Paul M. Finglas
Nutrition, Diet & Health Department, Institute of Food Research, Norwich Research Park, Colney, Norwich, NR4 7UA, UK
There is a need for improvements in the determination of vitamins in food, in particular, the establishment of properly validated and robust techniques that are applicable to a wide range of food matrices. This paper addresses three main topics: first, recent developments in methods including LC techniques and biospecific methods utilizing antibodies and naturally occurring vitamin binding proteins which are both based on the microtitration plate format; second, results from a European Union project under the Measurement and Testing Programme concerned with the improvement in vitamin analysis in food by intercomparisons of methods, optimization of extraction conditions and the preparation of food reference materials (RMs); and third, the impact of the improvements in methods on the quality of vitamin data currently presented in UK food tables by comparing calculated vitamin intakes obtained using both the 4th and 5th editions of McCance & Widdowson's The Composition of Foods and direct analysis of duplicate diets.
It is important that the method of analysis chosen for any vitamin should be that which most closely reflects the vitamin activity of the food in question since the primary objective for use of the data is for nutritional purposes. The term “vitamin” reflects a certain physiological activity which is related to the chemical substances or “vitamers” responsible for this activity (1). Ideally methods would be chosen that could determine each vitamer separately and then by calculating the sum of the individual activities, a total activity of the food could be obtained (2). However, in practice this is rarely possible as procedures are not specific for the compounds of interest.
The major advantage of this technique is that individual forms of the vitamin can be measured, and together with estimates of the biological activity of the various forms, can be used to provide better estimates of the vitamin activities of foods than are currently available in food tables.
Although LC techniques have been widely used for the determination fat-soluble vitamins in food, their application to water-soluble vitamin analysis has mainly been limited to vitamins B1, B2, B6 and C. This has largely been due the availability of sufficiently sensitive and specific detection systems that are capable of quantifying several vitamers from a complex mixture of compounds. The use of LC with UV detection has been used for the determination of these vitamins in food but this form of detection is generally not sufficiently sensitive nor specific due to the low levels found, especially in unfortified foods (3). LC with fluorescence detection has been preferred using either the natural fluorescence of the vitamin (e.g. riboflavin), or with derivatization to form a suitable fluorescent complex (e.g. thiamin and vitamin C). This form of detection gives better sensitivity and specificity compared to UV detection (3). Examples of LC procedures available for selected water-soluble vitamins are given in Table I.
Thiamin and Riboflavin. These vitamins are usually extracted from foods using dilute mineral acids and autoclaving at 121°C, followed by enzymatic hydrolysis to release the bound forms of the vitamins. Riboflavin can be measured directly with fluorescence but thiamin requires conversion to thiochrome with alkaline potassium ferricyanide solution. The latter can be performed either manually prior to injection onto the analytical column, or by post-column derivatization. Reverse phase LC analysis of thiochrome invariably involves the injection of high salt concentrations on to the analytical column necessitating frequent washing and much reduced column life (4). This can be overcome to some extent by the use of guard columns but this may be uneconomic. Alternatively, thiochrome can be selectively extracted into an organic solvent, normally isobutanol, allowing fluorometric determination after a normal phase LC separation of thiochrome from any remaining fluorescing interferences (5). This approach can give increased sensitivity and rapid automated analysis without the need for post-column derivatization.
Thiamin methods based on the thiochrome reaction after acid hydrolysis and treatment with enzyme to release the phosphorylated forms give total thiamin concentrations. The separation of thiamin and its phosphorylated forms [thiamin monophosphate (TMP), thiamin di- or pyro-phosphate (TPP) and thiamin triphosphate (TTP)] has also been reported (6).
Table I. Selected LC procedures available for some water-soluble vitamins
|Vitamin||Principle||Column||Mobile phase||Detectiona (nm)||Reference|
|Thiamin||1||Post-column oxidation to thiochrome||Silica (radpak)||0.05M Phosphate buffer: EtOH||F (265/418)||37|
|2||Pre-column oxidation to thiochrome||μ-Bondapak C-18||MeOH:H2O||F (365/435)||38|
|3||Free thiamin + phosphorylated forms||Micropak Ax5||Ammonium phosphate||UV (245)||6|
|Riboflavin||Native fluorescence (B2, FMN)||Apex ODS 2||MeOH:H2O||F (450/510)||37|
|Vitamin B6||1||Acid digestion with autoclaving (PM, PL, PN)||Spherisorb ODS 2||0.04M H2SO4:MeOH||F(290/395)||2|
|2||HC104 extraction + post-column reaction with bisulphate (PM, PL, PN, PMP, PLP)||Lichrosphere RP 18||0.03M Phosphate buffer: McOH:ion pair||F (340/400)||9|
|3||TCA extraction + post-column reaction with KH2PO4||Lichrosphere RP18||0.1M KH2PO4:EtOH acetonitrile: ion pair||F (325/385)||7|
|4||Acid phosphatase hydrolysis. PM→PL using glyoxylic acid/Fe2+; PL→PN using NaBH4/NaOH||Octysilyl||0.05 KH2PO4: acetonitrilc:ion pair||F (290/395)||10|
|Vitamin C||1||Acid extraction + oxidation with ascorbate oxidase) (AA + DHAA)||Hypersil ODS||0.8 M Phosphate buffer:MeOH||F (367/418)||39|
|2||Acid extraction (AA only)||Partisil P5||H2O:MeOH: acetic acid||UV (248)||Leatherhead Food RA|
|3||Acid extraction, homocysteine reduction (Total AA, AA)||Apex ODS 2||Acetate buffer:NaOH||EC||11|
|Folates||Hog kidney deconjugation, anion exchange cleanup, post-column reaction with Ca(OCl)2||μ-Bondapak phenyl||Phosphate buffer: acetonitrile||F(9295/365)||14|
a Fluorescence (excitation and emission wavelengths); EC, electrochemical; UV, ultraviolet
Results from the manual thiochrome procedure agree well with the LC fluorometric procedures but are generally lower when compared to the microbiological results (7). The main cause of variability between laboratories for the determination of thiamin was found to be the extraction/enzyme hydrolysis step (7).
Similarly, for riboflavin, LC methods generally give lower results compared to the microbiological assay (MA). Most test organisms used in the MA give an equivalent growth response to riboflavin and flavin mononucleotide (FMN), which are both vitamin active. In LC procedures, riboflavin is separated from FMN and the latter is not measured. In the same study as above several laboratories reported incomplete conversion of FMN to riboflavin and thus lower LC results compared to the MA. The enzyme conversion of FMN to riboflavin needs to be improved if LC is used for quantification (7).
Vitamin B6. LC provides a technique that is capable of measuring in a single chromatographic run all vitamin B6 forms [pyridoxamine (PM), pyridoxal (PL), pyridoxine (PN), pyridoxamine phosphate (PMP) and pyridoxal phosphate (PLP)] and is now more widely used than the microbiological assay in food composition work. All these five forms have equal biological activity (1).
A major limitation of the MA is the lower growth response to PM compared to PL and PN obtained using some organisms. This can obviously result in a substantial underestimation in the total vitamin B6 activity of foods that are high in the PM form (8). Two approaches have been used in the development of LC methods for vitamin B6 depending on whether the separation and quantification of the phosphorylated forms are required in addition to PM, PL and PN. The normal acid digestion with autoclaving and enzymatic digestion is used for the extraction of total amounts of PM, PL and PN and these forms can be quantified using a reverse phase LC system with fluorometric detection (2). This type of extraction gives an array of potential interfering compounds that can make accurate quantification difficult. In addition, the conversion of PMP to PM can be incomplete unless autoclaving for 2 hours is performed. However, such conditions may lead to thermal degradation of vitamin B6 vitamers (9).
The second type of extraction system developed for vitamin B6 employs less vigorous conditions in order to keep the phosphorylated forms (PMP and PLP) in tact. Typical acids used include sulfosalicylic acid, SSA (8), trichloroacetic acid, TCA (7) and perchloric acid, PER (9). The SSA procedure uses anion exchange sample clean-up to remove SSA and interfering compounds prior to separation and quantification on a similar column with fluorometric detection. All five B6 forms and the internal standard, 3-hydroxypyridine, are measured by means of detector wavelength switching. The inclusion of an internal standard before sample extraction is useful as corrections for any dilution errors and instrument variation can be made. This method has been compared to a microbiological assay and satisfactory agreement obtained for meat samples (8). Differences were found, however, between the two methods for fruits and vegetables. This was attributed to incomplete extraction of vitamers for the LC procedure (8).
The other two systems with TCA and PER use reverse-phase LC with ion-pair regents and a post-column reaction with potassium dihydrogen phosphate (7) and sodium bisulfite (9), respectively, to enhance the flurometric response of the various forms. The internal standard used was 4-deoxypyridine. Chromatograms for these procedures tend to give fewer interfering peaks and better resolution compared to those LC methods measuring total amounts of PM, PL and PN forms. For the determination of vitamin B6 in fortified foods where PN is predominantly present, LC systems can be greatly simplified.
