To investigate the possibility of correcting thrombocytopenia of chron
ic liver disease, 19 patients (6 male and 13 female) with long-term ch
ronic liver disease and platelet count less than or equal to 85 000/mu
l were studied. Either a short-term course (7-20 days) of recombinant
human erythropoietin, 4000 U daily SQ (12 patients) or placebo (7 pat
ients) was administered. Treatment was interrupted if the platelets ro
se to greater than or equal to 100 000/mu l or if no significant incre
ase was noted after 14 days. After treatment, platelets increased in t
he recombinant human erythropoietin group (from a baseline value of 70
000+/-11 184 to 101 250+/-37 625/mu l), while no difference was noted
in the placebo group (70 714+/-9928 vs 70 000+/-10 231/mu l). The inc
rease in the platelet count in the recombinant human erythropoietin gr
oup was significant, both compared to baseline values (paired Student'
s t-test, t=-3.80, p<0.005) and to the results of treatment in the pla
cebo group (unpaired Student's t-test, t= 2.71, p<0.02). Eight out of
12 recombinant human erythropoietin-treated patients (66%) reached gre
ater than or equal to 100 000/mu l platelets while four (33%) did not.
In comparison to responders, non-responders had a significantly lower
baseline platelet count (58 500+/-7937 vs 75 750+/-7498/mu l, t=-3.69
, p=0.004) and failed more frequently than responders to improve their
haematocrit in response to recombinant human erythropoietin (Pearson
chi(2)=4.687, p=0.03). When treatment was discontinued, the platelet c
ount reverted to baseline in a few weeks. In conclusion, recombinant h
uman erythropoietin treatment transiently corrected mild thrombocytope
nia in patients with chronic liver disease. The failure to increase ci
rculating thrombocytes with recombinant human erythropoietin treatment
occurred in patients with a lower steady-state value in the balance b
etween excessive platelet destruction and compensatory production. (C)
Journal of Hepatology.