Contents - Previous - Next

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

International food composition data

Nutrient intake data calculated using food composition tables: factors affecting accuracy
The status of food composition data in Asia
Food composition data in Sweden and the nordic countries
Food data in Canada: the Canadian nutrient file

Nutrient intake data calculated using food composition tables: factors affecting accuracy

Materials and methods


Laboratory of Nutritional Biochemistry, Department of Nutrition, Federal University of Pernambuco, Cidade Universitaria, Brazil


Specific nutrient deficiencies continue to be one of the world's major public health problems, especially in underdeveloped countries. In terms of number of individuals affected and geographical distribution, vitamin A and iron deficiencies are among the most prominent [25]. One very serious drawback for the design and implementation of nutrition intervention programmes is the inadequacy of dietary information, almost always plagued by the spectre of inaccuracy [21]. Thus, it is not uncommon to find reports describing high iron intakes in areas where iron-deficiency anaemia appears with undesirably high prevalence [2, 9, 11], or extremely low vitamin A intakes which are not accompanied by a compatibly high prevalence of eye lesions [2, 6, 9, 11]. Similar problems exist when trying to establish correlations between the intake of other nutrients and related clinical or biochemical indicators [3,12, 18, 20, 21].

The accuracy of nutrient information depends on the methods of collecting and handling the data. Some of these have been examined [3, 8,14, 20, 21] but, in general, attempts to reconcile dietary and biochemical or clinical information from nutrition surveys are still needed. The present study was designed to evaluate the relative contribution to the inaccuracy of dietary information of both regional differences in the nutrient composition of foods and the differences between those values obtained by calculation and those obtained by direct analysis of foods as eaten.

Materials and methods

Three recipes for the dishes most frequently consumed by the population in Northeast Brazil were selected for this study. These recipes are practically standard and appear with very little variation among users [19]; they can thus be considered representative of local alimentary practices. The proportion of raw ingredients and the per cent composition as actually eaten are shown in table 1. The culinary preparations, also in accordance with local practices, were as follows:

Table 1. Composition of regional dishes

Ingredients Raw
(% in serving)
Beans (mulatinho) 39.5 21. 1
Beef,dried 19.8 24.2
Bacon 4.9 10.1
Okra 2.0 2.7
Pumpkin 15.8 21.7
Green herbsa 14.8 18.9
Salt 1.0 -
Pork blood 39 5 39 5
Pork liver 12.5 12.5
Pork heart 12.9 12.9
Chard 11.9 11.8
Green herbsa 21.8 21.7
Salt 1.0 -
Beef 42.6 24.0
Pumpkin 9.1 8.9
Sweet potatoes 9.1 10 9
Potatoes 9.1 91
Plantain 7.3 116
Wild cabbage 3 7 7 4
Green herbsa 18.8 15 0
Cassava flour - 5 1
Salt 1.0 -

a. A mixture, in equal parts, of green pepper, coriander, green onions, tomatoes, and onions.

- Feijoada: the beans (mulatinho type) were soaked in water for one hour, after which the beef and bacon were added. The mixture was boiled for two hours. Then the vegetables were added for a final boiling for 30 minutes.

- Sarapatel de porco: the pork blood and viscera were cooked in salted water, cut in small pieces, and boiled for one hour with the vegetables.

- Cozido: the meat and aromatic herbs were boiled in water for two hours. When these were nearly cooked, the vegetables were added. Before serving, the solids and liquid were separated, the latter to be mixed with cassava flour.

Individual servings of each preparation, in accordance with local uses [19], were duly homogenized with a blender, and appropriate aliquots of the homogenates were taken for the analysis of moisture, ash [10], fibre [24], ether extract [16], protein [15], calcium [4], phosphate [5], iron [10], vitamin A and carotenoids [1], and vitamin C [23]. The minerals were analysed in aliquots of the ashes. Carbohydrates were calculated by difference. The net weight of raw and cooked ingredients and the relative contribution of the latter in individual servings were recorded. All assays were run in duplicate, using appropriate standards. Differences greater than 5 per cent between duplicates were considered unacceptable. All ingredients of the recipes were also analysed individually. The nutrient composition of the recipes was calculated using the values of the Food Composition Table of INCAP-ICCND [17] and those obtained from our analysis of the individual raw ingredients. Hence, two estimates of the nutrient composition of each recipe were made, along with direct analysis of appropriate aliquots of each dish. All reagents were analytical grade.


Table 2 permits comparison of the values of the INCAP table [17] with those obtained at our laboratory for the proximal composition, plus the iron, vitamin A, and vitamin C content of the 22 ingredients of the 3 dishes selected for this study. The values for fibre, calcium, and phosphate were not included in the table, as they add little to the objective of this work. It is readily apparent that agreement within a margin of + 20 per cent was obtained only in about one-third of the total number of analyses performed. Results were below 80 per cent of the table value in 38 per cent of the observations, while 28 per cent presented with differences of more than 20 per cent above the figure in the table.

In the case of iron and vitamin A, the tendency was for the INCAP table values to grossly overestimate the content of the foodstuffs analysed. Especially remarkable was the case of iron, where 90 per cent of our values were well below 80 per cent of the figure presented by the INCAP table. Most of the major differences in protein content were in foodstuffs that are unimportant as sources of this nutrient, like coriander, pepper, and onions.

Table 3 was constructed to show the practical implications of the differences in the nutrient compositions described above. A previous survey [2], carried out in a country village, was used as a source of data to recalculate nutrient intake using the values obtained at our laboratory. Protein consumption shows very little difference, as could be expected from table 2. Vitamin A intake, on the other hand, seems to have been slightly underestimated, and that of iron grossly overestimated, when the INCAP table values were used.

The problem of "foods as eaten" was approached by comparing the results of direct analysis of the dish with the nutrient composition calculated using the values in the INCAP table or the the values obtained by local analyses (table 2). The changes in the relative proportion of the ingredients after cooking were taken into account by directly weighing the ingredients before and after culinary processing.

Figure 1 shows that up to 22-fold differences could be found when comparing the nutrient composition of the food as eaten with that calculated from the composition of the raw ingredients ("recipe calculation"). As could be expected from the degree of "complexity" of the dishes, feijoada was the one containing the largest differences.

Contents - Previous - Next