On reviewing the literature about intratumoral hemorrhage in children [4,5], we observed an unexpectedly high inci dence of an ictal presentation. Hemorrhage is reported to occur in highly vascularized tumors, such as metastases. med- ulloblastomas. and neuroblastomas. Although ependymoma is a highly vascularized tumor. only a few cases of hemorrhage in ependymoma have been reported. Our patient had a tumor that does not often bleed. The sudden and dramatic clinical deterioration after ventricular decompression suggests that VP shunt insertion was the cause of intratumoral bleeding. In spite of avoiding cerebrospinal fluid leak during the operation and inserting a medium pressure valvular shunt, operation upset the delicate balance of pressures between the supra- and infratentorial compartments. Only three other cases of hem- orrhage after VP shunting and two cases of hemorrhage after the insertion of an external drainage system have been reported [12,13].

This case stirs up again the controversy about the use of precraniotomy shunts before tumor operation and the choice between internal and external shunts. Almost all published data indicate that precraniotomy shunts benefit children who are ill from hydrocephalus due to posterior fossa tumor by improving their general condition before operation and by reducing operative mortality and morbidity. This agreement lessens the significance of the controversy and argumentation for and against precraniotomy shunts. The major arguments against precraniotomy internal shunts are: (a) the risk of upward herniation after shunt insertion [6,7], (b) the possibility of transshunt metastasis [8]. (c) the belief that cortico- steroids restore patency to an obstructed cerebrospinal fluid pathway [9] and (d) the fact that most children do not need a permanent shunt after tumor removal.

Advances in technology have made external drainage com- parable to internal shunting in risk and benefit for children who need hydrocephalus treatment before craniotomy, so that in the U.S. the use ofa precraniotomy shunt seems to be anachronistic.

The risk of infection. generally accepted to be between 4% and I 0% [14], is an important argument for those, and we are among them, who use permanent shunts, although it has been reported that the risk of infection with current external drain- age systems is equivalent to that with internal shunts [10].

the occurrence of hemorrhage in ependymoma after VP shunting, although infrequent [4,11], is a complication that must be borne in mind when deciding between an internal and an external shunt.