Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Food science

Food fortification in the English-speaking Caribbean


William K. Simmons



This study was designed to collect and analyse information on food fortification in the Caribbean and its possible modification to improve the health of West Indians. Foods already fortified and imported into the Caribbean are not included except for pre-ground flour. The data were obtained by visiting flour mills and supermarkets and by discussing fortification programmes with health and agriculture officials All 17 member countries of tile Caribbean Food and Nutrition Institute and the Pan American Health Organization (CFNI/PAHO) were included in the study.



The purpose of food fortification is to maintain the health of individuals in a group or population by providing adequate levels of nutrient intake. While such a programme may be implemented primarily to combat a nutritional problem apparent in a particular target group, it may also have benefits outside the immediate target group, depending on the nutrient intake of other groups involved.

The potential benefits of fortification are greatest where the total amount of food is reasonably adequate but where the nature of the food supply is such that one or more nutrients are in limited supply (or where environmental factors affect requirements for particular nutrients). In this situation, a prime consideration is whether a change in the nature of the foods consumed, an improvement in their nutritional quality (through fortification), or a combination of these is most desirable. In any event, one of the primary goals of fortification is to provide the additional nutrients required to upgrade a population's diet, rather than to emphasize improvement of individual foods.

No study has ever compiled information on food fortification in the Caribbean and its possible modification to improve the health of West Indians. This study was done to collect and analyse such information. Foods already fortified and imported into the Caribbean are not included except for pre-ground wheat flour.

All 17 member countries and territories of the Caribbean Food and Nutrition Institute and the Pan American Health Organization (CFNI/PAHO) were chosen for the study. The author visited most of these countries. The information was obtained by visiting flour mills and supermarkets and by discussing fortification programmes with health and agriculture officials. Information from three of the countries was obtained by correspondence.

The author is a Public Health Nutritionist with the Caribbean Food and Nutrition Institute in Kingston, Jamaica.



Table 1 shows the programmes for the fortification of food products with vitamins and minerals operating at present in the CFNI member countries. Only four of the thirteen smaller countries and territories have such programmes: Barbados, Belize, Grenada, and Saint Vincent. Several foodstuffs are fortified in Barbados and Saint Vincent; in Belize and Grenada only wheat flour is fortified.

Three of the four larger countries have fortification programmes. A variety of foodstuffs are fortified in Jamaica and in Trinidad and Tobago, and wheat flour is fortified in Guyana. There is no fortification scheme in Suriname at present.

Possibilities for modifications or additions to these fortification programmes are listed in table 2. Among the smaller countries, Barbados and Saint Vincent each have a number of food processing facilities where more or different vitamins and minerals could be added to the various foods; Antigua and Saint Christopher-Nevis each have a central sugar processing plant, and Belize and Grenada each have a central wheat flour mill. In the remaining smaller countries and territories any fortification programme would be difficult because there are few facilities for food processing.

Of the larger countries, Jamaica, Guyana, and Trinidad and Tobago all have many food processing plants. In Suriname there is a central wheat flour mill, and a central rice flour mill may be opened soon.

With the exception of Guyana, where rice consumption is highest, wheat flour and wheat products are the principal foodstuffs in the Caribbean. As can be seen in figure 1, wheat and wheat flour are imported from the Federal Republic of Germany, the United States, Canada, and Puerto Rico. Some of the pre-ground wheat flour imported to the Caribbean is fortified with vitamins and iron.

All flour ground locally in the Caribbean from imported wheat is fortified with vitamins and iron except counter flour (coarse flour) ground in Grenada and all flour ground in Suriname. By law wheat flour must be fortified with iron and the B complex vitamins in Jamaica, Trinidad and Tobago, and Guyana. This fortification is usually at the following levels:

Fortification with calcium is optional and is usually not done.

The iron added to both pre-ground and locally ground wheat is in the form either of ferrous sulphate or of reduced iron, the latter usually in particles 44 microns in size. Ferrous sulphate has a higher bioavailability than does reduced iron of such large particle size. In one of the flour mills reduced iron with smaller particles is being used, which should be more available.

Wheat for baking flour is imported into Jamaica and ground at the Jamaica Flour Mills. This baking flour accounts for about 40% of the total consumption of wheat flour in the country, and its iron source is usually reduced iron of large particle size. In the past, counter flour, which makes up about 60% of the total flour consumed, was imported directly and was sometimes fortified with iron and B vitamins but usually not. Counter flour is the main type consumed by the lower socio-economic class.

