol concentrations). Partial Spearman's rank correlation coefficients were also calculated for plasma total ARs with total WG intake, WG wheat intake, and total AR intake with correction for sex, study center, and plasma triglycerides and NEFAs.
Data were also analyzed as quartiles of plasma AR concentrations. The percentages of agreement and misclassification were determined for these quartiles on the basis of the ability of plasma ARs to correctly predict quartiles of WG intake. Subjects placed in the opposite quartile were considered to be grossly misclassified. Cohen's weighted κ statistic was used to determine the agreement between plasma ARs and self-reported WG intakes. Differences and associations were considered to be significant at P < 0.05.
Because repeated plasma and dietary intake measurements were made in this study, it was possible to determine both the ICC and the deattenuation factor. ICC is a measure of biomarker reproducibility under similar conditions, whereas a deattenuation factor can help improve observed correlation coefficients in cases in which there is known within-subject variation in a biomarker measurement or questionnaire response. The linear mixed model described above was used to estimate the variance components σ2B (between-subject variance) and σ2W (within-subject variance) for the determination of the ICC:
 |
The deattenuation factor for WG intake and plasmas AR was determined by using the following formula:
 |
where rtrue is the corrected correlation coefficient, robserved is the observed correlation coefficient, λ is the within- to between-individual variance ratio, and k is the number of replicates per individual. SAS 9.2 (SAS Institute Inc) and NCSS for Windows 2007 were used for statistical analysis.
RESULTS
A total of 266 of the 316 subjects who started the study completed the study, and data from these subjects were used in the analyses reported here. Details for the subjects who dropped out were reported previously (19). Complete dietary records were available for 252 subjects.
Of the foods provided for the intervention, those based primarily on WG wheat contained the highest concentration of ARs (Table 1), although it was notable that there was variation between the AR concentrations in the WG wheat products and similar WG contents, which reflects the wide natural variation previously reported (1, 24). Some products based on mixed WG contained low concentrations of ARs, whereas those containing predominantly oats or rice contained little or no ARs. Associations of plasma ARs were performed for total WG, WG wheat, and AR intakes (which accounts for the low concentrations of ARs in refined wheat products; 2).
Total WG, WG wheat, and AR intakes at baseline did not differ between groups (Table 2), nor did plasma lipid or AR concentrations (Table 3). WG wheat accounted for most of the WG eaten: of those reporting eating WG, 65% (60%, 71%), 65% (61%, 68%), and 64% (60%, 68%) [geometric mean (95% CI)] of total WG intake was WG wheat at baseline, 8 wk, and 16 wk, respectively. Only 17 of 252 subjects who completed the study with complete dietary records reported eating WG, but no WG wheat. After 8 wk of the intervention, a significant difference in plasma total AR concentrations was observed between both interventions 1 and 2 (at 3 servings WG/d) and the control group (low-WG diet; P = 0.0073 compared with intervention 1; P = 0.002 compared with intervention 2) (Table 4). After 16 wk, subjects in both intervention groups 1 and 2 were different from those in the control group (P < 0.0001), and no difference was observed between the 2 different recommended WG intakes (P = 0.1103). The interaction between time and treatment was significant (P < 0.001). When the AR concentration was expressed as nmol/mmol total lipids, there were no differences from when results were expressed as nmol/L (data not shown).
TABLE 2.
WG, WG wheat, and AR intakes in the different intervention groups throughout the study1
| Baseline |
Week 8 |
Week 16 |
|||||||
| Control | Intervention 1 | Intervention 2 | Control | Intervention 1 | Intervention 2 | Control | Intervention 1 | Intervention 2 | |
| (n = 96) | (n = 87) | (n = 83) | (n = 96) | (n = 87) | (n = 83) | (n = 96) | (n = 87) | (n = 83) | |
| Total WG intake (g/d) | 17.9 (15.1, 21.3) | 18.7 (15.0, 23.3) | 16.4 (12.9, 20.8) | 18.4 (14.7, 23.0) | 61.7 (53.4, 71.3)2 | 76.3 (69.7, 83.6)2 | 20.5 (15.8, 26.6) | 57.6 (49.3, 67.3)2 | 98.4 (85.7, 112.9)23 |
| WG wheat intake (g/d) | 11.4 (9.3, 14.1) | 15.2 (12.2, 19.0) | 13.5 (10.7, 17.0) | 13.4 (10.5, 17.3) | 43.7 (37.4, 51.0)2 | 48.7 (42.3, 53.2)2 | 15.6 (11.8, 20.7) | 38.5 (32.0, 46.2)2 | 66.8 (57.2, 78.0)23 |
| AR intake (mg/d) | 11.8 (9.8, 14.1) | 10.4 (7.9, 13.7) | 10.2 (8.0, 13.1) | 10.4 (8.4, 12.8) | 30.5 (26.3, 35.4)2 | 36.2 (32.3, 40.6)2 | 12.5 (10.1, 15.5) | 30.3 (25.8, 35.5)2 | 46.8 (39.6, 55.2)23 |
All values are geometric means (95% CIs) based on estimations from food-frequency questionnaires. Differences in intake between groups were determined by using ANOVA on log-transformed data. Intervention 1 subjects were requested to eat three 20-g servings WG/d for 16 wk, and Intervention 2 subjects were requested to eat three 20-g servings WG/d for 8 wk and six 20-g servings WG for the following 8 wk. Control subjects maintained their normal diet. AR, alkylresorcinol; WG, whole grain.
Significantly different from the control group at the same time point, P < 0.001.
Significantly different from Intervention 1 at the same time point, P < 0.001.
TABLE 3.
Baseline characteristics of the subjects who completed the WHOLEheart Study1
| Control (n = 96) | Intervention 1 (n = 87) | Intervention 2 (n = 83) | |
| Age (y)2 | 45.6 ± 10.03 | 45.9 ± 10.1 | 45.7 ± 9.9 |
| Sex (% male)2 | 49.0 | 50.0 | 51.2 |
| BMI (kg/m2)2 | 30.0 ± 4.0 | 30.0 ± 3.7 | 30.3 ± 4.5 |
| Total cholesterol (mmol/L) | 5.3 ± 1.0 | 5.2 ± 0.8 | 5.4 ± 1.0 |
| HDL (mmol/L) | 1.3 ± 0.3 | 1.3 ± 0.3 | 1.3 ± 0.2 |
| LDL (mmol/L) | 3.3 ± 0.9 | 3.2 ± 0.7 | 3.4 ± 0.8 |
| Triglycerides (mmol/L) | 1.7 ± 1.0 | 1.6 ± 0.8 | 1.6 ± 0.8 |
| NEFAs (μmol/L) | 540.1 ± 227.5 | 560.7 ± 186.8 | 532.2 ± 202.8 |
| Total alkylresorcinols (nmol/L) | 87.7 ± 95.5 | 95.0 ± 120.1 | 76.1 ± 86.6 |
Intervention 2 subjects were requested to eat three 20-g servings of WG/d for 8 wk and six 20-g servings of WG for the following 8 wk. Control subjects maintained their normal diet. No significant differences were observed between groups at baseline (linear mixed model using log-transformed data). NEFAs, nonesterified fatty acids; WG, whole grain.
Used in participant randomization.
Mean ± SD (all such values).
TABLE 4.
Alkylresorcinol concentrations in the different intervention groups at baseline, 8 wk, and 16 wk1
| Control (n = 96) | Intervention 1 (n = 87) | Intervention 2 (n = 83) | |
| Baseline | 56.7 (68.3, 66.5) | 63.4 (51.9, 75.0) | 50.7 (42.1, 61.1) |
| Week 8 | 63.7 (53.9, 75.3) | 77.6 (63.5, 94.7)2 | 96.5 (80.2, 116.0)3 |
| Week 16 | 66.1 (55.8, 78.4) | 95.0 (78.8, 114.5)4 | 116.1 (92.9, 145.1)45 |
All values are geometric means (95% CIs). Differences due to group and time were assessed by using a linear mixed model including sex, study center, and plasma triglycerides and nonesterified fatty acids as fixed effects and subject as a random factor. Intervention 1 subjects were requested to eat three 20-g servings of WG/d for 16 wk, and Intervention 2 subjects were requested to eat three 20-g servings of WG/d for 8 wk and six 20-g servings of WG for the following 8 wk. Control subjects maintained their normal diet. There was a significant interaction between time and treatment, P < 0.0001. WG, whole grain.
Significantly different from control: 2P = 0.0073, 3P = 0.0020, 4P < 0.0001.
Nonsignificantly different from Intervention 1, P = 0.1103.
