Growth hormone (GH) assays are designed to determine the biological effectiveness of GH in a variety of preparations or in biological fluids (e.g. blood samples). Assays are important for the standardization of different GH preparations as well as for the evaluation and diagnosis of clinical disorders of growth. As shown in Table I , GH assays can be divided into four types.Two classical bioassays for GH were established over 40 years ago. The first depends on the body weight gain seen in hypophysectomized rats following a 10-day period of daily injections of GH. In the second, the 'tibia assay', the widening of the cartilage growth plate of the tibia of hypophysectomized rats is determined following four daily injections of GH. Purified or recombinant GH preparations are standardized on the basis of these two assays (1). Unfortunately, both assays have low sensitivity. Therefore, these assays are not suitable for the routine measurement of GH in serum samples.The development and application of radiommunoassays (RIAs) for GH dramatically changed our understanding of the secretion of GH in health and disease because, for the first time, circulating levels of GH could be measured relatively easily in a large number of samples. Using RIAs, basal levels of GH in the circulation were established, a sleep-related peak of GH was found, and hypoglycaemia and arginine infusion were found to be potent stimuli of GH secretion. Pharmacological agents, such as L-dopa-propranolol, were also shown to be potent stimuli for GH secretion. The RIA measurement of serum GH levels following provocative stimuli was widely used as a diagnostic test for GH deficiency (2). Whilst GH immunoassays were convenient and appeared satisfactory in identifying GHdeficient patients, a lingering concern persisted that, at least in some patients, there were discrepancies between the levels of the biologically active form of GH in the circulation and the immunoassay estimates. The most dramatic example of this was found by Ellis et al.. who reported much higher values of GH by the tibia assay than by RIA of plasma samples (3).In the absence of a specific and sensitive bioassay that matched the sensitivity of RIAs, the Table 1. Type of GH assay. Description Sensitivity (ng) Bioassays ( i l l virro) Hypophysectoniized rats Body weight gain Tibia width Radioimniunoassay Radioreceptor assay (liver) IM9 lymphocytes Receptor modulation assay NB2 node lymphoma cells (lactogenic)5000 ( X 10)' 2000 ( X 4)' 0.5 20.0 10.0 2.0 0.5 'Numhers in parentheses indicate that GH must be injected in the amounts shown for 10 or 4 days. respectively.