Development of a database of fatty acids in Australian foods. (Original Research).

Abstract

Food composition databases in general contain limited data on fatty acid profiles of foods, both in terms of foods represented and the range of fatty acids reported. The aim of this project was to compile a database of Australian and New Zealand foods for which complete and accurate fatty acid compositional data was available. The database constructed, contains 1044 food items with data on amounts of 31 fatty acids recorded to two decimal places as gram fatty acid per 100 g edible food portion. This fatty acid database will be of value to dietitians and research groups in estimating intake levels of specific fatty acids in individuals and population groups. Furthermore food companies will have a means of assessing the validity of making claims on n-3 polyunsaturated fatty acid content of their commodities in line with the stipulations of the new Australia and New Zealand Food Standards Code.

Key words: database, diet, dietary analysis, fatty acid, food composition, n-3 polyunsaturated fatty acid (PUFA)

Introduction

A major facet of the work of researchers and dietitians in the field of nutrition and food science is the determination of the likely nutrient intakes of individuals or population groups from a knowledge of dietary intake and the nutrient composition of foods. In Australia the most comprehensive food composition database available to analyse dietary intakes is NUTTAB95 (NUTTAB95 Data Tables for Use in Australia, Canberra, Australian Government Publishing Service, 1995). The supplement to the NUTTAB95 database (1) contains several files, one of which provides fatty acid profiles (g/100 g edible portion, to one decimal place) for a subset of the NUTTAB95 foods based on data previously published in Composition of Foods, Australia (2).

As health authorities endeavour to reduce the consumption of saturated fat, there is growing recognition of the need to distinguish between the different types of fatty acids, particularly unsaturated fatty acids in the diet. Dietary polyunsaturated fatty acids (PUFA), in particular the long chain fatty acids such as arachidonic acid (AA), eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), are present in only small amounts in individual foods but accumulate to significant levels of biological importance in the context of a whole diet. For most foods, these low values often round down to zero when reported to a single decimal place and so it is likely that such fatty acids will be consistently underreported in any dietary analysis. Hence there is a need to measure and report fatty acid concentrations in foods as mg/l00 g edible food portion or g/100 g to two decimal places (e.g. 40 mg EPA/100 g food would be reported as 0.04 g/100 g [two decimal places] or 0.0 g/100 g [one decimal place]).

This problem is compounded by the incomplete subset of foods, for which fatty acid data are available, in relation to the total number of foods reported in NUTTAB95. There is also an issue with inaccuracy of data, either in identifying specific fatty acids correctly or determining their quantity accurately, often due to the limitations of old technology such as packed column gas chromatography analysis, which had limitations in resolving some fatty acid peaks, leading to misidentification. Also, NUTTAB95 was released in 1995 and much of its analytical data were determined before 1992, therefore it is likely that a substantial amount of fatty acid data are out of date in relation to the current food supply.

The aim of this study was to extend and update the fatty acid file contained in the NUTTAB95 supplement (1) utilising the data expressed to two decimal places, provided by the Australian and New Zealand Food Authority (ANZFA), now Food Standards Australia and New Zealand, and used with permission, in conjunction with fatty acid compositional data from research groups, government bodies and industry, in both Australia and New Zealand.

Methods

Search strategy

An extensive search of the literature was conducted to locate published fatty acid analyses of commonly available foods from Australian and New Zealand research groups and published since 1990. Also research groups, government bodies and industry sources known to the authors were contacted directly to request any further fatty acid compositional data they had in their possession. Any data contained in the database have been included with the permission of the research group, government body or industry providing the data, but copyright remains with the relevant provider.

Criteria for inclusion or exclusion

Our database was constructed using Microsoft Excel 97 software (Microsoft Corporation, version SP-2 2002) and a template based on the customised version of NUTTAB95 fatty acid file provided by ANZFA (g/100 g edible portion, to two decimal places). This file initially consisted of fatty acid compositional data on 1116 foods. However, after being reviewed extensively for foods for which values were erroneous or incomplete (primarily long chain PUFA), 449 food items were eliminated, including the majority of the meat, eggs, prawns and some processed fish products. This process was carried out independently by two fatty acid specialists. The main reasons for exclusion were that some PUFAs identified as present are not found in those foods, or because there was no value recorded for important long chain PUFAs that are normally found in those foods. For example, many of the egg products were excluded because the data lacked values for arachidonic acid (20:4n-6), a predominant long chain fatty acid in egg yolk found at levels up to 400 mg/l00 g egg yolk (3).

