We hypothesize that infusion of chemotherapeutic agents directly into the fourth ventricle potentially may play a role in treating malignant posterior fossa brain tumors. Accordingly, we used a piglet model developed in our laboratory to test the safety of etoposide infusions into the fourth ventricle and to study the pharmacokinetics associated with these infusions. In 5 piglets, closed-tip silicone catheters were inserted into the fourth ventricle and lumbar cistern. Five consecutive daily infusions of etoposide (0.5 mg) were administered via the fourth ventricle catheter. Serum and CSF from both catheters were sampled for measurement of etoposide level by reversed-phase high performance liquid chromatography (HPLC). For CSF samples, area under the concentration-time curve (AUC) was calculated. Piglets underwent daily neurological examinations, a 4.7 Tesla MRI scan, and then were sacrificed for post-mortem brain examination. No neurological deficits or signs of meningitis were caused by intraventricular chemotherapy infusions. MRI scans showed catheter placement within the fourth ventricle but no signal changes in the brain stem or cerebellum. In all piglets, the mean fourth ventricular CSF peak etoposide level exceeded the mean peak lumbar etoposide levels by greater than 10-fold. Statistically significant differences between fourth ventricle and lumbar AUC were noted at peaks (DeltaAUC = 3384196 ng h/ml with 95%CI: 1758625, 5009767, P = 0.0044) and at all collection time points (DeltaAUC = 1422977 ng h/ml with 95%CI: 732188, 2113766, P = 0.0046) but not at troughs (DeltaAUC = -29546 ng h/ml (95%CI: -147526, 88434.2, P = 0.5251). Serum etoposide was absent at two and four hours after intraventricular infusions in all animals. Pathological analysis demonstrated meningitis, choroid plexitis, and ependymitis in the fourth and occasionally lateral ventricles. Etoposide can be infused directly into the fourth ventricle without clinical or radiographic evidence of damage. Autopsy examination revealed ventriculitis and meningitis which did not have a clinical correlate. Etoposide does not distribute evenly throughout CSF spaces after administration into the fourth ventricle, and higher peak CSF levels are observed in the fourth ventricle than in the lumbar cistern.