• QA 097
    How easy is it to sample for OTA in green coffee?
    Chapter 12 of the Guide discusses the prevention of Ochratoxin A (OTA) in green coffee and section 12.08.01 suggests that the sampling of green coffee to determine the presence of OTA is problematic. However, we hear that in fact a simple and fast method (10 minutes…) exists to determine the presence of OTA in green coffee arrivals at destination. What are we to make of this?
    Asked by:
    Association - Uganda

    We cannot caution strongly enough against simplification of the sampling issue because in fact this is the main hurdle in determining whether a shipment contains OTA or not. Of course there are a number of test processes available to determine the presence of OTA in green coffee and other products. Some may also perhaps work better/faster than others but we cannot comment on claims made by individual manufacturers. What we can comment on is the sampling issue…

    OTA in green coffee is caused by mould. Except for extremely rare cases where an entire consignment might have been affected, mould occurs on individual beans, or small pockets of beans which may be found in some parts of a bulk consignment (or a single bag even), yet not at all in other parts. In other words, no matter how the procedure is carried out, conventional sampling* cannot guarantee that the presence of mould will be discovered, irrespective of the test process used. * See footnote at end.

    As yet no internationally agreed sampling method to cope with this exists! Nevertheless, we could imagine that an accurate sampling cum testing method could be devised to determine the presence of OTA whilst discharging bulk or loose coffee into a silo. However, this ignores two major considerations…

    1. The logistical problem incumbent in the 100% accurate sampling of green coffee arrivals is unmanageable. To achieve this, before travelling to its final destination, all arriving bagged coffee and bulk coffee would first have to be discharged into holding silos to allow sampling and testing to take place. In 2005 the European Union alone imported the equivalent of between 46 and 50 million 60-kilo bags green coffee, i.e. close to three million tonnes amounting to approximately 150,000 individual container loads. All these would require individual discharge into individual silos, sampling, testing and separate storage pending the outcome of the testing procedure. This kind of silo storage capacity simply does not exist whereas the associated costs would surely be passed on to consumers or, more likely, will result in lower producer prices. 

    2. It is not necessary! It has been established through the work of the Food and Agriculture Organization of the United Nations (FAO-Rome) that if green coffee smells clean, cups clean and looks clean (especially free of skins, husks and beans in cherry), then the likelihood of it having an OTA level below 1 ppb is close to 90%! This is a better accuracy than provided by notoriously difficult sampling in combination with analysis that currently - even in well performing laboratories - may easily deviate by 25% plus or minus! A clean smell is particularly important since smell overcomes the problem of inhomogeneity in a container that sampling cannot deal with. 

    The best safeguard against OTA contamination is prevention! This has been stressed time and again by widely respected scientific and industry sources whose work is best illustrated by the FAO's Task Force at www.coffee-ota.org. Other good sources of information include the International Coffee Organization at www.ico.org whereas the European Coffee Federation shows the entire set of preventative recommendations under 'Publications' at www.ecf-coffee.org.

    In closing we would stress the need to keep the discussion around OTA realistic: green coffee likely to be contaminated will show the signs, basically because it will be of poor quality! Experience proves that sound coffee of clean appearance offers little or no risk and does not warrant the imposition of complicated, time-consuming and costly 'control mechanisms'. 

    * In coffee trade terms 'a representative sample' used to mean a sample drawn at random from at least ten percent, sometimes twenty, of a parcel. This worked well in the days when all coffee was shipped in bags and the only objective was to determine the general quality of a parcel. Of course all this changed when bulk shipments became the order of the day: bulk coffee can only be sampled accurately upon discharge, either into a holding silo or when being discharged into the receiving silo of a roasting plant. This type of automated sampling is much more accurate than the old method but, no one can afford to allow entry into a plant's receiving system of bulk coffee that subsequently turns out to be unacceptable. Obviously this has resulted in stricter control by the roasting sector over both suppliers and incoming quality.

    Posted 03 June 2006

    Related chapter(s):
    Related Q & A: