Oats are a rich source of β-glucan, a viscous, soluble fibre recognised for its cholesterol-lowering properties. The attenuation of blood cholesterol levels by oats was first reported in 1963 in a study that substituted white bread for oat bread containing 140 g of rolled oats( Reference de Groot, Luyken and Pikaar 1 ). Since then, a large number of studies have been conducted to assess the effects of oats on cholesterol levels, especially LDL-cholesterol, for the reduction of CVD risk. On the basis of the extensive evidence relating an inverse association between β-glucan intake and LDL-cholesterol, several countries have currently approved health claims of oat β-glucan and its LDL-cholesterol-lowering effect or CVD risk reduction( 2 – 6 ).
At present, the primary lipid target for CVD risk reduction is LDL-cholesterol, with non-HDL-cholesterol and apoB as alternate targets. However, it has been suggested that non-HDL-cholesterol and apoB may be more relevant targets as non-HDL-cholesterol contains all atherogenic cholesterol and there is one apoB on all atherogenic lipoprotein particles. Furthermore, both non-HDL-cholesterol and apoB have been shown to be highly correlated with CVD risk, especially when LDL-cholesterol appears to be within the normal range( Reference Saenger 7 ), and have been added to the Third Report of the National Cholesterol Education Program – Adult Treatment Panel and the Canadian Cardiovascular Society (CCS) lipid guidelines as alternate lipid targets for CVD risk reduction( 8 , Reference Anderson, Gregoire and Hegele 9 ).
In contrast to the established relationship between oat β-glucan and LDL-cholesterol, there is currently little understanding of the relationship between oat β-glucan and alternate markers of CVD risk – that is, non-HDL-cholesterol and apoB. The objective of this study was to conduct a systematic review and meta-analysis of randomised-controlled trials (RCT) to analyse the evidence of the effect of oat β-glucan on LDL-cholesterol, as well as for the first time on non-HDL-cholesterol and apoB, for CVD risk reduction.
Methods
Protocol and registration
The Cochrane Handbook for Systematic Reviews of Interventions was used to plan and conduct this meta-analysis( 10 ). Results were reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines( Reference Moher, Liberati and Tetzlaff 11 ). The review protocol is available online at ClinicalTrials.gov (registration no. NCT02068248).
Search strategy and data sources
MEDLINE, Embase, CINAHL and the Cochrane Central Register of Controlled Trials were searched, using the search strategy presented in the online Supplementary Table S1, through 5 November 2015, to identify RCT investigating the effects of oat β-glucan on LDL-cholesterol, non-HDL-cholesterol or apoB. Manual searches of references supplemented the electronic search. One unpublished trial from our group was included in the analysis( Reference Panahi 12 ). No language restrictions were imposed.
Study eligibility
All titles and abstracts were initially assessed according to inclusion and exclusion criteria outlined in the online Supplementary Table S2. In brief, only RCT that investigated the effects of supplementing β-glucan from oat products on LDL-cholesterol, non-HDL-cholesterol and/or apoB were included in the analysis( 13 , Reference Greer, Mosser and Logan 14 ). Trials that did not report non-HDL-cholesterol but provided enough information to permit the calculation of non-HDL-cholesterol (total cholesterol (TC)−HDL-cholesterol) were also considered. Included trials involved any population, had a minimum follow-up period of 3 weeks, as per the United States Food and Drug Administration (US FDA)( 13 , Reference Kris-Etherton and Dietschy 15 ), administered any dose of β-glucan and provided enough information to calculate a treatment effect.