A recent LC procedure using ionpair reagent and fluorometric detection has been reported (10). After dephosphorylation using acid phosphatase, PM is transformed to PL with glyoxylic acid/ Fe2+ as catalyst, followed by conversion of PL to PN using alkaline sodium borohydride. The conversion of PM to PL was reported to be 95 per cent but it was found the conversion of PL to PN was always less than 100 per cent and variable from food to food. The procedure was only tested on four foods [yeast, wheat germ and two types of breakfast cereals (with bran and muesli)] and no comparison with the MA was made. However, the method has the advantage of much simplified chromatograms with only the separation and quantification of PN required.
Vitamin C. Vitamin C activity is exhibited by L-ascorbic acid (AA) and Ldehydroascorbic acid (DHAA) but not by D-isoascorbic acid (erythorbic acid, IAA) (1). The latter is also used as an antioxidant in foods and it is therefore important to have methods that can measure both AA and DHAA either separately or together, and differentiating them from IAA (1).
A variety of LC methods have been reported for the determination of vitamin C in food and biological materials have been reported either as a total value (AA+DHAA), or with separate determinations of each form. The latter can increase the complexity and length of the procedure and introduce additional errors and thus LC methods which give total vitamin C values are preferred (1). Three types of detection systems have been used. Fluorescence detection requires the oxidation of AA to DHAA with ascorbate oxidase or charcoal followed by a reaction with o-phenylenediamine to form a fluorescent quinoxaline derivative (1). These methods measure total vitamin C, i.e. ascorbic and dehydroascorbic acids.
Those LC methods that employ UV detection are less specific and can only measure AA since DHAA is non-UV absorbing (2). In addition, they may not separate AA from IAA (2).
The most recent form of detection used for vitamin C is electrochemical which is extremely sensitive. DHAA is reduced to AA by homocysteine and the total AA measured electrochemicallychemically using an Ag/AgCl electrode (11). Concentrations of DHAA have been found to be fairly low in individual foods and mixed diets compared to AA (12). Good agreement between the vitamin C results obtained using various types of LC methods and the manual fluorometric procedure has been found (7).
Folates. Various LC methods have been developed for the determination of folates but most of these procedures have only been applied to standard mixtures and fortified foods such as infant foods and breakfast cereals where pteroylmonoglutamic acid (PGA) is the predominant form. When these methods have been applied to natural levels found in most foods, extensive sample clean-up and purification is required (13).
Most of the LC methods developed have concentrated on the analysis of folate monoglutamates which requires conjugase treatment of samples prior to analysis. The selection of a deconjugase enzyme and of reaction conditions that provide complete hydrolysis of folate polyglutamates to the monoglutamate level is essential for accurate LC quantification (14). Three types of deconjugase enzymes are available: hog kidney (HK), human plasma (HP) and chicken pancreas (CP). The first two enzymes produce mainly monoglutamate products whereas CP gives essentially diglutamates. Although CP may be less susceptible to inhibition by certain food components compared to HP and HK enzymes (15), it is not advisable to use this enzyme in chromatographic studies as identification and quantification of diglutamate products is difficult due the lack of availability of commercial standards for these forms.
Although HK enzyme gives essentially monoglutamate products, deconjugase activity can be low compared to CP necessitating extensive enzyme clean-up and purification before use (16).
There has been renewed interest in the use of HP as the deconjugase enzyme because it is readily available, it can be used without any purification, it contains low levels of folate and it gives a monoglutamate end-point (17). However, a recent collaborative study has found it may not be as effective as CP or HK in three foods (wheat flour, milk flour and yeast powder) (Lumley & Finglas, unpublished data). A triple enzyme combination consisting of CP (acetone preparation; Difco), α-amylase (Aspergillus oryzae; Sigma) and Pronase (Calbiochem-Behring Corp.) has been reported to give higher folate values in some starchy foods compared to CP alone (18). Similar results have also been found in cereal products using CP, a heat resistant α-amylase (Termamyl; Novo) and amyloglucosidase (Sigma) (19). It is clear that whatever conjugase enzyme is used for folate deconjugation, it should be fully optimized for each food analyzed.
Table II. Folate content (nmol/g) and distribution in selected foods using LC and microbiological assay (MA)a
|Whole cow's milk||ndb||0.07||nd||0.07||0.03|
a Adapted from (14)
b nd, not detected
There has also been recent interest in being able to use LC in quantifying not only any type of folate in respect to its carbon substitution but polyglutamate forms as well (20). This would obviously eliminate the need for the lengthy deconjugation step and facilitate the use of simplified extraction systems.
The most promising LC procedures are those based on methods incorporating anion-exchange clean-up followed by hog kidney deconjugation and quantification using the native fluorescence of the reduced folates in an acidic mobile phase (14). Extraction conditions were chosen so that 10-formyl-tetrahydrofolic acid (10-CHO-THF) was quantitatively converted to 5-formyl-tetrahydrofolic acid (5-CHO-THF). While tetrahydrofolic acid (THF) and 5-methyl-tetrahydrofolic acid (5-CH3THF) can be readily quantified using their natural fluorescence, the relatively weak fluorescence of 5-CHOTHF is a limiting factor and requires careful control of the slit widths of the fluorometer during the chromatographic run to give greater sensitivity. For PGA, THF and dihydro-folic acid, a oxidative post-column derivatiation system using calcium hypochlorite converts these compounds to highly fluorescent pterin compounds. The agreement between this LC procedure and the MA for six foods was found to be poor (Table II) and further work is needed to explain the reasons for this.
Despite many of the drawbacks concerning the use of microbiological assays for the analysis of water-soluble vitamins, they are still widely used for folate, vitamin B12, pantothenic acid, niacin and biotin. Over the years MAs have received a great deal of attention especially for folates (17). Although the MA has been modified and improved, the basic concept of the assay has not changed.
In our laboratory, the response of folate monoglutamates to L. rhamnosus (formerly L. casei) has been extensively studied (21–23) and a pH of 6–6.2 found to give equivalent response of the major forms over the 0–1 ng calibration range.
The most recent development in the MA for folates has been the introduction of the 96-well microtitration plate based assay (24, 25). In our experience this has improved assay reliability and increased sample through-put dramatically. Bacterial growth in each well is measured using a microplate reader at 600–670 nm, the entire plate taking 2–3 minutes to read completely. Seven levels of standard and 16 samples (all in quadruple) are included on each plate, with ten plates in an assay run. Considerable time is saved using the microtitration plate assay over conventional procedures using test tubes.
Table III. Biospecific methods currently developed for the determination of watersoluble vitamins in fooda
|2.||Avidin/Streptavidinb Biotin-HRP||2.||Breakfast cereal||2.39|
|Biotin||ELISA||1.||Anti-biotin antiserac Biotin-KLH||1.||Liver||1.-|
|2.||Anti-biotin antiserab Biotin-HRP|
|Pantothenic acid (PA)||ELISA||Anti-pantothenic acidc antisera, PA-KLH||Milk, eggs, bread, potatoes, liver, lettuce||26|
|Bread, milk, rice, vegetables, chicken, beef||30|
|Anti-pyridoxamine antiserac, PL-KLH R-protein-HRP, B12-KLH||Breakfast cereals||29|
|Folates, total||EPBA||1.||FBP-HRP, PGA-KLH||1.||Vegetables, animal feeds||1.27|
|2.||FBP-HRP, BSA-KLH||2.||Breakfast cereals, fruit juices||2.39|
|PGA||ELISA||3.||Anti-folate antiserab PGA-KLH||3.||Standard mixtures||3.-|
|4.||Anti-folic antiserab PGA-HRP||4.||Multivitamin preparations||4.39|
a Adapted from (44)
b Commercially available
c In-house antisera
During the last five years biospecific methods of analysis using either antibodies [enzyme-linked immunosorbent assay (ELISA) format], or naturally occurring vitamin binding proteins [enzyme protein-binding assay (EPBA) format], have been shown to be capable of producing sensitive and specific assays for a number of B-group vitamins (Table III). Both types of assay are based on the microtitration plate format which is ideally suited to high sample through-put. ELISAs have the greater potential for the measurement of specific vitamers, whereas EPBAs is more useful for broad specificity assays where the analysis of a group of vitamers is required (e.g. folates).
Most of the ELISAs developed at Norwich have used in-house antisera which can take several months to produce. Antisera have been raised against the major, naturally occurring forms of each vitamin. There is, however, a range of commercial antisera (polyclonal and monoclonal) currently available to several vitamins including PGA, 5-MTHF, 5-CHO-THF, vitamin B12 and biotin. This should facilitate the development of suitable ELISAs for food use.
There are limited comparative data available comparing results from biospecific methods to MA and HPLC procedures. We have compared the determination of pantothenic acid by ELISA and MA (26); biotin, folate and vitamin B12 by EPBA and MA (27–29) and PM by ELISA and LC in a range of foods (30). In general, the agreement between the results obtained using the biospecific methods and the more conventional techniques has been good for those foods analyzed. These techniques also offer the potential of developing simplified extraction systems as it should be possible to measure bound vitamin forms.
