Cerebrospinal fluid shunt placement for pseudotumor cerebri-associated intractable headache: predictors of treatment response and an analysis of long-term outcomes

Object. Cerebrospinal fluid (CSF) shunts effectively reverse symptoms of pseudotumor cerebri postoperatively, but long-term outcome has not been investigated. Lumboperitoneal (LP) shunts are the mainstay of CSF shunts for pseudotumor cerebri; however, image-guided stereotaxy and neuroendoscopy now allow effective placement of a ventricular catheter without causing ventriculomegaly in these cases. To date it remains unknown if CSF shunts provide long-term relief from pseudotumor cerebri and whether a ventricular shunt is better than an LIP shunt. The authors investigated these possibilities. Methods. The authors reviewed the records of all shunt placement procedures that were performed for intractable headache due to pseudotumor cerebri at one institution between 1973 and 2003. Using proportional hazards regression analysis, predictors of treatment failure (continued headache despite a properly functioning shunt) were assessed, and shunt revision and complication rates were compared between LP and ventricular (ventriculoperitoneal [VP] or ventriculoatrial [VAT]) shunts. Forty-two patients underwent 115 shunt placement procedures: 79 in which an LP shunt was used and 36 in which a VP or VAT shunt was used. Forty patients (95%) experienced a significant improvement in their headaches immediately after the shunt was inserted. Severe headache recurred despite a properly functioning shunt in eight (19%) and 20 (48%) patients by 12 and 36 months, respectively, after the initial shunt placement surgery. Seventeen patients without papilledema and 19 patients in whom preoperative symptoms had occurred for longer than 2 years experienced recurrent headache, making patients with papilledema or long-term symptoms fivefold (relative risk [RR] 5.2, 95% confidence interval [CI] 1.5-17.8; p < 0.01) or 2.5-fold (RR 2.51, 95% CI 1.01-9.39; p = 0.05) more likely to experience headache recurrence, respectively. In contrast to VP or VAT shunts, LP shunts were associated with a 2.5-fold increased risk of shunt revision (RR 2.5, 95% Cl 1.5-4.3; p < 0.001) due to a threefold increased risk of shunt obstruction (RR 3, 95% CI 1.5-5.7; p < 0.005), but there were similar risks between the two types of shunts for overdrainage (RR 2.3, 95% CI 0.8-7.9; p = 0.22), distal catheter migration (RR 2.1, 95% CI 0.3-19.3; p = 0.55), and shunt infection (RR 1.3, 95% CI 0.3-13.2; p = 0.75). Conclusions. Based on their 30-year experience in the treatment of these patients, the authors found that CSF shunts were extremely effective in the acute treatment of pseudotumor cerebri-associated intractable headache, providing long-term relief in the majority of patients. Lack of papilledema and long-standing symptoms were risk factors for treatment failure. The use of ventricular shunts for pseudotumor cerebri was associated with a lower risk of shunt obstruction and revision than the use of LP shunts. Using ventricular shunts in patients with papilledema or symptoms lasting less than 2 years should be considered for those with pseudotumor cerebri-associated intractable headache.