Data extraction and quality assessment
H. V. T. H. and A. Z. independently reviewed all studies that passed the initial assessment. A standardised proforma was used to extract relevant data including sample size, subject characteristics (health status, sex, age, weight, etc.), study setting (inpatient/outpatient), study design (parallel/cross-over), follow-up duration, β-glucan dose, comparator, background diet, energy balance and funding source. If the β-glucan content was not reported, oat bran and whole oats were estimated at 6·9 and 5·0 %( Reference Chen and Anderson 16 , Reference Anderson and Bridges 17 ) β-glucan, respectively, and oat soluble fibre was estimated at 92·5 % β-glucan( Reference Whitehead, Beck and Tosh 18 ). The mean and standard deviation values were extracted for LDL-cholesterol, non-HDL-cholesterol and apoB at baseline and follow-up for both control and intervention groups. When standard deviation values were not reported, they were derived from available data (95 % CI, P-values, t or F statistics, sem) using standard formulae( 10 ). If available, mean change from baseline and standard deviation values for both groups, mean end difference and standard deviation values, and/or mean change from baseline difference and standard deviation values between groups were also extracted.
The Heyland Methodological Quality Score (MQS) was used to assess study quality( Reference Heyland, Novak and Drover 19 ). Points were given on the basis of methods (randomisation, blinding and analysis), sample (selection, comparibility and follow-up) and intervention (protocol, co-intervention and cross-overs) and a maximum of 13 points could be received. Trials that received scores of≥8 were considered to be of higher quality.
The Cochrane Risk of Bias Tool was used to assess the study risk of bias( 10 ). Domains of bias assessed were sequence generation, allocation concealment, blinding, outcome data and outcome reporting. Trials were considered high risk when methodological flaws were likely to have affected the true outcome, low risk if the flaw was deemed inconsequential and unclear risk when insufficient information was provided to permit judgement. Authors were contacted for additional information where necessary. All disagreements on the MQS and Risk of Bias Tool were resolved by consensus.
Data management and analysis
Data were analysed using Review Manager (RevMan), version 5.3 (The Nordic Cochrane Centre, The Cochrane Collaboration), for primary analyses. The difference between the change from baseline values for the intervention and the control arms was derived from each trial for the end points of LDL-cholesterol, non-HDL-cholesterol and apoB. When non-HDL-cholesterol was not reported, it was calculated from aggregate data by subtracting HDL-cholesterol from TC. A previously developed formula was used to calculate sd for calculated values of non-HDL-cholesterol( Reference Ha 20 ). If change from baseline values were not available, end-of-treatment values were used. For trials containing multiple intervention or control arms, a weighted average was applied to combine them in order to create a single pair-wise comparison and to mitigate the unit-of-analysis error( 10 ). Paired analyses were conducted for all cross-over studies( Reference Elbourne, Altman and Higgins 21 ). Where necessary, a pooled correlation coefficient was derived and used for calculation of an imputed sd for the between-treatment difference. Correlation coefficients between baseline and end-of-treatment values within each individual cross-over trial were derived from the reported within- and between-treatment sd according to a published formula( Reference Elbourne, Altman and Higgins 21 ). These correlation coefficients were transformed into z-scores and sd, meta-analysed using inverse-variance weighing and back-transformed to derive the pooled correlation coefficient. For end points, when a pooled correlation coefficient for imputing missing sd could not be derived, a value of 0·50 was assumed, as it is a conservative estimate for an expected range of 0–1. The values derived from each trial were pooled and analysed for LDL-cholesterol, non-HDL-cholesterol and apoB using the generic inverse-variance method with random effects models, which were used even in the absence of statistically significant between-study heterogeneity, as they yield more conservative summary effect estimates in the presence of residual heterogeneity. Data are expressed as mean differences (MD) with 95 % CI. Furthermore, results are presented separately according to individual study inclusion criteria. The hypercholesterolic group included studies that recruited participants who were hypercholesterolaemic, and the unclassified group included studies that did not specify that participants had to be hypercholesterolaemic. A two-sided P-value<0·05 was set as the level of significance for comparisons of MD.