Table IV. Food reference materials (RMs) currently under stability testing for fatand water-soluble vitaminsa
|Wholemcal flour||-||B1, B2, B6, folate, niacin, biotin|
|Milk powderc||A, E, D, β-carotene||C, folate, B12, B1, B2, B6|
|Brussels sproutsd||-||C, B1, B2, folate, B12, niacin|
|Pig liverd||A, E, D, β-carotene||Folate, B12, C, B1, B6, biotin|
|Mixed vegetablesd||Carotenoids||C, B1, B2, folate, B12|
a Adapted from (45)
b Canned product
c Vtamin enriched
• Community Bureau of Reference (BCR)
In 1988, BCR undertook a research project to improve the quality of vitamin analysis carried out by its Member States for nutritional labeling purposes and the production of food composition data. The project involves about 50 European laboratories specialized in vitamin determinations and includes method intercomparison studies to identify and eliminate sources of error, and the preparation of suitably stable and homogeneous food reference materials (RMs).
Reference Materials (RMs)
The use of certified RMs is an essential part of quality assurance and provides a means by which results can be traced back to a certified value. RMs are especially useful in the development of new techniques. The range of RMs currently under stability testing are given in Table IV. The approach used in the preparation of RMs has been to produce dry foods but taking care to control the lyophilization conditions (time, temperature etc.) in order to minimize vitamin losses. The final materials have been packaged into food grade, aluminum sachets under an inert atmosphere. The homogeneity and stability of a wide range of vitamins are being vigorously tested for each RM.
Method Intercomparison Studies
The first intercomparison of methods for vitamins was organized to study the state of the art in a group of European laboratories experienced in vitamin determinations. A summary of the main results is given in Table V. The agreement between the participants for vitamins B1 and C, and niacin was good and indicative values were given for three candidate RMs.
For vitamin B6, two problem areas were identified. Firstly, the identification of the various vitamers proved difficult and consequently there were large variations in both individual vitamers and total B6 values obtained by the participants (Table VI). Secondly, it was found that glucoside derivatives of pyridoxine were not hydrolyzed by phosphatase and takadiastase enzymes used and required an additional β-glucosidase enzyme to release these forms for LC analysis. However, these forms are unlikely to be absorbed by man and should therefore not be included in the total B6 activity. The microbiological results for B6 were higher due to inclusion of the glucoside forms. There was also large variation in the results for vitamin B2 (Table V) and it was concluded that this was due to the hydrolysis/de-phosphorylation step of the procedure.
Table V. Summary of the results of the First BCR Intercomparison of Vitamins in the candidate food reference materials
|Vitamin||Methods usedb||Number of laboratories||Within-and between-laboratory variation|
|Thiamin||LC, fluorometric, MA||10||3–5||11–18||Indicative values given for three RMs|
|Riboflavin||LC,MA||12||4–7||28–74||Problems in extraction/dephosphorylation step|
|Vitamin B6||LC, MA||6||4–7||18–51||Problems in peak identification of extraction/hydrolysis|
|Niacin||MA||7||3–5||9–15||Indicative values given for three RMs|
|Vitamin C||LC, fluorometric||10||5||15||Indicative value given for one RM|
a adapted from (7)
b LC, liquid chromotography; MA, microbiological assay
Table VI. Results for B6 content and individual vitamers in three food reference materials (RM)a. Results for individual vitamers are means (with ranges in parentheses) from five laboratories using LC procedures.
(% of total B6)
(mg/100 g dry mass)
|Haricot vert beans||50|
a Adapted from (7)
b PM, pyridoxamine; PL, pyridoxal; PN, pyridoxine.
Figure 1. BCR-intercomparison on the determination of folates in food
Determination of folate in a brussels sprouts RM using microbiological assay (MA), enzyme protein-binding assay (EPBA), radioassay kit (RIA) and liquid chromatography (LC). (∆), chicken pancreas, (×) human plasma and (+) hog kidney deconjugase enzymes. Results are means +/- 1SD (Reprinted with permission from Food Chemistry, Copyright 1992, Elsevier Applied Science Publishers Ltd).
Few laboratories were able to perform analyses for folate and vitamin B12 and thus a second intercomparison was organized for folate analysis in a lyophilized brussels sprouts material using MA, LC, EPBA and radio-protein binding kit (RIA) (13). Three types of deconjugation were investigated: HP, CP and HK. Good agreement was obtained with laboratories using MAs (Figure 1). CP deconjugation gave about 20 per cent higher folate levels in this foodstuff compared to HP. The use of autoclaving followed by CP or HP deconjugation gave lower (10–20 per cent) levels determined by MA when compared to refluxing and deconjugation with the same enzymes. Problems in the identification of peaks and poor calibration were found in the LC procedures used. In general, EPBA and RIA results were higher than MA and LC values but much more variable. It was concluded that the response of the individual folate forms to the binding protein used in the assay is crucial and careful control of assay pH and type of calibrant is required if an equivalent response to the main folate forms is to be obtained. Further work is needed before LC and other techniques can be used for routine folate analysis in food.
The extraction conditions for the LC determination of vitamins B1, B2 and B6 in three food materials (pork muscle, milk powder and wholemeal flour) have also been investigated. The aim of this study was to optimize conditions for enzyme hydrolysis (pH, time, temperature and sample-enzyme ratio). Four commercially available enzymes which are most commonly used were selected: takadiastase (Pfaltz & Bauer, Serva), phosphatase (Sigma) and a mixture of takadiastase (Fluka) and phosphatase (Sigma). Optimum levels of B6 in pork and milk powder were found using either takadiastase (Pfaltz & Bauer) or phosphatase (Sigma) with a pH of 4.8–5.5, an overnight incubation and a temperature of 37–45C. The minimum amounts of enzyme needed were 500 mg takadiastase or 50 mg phosphatase per gram of sample. For B1 in pork and flour, optimum levels were only found using takadiastase (Pfaltz & Bauer) with a pH of 4 and incubation time of four hours. The minimum amount of enzyme needed was 100 mg/g pork and 20 mg/g flour. The conversion of thiamin monophosphate to free thiamin was only about 30 per cent using the takadiastase/phosphatase mixture and these enzymes should be avoided for LC work. Similar results were obtained for B2 in pork and flour except an overnight incubation and higher enzyme levels (100 mg takadiastase/g pork, 20 mg takadiastase/g flour) were required. The conversion of flavin mononucleotide (FMN) to free vitamin was not complete using both enzyme preparations. The takadiastase/phosphatase mixture gave about 60 per cent conversion whereas takadiastase (Pfaltz & Bauer) gave nearly 80 per cent conversion. This is not a problem in the MA as FMN gives an equivalent growth response compared to riboflavin.
• Impact of Improvements in the Determination of Water-Soluble Vitamins on the Quality of Food Composition Data in UK Food Tables
Many nutritional studies have relied upon the 4th edition of McCance and Widdowson's The Composition of Foods (MW4) for the calculation of nutrient intakes (31). This has been superseded by a revised 5th edition (MW5) (32) containing a much wider range of foods and new analytical data using improved methods of analysis.
One way of assessing the impact of improved vitamin methods is to compare vitamin intake values computed from food composition tables with values obtained by direct analysis of diets consumed. This is also important as many nutritional studies in the literature have relied on MW4 for the calculation of vitamin intakes and it is therefore essential to assess the likely impact on the interpretation of such studies if MW4 is now replaced with revised data given in MW5.
In a study in Norwich, nutrient intake data, collected from 54 adolescents using a 7-day weighed inventory (recorded every 6th day for 7 weeks), calculated using MW4 with the same data calculated using MW5. In addition, intake values obtained using MW4 and MW5 are compared with values obtained by direct analysis of duplicate diets collected on the same day of dietary recording (33, 34). In particular, MW5 versus analyzed values are examined for those nutrients where the agreement was found to be poor when calculated intake values were computed using MW4, for example folate and vitamin B6 (33). A summary of methods used to analyze duplicate diets for selected water-soluble vitamins are given in Table VII. Typical methods used for food composition data for these vitamins in MW4 and MW5 are also included.
Table VII. Summary of methods used to analyse duplicate diets of schoolchildren in Norwich study for energy and selected water-soluble vitamins and typical methods used for UK food composition tables (MW4 & MW5)a
|Nutrient||Methodsb used for duplicate diet analysis|
|Methodse used for UK food composition folates|
|Energy||Bomb calorimetry corrected to metabolizable energy (40)||Calculation from protein, fat, carbohydrate and alcohol||Calculation from protein, fat, carbohydrate and alcohol|
|Thiamin (B1)||LC/F (41, 42)||F (43), MA (35)||MA (34), F (43), LC/F (37)|
|Riboflavin (B2)||LC/F (41, 42)||MA (34), F (43)||MA (34), LC/F (35, 37)|
|Vitamin B6c||LC/F (2, 42)||MA (34)||MA (34), LC/F (2)|
|Vitamin Cd||LC/EC (11,42)||T (AA only; 43), F (Total; 43)||T (AA only; 43) F(total; 43)|
|Folate (total)||MA (21,42)||MA (34)||MA (34), MA (21)|
a (31); (32)
b LC/F, liquid chromatography with fluorometric detection; LC/EC, liquid chromatography with electrochemical detection and MA, microbiological assay using Lactobacillus rhamnosus at pH 6.2
c Total vitamin B6 = pyridoxamine (PM) + pyridoxal (PL) + pyridoxine (PN)
d Total vitamin C = ascorbic acid (AA) + dehydroascorbic acid (DHAA)
e F, fluorometric & T, titrimetric. For MW4, MA (riboflavin): Streptococcus zymogenes; MA (thiamin): Lactobacillus viridescens or L. fermienti; MA (B6): Saccharomyeces carlsbergensis and MA (folate): Lactobacillus casei. For MW5, MA (thiamin): L. viridescens; MA (B6): Kloeckera apiculata; MA (folate): Lactobacillus rhamnosus & MA (B2): Streptococcus zymogenes
The possible extent of under-recording of total food intake in this study was assessed using mean daily analyzed energy intake values and estimates of daily energy expenditure for similar subjects. It was concluded that there was no gross under-estimation of habitual energy intake and the calculated nutrient intake values represented a good reflection of their habitual intake (12). The results for energy and water-soluble vitamins are given in Table VIII for girls only. Similar trends were found for boys (34).
