Ultrafiltration method for direct radioimmunoassay measurement of free thyroxine and free tri-iodothyronine in serum.

Ultrafiltration at physiological pH and temperature of undiluted serum followed by direct radioimmunological determination of T3 and T4 in the protein-free ultrafiltrate offers the best possible approach towards estimation of in vitro plasma levels of free T3 and free T4. The major technical difficulties in meeting this apparently simple proposition are: establishing adequately sensitive radioimmunoassays; avoidance of adhesion to ultrafilters and glassware; removal from the ultrafilters of compounds which would cross-react or interfere in the radioimmunoassays; and avoidance of co-filtration of thyroid hormone binding proteins in serum, which would obviously imply spurious data. This methodological study describes the magnitude and significance of each of these obstacles and how to circumvent them. Practically all other available methods, including equilibrium dialysis, imply dilution of serum samples with buffer often leading to alterations in ionic composition to which thyroid hormone binding to proteins is peculiarly sensitive. Dilution itself alters the fraction of free thyroid hormones in serum especially when pharmaca or compounds are present which compete for the binding sites. These pitfalls are avoided in ultrafiltration of undiluted serum. This is illustrated through measurements on serum containing therapeutic concentrations of Fenclofenac which was found to displace 120% more T4 in undiluted than in diluted (1:28) serum. Using the described technique FT3 was 8.8 +/- 1.7 pmol/l and FT4 30.8 +/- 8.2 (SD) pmol/l in serum from 29 normal subjects. Pregnant women in their third trimester had lower levels: FT3 7.1 +/- 2.1 and FT4 17.6 +/- 5.8 pmol/l (SD, n = 24).