A multicenter evaluation of dysthyroxinemia in a defined patient cohort
2014
BACKGROUND: In the region Limburg (The Netherlands) almost all of the five participating laboratories use a different immunoassay platform to determine thyroid stimulating hormone (TSH) and free thryoxine (FT4). With the frequent transfer of patients within the region, harmonization of test result interpretation is necessary. In this study, we investigated dysthyroxinemia classification between participating laboratories and developed procedures for improvement. METHODS: Two ring surveys with an interval of 2 years were performed. Four patient groups (n=100) with different dysthyroxinemia classification were based on biochemical results of the Autodelphia analyzer. Samples were tested in five participating laboratories. In each group the percentage of patients classified with dysthyroxinemia was calculated and differences were analyzed by the Fisher's exact test. RESULTS: After the first survey, the percentage of patients with hyperthyroxinemia was more than 20% lower in three laboratories compared to the other two. Bhattacharya analysis revealed that the upper reference limit of FT4 was 20%-30% too high in two laboratories. Adjustments of reference ranges appeared to be effective in the second survey. The third laboratory reported significantly lower percentages of patients with hyperthyroxinemia in the second survey. New FT4 reference ranges were determined for this laboratory, resulting in adequate classification of hyperthyroxinemia. CONCLUSIONS: This study illustrates the potential of a multicenter evaluation of dysthyroxinemia in a biochemical-defined patient cohort. In particular, classification of hyperthyroxinemia differed between laboratories. Adjustments of reference ranges resulted in better agreement of dysthyroxinemia classification. Even using internal and external quality assurance programs, application of multicenter ring surveys is advised to prevent inadequate reference ranges.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
2
References
2
Citations
NaN
KQI