Complications of Ventriculoperitoneal Shunt for Idiopathic Intracranial Hypertension: A Single-Institution Study of 32 Patients

Background: Because there are no head-to-head studies of the efficacy of surgical options in the treatment of medically-intractable idiopathic intracranial hypertension (IIH), procedure selection is often based on expected complications. Cerebrospinal fluid (CSF) diversion by shunting has been reported to have a 23%-67% rate of shunt failure. But these figures derive from small cohorts or studies that do not separate the complication rates of the different shunt options-ventriculoperitoneal (VP), lumboperitoneal (LP), and ventriculoatrial (VA). The complication rate of LP shunts seems to be higher than that for VP shunts, the procedure currently in widest use. Our experience with VP shunts for IIH over the past decades suggests that the complication rate for that option is lower than reported series would suggest. Methods: Retrospective single-institutional study using a search engine that finds all text mentions of particular terms. We searched for the terms ('IIH' OR 'Idiopathic intracranial hypertension' OR 'PTC OR 'pseudotumor') AND ('VP Shunt' OR 'ventriculoperitoneal') over the period 1998 to 2018. From 490 'hits,' only 32 patients met entry criteria: diagnosis of IIH confirmed at our institution, including examination in a neuro-ophthalmology clinic showing papilledema, elevated opening pressure on lumbar puncture or a consistently elevated intraparenchymal pressure on Codman intracranial pressure (ICP) monitoring, neuroimaging and CSF studies consistent with a diagnosis of IIH, and at least 1 year of follow-up in our neuro-ophthalmology or neurosurgery clinics. Results: Shunt failures occurred in 6 (18.7%) of 32 patients, which included elevated ICP due to obstruction or discontinuity of the shunt, over-drainage, infection of the shunt system, or intractable abdominal pain or infection. None of these failures caused permanent worsening of vision, neurologic morbidity, or death. But they entailed considerable medical care. Placement of VP shunts also aroused patient fear of complications, precipitating 38 emergency visits in 14 patients for 'false alarms.' Conclusions: In this series of 32 patients, the largest retrospective study of VP shunts for IIH, there was an 18.7% failure rate, lower than previously published smaller series have disclosed. But among patients who suffered complications, a saga of intensive medical care often occurred. Placement of the shunt also aroused patient fear of complications, leading to many medical encounters required to rule out complications.

Original Contribution Section Editors: Clare Fraser, MD Susan Mollan, MD Complications of Ventriculoperitoneal Shunt for Idiopathic Intracranial Hypertension: A Single-Institution Study of 32 Patients Anthony J. Brune, DO, Taania Girgla, BS, Jonathan D. Trobe, MD Background: Because there are no head-to-head studies of the efficacy of surgical options in the treatment of medicallyintractable idiopathic intracranial hypertension (IIH), procedure selection is often based on expected complications. Cerebrospinal fluid (CSF) diversion by shunting has been reported to have a 23%–67% rate of shunt failure. But thesefigures derive from small cohorts or studies that do not separate the complication rates of the different shunt options—ventriculoperitoneal (VP), lumboperitoneal (LP), and ventriculoatrial (VA). The complication rate of LP shunts seems to be higher than that for VP shunts, the procedure currently in widest use. Our experience with VP shunts for IIH over the past decades suggests that the complication rate for that option is lower than reported series would suggest. Methods: Retrospective single-institutional study using a search engine that finds all text mentions of particular terms. We searched for the terms (“IIH” OR “Idiopathic intracranial hypertension” OR “PTC OR “pseudotumor”) AND (“VP Shunt” OR “ventriculoperitoneal”) over the period 1998 to 2018. From 490 “hits,” only 32 patients met entry criteria: diagnosis of IIH confirmed at our institution, including examination in a neuro-ophthalmology clinic showing papilledema, elevated opening pressure on lumbar puncture or a consistently elevated intraparenchymal pressure on Codman intracranial pressure (ICP) monitoring, neuroimaging and CSF studies consistent with a diagnosis of IIH, and at least 1 year of follow-up in our neuroophthalmology or neurosurgery clinics. Results: Shunt failures occurred in 6 (18.7%) of 32 patients, which included elevated ICP due to obstruction or discontinuity of the shunt, over-drainage, infection of the shunt system, or intractable abdominal pain or infection. None of these failures caused permanent