Download the poster

Marchis D., Lai J., Abete M. C., Squadrone S., Palmegiano P., Fragassi S., Decastelli L.
Istituto Zooprofilattico Sperimentale del Piemonte, Liguria e Valle d’Aosta
C.Re.A.A - National Reference Centre for the Surveillance and Monitoring of Animal Feed
NRL for animal proteins in feeding stuffs

Introduction:
Following the outbreak of BSE, processed animal proteins (PAP) were banned in animal feedstuffs in the EU and each individual member state was required to implement a feed quality programme to enforce this ban. An essential aspect of these programmes was the adoption of EU-approved methods for detecting PAP in feed. The official analytical method for the detection of processed animal proteins in feedstuffs is the microscopic examination technique described in Directive 2003/126/EC (EU, 2003). Since 1998, near infrared microscopy (FT-NIR) has been proposed as an attractive alternative method for the detection of meat and bone meal (MBM) in feeding stuffs.  Here is 4 months of routine control results in Piemonte (Italy), performed with both methods.

Materials and Methods:
From 11.05.07 to 10.09.07, 113 routine control samples for  the identification of constituents of animal origin in animal feeding stuffs were carried out both with classic microscopic examination, according to the Directive 2003/126/EC, and with FT-NIR technique, for a qualitative identification of PAP.
MICROSCOPIC ANALYSIS: Each sample was examined using the microscopic method, according to the Directive 2003/126/EC. Sediment and sieved fractions were analyzed. Any potential bone fragments were characterized by similar morphological features (colours, shape, lacunae shape, lacunae distribution, etc.).
FT-NIR ANALYSIS: Samples were processed applying a complete NIR protocol, that was developed and validated in order to analyze the raw and the sediment fractions of the feeding stuffs. The recognition of bone fragments among other constituents was made treating their vibrational spectra by a mathematics algorithm based on least-squares, resulting from the comparison between the spectrum of unknown sample and spectra of reference materials. Spectra examined were obtained in the reflectance mode, with 10000-4000 cm-1 range.  By setting a 95% level of similarity as acceptance threshold, good results were obtained in the detection of bone fragments.

Results:
Microscopic examination detected 4 positive samples for fish, tag declared, and two positive specimens, for not allowed terrestrial animals. These two not compliant samples were avian feed, MBM contaminated. The FT- NIR analysis showed six positive results, as in our experience this method fails to discriminate between fish meal and terrestrial animal (mammals and/or poultry meal).

Conclusions:
The results of analyses of routine samples by NIR and classical microscopy demonstrate that there is no difference between the outcome of both techniques, although the near infrared microscopy does not identify different kind of animal meal. The principal particles of animal origin that might be present in feeds are bones and muscle fibres. Additionally, cartilage, hairs, feather filaments, egg shells, fish scales and ligaments may also be present. Most of these particles show a limited number of characters, and fine structures are visible after microscopic inspection at different magnifications. Thus, besides the animal/non animal recognition, it is very important to distinguish bone tissues of different zoological origin. Experimental samples carried out by using a soft independent modelling of class analogies-SIMCA seemed to show a good differentiation between different species. The next future challenge could be how to apply these techniques to the daily routine samples.

Keywords:
Microscopic  examination, FT-NIR spectroscopy, feed, bone fragments.