CDTect-RIA and CDTect-EIA for determination of serum carbohydrate-deficient transferrin (CDT) by radioimmunoassay and enzyme immunoassay respectively were tested for equality and precision in four European laboratories. For correlational studies, serum samples with CDT concentrations up to 130 U/l were analysed in accordance with a uniform trial schedule. The regression of CDT values obtained by the two procedures was computed for each laboratory using the method of Passing and Bablok. Slopes and intercepts of the regression functions did not differ significantly from the values 1 or 0, as proved by the corresponding 95% confidence intervals. Precision studies were computed using analysis of variance. For CDT concentrations at the upper reference limit for men, the within-day coefficients of variation (CVs) ranged between 0.7 and 6.4% (median 5.2%) for CDTect-RIA and from 4.3 to 9.2% (median 6.2%) for CDTect-EIA. The corresponding pure between-day CVs were 5.0-18.5% (median 9.8%) and 3.5-14.5% (median 10.9%). The study demonstrates the equality of CDT values obtained by CDTect-RIA and CDTect-EIA. According to this study, the two methods can be used interchangeably without getting fluctuating CDT values, e.g. in longitudinal studies.