worsening of vision, neurologic morbidity, or death. But they entailed considerable medical care. Placement of VP shunts also aroused patient fear of complications, precipitating 38 emergency visits in 14 patients for “false alarms.” Department of Ophthalmology and Visual Sciences (AJB, JDT), Neurology (JDT), Kellogg Eye Center, Ann Arbor, Michigan; University of Michigan Medical School (TG), Ann Arbor, Michigan; and Neurosurgery (JDT), Kellogg Eye Center, Ann Arbor, Michigan. The authors report no conflicts of interest. Address correspondence to Jonathan D. Trobe, MD, Kellogg Eye Center, 1000 Wall Street, Ann Arbor, MI 48105; [e-mail] 224 Conclusions: In this series of 32 patients, the largest retrospective study of VP shunts for IIH, there was an 18.7% failure rate, lower than previously published smaller series have disclosed. But among patients who suffered complications, a saga of intensive medical care often occurred. Placement of the shunt also aroused patient fear of complications, leading to many medical encounters required to rule out complications. Journal of Neuro-Ophthalmology 2021;41:224–232 doi: 10.1097/WNO.0000000000000922 © 2020 by North American Neuro-Ophthalmology Society I diopathic intracranial hypertension (IIH) is a common vision-threatening condition in which high intracranial pressure (ICP) may cause damage to the optic nerves. Medication and weight loss are often successful in lowering ICP and preventing optic neuropathy (1), but when they are not, surgical intervention may be indicated, consisting of 3 options: cerebrospinal fluid (CSF) diversion, optic nerve sheath fenestration, and venous sinus stenting (2). No prospective comparative trials have been done to establish superior efficacy of any of these surgical options. Accordingly, choices are often based on the availability of experienced surgeons or interventional radiologists and the perceived safety of each procedure. The choice of optic nerve sheath fenestration or venous sinus stenting is partly based on the impression that ventriculo-peritoneal (VP) shunting carries a high failure rate. That impression derives mostly from small series reports and from other confounders (Table 1). Among the small series reports, Bynke et al (3) found that 7 (41%) of 17 patients required VP shunt revision after a mean of 6 months. Over a mean follow-up period of 21 months, Kandasamy et al (4) found that 5 (29%) of 17 patients undergoing VP shunts required revision. Maher et al (5) reported that 3 (23%) of 13 patients with VP shunts underwent revision for shunt obstruction. Abu-Serieh et al (6) found that 6 (67%) of 9 patients undergoing VP shunts required revision within an average of 44 months. Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution TABLE 1. Previously published studies of complications of cerebrospinal fluid diversion for idiopathic intracranial hypertension Shunt Type # of VP Shunt Patients VP 17 VP 17 VP 13 Abu-Serieh et al (6) Huang et al (7) VP 9 VP 19 6.5 years (1.8–12.8 years) 21 months (9–49 months) 15.2 months (1–38 months) 44.3 months (6–110 months) 21.2 months Woodworth et al (9) Tarnaris et al (10) VP, VA, VPl VP, LP 20 Not reported 9 Abubaker et al (11) VP, LP 10 28.9 months, including both shunt types 4 years (6 months–7 years), including both shunt types McGirt et al (12) Multicenter, retrospective Rosenberg et al (13) Administrative database Menger et al (14) All Not reported VP, LP, VA, VJ 8 LP, VP LP, VP Author(s) Single center, retrospective Bynke et al (3) Kandasamy et al (4) Maher et al (5) Azad et al (15) Mean Follow-up (Range) VP Shunt Failure Rate Limitations 41% Small cohort 29% Small cohort 23% Small cohort 67% Small cohort 31% Reported 48% survival at 36 months because of Kaplan–Meier analysis VP shunt complications not separately reported Small cohort, limited VP shunt-specific data reported Small cohort Not reported 11% 20% VP shunt complications not separately reported 30.9 months (1 month–15 years) Not reported VP shunt complications not separately reported 2,505 3.9% 735 31.7% Duration of follow-up and indications for shunting and re-operation unavailable Minimum follow-up of only 30 days, indications for shunting and reoperation unavailable LP, lumboperitoneal; VA, ventriculoatrial; VPl, ventriculopleural; VJ, ventriculojugular; VP, ventriculoperitoneal. In other reports, overestimation of shunt failure may be based on shunt survival data. Thus, among 19 patients who underwent VP shunting in the report of Huang et al (7), the proportion of surviving VP shunts had declined to 48% at 36 months of follow-up. However, a long survival time of CSF diversion procedures may not be important in most cases of IIH, because the disease process is commonly selflimited. The impression that CSF diversion