Chemical composition, digestibility and aflatoxin content of Brazil nut (Bertholletia excelsa) cake produced in north-eastern Bolivia

 C W Nagashiro, A Saucedo, E Alderson*, C D Wood** and  M J Nagler** 

Animal Science Research Institute, Faculty of Veterinary Medicine and Animal Science, Technical University of Beni, Cas. 314, Trinidad-Beni, Bolivia
carlosnagashiro@hotmail.com.
*Department for International Development, Misión Británica, FEDEPLE, Santa Cruz, Bolivia;
elizabeth.alderson@btinternet.com
**Natural Resources Institute, University of Greenwich, Central Avenue, Chatham Maritime, Kent ME4 4TB, UK  Tel: +44 1634 883540; 
c.d.wood@gre.ac.uk. 

Abstract

Brazil nut cake, a by-product of small scale oil extraction, is of potential use as a dry season feed supplement for cattle in North Eastern Bolivia. A survey indicated that over 800 tonnes/year of shelled Brazil nuts are used for oil extraction and cake production in Riberalta.  Oil extraction plants were also surveyed and samples taken for the analysis of chemical composition, in vitro digestibility and aflatoxin contamination. The nut cakes varied in composition depending on the quality of the nut used and oil extraction process. Cakes had high residual oil contents (100 to 318 g/kg DM) and high protein contents (238 to 442 g/kg DM). Crude fibre contents ranged from 25 to 108 g/kg DM, the higher contents being related to the addition of rice hull in the extraction process.  The cakes were highly digestible (69.6 to 94.4%), the lower digestibilities being related to rice hull inclusion. Brazil nut cakes were found to be rich sources of protein, phosphorus and sulphur amino acids. Brazil nut cakes produced from rotten nuts (black cakes) had relatively high levels of aflatoxins and could constitute a hazard to milk consumers if fed to dairy cattle.

Keywords: Bolivian Amazon,  Bertholletia excelsa, Brazil nut cake, chemical composition, digestibility, aflatoxins.

Introduction 

The Bolivian Amazon, the northern and northeastern regions influenced by the Amazon hydrographic basin, is about one third (360 000 km²) of the total area of Bolivia. This vast region has diverse physiographic units with different ecosystems, ranging from tropical rainforest to poorly drained savannas (Pereira and Salinas 1982; Cochrane et al 1985).

The Brazil nut tree (Bertholletia excelsa) grows mainly in the rainforest of the Amazon regions of Brazil, Bolivia and Peru, but can also be found in Colombia, Venezuela and the Guyanas, and is one of the largest and most important economic plants of the extractive reserves in the Amazon (Mori 1992).

The Brazil nuts, classified as one of the major edible nuts marketed commercially (Rosengarten 1984), are obtained from woody and thick-walled pods (pyxidium), each of which contains 18-30 seeds with a woody and bone-hard shell. These are mainly collected for export, as a high value edible nut used in the confectionery and baking trades, principally in north-western Amazon, the Acre State of Brazil and the Pando/Beni regions of Bolivia. In northern and northeastern Bolivia, the extraction, collection and processing of Brazil nuts for export, together with the latex of Hevea brasiliensis are the most important sources of income. Brazil nuts are highly nutritious, containing approximately 14-16 % protein, 65-70 % digestible fat or oil (for this reason it is considered as a minor oil crop), and 11 % carbohydrates in addition to calcium, phosphorus, potassium, vitamin B and the rare vitamin excelsine. The oil is rich in unsaturated fatty acids. The nut is also rich in the sulphur amino acids methionine and cysteine (FAO 1995). Surplus or damaged nuts are used for processing into Brazil nut oil, usually extracted in hand presses. The oil is used for cooking, soap-making and lighting. The residue of oil extraction, cake, can be used for animal feed. The pods are often utilised as a fuel source or are used to make cups and other household utensils (FAO 1992; Clay and Clement 1993).

