Thyroid function in pregnancy: Implications and applications of immunoassay design

Serum thyroid stimulating hormone (TSH) levels and the free thyroxine (FT4) concentration have both been shown to provide generally useful indices of thyroid function in many pathological and physiological conditions. In particular, measurements of these hormones in the maternal circulation during pregnancy should establish the range of normal maternal thyroid function. They may also be useful in clarifying the role of the maternal thyroid in fetal development. However the results of studies in this area have been inconsistent, largely due to inaccuracies of the analytical methods employed. The sensitivity and specificity of conventional immunoassay methods for TSH have not been adequate to provide reliable measurement of small changes in TSH concentration, particularly in the presence of relatively high concentrations of HCG. The accuracy of free thyroxine assay methods has also been questionable in circumstances, such as pregnancy, where the serum levels of binding proteins (and perhaps also of circulating binding inhibitors) vary widely. Labelled antibody immunoassay techniques have been claimed to provide higher potential sensitivity than that attainable by conventional radioimmunoassay. This claim has attracted equivocal experimental support and has been contested on theoretical grounds. Previously developed mathematical models of labelled antibody immunoassays have served as a useful focus for discussion but have failed, in practice, to offer useful guidelines to the design of an assay or to the thermodynamic characteristics required of the antibodies used in this context. The objectives of this investigation have therefore been. a. to establish a theoretical basis for labelled antibody immunoassays which identifies the critical parameters limiting their sensitivity b. to test the utility of this theoretical model by isolating appropriate antibodies and developing maximally sensitive and specific labelled antibody immunoassays for TSH c. to examine the theoretical basis and practical validity of several alternative FT4 assay methods d. to study the levels of both these analytes in pregnancy using a highly sensitive, specific TSH assay and a well-validated free thyroxine assay. In the course of this study, I have developed a theoretical model of noncompetitive labelled antibody immunoassays which shows that the sensitivity of this method is limited largely by the ability to detect the label used, and that, using labels of high specific activity, it may be several orders of magnitude greater than for conventional labelled analyte methods. In light of these physico-chemical considerations, I evaluated a range of antibodies and developed sandwich assays for TSH. However, HCG significantly interfered with the most sensitive assay configurations at levels of this hormone commonly found in pregnancy. To continue the study, I identified an appropriately sensitive and specific immunofluorometric assay (IFMA) for TSH (subsequently developed by a commercial manufacturer in accordance these principles) using monoclonal antibodies developed elsewhere. This kit was found to be both highly sensitive and sufficiently specific for use in the study. A review of principles governing serum free hormone measurement reveals theoretical and practical advantages for a simple "two-step" method as compared both with more recently-developed analogue-based kits and equilibrium dialysis/RIA. In a detailed comparative study undertaken for the UK Department of Health, I found that the two-step FT4 RIA yielded results which correlated well with those by ED/RIA in serum samples from subjects in a variety of physiological and pathological states. I used the commercially-available ultra-sensitive TSH IFMA and the in- house two-step FT4 RIA in a study of the levels of these hormones throughout pregnancy. The raised FT4 levels and lowered TSH levels I found in women early in pregnancy supports the view that other thyroid stimulating factors are present in the maternal serum during pregnancy. Indirect evidence supports the contention that HCG, per se, is not the putative additional thyroid stimulator