Autor(es):
Ratola, N. 
; Barros, P. ; Simões, T. ; Cerdeira, António ; Venâncio, Armando ; Alves, Arminda
Data: 2006
Identificador Persistente: http://hdl.handle.net/1822/5704
Origem: RepositóriUM - Universidade do Minho
Assunto(s): Ochratoxin A; Wines; Interlaboratory study; z-Scores; Outliers
Descrição
Interlaboratory studies are decisive tools to help the validation of a specific analytical methodology or to assess the reproducibility of the use of different methods to analyze a given compound or compounds in certain sample matrices. In this work, homogeneous samples of two white wines (“White Wine” and “White Liqueur Wine”) and one red wine (“Red Fortified Wine”) from Portugal with different production techniques and characteristics, namely in alcohol strength (10.5%, 16.0% and 19.0% ethanolic content, respectively), were analyzed for their contents in ochratoxin A (OTA), a mycotoxin generated from fungal contamination. White Liqueur Wine was naturally contaminated, whereas the other two wine type were spiked with ethanolic OTA solutions. The participation of 24 laboratories from 17 countries of five continents was ensured for this study. Although with no restrictions in terms of analytical methodology to employ, 75% of the laboratories resorted to immunoaffinity columns clean-up followed by high performance liquid chromatography with fluorescence detection (HPLC-FD), most of them in accordance with the European Standard EN 14133. For White Wine samples, the general mean OTA concentration was 1.96 μg/l (two outliers) with interlaboratorial standard deviation (sL) of 0.53 μg/l; for White Liqueur Wine, mean of 1.59 μg/l (one outlier), with sL = 0.59 μg/l; and for Red Fortified Wine, mean of 2.73 μg/l (no outliers), with sL = 0.96 μg/l. Outliers were determined by Cochran and Grubbs tests. The Horrat index, recommended by the Association of Official Analytical Chemists (AOAC) for the quality assurance of the collaborative study was, on average, 1.7. This study proved that OTA determination in wines is reproducible, regardless of the methodology employed.