Since July 1984 all baking flour and counter flour have been milled in Jamaica by the Jamaica Flour Mills. Both types of flour are fortified at the following levels:

TABLE 1. Present fortification programmes in the CFNI member countries

  Food fortified Nutrients added Amount

Smaller countries and territoriesa

Barbados wheat flour thiamine 4.1
    riboflavin 2.5
    niacin 30.1
    iron (reduced) 26.4
  orange juice vitamin C b
  evaporated milk vitamins A and D b
  margarine vitamins A and D b
Belize wheat flour thiamine 4.1
    riboflavin 2.5
    niacin 30.1
    iron (reduced, or Fe sulphate) 26.4
Grenada wheat flour (baking flour only) thiamine 4.1
    riboflavin 2.5
    niacin 30.1
    iron (reduced) 26.4
Saint Vincent wheat flour thiamine 4.1
    riboflavin 2.5
    niacin 30.1
    iron (reduced) 26.4
  milk. reconstituted, full-cream vitamins A, C, and D b
  chocolate milk vitamins A, C, and D b
  orange juice vitamin C c
  grapefruit juice vitamin C c
  snack pack (blend of tropical juices) vitamin C c

Larger countriesd

Jamaica wheat flour (baking and counter thiamine 6.3
  flour) riboflavin 3.9
    niacin 52.8
    iron (reduced) 44
  salt potassium iodide b
    fluoride 250
  Foska oats calcium 2,112
    phosphorus (calcium phosphate) 1,760
  condensed milk. sweetened (Nestle) vitamin A 2,133
    vitamin D 24.6
    thiamine 14.0
  milk, reconstituted homogenized, vitamins A and D b
Guyana wheat flour thiamine 4.1
    riboflavin 2.5
    niacin 30.1
    iron (reduced, or Fe sulphate) 26.4
Trinidad and wheat flour thiamine 4.1
Tobago   riboflavin 2.5
    niacin 30.1
    iron (Fe sulphate) 26.4
    calcium 1,100
  pineapple juice, canned vitamin C b
  grapefruit juice, canned vitamin C b
  cold drinks (sorrel. orange. passion vitamin C b
  fruit. grapefruit)    
  condensed milk. full-cream. vitamin A 2,133
  sweetened vitamin D 29.6
    thiamine 4.9
  milk powder, full-cream vitamin D 1,212
  margarine vitamins A and D b

a. There are no programmes in Antigua. the Bahamas. the British Virgin Islands. the Cayman Islands, Dominica. Montserrat, Saint Lucia, Saint Christopher-Nevis, or the Turks and Caicos Islands.
b. Amounts not given.
c. Minimum 30 mg per 0 25 litre.
d. There is no programme in Suriname.

Iron is added in the reduced form of large particle size and provides a 70% increase in the iron content of the flour. The use of electrolytically reduced iron is being considered.



Protein-energy malnutrition in the English-speaking Caribbean, although on the decline during the last two decades, remains a public health problem. From 2% to 15% of children were below 80% of the NCHS/ WHO standard in 12 Caribbean countries. In another country 24% of children were reported malnourished by this criterion. This malnutrition is found largely in specific geographic pockets in most countries [1].

In addition, a large proportion of the population in all age groups suffer from varying degrees of anaemia, which available evidence suggests is caused primarily by a deficiency of iron. Deficiencies in folate and vitamin B12 have been noted occasionally in Trinidad and Tobago and Guyana among nationals of East Indian origin, but their prevalence is much less than that of iron deficiency [2-13].

In a recent island-wide survey in Grenada, anaemia was found in all age groups in the population. Sixty-two per cent of pregnant women had hemoglobin levels below the WHO recommended haemoglobin standard. The high prevalence of iron deficiency as exhibited by low plasma ferritin levels was even more striking; it too was found in all age groups but was particularly high in pregnant and lactating women and pre-school children. Sixty per cent of pregnant women, 53.6% of lactating women, and 61.7% of preschool children had ferritin levels below 12 microg/litre, indicating no iron reserves [14]. In another study, conducted in 1989 on an eastern Caribbean island, 48% of pregnant women had plasma ferritin levels below 12 ,microg/litre (C. Ferris, personal communication), which again represents iron deficiency as a public health problem in the Caribbean.