The natural grasslands of the poorly drained savannas of the Bolivian Amazon region, known as “Pampas" or "Llanos de Moxos” (100,000 km²), have been used for sedentary cattle ranching since colonial times, and form the main basis for the region’s animal production (Rouse 1972). Due to the seasonal rainfall pattern they are flooded for several months, while droughts of varying severity and duration occur during the dry season leading to seasonal changes in quality and quantity of forage. The pasture management is usually limited to sparse grazing of large fenced paddocks and periodic burning by ranchers (Beck 1984). Natural grasses are the major if not the only feed resource. Seasonal live weight gains and losses in the dry season, similar to those reported elsewhere (Golding 1985), along with a low reproductive efficiency are observed in grazing cattle.

It is well known that in warm climates, both the quantity and quality of forage available in dry season pastures are diminished to levels that require supplemental feeding of grazing ruminants with energy and/or protein, to enable animals to survive, and to produce and reproduce. There are many factors (eg:  input costs, selling prices of products, possibility of conserving excess forage during the wet season, availability of supplemental feeds) that influence both the decision of whether or not to supplement ruminants during the dry season, and what types and amount of supplements to feed (Siebert and Hunter 1982; Golding 1985). One of the potential feed supplements, locally available, is Brazil nut cake, which is currently used for feeding dairy cattle. However, there is no basic information on its chemical composition and digestibility. Moreover, presence of aflatoxins in the cake might be expected as Brazil nut is highly susceptible to moulds (FAO 1992). It provides a rich medium for the growth of toxigenic Aspergillus species, especially in humid and warm tropical regions, as reported for other oilseeds and their meal or cake (Buckle and Scudamore 1990; Pier 1992; Devegowda et al 1998). The aflatoxins have been of intense interest to scientists because of their deleterious effects on animal health, and acute toxicological effects in humans (Wood 1992).  

This paper reports the results of a survey on the kind of nut and oil extraction techniques used to obtain the Brazil nut cake in north-eastern Bolivia, and how these influence the chemical composition, in vitro dry matter digestibility and aflatoxin content.  

Materials and methods 

Survey

Before taking any cake samples, a survey was carried out using a questionnaire to determine the possible factors which might influence the chemical composition, digestibility and aflatoxin content of the cakes readily available in the region, and to define how many and which samples should be taken. It was designed to identify and classify the factories according to activities such as nut processing for export, oil extraction and cake production, and to determine the kinds of nuts and processes used for oil extraction, the kind of cake obtained and its further management. The survey was conducted in April, 1997. 

Sampling

Based on the results of the survey, cake samples were taken according to the procedure described by Mallman (1995), taking into consideration the difficulties of taking representative samples because of the growth characteristics of mould, which is well recognised as the main source of error for quantitative estimation of mycotoxins concentration (Wood 1992). The cake samples were stored in air-tight aluminium bags and placed in a refrigerator until they were sent to the Laboratories of the Natural Resources Institute (NRI) of the University of Greenwich, England, for further analysis. 

Chemical composition, digestibility and aflatoxins analysis

Samples were analysed for dry matter (DM), crude protein (CP), crude fibre (CF), ash (CA), ether extract (EE), total phosphorus and calcium by methods described in Anonymous (1982).

Amino acids were analysed by separation using ion exchange chromatography on a Biotronik LC5001 amino acid analyser and detection photometrically by ninhydrin (Moore 1963; Moore and Stein 1963).

The in vitro dry matter digestibility (IVDMD) was determined using the method of Tilley and Terry (1963).

For aflatoxin analysis, all samples were ground using a coffee grinder and the 50 g analytical samples were taken using a small scoop x10 from throughout the mixed samples. Samples were analysed by the standard procedure for aflatoxin analysis, SOP3 (Bradburn etal 1990) with triplicate quantification by HPTLC using a scanning fluorodensitometer (Scanner II, CAMAG) controlled by software (CATS 3, CAMAG).   

Results 

Survey

Data obtained from the survey are summarized in Tables 1 and 2. Table 1 shows the approximate amount of in-shell raw nut reported by each of the 14 nut shelling factories which responded, of the 15 found in Riberalta City, northeastern Bolivia. An approximate total amount of 21,729 tonnes of in-shell raw nut was processed to obtain 7,171 tonnes of edible nut suitable for export. The processing period is variable among factories, ranging form 2 to 12 months/year. The amounts of rejected nuts (edible of low grade for export = 441 tonnes; rotten = 392 tonnes) from the selection and classification process reported by some 12 of the 15 factories are also shown.

