Hostname: page-component-586b7cd67f-l7hp2 Total loading time: 0 Render date: 2024-11-22T15:59:34.917Z Has data issue: false hasContentIssue false

Use of biomarkers to assess fruit and vegetable intake

Published online by Cambridge University Press:  28 March 2017

Jayne V. Woodside*
Affiliation:
Centre for Public Health, Queen's University Belfast, Belfast, UK CRC Centre of Excellence for Public Health Northern Ireland, Belfast, UK
John Draper
Affiliation:
Institute of Biological Environmental and Rural Sciences, Aberystwyth University, Aberystwyth, Ceredigion, UK
Amanda Lloyd
Affiliation:
Institute of Biological Environmental and Rural Sciences, Aberystwyth University, Aberystwyth, Ceredigion, UK
Michelle C. McKinley
Affiliation:
Centre for Public Health, Queen's University Belfast, Belfast, UK CRC Centre of Excellence for Public Health Northern Ireland, Belfast, UK
*
*Corresponding author: J. V. Woodside, fax 0044 2890 235900, email [email protected]
Rights & Permissions [Opens in a new window]

Abstract

A high intake of fruit and vegetables (FV) has been associated with reduced risk of a number of chronic diseases, including CVD. The aim of this review is to describe the potential use of biomarkers to assess FV intake. Traditional methods of assessing FV intake have limitations, and this is likely to impact on observed associations with disease outcomes and markers of disease risk. Nutritional biomarkers may offer a more objective and reliable method of assessing dietary FV intake. Some single blood biomarkers, such as plasma vitamin C and serum carotenoids, are well established as indicators of FV intake. Combining potential biomarkers of intake may more accurately predict overall FV intake within intervention studies than the use of any single biomarker. Another promising approach is metabolomic analysis of biological fluids using untargeted approaches to identify potential new biomarkers of FV intake. Using biomarkers to measure FV intake may improve the accuracy of dietary assessment.

Type
Conference on ‘New technology in nutrition research and practice’
Copyright
Copyright © The Authors 2017 

Fruit and vegetables and health

Diets rich in fruit and vegetables (FV) have been linked with a reduced risk of chronic disease( Reference Boeing, Bechthold and Bub 1 , Reference Wang, Ouyang and Liu 2 ). The evidence is particularly strong for CVD( Reference Boeing, Bechthold and Bub 1 Reference Crowe, Roddam and Key 4 ), is weaker for both diabetes and cancer( Reference Wang, Ouyang and Liu 2 , Reference Cooper, Sharp and Lentjes 5 10 ), and is relatively consistent for specific cancer sites( 10 ). FV are micronutrient and fibre-rich and therefore are recommended across all dietary guidelines( 11 14 ).

Although the evidence linking increased FV intake with a reduced risk of CVD is consistent and relatively strong, it is largely based on observational studies( Reference Boeing, Bechthold and Bub 1 Reference Woodside, Young and McKinley 3 ), with few randomised controlled trials with clinically-relevant endpoints( Reference Woodside, Young and McKinley 3 ). This observational evidence relies on traditional dietary assessment of FV intake, with the majority of studies using a FFQ( Reference Wang, Ouyang and Liu 2 , Reference Woodside, Young and McKinley 3 , Reference Boffetta, Couto and Wichmann 9 ).

Assessment of fruit and vegetable intake

Accurate estimation of dietary intake can be challenging, and traditional methods have been shown to be prone to both random and systematic errors. In terms of the specific problems associated with measuring FV intake through traditional methods, FV have been shown to be particularly prone to overreporting, as the participants know that they are known to be health-promoting foods, and therefore tend to exaggerate usual intake( Reference Jenab, Slimani and Bictash 15 , Reference Woodside, Young and McKinley 16 ). A second reason, which may impact on the accuracy of reporting is that, to report consumption of a particular number of portions daily requires a knowledge of what constitutes a portion of a range of FV, and such knowledge of what constitutes a portion has been recently shown to be lacking amongst a population of low FV consumers( Reference Rooney, McKinley and Appleton 17 ).

Accurate dietary assessment is extremely important for confirming the associations between overall FV intake and chronic disease risk and to inform quantitative dietary guidelines. Indeed, the optimum level of FV intake for health protection is still a topic of debate( Reference Woodside, Rooney and McKinley 18 , Reference Kypridemos, O'Flaherty and Capewell 19 ). Accurate dietary assessment is also needed to help elucidate if different types of FV have different health-promoting properties. For example, while the evidence for an association between increased overall FV intake and diabetes risk is equivocal, specific FV might be associated with risk; for example, leafy green vegetables and diabetes risk( Reference Carter, Gray and Troughton 6 Reference Li, Fan and Zhang 8 ). There is also some debate over whether fruit juice has less benefit to health than other forms of fruit( Reference Muraki, Imamura and Manson 20 ). Furthermore, the effect of particular cooking and processing methods on micronutrient content, and micronutrient bioavailability and the resulting effects on health are still uncertain( Reference Oude Griep and Geleijnse 21 ). Finally, the concept of the need to consume a variety of FV, and the association between FV variety and health has been a focus of recent interest( Reference Bhupathiraju and Tucker 22 Reference Cooper, Sharp and Lentjes 24 ), but, again, to determine the true value of variety does rely on accurate dietary assessment methods.