The median plasma AR 17:0:21:0 ratio, which can distinguish between wheat and rye ARs (∼0.1 for wheat and ∼0.6 for rye) (23), was 0.066 (25th–75th percentiles: 0.033–0.068), which confirmed that WG rye was not commonly consumed by subjects during the study, as reported in the FFQ. As was found previously, an analysis of the individual homologs did not alter the outcomes; therefore, only the results for total ARs are reported.
Classification of subjects into the same or adjacent quartiles of WG intake ranged from 70% to 79% at 16 wk (Table 5), with a gross misclassification rate (subjects in opposite quartile to that predicted) of 9% to 12.1%. Agreement as rated by Cohen's weighted κ was 0.131 at baseline and 0.238 at 16 wk, which suggested that plasma ARs in this study were slight to fair at classifying subjects into the correct quartile of WG intake.
TABLE 5.
Subjects grouped into quartiles of plasma total AR concentrations and corresponding self-reported WG intake after 16 wk of the WHOLEheart Study intervention1
| Q1 (n = 65) | Q2 (n = 64) | Q3 (n = 64) | Q4 (n = 65) | κ2 | |
| AR (nmol/L) | 28.4 (25.8, 31.3)3 | 61.0 (58.4, 63.8) | 117.6 (112.0, 123.6) | 307.8 (276.5, 342.6) | |
| Total WG intake (g/d) | 32.6 (24.6, 43.2) | 45.5 (34.2, 60.5) | 49.0 (36.6, 65.7) | 86.0 (74.0, 99.8) | |
| Same or adjacent Q (%) | 72.7 | 79.2 | 74.2 | 70.2 | 0.238 |
| Gross misclassification (%) | 12.1 | — | — | 9.0 |
AR, alkylresorcinol; Q, quartile; WG, whole grain.
Cohen's weighted kappa statistic.
Geometric mean; 95% CI in parentheses (all such values).
The ICC determined for plasma ARs in this study was 0.48, with women having a slightly higher ICC than men (0.47 compared with 0.43). The deatteunuation factor for WG intake and plasma ARs was 1.61.
Study center (Newcastle compared with Cambridge; P = 0.0056), sex (P = 0.0012), fasting plasma NEFAs (P = 0.0015), and triglycerides (P < 0.0001) were significant predictors of AR concentration (Table 6). Notable estimated effect sizes included a 10-g increase in WG intake leading to a 6% increase in plasma AR, women having an average 23.3% lower plasma AR concentration, and a 100-μmol/L increase in NEFAs leading to a 3.8% decrease in plasma ARs.
TABLE 6.
Predictors of AR concentrations: effect size of studied variables on plasma AR concentrations1
| Estimated effect size2 | P value | |
| % | ||
| WG intake (per 10-g increase) | 6 (4, 7) | <0.0001 |
| Sex (females vs males) | −23.3 (−34, −9.9) | 0.0012 |
| Study center (Newcastle vs Cambridge) | −23.7 (−37, −7.7) | 0.0056 |
| NEFAs (per 100-μmol/L increase) | −3.8 (−5.4, −1.7) | 0.0015 |
| Triglycerides (per 1-mmol/L increase) | 27 (17.8, 37) | <0.0001 |
Only significant variables are shown. AR, alkylresorcinol; NEFAs, nonesterified fatty acids; WG, whole grain.
95% CI in parentheses. Effect size was determined by using a linear mixed model on log-transformed values, with potential predictors as fixed factors and subject as a random factor. Only significant predictors were included in the final model.
Total WG, WG wheat, and total AR intakes were associated with plasma AR concentrations at all time points. The strongest correlations were found when there was the greatest range of WG intakes (ie, week 16) (Table 7). In general, associations for plasma AR concentrations with WG wheat intake only, or AR intake, were stronger than for total WG intake. Correction for the correlations for identified covariates (partial Spearman's rank correlation coefficient) only increased correlation coefficients in male subjects. Application of the deattenuation factor increased the correlation between WG intake and plasma ARs for all subjects to 0.56.
TABLE 7.