A summary of the use of rejected nuts for oil extraction is given in Table 2. In the region there are 3 oil extraction plants (A, B, C) that acquired the rejected nut from nut shelling factories. The rejected nuts used for oil extraction are classified as: broken or out of size still edible, but not suitable for export; and rotten. The edible broken, the edible out of size and rotten nuts are used by all plants. Hand presses are used for oil extraction. To facilitate the oil extraction some plants add rice hull to the nut in different and variable proportions (Nut : rice hull ratio for A 2:1; for B 8 : 1), but at plant C or, for certain oils at plant B, no rice hull is added. Two kinds of oils are normally extracted:  edible oil and oil for other uses (energy sources, making soap, etc.). In all plants the edible oil is usually extracted from edible nuts, but conversely, the oil for other uses is extracted from rotten nuts only in plants that buy them (A and B). Three kinds of cake are obtained: white, black and mixed. The white cake results from the extraction of edible oil using edible nuts, and the black cake is obtained from the extraction of oil for other uses using only rotten nuts. The mixed cake, produced in plant B, refers to that obtained for mixing both, white and black cakes, after separately extracting the oil.     

 Sample characteristics 

Details of the samples taken in the 3 plants (2 for plant A, 3 for B and 1 for C), after taking into account the results of the survey, are given in Table 3. They have been characterized according to the kind of nut and oil extraction process used, and the kind of cake obtained.

Three white cakes, obtained from rejected edible nuts, were found. One of the white cakes was obtained using rice hull (sample 1 belonging to the plant A), and the other two do not (samples 3 and 6 of plants B and C, respectively). The two black cakes (samples 2 and 4), obtained from only rotten nuts as was shown above, and had rice hull in different proportions according to the plant (2:1 for plant A, and 8:1 for B). The mixed cake (5) of the plant B, is that obtained from mixing the different kinds of cake after oil extraction, which are thereafter piled up outside of the plant and marketed in this way. 

Chemical composition and IVDMD 

Table 4 shows the chemical composition and IVDMD for each sample and the average of all samples.The mean CP content was 376 g/kg DM, ranging from 238 to 442 g/kg DM. The CP content among samples was negatively correlated (r = -0.83, P<0.05) to the EE (residual oil) content. The mean EE content was high, 216 g/kg DM, and variable among samples (range of 100 to 381 g/kg DM). The lowest and highest EE values, both corresponded to the white and black cakes, respectively, of the plant A. The EE content of the three samples of plant B (samples 3, 4 and 5) tended to be more uniform (range of 181 to 187 g/kg DM). The mean CF content was 64 g/kg DM (range of 25 to 108 g/kg DM). Samples with higher values were 1 and 2, 108 and 91 g/kg DM, respectively, of plant A, which uses a nut : rice hull ratio of 2 : 1 during nut pressing, for both (white and black) cake. Among samples of plant B, sample (4) obtained using rice hulls tended to have a higher CF content. Since rice hull was used in plant B, but in less proportion than that in plant A, sample 4 shows a lower CF content (70 g/kg DM) than samples 1 and 2, 108 and 91 g/kg DM, respectively. Sample 3 of plant B and sample 6 of plant C, obtained without adding rice hulls, show a lower CF content, 36 and 25 g/kg DM, respectively. The mean content of ash, Ca and total P were 92, 3.7 and 17.3 g/kg DM, and it was also variable among samples, as was observed in other chemical constituents. 

In general, the mean IVDMD of the samples was high (86.3 %). Samples 1 and 2 of plant A, in spite of the fact that these are different kinds of cake (white and black), obtained by pressing nut with a high proportion of rice hull, show comparatively lower IVDMD values, 77.6 and 69.9 %, respectively. The IVDMD of the 3 different samples (3, 4 and 5) of plant B and sample 6 of plant C was up to 90 %. Although sample 4 was obtained with the addition of rice hulls, but in a less proportion (nut : rice hull=8:1) to samples 1 and 2 of plant A,  it tends to show a high IVDMD (92.3 %), which is almost similar to that of samples obtained without adding rice hulls, sample 3 of plant B (94.4 %) and 6 of plant C (92.4 %). Among samples obtained with rice hulls, the IVDMD was higher in sample 4 than that of the two samples (1 and 2) of plant A  (92.3 % vs. 77.6 and 69.9 %).    