is an unfavorable option may also derive from the fact that lumboperitoneal (LP) shunting carries a relatively high failure rate. A review Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 of 14 reports (8) of LP shunting for IIH disclosed that 128 patients had a revision rate of 38%. Many reports do not separate out the failure rates of VP, LP, and ventriculoatrial (VA) shunts. For example, in the study of Woodworth et al (9) of 21 patients who had undergone 32 ventricular shunts, only 20 were VP shunts. There was an aggregate 50% failure rate at 12 months, but the fate of the VP shunts was not described. Tarnaris et al (10) reported a 12-year experience in 34 patients, 9 of whom underwent VP shunting, 4 of which were lost to follow-up. There was only one case (11%) of revision due to ventricular catheter 225 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution obstruction in the VP shunt group. In a study that included 18 patients treated with LP shunts and 10 patients treated with VP shunts, Abubaker et al (11) found that LP shunts had a revision rate of 60%, whereas the VP shunts had a revision rate of only 20%. McGirt et al (12) reported 36 patients who had undergone VP or VA shunts. In that cohort, 16 (44%) shunts failed at a mean follow-up of 13 months for nonfailing shunts, but the authors did not separately report the failure rates of the VP and VA shunts. Rosenberg et al (13) reported a six-center series of 83 shunt procedures for IIH, in which 56 (67%) required revision or replacement, but only 8 were VP shunts. The VP shunt failure rate was not separately reported. Large data sets have also been the source of information on reoperation rates, but criteria for patient selection and indications for reoperation were not described (14,15). Menger et al (14) compared LP and VP shunting in IIH from a nationwide database including 2,505 patients. The VP shunt revision rate was only 3.9% compared with 7.0% for LP shunts. Reasons for revision and follow-up periods in patients who did not undergo shunt revision were not documented. Azad et al (15) used a national administrative database to elicit the need for repeat surgery in patients undergoing VP or LP shunts for IIH. Of 735 patients undergoing VP shunts over a mean follow-up of 680 days, 233 (31.7%) required at least one more procedure. Indications for shunt revisions were not reported. The surgical alternatives to CSF diversion in refractory IIH have their own complications. In a review of 341 optic nerve sheath fenestrations (8), 15.9% of eyes sustained a decline in visual field or visual acuity postoperatively. In addition, 69.8% of cases required a repeat procedure. Apart from vision loss, complications were reported in 72 (26%) of 278 patients, including 37 cases of diplopia, 2 orbital hematomas, 1 orbital apex syndrome, and 1 traumatic optic neuropathy. Our clinical impression from more than 30 years’ experience in the care of patients with IIH has been that the complication rate of VP shunting is much lower than many previous publications have suggested. We now report those complications based on a retrospective study of 32 patients who underwent VP shunts for IIH at our institution between 1996 and 2017. METHODS We conducted a single-center retrospective analysis to determine the incidence of VP shunt complications in patients with IIH who had undergone VP shunting and were examined between 1998 and 2018 at the University of Michigan. We did not include any patients who had undergone LP or VA shunts, because those procedures are rarely performed at our institution. We used EMERSE (Electronic Medical Record Search Engine) to search patient records for the terms (“IIH” OR “Idiopathic intra226 cranial hypertension” OR PTC OR pseudotumor) AND (“VP Shunt” OR ventriculoperitoneal) (16). That search identified 490 potential patients. From that cohort, we included only 32 patients whose diagnosis of IIH had been confirmed at our institution, including examination in a neuro-ophthalmology clinic showing papilledema, elevated opening pressure on lumbar puncture or a consistently elevated intraparenchymal pressure on Codman ICP monitoring, neuroimaging and CSF studies consistent with a diagnosis of IIH (17), and at least 1 year of follow-up in our neuro-ophthalmology or neurosurgery clinics. Shunt failures were defined as those requiring reprogramming, revision, or removal of the shunt owing to obstruction or discontinuity of the shunt, over-drainage, infection of the shunt system, or intractable abdominal pain or infection. We also collected data on emergency room encounters prompted by patient concern for shunt complications even if they were ruled out with appropriate studies. Although these “false alarm” encounters were not considered