Occurrence of ochratoxin A in commodities and processed food – A review of EU occurrence data

Abstract

A brief review on the occurrence of ochratoxin A in commodities and processed food on the European market (meat and meat products, cereal and cereal products, spices, beer, cocoa and derived products, coffee, wine, dried vine fruits, grape juice) is given in an historical perspective based on two EU project reports from the activity initiated by the European Commission: Scientific cooperation on questions relating to food (SCOOP). The most important commodities contaminated with ochratoxin A are known and the amount of occurrence data is in most cases comprehensive. However, gaps of knowledge exist, e.g. possible year-to-year variations for wine, dried vine fruits and grape juice are not well investigated. In addition, a follow-up on possible improvements in agricultural and processing practices is needed for some of the commodities recently discovered as being contaminated with ochratoxin A, such as coffee, wine, dried vine fruits and grape juice.

Keywords:

• Ochratoxin A
• occurrence
• EU data

Introduction

In 1995, the European Commission (EC) initiated the activity: Scientific cooperation on questions relating to food (SCOOP). One of the first projects was SCOOP task 3.2.2, coordinated by Denmark, to provide data on the occurrence of ochratoxin A in commodities on the European market and on dietary exposure to ochratoxin A in the EU member states. After the first SCOOP report (SCOOP-1) (European Commission 1997) was finalized and during the ongoing debate on possible EC legislation, many surveys of ochratoxin A in different commodities were initiated in several European countries, e.g. in Germany (Wolff et al. 2000). Several activities were also initiated by the private sector such as the coffee and wine industry, e.g. activities related to improvement of analytical methods, generation of occurrence data and prevention measures to decrease contamination by ochratoxin A.

Many new data were therefore generated, and it was recognized that the SCOOP-1 report was based on a limited amount of data. Therefore the EC initiated a second SCOOP task in 1999, 3.2.7, coordinated by Italy, in order to determine if more recent data would change the conclusions of the first report. This project was finalized in 2002 with the second SCOOP report (SCOOP-2) (European Commission 2002). The SCOOP reports contain references to most of the reported occurrence data.

During the years, new commodities have been added to the list of possible ochratoxin A-contaminated commodities and thereby to the list of possible contributors to human intake. By use of the database Food Science and Technology Abstract (FSTA) an historical view of the food commodities that had been in focus in relation ochratoxin A can been seen (Table I). Ochratoxin A was discovered in the mid-1960s and was for the first time found as a natural contaminant in cereals at the end of the 1960s (Shotwell et al. 1969). FSTA includes publications from 1969 and forward and is therefore chosen as an indicator for the activities related to ochratoxin A in the different commodities.

Table I. Number of publications in the database Food Science and Technology Abstract (1969-7/2005); searching: ‘commodity and ochratoxin A’.

Years of publication	Meat	Cereals	Beer	Spices	Coffee	Cocoa products	Wine	Dried vine fruits	Grape juice
1969–1974	3	4	2	0	1	0	0	0	1
1975–1979	10	14	2	2	4	1	0	0	0
1980–1984	27	25	2	2	6	1	0	0	0
1985–1989	18	23	3	1	6	2	0	0	0
1990–1994	20	38	2	6	7	2	0	1	0
1995–1999	23	43	13	6	25	4	7	3	3
2000–7/2005	24	43	27	11	53	16	63	16	14

Based mainly on the two SCOOP reports and due to this based on occurrence data for products on the market in the old EU Member States, the list of commodities is briefly reviewed and any possible gaps of knowledge highlighted. However, before that some aspects concerning the quality and reliability of occurrence data are emphasized.

Evaluation of occurrence data

It is important to consider sampling procedures when occurrence data are used for different purposes (Whitaker 2003, European Commission 1997). Year-to-year variations and possible changes in agricultural and processing practices make it important to consider if the occurrence data can be considered to be representative or not, e.g. in relation to possible use for the estimation of human intake for a given population. Year-to-year variation is an important issue, because the occurrence of ochratoxin A in cereals and other vegetable products may depend on climatic conditions (Jørgensen et al. 1996, Jørgensen & Jakobsen 2002, Stefanaki et al. 2003). In addition, any possible changes in agricultural and processing practices are important and activities initiated with the goal of decreasing the contamination levels should be taken into consideration. Such improvements have, for instance, been seen in the coffee industry (van der Stegen 2003), in the Danish cereal sector (Jørgensen et al. 1996, Jørgensen & Jakobsen 2002) and in the Danish pig sector (Jørgensen & Petersen 2002).