The amino acid composition is given in Table 5.  The variation in the content of each amino acid  among samples tends to be similar to that of CP content.   

 Aflatoxin content 

The aflatoxin content is shown (Table 6) separately for aflatoxins B1, B2, G1 and G2,  and as total content for each sample. From the mean of all samples, the highest content was that of G1 (164 µg/kg) followed by B1 (93 µg/kg). The samples with greater total aflatoxin content were 4  (746 µg/kg) and 5 (608 µg/kg), followed by sample 2 (248 µg/kg), which correspond to the black and mixed cake of plants B, and to the black cake of plant A, respectively.  The white cakes (samples 1, 3 and 6), showed relatively low total aflatoxin content (range of 56 to 88 µg/kg).    

Discussion

Survey

The estimate of in-shell raw nut production obtained from the survey was higher than the value reported by Mendoza (1988) for the same region, which might be attributed to the number of factories found in the present survey (N=15, reported only 14), compared to that in his study (N=7). The difficulties of getting information from the factories, reported by Mendoza (1988), were also found in the present survey. Therefore, the values of in-shell nut and edible nut for export, should be considered as rough estimates. 

Chemical composition and IVDM digestibility

The CP content of the samples in general was relatively high, and it appeared to be more affected by the residual oil content, measured by EE, than by the addition of rice hull. This effect was observed in samples 1 (EE, 100 g/kg DM) and 2 (EE, 388 g/kg DM) of plant A, which had CP contents of 393 and 238 g/kg DM, respectively. Since the sample with less residual oil content (1) is a white cake and sample 2 is a black cake, the reason for the difference in residual oil in these samples might be associated with the need to extract more edible oil from edible nuts. However, this is not observed in samples 3, 4 and 5 from plant B, which had similar residual oil contents and similar CP contents. This reflects the lack of a standard oil extraction process.

The residual oil content of samples was high enough for the occurrence of oxidative rancidity due to the high unsaturated fatty acids content (ca. 75 %) reported for the oil (Adams 1975, cited by Clay and Clement 1993).  This is a risk, and particularly, if the cakes are stored and transported under warm and humid conditions, prevalent in the region. It could be a factor which may lower the actual nutritive value as well as palatability, as has been reported for other feeds of high residual oil content (Penz 1994; Valenzuela 1995). This problem could be overcome by extracting a greater amount of oil.

The total P content in general is comparatively high, and was well correlated to CP content (r=0.99, P<0.01). This fact is very important particularly in the Bolivian tropical lowlands where 100 % of the forage samples were reported to be below the minimum P requirement (0.25 %) for grazing ruminants (Peducassé et al 1983). It is well recognized that the lack of P in forages is the most prevalent mineral element deficiency throughout the world, but specially in tropical regions, for grazing livestock (McDowell 1997). Some mineral supplementation experiments carried out in the tropical lowland of Bolivia, confirmed the fact that the low P content of grasses could be a factor for the low productive and reproductive efficiency observed in animals grazing natural grasslands (McDowell et al 1982).

Compared to the amino acid content of soya bean meal (summarised by Novus International 1994 and Degussa 1997), the most used protein supplement of vegetable origin for animal feeding in Bolivia, the sulphur amino acids content is higher in Brazil nut cake, which confirms the report by FAO (1995) and data reviewed by Guimaraes Cruz and Macedo Junior (1995) in Brazil. In addition to the high sulphur amino acid content, Guimaraez Cruz and Macedo Junior (1995) reported that tryptophan content is also high. In the present work tryptophan was not determined.

The IVDMD appeared to decrease due to rice hull addition during the oil extraction process.  Nut quality appeared to make little difference even though black cake samples (4 and 5) had a greater aflatoxin content.  

Aflatoxin content

In all nut cake samples the level of aflatoxin G1 exceeded that for aflatoxin B1. This is in contrast to that found in maize, cottonseed and peanuts, considered as the most important sources of aflatoxins in animal feed, in which the concentration of aflatoxin B1, recognized as the most biologically active member of the aflatoxin family, is greater (Pier 1992). However, it is well known that the resulting profile of aflatoxins and their individual concentration will vary greatly according to the existing environmental conditions (temperature, moisture, aeration), the substrate, and the type of mould involved (Cheeke and Shull 1985).

