Hypersplenism is of great relevance in the management of cirrhosis bec
ause of the widespread use of myelodepressant drugs such as interferon
or antineoplastic agents. Because no standard therapy exists for this
complication, we have evaluated the efficacy and risks of splenic emb
olization in the treatment of hypersplenism in cirrhosis. Partial sple
nic embolization was performed in 40 consecutive patients with the fol
lowing indications: 25 patients with active viral cirrhosis before int
erferon therapy, 8 patients with unresectable hepatocellular carcinoma
before antineoplastic chemotherapy and 7 patients with thrombocytopen
ia associated with spontaneous bleeding events, with high risk of cent
ral nervous system hemorrhage or facing major surgery. After selective
catheterization of the splenic artery, partial splenic embolization w
as performed by means of repeated injections of gelatin sponge until a
40% to 60% reduction in the splenic blood flow was achieved. After pa
rtial splenic embolization a significant and sustained increase in pla
telet and white blood cell count was observed during the follow-up per
iod (mean = 13.9 +/- 2.2 mo; range = 1 to 36 mo). The goal of partial
splenic embolization was achieved in all but two patients in whom a do
se reduction of interferon was needed. Hypersplenism relapsed in only
seven patients, and all of them exhibited an embolization of less than
50% of the splenic mass. Postembolization syndrome was the main side
effect, but no life-threatening complications were detected. In conclu
sion, partial splenic embolization is a safe and effective therapy of
hypersplenism in cirrhosis.