TABLE 2. Possibilities for modifications or additions to fortification programmes

  Possible modifications

Smaller countries and territories

Antigua a central sugar processing plant
Bahamas difficult
Barbados a central wheat flour mill
2 cornmeal processing plants
5 sugar processing plants
2 macaroni and spaghetti pro- cessing plants
Belize a central wheat flour mill
British Virgin Islands difficult
Cayman Islands difficult
Dominica difficult
Grenada a central wheat flour mill
Montserrat difficult
Saint Lucia difficult
Saint Christopher- Nevis a central sugar processing plant
Saint Vincent a central wheat flour mill
5 arrowroot processing plants
a central sugar processing plant
a central milk processing plant
Turks and Caicos Islands difficult

Larger countries

Jamaica a central wheat flour mill
a central cornmeal processing plant
a central salt processing plant
11 sugar processing plants
many food processing plants
Guyana a central wheat flour mill
11 sugar processing plants
a food processing plant
rice mills-polished rice, rice flour
Suriname a central wheat flour mill
a possible future rice flour mill
Trinidad and Tobago a central wheat flour mill
2 sugar processing plants
many food processing plants
a central milk processing plant
soft drink plants

Vitamin-A deficiency severe enough to cause eye lesions is extremely rare. Among the B vitamins, only riboflavin shows evidence of clinical deficiency [15]. Scurvy (vitamin-C deficiency) and rickets (vitamin-D deficiency) are virtually unknown [1]. The fortification of cow's milk with vitamin D has been credited as a major factor in removing infantile rickets as a public health problem. Non-fat milk often has added vitamins A and D. Most milk and milk products in the Caribbean, including butter, as well as margarine, are fortified with vitamins A and D.

There is a high prevalence of dental caries in some Caribbean countries. A survey conducted in 1984 showed a very high prevalence of caries, according to the classification established by WHO, among Jamaican children [16]. Approximately 60 countries throughout the world add fluoride to either drinking water or salt. In September 1987 the government of Jamaica started adding fluoride as well as iodine to all salt in the country. The level was 250 mg of fluoride per kilogram of salt, which increased the cost of salt to the consumer by four cents per kilogram. There is one salt factory in Jamaica and some of this fortified salt is exported to Belize, Grenada, Saint Vincent and the Grenadines, Barbados, Trinidad and Tobago, and Saint Lucia.

Iron has been added to fish sauce, sugar, salt, pasta, cornmeal, infant foods, breakfast cereals, and MSG, but it is most widely used to fortify wheat flour [17; 18]. The technology for adding iron to cereal foods (generally as a component of a vitamin-mineral blend) has been well developed over the past 40 years and has been greatly facilitated by the fact that these foods are centrally processed in a manner that easily lends itself to control of the addition of nutrients. There still remain questions as to which form of iron can and should be used by the manufacturers in various applications. The processors primarily use reduced iron or ferrous sulphate; some use insoluble phosphates such as ferric orthophosphate or sodium-iron pyrophosphate.

Reduced iron can be produced by either hydrogen or electrolytic reduction of iron oxide. Its bioavailability and chemical reactivity depend on the particle size. The smaller particles are better absorbed. One form of reduced iron is the electrolytically reduced type that has a very fine particle size (20 microns) and good bioavailability.

Hydrogen-reduced iron is reduced mechanically to a particle size of about 44 microns. This large-particle iron has poor bioavailability in comparison to that of ferrous sulphate or electrolytically reduced iron [19]. However, it is what is usually used to fortify foodstuffs because it is one of the least expensive forms of iron available.

FIG. 1. The fortification of wheat flour in the English-speaking Caribbean (January 1989)

Ferrous sulphate is widely used for food fortification when soluble iron is considered necessary and also for therapeutic doses in the treatment of acute iron-deficiency anaemia. Iron sulphate tends to cause rancidity and discoloration in flour when used at high levels or when the flour is stored for a long period under high humidity. However, it is inexpensive, has excellent bioavailability, and is often used for fortification of wheat flour.

Another type of iron that could be used is iron sodium EDTA. Experiments so far have shown that this compound has an absorption rate similar to that of haemoglobin iron in both animals and pre-school children. This iron does not seem to enter the inorganic iron pool of the diet, and therefore its availability is not influenced by inhibiting substances in food. Although it is much more expensive than ferrous sulphate or reduced iron, this is compensated for by its higher availability.