Inter-study heterogeneity was tested using Cochran’s Q statistic and quantified using the I 2-statistic with a significance level set at P<0·10. I 2 values <50, ≥50 to <75 and ≥75 % were considered to be evidence for ‘moderate,’ ‘substantial’ and ‘considerable’ heterogeneity, respectively( 10 ). Sources of heterogeneity were explored using sensitivity and subgroup analyses. To determine whether a single trial exerted undue influence on the overall results, sensitivity analyses were performed in which each individual trial was removed from the meta-analysis and the effect size was re-calculated with the remaining trials. Sensitivity analyses were also undertaken using correlation coefficients of 0·25, 0·50 and 0·75 to determine whether the overall results were robust to the use of different derived correlation coefficients in paired analyses of cross-over trials. A priori subgroup analyses (continuous and categorical) were conducted for baseline values of LDL-cholesterol, non-HDL-cholesterol and apoB within the intervention arm, dose, design, follow-up and study quality. Meta-regression was performed to assess the significance of subgroup effects with STATA software, version 13 (StataCorp LP), with a significance level set at P<0·05.
Publication bias was investigated by visual inspection of funnel plots and quantitatively assessed using Egger’s and Begg’s tests, where P<0·05 was considered evidence for small study effects.
Funnel plots were used to display the relative treatment effect and its 95 % CI for each trial and dose amount and for the overall random-effects meta-analyses.
Results
Search results
The search strategy initially yielded 8190 publications, of which 269 were reviewed in full and fifty-eight (n 3974) were included in the final meta-analysis (Fig. 1). In total, fifty-six trials reported data on LDL-cholesterol (n 3745) and seventeen on apoB (n 1070). Only one trial reported data on non-HDL-cholesterol; however, fifty-six other trials reported enough information to calculate non-HDL-cholesterol (n 3926).
Flow of literature. Summary of search and selection process.
Trial characteristics
The characteristics of the included trials are summarised in Table 1. Trials were conducted in both in-patient and out-patient settings with twenty-five in North American (nineteen in USA, five in Canada and one in Mexico), nineteen in Europe (six in Sweden, four in England, three in the Netherlands, two in France and one each in Denmark, Finland, Germany and Greece), eight in Australia and New Zealand, three in Asia (two in China and one in Thailand), one in South America (Venezuela) and one in the Middle East (Iran). All trials were randomised, with 66 % (thirty-eight trials) utilising a parallel design and 34 % (twenty trials) utilising a cross-over design. Participants were generally middle aged (median age=50·6 (range: 10–67) years) with an approximately equal number of men and women. Participants were slightly overweight (median BMI=26·8 (range: 22·8–32·2) kg/m2), despite only 4 four trials recruiting on the basis of overweight/obese. Two-thirds of the trials (thirty-nine trials) were conducted in hypercholesterolaemic individuals. The dose of oat β-glucan ranged from 0·9 to 10·3 g/d with a median dose of 3·5 g/d. Treatment duration ranged from 3 to 12 weeks with the median length being 6 weeks for trials reporting LDL-cholesterol and non-HDL-cholesterol and 5 weeks for trials reporting apoB.
Characteristics of included studies
MQS, Heyland Methodological Quality Score; M, male; F, female; C, cross-over; SB, single blind; AHA, American Heart Association; A-I, agency-industry; OP, outpatient; P, parallel; NB, not blinded; A, agency; IP, inpatient; NCEP, National Cholesterol Education Program; N/R, not reported; I, industry; DB, double blind; RTE, ready to eat; MW, molecular weight.
† Whole oats can be oatmeal, instant oats, oat flakes or whole oat flour.
‡ The number of participants listed for each trial is the number of participants that completed the trial, and therefore the number used in our analyses and the number used for the reported baseline data (age and BMI), unless otherwise indicated with ‘*’.
§ Dose of β-glucan.
|| Trials with an MQS≥8 were considered to be of higher quality.
¶ Agency funding is that from government, university or not-for-profit health agency sources.
