Information on the nutrient composition of South African foods is available from the South African Food Database (SAFOOD). The data can be accessed in the following ways:
The metabolisable energy value of a food item was derived by multiplying the amounts of protein, total carbohydrate (available carbohydrate plus dietary fibre) total fat, and alcohol (when applicable) with the conversion factors shown below. The results are tallied to give the energy value of the food item.
1 g protein = 17 kilojoules (kJ) or 4 kilocalories (kcal)
1 g total carbohydrate (available carbohydrate plus dietary fibre) = 17 kJ or 4 kcal
1 g fat = 37 kJ or 9 kcal
1 g alcohol = 29 kJ or 7 kcal
Each of the analytical methods used for the individual proximates has an acceptable degree of variation from the true value. If, for example, the proximates are at the upper-end of their acceptable range, the sum of the proximates will be more than 100 g (and vice versa).
The rounding off of reported values adds to the compounding error of a value. Another possible reason is that organic acids and non-nutrient factors, which also contribute to the weight of the food, may not be included. The range of 97-103 g for the sum of the proximates could be considered acceptable (Greenfield and Southgate, 2003). Values falling outside this range were evaluated for valid reasons for the discrepancy.
In most foods ‘total fat’ (total lipids) is made up of fatty acids plus non-fatty acid material such as glycerol, phosphate moiety of phospholipids and sterols. To calculate what proportion of the ‘total fat’ content in 100 g of a specific food item is fatty acids, fatty acid conversion factors can be used. The fatty acid conversion factors for different foods are: (a) Milk and milk products, 0.945; (b) Eggs, 0.830; (c) Meat and meat products, 0.910-0.945 (organ meats: 0.561-0.789) (Holland et al., 1991).
Fatty acid conversion factors for fruit and vegetables range from 0.534-0.950 g fatty acids/g of total fat (US Department of Agriculture, 1998).
Conclusions cannot be drawn from the direct comparison between the vitamin and mineral composition of different forms of a specific food (raw, boiled, frozen, canned, etc.). The destruction of some nutrients by heat is only one of the factors influencing nutrient composition during processing. Moisture loss during cooking leads to the concentration of all nutrients. In addition, most often it is not samples of food from the same crop that were analysed raw, frozen, boiled, canned or dried. Differences in growing conditions, sample handling, etc. could therefore result in nutrient differences. Analyses for the various forms (raw, boiled, frozen, canned, etc.) of a food item may also have been done at different analytical laboratories, possibly using different analytical methodologies and thus producing different values. With inter-laboratory studies the impact on values, using different analytical methods, can be minimised. Adding additional ingredients to processed food as well as the fortification of the food item may also influence the final analytical values.
Analytical methodology for nutrients are continually improving and becoming more accurate. A zero value was used when it is known that the specific nutrient is normally not present in that food.
The term trace value (“tr”) means that the nutrient is present in that food, but at such low levels that it cannot be adequately measured. A trace value can also be regarded as nutritionally insignificant as it is too small to be expressed in the unit of measurement used in the specific food database (Greenfield and Southgate, 2003).
As a user of food composition data, you need to be aware of missing values in the food composition database. Missing values should not be regarded as a zero value. A missing value means that no information is yet available for the nutrient. When analysing dietary intake data, missing values should be flagged or the results interpreted keeping in mind that missing values lead to an underestimation of nutrient intake of an individual or population.
In order to reduce the number of missing values in a food composition database it is practice to estimate values by using other reference sources or by deducing the values from related foods. Estimated values prepared by careful interpretation of the data on related foods are acceptable in nutritional studies, provided that their use is clearly noted.A useful tool is the SAMRC Food Quantities Manual, which has the weight of foods in common units and household measures for many of the foods listed in the FCT.
Guidelines on how to use a recipe calculation to estimate the nutrient content of a food item is given in the Recipe Calculations section of the electronic FCT.
Before any part of the South African Food Database can be reproduced, permission must be granted by the South African Medical Research Council (SAMRC) in writing. Even if the use of the reproduced copy is for educational purposes, written permission from the SAMRC is still required.
Written requests can be forwarded to safoods@mrc.ac.za