Nonspecific binding as a source of error in thyrotropin radioimmunoassay with polyethylene glycol as separating agent.

Abstract

Abstract We investigated the effects of nonspecific binding on thyrotropin values obtained by radioimmunoassay in which polyethylene glycol is used as precipitant. Differences in nonspecific binding among individual samples were significant (F-test, p less than 0.001, range 5.5 to 14.1%). Non-specific binding and total serum protein were directly correlated (r = 0.472, n = 59; p less than 0.001). Nonspecific binding increased with increasing concentrations of globulins but showed no relation to albumin concentration. If globulin concentration was less than 15 g/L, precipitation of the antigen—antibody complex by polyethylene glycol was incomplete. The mean value for thyrotropin in sera from 67 healthy subjects was 2.7 (SD 0.3) milli-international units per liter (milli-int. unit/L) without individual serum nonspecific binding correction, significantly (p less than 0.005) higher than that with nonspecific binding correction (1.6, SD 0.1, milli-int. unit/L). Evidently, inter-sample variations in nonspecific binding may cause significant errors under these conditions, which can be minimized by taking into account the individual nonspecific binding of each serum sample.