Very few studies (nine trials, 16 %) were considered to be of higher quality (MQS≥8). Lack of or poor description of randomisation, patient selection, protocol analysis and absence of double-blinding contributed to lower scores (online Supplementary Table S3). The Cochrane Risk of Bias Tool (online Supplementary Fig. S1 and Table S4) showed that seventeen trials (29 %) had low risk of bias and forty-two trials (71 %) had unclear risk of bias for random sequence generation. A total of thirteen trials (22 %) had low risk of bias, and forty-six trials (78 %) were unclear for allocation concealment. Moreover, thirty trials (50 %) had high risk of bias, twenty-one trials (36 %) had low risk of bias and eight trials (14 %) had unclear performance bias (blinding of participants and personnel); five trials (8 %) has high risk of bias, forty-nine trials (84 %) had low risk of bias and five trials (8 %) had unclear risk of bias for attrition bias. The majority of trials (93 %) had low risk of bias for reporting bias, whereas the remainder of the trials (7 %) had unclear risk of bias for these items. Funding of trials included agency (26 %), agency-industry (16 %), industry (34 %) sources or were not reported (24 %).
Effect on LDL-cholesterol
The effect of oat β-glucan on LDL-cholesterol is shown in Fig. 2. Overall, a significant LDL-cholesterol reduction was observed with a median dose of 3·5 g/d for a median duration of 6 weeks (MD=−0·19 mmol/l; 95 % CI −0·23, −0·14; P<0·00001). However, substantial evidence of inter-study heterogeneity was present in the overall analysis (I 2=79 %; P<0·00001). Systematic removal of individual trials did not alter the results.
Forest plot of randomised-controlled trials investigating the effect of oat β-glucan on LDL-cholesterol. Pooled effect estimate (
) for LDL-cholesterol (mmol/l). Values are mean differences (MD) with 95 % CI, using the generic inverse-variance random effects models. Inter-study heterogeneity was quantified by I
2 at a significance of P<0·10. N, number of participants in each treatment group.
Categorical a priori subgroup analyses revealed that the LDL-cholesterol lowering effect of oat β-glucan was modified by both study design (between-group MD=0·09 mmol/l; 95 % CI 0·01, 0·17; P=0·03) – studies that utilised a cross-over design demonstrated an MD of −0·25 mmol/l (95 % CI −0·31, −0·18), whereas studies that utilised a parallel design showed an MD of −0·16 mmol/l (95 % CI −0·20, −0·11) – and study duration (between-group MD=0·09 mmol/l; 95 % CI 0·02, 0·17; P=0·03) – studies where oat β-glucan was administered for <6 weeks demonstrated an MD of −0·24 mmol/l (95 % CI −0·29, −0·18), whereas studies that administered oat β-glucan for 6 weeks or more showed an MD of −0·15 mmol/l (95 % CI −0·20, −0·09), (online Supplementary Fig. S2). Continuous meta-regression analyses demonstrated an inverse association between baseline LDL-cholesterol and treatment differences for LDL-cholesterol (β=−0·09 mmol/l; 95 % CI −0·15, −0·03; P=0·004) (online Supplementary Table S5). Heterogeneity remained significant, and could not be explained by subgroup analyses.
Effect on non-HDL-cholesterol
The effect of oat β-glucan on non-HDL-cholesterol is shown in Fig. 3. Overall, non-HDL-cholesterol was significantly reduced by −0·20 mmol/l (95 % CI −0·26, −0·15), P<0·00001, with a median dose of 3·5 g/d for a median duration of 6 weeks. Considerable evidence of inter-study heterogeneity was present in the overall analysis (I 2=99 %; P<0·00001). Systematic removal of individual trials did not alter the results.
Forest plot of randomised-controlled trials investigating the effect of oat β-glucan on non-HDL-cholesterol. Pooled effect estimate (
) for non-HDL-cholesterol (mmol/l). Values are mean differences (MD) with 95 % CI, using the generic inverse-variance random effects models. Inter-study heterogeneity was quantified by I
2 at a significance of P<0·10. N, number of participants in each treatment group.
Categorical a priori subgroup analyses revealed that the non-HDL-cholesterol lowering was not modified by dose, study duration, study design, MQS scores or baseline non-HDL-cholesterol levels (online Supplementary Fig. S3). Furthermore, continuous meta-regression analyses did not reveal associations between dose, treatment duration or baseline non-HDL-cholesterol levels (online Supplementary Table S5).
