The performance characteristics of a radioimmunoassay employing a 125I-labelled tracer for determination of calcitriol in human sera are reported. The assay is based on an immunoextraction step employing a monoclonal antibody against calcitriol followed by a radioimmunoassay (using 125I-labelled calcitriol and sheep antiserum against calcitriol). Bound/free separation is performed with an anti-sheep IgG antibody bound to cellulose. Within-run imprecision (n = 20) was 12.8% (mean = 9.4 ng/l) and 11.1% (mean = 48.8 ng/l), between-day imprecision (n = 11) was 27.1% (mean = 9.6 ng/l) and 17.2% (mean = 47.2 ng/l). Linearity of dilution as investigated by mixing pooled sera containing 62.0 ng/l and 4.7 ng/l calcitriol, respectively. The relationship between measured and expected concentrations was characterized by a linear correlation coefficient of r = +0.990. The values obtained with the 125I-based radioimmunoassay were compared with those obtained with a radioreceptor-assay using a 3H-labelled tracer; the regression line was y = 1.091 x -4.545 (n = 84; r = +0.935), where y = calcitriol [125I] [ng/l] and x = calcitriol [3H] [ng/l]. Mixing 9 volumes of sera from patients with renal insufficiency (n = 7) with 1 volume of 'calibrator F' (assigned value: 227 ng/l) yielded recovery rates of 90 +/- 8% (mean +/- SD). The detection limit was 3.0 ng/l. The cross-reactivity of cholecalciferol metabolites was found to be < 0.00003 for 25-hydroxycholecalciferol, 24R,25-dihydroxycholecalciferol and 25S,26-dihydroxycholecalciferol. A preliminary reference interval (5th to 95th percentile) was established in 40 apparently healthy persons (17 males and 23 females; age range: 20-61 [mean: 32] years) (19-74 ng/l). The method presented shows high practicability and may therefore be considered as a useful alternative to cumbersome assays using 3H-labelled tracers.