The major differences in vitamin intake values calculated from MW4 and MW5, and compared to analyzed intake values, were found for folate and vitamin B6.
In the UK, the microbiological assay conditions of Bell (35) were used to obtain values for the folate content of foods given in MW4. Under these conditions, i.e. initial pH of 6.8 for organism growth and 0–1 ng calibration range, the growth response of Lactobacillus rhamnosus (casei) to 5-MTHF was poor in comparison to PGA, the normal folate used to calibrate the assay. If the pH is lowered to 6.2, the response of 5-MTHF and PGA is the same (21). Although PGA is used for food fortification because of its greater stability and lower cost, it does not occur naturally. The implication of this is that in foods that contain appreciable amounts of natural folates, the folate values given in MW4 are likely to be grossly under-estimated. This has subsequently led to an apparent underestimate in the true folate intake in the UK (22, 23). Analysis of 128 vegetables for folate using an improved MA found that the revised folate values were about two-fold higher compared to MW4 data (36).
Table VIII. Average daily intakes of energy and water-soluble vitamins calculated from MW4, MW5 and direct analysis of duplicate diets
Values are means with standard errors for 35 girls (13–14 year olds) with ranges in parentheses.a
|Analyzedb||Correlation coefficient, rc||Calculated|
|Analyzedb||Correlation coefficient, rc|
|Vitamin B6 (mg)||1.1±0.04*|
|Vitamin C (mg)||75±7|
|Folate (mg)||145±8 **|
a Adapted from (42)
b For details of methods see Table VII
c Correlation coefficient (r) between calculated MW5 & MW4 and analyzed data for girls [statistical significance of r=P<0.05]
d Calculated MW4/MW5 intake was significantly different from analyzed data:
e NS, not significant
In this study, calculated folate intake values using MW5 are considerably higher (50 per cent) than MW4 intake values and much nearer the values obtained by duplicate diet analysis. The revised folate content of several breakfast cereals, especially cornflakes and muesli, using the improved MA were much higher and made a significant contribution to the total daily intake for this vitamin. Although a weak correlation was found for between MW4 calculated and analyzed intake values for girls, no association was found for boys (33). However, the correlation between MW5 and analyzed intake values is improved, and significant for both sexes.
Some doubt has been expressed over the vitamin B6 data appearing in MW4, which has largely been obtained by MA using a variety of test organisms, some of which may not respond equally to all forms and thus total vitamin B6 activity may be underestimated (34). This possible underestimation will vary from food to food depending on the relative amounts of each vitamer present. Much of the newer data for vitamin B6 appearing in MW5 has been obtained using the MA with Kloeckera apiculata which should give better estimates for the total vitamin activity than other organisms used. This is likely to have contributed to the large increase in calculated intake values found when using MW5 compared to MW4. The vitamin B6 MW5 values of muesli and cornflakes, which are two of the major contributors to total daily intake of this vitamin, are considerably higher than values appearing in MW4 for these foods.
Some of the vitamin B6 values appearing in MW5, for example raw and cooked vegetables, have been obtained by an LC procedure which permits the separation and quantification of PM, PL and PN forms and the sum of which gives the total vitamin B6 activity (36). Although the analyzed intake values are about 50 per cent lower than the revised MW5 calculated intake values for total vitamin B6, there is now a significant correlation between the data for both sexes, which was not found for the MW4 calculated intake versus analyzed values. Clearly further improvements in LC methodology is needed for this vitamin, especially extraction and peak identification, before more reliable food composition data can be obtained.
There has been a steady improvement in methods for the determination of watersoluble vitamins in food over recent years notably with the development of LC and biospecifc procedures. These techniques allow the quantification of individual vitamers and thus give better estimates of the vitamin activities of foods than is currently available in food composition tables. Further work is needed on the optimization of extraction/dephosphorylation conditions and comparative data between methods. The availability of a range of certified RMs will greatly assist in method validation and improving the quality of data generated for food composition tables. Care should also be exercised when interpreting dietary intake data from nutritional studies particularly the source, limitations and reliability of the analytical techniques used to acquire them.
Various parts of this work were supported by the Office of Science and Technology and the European Communities' Community Bureau of Reference.
(1) Greenfield, H., & Southgate, D.A.T. (1992) in Food Composition Data, Elsevier Science Publishers, London
(2) Brubacher, G., Müller-Mulot, W., & Southgate, D.A.T. (1985) Methods for the Determination of Vitamins in Food, Elsevier Applied Science, London, pp. 129–140
(3) Finglas, P.M., & Faulks, R.M. (1987) J. Micronutr. Anal. 3, 251– 283
(4) Ang, C.Y.W., & Moseley, F.A. (1980) J. Agric. Food Chem. 28, 483–486
(5) Bailey, A.L., & Finglas, P.M. (1990) J. Micronutr. Anal. 7, 147– 157
(6) Hilker, D.M., & Clifford, A. J. (1982) J. Chromatogr. 231, 433– 438
(7) Hollman, P. C. H., Slangen, J.H., Wagstaffe, P.J., Faure, U., Southgate, D.A.T., & Finglas, P.M. (1993) Analyst. 118, 481–488
(8) Polansky, M.M., Reynolds, R.D., & Vanderslice, J.T. (1985) in Methods of Vitamin Assay, J.A. Augustin, B.P. Klein, D. Becker & P.B. Venugopal (Eds.), John Wiley & Sons, New York, pp. 427–428
(9) Bitsch, R., & Moller, J. (1989) J. Chromatogr. 463, 207–211
(10) Reitzer-Bergaentzle, M., Marchioni, E., & Hasselmann, C. (1993) Food Chem. 48, 321–324
(11) Behrens, W.A., & Madere, R. (1987) Anal. Biochem. 165, 102– 107
(12) Finglas, P.M., Bailey, A.L., Walker, A., Loughridge, J.M., Wright, A.J.A., & Southon, S. (1993) Br. J. Nutr. 69, 563–576
(13) Finglas, P.M., Faure, U., & Southgate, D.A.T. (1993) Food Chem. 46, 199–213
(14) Gregory, J.F., Sartain, D.B., & Day, B.P.F. (1984) J. Nutr. 114, 341–353
(15) Eigen, E., & Shockman, G.D. (1963) Fed. Proc. 42, 2105
(16) Keagy, P.M. (1985) in Methods of Vitamin Assay, J. Augustin, B.P. Klein, D.A. Becker & P.B. Venugopal (Eds.), John Wiley & Sons, New York, pp. 450–452
(17) Tamura, T. (1990) in Folic Acid Metabolism in Health & Disease, M.P. Picciano, E.L.R. Stokstad & J.F. Gregory (Eds.), Wiley-Liss, New York, pp. 121–137
(18) De Souza, S., & Eitenmiller, R. (1990) J. Micronutr. Anal. 7, 37–57
(19) Goli, D.M., & Vanderslice, J.T. (1992) Food Chem. 43, 57–64
(20) Gregory, J.F. (1985) in Methods of Vitamin Assay, J. Augustin, B.P. Klein, D.A. Becker & P.B. Venugopal (Eds.), John Wiley & Sons, New York, pp. 473–496
(21) Phillips, D.R., & Wright, A.J.A (1982) Br. J. Nutr. 47, 183–189
(22) Wright, A.J., & Phillips, D.R. (1985) Br. J. Nutr. 53, 569–573
(23) Finglas, P.M., Wright, A.J.A., Faulks, R.M., & Southgate, D.A.T. (1990) in Recent Knowledge on Iron and Folate Deficiencies in the World, Vol. 197, S. Hercberg, P. Calan & H. Dupin, (Eds.), Inserm, Paris, pp. 385–392
(24) Newman, E.M., & Tsai, J.F. (1986) Anal. Biochem. 154, 509–515
(25) Horne, D.W., & Patterson, D. (1988), Clin. Chem. 34, 2357–2359
(26) Finglas, P.M., Faulks, R.M., Morris, H.C., Scott, K.J., & Morgan, M.R.A. (1988) J. Micronutr. Anal. 4, 47–59
(27) Finglas, P.M., Faulks, R.M., & Morgan, M.R.A. (1986) J. Micronutr. Anal. 2, 247–257
(28) Finglas, P.M., Kwiatkowska, C., Faulks, R.M., & Morgan, M.R.A. (1988) J. Micronutr. Anal. 4, 309– 322
(29) Alcock, S.A., Finglas, P.M., & Morgan, M.R.A. (1992) Food Chem. 45, 199–203
(30) Alcock, S.A., Finglas, P.M., & Morgan, M.R.A. (1990) Food Agric. Immunol. 2, 197–204
(31) Paul, A.A., & Southgate, D.A.T. (1978) in McCance & Widdowson's The Composition of Foods, 4th Ed., HMSO, London
(32) Holland, B., Welch, A.A., Unwin, I.D., Buss, D.H., Paul, A.A., & Southgate, D.A.T. (1991) in McCance & Widdowson's The Composition of Foods, 5th Ed., Royal Society of Chemistry, Cambridge
(33) Southon, S., Wright, A.J.A., Finglas, P.M., Bailey, A.L., & Belsten, J. (1992) Proc. Nutr. Soc. 51, 315–324
(34) Bell, J.G. (1974) Lab. Pract. 23, 235–242, 252
(35) Kwiatkowska, C.A., Finglas, P.M., & Faulks, R.M. (1989) J. Hum. Nutr. Diet 2, 159–172
(36) Schrijver, J., Speek, A.J., Klosse, J.A., van Rijn, H.J.M., & Schreurs, W.H.P. (1982) Ann. Clin. Biochem. 19, 52–56
(37) Finglas, P.M., & Faulks, R.M. (1984) Food Chem. 18, 37–44
(38) Speek, A.J., Schrijver, J., & Schreurs, W.H.P. (1984) J. Agric. Food Chem. 32, 352–355
(39) Rubach, K., & Reichert, N. (1991) Deutsche Lebensmittel-Rundschau 88, 341–347
(40) Miller, D.S., & Payne, P.R. (1959) Br. J. Nutr. 13, 501–508