The importance of dietary assessment method when determining the association between FV intake and disease risk is exemplified by the work of Bingham et al. ( Reference Bingham, Luben and Welch 25 ), who examined the association between FV intake and IHD risk, in a cross-sectional analysis of the EPIC Norfolk cohort study. Whilst there were strong associations between vitamin C intake assessed by food diary and plasma vitamin C status, coefficients were attenuated for vitamin C intake assessed by FFQ. Similarly, when examining risk of IHD, plasma vitamin C and FV intake assessed by food diary were associated with risk of IHD, but not FV intake assessed by FFQ( Reference Bingham, Luben and Welch 25 ). Therefore the choice of dietary assessment method can affect the observed association with disease risk, and selection of an appropriate method is vital. The fact that a food diary and plasma vitamin C reflect recent intake, whilst an FFQ will typically reflect intake over the previous year, is likely to have a bearing on diet–disease associations in observational studies and highlights the need to consider the timescale of the various intake or status assessment methods( Reference Willett 26 ).

Thus, there is a need to explore and develop new methods of accurately estimating FV in order to better capture intake and be able to answer important research questions, such as those mentioned earlier, by allowing better evaluation of the association between intake and disease risk, and the measurement of compliance in intervention studies.

Biomarkers of fruit and vegetable intake

As outlined earlier, traditional methods of assessing FV consumption have significant limitations, and an alternative, more objective way of estimating FV intake may be to measure the levels of compounds found in FV in biological samples, such as plasma, serum and urine. The use of biomarker methods in nutritional epidemiology in general has developed greatly in the past 20 years, with Bingham stating in 2002( Reference Bingham 27 ) that ‘The collection of biological samples to improve and validate estimates of exposure, enhance the pursuit of scientific hypotheses, and enable gene–nutrient interactions to be studied, should become the routine in nutritional epidemiology.’ However, there are knowledge gaps, and in 2007 the Institute of Medicine recognised the lack of nutritional biomarkers, and confirmed a need for both biomarkers that can predict functional outcomes and chronic diseases, and those that can improve dietary assessment, but which are non-invasive, inexpensive and specific( 28 ). Hedrick et al.( Reference Hedrick, Dietrich and Estabrooks 29 ) reacted to this recommendation, suggesting a need to emphasise the development of biomarkers for evaluating adherence to national recommendations for specific food groups, e.g. wholegrains, fruit and vegetables.

Biomarkers are constituents in the blood, urine or saliva that can be used to indicate dietary exposure and compare this to intake estimated by dietary assessment. Depending on the food group and particular marker used, biomarkers can be classified into three main classes: recovery biomarkers are based on the total excretion of the marker over a specific time period and can estimate absolute intake, but only a few of these recovery biomarkers exist in nutrition, e.g. urinary potassium and urinary nitrogen( Reference Kuhnle 30 ). A further class of markers are predictive markers: these have incomplete recovery, but have a stable, time-dependent and strong association with intake, the main example being urinary sucrose and fructose as a marker of sugar intake( Reference Tasevska, Runswick and McTaggart 31 ). Concentration markers cannot estimate absolute intake, but are correlated with intake and therefore can rank intake of specific nutrients( Reference Kuhnle 30 ), while replacement biomarkers are closely related to concentration biomarkers, but are specifically where information from food databases is unsatisfactory or unavailable( Reference Jenab, Slimani and Bictash 15 ). A number of potential biomarkers of FV intake have been suggested, which are compounds found within FV, including a range of serum carotenoids (lutein, zeaxanthin, β-cryptoxanthin, α- and β-carotene and lycopene), and plasma vitamin C, and also urinary potassium, flavonoids in both urine and serum, and glucosinolates. All of these biomarkers of FV intake would be classified as concentration markers; therefore they will not reflect the exact dietary intake, but are likely to be highly correlated with intake.

Vitamin C and carotenoids are the most commonly used biomarkers, but the complexity of the FV food group makes these compounds potentially less useful as biomarkers of the overall food group, because of the variability of content within different fruit and vegetables( Reference Kuhnle 30 ). For example, the amount of vitamin C found within one portion of green pepper is equivalent to that found in about twenty portions of carrots and, conversely, the amount of total carotene found in one portions of carrots is equivalent to that found in more than forty-five portions of green pepper( Reference Kuhnle 30 ). Kuhnle concluded that, given this variability, it is important to use a combination of biomarkers or to develop new biomarkers; for example, total phenols have been suggested as a potential biomarker which, unlike vitamin C and carotenoids, has much lower variation across different types of FV( Reference Kuhnle 30 ). However, the use of total phenols as a biomarker of FV intake, while plausible based on food analysis, has yet to be explored in detail in human studies( Reference Kuhnle 30 , Reference Medina-Remón, Barrionuevo-González and Zamora-Ros 32 ).

