Background Thyroglobulin (Tg) assay in washout fluids of fine needles, after cervical lymph nodes aspiration, is used for detecting metastases from differentiated thyroid carcinomas. Assay methods are the same as for Tg in serum. However, with non-serum samples, methods require extensive validation to notably check for the absence of matrix effect. This study fits this context. Our objectives were to assess analytic performances, in washout fluid, of eight different Tg assay methods and to compare them to validated data in serum. Methods Eleven medical laboratories participated in this study. The matrix tested was phosphate-buffer saline containing 1% bovine serum albumin (PBS-1% BSA). Samples used were dilutions, in this buffer, of Certified Reference Material (CRM 457). We verified, for all methods, the limit of detection, precision, linearity, trueness and accuracy. Results In PBS-1% BSA, the functional sensitivities (FS) were comparable to those expected for serum. All the methods were linear. The relative biases of trueness were between –24.5 and 10.2% around 1 µg/L. Total analytical error was ≤40% near the functional sensitivity values. Conclusion No quantitatively important matrix effect was observed. All the methods showed their ability to measure Tg in PBS-1% BSA, over the concentration range of interest, with acceptable total analytical error. We validated the functional sensitivity value as a decision threshold in thyroidectomized patients after treatment and with low concentrations of serum Tg.
“Sampling vs analytical error: where the money is …”
