The presence of ochratoxin A in cocoa products has been reported. This chapter describes work by the European chocolate and cocoa industry and trade to determine where ochratoxin A enters the cocoa supply chain and how to minimize it. Ochratoxin A can be found in beans from most producing countries. Ochratoxin A is produced in the beans in the producing countries and toxin levels do not seem to increase during shipping or during storage in consuming countries. Och-ratoxin A levels increase at the end of the main West-African cocoa harvest when damaged pods are implicated in the contamination. When ochratoxin A contamination occurs, the toxin is located mainly on the shell of the bean, so a large portion of the toxin present is removed during bean processing. Analysis of cocoa products on the European market confirms that only low levels of ochratoxin A are present in cocoa-containing products as consumed.
Ochratoxin A, sometimes abbreviated as OTA, is a secondary metabolite produced by some Aspergillus and Penicillium spp. Ochratoxin A is found in a range of foods such as cereals, dried fruits, grape juice, coffee, cocoa, wine and beer with cereals providing the largest contribution to the intake of ochratoxin A in Europe (DG Health and Consumer Protection, 2002). Ochratoxin A's toxicological effects have been evaluated on several occasions by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) and the EU Scientific Committee for Foods (SCF) (DG Health and Consumer Protection, 1998; JECFA, 2001). SCF concluded in 1998 that ochratoxin A possesses carcinogenic, nephrotoxic, teratogenic, immunotoxic and possibly neurotoxic properties. Based on this evaluation SCF recommended that ochratoxin A exposure should be reduced as much as possible ensuring that exposures are < 5 ng/kg bw per day (DG Health and Consumer Protection, 1998). Since 1998 further research, especially on
the mechanism(s) of ochratoxin A-induced carcinogenicity has been conducted. Based on the new research the European Food Safety Agency (EFSA) re-evaluated risks from ochratoxin A in 2006 and established a Tolerable Weekly Intake of 120 ng/kg bw and recommended that all efforts should be made to continue to reduce ochratoxin A levels in foods (EFSA, 2006).
Cocoa beans are a key raw material for the manufacture of chocolate. Approximately 65% of the world supply of cocoa beans comes from West Africa, especially Côte d'Ivoire, Ghana and Nigeria. Cocoa also is produced in Asia and Latin America. As cocoa is a crop produced by smallholders, it is a valuable non-perishable cash crop for hundreds of thousands of farmers in cocoa-producing countries and is of great importance to the economies of these countries.
Cocoa pods are harvested and broken manually. The wet cocoa beans from the pods are fermented, usually under banana leaves, and then dried in the sun. The fermented and dried cocoa beans are collected from farms and buying stations, checked for quality, graded and put into export bags. Most of the cocoa beans are exported to Europe and North America to be made into cocoa liquor, cocoa butter and cocoa powder (Fig. 1).
The CAOBISCO, ECA and FCC work program on ochratoxin A
The presence of low levels of ochratoxin A in cocoa or cocoa products was reported in the late 1990s (Engel, 2000). In 1999, CAOBISCO (The Association of the Chocolate, Biscuit and Confectionery Industries of the EU), ECA (European Cocoa Association) and FCC (Federation for Cocoa Commerce) agreed to initiate work on ochratoxin A in cocoa and a joint working group was established. The objective of the work was to "specify as precisely as possible conditions favoring ochratoxin A production in cocoa and identify solutions to minimize its production".
The stage at which ochratoxin A contamination of cocoa occurs and conditions that stimulate (or discourage) ochratoxin A production were not known at this time. Studies were therefore initiated at different steps in the supply chain (Fig. 1), e.g., fermentation, drying, storage in export warehouses, marine shipment, storage in warehouses in Europe, and the conversion to cocoa products and chocolate, were evaluated. Samples were collected from each step and analyzed for ochratoxin A level, to identify the steps at which ochratoxin A contamination occurred. The program activities were built on knowledge and experience from other crops contaminated with ochratoxin A or other mycotoxins. The bulk of the research in producing countries was contracted to CIRAD, France. Work also was contracted to CABI Bioscience and University of Aberystwyth, U.K. Studies of the processing steps in Europe were carried out by the cocoa and chocolate industry.
Mycotoxins are not homogenously distributed in commodities and therefore special sampling plans are defined for regulatory purposes. The purpose of the work described here was to identify the critical steps for production of ochratoxin A and therefore a more pragmatic approach to sampling was implemented. In the studies of cocoa bean processing at cocoa farms and of beans in collection centers where the batches were small (at maximum a few hundred kg), the samples were ~1 kg. If the samples were wet, they were immediately dried in an oven. All samples were sent to Europe for analysis.
Studies of storage in export warehouses in Africa and in warehouses in Europe were made on batches of 16 big bags (1 t each) or 100 standard bags (60 kg each). A sample of 1 kg was taken from each big bag and mixed to form a composite sample. From the standard bags, samples of ~100 g were taken from each of the 100 bags and mixed to make a composite sample. The composite samples were run through a splitter and reduced to a sample of 1 kg which was used as the analytical sample. Arbitrage samples, i.e., samples taken routinely for quality control, from lots of cocoa beans arriving in Europe were used to survey the amount of ochratoxin A in cocoa beans. The arbitrage samples are composite samples of subsamples taken randomly from at least 30% of the bags in the lot. In the finished products (chocolate, cocoa drink powder and cocoa powder) the batches are expected to be homogenous with each finished sample weighing 100-200 g.
