C. Riccioli et al., J. Near Infrared Spectrosc. 20, 623–633 (2012) Received: 5 September 2012 n Revised: 15 October 2012 n Accepted:17 October 2012 n Publication:18 October 2012
623
JOURNAL OF NEAR INFRARED SPECTROSCOPY
Detection and quantification of ruminant meal in processed animal proteins: a comparative study of near infrared spectroscopy and near infrared chemical imaging Cecilia Riccioli,a Dolores Pérez-Marín,a,* José Emilio Guerrero-Ginel,a Tom Fearnb and Ana Garrido-Varoa,* a
Non Destructive Spectral Sensors Unit. Faculty of Agricultural and Forestry Engineering, University of Córdoba, 14071, Córdoba, Spain. E-mail:
[email protected],
[email protected] b
Department of Statistical Science, University College London, UK
This study compared the performance of single-point near infrared spectroscopy (NIR) and near infrared chemical imaging (NIR-CI) for the detection and quantification of ruminant meat meal in processed animal proteins (PAPs). A set of 126 fish-meal samples adulterated with controlled amounts (0.25% to 16%) of ruminant meal were analysed using the two techniques. Comparison of results showed that spectra obtained by NIR-CI provided better qualitative information, whereas more accurate quantitative predictions were obtained using NIR spectroscopy. NIR-CI thus offers greater potential for species discrimination/identification, whilst NIR spectroscopy is better suited for the quantification of meal derived from a given species in PAPs. These findings represent a first step in the analysis of mixed-species processed animal proteins and suggest that NIR-CI, by providing valuable information on species origin, is a promising tool that could be used as part of the EU feed control programme aimed at eradicating and preventing Bovine Spongiform Encephalopathy (BSE) and related diseases. Keywords: NIR spectroscopy, NIR-CI, hyperspectral analysis, processed animal proteins, fish meals, ruminant meals
Introduction
A total EU-wide suspension on the use of processed animal proteins (PAP) in feed for any animals farmed for the production of food has existed since January 2001 with some exceptions such as the use of fish meal for non ruminants. However, in March 2011, growing concern at the protein deficit affecting the compound-feed manufacturing sector prompted the European Union (EU) Directorate General for Health & Consumers (DG SANCO) to approve a partial lifting of the EU ban on the use of PAPs in feeds intended for farmed animals; draft amendments have already been produced to Regulation (EC) No. 999/2001.1 Clearly, the rendering industries and the producers of fish meal as well as the makers of compound animal feed stand to ISSN: 0967-0335 doi: 10.1255/jnirs.1027
benefit greatly from any relaxation of the total ban. However, any such relaxation carries with it some increased risk of accidental or even deliberate contamination. Serious food frauds, with consequent damage to consumers, have occurred even in situations where bans are supposedly accompanied by systems of inspection and control, underlining the need for intensive monitoring. One important point is that fraudulent adulterations do not normally involve the addition of trace amounts of the banned ingredient. The economic motive behind the adulteration can usually only be achieved by the addition of significant quantities of the adulterant. With this in mind, it is not clear that the Commission’s chosen strategy over methods of inspection and control for the presence of © IM Publications LLP 2012 All rights reserved
624
prohibited PAPs in animal feed is the most appropriate one. The Commission’s objective is protection with zero tolerance. This brings with it various difficulties for the implementation of an extensive and systematic scheme of inspection and control to detect fraudulent activities. In particular, the classical methods of analysis are mostly not able to detect trace levels, whilst those that can are not suitable for handling large numbers of samples because of their high cost and long analysis time. Despite this, the Commission clearly states that “Before release for free circulation in the Union, each consignment of imported animal by-products of non-ruminant origin intended for the production of processed animal proteins for use in feedingstuffs for non-ruminant farmed animals shall be sampled and analysed to verify the absence of proteins of ruminant origin.” At the same time, the need for a quantitative or semi-quantitative method of analysis in order to implement these controls is acknowledged.2 Light microscopy is currently the only method authorised for official inspections. However, this method is not without its limitations, particularly when used for the routine inspection and monitoring of large volumes of PAP; one insurmountable constraint, if the EU ban on intraspecies recycling is to be partially lifted, is its inability to detect and quantify the component species3,4 in feeds. As part of their active involvement in the EU STRATFEED and SAFEED-PAP projects,5 the authors have focussed on the use of NIR technology—also known as single-point or bulk spectroscopy—either alone or in conjunction with other techniques (microscopy and image analysis), to enhance or, indeed, replace light microscopy in the detection and quantification of animal species in compound feeds and PAPs. Research has shown that NIR technology can not only identify the various species present in a blend of PAPs,6,7 but also determine the percentage of each species in the mixture;8 to date, however, the error-rates displayed by prediction models have not been sufficiently low to enable the identification of trace amounts of adulteration of one species by another. Nonetheless, NIR spectroscopy could be used as an effective screening technique, with enormous throughput potential (hundreds of samples per day), for the instant detection of suspicious samples which could later be analysed using other forensic methods, including a NIR-imaging combination. In conventional NIR spectroscopy, a spectrum reflects the integrated spectral information of the sample surface, which depends on the spot size generated by the beam of light; the analyst, therefore, acquires only partial information on the sample. In the late 1990s, scientific instruments became available which combined digital representation and molecular spectroscopy or hyperspectral chemical imaging (HCI), thus enabling spectral and spatial information to be acquired simultaneously. Hyperspectral chemical imaging is a growing field of research for agri-food applications.9–13 Analysis of the literature in this field shows that many applications make use of sensors which measure in the so-called “visible and near infrared” (vis-NIR) region (