Added to our database was any food with a complete fatty acid profile from a reputable Australian or New Zealand source that was reported in concentration terms (g/100 g edible portion to two decimal places), performed by capillary column gas chromatography. Similarly, fatty acid compositional data from reputable sources reported as % total fatty acids were incorporated into the database if total lipid content was reported and a relevant lipid conversion factor was available (4) allowing calculation of fatty acid concentrations. The data gathered from the various sources such as research groups, government bodies and the scientific literature, were newly added to the remnant fatty acid file data from the NUTTAB95 supplement, rather than attempting to cross match the data with the foods already listed in the primary NUTTAB95 database.

As the various data sets came from a number of sources that did not use a standardised set of fatty acid descriptors, the identification, naming and recording of trace fatty acids varied considerably. A standardised set of fatty acid descriptors was needed to more accurately reflect the true identity of the individual fatty acids and overall accuracy of the composition. Based on their abundance in foods and their biological significance, a comprehensive range of 31 fatty acids were selected, by two independent fatty acid specialists, for inclusion in the new database. The major additional fatty acids included in our database were trans fatty acids, for which there is an increasing interest but little data. However, for many foods in our database, the analytical method either did not allow discrimination of trans fatty acids or these acids were not specifically identified and reported. Many such foods are likely to contain traces of trans fatty acids but there are no entries for them. Caution, therefore, is n eeded when using this or any other fatty acid database to calculate total trans fatty acid intakes, as values calculated will be minimum estimates rather than total intakes.

Fatty acids selected

The range of fatty acids selected for inclusion in our database are represented in 31 fatty acid columns, which include all 11, even-numbered, carbon-saturated fatty acids from 4:0 to 24:0; nine monounsaturated fatty acids (14:1, 15:1, 16:1, 16:1trans, 17:1, 18:1, 18:1trans, 20:1, 22:1); and 11 PUFA (l8:2n-6, l8:2n-6 trans, 18:3n-3, 20:2n-6, 20:3n-6, 20:4n-6, 20:5n-3, 20:5n-6, 22:4n-6, 22:5n-3, 22:6n-3). Total saturates, monounsaturates and polyunsaturates are also included, along with total fat and moisture columns where this data was available. Discrimination of positional isomers of trans fatty acids was not evident in most data collected, hence the figures given represent total concentration of all positional isomeric forms for each trans fatty acid. Similar food items were cross checked between data sources and international fatty acid tables (closest food item to Australian item). Any food missing a fatty acid common to that food or containing a fatty acid not found in that food by other sources was exc luded.

Results

Our database at this point in time contains fatty acid profiles for 1044 foods from a broad range of food categories as indicated in Table 1. A number of Australian research groups contributed data to this project along with substantial input from ANZFA, the Health Department of Western Australia -- Chemistry Centre of WA, CSIRO Division of Marine Science, Crop & Food Research, New Zealand and Unilever Foods Australia. The number of food items analysed and represented in our database by source are indicated in Table 2. As can be seen the largest contributions of data were from the fatty acid supplement to NUTTAB95 provided by ANZFA (1), followed by CSIRO Division of Marine Science (5), Sinclair and Mann, Lipid Research Group at RMIT (3,6-10), Health Department of Western Australia (personal communication, Spadek T, Chemistry Centre of WA, 1991) and New Zealand FOODfiles (11).

Discussion

The literature search did not reveal as many articles as was originally envisaged, with most papers listing only a few selected foods. The range of foods also was quite limited and tended to concentrate on fish, due to fish being an excellent source of long chain n-3 PUFAs, and a great deal of current research is focussing on the benefits of including long chain n-3 PUFAs in the diet.

A number of specific problems were evident as a result of this search for fatty acid compositional data on common foods. Few research papers were written with the aim of showing a comprehensive fatty acid profile for the foods. Most were written for dietary studies or concentrated on particular fatty acids, particularly PUFA, trans and n-3 PUFAs. Other studies simply reported total saturated, monounsaturated or polyunsaturated fatty acids, although complete analysis would have been performed. Further to this, the values were also frequently reported as % total fatty acid but without the total fat content that is necessary to convert % total fatty acid into a meaningful quantitative value (g/100 g edible portion). In a number of cases the documentation of the food was inadequately described because specific details such as brands or varieties were missing. A further problem encountered in a number of studies was that the authors often did not identify the monounsaturated fatty acid (omega n-7 or n-9) and PUFA family groupings (n-3 or n-6). All such incomplete data were discarded and only complete data were included in our database.

Permission was obtained to include values from the Crop & Food (New Zealand) database, FOODfiles (11). This database is a composite of data from overseas or derived from primary analysis by New Zealand laboratories, the latter being considered as an appropriate source of data for inclusion in our database. Most of the foods added from the New Zealand FOODfiles were processed foods, such as biscuits and snack foods which are also available in Australia.