Two separate systematic reviews (SR) have examined the use of FV biomarkers used in human intervention studies. The first, published in 2011 by Baldrick et al.( Reference Baldrick, Woodside and Elborn 33 ), aimed to examine the utility of the main biomarkers of FV intake to act as objective indicators of compliance in dietary intervention studies. Therefore, this review was particularly focused on identifying compliance markers for intervention studies and reviewed usual practice in this area. The search identified a total of ninety-five studies as suitable for inclusion according to pre-defined criteria and classified the interventions as being whole-diet interventions, individual FV intervention studies or mixed FV studies. Data were extracted and summarised for each study type. This review concluded that it was rarely possible to rely on assessment of a single biomarker as an indicator of dietary change in human intervention studies, but that single biomarkers could be good predictors of single classes of FV, e.g quercetin has been demonstrated to be a reasonable indicator of onion consumption. Similarly, for ‘fruit only’ intervention studies, assessment of vitamin C alone may suffice. However, the authors concluded that given the complexity of FV, and the large number of bioactive compounds they contain, a panel of biomarkers should be measured in FV trials, and this was likely to include a panel of carotenoids and vitamin C, but that further research should continue to explore more novel biomarker approaches( Reference Baldrick, Woodside and Elborn 33 ).

A more recent SR, in contrast to the more qualitative review of Baldrick et al.( Reference Baldrick, Woodside and Elborn 33 ), examined plasma vitamin C and serum carotenoids as indicators of FV intake, conducting both a SR and meta-analysis of randomised controlled trials and examining their comparative validity( Reference Pennant, Steur and Moore 34 ). Nineteen FV interventions, with 1382 participants in total, measures at least one biomarker, and nine trials, with n 667 participants, measured the five main carotenoids (lutein, β-cryptoxanthin, α-, β-carotene and lycopene) and vitamin C. Vitamin C and carotenoids (except lycopene) were responsive to general changes in FV intake at the group level, but there was no clear evidence of dose–response, so that those groups consuming higher number of portions of FV did not have more marked increases in these biomarkers. There was also no convincing evidence that any single biomarker was more responsive than others, with all CI overlapping, whilst there was high heterogeneity in responses, suggesting a lack of consistency in the size of response between studies. Owing to the high heterogeneity and lack of dose–response, the authors concluded that individual-level biomarker responses would be highly variable and could not be relied on( Reference Pennant, Steur and Moore 34 ). Moreover, the randomised controlled trials included in the SR were of low quality, as assessed using the GRADE (grades of recommendation, assessment, development and evaluation) system. This is not unexpected, as blinding is not possible in these whole-food studies, while many of the trials included were not originally designed to develop biomarkers and therefore included participants consuming nutritional supplements and those who smoked, or did not collect samples in the fasting state. Few trials stated whether there was allocation concealment, and the level of dietary control or monitoring of adherence was low, leading to uncertainty about actual FV intake, which is crucial for biomarker response. As with the previous SR, the authors concluded that further work is required to understand the determinants of biomarker variation among individuals( Reference Pennant, Steur and Moore 34 ).

Novel biomarker approaches

Given the challenges of the complexity of the FV food group, a number of novel biomarker approaches have been suggested. It is possible to consider the assessment of a range of biomarkers and statistically combining them to better predict overall FV intake. One approach to this is simply to sum individual biomarkers, e.g. carotenoids, to give a total carotenoid figure( Reference Woodside, Young and Gilchrist 35 ), but this leads to the total being dominated by the carotenoids present at the highest concentrations, e.g. lycopene. To overcome this potential issue, Cooper et al.( Reference Cooper, Sharp and Luben 36 ) have recently summed the biomarkers identified within a previous SR as most likely to respond to increased FV intake( Reference Baldrick, Woodside and Elborn 33 ), and calculated the sum of standardised variables of vitamin C, β-carotene and lutein, examining resulting associations with type 2 diabetes risk in the EPIC-Norfolk study( Reference Cooper, Sharp and Luben 36 ).

McGrath et al.( Reference McGrath, Hamill and Cardwell 37 ) have examined the effect of increased FV intake on biomarkers of FV consumption, both singly and in combination, but using data from dietary intervention studies and applying more complex statistics to combine the biomarkers. They conducted the BIOFAV study, a tightly controlled FV dietary intervention (all food provided and two meals daily on weekdays consumed under supervision) in low FV consumers. A total of thirty participants, who usually consumed fewer than two portions of FV daily, were randomised to either two, five or eight portions of FV daily for 4 weeks. Blood and urine samples were collected at baseline and 4 weeks, and plasma vitamin C and serum carotenoid analysis conducted. A combined model containing all carotenoids and vitamin C, when predicting allocated FV group, was a better fit than a model containing vitamin C only (P < 0·001) or lutein only (P = 0·006). The C-statistic was lower in the lutein only model (0·85) and the vitamin C model (0·68) than the full model (0·95)( Reference McGrath, Hamill and Cardwell 37 ).

The authors then applied this approach to three other previously conducted FV interventions. They observed a similar pattern of results, but the differences between the combined biomarker and individual biomarker models were less marked, perhaps due to the lower levels of dietary control in these other studies( Reference McGrath, Hamill and Cardwell 37 ). This approach needs to be replicated, and the effect of adding additional potential biomarkers, e.g. urinary flavonoid excretion, to the models to potentially increase the predictive capacity of the model needs to be explored. The utility of such an approach in observational studies also needs to be tested. An issue is that examining the potential of a combined biomarker panel in observational studies will require a ‘true’ measure of FV intake to compare the biomarker against, and most observational studies will have used FFQ-based data collection, which may not be accurate enough to reflect intake comparable with the timescale of the biomarker, i.e. reflect recent intake.