Raw and partial Spearman's rank correlation coefficients between plasma total AR concentrations and measures of WG intake and significant plasma determinants of AR concentrations after the 16-wk intervention1
| All subjects | Women | Men | |
| Total WG intake | |||
| n | 252 | 124 | 128 |
| Raw Spearman | 0.35* | 0.33* | 0.37* |
| Partial Spearman2 | 0.36* | 0.33* | 0.44* |
| WG wheat intake | |||
| n | 252 | 124 | 128 |
| Raw Spearman | 0.43* | 0.42* | 0.44* |
| Partial Spearman2 | 0.45* | 0.43* | 0.49* |
| AR intake | |||
| n | 252 | 124 | 128 |
| Raw Spearman | 0.39* | 0.32* | 0.42* |
| Partial Spearman2 | 0.39* | 0.32* | 0.50* |
| Plasma triglycerides | |||
| n | 266 | 133 | 133 |
| Raw Spearman | 0.20** | 0.02 | 0.34* |
| Partial Spearman2 | — | — | — |
| Plasma NEFAs | |||
| n | 266 | 133 | 133 |
| Raw Spearman | −0.17*** | −0.14 | −0.19*** |
| Partial Spearman2 | — | — | — |
*P ≤ 0.001, **P ≤ 0.01, ***P ≤ 0.05. AR, alkylresorcinol; NEFAs, nonesterified fatty acids; WG, whole grain.
Corrected for study center, plasma triglycerides, plasma NEFAs, and sex when appropriate.
DISCUSSION
This is the first time that the response of plasma AR concentration to a WG intervention has been investigated in a population of >60 subjects and the first time in a large-scale intervention in a population not eating a high amount of rye. In this study, plasma AR concentrations increased compared with control subjects with 3 servings WG/d (60 g/d) after 8 wk, although no further significant increase was observed from 3 to 6 servings WG/d (120 g/d) after another 8 wk of the intervention. The lack of a significant increase in plasma ARs for 3 compared with 6 servings WG/d may have been due to an increase in the intake of nonwheat-based foods (eg, oats or rice) to reach the target of 6 servings/d. Despite the considerable inter- and intraindividual difference in the response of plasma ARs to the intervention, it is clear that they do respond to a WG intervention in a free-living setting. All WG-based intervention studies that have reported plasma AR concentrations have found similar results, ie, whereas mean AR concentrations increased, there was considerable interpersonal variation for the response to an increase in intake of ARs (6, 23, 25–27) and overlap between different individuals on either a high- or low-WG diet because of interindividual variation (6). Because subjects in the control group were not told to avoid WG foods, and the criterion for inclusion was a WG intake <31 g /d, some control subjects did report eating more WG food during the study, because they were not told to avoid this. Mean AR concentrations were relatively high compared with those who reported low- or WG-free diets in previous studies (28).
The reproducibility (based on the ICC) of AR measurements in this study was similar to that found in other studies in free-living subjects (12, 29) for fasting plasma samples taken 2–4 mo apart. As in a previous Swedish study, no sex difference in the ICC was found; however, German women were found to have a much higher ICC than German men (0.55 compared with 0.17) (29).
When subjects were grouped into quartiles on the basis of plasma AR concentrations, they were moderately successful at correctly classifying subjects into quartiles of WG intake, and this was best at the greatest range of AR concentrations (after 16 wk). The relatively high percentage of gross misclassification (9–12.1%) suggests that a single measurement of plasma ARs is less robust for estimating an individual's self-reported WG intake under relatively free-living conditions. What proportion of this classification error is due to variation in individual plasma AR responses and what proportion is due to measurement error of WG intake remains to be established.
The predictors of plasma AR concentrations determined in this study were WG intake (including WG wheat and AR intakes), sex, study center, and plasma NEFAs and triglycerides. Mean plasma AR concentrations were higher in men than in women, similar to recent studies on plasma AR concentrations in free-living German and Swedish adults (12, 29). Adjustment of differences in AR concentrations for total WG intake did not change the result for sex, which suggests that this difference is not related to a greater overall intake of AR. Sex differences have been found for γ-tocopherol metabolism, with females having an apparently faster metabolism of this E vitamer (30), and because ARs are proposed to be metabolized in a manner similar to that of γ-tocopherol (31), it is possible that similar mechanisms are at play. No studies have looked at sex differences for the AR metabolites 3,5-dihydroxybenzoic acid and 3,5-dihydroxyphenylpropianoic acid (32). Despite the significantly larger increase in plasma AR concentrations of men in response to the WG intervention, this did not have a large effect on the overall results for intervention effects or correlations, with similar results being found for both sexes. Differences in fat distribution may also explain some of the sex differences as discussed below. An association with plasma lipids was expected, because plasma ARs are transported in lipoproteins (23). The lipid fractions measured were associated only with plasma triglycerides and not with lipoproteins, which suggested that factors that would influence transient increases in plasma triglycerides (eg, time since last meal) may be important for the repeatability of plasma AR measurements (11). Plasma AR were negatively associated with plasma NEFAs—a finding that is interesting in the context of recent in vitro work that suggests that high concentrations of AR may inhibit hormone-sensitive lipase (33), which mobilizes free fatty acids from adipose tissue. The results may also be due to differences in fasting time (greater NEFA release during fasting coinciding with lower circulating concentrations of AR). More work is needed to understand whether ARs play a role in inhibiting lipid mobilization in vivo.