The analytical methods for ochratoxin A have been improved considerably in the last ten years with the use of immunoaffinity columns and high-performance liquid chromatography, e.g. the detection limits have been lowered. In the past it was not possible to detect natural levels in many commodities. When occurrence data are compared, frequencies of contamination are often listed but these are most of the time of little informative value. For instance in the SCOOP-1 report 19% of all data points were positive and in the SCOOP-2 report 49% of all data points were positive; this is the consequence of the improvement of detection limits. Many of the commodities are natural places for the ochratoxin A producing fungi to grow, so with even better analytical methods 100% positive samples is a possibility. The division into ranges of concentrations and/or reporting of ranges of percentiles are of more value when occurrence data are reported. There is a general need for occurrence data to be reported in a uniform format including some minimum requirements of quality assurance data for the analytical methods used.

Food commodities and processed food

In the following, the main groups of commodities and processed food are briefly reviewed mainly based on the SCOOP-2 report.

Meat and meat products

From the FSTA overview (Table I) it can be seen that publications on ochratoxin A and meat products have been rather evenly distributed through the years. In the first years after ochratoxin A was found to be an important food contaminant, meat products were regarded as one of the main contributors to the human intake. For instance in Denmark it was not until the mid-1980s that the Danish food authorities recognized that it was more likely that cereals for human consumption were the main contributor to the human intake.

The overall mean level of ochratoxin A in meat products was 0.2 µg/kg (n = 1828) in the SCOOP-2 report. Pig products were the most contaminated commodities; especially pig products containing blood, kidney and liver. The data on meat and meat products are comprehensive. At present, the contribution of those products to the total human intake of ochratoxin A is limited; the contribution from meat may have been larger in the past. This is indicated by the results from the Danish control with slaughter pigs (Jørgensen & Petersen 2002). Until around 1990 the intake of ochratoxin A through pig products was probably of some importance in Denmark. However, data for mean levels in Danish pig products before 1990 are not available.

Cereals and cereal products

A lot of data for cereals was already reported in the SCOOP-1 report. However, some of the data was relatively old and analytical methods had high detection limits. Together with meat and meat products, cereals and cereal products are the only commodities where the SCOOP-1 report included more data than the SCOOP-2 report. Data from the SCOOP-2 report on cereal grains are shown in Table II. Rye seems to be the most contaminated species of grain. The amount of occurrence data for cereals is again comprehensive.

Table II. Contents of ochratoxin A in cereal grains reported in the SCOOP-2 report (European Commission 2002).

Cereal grains	Mean (µg/kg)	Maximum (µg/kg)	>EU limit of 5 µg/kg^1 (%)	Number of samples
Wheat	0.27	32	0.8	979
Rye	0.60	33	2.0	444
Barley	0.30	6.4	0.7	142
Maize	0.17	4.9	0	267
Oat	0.19	5.9	0.6	164
^1EU maximum limit for raw cereal grains (European Commission 2005).

Some investigations have shown higher levels in organically grown cereals when compared with conventionally grown cereals (Jørgensen et al. 1996, Jørgensen & Jakobsen 2002, Czerwiecki et al. 2002).

Spices

In the SCOOP-2 report, the overall mean level of ochratoxin A in spices was 1.2 µg/kg (n = 361) and the maximum content reported was 24 µg/kg. Some of the most contaminated products were nutmeg, paprika, chilli, coriander and pepper. The amount of data for spices is limited. However, spices are of minor importance for the total human intake of ochratoxin A because of the low consumption of spices.

Beer

The first investigations of occurrence of ochratoxin A in beer were conducted at the beginning of the 1970s and several investigations have been published (see Table I). However, most available data has been generated since the middle of the 1990s when ochratoxin A was in focus in relation to food safety and possible EU regulation and the analytical methods could measure the levels found in beer. The overall mean level of ochratoxin A in beer was 0.03 µg/kg (n = 496) in the SCOOP-2 report and the maximum content reported was 0.29 µg/kg. The mean of 0.03 µg/kg is probably a good estimate for the overall mean level of ochratoxin A in beer on the European market today. This has been confirmed by a recent publication (Tangni et al. 2002).

Cocoa and derived products

The current data on cocoa and derived products have mainly been generated in the last five years (Table I). In the SCOOP-2 report the overall mean level of ochratoxin A in all cocoa and derived products was 0.24 µg/kg (n = 547) and the maximum content reported was 3.6 µg/kg. Some new publications since the SCOOP-2 project was finalized have indicated that the levels in cocoa products could be higher (Amézqueta et al. 2004, Bonvehi 2004, Tafuri et al. 2004). Therefore more data are needed.