Other studies have also explored the combined biomarkers approach, and have similarly demonstrated an indication of its utility, although each study has used different biomarkers and approached the ‘combining’ in a different way. Analysis of the FLAVURS study, a study testing sequential increases of 2·3, 3·2 and 4·2 portions of FV every 6 weeks across 18 weeks in n 154 male and female participants at increased risk of CVD, suggested that an integrated plasma biomarker (including vitamin C, total cholesterol-adjusted carotenoids and FRAP (ferric reducing ability of plasma) values) was better correlated with FV intake (r = 0·47, P < 0·001) than individual biomarkers( Reference Jin, Gordon and Alimbetov 38 ). Inclusion of urinary potassium into the integrated biomarker panel did not further improve the correlation. This integrated plasma biomarker could therefore, the authors suggest, be used to distinguish between high and moderate FV consumers. No further indicators of model performance were included, which makes further comparisons with other studies difficult.

In another study, a prediction model was developed from twelve FV intervention studies( Reference Souverein, de Vries and Freese 39 ). The prediction model was developed based on a total of 526 male and female participants and was conducted as an individual participant data meta-analysis examining FV intake both including and then excluding FV juices. What was also important was that adjustments were included for important potential characteristics, such as age, BMI and smoking, that may have affected biomarker response, and this is the only study combining biomarkers to have explored the effect of such adjustment to date. Measures of performance for the prediction model were calculated using cross-validation. The final prediction model included carotenoids, folate and vitamin C, and these were positively correlated with FV intake( Reference Souverein, de Vries and Freese 39 ). For the prediction model of fruit, vegetable and juice intake, a reduced model which included only statistically significant predictors, selected using multivariable fractional polynomials performed best. For this model, a number of measures of performance were presented: the root-mean-squared error (258·0 g, the correlation between observed and predicted intake (0·78) and the mean difference between observed and predicted intake (−1·7 g limits of agreement: −466·3, 462·8 g). For the prediction of FV intake (excluding juices), the root-mean-squared error was 201·1 g, the correlation was 0·65 and the mean bias was 2·4 g (limits of agreement: −368·2, 373·0 g). The authors concluded that these models could be used to predict ranking of FV intake when validating questionnaires or to estimate FV intake at the group level. However, low levels of agreement meant that the prediction model should not be used to estimate individual intake( Reference Souverein, de Vries and Freese 39 ).

Therefore combining already known biomarkers of FV intake may be useful in improving the use of biomarkers to accurately estimate FV intake, but only a limited number of studies have, to date, examined this approach.

Metabolomics is an emerging analytical technique that identifies and quantifies small metabolites( Reference Gibney, Walsh and Brennan 40 , Reference O'Gorman, Gibbons and Brennan 41 ). Traditional biomarker approaches have assessed mainly the concentration in biofluids of phytochemicals measured previously in uncooked FV. In contrast, metabolomics has been used to identify biotransformation products (e.g., glucuronide and sulphate conjugates or colon microbiota fermentation products) of diet-derived chemicals that are both stable, more abundant and easily quantified by the standardised methods( Reference Lloyd, Favé and Beckmann 42 , Reference Scalbert, Brennan and Manach 43 ). The ability to comprehensively analyse metabolites in biological fluids to look for novel dietary exposure biomarkers in an untargeted way is likely to enhance the ability of researchers to characterise dietary exposure, with many potential applications in nutritional epidemiology. Challenges, however, exist, both in terms of the technology required to identify unknown metabolites and to deal with the large amounts of data produced during this type of analysis. Although a number of studies have examined specific FV classes and used metabolomics to identify potential novel biomarkers, e.g. proline betaine as a biomarker of citrus intake( Reference Heinzmann, Brown and Chan 44 , Reference Lloyd, Beckmann and Favé 45 ), and S-methyl-l-cysteine sulphoxide and metabolic derivatives as biomarkers of cruciferous vegetable intake( Reference Edmands, Beckonert and Stella 46 ), the use of metabolomics to assess overall FV intakes is, as yet, uncertain.

Another approach that has been suggested is the optical detection of carotenoids in the skin using a range of methods, including resonance Raman spectroscopy, reflection spectroscopy and pressure-mediated reflectance spectroscopy( Reference Whigham and Redelfs 47 ). Such a method would be non-invasive, simple and relatively inexpensive and would provide estimates on the spot without the need for collection of biological samples, which are then analysed in a laboratory. Whether such a technique is sensitive enough to pick up changes in FV intake within normal diet ranges remains to be established. However, a recent study has demonstrated a statistically significant association between carotenoid intake and skin carotenoids in 9–12-year-old children; hence, the authors suggest the potential for such a non-invasive method to measure FV intake in this population( Reference Nguyen, Scherr and Linnell 48 ).

