Contribution of Sample Processing to Variability and Accuracy of the Results of Pesticide Residue Analysis in Plant Commodities A ́ rpa ́ d Ambrus,* ,† Judit Buczkó , § Kamira ́ nA ́ . Hamow, ⊗ Viktor Juha ́ sz, ⊗ Etelka Solymosne ́ Majzik, ⊗ Henriett Szema ́ nne ́ Dobrik, § and Ró bert Szita ́ s § † Retired Scientific Adviser, Hó mező u 41, HU1221 Budapest, Hungary § National Food Chain Safety Office Directorate of Plant Protection, Soil Conservation and Agri-environment Pesticide Residue Analytical Laboratory, Miskolc, Blaskovics u 24, H-3500 Miskolc, Hungary ⊗ National Food Chain Safety Office Directorate of Plant Protection, Soil Conservation and Agri-environment Pesticide Analytical Laboratory, Velence, Orsza ́ g u 23, H-2481 Velence, Hungary ABSTRACT: Significant reduction of concentration of some pesticide residues and substantial increase of the uncertainty of the results derived from the homogenization of sample materials have been reported in scientific papers long ago. Nevertheless, performance of methods is frequently evaluated on the basis of only recovery tests, which exclude sample processing. We studied the effect of sample processing on accuracy and uncertainty of the measured residue values with lettuce, tomato, and maize grain samples applying mixtures of selected pesticides. The results indicate that the method is simple and robust and applicable in any pesticide residue laboratory. The analytes remaining in the final extract are influenced by their physical-chemical properties, the nature of the sample material, the temperature of comminution of sample, and the mass of test portion extracted. Consequently, validation protocols should include testing the effect of sample processing, and the performance of the complete method should be regularly checked within internal quality control. KEYWORDS: pesticide residues, uncertainty of pesticide residue results, stability of pesticide residues, internal quality control, method validation ■ INTRODUCTION Pesticide residues are unevenly distributed in/on treated objects, and hundred-fold differences were found in field- treated natural crop units. 1-3 Consequently, sample processing, including disintegrating by chopping, grinding, or blending the sample material, may not result in a homogeneous matrix. Therefore, test portions withdrawn from the “homogenized” laboratory sample may contain varying residues. The uncertainty, expressed as relative standard deviation, of the pesticide residue concentrations comprises the variability deriving from the inhomogeneity of the analyte in the “well- mixed” laboratory sample after the sample processing step (CV SP ) and the analysis of residues (CV A ) present in the test portions withdrawn from the processed laboratory sample. The combined uncertainty of the laboratory phase (CV L ) of the determination of pesticide residues, provided that the entire laboratory sample is homogenized, can be described as = + CV (CV CV ) L Sp 2 A 2 (1) The variability of residues in test portions is substantially affected by the particle size distribution of the homogenized matrix. 4 The sampling constant (K s ) defined by Wallace and Kratochwil 5 describes the relationship between the mass (m) of single increments (TP, test portion) to be withdrawn from a well-mixed material and the relative standard deviation (CV) of the analyte concentration in repeatedly withdrawn test portions: = × K m CV s TP 2 (2) Maestroni and co-workers, 6 applying 14 C-chlorpyrifos, exam- ined the effect of sample material and different chopping devices and concluded that the K s value depends on the type of equipment used for comminution of the sample material, the texture of the sample, and the temperature of comminution (ambient or deep-frozen). For instance, withdrawing 10 g of orange and apple test portions from sample matrix comminuted in the presence of dry ice would result in, on average, CV Sp values of 12 and 15%, respectively. These findings were also confirmed in the case of cucumber. 7 Tomato represents one of the worst cases because of its hard peel and soft pulp. To keep the CV Sp between 6 and 7%, 110 and 5 g of tomato should be homogenized with a vertical cutter mixer at room temperature and cryogenic processing applied (< -20 °C), respectively. 8 A two-step procedure comprising chopping the whole laboratory sample material with a blender and further homogenizing about 400-500 g portions with Ultra Turrax can reduce the CV Sp to about one-third, compared to the one-step procedure. 6 The extended homogenization process may have an undesirable adverse effect on the accuracy of the results, namely, the decomposition of sensitive analytes. This phenomenon has been observed a long time ago in the case Special Issue: 52nd North American Chemical Residue Workshop Received: December 6, 2015 Revised: December 30, 2015 Accepted: January 11, 2016 Article pubs.acs.org/JAFC © XXXX American Chemical Society A DOI: 10.1021/acs.jafc.5b05779 J. Agric. Food Chem. XXXX, XXX, XXX-XXX