Desenvolvimento e validação de metodologia multirresíduos por cromatografia líquida acoplada a espectrometria de massas para análise de fluorquinolonas em músculo e rim de frangos

Abstract

This work involved the optimization and validation of two methods for determining fluoroquinolones in samples of muscle and kidney of chicken: nalidixic acid, oxolinic acid, pipemidic acid, danofloxacin, difloxacin, ciprofloxacin, enrofloxacin, flumequine, marbofloxacin, norfloxacin, ofloxacin and sarafloxacin. The QuEChERS extraction procedure was used for both methods, whose optimization for extraction of 2.0000 g of the sample was based on the use of a Box- Behnken factorial design 33 with statistical evaluation by ANOVA. As extraction phase it was used 10.0 mL of acetonitrile acidified with 5 % glacial acetic acid to chicken kidney, and to chicken muscle it was employed a mixture of 10.0 mL of the previous extraction phase with ultrapure water (80:20 % v/v). The clean-up step was performed by dispersive solid phase extraction, using 50.0 mg of a mixture of C18 and PSA sorbent (50:50 % m/m). The first validated method involved the use of LC-MS/MS technique for separation, identification and quantification of the twelve analytes by evaluating each fluoroquinolone using a transition mass for confirmation and another for quantification by multiple reaction monitoring mode. The figures of merit measured during validation showed good performance, evaluating accuracy (recoveries above 88.5%), linearity (R2 > 0.931), precision (coefficient of variation less than 12.3%), besides appropriate selectivity and low values of expanded uncertainties. The second validated method employed the high performance liquid chromatography coupled to high resolution mass spectrometry technique, in which the identification and quantification of nine fluoroquinolones was performed in the single ion monitoring mode. The figures of merit were also evaluated as satisfactory, with recoveries above 87.0%, coefficients of determination higher than 0.960, coefficients of variation less than 8.7%, and good selectivity and low values of expanded uncertainty. The two methods were statistically compared using the unpaired t Student test and pooled variance, at the confidence level of 95 %, and showed no significant difference.