Since iron deficiency anaemia is a major public health problem, programmes for changing the iron status of West Indians should receive attention. A programme that would have wide usefulness would be to change the type and level of iron added to wheat flour. In Jamaica the level of iron has been increased from 26 to 44 mg per kilogram of wheat flour. Careful consideration should be given to changing the level of iron at other flour mills in the Caribbean. Given the high level of iron deficiency in the Caribbean countries and the fact that wheat flour is the staple food of the population, such a programme should merit high priority.



1. Sinha D. Children of the Caribbean. Kingston, Jamaica: CFNI, 1989.

2. Simmons WK. Nutritional anaemia in Jamaica. West Ind Med J 1979;28:199-207.

3. Simmons WK, Sinha D. Anaemia in the Cayman Islands: its prevalence and control. (CFNI-J-24-80) Kingston, Jamaica: CFNI, 1980.

4. Simmons WK, Jutsum PJ, Fox K et al. A survey of the anaemia status of pre-school age children and pregnant and lactating women in Jamaica. Am J Clin Nutr 1982;35:319-26.

5. Simmons WK, Gurney JM. Nutritional anaemia in the English-speaking Caribbean and Suriname. Am J Clin Nutr 1982;35:327-37.

6. Simmons WK. Anaemia in the Caribbean: its prevalence and causes. (CFNI-J-10-83) Kingston, Jamaica: CFNI, 1983.

7. Serjeant GR. The clinical features of sickle cell disease. Amsterdam: North Holland Publishing Co, 1974.

8. Simmons WK, Gallagher P, Patterson AW. Anaemia in ante-natals in the Turks and Caicos Islands. West Ind Med J 1989;36:210-15.

9. Simmons WK. The control of anaemia in the English-speaking Caribbean. (CFNI-J-34-83) Kingston, Jamaica: CFNI, 1983.

10. Simmons WK. Programmes for the prevention of 17. anaemia in Jamaica. West Ind Med J 1980;29:1521.

11. Bramble D, Simmons WK. Anaemia in ante-natals in Montserrat. West Ind Med J 1984;33:92-96.

12. Simmons WK, Sinha D. Anaemia status and current methods for its control in Antigua. Kingston, Jamaica: CFNI, 1981.

13. Simmons WK. Nutritional anaemia in ante-natals in the English-speaking Caribbean. (CFNI-J-22-85) Kingston, Jamaica: CFNI. 1985.

14. Assessment of iron status of the Grenadian population. (CFNI-J-17-86) Kingston, Jamaica: CFNI, 1986.

15. Golden B, Ramdath D, Appleby J, Charley L, Golden MHN. Erythrocyte glutathione reductase activity and riboflavin status in severely malnourished children. Proceedings of the 32nd Scientific Meeting, CC MRC. West Ind Med J 1987;(suppl)36:31.

16. Warpa R. Dental caries and salt fluoridation. Assoc Gen Pract Jamaica Newsletter Dec 1985:6-7.

17. INACG. Guidelines for the eradication of iron deficiency anaemia. A report of the International Anaemia Consultative Group. Washington, DC, 1977.

18. WHO. Control of nutritional anaemia with special reference to iron deficiency. Technical Report Series, no. 508. Geneva: World Health Organization, 1975.

19. Shah BG, Belonje B. Bioavailability of reduced iron. Nutr Rep Int 1973;3:151-56.


Provitamin A determination problems and possible solutions


Delia B. Rodriguez-Amaya


Although the grave consequences of vitamin A deficiency have long been recognized and widely discussed, this serious public health problem continues to exist in many developing areas of the world. Experts agree that the long-term solution lies in increased dietary intake of this nutrient. Due to the much higher cost of animal foods, which provide preformed vitamin A (retinol, retinyl ester, retinal, 3-dehydroretinol, retinoic acid), developing countries rely largely on the provitamins of plant foods, which contribute about 82% of dietary vitamin A [1]. Thus, data on available and potential sources of vitamin A-active carotenoids is urgently needed to implement programmes and projects to alleviate the present situation and eventually eliminate the problem. Unfortunately, a good part of existing data on the provitamin A content of foods are now considered unreliable.

This review will discuss the major difficulties and suggest possible solutions in provitamin A determination, an admittedly complicated analysis. For a detailed discussion of this subject, the reader is referred to an earlier paper [2].

