CHAPTER 2: METHODS OF FOOD ANALYSIS

Despite efforts over the past half-century, there is still a need for internationally harmonized methods and data. In fact, as described in Chapter 1, the development of new methods for analysing specific components of the energy-yielding macronutrients has increased the complexity and made this need greater than ever.

This chapter discusses the commonly used analytical methods for protein, fat and carbohydrate, and makes recommendations regarding the preferred methods for the current state of the art and available technology. Methods that continue to be acceptable when the preferred methods cannot be used are also noted. Analytical methods for alcohol, which can be a significant source of energy in some diets, polyols and organic acids were not discussed, and hence no recommendations for methods are made.

2.1 ANALYTICAL METHODS FOR PROTEINS IN FOODS

2.1.1 Current status

For many years, the protein content of foods has been determined on the basis of total nitrogen content, while the Kjeldahl (or similar) method has been almost universally applied to determine nitrogen content (AOAC, 2000). Nitrogen content is then multiplied by a factor to arrive at protein content. This approach is based on two assumptions: that dietary carbohydrates and fats do not contain nitrogen, and that nearly all of the nitrogen in the diet is present as amino acids in proteins. On the basis of early determinations, the average nitrogen (N) content of proteins was found to be about 16 percent, which led to use of the calculation N x 6.25 (1/0.16 = 6.25) to convert nitrogen content into protein content.

This use of a single factor, 6.25, is confounded by two considerations. First, not all nitrogen in foods is found in proteins: it is also contained in variable quantities of other compounds, such as free amino acids, nucleotides, creatine and choline, where it is referred to as non-protein nitrogen (NPN). Only a small part of NPN is available for the synthesis of (non-essential) amino acids. Second, the nitrogen content of specific amino acids (as a percentage of weight) varies according to the molecular weight of the amino acid and the number of nitrogen atoms it contains (from one to four, depending on the amino acid in question). Based on these facts, and the different amino acid compositions of various proteins, the nitrogen content of proteins actually varies from about 13 to 19 percent. This would equate to nitrogen conversion factors ranging from 5.26 (1/0.19) to 7.69 (1/0.13).

In response to these considerations, Jones (1941) suggested that N x 6.25 be abandoned and replaced by N x a factor specific for the food in question. These specific factors, now referred to as “Jones factors”, have been widely adopted. Jones factors for the most commonly eaten foods range from 5.18 (nuts, seeds) to 6.38 (milk). It turns out, however, that most foods with a high proportion of nitrogen as NPN contain relatively small amounts of total N (Merrill and Watt, 1955; and 1973).^[4] As a result, the range of Jones factors for major sources of protein in the diet is narrower. Jones factors for animal proteins such as meat, milk and eggs are between 6.25 and 6.38; those for the vegetable proteins that supply substantial quantities of protein in cereal-/legume-based diets are generally in the range of 5.7 to 6.25. Use of the high-end factor (6.38) relative to 6.25 increases apparent protein content by 2 percent. Use of a specific factor of 5.7 (Sosulski and Imafidon, 1990) rather than the general factor of 6.25 decreases the apparent protein content by 9 percent for specific foods. In practical terms, the range of differences between the general factor of 6.25 and Jones factors is narrower than it at first appears (about 1 percent), especially for mixed diets. Table 2.1 gives examples of the Jones factors for a selection of foods.

Because proteins are made up of chains of amino acids joined by peptide bonds, they can be hydrolysed to their component amino acids, which can then be measured by ion-exchange, gas-liquid or high-performance liquid chromatography. The sum of the amino acids then represents the protein content (by weight) of the food. This is sometimes referred to as a “true protein”. The advantage of this approach is that it requires no assumptions about, or knowledge of, either the NPN content of the food or the relative proportions of specific amino acids - thus removing the two problems with the use of total N x a conversion factor. Its disadvantage is that it requires more sophisticated equipment than the Kjeldahl method, and thus may be beyond the capacity of many laboratories, especially those that carry out only intermittent analyses. In addition, experience with the method is important; some amino acids (e.g. the sulphur-containing amino acids and tryptophan) are more difficult to determine than others. Despite the complexities of amino acid analysis, in general there has been reasonably good agreement among laboratories and methods (King-Brink and Sebranek, 1993).

TABLE 2.1
Specific (Jones) factors for the conversion of nitrogen content to protein content (selected foods)

Source: Adapted and modified from Merrill and Watt (1973).

2.1.2 Recommendations

1) It is recommended that protein in foods be measured as the sum of individual amino acid residues (the molecular weight of each amino acid less the molecular weight of water) plus free amino acids, whenever possible. This recommendation is made with the knowledge that there is no official Association of Analytical Communities (AOAC)^[5] method for amino acid determination in foods. Clearly, a standardized method, support for collaborative research and scientific consensus are needed in order to bring this about.

2) Related to the previous recommendation, food composition tables should reflect protein by sum of amino acids, whenever possible. Increasingly, amino acid determinations can be expected to become more widely available owing to greater capabilities within government laboratories and larger businesses in developed countries, and to the availability of external contract laboratories that are able to carry out amino acid analysis of foods at a reasonable cost for developing countries and smaller businesses.

3) To facilitate the broader use of amino acid-based values for protein by developing countries and small businesses that may lack resources, FAO and other agencies are urged to support food analysis and to disseminate updated food tables whose values for protein are based on amino acid analyses.

4) When data on amino acids analyses are not available, determination of protein based on total N content by Kjeldahl (AOAC, 2000) or similar method x a factor is considered acceptable.

5) A specific Jones factor for nitrogen content of the food being analysed should be used to convert nitrogen to protein when the specific factor is known. When the specific factor is not known, N x the general factor 6.25 should be used. Use of the general factor for individual foods that are major sources of protein in the diet introduces an error in protein content that is relative to the specific factors and ranges from -2 percent to +9 percent. Because protein contributes an average of about 15 percent of energy in most diets, the use of N x 6.25 should introduce errors of no more than about 1 percent in estimations of energy content from protein in most diets ([-2 to +9 percent] x 15).

6) It is recommended that only amino acid analysis be used to determine protein in the following:

• foods used as the sole source of nourishment, such as infant formula;
• foods/formulas designed specifically for special dietary conditions;
• novel foods.

2.2 ANALYTICAL METHODS FOR FATS IN FOOD

2.2.1. Current status

There is perhaps more agreement on standardized methods of analysis for fat than for protein and carbohydrate. Most fat in the diet is in the form of triglyceride (three fatty acids esterified to a glycerol molecule backbone). There are also non-glyceride components such as sterols, e.g. cholesterol. While there is considerable interest in the roles that these non-glyceride components may play in metabolism, they are not important sources of energy in the diet (FAO, 1994).

There are accepted AOAC gravimetric methods for crude fat, which includes phospholipids and wax esters, as well as minor amounts of non-fatty material (AOAC, 2000). Total fat can be expressed as triglyceride equivalents determined as the sum of individual fatty acids and expressed as triglycerides (FAO, 1994). This method is satisfactory for the determination of fat in a wide variety of foods.

2.2.2 Recommendations

1) For energy purposes, it is recommended that fats be analysed as fatty acids and expressed as triglyceride equivalents, as this approach excludes waxes and the phosphate content of phospholipids, neither of which can be used for energy (James, Body and Smith, 1986).

2) A gravimetric method, although less desirable, is acceptable for energy evaluation purposes (AOAC, 2000).

2.3 ANALYTICAL METHODS FOR CARBOHYDRATES IN FOODS

2.3.1 Current status

FAO/WHO held an expert consultation on carbohydrate in 1997. The report of this meeting (FAO, 1998) presents a detailed description of the various types of carbohydrates and a review of methods used for analysis, which is summarized conceptually in the following paragraphs. Other recommendations from the 1997 consultation, e.g. the nomenclature of carbohydrates, were considered by the current technical workshop participants.

Total carbohydrate content of foods has, for many years, been calculated by difference, rather than analysed directly. Under this approach, the other constituents in the food (protein, fat, water, alcohol, ash) are determined individually, summed and subtracted from the total weight of the food. This is referred to as total carbohydrate by difference and is calculated by the following formula:

100 - (weight in grams [protein + fat + water + ash + alcohol] in 100 g of food)

It should be clear that carbohydrate estimated in this fashion includes fibre, as well as some components that are not strictly speaking carbohydrate, e.g. organic acids (Merrill and Watt, 1973). Total carbohydrate can also be calculated from the sum of the weights of individual carbohydrates and fibre after each has been directly analysed.