The CSIRO Division of Marine Research made available the database of fatty acid profiles of Australian seafood analysed with Fisheries Research Development Corporation (FRDC) funding and published in 'Seafood the good food' (5). This is a very comprehensive database with excellent documentation and data. A copy of this report on seafood consumption (12) was obtained and used in conjunction with NUTTAB95 (1) in order to select the most common seafoods consumed in Australia, to determine the appropriate fish data to be included in the database.

Our fatty acid database will be updated as more food compositional data becomes available, either from research groups or food companies. Persons or organisations with complete fatty acid compositional data on well-described food items should submit these to the primary author for consideration for inclusion. It is anticipated that the database will be included in commercial software packages, to aid dietitians and researchers in dietary analysis of fatty acid intakes of individuals or groups. The database is currently available on the dietary analysis software platform, FoodWorks (Xyris Software, Brisbane, Food Works, version 3.01 2002) as a stand-alone database for analyses of fatty acid composition from dietary intake information in a similar manner to NUTTAB95 database, but unfortunately not simultaneous with this database as many of the foods listed are different to those entered in NUTTAB95. This is unfortunate but unavoidable given that the fatty acid data are a collection of foods analysed by research groups for their own specific projects and not related to either of the existing databases of Australian food composition. Another limitation of the database is the incomplete list of foods with fatty acid data compared with NUTTAB95, which leads to an underestimate of total fatty acid intake when analysing food intake records. However, at this point in time we believe the database contains fatty acid composition on most foods in the Australian diet containing long chain PUFA, hence it should give an accurate indication of the intake levels of these fatty acids involved in a wide range of health-related outcomes.

A further application of this fatty acid database is for the purposes of food labelling. Volume 2 of the (Australia New Zealand) Food Standards Code (clause 13, Standard 1.2.8, Food Standards Australia and New Zealand, 2002) sets conditions for several types of fatty acid claims including those for source and good source of omega-3 fatty acids (13). The reliable and complete fatty acid compositional data of our database will guide manufacturers to the likely fatty acid composition of their products and aid them in decision making in relation to the making or otherwise of specific fatty acid claims.

Table 1

Fatty acid database breakdown by food category, indicating the number of
food items represented in each category

Food Category                Number of entries

Breakfast cereals                    12
Oils, margarine, spreads            105
Fruit and vegetables                 44
Cheese                               50
Milk and milk products               43
Noodle and pasta dishes              19
Soy products                          5
Infant foods                         50
Eggs and egg dishes                  28
Meat and meat products              120
Seafood                             265
Fast foods                           50
Breads                               28
Soups, sauces and dressings          42
Desserts and sweets                  17
Biscuits and snack foods             76
Pies, cakes and buns                 31
Pastry                                8
Nuts and seeds                       25
Miscellaneous                        23

Total                              1044

Table 2

Fatty acid database breakdown by data source, indicating the number of
food items analysed reported by each group

Data source                         Foods analysed

ANZFA (NUTTAB95 supplement)               667
1999 (1)
Mann et al. 1995 (3)                       20
Nichols 1998 (5)                          126
Attard 1997 (6)                             3
Mansour, Sinclair 1993 (7)                  6
Khoh 1997 (8)                               9
Li 2001 (9)                                20
Periera 1999 (10)                           5
Crop & Food Research NZ 1998 (11)          45
Belling et al. 1997 (14)                   11
Siebert et al. 1996 (15)                    1
Report on the Nutritional Analysis          1
of Australian Beef 1995 (16)
'Entertainers' lamb composition             4
study 1995 (17)
Sadler, Lewis, Buick 1993 (18)              4
Imbeah 1997 (19)                            5
Armstrong, Wyllie, Leach 1994 (20)         11
Gibson, Makrides, Hawkes 1998 (21)          2
Ashton 2000 (22)                            8
Beilin, Mori (a)                            7
Makrides (b)                                3
Spadek 1991 (c)                            63
Unilever Foods Australia 2001 (d)          23

Total                                    1044

(a)personal communication, Beilin L, Mori T, Department of Medicine,
University of Western Australia, Perth, 2000.

(b)pers. comm.,  Makrides M, Child Nutrition Research Unit, Womens and
Childrens Hospital, Adelaide, 2000.

(c)pers. comm., Spedek T, 1991.

(d)pers. comm., Unilever Foods Australia, 2000.

Acknowledgements

The authors wish to acknowledge all groups listed in the reference section who supplied fatty acid data. The data used in this database were sourced and compiled by Dr Paul Percival (RMIT) with the assistance of Greg Milligan (FSANZ formerly ANZFA) and Fiona Kelly (Department of Food Science, RMIT University). The project was made possible by funding support from Unilever Australia and Meadow Lea Foods/Grains Research Development Corporation.

References

(1.) Australia New Zealand Food Authority. Supplement to NUTTAB95 database. Canberra: Australia New Zealand Food Authority; 1999.