Effect on apoB
The effect of oat β-glucan on apoB is shown in Fig. 4. Overall, there was evidence of a significant lowering of apoB with a median dose of 3·5 g/d for a median duration of 5 weeks (MD=−0·03 g/l; 95 % CI −0·05, −0·02; P<0·0001) with moderate evidence of heterogeneity (I 2=38 %; P=0·06). Systematic removal of individual trials did not alter the results.
Forest plot of randomised-controlled trials investigating the effect of oat β-glucan on apoB. Pooled effect estimate (
) for apoB (g/l). Values are mean differences (MD) with 95 % CI, using the generic inverse-variance random effects models. Inter-study heterogeneity was quantified by I
2 at a significance of P<0·10. N, Number of participants in each treatment group.
Categorical a priori subgroup analyses revealed that the apoB lowering by oat β-glucan was not modified by dose, study duration, study design, MQS scores or baseline apoB levels (online Supplementary Fig. S4). Furthermore, continuous meta-regression analyses did not reveal associations between dose, treatment duration or baseline apoB levels (online Supplementary Table S5).
Publication bias
Funnel plots for LDL-cholesterol, non-HDL-cholesterol and apoB are shown in Fig. 5. Visual inspection of funnel plots suggested minor asymmetry in the LDL-cholesterol and non-HDL-cholesterol analyses, with tendencies for the publication of small and/or imprecise trials favouring oat β-glucan for both. This was confirmed by Begg’s tests (P=0·061) for LDL-cholesterol; however, neither Egger’s (P=0·381) nor Begg’s (P=0·528) test was significant for non-HDL-cholesterol.
Publication bias funnel plots. Funnel plots assessing publication bias and effect of small and/or imprecise study effects for (a) LDL-cholesterol, (b) non-HDL-cholesterol and (c) apoB. 
, the pooled effect estimate expressed as the mean difference for each analysis; 
, pseudo-95 % CI. P-values are derived from quantitative assessment of publication bias by Egger’s and Begg’s tests.
Discussion
The present systematic review and meta-analysis of fifty-eight trials involving 3974 participants assessed the effects of oat β-glucan on clinical lipid targets for CVD risk reduction (LDL-cholesterol, non-HDL-cholesterol and apoB). Diets enriched with a median dose of 3·5 g/d of β-glucan were found to modestly improve LDL-cholesterol (−4·2 %), non-HDL-cholesterol (−4·8 %) and apoB (−2·3 %), compared with control diets.
Brown et al.( Reference Brown, Rosner and Willett 79 ) were the first to undertake a comprehensive meta-analysis of all viscous, soluble fibre types on cholesterol. Although the main objective was to study the cholesterol-lowering effect of all viscous, soluble fibre types, it was, nevertheless, the first to consolidate data on oats and LDL-cholesterol levels. In total, twenty-five studies investigating the cholesterol-lowering effect of oats were included in a subgroup analysis, and the authors reported a significant overall LDL-cholesterol reduction of −0·037 mmol/l (95 % CI−0·047, −0·017) per g of oat fibre. This is approximately equivalent to −0·13 mmol/l per 3·5 g, 30 % less than what was observed in our current study (Fig. 2). However, as the results from this meta-analysis were reported as mmol/l of LDL-cholesterol reduction per gram of soluble fibre, they cannot be directly compared with the results of the current study.
