Summary
Bovine thrombins (Gel-T) isolated from bioactivated crude prothrombin, obtained from
Holstein, Jersey and Hereford single animal plasmas, occur in two distributions: either
T1, T2 and T3 as 1:2:1 (Type I), or T1 and T2 as 1:1, with T3 absent or present to the extent of 5% of total Thrombin (Type II). Since components
differentially denature during procedures used to fractionate, examinations of Gel-T
were conducted so that statistical procedures could be used to evaluate particular
variances. The variance due to the Gel-T preparative procedure and Gel-T intrinsic
specific activity yield a coefficient of variance (Sc) of 0.55%. Thus 95% of single animal Gel-T specific activities fall within 1.1% of
the average value. There are no significant differences between animals of different
breed, or of bloods drawn from the same animal at different times. Evidently, Gel-T
represent a reproducible reference standard. Analysis of distributions shows that
the intrinsic specific activities of thrombin components are identical to within 1
%, and are approximately 1.0 NIH unit per 6.0 × 10−4 absorbance unit (3,250 ±450 NIH units per mg). The clotting time (τ) - thrombin concentration
(T) relation, log τ = — ã log T + C, is examined. Over the range of τ from 15 to 60
sec it is shown that a T series can be reproduced to within 1 % in thrombin concentration.
The variances in the assay system are then due almost entirely to variance in the
assay substrate, fibrinogen. These are an Sc of 1.7% for τ̄ between sets of tubes containing dilute fibrinogen (τ-tubes) from
one lot of Stock-F, and Sc of 12% between τ̄ for sets of τ-tubes from Stock-F prepared from different lots of
Armour Fraction-I, and an Sc of 1.9% for ã̄( =0.679) for sets of T-tubes, obtained either from one Stock-F or
Stock-F obtained from different lots of Fraction-I. Errors in several procedures for
determining unknown thrombin concentrations are examined.