Ochratoxin A analyses were made by laboratories that had participated in a proficiency test organized by the working group and had a z-score between ±2. The proficiency test was made with one sample each of cocoa beans and cocoa powder with ochratoxin A levels 1.1 ng/g and 2.5 ng/g, respectively. The principle analytical steps used by the laboratories involved were: extraction of the toxin, sample clean up on an immunoaffinity column, and HPLC on reversed phase column with fluorescence detection. The laboratories worked with different minimal detection limits, 0.1, 0.2 or 0.5 ng/g. When calculating averages, half of the limit of detection value was used for the samples that had ochratoxin A present below the detection limit.
Cocoa bean processing in producing countries
Some initial studies were conducted both in Ghana and Côte d'Ivoire, but after the initial phase the work concentrated in Côte d'Ivoire.
As with other fruits, the beans inside an undamaged pod were thought to be sterile until the pod wall is broken. Analysis of beans from nine healthy pods are consistent with this hypothesis, and ochratoxin A was not detected in any of the samples.
Mycoflora of cocoa samples collected in Ghana and Côte d'Ivoire
During field visits to Ghana and Côte d'Ivoire in November 2000, samples for mycological studies were collected from every stage of the cocoa processing system: fermentation heaps, drying beans, and bagged beans at both village and regional depot stores. In addition, swabs were taken from bamboo drying mats, tarpaulin drying sheets, plantain leaves covering the fermentation heaps, and air samples taken around fermentation heaps. Thirty-seven samples were analyzed for their fungal flora. All Penicillium and Aspergillus isolates were identified to species, and all other genera were ignored as not being ochratoxin A producers.
Yeast and fungi other than Aspergillus and Penicillium dominated the samples, but no ochratoxin A-producing Penicillium species were found. In samples from wooden boxes (sometimes used at larger farms for fermentation), bamboo mats and tarpaulin sheets, mostly field fungi were found, but in one sample Aspergillus niger, which can produce ochratoxin A, also was found. Aspergillus niger also was found in air samples around fermentation heaps after 3 days of fermentation and in some samples of visibly moldy beans.
Some of the samples of beans taken during drying and storage of dried beans had high ochratoxin A levels, but from only one of these samples was a species capable of producing ochratoxin A (Aspergillus carbonarius) recovered. Other non-ochratoxin A producing fungi were found on most of the samples. This pattern is not unusual since fungal vegetative growth often stops after toxin is produced since the environmental conditions may no longer be suitable for fungal growth, e.g., when a product is dried. However, the mycotoxin produced will remain in the product even after the death of the organism that produced it. This limited mycological survey of cocoa beans and farm environment shows that while yeasts and fungi that do not produce ochratoxin A dominate, fungi that can produce ochratoxin A also are present in cocoa bean samples, in the on-farm environment and on the equipment.
Ochratoxin A development during fermentation and drying at a large commercial farm in Côte d'Ivoire
The first study of ochratoxin A development during fermentation and drying was conducted on a commercial farm (about 70 hectares). The samples were fermented in bags (micro-fermentation) placed in the middle and on the top of wooden fermentation boxes. The boxes had no visible mold contamination. Samples were taken after 2, 4 and 6 days of fermentation and sun dried. When the fermentation was complete, the content of the boxes was dried at two different depths (3 and 8 cm) in drying beds. The experiments were repeated four times between November 2000 and February 2001. No ochratoxin A was detected in any of the samples. Ochratoxin A also was not detected in samples from any of five micro-fermentation tests carried out with beans from black pod (Phytophtora) infected pods.
Fermentation boxes normally are covered, e.g., with jute bags. Two trials were made in boxes that were not covered. Samples were taken after 2, 4 and 6 days of fermentation. In trials in February, 1.6 ng/g of ochratoxin A was detected after 2 days with a slight increase to 2.5 ng/g after 6 days. Repeat trials in June produced no detectable ochratoxin A.
These experiments indicate that very little ochratoxin A is produced during well-controlled fermentations in big boxes followed by sun drying.
Ochratoxin A development under smallholder conditions in Côte d'Ivoire
From February to July 2001, 62 samples of beans produced under smallholder conditions were analyzed for ochratoxin A. Unlike the results from the experiments from industrial scale fermentation in large boxes, the beans produced under small holder conditions (heap fermentation, small batches) contained ochratoxin A in many of the samples. The ochratoxin A level was > 0.5 ng/g in 24 of the samples (39%) and > 2 ng/g in 11 samples (18%). Higher levels of ochratoxin A were found in February and March than in the later April to July period.