In comparison to white cakes, the higher aflatoxin content was found in the black cake samples and also in the cake resulting from the mixing of white and black cakes and stored outside of the plant. Since rotten nuts were used to obtain those black cakes, the raw material appeared to be the main source of contamination. The origin of nut spoilage might be attributed to the susceptibility to insect damage in the field before harvest, although the seeds have a very hard shell which, moreover are contained in a round capsule or pod (pyxidium) with a hard woody capsule wall (FAO 1992; Clay and Clement 1993). The penetration of Aspergillus spp. into the edible kernel, through the damaged areas, could occur, in a similar way to that reported for cotton (Cheeke and Shull 1985) and groundnut (Buckle and Scudamore 1990).

The aflatoxin level of white cakes is lower than the maximum level accepted for cottonseed as a feed ingredient (300 µg/kg), but it is higher than the maximum level for maize used for immature animals and dairy cattle or for all feedstuff (20 µg/kg) established by FDA (Wood 1992). However, the aflatoxin content of black cakes was higher than the maximum level (300 µg/kg). As cake samples were taken at the moment of oil extraction, excepting mixed samples, the values represent the aflatoxin content of fresh cakes. Therefore, higher values could be expected in cakes that are being transported for long distances to farms and stored for a long time under tropical climate conditions. A major concern in the region is the use of the cakes for feeding dairy cattle, because the well known transfer of the aflatoxin M1, a metabolite of aflatoxin B1, into the milk, which would be a hazard for humans. 

Conclusions

Brazil nut cakes are rich in protein, phosphorus and sulphur amino acids.  They are highly digestible.  As a supplement to dairy cattle the cakes are potentially very useful.  However, black cakes produced from rotten Brazil nuts contain relatively high levels of aflatoxins, which could be hazardous to consumers of milk. 

Acknowledgements 

This publication is an output from a research project partly funded by the United Kingdom Department for International Development (DFID) for the benefit of developing countries. The views are not necessarily those of DFID.

Dr C Nagashiro wishes to thank the German Academic Exchange Service (DAAD), because the first draft was written during a study visit to the JLU Giessen University sponsored by them. The valuable comments and suggestion on the manuscript from Professor Dr Joerg Steinbach, Department of Livestock Ecology of JLU Giessen University, are also gratefully acknowledged. 

References 

Anonymous 1982 Feeding Stuffs (Sampling and Analysis) Regulations. Her Majesty's Stationary Office (HMSO), London, UK.

Beck S 1984 Comunidades vegetales de las sabanas inundadizas en el NE de Bolivia. Phytocoenologia 12, 321-350.

Bradburn N R, Coker R D and Jewers K 1990 A comparative study of phenyl bonded-phase, CB and Romer clean-up procedures for determining aflatoxin levels in maize by bi-directional HPTLC. Chromatographia 29: 177-181.

Buckle A E  and Scudamore K A 1990 The occurrence, detection and significance of mycotoxins in animal feeding stuffs. In : Feedstuff evaluation (Wiseman J and Cole D J A, eds) pp. 411-428. London : Butterworths.

Cheeke P R and Shull L R 1985 Natural toxicants in feed and poisonous plants. Pp. 393-477. Connecticut: Avi Publishing Company.

Clay J W and Clement C R 1993 Brazil nut. Selected species and strategies to enhance income generation from Amazonian forest Misc/93/6 working paper. Rome: FAO.

Cochrane T T, Azevedo L G de, Thomas D, Madeira Netto J, Adamoli J and Verdesio J J 1985 Land use and productive potential of American Savannas pp 114 - 124. In: Ecology and management of the world's savannas (Tothill J C and. Mott J C, eds) 358 pp.  Pub by Australian Academy of Science, Canberra, Australia

Degussa 1997 AminoDat 1.1, The amino acid composition of feedstuffs (A PC-software). Frankfurt am Main : Degussa AG.

Devegowda G, Raju M V L N, Nazar Afzali and Swamy H V L N 1998 Micotoxin picture worldwide: Novel solutions for their counteraction. In : Biotechnology in the feed industry, Proceedings of Alltech’s fourteenth annual symposium (Lyons T P and Jacques K A, eds) pp. 241-255. Nottingham : Nottingham University Press.

FAO 1992 Brazil nut. In : Minor oil crops. FAO Agricultural service bulletin No. 94. Rome : FAO.

FAO 1995 Brazil nut. In : Edible nuts. Non – wood forest products 5. Rome : FAO.

Guimaraes Cruz F G and Macedo Junior R A 1995 Utilizacao da farinha de castanha do Brasil (Bertholletia excelsa H.B.K.) em racoes de frangos de corte. In: Congreso de APINCO pp. 47-48. Campinas: Fundacao APINCO de Ciencia e Tecnologia Avícolas (FACTA).

Golding E J 1985 Providing energy-protein supplementation during the dry season. In: Nutrition of grazing ruminants in warm climate (McDowell L R ed) pp. 129-163. London: AP Inc.

Mallman C A 1995 A mostragen de graos e racoes para a pesquisa de micotoxinas. In: Simposio internacional sobre micotoxinas y micotoxicosis en aves, pp. 21-33.

McDowell L R, Bauer B, Galdo E, Koger M, Loosli J K and Conrad J H 1982 Mineral supplementation of beef cattle in the Bolivian tropics. Journal of Animal Science. 55 (4): 964-970.

McDowell L R 1997 Mineral for grazing ruminants in tropical regions, Third edition Florida University, Florida Cooperative Extension Service.

Mendoza G 1988 Estudio sobre la comercialización de la castańa en Bolivia. IICA.

Moore S 1963  On the determination of cystine as cysteic acid.  Journal of Biological Chemistry 238, 235 - 237.

Moore S and Stein W H 1963 Chromatographic determination of amino acids by the use of automatic recording equipment.  P819 - 831. In: Methods in enzymology Vol. 6 (Colwick S  P and Kaplan N O, eds) New York: Academic Press.  1054 pp .

Mori S A 1992 The Brazil nut industry – Past, present and future. P241 - 251 In: Sustainable harvest and marketing of rain forest products (Plotkin M and Famolare L, eds) 325 pp. Washington D.C. & Covelo, California: Island Press.

Novus 1994 Raw material compendium-A compilation of worldwide data sources (2^nd edition) pp. 310-321. St. Louis, MO: Novus International Inc.

Peducassé C A, McDowell L R, Parra L A, Wilkins J V, Martin F G, Loosli J K and Conrad J H 1983 Mineral status of grazing beef cattle in the tropics of Bolivia. Tropical Animal Production 8: 118-130.

Penz A M 1994 Valor nutricional y toxicidad de las grasas en el alimento. In: VIII Seminario internacional de patología aviar (Villegas P ed) pp. 235-258. Athens-Georgia: Georgia University.

Pereira F and Salinas J G 1982 General evaluation of the agricultural potential of the Bolivian Amazon. In: Amazonia, agriculture and land use research (Hecht S B ed) pp. 17-32. Colombia: CIAT.

Pier A C 1992 Major biological consequences of aflatoxicosis in animal production. Journal of Animal Science 70: 3964-3967.

Rosegarten F 1984 The book of edible nuts. New York: Walker and Company.

Rouse J E 1972 Bolivia. In: World cattle, Cattle of Europe, South America, Australia and New Zealand (Rouse J E ed) pp. 367-381. Oklahoma: Univ. Oklahoma, Press- Norman.

Siebert B D and Hunter R A 1982 Supplementary feeding of grazing ruminants. In: Nutritional limits to animal production from pastures (Hacker J B ed) pp. 409-429. London: CAB.

Tilley J M A and Terry R A 1963 A two-stage technique for the in vitro digestion of forage crops.  Journal of the British Grassland Society 18, 104-111.

Valenzuela A 1995 Natural antioxidants: A new perspective for the problem of oxidative rancidity of lipids. In: Biotechnology in the feed industry, Proceedings of Alltech’s eleventh annual symposium (Lyons T P and Jacques K A, eds) pp. 207-220. Nottingham: Nottingham University Press.

Wood G E 1992 Mycotoxins in food and feeds in the United States. Journal of Animal Science 70: 3941-3949.

  Received 18 August 2000

Go to top