complications, they were deemed worthy of inclusion because they added medical costs derived from emergency room visits eliciting head computed tomography (CT), shunt series x-rays, neurosurgical consultations, ophthalmologic consultation to rule out papilledema, lumbar punctures, or ICP monitoring. RESULTS Among the 32 patients in our cohort, the mean follow-up was 65 months (median 44 months, range 14–246 months). There were 8 patients who had undergone at least one optic nerve sheath fenestration procedure before shunt placement. Among the 32 patients, 25 had undergone VP shunting months after the initial clinical encounter and 7 had undergone VP shunting for severe optic neuropathy within days of the initial encounter. The initial procedures were performed by 13 different neurosurgeons. VP shunt failures occurred in 6 (18.7%) patients, who altogether had 12 shunt failures distributed without clustering for any particular surgeon (Table 2). Two patients experienced 3 events of obstruction causing increased ICP (Patient 1 and Patient 3 twice), neither of whom suffered a persistent decline in vision as the result of the complications. One patient had over-drainage resulting in a low CSF pressure headache (Patient 5). Four patients (Patients 1, 2, 4, and 6) had 5 shunt infections requiring shunt removal or replacement, including one patient with an abdominal pseudo-cyst (Patient 1). One of these infections was associated with an erosion of the peritoneal catheter into the colon (Patient 6). Three patients had abdominal wounds potentially communicating with the shunt. Two wounds occurred after shunt placement (Patient 2 twice), one of which resulted in shunt infection, and 2 occurred after unrelated abdominal surgery (Patients 1 and 2), one of which led to shunt infection (Patient 1). Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution TABLE 2. Clinical data on our cohort of 32 patients Sex Age at Diagnosis (yrs) Interval From Diagnosis of IIH to Initial VP Shunt 1 F 29 37 mo 2 F 22 148 mo 3 F 16 2d 4 F 18 6 mo 5 F 24 22 mo 6 F 28 3 mo 7 F 24 4 mo 8 F 37 36 mo 9 F 27 15 d 10 F 16 8 mo 11 F 27 3 mo 12 F 42 35 mo 13 F 25 3d 14 F 24 3d 15 F 40 26 mo 16 F 42 18 mo 17 F 23 5 mo 18 F 24 339 mo 19 F 19 2d 20 F 29 21 mo 21 F 23 3d 22 F 28 12 mo 23 M 39 2 mo 24 F 17 6d 25 F 29 46 mo Patient # Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 Interval Between VP Shunt Placement and Last Follow-up Visit Episodes of Shunt Failure Refractory to medical management; failed ONSF Intolerant of medical management; ONSF and LPS failure 246 mo 3 238 mo 4 Severe vision loss at presentation Intolerant of medical management; failed ONSFs Intolerant of medical management Intolerant of medical management Refractory to medical management Intolerant of medical management Intolerant of medical management Intolerant of medical management; failed ONSF Intolerant of medical management; failed ONSF Refractory to medical management Severe vision loss at presentation Severe vision loss at presentation Intolerant of medical management Intolerant of medical management; ONSF failure Refractory to medical management Intolerant of medical management; LPS failure Severe vision loss at presentation Refractory to medical management; difficult to monitor (congenital nystagmus) Severe vision loss at presentation Refractory to medical management Refractory to medical management Severe vision loss at presentation Refractory to medical management 147 mo 2 41 mo 1 45 mo 1 23 mo 1 57 mo 0 3 87 mo 0 0 59 months 0 8 43 mo 0 1 24 mo 0 0 15 months 0 2 108 mo 0 2 36 mo 0 1 80 mo 0 1 32 months 0 33 mo 0 1 82 mo 0 0 61 mo 0 2 50 mo 0 4 25 mo 0 0 60 mo 0 0 13 mo 0 3 30 mo 0 0 35 mo 0 0 Indications for VP Shunt Causes of Shunt Failure Number of False Alarms Nonhealing abdominal wound, infection, proximal obstruction Cerebrospinal fluid cutaneous fistula and infection, poorly healing abdominal wound and infection, chest wall adhesions Proximal obstruction and infection Skin dehiscence and infection near reservoir Low pressure headache 0 Distal obstruction and infection 0 0 8 1 0 227 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution (Continued ) Sex Age at Diagnosis (yrs) Interval From Diagnosis of IIH to Initial VP Shunt 26 F 17 66 mo 27 M 55 7 mo 28 F 22 2 mo 29 F 31 10 mo 30 F 24 92 mo 31 F 53 5d 32 M 39 4 mo Patient # Indications for VP Shunt Refractory to medical management Refractory to medical management; failed ONSF Refractory to medical management Refractory to medical management; failed ONSF Intolerant of medical management Severe vision loss at presentation Refractory to medical management Interval Between VP Shunt Placement and Last Follow-up Visit Episodes of Shunt Failure 108 mo 0 1 16 mo 0 0 18 mo 0 0 24 mo 0 0 177 mo 0 0 