Based on these results, we focused on ochratoxin A formation at the smallholder level. During the 2001/02 season a general survey of cocoa post-harvest practices was conducted and 168 samples were taken at small farms in three cocoa producing regions of Côte d'Ivoire (two sites in the West region, two sites in the Centre-West region and one site in the East region). Sixty-three of the 168 samples (38%) had > 0.5 ng/g ochratoxin A and 28 samples (17%) had > 2 ng/g of ochratoxin A. These results are very similar to those from the samples taken in spring 2001. Ochratoxin A was not found in 48 of the samples (26%). Very few samples were severely contaminated and only seven samples contained > 10 ng/g ochra-toxin A. The average ochratoxin A level increased steadily from 1.4 ng/g in November to a maximum of 4.1 ng/g in February. In March the ochratoxin A level fell to < 1.0 ng/g ochratoxin A and stayed at these low levels through June. Clear conclusions regarding the regional distribution of ochratoxin A contamination cannot be drawn from the results.
For each field sample of dried cocoa beans produced by the farmer, control samples also were collected just after fermentation. The control samples were frozen immediately, transported to the lab, and dried. The only difference between the control samples and the field samples was the drying method. Twenty-four of the control samples were analyzed for ochratoxin A, all of which had less, but still significant, ochratoxin A contamination than did the corresponding samples dried on the farm. Thus ochratoxin A may occur prior to the drying stage. From these studies we concluded that:
• Ochratoxin A often is present in beans that have just been fermented and usually increases during the drying process. The initial colonization of cocoa beans by ochratoxin A producing fungi probably occurs between the time the pod is opened and the end of the fermentation.
• In Côte d'Ivoire, the end of the main harvest season (January and February) corresponds to the end of the dry season and appears to be the most critical period for ochra-toxin A contamination. Thus, the more difficult fermentation conditions that occur at that time, due to climatic conditions and the nature of the mucilage, could facilitate mold growth and ochratoxin A production.
• Drying conditions alone are not responsible for the ochratoxin A level, which depends on interactions between harvesting, fermentation and drying conditions.
A working hypothesis for higher ochratoxin A levels at the end of the main crop was developed, based on the following premise: smaller amounts of beans are available resulting in smaller fermentation heaps; the pulp is drier resulting in drier heaps due to the dry season; there is a large variation in ambient day/night temperature influencing the temperature in the heaps; there could be a long time between harvest and pod opening in order to collect sufficient amount of beans for a proper heap. Trials were conducted to test the factors associated with this hypothesis in February, March and April 2003.
Three treatments were set up to test the influence of heap size and of moisture: a) fermentation in a small heap (less than 50 kg of fresh beans), b) fermentation in a large heap (more than 200 kg of fresh beans), and c) fermentation in a small heap with wetted beans. The trial heaps were inoculated with moldy beans to strengthen eventual effects. The size of the inoculum was 1% (w/w) and it consisted of beans from damaged pods with internal mold. The first trial was conducted in March with a 3-day fermentation period, and was repeated in April with 5-day fermentations. All trial heaps fermented normally and no significant difference was found in the temperature profile of the heaps. None of the samples from these trials (four samples per heap in March and five samples per heap in April) had more than traces of ochratoxin A. It seems that neither the heap size nor the moisture in the heap have a significant influence.
The effect of pod storage time was studied in four experiments with healthy and damaged pods (damage by rodents, by cutlasses during harvesting and/or by Phytophthora). The pods were stored separately for 5 days or 4 weeks before fermentation. In each trial, beans from 1500 pods were used. Samples were taken 1 and 3 days into fermentation, from the middle and edges of the heap and immediately sun dried. The trials were started in February and beans after 5 days pod storage were fermented in February and March. Beans from the pods that had been stored for 4 weeks were fermented in March and April.
Ochratoxin A levels (Table 1) were higher in beans from damaged pods than in beans from healthy pods after 5 days of pod storage. Contamination of the beans from damaged pods began on the first day of fermentation, with a higher contamination levels in the middle of the heap. Three days into the fermentation, the trend was reversed and contamination was clearly greater at the edges. This reversal was accompanied by considerable mold growth on the surface of the heap. Five days into the fermentation, the ochratoxin A content increased further. Only traces of ochratoxin A were found after fermentation and drying of beans from healthy pods stored for 5 days.
After 4 weeks of pod storage the ochratoxin A levels were low and there was only a small difference between beans from healthy and damaged pods. During the longer storage, the inside of the pods dried out around the beans. It may be that the environment became unsuitable for the growth of ochratoxin A producing molds on the beans during this time. In these trials molds were recovered from placentas and beans affected by an undetermined black rot and from fermented beans severely contaminated with ochratoxin A. Aspergil-lus strains were recovered from five of the samples, and strains from two of the five samples produced ochratoxin A. Thus, ochratoxin A-producing Aspergillus strains are present on placenta and bean debris affected by black rot.
Delaying pod opening does not affect ochratoxin A development if the pods are healthy. If the pods are damaged, then storage increases ochratoxin A content. Thus pod damage may contribute to the ochratoxin A peak found at the end of the main harvest in West Africa, where all damaged pods are routinely included in the end-of-harvest picking.
Ochratoxin A (ng/g)
Was this article helpful?