Furthermore, the use of multiple dietary assessment methods and/or biomarker approaches in combination may strengthen the investigation of diet–disease relationships and increase statistical power( Reference Freedman, Kipnis and Schatzkin 49 , Reference Freedman, Midthune and Carroll 50 ). The approach has then been used in relation to the carotenoids lutein and zeaxanthin, the carotenoids, which are potential biomarkers of FV intake, and are of particular interest in eye disease as they are the only components of the macular pigment( Reference Freedman, Tasevska and Kipnis 51 ). In their study, Freedman et al.( Reference Freedman, Tasevska and Kipnis 51 ) explored the difference in statistical power produced when examining either: (i) self-reported dietary intake of lutein and zeaxanthin from a FFQ; (ii) serum lutein and zeaxanthin concentration; or (iii) a combined method summing the ranking of participants from (i) and (ii). The combined measure, when examining the association between lutein and zeaxanthin and risk of nuclear cataracts, provided higher statistical significance that the dietary measure or serum measure alone. The authors suggest a saving of 8–53 % over analysis with dietary intake alone and 6–48 % for the serum level alone in terms of required sample size( Reference Freedman, Tasevska and Kipnis 51 ). Such an increase in power or reduction in required sample size is sizeable and indicates the potential utility of this approach.

Considerations when using biomarkers of fruit and vegetable intake

There are a number of important considerations when using biomarker approaches, and these will be common to all biomarkers. Consideration of the chronology of exposure is important for both traditional dietary assessment and biomarkers, with the likely time frame covered by different dietary assessment methods and biomarkers being considered when comparing methods (Fig. 1). There are a number of further factors, which will affect the ability of biomarkers to predict intake. These have been summarised by Jenab et al.( Reference Jenab, Slimani and Bictash 15 ) for dietary assessment and biomarkers in general (adapted in Table 1), and will include a range of pre-analytical factors, which need to be considered( Reference Lippi, Guidi and Mattiuzzi 52 , Reference Blanck, Bowman and Cooper 53 ).

Fig. 1. (Colour online) Timescale of nutritional biomarkers from different biological sources (adapted from Kuhnle( Reference Kuhnle 30 )).

Table 1. Factors affecting nutritional biomarker response (adapted from Jenab et al.( Reference Jenab, Slimani and Bictash 15 )), with specific examples added for proposed fruit and vegetable (FV) biomarkers

Specifically, vitamin C is a particularly labile vitamin and therefore sample collection and stabilisation has to be conducted carefully, according to protocols, which involve the precipitation of proteins, usually with metaphosphoric or TCA( Reference Vuilleumier and Keck 54 , Reference Salminen and Alfthan 55 ). Such stabilisation is not commonly carried out within large-scale epidemiological studies. Similarly, carotenoids can be light-sensitive, and therefore exposure to light during processing and storage should be minimised( Reference Craft, Wise and Soares 56 ).

Genetic differences in biomarker responses have been observed, although to date these have only been analysed within observational studies( Reference Timpson, Forouhi and Brion 57 , Reference Borel 58 ). For example, Timpson et al.( Reference Timpson, Forouhi and Brion 57 ) examined variation at the SLC23A1 locus in five independent population studies and found that each additional rare allele was associated with a reduction in circulating ascorbic acid concentrations (−5·98 (95 % CI −8·23, −3·73) μm/l, P = 2·0 × 10(−7) per minor allele). Similarly, carotenoid status has been suggested to depend on range of genotypes, including phase two enzyme glutathione S-transferase M1 and T1 polymorphisms, and this has been reviewed( Reference Borel 58 ). The effect of such polymorphisms on biomarker responses within FV intervention studies is not known, but to test this will require careful study design consideration and likely increase in required sample size.

Differences in biomarker responses have been observed based on baseline concentration( Reference Jenab, Slimani and Bictash 15 ), inflammation( Reference Tomkins 59 ), status of other nutrients, including other carotenoids( Reference Reboul, Thap and Tourniaire 60 ), BMI( Reference Vioque, Weinbrenner and Asensio 61 ) and smoking( Reference Alberg 62 ). For example, plasma carotenoids and vitamin C were less strongly associated with dietary intake in obese older subjects than in those of normal weight( Reference Vioque, Weinbrenner and Asensio 61 ). Furthermore, plasma vitamin C tends to plateau at higher levels of intake (>120 mg/d), and therefore may not accurately reflect higher exposure( Reference Padayatty and Levine 63 ). A recent study examining carotenoids as biomarkers of FV intake in men and women, and using data from FV interventions, suggested that plasma β-cryptoxanthin and lutein concentrations were reliable biomarkers of FV consumption, but that there were significant sex differences in biomarker response following FV consumption( Reference Couillard, Lemieux and Vohl 64 ), suggesting that sex must be considered when monitoring biomarker responses. These factors are also considered in Table 1.

What has been less fully explored and which will be challenging, is whether biomarkers can ever be sensitive enough to pick up on differences in response by FV class, cultivar, production, processing and storage factors, which may impact on micronutrient content of the specific fruit or vegetable, and, affect health status. For example, cooking of FV leads to a reduction in vitamin C content( Reference Zeng 65 ), but the degree of loss will depend on the cooking procedure and length of cooking time. Miglio et al.( Reference Miglio, Chiavaro and Visconti 66 ) examined the effect of different cooking methods on phytochemical properties, total antioxidant capacity and physicochemical properties of carrots, courgettes and broccoli, and highlighted that the modifications by cooking are strongly dependent on the vegetable species. Similarly the consumption of fat alongside carotenoid-rich foods increases bioavailability of the carotenoids( Reference Priyadarshani 67 ). While it is perhaps unlikely that FV biomarkers will ever be sensitive enough to measure the impact of some of these factors, what is likely is that there will be an improvement of accuracy in terms of global FV assessment.