In the most recent meta-analysis of oat β-glucan and LDL-cholesterol( Reference Whitehead, Beck and Tosh 18 ), the authors included twenty-eight RCT and reported an LDL-cholesterol reduction of −0·25 mmol/l (−6 %), whereas this study demonstrated a reduction of −0·19 mmol/l (−4·2 %). This discrepancy could be due to differences in study selection criteria. Whitehead et al. only included RCT that administered ≥3 g/d of oat β-glucan, which resulted in a median daily dose of 5·1 g, whereas the current meta-analysis included studies of all doses and observed a median dose of 3·5 g/d. When the results were examined on a per gram basis, LDL-cholesterol reductions were on par (Whitehead et al.: −0·050 mmol/l v., our study: −0·054 mmol/l per g of oat β-glucan) despite the differences in dose. Interestingly, our meta-regression analysis indicated a significant inverse association between dose and LDL-cholesterol levels (online Supplementary Table S4). Furthermore, when dose was categorised according to Health Canada and US FDA recommendations (<3·0 v. ≥3·0 g/d), there was a trend towards treatment modification by dose (P=0·051), such that LDL-cholesterol reduction was almost double in trials that administered ≥3·0 g/d of oat β-glucan compared with those that administered <3·0 g/d (online Supplementary Fig. S2). These results further support the health claims set by Health Canada and US FDA that cholesterol lowering can be achieved with a minimum of 3 g/d of oat β-glucan.
This is the first meta-analysis of RCT yielding information on the effect of oat β-glucan on non-HDL-cholesterol and apoB. These markers have been added to clinical practice guidelines( 8 , Reference Anderson, Gregoire and Hegele 9 ) on the basis that they are more highly associated with CVD risk than LDL-cholesterol( Reference Saenger 7 ). Furthermore, the appreciation of these markers for CVD risk is especially important in adults with the metabolic syndrome and/or diabetes as LDL-cholesterol is not typically elevated in this population. Pooled analyses demonstrated significant reductions of non-HDL-cholesterol (−0·20 mmol/l (95 % CI −0·26, −0·15)) and apoB (−0·03 g/l (95 % CI −0·05, −0·02)); however, the results are compromised by considerable unexplained heterogeneity. Interestingly, when trials were classified into the hypercholesterolaemic or unclassified group, of which more than a quarter of the studies were conducted in type 2 diabetes mellitus, both categories demonstrated significant reductions in non-HDL-cholesterol and apoB. This is an important finding, considering that type 2 diabetes mellitus is generally not associated with increased LDL-cholesterol. Therefore, focusing on interventions that reduce non-HDL-cholesterol and apoB may be more practical and reliable for addressing the increased risk of CVD in type 2 diabetes mellitus.
Effect modification by baseline cholesterol levels has been previously described, such that cholesterol lowering by β-glucan is generally greater in those with hypercholesterolaemia( Reference Anderson, Gregoire and Hegele 9 ). This was confirmed by our meta-regression analysis demonstrating a significant inverse association between baseline LDL-cholesterol levels and the extent of LDL-cholesterol reduction (online Supplementary Table S4). However, higher baseline levels of non-HDL-cholesterol or apoB were not significantly associated with greater reductions.
There are several limitations to the present meta-analysis that complicate the interpretation of the results. The first one being that the β-glucan content of oats was estimated for the majority of trials as it was not routinely analysed and reported. As β-glucan content varies significantly depending on genetics and environmental growing conditions( Reference Bartlomiej, Justyna and Ewa 80 , Reference Lee, Horsley and Manthey 81 ), it is difficult to precisely measure the treatment effect when the majority of trials did not conduct a chemical analysis of the β-glucan content of their study products.
Second, the considerable heterogeneity that was observed in LDL-cholesterol and non-HDL-cholesterol was not explained by any of the a priori subgroup analyses. Nevertheless, considering the large number of studies included in this meta-analysis, high heterogeneity is inevitable. The studies included a wide range of food matrices that were used to administer oat β-glucan, several different processing and storage methods, varying molecular levels of β-glucan, etc., all of which are interrelated and significantly impact viscosity of the β-glucan, and thus its cholesterol-lowering potency. Furthermore, nutrition studies have not yet incorporated non-HDL-cholesterol into their primary analysis, despite the simple calculation. Therefore, in addition to all the previously mentioned sources of heterogeneity, the entire set of non-HDL-cholesterol data was mathematically imputed, which may have contributed to the increased heterogeneity.