(2.) National Food Authority. Composition of Foods, Australia (COFA), volumes 1-7, Canberra: Australian Government Publishing Service; 1995.

(3.) Mann NJ, Johnson LG, Warrick GE, Sinclair AJ. The Arachidonic acid content of the Australian diet is lower than previously estimated. J Nutr 1995;125:2528-35.

(4.) Cashel K, English R, Lewis J. Department of Community Services and Health. Composition of Foods Australia (Appendix 5). Canberra: Australian Government Publishing Service; 1989.

(5.) Nichols P. Seafood the good food: the oil (fat) content and composition of Australian commercial fishes, shellfish and crustaceans. Hobart: CSIRO Division of Marine Research; 1998.

(6.) Attard D. Trans-fatty acid content of some high fat foods [Honours thesis]. Melbourne: RMIT University; 1997.

(7.) Mansour MP, Sinclair AJ. The trans fatty acid and positional (sn-2) fatty acid composition of some Australian margarines, dairy blends and animal fats. Asia Pac J Clin Nutr 1993;3:155-63.

(8.) Khoh SO. Omega-3 polyunsaturated fatty acids (PUFA) in canned fish products [Honours thesis]. Melbourne: RMIT University; 1997.

(9.) Li D. The influence of diet on atherosclerotic and thrombotic risk factors in healthy men [PhD thesis]. Melbourne: RMIT University; 1998.

(10.) Pereira C, Li D, Sinclair AJ. The alpha-linolenic acid content of green vegetables commonly available in Australia. Int J Vitam Nutr Res 2001;71:223-8.

(11.) Crop & Food Research. FOODfiles. New Zealand Food Composition Database. Palmerston North, New Zealand: Crop & Food Research; 1995.

(12.) Ruello NV. A study of the demand and importance of seafood sourced in NSW and elsewhere to the catering and tourism industries in NSW. Prepared for the FRDC by Ruello and Associates, Sydney; 1996.

(13.) Food Standards Australia and New Zealand. Food Standards Code Volume 2, up to and including Amendment 61. <URL: (http://www.foodstandards.gov.au/foodstandards-code/standard12/standa rd128.cfm>accessed September 2001.

(14.) Belling GB, Abbey M, Campbell JH, Campbell GR. Lipid content and fatty acid composition of 11 species of Queensland (Australia) fish. Lipids 1997;32:621-5.

(15.) Siebert BD, Deland MP, Pitchford WS. Breed differences in the fatty acid composition of subcutaneous and intramuscular lipid of early and late maturing, grain-finished cattle. Aust J Agric Res 1996;467:943-52.

(16.) Report on the nutritional analysis of Australian beef, (as performed by) Australian Government Analytical Laboratories, commissioned by Australian Meat and Livestock Corporation 1995. Sydney: Meat and Livestock Corporation; 1995.

(17.) Australian Government Analytical Laboratories. 'Entertainers' lamb composition study. Report to Meat and Livestock Australia (MLA). Sydney: Meat and Livestock Corporation; 1995.

(18.) Sadler M, Lewis J, Buick D. Nutrient composition of trim lamb and lean beef. Food Australia 1993;45(11)Suppl:1S-19S.

(19.) Imbeah M. Fatty acid composition of five sausages in Australia. Food Australia 1997;49:408-10.

(20.) Armstrong SG, Wyllie SG, Leach DN. Effects of season and location of catch on fatty acid compositions of some Australian fish species. Food Chem 1994;51:295-315.

(21.) Makrides M, Hawkes JS, Neumann MA, Gibson RA. Nutritional effect of including egg yolk in the weaning diet of breast-fed and formula-fed infants: a randomized controlled trial. Am J Clin Nutr 2002;75:1084-92.

(22.) Ashton Y. The contribution of food groups to the total fat and fat type in various habitual diets [Honours thesis]. Melbourne: RMIT University; 2000.

Department of Food Science, RMIT University, Melbourne

N.J. Mann, B AppSc, BSc (Hons), DipEd, PhD, Senior Lecturer 

A. J. Sinclair, BSc (Hons), PhD, Research Professor

P. Percival, BSc (Hons), PhD, Project officer

Food Standards Australia and New Zealand, Canberra

J.L. Lewis, BSc, Grad Dip Nutrition and Dietetics, Principal Nutritionist 

Department of Biomedical Science & Smart Foods Centre, University of Wollongong 

B.J. Meyer, BSc (Hons), PhD, Senior Lecturer

P.R.C. Howe, BSc, MSc, PhD, previously Professor and Scientific Director; Professorial Research Fellow, University of Adelaide and University of South Australia 

Correspondence: N.J. Mann, Department of Food Science, RMIT University, GPO Box 2476V, Melbourne, VIC 3001. Email: [e-mail]