Correlations for total WG, WG wheat, and AR intakes were in the range of those found previously (Table 8). The type of dietary assessment method and range of intervention appear to be key predictors of the correlation, with weighed food diaries or specific WG FFQs appearing to lead to higher raw correlations (r > 0.5), whereas diet records and general dietary FFQs tend to lead to significant but lower correlations (r = 0.25–0.43) (Table 8). Deattenuated correlation coefficients for WG intake were similar to those found with the use of weighed food diaries to estimate WG intake. Even with low intakes of WG (<30 g WG/d) at baseline, moderate associations were found between WG wheat and AR intakes and plasma AR concentrations, which indicates that they are still associated with WG intakes at low levels of consumption. Correlation coefficients differed between the sexes, especially for AR intake. This may be due to differences in dietary reporting between the sexes or may be linked to the difference in correlations between plasma ARs and plasma triglycerides. Correction for the identified covariates had little effect on the observed correlations for women, but did lead to substantial improvements in the correlations observed for men. The difference in plasma triglyceride concentrations may explain this improvement, because male subjects had higher mean plasma triglycerides at all time points, and plasma ARs were only correlated with plasma triglycerides in male subjects. Plasma triglycerides are known to be influenced by fat distribution, including visceral fat deposits (34, 35), and these differences may explain variation in plasma ARs, even if BMI was not found to be associated with plasma As in this study. The presence of AR in reasonable quantities in adipose tissue could support this idea, although this has only been measured in female subjects thus far (36). The apparent sex difference observed in this study may explain why correction for plasma lipids has not improved correlations in past studies in largely female subjects (6, 11), and future studies should consider correcting for plasma triglyceride concentrations in males.
TABLE 8.
Correlations of plasma AR concentrations with different measurements of WG intake from previously published studies and this study1
| Reference | No. of subjects | Sex | Country | Study type | Dietary assessment method | Food or nutrient | Plasma AR correlation | P value |
| Aubertin-Leheudre et al (14) | 56 | F | Finland | Free-living | 5DFR | Cereal fiber | 0.38 | 0.004 |
| Linko et al (13) | 39 | F | Finland | Intervention | 4DFR | Insoluble fiber | 0.39 | 0.013 |
| Landberg et al (6) | 28 | 20 F, 8 M | Sweden | Intervention | 3DWFD | AR | 0.58 | <0.0001 |
| Ross et al (10) | 29 | 17 F, 12 M | Switzerland | Free-living | 3DWFD for F, WG-FFQ for M | Total WG | 0.54–57 | <0.0001 |
| Landberg et al (11) | 360 | F | Denmark | Free-living | FFQ | Rye bread | 0.25 | <0.0001 |
| Andersson et al (12) | 51 | M and F | Sweden | Free-living | 3DWFR | Total WG | 0.24 | NS |
| Andersson et al (12) | 51 | M and F | Sweden | Free-living | 3DWFR | WG rye and wheat | 0.53 | <0.001 |
| Current study | 252 | 124 F, 128 M | United Kingdom | Intervention2 | FFQ | Total WG | 0.32 | <0.001 |
| Current study | 252 | 124 F, 128 M | United Kingdom | Intervention2 | FFQ | WG wheat | 0.39 | <0.001 |
| Current study | 252 | 124 F, 128 M | United Kingdom | Intervention2 | FFQ | AR | 0.38 | <0.001 |
AR, alkylresorcinol; WG, whole grain; WG-FFQ, whole-grain food-frequency questionnaire; 3DWFD, 3-d weighed food diary; 4DFR, 4-d food record; 5DFR, 5-d food record.
After 16 wk of intervention.
