European proficiency study with control serum for the tumor marker CA 19-9 measured on different test systems.

BACKGROUND: Reliable and precise CA 19-9 testing is required for the long-term follow-up of patients with pancreatic carcinoma during therapy. The aim of this longitudinal proficiency study was to evaluate the comparability, linearity, and precision of CA 19-9 determinations performed in different laboratories using currently available test systems under routine conditions. METHODS: During the one year study period, 15 laboratories applied 7 different tests and included a liquid BIOREF control serum with pancreatic carcinoma derived CA 19-9 in their routine testing and quality control procedures. The results were collected centrally and evaluated statistically. RESULTS: The comparability of CA 19-9 results is limited especially when different tests are used, albeit, some tests show a good correlation: The CA 19-9 values obtained by different laboratories using different test systems vary up to a factor of 2. The precision of CA 19-9 determinations was acceptable in most laboratories with coefficients of variation ranging between very low 3.2% and high 17.8%. The imprecision was slightly increased when automatic dilution procedures of the analysers were used. CONCLUSIONS: The comparability of CA 19-9 test results must be improved. The precision is acceptable in most cases. In order to monitor key performance parameters, every laboratory should participate in external quality assessment schemes and should perform a routine internal quality control with a control serum independent from the test kit manufacturer.

Precision and long-term stability of different estradiol immunoassays assessed in a multi-center quality control study.

BACKGROUND: Because of the vast range of physiological relevant estradiol concentrations the requirements to be met by an estradiol assay are high. In the present study the performance of various commercially available estradiol assays was evaluated with regard to imprecision and long-term stability. METHODS: Precision and long-term stability of 7 commercially available estradiol immunoassays were assessed in a multi-centre quality control study based on the repeated measurement of liquid BIOREF estradiol control sera by 18 laboratories during a 14-month study period. RESULTS: The mean estradiol concentrations determined in 594 runs performed for each control level were 71 pg/ml, 349 pg/ml and 676 pg/ml. A high variation was found for the method specific mean values calculated from all results measured with the same method, which ranged between 32 - 90 pg/ml, 187 - 392 pg/ml and 373 - 790 pg/ml, resulting in a similar high inter-laboratory variation with coefficients of variation (CVs) of 25.0%, 16.7% and 17.5%. In contrast, the intra-laboratory variation of estradiol values as well as the variation of values measured with the same method were found to be considerably lower with coefficients of variation < 10% for most laboratories and methods; only the low control level was measured with CV values > 10% by the majority of laboratories and methods. For none of the laboratories a tendency was observed in the results from beginning to end of the 14 month study period indicating a high uniformity in assay production and a good long-term stability of the control material used. CONCLUSIONS: The present data demonstrate that also with the currently available estradiol immunoassays the comparability of results measured with different methods is limited. With most assays very low estradiol concentrations, as observed in postmenopausal women, can be determined only with a precision which is not adequate for clinical assessment.