59 mo 0 0 14 mo 0 0 Causes of Shunt Failure Number of False Alarms IIH, idiopathic intracranial hypertension; LPS, lumboperitoneal shunt; ONSF, optic nerve sheath fenestration; VP, ventriculoperitoneal. One patient had an incidental CSF cutaneous fistula (Patient 2). One patient required revision of the shunt tubing due to abdominal adhesions (Patient 2). Two patients experienced operative complications during the shunt procedure. One of these 2 patients had an accidental enterotomy requiring intravenous antibiotic treatment (Patient 1), delaying shunt placement, and prolonging hospitalization. Two patients (Patients 1 and 9) had intracerebral hemorrhages related to the ventricular catheter. In one case, the hemorrhage produced no clinical manifestations (Patient 9), but in the other case (Patient 1), it caused cerebral edema, temporary hemiparesis, and seizures requiring continuing antiepileptic medication. Among the 32 patients who had undergone VP shunting, 14 (44%) came to the emergency room 38 times with manifestations that prompted consideration of shunt malfunction, but appropriate evaluations were negative. Because of significant patient-to-patient variability in the pre-operative and postoperative clinical course, we have provided a narrative summary of each of the 6 patients who suffered shunt complications. Patient #1 At age 29, this woman received a diagnosis of IIH. Visual acuities were 20/15 OU, but Humphrey Visual Field mean deviations (MDs) were 220 dB right eye and 216 dB left eye. Opening pressure on lumbar puncture was 62 cmH2O. She was treated with acetazolamide and later optic nerve sheath fenestration right eye. Optic disc edema lessened, so acetazolamide was discontinued. She was lost to followup for more than 2 years, returning with abrupt loss of vision to hand movements (HM) right eye and 20/70 left eye with a MD of 232.00 dB left eye. The vision loss was attributed to initiation of antihypertensive medication. 228 Optic disc edema had recurred OU, so acetazolamide was restarted. Placement of a VP shunt was attempted 6 days later, but aborted due to accidental enterotomy during the abdominal portion of the procedure. Six weeks later, a VP shunt was placed successfully, but persisting abdominal pain led to a laparotomy. The abdominal incision of the laparotomy broke down, and CSF began draining externally for 2 weeks. Given the risk of infection, the shunt was removed and later replaced. One week after that surgery, she presented with clear yellow fluid leaking from an abdominal incision. Oral antibiotic treatment led to wound healing, but continuing abdominal pain prompted a CT that revealed a cyst associated with the shunt catheter. A shunt tap revealed gram-positive rods, so the VP shunt was removed. A third VP shunt later led again to abdominal pain, which eventually subsided. She was followed for 15 years with stable visual function. She eventually underwent gastric bypass surgery complicated by abdominal infection. Visual acuity declined to NLP right eye and 20/500 left eye. ICP monitoring showed pressures ranging between 20 cmH2O and 50 cmH2O. There was no flow through the ventricular catheter, so it and the valve were replaced. On routine head CT performed post-operatively, she had a right frontal intraparenchymal hematoma. CT venography was negative for dural venous sinus thrombosis. ICP monitoring demonstrated pressures above 20 cmH2O despite shunt revision. She was started on hypertonic saline for cerebral edema associated with the intraparenchymal hemorrhage that caused new left face and upper extremity weakness. Visual acuity left eye declined to light perception. Head CT was unchanged and a nuclear medicine shuntogram showed adequate shunt flow. Five months later, she presented after a generalized tonic-clonic seizure. Head CT showed gliotic changes in the right frontal lobe. She was Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution admitted to the Neurointensive Care Unit for 3 days and started on levetiracetam. Sixteen months after the last shunt revision, visual acuity had improved to HM right eye and 20/40 left eye with a MD of 232 dB. There were no lingering neurologic deficits. Patient #2 At age 22, this woman was diagnosed with IIH. She was intolerant of acetazolamide, which was eventually discontinued after 1 year. Manifestations of IIH returned, so acetazolamide was restarted, ultimately leading to pancreatic inflammation. Despite furosemide treatment, visual field loss progressed. Eight years after the original diagnosis, she underwent bilateral optic nerve sheath fenestrations, after which visual function continued to decline. Ten years after diagnosis, she suffered trauma to the left eye, resulting in fungal endophthalmitis that led to enucleation