Five immunochemical assays for determining low concentrations of albumin were investigated. These were a radioimmunoassay (RIA); turbidimetric immunoassays (TIA) both according to end-point measuring principle on a Cobas Fara and Hitachi 717 analysers, and according to kinetic measuring principle on a Turbitimer instrument; and a nephelometric immunoassay (NIA). All achieved the analytical goal necessary for optimal patient care. The correlations between the albumin concentrations measured with the different techniques were very good. In vitro glycation of albumin did not influence albumin concentrations measured by the five assays. Urine albumin excretion measured over 3 consecutive days showed considerable day-to-day variation. This was highest for spot-urine specimens and significantly lower for 24 h and timed-overnight samples. Variation of storage temperature (room temperature, 4 degrees C, -20 degrees C), time (up till 3 months), and pH (within the range pH 5-8) of the urine samples did not change significantly the measured albumin concentrations. Different sample preparations (vortex-mixing, centrifugation, and thawing) had no influence on the measured albumin concentration. In conclusion, a maximum standardization of the collection of timed-overnight urine samples for screening and 24 h urine samples for confirmation of microalbuminuria during 3 consecutive days is more crucial than the choice of the immunological technique.