The subjects in this study were noted to have added WG to their normal diets, rather than to have replaced refined carbohydrates in their diets as requested (19, 37), and were all overweight or obese [BMI (in kg/m2) > 25; a population known to have problems with diet reporting; 38], which suggests that variations in both biomarker response and self-reported WG intake contribute to the overall wide variation in plasma ARs observed. However, because the use of plasma AR concentrations as nonsubjective markers of WG wheat and rye intakes (and potentially total WG intake) is mainly targeted for use in epidemiologic settings (5, 11, 29), errors in dietary reporting are expected. Because the United Kingdom generally has a low consumption of WG (18), the finding that plasma AR concentrations reflect both WG wheat and total WG intake in a population of low-WG consumers is an important complement to those studies based in the Nordic countries, where WG intake is generally higher and more often based on rye (15).
In an analysis from a large intervention study with WG, the plasma AR concentration was found to correlate with WG intake—both before and during the intervention. The phenotypic determinants of plasma AR identified were sex and plasma triglyceride and NEFA concentrations. Adjustment for these factors had no effect on the performance of plasma ARs for differentiating between different WG intakes, although the link between plasma triglycerides and plasma ARs in men warrants further investigation. Overall, the results from this study support the idea that plasma AR concentrations can function as biomarkers of WG intake.
Acknowledgments
We thank Andreas Rytz for his helpful discussions on statistical aspects of this work.
The authors’ responsibilities were as follows—CJS and ABR: designed the analytic approach; IAB, CJS, and SAJ: designed the intervention study and provided the samples; ABR and AB: performed the laboratory analyses; HNM and ABR: performed the statistical analyses; and SK: provided resources for the analysis. All authors were involved in the data interpretation and manuscript preparation. ABR, AB, HNM, and SK all work for Nestec SA (part of the Nestlé food and beverage company), which produces WG products. None of the other authors declared any conflict of interest.
Footnotes
Abbreviations used: AR, alkylresorcinol; FFQ, food-frequency questionnaire; ICC, intraclass correlation coefficient; NEFA, nonesterified fatty acid; WG, whole grain.
REFERENCES
- 1.Ross AB, Shepherd MJ, Schupphaus M, Sinclair V, Alfaro B, Kamal-Edlin A, Åman P. Alkylresorcinols in cereals and cereal products. J Agric Food Chem 2003;51:4111–8 [DOI] [PubMed] [Google Scholar]
- 2.Ross AB, Kochhar S. Rapid and sensitive analysis of alkylresorcinols from cereal grains and products using HPLC-Coularray-based electrochemical detection. J Agric Food Chem 2009;57:5187–93 [DOI] [PubMed] [Google Scholar]
- 3.Ross AB, Kamal-Eldin A, Lundin EA, Zhang JX, Hallmans G, Åman P. Cereal alkylresorcinols are absorbed by humans. J Nutr 2003;133:2222–4 [DOI] [PubMed] [Google Scholar]
- 4.Linko AM, Parikka K, Wähälä K, Adlercreutz H. Gas chromatographic-mass spectrometric method for the determination of alkylresorcinols in human plasma. Anal Biochem 2002;308:307–13 [DOI] [PubMed] [Google Scholar]
- 5.Ross AB, Kamal-Eldin A, Åman P. Dietary alkylresorcinols: absorption, bioactivities, and possible use as biomarkers of whole-grain wheat- and rye-rich foods. Nutr Rev 2004;62:81–95 [DOI] [PubMed] [Google Scholar]
- 6.Landberg R, Kamal-Eldin A, Andersson A, Vessby B, Åman P. Alkylresorcinols as biomarkers of whole-grain wheat and rye intake: plasma concentration and intake estimated from dietary records. Am J Clin Nutr 2008;87:832–8 [DOI] [PubMed] [Google Scholar]