(41) Bailey, A.L., & Finglas, P.M. (1990) J. Micronutr. Anal. 7, 147– 157
(42) Southon, S., Wright, A.J.A., Finglas, P.M., Bailey, A.L., & Loughridge, J.M. (1994) Br. J. Nutr. (in press)
(43) Official Methods of Analysis (1975) 12th Ed., AOAC, Arlington, VA
(44) Finglas, P.M., Alcock, S.A., & Morgan, M.R.A. (1992) in Food Safety and Quality Assurance: Applications of Immunoassay Systems, M.R.A. Morgan, C.J. Smith, & P.A. Williams (Eds.), Elsevier Applied Science, London, 401–409
(45) Finglas, P.M., Faure, V., & Wagstaffe, P.J. (1993) Fres. J. Anal. Chem. 345, 180–184
Andrew J. Sinclair
Department of Medical Laboratory Science, Royal Melbourne Institute of Technology, GPO Box 2476 V, Melbourne, Vic 3001, Australia
In the last 20 years, there has been an increasing awareness of the nutritional importance of lipids in foods. This has led to a requirement to improve the quality and quantity of data on food lipids in food databases. This paper discusses the extraction of lipids from different food matrices, the use of manual versus automated procedures and problems which occur during the extraction process. Methodological approaches are discussed for the analysis of fatty acid composition and concentration, and the analysis of the sterols by gas chromatography and high pressure liquid chromatography. This paper raises the future requirement for a wider range of food lipid data including quantitative information on tocopherols and tocotrienols, molecular species of triacylglycerols, distribution of fatty acids on the 2-position of the triacylglycerols, cholesterol oxides and other lipid oxidation products.
Lipids are an important group of substances found in food where they make major contributions to taste, flavor and the energy content of the food. Lipids are a heterogeneous class of compounds which makes it difficult to provide a precise definition, however they are classified as those substances insoluble in water and soluble in a range of organic solvents. The information required on lipids for a food database include the total lipid content for the calculation of energy content as well as comparison between foods, the fatty acid types which include the saturates, cis- and trans-monounsaturates and ω-6 and ω-3 polyunsaturates, and the sterol content and composition, including the proportion of cholesterol in the food. While greater than 95 per cent of most food lipids in westernized countries consist of triacylglycerols (TAG), there are other lipids which need to be considered because of their presence in high concentration in certain foods. These include wax esters (found in high concentration in certain species of fish (1) and phospholipids (found in high concentration in eggs). Fatty acids, although rarely present in foods as such, are major components of all food lipids apart from sterols. Interested readers are referred to specialized books on lipids and lipid analysis (2–5).
• Extraction of Samples
The choice of technique to be adopted for the analysis of total lipids can depend on a number of factors including the type of food being analyzed and therefore the lipid classes present, the number of samples to be analyzed and the laboratory facilities. The general approach to the extraction of lipids from biological materials is to denature lipoproteins and enzymes in alcohol and to extract the lipids into an organic solvent. Simple soxhlet extraction in ether has been used in the case of foods with very high concentrations of TAG, such as meat fat and milk fat however this can lead to poor extraction of the more polar lipids (6). There have been many different procedures published for the extraction of lipids (2, 5) including refluxing liver tissue in ethanol (7), use of n-butanol saturated with water for cereals (8), the use of isopropanol as a preliminary extractant for plant tissue to inhibit phospholipase D activity (2), and a 5-stage extraction process using chloroform and methanol as well as acidic and basic solvents systems (9). A dry column method for the extraction of meat ground up with anhydrous sodium sulfate using columns containing celite followed by elution of the lipids from the column with a variety of solvents has been described (10). Extraction of lipids from lyophilized samples is known to result in difficulty of complete lipid extraction, as illustrated recently by poor recovery of TAG from oysters (11). The most common process for animal and fish tissues, however, is to blend samples in ten volumes of methanol followed by the addition of 20 volumes of chloroform, with additional re-extraction of the sample (2, 5). This approach was developed by Folch et al. (12) and Bligh and Dyer (13). The addition of antioxidants to the solvents is recommended, at a level of 50–100 mg/L in the case of butylated hydroxytoluene. Use of glass containers with glass stoppers or Teflon-lined caps is mandatory for the extraction and subsequent processing steps and use of redistilled solvents has been recommended, although highly purified solvents are available from suppliers.
Supercritical CO2 can also be used to extract lipids from foods followed by weighing of the residue (5, 14). This technique has a number of advantages including extraction of the samples at relatively low temperatures, and use of a non-toxic and inert gas which may satisfy regulatory authorities who are concerned with the use and disposal of hazardous solvents in laboratories.
• Removal of Non-Lipid Contaminants
Most polar solvents also extract non-lipids such as sugars, urea, amino acids and salts. It has been claimed that pre-extraction with 0.25 per cent acetic acid will remove contaminants and destroy lipolytic enzymes (15), however this procedure has not been widely adopted. Most contaminants can be removed by washing the organic solvent extract with water or dilute salt solution, with the proportions of solvents to aqueous phase being important to prevent losses of polar lipids (2, 12, 13). These contaminants can also be removed by passing the solvent through Sephadex G25, a procedure which is strongly recommended by Nelson (16). He showed that the extraction of 20 mL of plasma yielded 83.2 mg of organic solvent soluble material before passing through Sephadex where the yield was only 50.1 mg of lipid. Since the total lipid content of foods is usually estimated gravimetrically, this example highlights the importance of removal of non-lipid material. This is particularly important when the individual fatty acid content of foods is based on total lipid value rather than using an internal standard to quantitate the fatty acids (17). Following this clean-up step, the solvent is removed in a rotary evaporator with the water bath set at or near room temperature to protect the lipids which are then transferred in chloroform to a storage tube/flask. The total lipid content is estimated by weighing an aliquot from which the solvent has been evaporated using a stream of nitrogen gas. The remaining lipids should then be stored at -20°C, prior to future analysis of lipid classes, fatty acids, and sterols. The procedures described above are very time consuming, labor intensive and require the use of large volumes of solvents.
• Automated Lipid Extraction Procedures
There are a number of different automated techniques which can be used to analyze large numbers of food samples which would be ideally suited for the analysis of similar materials for quality control purposes. These include the soxhlet process either with or without acid digestion using solvents such as diethyl either or chloroform: methanol (6). This technique can allow samples to be analyzed in a batch process, and following extraction and evaporation of the solvent, the weight of the total fatty acids (after acid treatment) or total lipids is determined gravimetrically.
Another technique is to use a rapid automated procedure to remove water, extraction of the dry residue with dichloromethane and re-weighing the defatted residue which gives a figure for total lipids and percent water. This technique, using an instrument known as the CEM automated analyzer (CEM Corp., Indiana Trail, NC), allows the analysis of one laboratory sample at a time and is rapid (about 5 min/sample), however this process is not suited for further processing of the extracts due to loss of the lipid in the process. We have shown that this procedure does not extract all the phospholipids from meat and this can influence the total lipid value significantly in the case of the analysis of lean meats, such as lean pork, chicken breast and certain cuts of beef (18).
Supercritical CO2 can be used to extract lipids from foods in a continuous process. The extraction apparatus can be connected to a gas chromatograph (GC) or super-critical fluid chromatography unit for the analysis of the component lipids or for pesticide analysis (5, 14).
Foods or feeds can also be analyzed using infrared analysis, which is a nondestructive technique, using a single sample at a time and which provides data on total lipids, water and protein.