Coffee

At the beginning of the 1970s, high levels of ochratoxin A were found in green coffee samples (Levi et al. 1974). However, most investigations concerning ochratoxin A in coffee products have been carried out in the last ten years (Table I) following the discussion about possible EU maximum limits for coffee products. In the SCOOP-2 report the overall mean level of ochratoxin A in roasted coffee was 0.62 µg/kg (n = 788) and 1.3 µg/kg (n = 226) in instant coffee. The EU maximum limits in force from April 2005 are 5 µg/kg for roasted coffee and 10 µg/kg for instant coffee (European Commission 2005). One percentage of the samples in the SCOOP-2 report exceeded the EU maximum limit for roasted coffee; no samples of instant coffee exceeded the EU maximum limit.

Recent surveys seem to show that improvements in agricultural and processing practices have decreased the average level in coffee on the European market (van der Stegen 2003). Due to the EU maximum limits in force from April 2005 it can be expected that further data will be generated in the EU member states in the next years; this may show if the average level in coffee on the European market has been lowered.

Wine, dried vine fruits and grape juice

Wine and other grape products are the latest group of products that have been found to contain ochratoxin A. Occurrence of ochratoxin A in wine was reported for the first time in 1995 (Zimmerli & Dick 1995) and much occurrence data has been generated since then for wine, dried vine fruits and grape juice; especially in the last five years. The ecology of the involved fungi, Aspergillus niger and Aspergillus carbonarius, in relation to ochratoxin A production is now under investigation (Battilani & Pietri 2002, Mitchell et al. 2004, Bellí et al. 2004).

The overall mean level of ochratoxin A in wine was 0.36 µg/kg (n = 1470) in the SCOOP-2 report. The content in European wine from the south of Europe is higher than in wine from north Europe. The content is higher in red wine than in rosé and white wine. Surveys indicate lower levels in wine from North America compared with European wine (Ng et al. 2004). An EU maximum limit of 2 µg/kg for wine will be in force from the harvest 2005 (European Commission 2005). From the SCOOP-2 report it was not possible to give the exact number of samples above the coming maximum limit. However, 2–4% of the total number of wine samples exceeded the coming EU maximum limit.

The overall mean level of ochratoxin A in dried vine fruits was 3.1 µg/kg (n = 593) in the SCOOP-2 report. The EU maximum limit of 10 µg/kg for dried vine fruits has been in force since 2002 (European Commission 2005); 7% of the samples in the SCOOP-2 report exceeded the EU maximum limit.

The overall mean level of ochratoxin A in grape juice was 0.56 µg/kg (n = 146) in the SCOOP-2 report. An EU maximum limit of 2 µg/kg will be in force for grape juice from the harvest 2005 (European Commission 2005). Approximately 5% of the samples in the SCOOP-2 report exceeded the coming EU maximum limit.

Grape juice is special in the sense that it is quite a niche product on the market. However, the relatively few consumers are often children, so it might have some influence on this group in relation to the total intake of ochratoxin A, especially for high consumers of the product. Dried vine fruits are also products that are often eaten by children and the possible health risk caused by ochratoxin A by regularly consumption of dried vine fruits by children has caused public debate in some countries.

One investigation has indicated a year-to-year variation in the content of ochratoxin A in wine produced in Greece (Stefanaki et al. 2003). There is a need for more investigations of year-to-year variations of occurrence in wine, grape juice and dried vine fruits. Work is going on in relation to possible improvements of agricultural and processing practices for these products. It can be expected to lead to a decrease in the contamination levels, so a follow-up is needed for these products in the coming years. The introduction of EU maximum limits for wine and grape juice will probably initiate a generation of further occurrence data.

Conclusions

The most important commodities contaminated with ochratoxin A are known. This is indicated by the fact that human intake estimates based on occurrence and consumption data and intake estimates based on human fluids data seem to be at the same level. For many commodities the amount of occurrence data is comprehensive. However, some gaps of knowledge exist for some commodities, e.g. possible year-to-year variations are not well investigated for all. In addition a follow-up on possible improvements in agricultural and processing practices is needed for some of the commodities recently discovered as contaminated with ochratoxin A.

This brief review is based on occurrence data for commodities and processed food on the European market in the old EU member states and the amount of occurrence data are in most cases comprehensive. However, there are still limitations for some commodities from some of the old EU member states and there are limited amounts of occurrence data for many of the new EU member states. In other parts of the world the amount of occurrence data is very limited. A recent collection of occurrence data from the whole world can be found in JECFA (2001).