Accurate provitamin A data are not readily obtained because there are a large number of naturally occurring carotenoids, and foods vary markedly in their carotenoid composition, both qualitatively and quantitatively. In addition, not all carotenoids are precursors of vitamin A, and those which are vary in their biological activities. Thus, any method will have to be adapted to the sample being analysed, especially in relation to the extraction (e.q. sample size, volume of extracting solvent) and chromatographic steps. It is necessary to remove interfering inactive carotenoids and to separate and quantify the provitamins individually. Methods which do not make provision for this separation lead to over- or under-estimation.

This apparently confused picture could be somewhat simplified. From the known carotenoid distribution of foods, some patterns can be discerned which can facilitate the definition of the methodology, particularly the chromatographic separation. On the basis of the provitamin A composition, plant foods may be classified into three main groups: (1) those in which the vitamin A value is due almost exclusively to b-carotene (e.g. leafy vegetables, peas, broccoli, sweet potato, tomato, red-fleshed guava, watermelon, mango), (2) those in which a- and b-carotene are the major contributors (e.g. carrots, some varieties of squash, palm oil), and (3) those in which ,b-cryptoxanthin along with b-carotene primarily account for the vitamin A value (e.g. cashew-apple, peach, persimmon, loquat, Cyphomandra betacea), sometimes accompanied by appreciable amounts of 5,6- or 5,8-monoepoxy-b-cryptoxanthin (cryptoflavin) (e.g. papaya, loquat, Spondias lutea) and a-carotene. Other provitamins are also encountered but in amounts too small to affect the vitamin A values significantly [2].

The author is a professor in the Department of Food Science, Faculty of Food Engineering, Universidade Estadual de Campinas, in Campinas, Sao Paulo, Brazil.

This paper is reprinted from Food Laborarory News, no. 19 (March 1990).

Mention of the names of firms and commercial products does not imply endorsement by the United Nations University.

A further complication is the possibility of quantitative losses and artifact formation during analysis as a result of isomerization and oxidation of the highly unsaturated carotenoid molecule. Precautionary measures should therefore be standard practices in the laboratory. These include the use of distilled or analytical grade solvents free from damaging impurities (e.g. peroxide-free ethyl ether and tetrahydrofuran, acid-free chloroform), protection from light and high temperature (e.g. working under subdued light and heating avoided), short analysis time, maintenance of inert atmosphere (e.g. flushing with N2 and utilization of antioxidant). The last two precautions would increase the cost and may not be necessary if the analysis is accomplished within the shortest possible time. It is also a good policy to analyse fresh samples on arrival.

FIG. 1. Structures of common provitamins and interfering inactive carotenoids


Confirmation of the identity of provitamins

An obvious prerequisite to accurate quantitation is the conclusive identification of the provitamins. In order to have vitamin A activity, a carotenoid should have at least half of the molecule of b-carotene, i.e. an unsubstituted b-ionone ring having an 11-carbon polyene side chain (fig. 1).

For known carotenoids, confirmation of identity may be achieved by the judicious combination of traditional parameters, i.e. chromatographic behaviour, absorption spectra, and specific chemical reactions. Chromatographic data (order of elusion in gravity-flow column, retention times in HPLC, RF value and co-chromatography in TLC) give the initial clues to identification but should never be used as sole criteria. The visible spectrum remains the diagnostic tool most available to analysts, the characteristic spectrum being a consequence of the conjugated polyene chain and other structural features [3]. The type, location, and number of functional groups in xanthophylls can be confirmed by specific group reactions [4].

The vitamin A-active carotenes alfa, b, g -carotene and, b-zeacarotene and the inactive z and d-carotene and lycopene can be distinguished by their visible absorption spectra (table 1). The spectrum of b-cryptoxanthin is similar to that of b-carotene, and those of a-cryptoxanthin and zeinoxanthin resemble that of a-carotene. The hydroxycarotenoids, however, are much more adsorbed on normal phase columns and silica thin layer, and the presence of the hydroxyl group can be confirmed by acetylation with acetic anhydride. This reaction will not differentiate the monohydroxy derivatives of a-carotene which differ only in the position of the hydroxyl substituent (fig. 1) and exhibit similar chromatographic behaviour and identical spectra. Although distinguishable by NMR and MS, they are easily differentiated by methylation with acidified methanol to which the inactive zeinoxanthin responds negatively and the provitamin a-cryptoxanthin positively due to the allylic position of the hydroxyl group. a-Cryptoxanthin has been shown to be more widespread than b-cryptoxanthin [5], but it has not been found as a principal pigment and its contribution to the vitamin A value may often be regarded as insignificant. Also widely occurring, zeinoxanthin generally presents higher concentrations and, if taken for a-cryptoxanthin, will lead to overestimation of the vitamin A activity. In the Brazilian fruit Spondias lutea, for example, such a mistake signifies an overestimation of about 19% [6].