- 7.Wild CP, Andersson C, O'Brien NM, Wilson L, Woods JA. A critical evaluation of the application of biomarkers in epidemiological studies on diet and health. Br J Nutr 2001;86(Suppl 1):S37–53 [DOI] [PubMed] [Google Scholar]
- 8.Kaaks RJ. Biochemical markers as additional measurements in studies of the accuracy of dietary questionnaire measurements: conceptual issues. Am J Clin Nutr 1997;65(suppl):1232S–9S [DOI] [PubMed] [Google Scholar]
- 9.van Dam RM, Hu FB. Are alkylresorcinols accurate biomarkers for whole grain intake? Am J Clin Nutr 2008;87:797–8 [DOI] [PubMed] [Google Scholar]
- 10.Ross AB, Pineau N, Kochhar S, Bourgeois A, Beaumont M, Decarli B. Validation of a FFQ for estimating whole-grain cereal food intake. Br J Nutr 2009;102:1547–51 [DOI] [PubMed] [Google Scholar]
- 11.Landberg R, Kamal-Eldin A, Åman P, Christensen J, Overvad K, Tjønneland A, Olsen A. Determinants of plasma alkylresorcinol concentration in Danish post-menopausal women. Eur J Clin Nutr 2011;65:94–101 [DOI] [PubMed] [Google Scholar]
- 12.Andersson A, Marklund M, Diana M, Landberg R. Plasma alkylresorcinol concentrations correlate with whole grain wheat and rye intake and show moderate reproducability over a 2- to 3-month period in free-living Swedish adults. J Nutr 2011;141:1712–8 [DOI] [PubMed] [Google Scholar]
- 13.Linko AM, Juntunen KS, Mykkanen HM, Adlercreutz H. Whole-grain rye bread consumption by women correlates with plasma alkylresorcinols and increases their concentration compared with low-fiber wheat bread. J Nutr 2005;135:580–3 [DOI] [PubMed] [Google Scholar]
- 14.Aubertin-Leheudre M, Koskela A, Marjamaa A, Adlercreutz H. Plasma alkylresorcinols and urinary alkylresorcinol metabolites as biomarkers of cereal fiber intake in Finnish women. Cancer Epidemiol Biomarkers Prev 2008;17:2244–8 [DOI] [PubMed] [Google Scholar]
- 15.Ross AB, Becker W, Chen Y, Kamal-Eldin A, Åman P. Intake of alkylresorcinols from wheat and rye in the United Kingdom and Sweden. Br J Nutr 2005;94:496–9 [DOI] [PubMed] [Google Scholar]
- 16.Thane CW, Jones AR, Stephen AM, Seal CJ, Jebb SA. Whole-grain intake of British young people aged 4-18 years. Br J Nutr 2005;94:825–31 [DOI] [PubMed] [Google Scholar]
- 17.Thane CW, Jones AR, Stephen AM, Seal CJ, Jebb SA. Comparative whole-grain intake of British adults in 1986-7 and 2000-1. Br J Nutr 2007;97:987–92 [DOI] [PubMed] [Google Scholar]
- 18.Lang R, Thane CW, Bolton-Smith C, Jebb SA. Consumption of whole-grain foods by British adults: findings from further analysis of two national dietary surveys. Public Health Nutr 2003;6:479–84 [DOI] [PubMed] [Google Scholar]
- 19.Brownlee IA, Moore C, Chatfield M, Richardson DP, Ashby P, Kuznesof SA, Jebb SA, Seal CJ. Markers of cardiovascular risk are not changed by increased whole-grain intake: the WHOLEheart study, a randomised, controlled dietary intervention. Br J Nutr 2010;104:125–34 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Landberg R, Åman P, Kamal-Eldin A. A rapid gas chromatography-mass spectrometry method for quantification of alkylresorcinols in human plasma. Anal Biochem 2009;385:7–12 [DOI] [PubMed] [Google Scholar]
- 21.Seal CJ, Brownlee IA, Kuznesof SA, Chatfield M, Moore C, Jebb SA. Food Standards Agency Final Report: randomised controlled trial to test the impact of increased consumption of wholegrain foods on cardiovascular disease risk (the WHOLEheart study). Food Standards Agency UK, 2010. Available from: http://www.foodbase.org.uk//admintools/reportdocuments/547-1-954_N02036_final_report.pdf [Google Scholar]
- 22. American Association of Cereal Chemists AACC members agree on definition of whole grain. (29-12-1999). Available from: http://www.aaccnet.org/news/pdfs/wgPR.pdf (cited 30 May 2011) [Google Scholar]
