We describe an improved method for the determination of islet amyloid polypeptide (IAPP) levels in plasma. Plasma is first extracted with acid-acetone, followed by a specific and sensitive radioimmunoassay (RIA) for IAPP using rabbit-anti-human-IAPP serum. Recovery of synthetic IAPP from plasma was 82±6% (n = 16). Standard samples, prepared in ‘hormone-free’ serum, were also extracted with acid-acetone. Displacement curves of serially diluted acid-acetone extracted plasma samples were parallel to the standard curve. The lower detection limit of the RIA was 2·3 ± 0·1 fmol/sample (n = 5). Intra-assay variations for IAPP concentrations of 4, 17 and 32 pM were 16·3% (n = 10), 9·2% (n = 10) and 6·2% (n = 10); inter-assay variations were 35·9% (n = 14), 19·9% (n = 15) and 15·4% (n = 15), respectively. Non-stimulated IAPP levels ranged from 2·4 to 12 pM (mean 6±4pM, n = 10) in healthy control subjects. IAPP was not detectable in type 1 (insulin-dependent) diabetic patients before and after glucagon administration. In type 2 (non-insulin-dependent) diabetic patients basal levels ranged from 2·2 to 14·5 pM and glucagon-stimulated levels ranged from 2·2 to 38·9 pM. The increase in IAPP varied from 0 to 24·4 pM. The anti-human-IAPP serum had full cross-reactivity with rat IAPP (= mouse IAPP). Transgenic mice overexpressing the human IAPP gene showed elevated plasma IAPP levels as compared to (non-transgenic) control mice. It is concluded that the method presented for the determination of IAPP in plasma is reliable and easy to perform, yielding reproducible results. The ability to measure plasma IAPP levels will help to elucidate the physiological function of IAPP as well as its pathophysiological role in type 2 diabetes mellitus.