The location of the epoxide group in b-cryptoxanthin monoepoxide determines whether or not it is a provitamin. The spectrum and epoxide tests (table I) could only demonstrate its presence either at the 5,6- or 5',6'-position. NMR and MS were used to establish its location at the 5,6-position (fig. 1) in the carotenoid extracted from papaya, thereby confirming that it is vitamin A-active [7]


Quantitative analysis

The widely used AOAC method 43.013[8] has been regarded as inappropriate for provitamin A determination of foods, with the possible exception of green vegetables. In this method, the apolar carotenoids are separated from the more polar pigments, and the single fraction collected is assumed to be b-carotene. This fraction would in fact contain, if present in the sample, the less active a-carotene, b-zeacarotene, a- and b-cryptoxanthin, 5,6-mono-epoxy-b-cryptoxanthin, and the inactive carotenoids z-carotene, d-carotene, a-zeacarotene, and zeinoxanthin [9] Consequently, the vitamin A value can be grossly overestimated.

TABLE 1. Some properties of common provitamins and interfering inactive carotenoids

  Provitamin A l max (nm) in  
Carotenoid Activity (%) petroleum ethera Chemical testsb
b -carotene 100 (425). 448, 475
a -carotene 50-54 422, 444, 473
z -carotenes Inactive 378, 400, 425
b -zeacarotene 20-40 406, 428.454
d -carotene Inactive 431. 456, 489
g -carotene 42-50 437. 462.494
lycopene inactive 446, 472, 505
5,6-monoepoxy- b -carotene 21 (423). 444. 473 + epoxide test
20 nm hypsochromic shift on addition of dilute HCI
mutatochrome 50 409. 428, 452 + epoxide test
a -cryptoxanthin active (not quoted) 422. 444, 473 + acetylation
+ methylation
zeinoxanthin inactive 422. 444.473 + acetylation
- methylation
b -cryptoxanthin 50-60 (425). 449.476 + acetylation
- methylation
5,6-monoepoxy- b -crypto- xanthin active (not quoted) (422). 444, 473 + acetylation
- methylation
20 nm hypsochromic shift on addition of dilute HCI

a. Figures in parentheses represent a shoulder.
b. "+ epoxide test" means that the yellow colour turns blue or blue-green on exposure of the TLC plate to HCI vapour

The COST methods [10] assume that the provitamins other than b-carotene can be disregarded. The vitamin A value will thus be overestimated in samples having a-carotene (quantified with b-carotene) and underestimated in samples having , b-cryptoxanthin (not collected).

Chromatography in descending, gravity-flow (often with slight pressure provided by a water aspirator) columns coupled with visible absorption spectrophotometry has long been used to determine the carotenoid composition of foods. The procedure can be shortened so as to quantify only the provitamins. The simplified version used in our laboratory [9] consists of macerating a representative sample (the weight depending on the sample's carotenoid content) with cold acetone and celite in a Waring blender for 1 to 2 min, followed by filtration through a Buchner or glass sintered funnel. This operation is repeated until the residue is devoid of colour (2 or 3 times). The carotenoids are transferred to petroleum ether in a separatory funnel. To avoid formation of emulsion, which is difficult to break and can lead to loss of pigments to the lower aqueous phase, a small portion of the acetone extract is added each time, followed by water poured gently without agitation. The two phases are allowed to separate and the lower layer is discarded. When the entire extract has been added, washing 4 or 5 times with water is carried out to remove residual acetone, and the petroleum ether layer is collected. Anhydrous Na2SO4 is added to remove traces of water, the solution is concentrated in a rotary evaporator (< 35° C), and separation is accomplished on a MgO:HyfloSupercel (1:1) column developed with petroleum ether, 1%, 2%, 5% ethyl ether, and 1%, 2%, 5%, 8% acetone in petroleum ether. One or more of these solution could be deleted and the volumes adjusted according to the carotenoids found in the sample. Each separated band corresponding to a provitamin is collected and transferred to a suitable volumetric flask, the volume completed with petroleum ether, and the spectrum recorded from 550 to 350 nm. Because the absorbance is solvent-dependent, fractions containing acetone are washed 4 times with water and dried over Na2SO4 before transferring to a volumetric flask. The concentration of each provitamin is calculated, using the maximum absorbance and tabulated absorption coefficient [3].