Available carbohydrate represents that fraction of carbohydrate that can be digested by human enzymes, is absorbed and enters into intermediary metabolism. (It does not include dietary fibre, which can be a source of energy only after fermentation - see the following subsections.) Available carbohydrate can be arrived at in two different ways: it can be estimated by difference, or analysed directly.^[6] To calculate available carbohydrate by difference, the amount of dietary fibre is analysed and subtracted from total carbohydrate, thus:

100 - (weight in grams [protein + fat + water + ash + alcohol + dietary fibre] in 100 g of food)

This yields the estimated weight of available carbohydrate, but gives no indication of the composition of the various saccharides comprising available carbohydrate. Alternatively, available carbohydrate can be derived by summing the analysed weights of individual available carbohydrates. In either case, available carbohydrate can be expressed as the weight of the carbohydrate or as monosaccharide equivalents. For a summary of all these methods, see Table 2.2.

Dietary fibre is a physiological and nutritional concept relating to those carbohydrate components of foods that are not digested in the small intestine. Dietary fibre passes undigested from the small intestine into the colon, where it may be fermented by bacteria (the microflora), the end result being variable quantities of short-chain fatty acids and several gases such as carbon dioxide, hydrogen and methane. Short-chain fatty acids are an important direct source of energy for the colonic mucosa; they are also absorbed and enter into intermediary metabolism (Cummings, 1981).

TABLE 2.2
Total and available carbohydrate

* May be expressed as weight (anhydrous form) or as the monosaccharide equivalents (hydrous form including water).

Chemically, dietary fibre can comprise: cellulose, hemicellulose, lignin and pectins from the walls of cells; resistant starch; and several other compounds (see Figure 2.1). As more has been learned about fibre, a variety of methods for analysis have been developed. Many of these measure different components of fibre, and thus yield different definitions of, and values for, it. Three methods have had sufficient collaborative testing to be generally accepted by such bodies as AOAC International and the Bureau Communautaire de Reference (BCR) of the European Community (EC) (FAO, 1998): the AOAC (2000) enzymatic, gravimetric method - Prosky (985.29); the enzymatic, chemical method of Englyst and Cummings (1988); and the enzymatic, chemical method of Theander and Aman (1982). Monro and Burlingame (1996) have pointed out, however, that at least 15 different methods are applied for determining the dietary fibre values used in food composition tables. Their publication, and the FAO/WHO report on carbohydrates in human nutrition (FAO, 1998), discuss these issues in more detail. The effect of having such a variety of methods for dietary fibre, each giving a somewhat different value, affects not only the values in food composition tables for dietary fibre per se, but also those for available carbohydrate by difference.

2.3.2 Recommendations

1) Available carbohydrate is a useful concept in energy evaluation and should be retained. This recommendation is at odds with the view of the expert consultation in 1997, which endorsed the use of the term “glycaemic carbohydrate” to mean “providing carbohydrate for metabolism” (FAO, 1998). The current group expressed concerns that “glycaemic carbohydrate” might be confused or even equated with the concept of “glycaemic index”, which is an index that describes the relative blood glucose response to different “available carbohydrates”. The term “available” seems to convey adequately the concept of “providing carbohydrate for metabolism”, while avoiding this confusion.

2) Carbohydrate should be analysed by a method that allows determination of both available carbohydrate and dietary fibre. For energy evaluation purposes, standardized, direct analysis of available carbohydrate by summation of individual carbohydrates (Southgate, 1976; Hicks, 1988) is preferred to assessment of available carbohydrate by difference, i.e. total carbohydrate by difference minus dietary fibre. This allows the separation of mono- and disaccharides from starches, which is useful in determination of energy content, as discussed in Chapter 3.

3) Determination of available carbohydrate by difference is considered acceptable for purposes of energy evaluation for most foods, but not for novel foods or food for which a reduced energy content claim is to be made. In these cases, a standardized, direct analysis of available carbohydrate should be carried out.

4) “Dietary fibre” is a useful concept that is familiar to consumers and should be retained on food labelling and in food tables. Because the physical characteristic of solubility/insolubility does not strictly correlate with fermentability/non-fermentability, the distinction between soluble and insoluble fibre is not of value in energy evaluation, nor is it of value to the consumer.

5) The AOAC (2000) analysis - Prosky (985.29) or similar method should be used for dietary fibre analysis.

6) Because dietary fibre can be determined by a number of methods that yield different results, when the Prosky method is not used the method used should be stated and the value should be identified by INFOODS tagnames^[7] (Klensin et al., 1989). In addition, the method should be identified with the tagname in food composition tables.

7) Further research and scientific consensus are needed in order to develop standardized methods of analysis of resistant starch.

Figure 2.1 - Dietary fibre: constituents and associated polysaccharide fractions

Source: Monro and Burlingame (1996).