Conclusion

In conclusion, eating more FV is associated with better health status, but some uncertainties exist regarding the optimum number of portions, type, cooking and processing methods and effects on specific disease/health outcomes, particularly for different types of FV, and to what extent variety is important. Accurate assessment of dietary intake is, in general, difficult, and there are particular challenges for FV as it is a complex food group, with a range of bioactive compounds. Novel biomarker methods are a focus of interest and are potentially important in order to improve the accuracy of intake assessment and so advance research related to FV.

Financial support

None.

Conflicts of interest

None.

Authorship

J. V. W. drafted the manuscript and produced the final version after critical review by J. D., A. L. and M. C. M.

References

1. Boeing, H, Bechthold, A, Bub, A et al. (2012) Critical review: vegetables and fruit in the prevention of chronic diseases. Eur J Nutr 51, 637663.CrossRefGoogle ScholarPubMed
2. Wang, X, Ouyang, Y, Liu, J et al. (2014) Fruit and vegetable consumption and mortality from all causes, cardiovascular disease, and cancer: systematic review and dose-response meta-analysis of prospective cohort studies. BMJ 349, g4490.Google Scholar
3. Woodside, JV, Young, IS & McKinley, MC (2013) Fruit and vegetable intake and risk of cardiovascular disease. Proc Nutr Soc 72, 399406.Google Scholar
4. Crowe, FL, Roddam, AW, Key, TJ et al. (2011) Fruit and vegetable intake and mortality from ischaemic heart disease: results from the European Prospective Investigation into Cancer and Nutrition (EPIC)-Heart study. Eur Heart J 32, 12351243.Google Scholar
5. Cooper, AJ, Sharp, SJ, Lentjes, MA et al. (2012) A prospective study of the association between quantity and variety of fruit and vegetable intake and incident type 2 diabetes. Diab Care 35, 12931300.Google Scholar
6. Carter, P, Gray, LJ, Troughton, J et al. (2010) Fruit and vegetable intake and incidence of type 2 diabetes mellitus: systematic review and meta-analysis. BMJ 341, c4229.Google Scholar
7. Cooper, AJ, Forouhi, NG, Ye, Z et al. (2012) Fruit and vegetable intake and type 2 diabetes: EPIC-InterAct prospective study and meta-analysis. Eur J Clin Nutr 66, 10821092.Google Scholar
8. Li, M, Fan, Y, Zhang, X et al. (2014) Fruit and vegetable intake and risk of type 2 diabetes mellitus: meta-analysis of prospective cohort studies. BMJ Open 4, e005497.Google Scholar
9. Boffetta, P, Couto, E, Wichmann, J et al. (2010) Fruit and vegetable intake and overall cancer risk in the European Prospective Investigation into Cancer and Nutrition (EPIC). J Natl Cancer Inst 102, 529537.Google Scholar
10. World Cancer Research Fund (2016) http://www.wcrf.org/int/research-we-fund/continuous-update-project-cup (accessed October 2016).Google Scholar
11. Department of Health (2016) http://www.nhs.uk/LiveWell/5ADAY/Pages/5ADAYhome.aspx (accessed October 2016).Google Scholar
14. World Health Organisation (2004) Fruit and vegetables for health. http://www.who.int/dietphysicalactivity/publications/fruit_vegetables_report.pdf (accessed October 2016).Google Scholar
15. Jenab, M, Slimani, N, Bictash, M et al. (2009) Biomarkers in nutritional epidemiology: applications, needs and new horizons. Hum Genet 125, 507525.Google Scholar
16. Woodside, JV, Young, IS & McKinley, MC (2013) Fruits and vegetables: measuring intake and encouraging increased consumption. Proc Nutr Soc 72, 236245.Google Scholar
17. Rooney, C, McKinley, MC, Appleton, KM et al. (2017) How much is ‘5-a-day’? A qualitative investigation into consumer understanding of fruit and vegetable intake guidelines. J Hum Nutr Diet 30, 105113.Google Scholar
18. Woodside, JV, Rooney, C & McKinley, MC (2014) The 5-a-day message – should we be aiming higher? Nutr Bull 39, 351353.Google Scholar
19. Kypridemos, C, O'Flaherty, M & Capewell, S (2014) Fruit and vegetable consumption and non-communicable disease: time to update the ‘5 a day’ message? J Epidemiol Commun Health 68, 799800.Google Scholar
20. Muraki, I, Imamura, F, Manson, JE et al. (2013) Fruit consumption and risk of type 2 diabetes: results from three prospective longitudinal cohort studies. BMJ 347, f5001.Google Scholar
21. Oude Griep, LM, Geleijnse, JM et al. (2010) Raw and processed fruit and vegetable consumption and 10-year coronary heart disease incidence in a population-based cohort study in the Netherlands. PLoS ONE 5, e13609.Google Scholar
22. Bhupathiraju, SN & Tucker, KL (2011) Greater variety in fruit and vegetable intake is associated with lower inflammation in Puerto Rican adults. Am J Clin Nutr 93, 3746.Google Scholar