- 23.Linko-Parvinen AM, Landberg R, Tikkanen MJ, Adlercreutz H, Penalvo JL. Alkylresorcinols from whole-grain wheat and rye are transported in human plasma lipoproteins. J Nutr 2007;137:1137–42 [DOI] [PubMed] [Google Scholar]
- 24.Andersson AAM, Kamal-Eldin A, Fras A, Boros D, Åman P. Alkylresorcinols in wheat varieties in the HEALTHGRAIN Diversity Screen. J Agric Food Chem 2008;56:9722–5 [DOI] [PubMed] [Google Scholar]
- 25.Landberg R, Åman P, Friberg LE, Vessby B, Adlercreutz H, Kamal-Eldin A. Dose response of whole-grain biomarkers: alkylresorcinols in human plasma and their metabolites in urine in relation to intake. Am J Clin Nutr 2009;89:290–6 [DOI] [PubMed] [Google Scholar]
- 26.Landberg R, Kamal-Eldin A, Andersson SO, Johansson JE, Zhang JX, Hallmans G, Åman P. Reproducibility of plasma alkylresorcinols during a 6-week rye intervention study in men with prostate cancer. J Nutr 2009;139:975–80 [DOI] [PubMed] [Google Scholar]
- 27.Ross AB, Bruce SJ, Blondel-Lubrano A, Oguey-Araymon S, Beaumont M, Bourgeois A, Nielsen-Moennoz C, Vigo M, Fay LB, Kochhar S, et al. A whole-grain cereal-rich diet increases plasma betaine, and tends to decrease total and LDL-cholesterol compared with a refined-grain diet in healthy subjects. Br J Nutr 2011;105:1492–502 [DOI] [PubMed] [Google Scholar]
- 28.Ross AB. Present status and perspectives on the use of alkylresorcinols as biomarkers of wholegrain wheat and rye intake. J Nutr Metab ( in press ) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Montonen J, Landberg R, Kamal-Eldin A, Åman P, Knueppel S, Boeing H, Pischon T. Reliability of fasting plasma alkylresorcinol concentrations measured 4 months apart. Eur J Clin Nutr 2010;64:698–703 [DOI] [PubMed] [Google Scholar]
- 30.Frank J, Lee S, Leonard SW, Atkinson JK, Kamal-Eldin A, Traber MG. Sex differences in the inhibition of gamma-tocopherol metabolism by a single dose of dietary sesame oil in healthy subjects. Am J Clin Nutr 2008;87:1723–9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ross AB, Chen Y, Frank J, Swanson JE, Parker RS, Kozubek A, Lundh T, Vessby B, Åman P, Kamal-Eldin A. Cereal alkylresorcinols elevate gamma-tocopherol levels in rats and inhibit gamma-tocopherol metabolism in vitro. J Nutr 2004;134:506–10 [DOI] [PubMed] [Google Scholar]
- 32.Ross AB, Åman P, Kamal-Eldin A. Identification of cereal alkylresorcinol metabolites in human urine-potential biomarkers of wholegrain wheat and rye intake. J Chromatogr B Analyt Technol Biomed Life Sci 2004;809:125–30 [DOI] [PubMed] [Google Scholar]
- 33.Andersson U, Dey ES, Holm C, Degerman E. Rye bran alkylresorcinols suppress adipocyte lipolysis and hormone-sensitive lipase activity. Mol Nutr Food Res 2011;55(suppl):S290–3 [DOI] [PubMed] [Google Scholar]
- 34.Halkes CJM, Castro Cabezas M, Van Wijk JPH, Erkelens DW. Gender differences in diurnal triglyceridemia in lean and overweight subjects. Int J Obes Relat Metab Disord 2001;25:1767–74 [DOI] [PubMed] [Google Scholar]
- 35.Couillard C, Bergeron N, Prud'homme D, Bergeron J, Tremblay A, Bouchard C, Mauriège P, Després JP. Gender difference in postprandial lipemia: Importance of visceral adipose tissue accumulation. Arterioscler Thromb Vasc Biol 1999;19:2448–55 [DOI] [PubMed] [Google Scholar]
- 36.Jansson E, Landberg R, Kamal-Eldin A, Wolk A, Vessby B, Aman P. Presence of alkylresorcinols, potential whole grain biomarkers, in human adipose tissue. Br J Nutr 2010;104:633–6 [DOI] [PubMed] [Google Scholar]
- 37.McKeown NM, Jacobs DR. In defence of phytochemical-rich dietary patterns. Br J Nutr 2010;104:1–3 [DOI] [PubMed] [Google Scholar]
- 38.Willett WC. Nutritional epidemiology. New York, NY: Oxford UniversityPress, 1998 [Google Scholar]