left eye. A lumboperitoneal shunt was placed for progressive visual field loss in the right eye, headache, and a persistently elevated ICP on lumbar punctures. LP shunt placement caused back and left leg pain, perineal numbness, and a neurogenic bladder, so it was replaced with a VP shunt. Two years after VP shunt placement, minor trauma resulted in a laceration over the shunt reservoir. She returned the next day with increased headache, which was evaluated by head CT, shunt series x-rays, and a shunt tap, all of which were negative. Three weeks later she returned to the Neurosurgery Clinic with persistent postural headache. A shunt tap with dynamic pressure testing was normal. Abdominal ultrasound to evaluate for pseudocyst was negative. Because of persistent headache, head CT, shunt series x-rays, abdominal ultrasound, and lumbar puncture were performed days later, again negative. Three years after VP shunt placement, she presented again with headache. Head CT and shunt series x-ray were unremarkable. Four years after VP shunt placement, she came with abdominal pain. Laparoscopy showed numerous dense abdominal adhesions. A pinpoint CSF cutaneous fistula behind her right ear later led to removal of the VP shunt. CSF cultures grew Proprionibacterium acnes requiring intravenous antibiotic treatment. A lumbar puncture showed evidence of meningitis, triggering more intensive antibiotic treatment. One month later, the visual field right eye had declined. Blood cultures grew gram positive rods. Over the next several months, she was repeatedly admitted for septicemia of unclear origin, with one lumbar puncture showing an opening pressure of 39 cmH2O. Visual field loss progressed. One month after shunt replacement, she was noted to have a poorly healing and infected abdominal incision from placement of the peritoneal portion of the catheter. The shunt was externalized, and the abdominal wound was debrided. She was treated with intravenous antibiotics for 5 days and the shunt was reinserted. Two weeks later, she again had an infection of the abdominal surgical wound. Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 The shunt was externalized, and the abdominal wound was debrided. The shunt was subsequently removed; purulent material was noted along the shunt tract. CSF cultures grew Pseudomonas aeruginosa and later Candida and Achromobacter xylosoxidans secondary to PICC infection. Seven months after removal of the second VP shunt, a third VP shunt was placed, but 7 months later she underwent revision of the shunt tubing as adhesions had developed in the upper chest wall. Twenty-one months later she underwent appendectomy, after which she developed a wound infection requiring antibiotic treatment for Enterobacter. Over the next year, she was admitted 6 times for a nonhealing abdominal wound and bacteremia. Those visits involved 3 neurosurgical consultations, an additional shunt series, 6 head CT scans, and a lumbar puncture. The VP shunt was replaced with a ventriculopleural shunt. Visual function remained stable for the next 8 years. Patient #3 At age 16, this woman was diagnosed with IIH. Visual acuity was NLP right eye and 20/40 with an inferior altitudinal defect left eye. Imaging was negative and LP showed an opening pressure of 82 cmH2O. She was treated with acetazolamide and a VP shunt 2 days after the initial encounter. Five months later, she presented with headache. Diagnostic studies were negative, but the shunt was set to a lower resistance. As headache continued, shunt exploration was undertaken, showing no spontaneous flow between the ventricular catheter and reservoir. The ventricular catheter was replaced. Intraoperative CSF cultures grew Propionibacterium acnes. Three weeks after the shunt revision, she developed somnolence and increasing headache. Lumbar puncture opening pressure was 34 cmH2O with normal CSF constituents. Acetazolamide was restarted, but headache continued. CSF studies showed a pleiocytosis, but CSF cultures were negative. Twenty-five months after the first shunt revision, she developed acute vision loss in the left eye considered psychogenic. Forty-two months after the first shunt revision, she developed fever and CSF studies showed a high white cell count and high protein content, but ICP monitoring was normal. Ten years after the first shunt revision, she presented with headache, nausea, and vomiting. The shunt valve, found to be obstructing flow, was replaced. Visual function returned to baseline. Three years after the second shunt revision, visual function was stable and there were no further complications. Patient #4 At age 18, this woman presented with neck pain, diplopia, and blurred vision. Visual acuities were 20/60 right eye and 20/100 left eye and she had