23. Oude Griep, LM, Verschuren, WM, Kromhout, D et al. (2012) Variety in fruit and vegetable consumption and 10-year incidence of CHD and stroke. Public Health Nutr 15, 22802286.CrossRefGoogle ScholarPubMed
24. Cooper, AJ, Sharp, SJ, Lentjes, MA et al. (2012) A prospective study of the association between quantity and variety of fruit and vegetable intake and incident type 2 diabetes. Diab Care 35, 12931300.Google Scholar
25. Bingham, S, Luben, R, Welch, A et al. (2008) Associations between dietary methods and biomarkers, and between fruits and vegetables and risk of ischaemic heart disease, in the EPIC Norfolk Cohort Study. Int J Epidemiol 37, 978987.Google Scholar
26. Willett, WC (2008) Flawed study designs are not salvaged by large samples. Int J Epidemiol 37, 987988.Google Scholar
27. Bingham, SA (2002) Biomarkers in nutritional epidemiology. Public Health Nutr 5, 821827.Google Scholar
28. Institute of Medicine of the National Academies (2007). Dietary Reference Intakes; Research Synthesis Workshop Summary. Washington, DC: The National Academies Press.Google Scholar
29. Hedrick, VE, Dietrich, AM, Estabrooks, PA et al. (2012) Dietary biomarkers: advances, limitations and future directions. Nutr J 11, 109.Google Scholar
30. Kuhnle, GGC (2012) Nutritional biomarkers for objective dietary assessment. J Sci Food Agric 92, 11451149.Google Scholar
31. Tasevska, N, Runswick, SA, McTaggart, A et al. (2005) Urinary sucrose and fructose as biomarkers for sugar consumption. Cancer Epidemiol Biomarkers Prev 14, 12871294.Google Scholar
32. Medina-Remón, A, Barrionuevo-González, A, Zamora-Ros, R et al. (2009) Rapid Folin-Ciocalteu method using microtiter 96-well plate cartridges for solid phase extraction to assess urinary total phenolic compounds, as a biomarker of total polyphenols intake. Anal Chim Acta 634, 5460.Google Scholar
33. Baldrick, FR, Woodside, JV, Elborn, JS et al. (2011) Biomarkers of fruit and vegetable intake in human intervention studies: a systematic review. Crit Rev Food Sci Nutr 51, 795815.Google Scholar
34. Pennant, M, Steur, M, Moore, C et al. (2015) Comparative validity of vitamin C and carotenoids as indicators of fruit and vegetable intake: a systematic review and meta-analysis of randomised controlled trials. Br J Nutr 114, 13311340.Google Scholar
35. Woodside, JV, Young, IS, Gilchrist, SE et al. (2013) Factors associated with serum/plasma concentrations of vitamins A, C, E and carotenoids in older people throughout Europe: the EUREYE study. Eur J Nutr 52, 14931501.Google Scholar
36. Cooper, AJ, Sharp, SJ, Luben, RN et al. (2015) The association between a biomarker score for fruit and vegetable intake and incident type 2 diabetes: the EPIC-Norfolk study. Eur J Clin Nutr 69, 449454.CrossRefGoogle ScholarPubMed
37. McGrath, AJ, Hamill, LL, Cardwell, CC et al. (2016) Combining vitamin C and carotenoid biomarkers better predicts fruit and vegetable intake than individual biomarkers in dietary intervention studies. Eur J Nutr 55, 13771388.Google Scholar
38. Jin, Y, Gordon, MH, Alimbetov, D et al. (2014) A novel combined biomarker including plasma carotenoids, vitamin C, and ferric reducing antioxidant power is more strongly associated with fruit and vegetable intake than the individual components. J Nutr 144, 18661872.Google Scholar
39. Souverein, OW, de Vries, JH, Freese, R et al. (2015) Prediction of fruit and vegetable intake from biomarkers using individual participant data of diet-controlled intervention studies. Br J Nutr 113, 13961409.Google Scholar
40. Gibney, MJ, Walsh, M, Brennan, L et al. (2005) Metabolomics in human nutrition: opportunities and challenges. Am J Clin Nutr 82, 497503.Google Scholar
41. O'Gorman, A, Gibbons, H & Brennan, L (2013) Metabolomics in the identification of biomarkers of dietary intake. Comput Struct Biotechnol J 4, e201301004.Google Scholar
42. Lloyd, AJ, Favé, G, Beckmann, M et al. (2011) Use of mass spectrometry fingerprinting to identify urinary metabolites following consumption of specific foods. Am J Clin Nutr 94, 981991.Google Scholar
43. Scalbert, S, Brennan, L, Manach, C et al. (2014) The food metabolome: a window over dietary exposure. Am J Clin Nutr 99, 12861308.Google Scholar
44. Heinzmann, SS, Brown, IJ, Chan, Q et al. (2010) Metabolic profiling strategy for discovery of nutritional biomarkers: proline betaine as a marker of citrus consumption. Am J Clin Nutr 92, 436443.Google Scholar
45. Lloyd, AJ, Beckmann, M, Favé, G et al. (2011) Proline betaine and its biotransformation products in fasting urine samples are potential biomarkers of habitual citrus fruit consumption. Br J Nutr 106, 812824.CrossRefGoogle ScholarPubMed