The advantages of these automated techniques are speed of analysis and the use of less solvent or no solvent at all. Disadvantages include the small sample size, the necessity to standardize the infrared analyzer and the CEM analyzer for each food type and the fact that it would be necessary to use a separate conventional extraction technique for the analysis of fatty acids and sterols.
A major problem exists in relation to determining a representative sample of food to be analyzed when the foods are not homogeneous, as is the case with fresh meat and fish. This problem is exacerbated if the technique adopted requires the use of only a small analytical portion. For example, we have experienced difficulty in taking representative samples for the analysis of meat using the CEM analyzer which uses only 3–5 g meat/analysis. The fat content of these retail meat samples can range from 4 to 50 per cent lipid (19) and the problem occurs at the stage where the lean meat and visible fat are mixed and blended together, and is most evident in meat with a high lipid content. It may be necessary to use a larger food sample to analyze foods which are not homogeneous.
• Analysis of Lipid Classes
Prior to the analysis of unfamiliar tissues for fatty acid composition and sterols, it is considered essential to examine the lipid extract by thin layer chromatography using a solvent system to separate the non-polar lipids from the phospholipids which remain at the origin. The lipids can be visualized using cupric acetate-orthophosphoric spray (20), followed by heating in an oven at 100°C which leads to sterols showing a characteristic purplish color initially and finally all the lipids char. This procedure can assist in determining the likely identity of the lipid classes (e.g. presence of wax esters in fish flesh) as well as the approximate proportions of each class which is useful to estimate amounts of internal standards to be used in the fatty acid and sterol determinations. More sophisticated options include use of liquid chromatography (LC) (3, 5) or the Chromarod-Iatroscan technique to separate and quantitate lipid classes (5).
• Determination of Fatty Acid Composition and Content of Foods
For a food database, information is required on chain length of saturates, cis/trans isomers of the monounsaturates, and ω-6 and ω-3 polyunsaturated fatty acids (PUFA) (21). There is also a developing interest in the positional distribution of fatty acids on TAG molecule (22) and the molecular species of TAG present in the sample (23). The general procedure for the analysis of the fatty acids is to hydrolyze the fatty acids, form derivatives and then analyze these by GC or LC with quantitation by use of internal standards. At the present time, GC is the method of choice for analysis of fatty acids as methyl esters. There are many different methods for forming the fatty acid methyl esters (FAME), including the use of BF3 in methanol, 5 per cent anhydrous HCl in methanol, and 1–2 per cent H2SO4 in methanol (2). It is usually essential to purify the esters on a small column of fluorisil which removes sterols and oxidized material (2). The most common internal standard is heptadecanoic acid (C17:0), although other odd-chain fatty acids have been used including C13:0, C15:0, C19:0, C21:0 and C23:0. The choice depends on the fatty acids found in the food and the separation of the internal standard from the sample fatty acids. It is preferable to add the internal standard as the TAG if this is the main lipid class, rather than as a free fatty acid. There has been some discussion on the choice of internal standards when the food is rich in long-chain PUFA with Ackman (24) arguing for the use of a standard with a similar retention time (e.g. C23:0) as the fatty acids of interest.
The separation of the FAME is best achieved using polar capillary columns, with a wide range of columns of different lengths, internal diameters and phase thicknesses being commercially available. Columns designed for optimal separation of FAME are available (e.g. BPX70a, CPSil 88a, SP-2560a and Omegawaxa) and the choice depends on the specific separations required. The separation of cis- and trans-monounsaturated FAME as found in margarines can be achieved successfully on 50–100m columns of BPX70 (25) (Figure 1) and SP2560a (26). Standards for determination of retention times are available commercially with a wide range of mixtures of known composition being available from Nu Chek Prep. Inc. (Elysian, MN). Craske and Bannon (27), amplifying the work of Ackman and Sipos (28) demonstrated that, for a number of common saturated and unsaturated fatty acids, the flame ionization detector theoretical response factors can be calculated from the content of carbon bonded to hydrogen in the molecule. However, primary standards are still essential to determine that both the gas chromatographic parameters and the analyst's manipulative skills are efficiently optimized. Assistance in the confirmation of the identity of unknown FAME can be gained by using non-polar capillary columns (e.g.. methyl silicone) since the unsaturated FAME elute before the saturated FAME using these phases (4). The use of GC mass spectrometry (MS) to confirm the identity of FAME is becoming more common with the wider access to modestly priced instruments. The use of picolinyl esters (13-OH methyl pyridine) of fatty acids for the GC-MS has been recommended for FAME with unusual structures (29).
• Determination of Sterol Composition and Content in Foods
The main sterol of interest from a nutritional viewpoint in foods is cholesterol and it is now widely acknowledged that the earlier spectrophotometric methods employed for the estimation of the cholesterol content of foods led to an overestimate of the true amount present (30). The general approach to the analysis of food sterols is to saponify the lipids, isolate the non-saponifiable fraction and analyze the sterols by GC or LC using appropriate internal standards. Cholesterol can also be determined enzymatically (31) using kit methods which are widely available.
In a review of the analysis of sterols by GC or LC (30), details of the direct saponification methods for the analysis of food sterols are discussed. The advantages include reduced solvent volumes and sample preparation times, excellent recoveries compared with the standard AOAC method (32), and application of this method to the analysis of eggs, meats and milk products. A disadvantage could be the very small analytical portion weight which is likely to be problematic with heterogeneous foods (e.g. in the case of eggs, the sample weight amounts to only 200 mg). The compounds present in the non-saponifiable fraction of foods, apart from cholesterol, include plant sterols, tocopherols and tocotrienols, cholesterol oxides and other hydrocarbons such as squalene (30).
A common approach to the separation of sterols in foods is the use of capillary GC with non-polar columns and using a GC with a flame ionization detector. Internal standards of 5 α-cholestane or 5 α-cholestanol have been used, however it is more difficult to separate cholesterol and 5 α-cholestanol, compared with 5 α-cholestane and cholesterol. Separation of the sterols as trimethyl silyl (TMS) ether derivatives has been the preferred method since it is regarded that this improves peak shape, decreases retention times and improves sensitivity, however problems associated with the use of TMS derivatives include hydrolysis and the fact that the reagents are toxic, flammable and corrosive (30). With the development of inert fused silica capillary GC columns, many workers no longer derivatize sterol samples prior to GC analysis (30). LC has been used to separate food sterols, however, while cholesterol does not have a strong absorption in the UV region, absorption at 205 nm can be used (30).
Fig 1 (b)
Figure 1. Gas chromatograms on 50 metre × 0.22mm BPX70 column of (a) margarine and (b) butter showing separation of the main trans 18:1 and cis 18:1 positional isomers (adapted from Mansour & Sinclair, 25)
• The future
Cardiovascular disease is still the leading cause of death in most industrialized societies. Recent research into dietary aspects of cardiovascular disease have indicated that it will no longer be sufficient for food databases to have information on the total fat, cholesterol and major fatty acids types of different foods. Because of the increasing interest in oxidation of low density lipoproteins (33), oxidation of cholesterol in food (34, 35), the role of trans fatty acids in lipoprotein metabolism (36) and the beneficial effects of the ω-3 PUFA on various aspects of cell metabolism (37), in the future food databases will require detailed information on the tocopherol and tocotrienol isomers, cis and trans fatty acid isomers, cholesterol oxides and other oxidation products of food lipids, and molecular species and positional distribution of fatty acids in TAG (21–23). Another issue which is likely to emerge as a challenge for analysts in the near future will be the development of standard methods for the wide range of different fat substitutes which have been developed in recent years (38).