Often referred to in the literature as tedious, the simplicity of separation in gravity-flow column (also called open column) can be demonstrated by considering the three groups described previously. For group 1, only the first coloured band, which is b-carotene, is eluted with pure petroleum ether or up to at most 5% ethyl ether in petroleum ether, depending on its concentration. This fraction, after volume adjustment, is directly measured spectrophotometrically. For group 2, a-carotene elutes before b-carotene, and these first two bands are collected separately. In both cases, all the other carotenoids (and chlorophylls when present) are left on the column.

The situation is more complicated when group 3 is considered. Hydroxycarotenoids in fruits are mostly esterified with fatty acids (traces of free carotenols are also found). Saponification is necessary to free the carotenols, allowing clean and simple separation, the total amount of a given carotenol, such as, b-cryptoxanthin, being determined as a single fraction. Since, b-cryptoxanthin is much more adsorbed than, b-carotene, chromatography should be carried out further, including the separation of interfering inactive carotenoids. In Spondias lutea, for example, a-carotene is eluted first, followed by b-carotene, z-carotene, zeinoxanthin, b-cryptoxanthin, and cryptoflavin; thus fractions 1, 2, 5, and 6 are collected and quantified. Carotenoids such as lycopene, lutein, violaxanthin, zeaxanthin, and neoxanthin, found in some foods in large amounts, are tightly adsorbed on the column and are left uneluted. For analysts who wish to determine the complete provitamin A composition, the technique permits the separation and quantitation of even minor provitamins, although this will prolong and complicate the analysis.

Saponification is also needed for oily samples to get rid of unwanted lipids. It is best done after transferring the carotenoids to petroleum ether, by adding an equal volume of 10% methanolic KOH. The mixture is left overnight at room temperature in the dark. After washing 5 times with water and drying with Na2SO4, the carotenoid solution is concentrated for column chromatography. This step could result in degradation and artifact formation, depending on the condition employed, especially at high temperature and KOH concentration [11].

Commercially available adsorbents are known to vary in their adsorptive properties, and even minute amounts of impurities may alter the solvent's eluting strength. Although variations are greater between brands, lot-to-lot differences also exist; these variations tend to be much greater in developing countries. It is therefore not surprising that a laboratory's first attempt may not duplicate reported separation, and adjustments should be made when needed. The adsorptive capacity can be increased by activating the adsorbent for 4 hours at 110° C or decreased by increasing the proportion of HyfloSupercel (e.g. 1: 2). The volumes of eluting solvents can also be changed. For example, to increase the separation of a- and b-carotene, activated MgO: HyfloSupercel ( 1: 1 ) can be used or the volumes of the initial solvents can be increased. Fortunately, since carotenoids are pigments, the separations are monitored visually and alterations can be made easily without resorting to the collection and scanning of numerous fractions as is needed for colourless compounds.

For samples having appreciable levels of cis isomers, separation of these isomers from the bans form of the provitamins may be required since they exhibit lower activity (e.g., 13-cis-b-carotene has 53% and 9-cis-b-carotene 38% vitamin A activity). The provitamin A fractions obtained from the MgO: HyfloSupercel column are individually rechromatographed on a Ca(OH)2 (Mallinkrodt) column developed with different concentrations of diethyl ether in petroleum ether [12].

Tsai et al. [13] recommended a reusable open column packed with preparative HPLC RP-C18 material (50 mm) pressurized with N2 and eluted isocratically with acetonitrile-methanol-chloroform (47 :47: 6) as an alternative for HPLC in developing countries. To avoid the use of the highly toxic and expensive acetonitrile, elusion can be done with acetone containing decreasing water concentration (15%, 10%, 0%), resulting in better recovery of b-carotene [14]. The latter system has been recommended for leafy vegetables; the order of elusion is xanthophylls, chlorophylls, and b-carotene. Separation of the carotenoids of other food samples, however, is inefficient, thus limiting its use. a- and b-carotene are not separated or are separated with difficulty; lycopene diffuses on the column and mixes with part of b-carotene and, b-cryptoxanthin.