46. Edmands, WM, Beckonert, OP, Stella, C et al. (2011) Identification of human urinary biomarkers of cruciferous vegetable consumption by metabonomic profiling. J Proteome Res 10, 45134521.Google Scholar
47. Whigham, LD & Redelfs, AH (2015) Optical detection of carotenoids in living tissue as a measure of fruit and vegetable intake. Conf Proc IEEE Eng Med Biol Soc 2015, 81978200.Google Scholar
48. Nguyen, LM, Scherr, RE, Linnell, JD et al. (2015) Evaluating the relationship between plasma and skin carotenoids and reported dietary intake in elementary school children to assess fruit and vegetable intake. Arch Biochem Biophys 572, 7380.Google Scholar
49. Freedman, LS, Kipnis, V, Schatzkin, A et al. (2010) Can we use biomarkers in combination with self-reports to strengthen the analysis of nutritional epidemiologic studies? Epidemiol Perspect Innov 7, 2.Google Scholar
50. Freedman, LS, Midthune, D, Carroll, RJ et al. (2011) Using regression calibration equations that combine self-reported intake and biomarker measures to obtain unbiased estimates and more powerful tests of dietary associations. Am J Epidemiol 174, 12381245.CrossRefGoogle ScholarPubMed
51. Freedman, LS, Tasevska, N, Kipnis, V et al. (2010) Gains in statistical power from using a dietary biomarker in combination with self-reported intake to strengthen the analysis of a diet-disease association: an example from CAREDS. Am J Epdemioli 172, 836842.Google Scholar
52. Lippi, G, Guidi, GC, Mattiuzzi, C et al. (2006) Preanalytical variability: the dark side of the moon in laboratory testing. Clin Chem Lab Med 44, 358365.Google Scholar
53. Blanck, HM, Bowman, BA, Cooper, GR et al. (2003) Laboratory issues: use of nutritional biomarkers. J Nutr 133, 888S894S.Google Scholar
54. Vuilleumier, JP & Keck, E (1989) Fluorometric assay of vitamin C in biological materials using a centrifugal analyser with fluorescence attachment. J Micronutr Anal 5, 2534.Google Scholar
55. Salminen, I & Alfthan, G (2008) Plasma ascorbic acid preparation and storage for epidemiological studies using TCA precipitation. Clin Biochem 41, 723727.Google Scholar
56. Craft, NE, Wise, SA & Soares, JH (1992) Optimisation of an isocratic high performance liquid chromatography separation of carotenoids. J Chromatogr 589, 171176.Google Scholar
57. Timpson, NJ, Forouhi, NG, Brion, MJ et al. (2010) Genetic variation at the SLC23A1 locus is associated with circulating concentrations of L-ascorbic acid (vitamin C): evidence from 5 independent studies with >15,000 participants. Am J Clin Nutr 92, 375382.Google Scholar
58. Borel, P (2012) Genetic variations involved in interindividual variability in carotenoid status. Mol Nutr Food Res 56, 228240.Google Scholar
59. Tomkins, A (2003) Assessing micronutrient status in the presence of inflammation. J Nutr 133, 1649S1655S.Google Scholar
60. Reboul, E, Thap, S, Tourniaire, F et al. (2007) Differential effect of dietary antioxidant classes (carotenoids, polyphenols, vitamins C and E) on lutein absorption. Br J Nutr 97, 440446.Google Scholar
61. Vioque, J, Weinbrenner, T, Asensio, L et al. (2007) Plasma concentrations of carotenoids and vitamin C are better correlated with dietary intake in normal weight than overweight and obese elderly subjects. Br J Nutr 97, 977986.Google Scholar
62. Alberg, A (2002) The influence of cigarette smoking on circulating concentrations of antioxidant micronutrients. Toxicology 180, 121137.Google Scholar
63. Padayatty, S & Levine, M (2008) Fruit and vegetables: think variety, go ahead, eat! Am J Clin Nutr 87, 57.Google Scholar
64. Couillard, C, Lemieux, S, Vohl, M-C et al. (2016) Carotenoids as biomarkers of fruit and vegetable intake in men and women. Br J Nutr 116, 12061215.Google Scholar
65. Zeng, C (2012) Effects of different cooking methods on the vitamin C content of selected vegetables. Nutr Food Sci 43, 438443.Google Scholar
66. Miglio, C, Chiavaro, E, Visconti, A et al. (2008) Effects of different cooking methods on nutritional and physicochemical characteristics of selected vegetables. J Agric Food Chem 56, 139147.Google Scholar
67. Priyadarshani, AM (2017) A review on factors influencing bioaccessibility and bioefficacy of carotenoids. Crit Rev Food Sci Nutr 57, 17101717.Google Scholar
Figure 0

Fig. 1. (Colour online) Timescale of nutritional biomarkers from different biological sources (adapted from Kuhnle(30)).

Figure 1

Table 1. Factors affecting nutritional biomarker response (adapted from Jenab et al.(15)), with specific examples added for proposed fruit and vegetable (FV) biomarkers