Use of a quality control approach to assess measurement uncertainty in the comparison of sample processing techniques in the analysis of pesticide residues in fruits and vegetables

Abstract

In routine monitoring of foods, reduction of analyzed test portion size generally leads to higher sample throughput, less labor, and lower costs of monitoring, but to meet analytical needs, the test portions still need to accurately represent the original bulk samples. With the intent to determine minimal fit-for-purpose sample size, analyses were conducted for up to 93 incurred and added pesticide residues in 10 common fruits and vegetables processed using different sample comminution equipment. The commodities studied consisted of apple, banana, broccoli, celery, grape, green bean, peach, potato, orange, and squash. A Blixer® was used to chop the bulk samples at room temperature, and test portions of 15, 10, 5, 2, and 1 g were taken for analysis (n = 4 each). Additionally, 40 g subsamples (after freezing) were further comminuted using a cryomill device with liquid nitrogen, and test portions of 5, 2, and 1 g were analyzed (n = 4 each). Both low-pressure gas chromatography-tandem mass spectrometry (LPGC-MS/MS) and ultrahigh-performance liquid chromatography (UHPLC)-MS/MS were used for analysis. An empirical approach was followed to isolate and estimate the measurement uncertainty contribution of each step in the overall method by adding quality control spikes prior to each step. Addition of an internal standard during extraction normalized the sample preparation step to 0% error contribution, and coefficients of variation (CVs) were 6-7% for the analytical steps (LC and GC) and 6-9% for the sample processing techniques. In practice, overall CVs averaged 9-11% among the different analytes, commodities, batches, test portion weights, and analytical and sample processing methods. On average, CVs increased up to 4% and bias 8-12% when using 1-2 g test portions vs. 10-15 g. Graphical abstract Efficient quality control approach to include sample processing.

Assessment of Test Portion Sizes after Sample Comminution with Liquid Nitrogen in an Improved High-Throughput Method for Analysis of Pesticide Residues in Fruits and Vegetables

“…13 All pesticides found match those often reported elsewhere in the given commodities. 13 Despite higher variabilities of the incurred pesticide results, largely due to lower concentrations, 4 incurred and spiked pesticides were compiled together in the study to yield average RSD Proc used for calculations of CV Proc .…”

“…The cleanup was performed just-in-time for injection in parallel as the previous cleaned-up sample was being analyzed via LPGC-MS/MS as described previously. 4,5,[24][25][26] Preparation of calibration standards.…”

“…As described in Materials and Methods and previously, 4,5 the RSDs of the results were used to calculate the isolated CV components in the method(s) to estimate measurement uncertainties with and without use of an int. std.…”

“…Recently, a series of articles have been published in an attempt to change that oversight among analytical chemists in the pesticide residue analysis community. [1][2][3][4][5][6][7][8][9][10] Regulatory purposes require that bulk commodities typically >1 kg be collected and processed to yield pesticide determinations that accurately correspond to the original sample lot (e.g. shipment or warehouse).…”

What are the Main Sensor Methods for Quantifying Pesticides in Agricultural Activities? A Review

“…They concluded that the preparation of the sample impacts more significantly the results obtained by any method evaluated. In addition, they evaluated the sample size, between 1–15 grams, and concluded that the bias increased by ≈10% when using 1–2 g portions vs. 15 g portions [125].…”