HPLC is the most modern technique applied to provitamin A determination. Most methods utilize reversed-phase C18 columns developed isocratically with combinations of acetonitrile, chloroform, dichloromethane, tetrahydrofuran, methanol, and hexane. Cited advantages are sensitivity, resolution, reproducibility, inert conditions, and speed. With all these potentials, however, most HPLC work involves long sample preparation and quantifies only, b-carotene even in foods containing a-carotene and b-cryptoxanthin, and results reported are widely diverging [2]. These flaws will obviously be corrected as analysts become more familiar with the application of the technique to provitamin A analysis. In developing countries, however, a great obstacle will remain-the high cost of the instrument and its maintenance. Two reports were encountered on HPLC determination of provitamin A in these areas [15; 16]; both were made possible through external support. For laboratories depending on their own resources, the option is still gravity-flow column chromatography.

Even if the analysis per se is accomplished accurately, two other problems confront the analyst. The carotenoid composition of a certain food varies with the stage of maturity, cultivar, climate, soil composition, part of the plant utilized, time after harvest, and storage and shipment conditions. Thus, results obtained from single sample lots are of doubtful validity, and specification of cultivar, stage of maturity, portion analysed, etc. is useful. There is also an urgent need to better define conversion factors to transform provitamin A data to vitamin A values or activity. The current practice of expressing retinol equivalent as equal to 1 ,microg retinol, or 6 microg, b-carotene, or 12 microg of other provitamins [17] for all foods is an obvious oversimplification. For as long as this definition is not done, analysts should report their data not only in terms of the vitamin A values but also in terms of provitamin A concentrations to allow recalculations in the future.



1. Simpson KL. Relative value of carotenoids as precursors of vitamin A. Proc Nutr Soc 1983;42:7-17.

2. Rodriguez-Amaya DB. Critical review of provitamin A determination in plant foods. J Micronutr Anal 1989; 5: 191-225.

3. Davies BH. Carotenoids. In: Goodwin TW, ed. Chemistry and biochemistry of plant pigments. 2nd ed. London: Academic Press, 1976;2:38-6S.

4. Liaanen-Jensen S. Isolation, reactions. In: Isler O, ed. Carotenoids. Basel, Switzerland: Berkhauser Verlag, 1971 :61-188.

5. Rodriguez-Amaya DB, Cecchi HM, Padula M et al. Distribution of carotenoids in Brazilian foods. Paper presented at the 8th International Symposium on Carotenoids, Boston, Mass, USA, 1987.

6. Rodriguez-Amaya DB, Kimura M. Carotenóides e valor de vitamina A em cajá (Spondias lutea). Cienc e Tecnol Aliment 1989.

7. Godoy HT, Rodriguez-Amaya DB, Connor AK, Britton G. Confirmation of the structure of papaya b-cryptoxanthin monoepoxide. Food Chem (in press).

8. AOAC. Official methods of analysis. 14th ed. Arlington, Va, USA: Association of Official Analytical Chemists, 1984:834-35.

9. Rodriguez-Amaya DB, Kimura M, Godoy HT, Arima H. Assessment of provitamin A determination by open column chromatography/visible absorption spectrophotometry. J Chromatogr Sci 1988;26:624-29.

10. Brubacher G, Müller-Mulot W, Southgate DAT. Methods for the determination of vitamins in food. London: Elsevier Science Publishers, 1985:33-50.

11. Kimura M, Rodriguez-Amaya DB, Godoy HT. Assessment of the saponification step in the quantitative determination of carotenoids and provitamins A. Food Chem (in press).

12. Tavares CA, Rodriguez-Amaya DB. Determinação de provitamins A e carotenóides em alimentos "in nature" e processados. Paper presented at the IV Encontro Nacional de Analistas de Alimentos, Brasil, 1988.

13. Tsai S-W, Tsou SCS, Simpson KL. Reversed phase flash column analysis of provitamin A carotenoids. J Micronutr Anal 1989;5:171-79.

14. Mercadante AZ, Rodriguez-Amaya DB. Comparison of normal-phase and reversed-phase gravity-flow column methods for provitamin A determination. Chromatographia 1989;28:249-52.

15. Speek AJ, Speek-Saichua S, Schreurs WHP. Total carotenoid and b-carotene contents of Thai vegetables and the effect of processing. Food Chem 1988;27:24557.

16. Pepping F, Vencken CMJ, West CE. Retinol and carotene content of foods consumed in East Africa by high performance liquid chromatography. J Sci Food Agric 1988;45:359-71.

17. NAS-NRC. Recommended dietary allowances. 9th ed. Washington, DC: National Academy of Science, 1980: 55-71.

Contents - Previous - Next