(1) Vlieg, P., Body, D.R., & Burlingame, B. (1991) J.N.Z. Diet. Assoc. 45, 29–30
(2) Christie, W.W. (1982) Lipid Analysis, 2nd Ed., Pergamon Press, New York, NY
(3) Christie, W.W. (1987) High Performance Liquid Chromatography and Lipids, Pergamon Press, New York, NY
(4) Christie, W.W. (1989) Gas Chromatography and Lipids, The Oily Press, Ayr, Scotland
(5) Analysis of Fats, Oils & Derivatives (1993) E.G. Perkins (Ed.), AOCS Press, Champaign, IL
(6) Sahasrabudhe, M.R., & Smallbone, B.W. (1983) J. Am. Oil Chem. Soc. 60, 801–805
(7) Lucas, C.C., & Ridout, J.H. (1970) Prog. Chem. Fats 10, 1–150
(8) Morrison, W.R., Tan, S.L., & Hargin K.D. (1980) J. Sci. Food Agric. 31, 329–340
(9) Rouser, G., Kritchevsky, G., & Yamamoto, A. (1967) in Lipid Chromatographic Analysis, Vol 1, G.V. Marinetti (Ed.), Edward Arnold Ltd., London, pp.99–162
(10) Marmer, W.N., & Maxwell, R.J. (1981) Lipids 16, 365–371
(11) Dunstan, G.A., Volkman, J.K., & Barrett, S.M. (1993) Lipids 28, 937–944
(12) Folch, J., Lees, M., & Sloane-Stanley, G.H.S. (1957) J. Biol. Chem. 226, 497–509
(13) Bligh, E.G., & Dyer, W. (1959) Can. J. Biochem. Physiol. 37, 911– 917
(14) King, J.W. (1993) INFORM (J. Am. Oil Chem. Soc.) 4 1089–1098
(15) Phillips, F.C., & Privett, O.S. (1979) Lipids 14, 949–952
(16) Nelson, G.J. (1993) in Analysis of Fats, Oils & Derivatives, E.G. Perkins (Ed.), AOCS Press, Champaign, IL, pp. 20–59
(17) Weirauch, J.L., Posati, L.P., Anderson, B.A., & Exler, J. (1977) J. Am. Oil Chem. Soc. 54, 36–40
(18) Mann, N.J., Sinclair, A.J., Watson, M.J., & O'Dea, K. (1991) Food Aust. 43, 67–69
(19) Watson, M.J., Mann, N.J., Sinclair, A.J., & O'Dea, K. (1992) Food Aust. 44, 511–514
(20) Fewster, M.E., Burns, B.J., & Mead, J.F. (1969) J. Chromat. 43, 120–126
(21) Sinclair, A.J. (1993) Food Aust. 45, 226–231
(22) Kritchevsky, D. (1988) Nutr. Rev. 46, 177–181
(23) Currie, G.J., & Kallio, H. (1993) Lipids 28, 217–222
(24) Ackman, R.G. (1991) Lipids 27, 858–862
(25) Mansour, M.P., & Sinclair, A.J. (1993) Asia Pacific J. Clin. Nutr. 3, 155–163
(26) Firestone, F., & Sheppard, A. (1992) in Advances in Lipid Methodology, Vol. 1, W.W. Christie (Ed.), The Oily Press, Ayr, Scotland, pp. 273–322
(27) Craske, J.D., & Bannon, C.D. (1987) J. Am. Oil Chem. Soc. 64, 1413–1417
(28) Ackman, R.G., & Sipos, J.C. (1964) J. Am. Oil Chem. Soc. 41, 377–383
(29) Harvey, D.J. (1992) in Advances in Lipid Methodology, Vol. 1, W.W. Christie (Ed.), The Oily Press, Ayr, Scotland, pp. 19–80
(30) Fenton, M. (1992) J. Chromat. 624, 369–388
(31) Shen, C.J., Chen I.S., & Sheppard, A.J. (1982) J. Assoc. Off. Anal. Chem. 65, 1222–1224
(32) Official Methods of Analysis (1980) 13th Ed., AOAC, Washington, DC, sec. 43.235
(33) Steinberg, D., Parthasarathy, S., Carew, T.E., Khoo, J.C., & Witztum, J.L. (1989) New Eng. J. Med. 320, 915–924
(34) Chisolm, G.M. (1991) Current Opinion in Lipidology. 2, 311–316
(35) Sarantinos, J., O'Dea, K., & Sinclair, A.J. (1993) Food Aust. 45, 485–490
(36) Mensink, R.P., Zock, P.L., Katan, M.B. & Hornstra, G. (1992) J. Lipid Res. 33, 1493–1501
(37) Sinclair, A.J. (1993) Asean Food J. 8, 3–13
(38) Haumann, B.F. (1993) INFORM (J. Am. Oil Chem. Soc.) 4, 1227–1235
David A. T. Southgate
formerly of AFRC Institute of Food Research, Norwich Laboratory, Colney, Norwich, NR4 7LY, UK
The modes of expression and the conventions used in citing analytical data in nurtritional databases are critical for the accurate use of the database. This is especially true for the nutritionist who wishes to use data from a number of different databases or who wishes to merge data from databases in different countries for a specific study. While international agreement on the modes of expressions and conventions is the preferred approach, at the very least it is essential that the documentation of all databases describes in detail the modes of expression and conventions explicitly so that users will know where data are compatible. The major conventions in use are discussed and proposals made for establishing some common positions in the preparation of nutritional databases.
The nomenclature, conventions and modes of expression used to describe the nutrient values have a profound influence on the accurate use of nutritional databases. They are especially important when one is working using a series of different databases. Many nutritional epidemiological studies are being made internationally. For such studies it is often necessary to construct a special database from a number of different sources, and it becomes critically important to ensure that compatible data are being combined. This involves consideration of, first, the compatibility of the analytical methods used to generate the data (1) and, second, that the modes of expression, and for many nutrient values, the conventions used in deriving the values, are also compatible.
Where the analytical methods for a nutrient are specific and the range of methods in use is known to produce comparable values then expression on an appropriate weight basis is straight-forward (2).
There are three major categories of nutrients where the conventions and modes of expression are a major cause of incompatibility. First, where the nutrient values are calculated from some other analytical measurement. Second, where the nutrient is a complex mixture and some analytical compromises have been adopted in the measurement of the mixture. The third category includes those vitamins for which there are a number of active forms (vitamers) which differ in biological activity.
• Nutrient Values Calculated from Other Analytical Measurements
These cover two of the most important conventions used in the expression of nutrient values in databases; the expression of values for energy and protein.
Energy Values of Foods
In nutritional databases (and in general nutritional usage) the energy values of foods are, strictly, the “metabolizable energy values” in other words, the energy that is available for use by the body. In formal terms metabolizable energy of the dietary intake is equal to the gross (the heat of combustion) energy intake minus the energy lost in feces and urine (other losses should also be included, for example, energy lost in other secretions or gases but in human nutrition these are customarily discounted). Metabolizable energy values thus are, in strict terms, an attribute of the dietary intake and the calculation of values for foods is a pragmatic approximation.
Metabolizable values for foods are calculated using energy conversion factors applied to the amounts of protein, fat, carbohydrates and alcohol in the food.
The energy conversion factors in use in most databases are based on the studies made by Atwater and his colleagues in the early years of this century and it is a tribute to their experimental skills that the system, despite the assumptions which had to be made at that time, remains a practical approximation (3). The Atwater factors adjust the protein, fat and carbohydrate heats of combustion to allow for the fecal losses of these constituents by multiplying the average heats of combustion for mixed proteins, fats and carbohydrates, by the respective apparent digestibilities of the three components. The energy loss in urine was allowed for by correcting the protein value for the energy per unit nitrogen lost in urine.
These calculations give the familiar, 4, 9, 4 factors (kcal per g of protein, fat, and carbohydrate, respectively), which are still widely used.
There are two other conventions used in nutritional databases.
Specific Energy Conversion Factors. This was originally developed by Atwater but not used by him. It was later advocated by Merrill and Watt (4). This convention is based on the premise that it would be better to use heats of combustion values specific for the different foods (or food groups) and to use specific apparent digestibility values for the constituents of each food (or group), rather than to use average values derived for mixed diets. This system depends on having experimental data for the heats of combustion for all components of foods and extensive data from human metabolic studies, and at the present time relies heavily on early data, some from Atwater himself. It is important that databases using this system document, precisely, which factors have been used for specific foods as the published account can be interpreted in different ways
Modified Atwater System Used in the UK Nutritional Database. In the United Kingdom a different system was developed by McCance and Widdowson (5). The need for this arose because these authors measured carbohydrates directly (as opposed to Atwater's use of the “by difference” method) and furthermore, they divided the food carbohydrates into two categories; “available” the digestible sugars and starches, which were glucogenic in man and “unavailable”, those carbohydrates not digested in the small intestine and not providing the body with absorbable carbohydrates. For the available carbohydrates they assigned the energy conversion factor 3.75 kcal per g (this is the heat of combustion of monosaccharides) because available carbohydrates were expressed in this way. Unavailable carbohydrates were assigned a zero energy value, not because they did not provide energy but because these components were known to reduce the apparent digestibility of proteins and fats and therefore any energy from the short chain fatty acids produced by fermentation in the large bowel was discounted.
Detailed evaluation of this approach in experimental studies on a large number of subjects showed that it gave a good prediction of metabolizable energy intakes (6). The UK system gives lower energy values for plant foods that are rich in plant cell wall material (and incidentally organic acids).
Comparisons of the Three Conventions. In practical terms the three conventions give similar values. The Atwater system tends to give over-estimates of metabolizable energy and the UK system under-estimates energy intakes at very high unavailable carbohydrate intakes because of the increasing importance of the energy from the fermentation products of the unavailable carbohydrates. It has been suggested (7) that a radically different system based on measured heats of combustion of foods would be intellectually more satisfying but in practical terms the differences produced by the three systems are much smaller than the errors inherent in ignoring individual differences in digestibility and the errors in measuring food intakes. Provided that it is recognized that these conventions are approximations and energy values are not cited to four significant figures the practical incompatibilities are not significant (2).
Virtually all nutritional databases give values for protein that are derived by calculation from measured total nitrogen values. These have been customarily measured by the Kjeldahl method. These calculated values should be called “crude protein” and recognized as a approximations and not as estimates of protein in a biochemical sense. The factors used for calculating protein are based on the percentage of nitrogen in a typical proteins. Thus 6.25 is appropriate for proteins with 16 g N per 100 g protein.
The FAO expert report on protein requirements in 1973 (8) recommended that different factors should be used for calculating “crude” protein values for different food groups to take account of the N-content of different food proteins and this approach has been followed in the UK nutritional database (5, 13). When the protein and energy requirements were reviewed again in 1985 (9), it was clear that the use of these factors produced anomalies when making recommendations for protein intakes. This arises because all the experimental studies on protein requirements have been based on the measurement of nitrogen metabolism and it has been argued that it would be better to avoid confusion by using only 6.25 as a conversion factor.
