[Advantages, disadvantages and selection methods of thyroxine measurement--comparison of CPBA and RIA].

Although recent progress in radioimmunoassay (RIA) permitted direct RIA for thyroxine (T4), there appeared few reports evaluating clinical applicability of T4 RIA as routine method. The authors evaluated advantage, disadvantage of methods for T4 measurement in comparison between CPBA (competitive protein binding analysis) and RIA, and tried to discuss about selection of methods for T4 measurement. Res-O-Mat T4 was chosen as CPBA, and RIAMAT-T4(St) and T4-RIAKIT(Sp) were chosen as RIAs. Detailed descriptions about both RIA methods appeared elsewhere [2, 3]. From the basic experiments, reproducibility of CPBA, St and Sp within assay were 5.8% (C.V.), 4.0% (C.V.), and 8.5%(C.V.) respectively, and mean recovery rates were 92.2%, 100.7% and 96.1% respectively. T4 value measured by both RIAs correlated well with CPBA, and St showed slightly lower value and Sp showed slightly higher value than CPBA. The normal ranges for RIAs were 3.7-11.7 microng% (St) and 4.6-15.2 microng% (Sp). Diagnostic accuracy for hyperthyroid state were 83.3% (CPBA), 85.7%(St) and 69.2% (Sp), and for euthyroid state were 92.0% (CPBA), 98.8% (St) and 95.7% (Sp). The diagnostic accuracy for hypothyroid state in three methods were lowest. From the economical point of view, cost performance, time performance and total time for single assay were analysed and mathematical equations were obtained from actual assay by each methods. When the number of samples per single assay was less than 10, CPBA was most economic in terms of cost and time performance and total time required. In the assay of more than 20 samples, RIAs were efficient in each point of views. In the point of view of sample volumes required, RIAs were most promising. The authors concluded that guides for selection of methods in T4 assay were as follows: A) CPBA was most efficient in local use especially when less than 10 samples were frequently to be assayed. B) RIAs were efficient in batch processing more than 20 samples, especially in automated laboratories.