We compared concentrations of antithrombin III (AT-III) in plasma, as determined by an immunological method and by a functional thrombin inhibition method, in the presence of heparin in 160 blood samples from Type I diabetics. Although the correlation was highly significant (P less than 0.001) between the results obtained by the two methods, our data demonstrated that results by the thrombin inhibition assay, 121 (SD 15)%, expressed as percentages of the results for a normal plasma pool, were significantly (P less than 0.001) higher than by the immunoreactive method, 104 (SD 15)%, indicating an overestimation of functionally active AT-III. Concentrations of functionally active AT-III determined by a factor Xa inhibition assay, 105 (SD 13)%, were in the same range as immunoreactive AT-III. Addition of IgG antiserum to normal pooled plasma quenched only about 90% of the AT-III activity determined by the thrombin inhibition assay, but all of the AT-III activity determined by a factor Xa inhibition assay. These results demonstrate that the factor Xa inhibition assay is more specific for the determination of AT-III than the thrombin inhibition assay. We suggest that the high concentrations of heparin cofactor II, 117 (SD 17)%, might have caused an overestimation of AT III in this group of patients with diabetes Type I, and should not be overlooked in other clinical situations.