Irrespective of the large heterogeneity associated with including studies that were conducted in a wide range of participants, in numerous countries, and used various common food products to administer the oat β-glucan, the results can be considered largely generalisable and indicative that the cholesterol-lowering benefits can be achieved by supplementing oat β-glucan into commonly consumed foods.
In conclusion, this systematic review and meta-analysis supports the dose-dependent intake of oat β-glucan for the reduction of LDL-cholesterol, non-HDL-cholesterol and apoB in middle-aged participants. Because of considerable unexplained heterogeneity, caution should be taken when interpreting the results. There is a need for larger, longer, high-quality RCT on the effect of oat β-glucan on blood cholesterol levels, especially non-HDL-cholesterol and apoB end points, and in participants with different metabolic phenotypes. Special attention should be paid to β-glucan molecular weight and content in these trials to allow for a more accurate assessment of the cholesterol-lowering properties of β-glucan.
Acknowledgements
The authors thank Teruko Kishibe of Li Ka Shing’s International Healthcare Education Centre at St. Michael Hospital for her help in the development of the search strategy.
The authors’ responsibilities were as follows: J. L. S., E. J., A. L. J. and V. V. designed the study; H. V. T. H. and A. Z. conducted the study; H. V. T. H. and S. B. M. analysed data or performed statistical analysis; H. V. T. H. wrote the paper; H. V. T. H. and V. V. had primary responsibility for the final content; all the authors contributed to the critical revision of the article for important intellectual content and approved the final manuscript.
H. V. T. H., A. Z., S. B. M., E. J., F. A.-Y. have no declared conflicts of interest related to this manuscript. J. L. S. has received research support from the Canadian Institutes of Health Research, Canadian Diabetes Association (CDA), PSI Foundation, Calorie Control Council, American Society of Nutrition (ASN), The Coca-Cola Company (investigator initiated, unrestricted), Dr Pepper Snapple Group (investigator initiated, unrestricted), Pulse Canada, The International Tree Nut Council Nutrition Research & Education Foundation and the INC International Nut and Dried Fruit Council. He has received reimbursement of travel expenses, speaker fees and/or honoraria from the American Heart Association, American College of Physicians, ASN, National Institute of Diabetes and Digestive and Kidney Diseases, CDA, Canadian Nutrition Society, University of South Carolina, University of Alabama at Birmingham, Oldways Preservation Trust, Nutrition Foundation of Italy, Calorie Control Council, Diabetes and Nutrition Study Group (DNSG) of the European Association for the Study of Diabetes (EASD), International Life Sciences Institute (ILSI) North America, ILSI Brazil, Abbott Laboratories, Pulse Canada, Canadian Sugar Institute, Dr Pepper Snapple Group, The Coca-Cola Company, Corn Refiners Association, World Sugar Research Organization, Dairy Farmers of Canada, Società Italiana di Nutrizione Umana, III World Congress of Public Health Nutrition, C3 Collaborating for Health, White Wave Foods, Rippe Lifestyle, mdBriefcase. He has ad hoc consulting arrangements with Winston & Strawn LLP, Perkins Coie LLP and Tate & Lyle. He is on the Clinical Practice Guidelines Expert Committee for Nutrition Therapy of both the CDA EASD, and CCS, as well as being on an ASN writing panel for a scientific statement on sugars. He is a member of the International Carbohydrate Quality Consortium and Board Member of the DNSG of the EASD. He serves as an unpaid scientific advisor for the Food, Nutrition, and Safety Program and the Technical Committee on Carbohydrates of the ILSI North America. His wife is an employee of Unilever Canada. V. V. holds a research grant from CDA for study of dietary intervention including viscous, soluble fibre and holds the Canadian (2,410,556) and American (7,326.404) patent on medical use of viscous fibre blend for reducing blood glucose for treatment of diabetes, increasing insulin sensitivity, reduction of systolic blood pressure and blood lipids. A. L. J. is director of research and part owner of Glycemic Index Laboratories, a clinical research organisation.
Supplementary Material
For supplementary material/s referred to in this article, please visit http://dx.doi.org/10.1017/S000711451600341X
