Congenital protein S (PS) deficiency is associated with increased risk of v
enous thrombosis. To investigate the possibility of automating PS testing w
ith decreased turnaround time, a clotting-based functional protein S assay
was evaluated and compared to an antigenic method. Samples were collected f
rom 126 patients within 5 days of their first acute cerebral infarction, fr
om 62 controls and from 47 consecutive samples for thrombophilia investigat
ion. The normal range for the clotting-based kit, calculated from the resul
ts of 20 healthy controls, was 62-136% (mean +/- 2 SD). Intra- and inter-as
say co-efficients of variation were <3.0 and 10.0%, respectively. There was
no significant correlation between the two methods (r = 0.30, P > 0.05). T
wo patients had low PS antigen results with normal functional levels. Both
techniques were used to compare a further group of 53 patients with defined
abnormalities which included nine antigenic protein S deficiencies, five p
rotein C deficient patients, 10 patients with a lupus anticoagulant (LA), 1
7 Factor V Leiden (FVL) heterozygotes, two FVL homozygotes and 10 patients
on therapeutic levels of heparin. In this group we found that four of nine
antigenic PS deficient patients had normal functional PS levels. The test w
as susceptible to the FVL mutation with four of 17 FVL heterozygotes and bo
th of two FVL homozygotes giving low levels. One of five protein C-deficien
t patients also had a low functional PS result with a normal antigenic leve
l. Normal results were obtained by both methods for all of the LA and patie
nts on therapeutic heparin, We concluded that the automated protein S clott
ing assay was rapid and simple to perform but appeared to be influenced by
factors other than PS deficiency. Results need to be interpreted with cauti
on but may be useful as part of a full thrombophilia investigation.