Statistical Analysis for Quality Adjustment of the Analytical Curve for Determination of Pesticide Multiresidue in Pineapple Samples

This work describes the development of an analytical methodology based on the modified QuEChERS method and GC/MS detection involving the analysis of 38 pesticides in pineapple from a local market in Fortaleza-Brazil. Matrix-standard calibration and heteroscedasticity have been studied by a weighted least squares linear regression model application. The analytical curves for most of the analytes (except cyproconazole and pyraclostrobin) presented heteroscedasticity by weighted least square regression (WLS). The weighted correlation coefficient values were in the range of 0.9690 to 0.9987 (linear range of 0.03 to 0.50 mg kg(-1)). The calibration curves obtained by applying regression by ordinary least squares (OLS) showed relative errors of approximately 32.3% for fenpropathrin, 23.1% for buprofezin, 17.4% for methyl parathion, and 13.9% for ametryn. Analysis of commercial pineapple samples revealed the presence of methyl parathion, buprofezin, ametryn, and fenpropathrin at 0.023 +/- 0.0037, 0.013 +/- 0.0047, 0.043 +/- 0.0156, and 0.031 +/- 0.0049 mg kg(-1), respectively.