Determination of 2,3,7,8 Chlorine-Substituted Dibenzo-P-Dioxins and Furans at the Part Per Trillion Level in United States Beef Fat Using High-Resolution Gas Chromatography/High-Resolution Mass Spectrometry
As part of the U.S. EPA Dioxin Reassessment Program, the 2,3,7,8-chlorine-substituted dibenzo-p-dioxins and furans were measured at part per trillion (ppt) levels in beef fat collected from slaughter facilities in the United States. This is the first statistically designed national survey of these compounds in the U.S. beef supply. Analyte concentrations were determined by high resolution gas chromatography/high resolution mass spectrometry, using isotope dilution methodology. Method limits of detection on a whole weight basis were: 0.05 ppt for TCDD and 0.10 ppt for TCDF, 0.50 ppt for pentas (PeCDDs/PCDFs)/ hexas (HxCDDs/HxCDFs)/ heptas (HpCDDs/HpCDFs), and 3.00 ppt for octas (OCDD/OCDF). Method detection and quantitation limits were established based on demonstrated performance criteria utilizing fortified samples rather than by conventional signal-to-noise or variability of response methods. The background subtraction procedures developed for this study minimized the likelihood of false positives and increased the confidence associated with reported values near the detection limits. Mean and median values for each of the 2,3,7,8-Cl-substituted dioxins and furans are reported along with the supporting information required for their interpretation. The mean toxic equivalence values for the samples are 0.35 ppt (nondetects = 0) and 0.89 ppt (nondetects = ? LOD). All analytical methods employed were validated prior to sample analysis, and rigorous QA/QC procedures were maintained throughout the survey. The LODs/LOQs were developed and verified using the results from the analyses of fortified samples and all of the required information (i.e. precision, accuracy) is provided. This practical approach for the determination of LODs/LOQs based on demonstrated performance criteria differs conceptually and in practice from the conventional methods employing signal-to-noise (S/N) ratios or variability of response. These later methods rely on extrapolations from an instrumental response, most often in the absence of the analyte of interest, and the derived LODs are not verified before being reported.