Four functional assays for protein S were evaluated by 4 different laboratories, each center using its own method. The aim of this study was to compare these different assays and to establish a relationship with results of immunological assays of total and free protein S antigen and C4bBP. The same plasma samples were distributed to each center and tested in blind. In 47 normal subjects, there was no significant difference between the 4 functional assays, with mean values ranging from 93 to 100%. These values were in good agreement with those of free and total protein S antigen. In 34 patients with a quantitative congenital deficiency of protein S the mean values of protein S activity were decreased with the 4 assays, ranging from 25 to 40%. Free protein S antigen was reduced to a similar extent, whereas total antigen was either normal or decreased. The correlation of protein S activity with free protein S antigen was satisfactory for 3 methods, with coefficients of correlation varying from 0.84 to 0.92 whereas it was only 0.70 in one lab. When total protein S antigen was reduced, protein S activity was decreased in all the patients with the 4 assays. In contrast when total protein S antigen was normal an important overlap of protein S activity between normals and patients was observed in one lab with 12 patients misclassified. In 8 patients with a functional defect, results of protein S activity differed substantially according to the assay used and about half of these patients were misclassified.(ABSTRACT TRUNCATED AT 250 WORDS)