464 
HEMATOLOGY 
wards fragmentation of circulating platelets; 
these heparin-induced changes were particu- 
larly noticeable in kidney perfusions. Platelet 
counts in the citrated perfusion system adopted 
subsequently, remained relatively constant at 
40,000-100,000/cu mm throughout the perfu- 
sion period. As described above, an initial hour 
of perfusion, the "flush" period, was included 
and followed by a second perfusion period of 4 
hr with a fresh aliquot of perfusate.^ Perfusate 
was circulated through the organs at a rate of 
7-10 ml/min; greater flow rates were asso- 
ciated with platelet damage, hemolysis, and 
nonspecific activation of coagulation factors. 
With this technique, perfused organs main- 
tained their physiologic function; livers incor- 
porated radio-labeled leucine into albumin and 
kidneys sustained the uptake of radioactive vi- 
tamin Bi2 bound to transcobalamin II.* Pro- 
duction of coagulation factor VII, VIII, and IX 
activities in livers and spleens increased with 
time and was sensitive to puromycin, cyclohex- 
imide, and actinomycin D.^"^ Subsequent inves- 
tigation of these factor VIII and IX activities 
demonstrated properties similar to plasma co- 
agulation factors VIII and IX as well as the ca- 
pacity to neutralize human anti-factor VIII and 
anti-factor IX, respectively.^^ Perfused kid- 
neys also generated factor VII, VIII, and IX 
activities but, of these, only factor VII was sensi- 
tive to inhibitors of protein synthesis.^ In addi- 
tion, kidney factor VIII activity was not 
inactivated by human anti-factor VIIL^^ 
In the present studies, 5 ml of citrated 
plasma from normal individuals or from those 
with congenital or acquired defects of coagula- 
tion were added to hepatic and splenic perfus- 
ates at 30 min of the second perfusion period. 
The plasmas used were either fresh or fresh- 
frozen and of the following types : normal rab- 
bit or human; warfarin-treated rabbit; factor 
VIII and IX deficient with and without de- 
monstrable cross-reacting material (CRM) ; 
factor VIII deficient from patients with von 
Willebrand's disease (VWD). CRM(-) and 
( + ) hemophilia A plasmas were obtained from 
patients previously reported by one of us.'''^ 
CRM ( - ) and ( + ) hemophilia B plasmas were 
kindly supplied by Dr. H. R. Roberts of Chapel 
Hill, N. Von Willebrand's disease plasmas 
were generously supplied by Dr. E. J. W. Bowie ■ 
of Rochester, Minn." Normal rabbit and human K 
plasmas had between 80 and 160% of the coag- I 
ulation activity found in a pooled standard 1, 
plasma obtained from rabbits and humans, re- I 
spectively. Paired experiments were performed I 
with fresh normal rabbit plasma added to per- " 
fusates, and then with plasma from the same j 
rabbit after pretreatment with 2.5 mg/kg so- m 
dium warfarin (Endo Laboratories, Inc., Gar- I 
den City, N. Y.) twice daily for 3-5 days. The I 
prothrombin time of this plasma was > 2 min " 
and factor VII and IX activities were reduced 
to < 3 % of normal ; factor VIII activity varied 
from 48 to 200% with an average of 124%. , 
Normal human plasma used in perfusions was i 
obtained from 4 healthy volunteers who had not I 
taken any medication for at least 7 days. Each i 
plasma was divided into equal 4-5 ml aliquots ; 
one aliquot was used immediately and the oth- 
ers frozen and stored at —70° C for subsequent ' 
experiments. All deficient plasmas were fresh- 
frozen, stored at —70° C until used, and were i 
compared in assays to a pooled standard human j 
plasma. CRM ( — ) factor VIII deficient plasma i 
had < 1 % factor VIII activity, but normal lev- 
els of the other coagulation factors. CRM ( + ) 
factor VIII deficient plasma had 3-9% factor j 
VIII with normal factor IX activity. CRM ( - ) i 
and ( + ) factor IX deficient plasmas had < 1% I 
and 7-9 % factor IX activity, respectively, with j 
low normal (45-78% ) factor VIII activity. Von i 
Willebrand's disease plasma had 2-8% factor ! 
VIII activity with normal factor IX activity. In 
some experiments, factor VIII concentrates, 
prepared by cryoprecipitation of normal human 
plasma, were dissolved in the 5-ml aliquot of 
normal or deficient plasma added to splenic per- 
fusates. This was accomplished by centrifuga- 
tion of the cryoprecipitate, careful removal of 
all supernatant plasma, and dissolving the pre- 
cipitate in the appropriate sample. 
Perfusate samples (3 ml) were taken at the 
end of the initial hour of perfusion, termed 
"flush," and at zero, Vz, H, 1, IV2, 2, 3, and 4 hr 
of the final perfusion period. Perfusion volume 
was maintained after sampling by addition of 3 
ml of fresh perfusate. All samples were pre- 
pared and assayed for factor VIII and IX 
activities with canine deficient plasmas in a one- 