In nutritional terms it would be preferable to move to more biochemically coherent measures where protein values were based on estimates of amino acid nitrogen content using amino acid composition data. The pratice of correcting for non-protein, non-amino acid nitrogen is desirable for some foods especially cartilaginous fishes and some fungi in order to avoid over-estimating the protein in these foods.
As a general principle it is desirable that all databases include total nitrogen values so that when merging data a consistent approach to calculating crude protein values can be adopted. It is important to ensure that users do not equate (N × a factor) with a biochemical concept of protein as functional polymers of amino acids and to recognize the conventional approximation.
• Complex Nutrients Involving Analytical Compromises
The nutrients which fall into this category include fat, the carbohydrates and the folates. The essential difficulties arise because these nutrients are very complex mixtures the determination of which requires time-consuming separations which are not widely used in routine food analysis. This means that most of the data available to database compilers is based on simple methods and the values obtained tend to be method-dependent.
The values for fat in most databases are for “total lipid-solvent soluble material”, which includes the triacylglycerol compounds, sterols and depending on the method used, phospholipds. Several of the methods in common use do not extract the lipids efficiently from some food matrices and the values cited in databases may be incompatible because of this method-dependence. In the compilation of databases preference should be given to values obtained using methods that give complete extraction such as those used to prepare extracts for fatty acid analysis (2) with all values documented by method.
The main difficulties arise from the incompatibility between carbohydrates values obtained “by difference” and by direct analysis. Direct analysis provides values for the different classes of sugars, mono-, di- and oligo-saccharides starch and non-starch polysaccharides (the major components of dietary fiber) (10). The sum of the individual components approximates to the total carbohydrate “by difference” if the food does not contain substantial amounts of non-carbohydrate components. It is important to recognize that “by difference” values also include errors in the measurement of water content, ash, protein and total lipid.
In the UK database the available carbohydrates are expressed as monosaccharides, which means that the summation of the values will exceed the total carbohydrate “by difference” by a significant amount especially in starch-rich foods because of the addition of water on hydrolysis of disaccharides and polysaccharides.
The ideal from the analytical point of view is to move away from the technically obsolete “by difference” method; but at the present time it is essential to identify the method used when giving carbohydrate values. This is especially true for dietary fiber where the methods in use measure conceptually different fractions. Thus dietary fiber values should be cited (measured as total dietary fiber, AOAC, or as non-starch polysaccharides etc.).
A large number of folate vitamers exist in foods. The biological availability of them is an active topic of research and there are chromatographic methods for the separation of the different forms, although these are technically exacting and not suitable for routine use at the present time. Most values in nutritional databases are based on microbiological determinations using either Streptococcus faecalis or Lactobacillus caseii (rhamnosus). Deconjugation of the polyglutamyl chain has to be used to measure total folates. At one time it was assumed that a value measured before deconjugation represented free monoglutamyl folates but is now clear that this value includes contributions from the polyglutamyl forms and has no value for the nutritional classification of the folates. S. faecalis does not respond to some folates and the values obtained are significantly lower than when using L. rhamnosus. It is thus critically important to identify the organism used in the assay when citing folate values (11).
• Assigning Biological Activity to Different Vitamers
This concerns those vitamins where there are a range of vitamers in foods which have differing biological activities and where it is customary in nutritional databases to give some kind of aggregated value for the vitamin activity of the food. The major vitamin which falls into this category is vitamin A, where it has become accepted to give a vitamin A value expressed as micrograms retinol equivalents, based on summation of the amounts of retinol and the activity of different carotenoids. An analogous convention has been proposed for vitamin E where the aggregated α-tocopherol equivalents are being estimated by summing the weighted values of the different tocopherols and tocotrienols.
Vitamin A Activity
The accepted convention is to take the values for the provitamin carotenoids and divide by factors which reflect the efficiency of conversion to retinol. These conversion factors are based on studies with mixed diets and there is growing evidence that the efficiency of conversion is profoundly affected by the amount of fat in the diet and on the structure of the food matrix, so that under some conditions the carotenoids may not contribute to vitamin A status. The convention is to divide β-carotene values by 6 and other carotenoids by 12 to convert them to retinol equivalents. Some authors argue that the conversion efficiency of the β-carotene in dairy foods is under-estimated by the use of 6 and suggest 3 but the evidence for this is not totally convincing (12).
At the present time there is growing evidence that the carotenoids are important nutrients in addition to their role as provitamin A compounds and this suggests that databases should firstly, give values for retinol and the different carotenoids separately. If a total vitamin A activity is deemed useful in addition, the convention used must be described. Such an approach permits users to assign their own convention and to use different conventions as understanding of the factors determining the efficiency of conversion improves.
Vitamin E Activity
The biological activity of the tocopherols and tocotrienols vitamers is related in this convention to the activity of α-tocopherol. The biological evidence for calculating the biological activity factors is limited but the approach is being used in the UK database (13).
Once again it is desirable that the databases contain the values for the different forms so that users can apply their own consistent conversion conventions. Should the compilers wish, in addition, to give a total vitamin E activity value in the database then the convention must be documented.
• Development of Agreed Conventions and Modes of Expression
The interchange systems developed by INFOODS (14) allow for the documentation of data values within the exchange tags which incorporate units and details of method where values are method-dependent. In the context of this paper “method-dependence” is restricted to those instances where different methods give different values, not to differences arising because of differences in the performance of the method. Values that have been calculated by conventions, such as energy and protein are similarly assigned different tags where this is appropriate. One has therefore within the INFOODS interchange scheme a system that should prevent the aggregation of incompatible data. The primary need at present for those involved in merging data from databases where different conventions have been used is to resolve the differences so that they can bring the values into a consistent compatible form. This depends on having access to complete documentation of all databases, and at the primary data source level, avoiding any tendency to aggregate original analytical data.
Thus if a database has values for protein, fat and carbohydrate (however expressed) it is possible to recalculate energy values de novo to avoid incompatibilities. Similarly total nitrogen values give the possibility of calculating protein values using either one or several factors.
At the present time some carbohydrate values are incompatible without more detailed analytical information than is currently available. This is especially true for dietary fiber values because the crude fiber values still given in some databases are incompatible with dietary fiber values. In principle TDF values should be higher than NSP values but taken across the whole range of plant foods resistant starch may be of the order of 3–6 g per day and lignin intakes rarely exceed 1 g per day so in practical terms great errors will not flow from combining these two types of values.
In the case of the vitamins the biological equivalents will undoubtedly change as more evidence emerges and it is essential to accumulate and give the actual analytical values for the different vitamers in databases to permit reappraisal in the future.
Finally documentation as part of the database should be seen as the ideal. We really should be aiming for imbedded information (15) so that the user can easily establish what a nutrient name implies and where analytical methods may make the values method-dependent.
Conventions should be recognized as such and the values they provide recognized as approximations of biological phenomena and treated as one would treat all derived values, with caution.
(1) Southgate, D.A.T. (1985) Ann. Nutr. Metab. 29, Suppl., 49–53
(2) Greenfield, H., & Southgate, D.A.T. (1992) Food Composition Data: Production, Management and Use, Elsevier Applied Science, London
(3) Allison, R.G., & Senti, F.R. (1983) A Perspective on the Application of the Atwater System of Food Energy Assessment, Federation of American Societies for Experimental Biology, Bethesda, MD
(4) Merrill, A.L., & Watt, B.K. (1955) Energy Value of Foods, US Dept of Agriculture, Handbook 74, USDA, Washington, DC
(5) Paul, A.A., & Southgate, D.A.T (1978) McCance and Widdowson's The Composition of Foods, 4th Ed., HMSO, London
(6) Southgate, D.A.T., & Durnin, J.V.G.A. (1970) Br. J. Nutr. 24, 517–535.
(7) Livesey, G. (1991) Proc. Nutr. Soc. Aust. 16, 79–87
(8) FAO/WHO Expert Group (1973) FAO Nutrition Series, No. 52, FAO, Rome
(9) FAO/WHO/UNU (1985). WHO Technical Report Series 724, WHO, Geneva
(10) Southgate, D.A.T (1991) Determination of Food Carbohydrates, 2nd Ed., Elsevier Applied Science, London
(11) Finglas, P.M., Faure, U., & Southgate, D.A.T. (1993) Food Chem. 46, 199–213
(12) WHO (1967) WHO Technical Report Series 362, WHO, Geneva
(13) Holland, B., Welch, A.A., Unwin, I.D., Buss, D.H., Paul, A.A., & Southgate, D.A.T. (1991) McCance and Widdowson's The Composition of Food, 5th Ed., Royal Society of Chemistry, Cambridge
(14) Klensin, J.C., Feskanich, D., Lin, V., Truswell, A.S., & Southgate, D.A.T. (1989) Identification of Food Components for INFOODS Data Interchange, UNU Press, Tokyo
(15) Southgate, D.A.T. (1992) in The Contribution of Nutrition to Human and Animal Health, E.M. Widdowson & J.C. Mathers, (Eds.), Cambridge University Press, Cambridge, pp. 369–378