ASSESSMENT OF VALIDITY OF STEROID RADIOIMMUNOASSAYS

According to the generally accepted definition an assay is valid if the recognized criteria of validity are fulfilled, such as specificity as assessed by cross-reaction studies, parallelism of the dose response relationship for the standard an unknown, and agreement of results with the physiological state of the subject. However, these criteria are not stringent enough for the radioimmunoassay of steroids. According to another definition an assay is valid only if it is intrinsically completely pure i.e. if the substance measured is structurally unique, homogeneous and identical with the standard. However, this definition of validity has a limited practical meaning since probably no steroid radioimmunoassay of samples of biological origin can fulfil this condition. As a practical solution to this problem it is proposed to couple the term validity to the outcome of the assays and to define a valid assay as the one which yields results identical with the true value within the limits of experimental error. It is suggested that in practice, when various antisera and purification techniques are compared, the assay procedure yielding the lowest estimates may usually be assumed to be closest to validity. However, the experiments of this type cannot result in a proof of validity. Such a proof may be obtained by a validity test based on a comparison of results yielded by the radioimmunoassay tested for validity on the one hand, with those obtained by a reference radioimmunoassay performed on the sample purified to radiochemical purity on the other hand. Since the radiochemically pure sample can be assumed to represent the true value, the reference assay is valid. The assay tested is to be considered valid as well, if the results yielded by this assay cannot be distinguished statistically from those obtained by the reference assay. The advantage of this test of validity is that it can be performed in every radioimmunoassay laboratory. The above concepts are illustrated by references to previous studies on progesterone, testosterone and norethisterone assays. An example of the validity test for a progesterone radioimmunoassay is given. © 1979.