severe papilledema. Brain MRI was negative and lumbar puncture showed an opening pressure of greater than 60 cmH2O. She was treated with intravenous methylprednisolone and then prednisone and 229 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution acetazolamide. Six days later, a right optic nerve sheath fenestration was performed. Visual function improved slightly, but the patient required escalating doses of acetazolamide, so VP shunting was performed 4 months later. Three months after VP shunt placement, she was admitted for progressive headache, nausea, and vomiting. Diagnostic studies were negative. She was treated for a urinary tract infection with resolution of symptoms. Two years after VP shunt placement, minor head trauma resulted in dehiscence of the skin above the shunt reservoir. A shunt tap showed staphylococcal and Bacteroides organisms within the CSF. She was treated with intravenous antibiotics and the VP shunt was removed. She was weaned off acetazolamide and followed for an additional 16 months without recurrence of IIH or further complications. Patient #5 At age 24, this woman was diagnosed with IIH, but was intolerant of acetazolamide. Two years after diagnosis, she continued to have headache. A repeat lumbar puncture revealed an opening pressure of 37 cmH2O. Visual function was normal. ICP monitoring disclosed ICP as high as 60 cmH2O. Therefore, she underwent VP shunt placement with a fixed pressure valve. Because of persistent postural headaches postoperatively, she underwent revision of the shunt with placement of a programmable antisiphon valve 1 week after the initial surgery. Two weeks after shunt revision, she presented with abdominal pain. CT showed an abdominal fluid collection, but ultrasound-guided biopsy revealed no infection. Headache persisted, as did mild optic disc elevation, but visual function remained normal. Eight weeks later, she developed dizziness. A shunt series performed elsewhere was interpreted as showing kinking of the distal shunt tubing, but that finding was not confirmed at our institution. Five months later, she reported increasing eye pain, blurred vision, and vomiting. Neuro-ophthalmologic examination and neuroimaging were normal. Eleven months later, she appeared with chest pain and headache. All studies were again negative. At 1 year after shunt revision, she underwent ligation of the shunt. When ICP monitoring was normal, the shunt was removed. Multiple additional emergency room visits occurred over the ensuing months, during which acetazolamide was briefly prescribed and another round of ICP monitoring was negative. Patient #6 At age 27, this woman was diagnosed with IIH after presenting with episodic vision loss right eye. MRI and MR venography were negative. Lumbar puncture opening pressure was 31 cmH2O. With escalating doses of acetazolamide, she experienced increasing generalized weakness, believed to be an exacerbation of a pre-existing mitochondrial myopathy. After discontinuation of acetazolamide, she was noted to have worsening optic disc edema. She was 230 admitted for ICP monitoring, which demonstrated pressures as high as 40 cmH2O. A VP shunt was placed. One month later, she presented with severe abdominal pain, but evaluation was negative. One year later, she presented again with severe abdominal pain. Physical examination demonstrated a 4–5 cm area of erythema, induration, and tenderness superior to the abdominal surgical scar, diagnosed as cellulitis. Shunt-related diagnostic studies were negative. After antibiotic treatment, the cellulitis resolved. Head CT, shunt series x-rays, and abdominal ultrasound were unremarkable. She was treated with broad-spectrum antibiotics. Eighteen months after shunt placement, she presented with dizziness, nausea, and pain over the shunt reservoir. Diagnostic studies were negative. The shunt was reprogrammed to drain at 18 cmH2O. Two weeks later she came again with severe headache, nausea, photophobia, and gait instability. The shunt was reprogrammed to 13 cmH2O. Twenty-two months after VP shunt placement, she presented again with severe headache, nausea and vomiting, and photophobia. A complete blood count revealed a leukocytosis of 18.8. CT Head showed an interval increase in ventricular size. The shunt was reprogrammed to 8 cmH2O. Subsequently she became febrile. Minimal CSF could be obtained via shunt tap, but a lumbar puncture showed a pleiocytosis and a negative CSF gram stain and culture. She was treated with intravenous antibiotics for meningitis and later underwent shunt revision. At surgery, the distal catheter could not be flushed even at pressures of 40 cmH2O. When the catheter was removed, it was discolored, and the distal end contained fecal material. The ventricular catheter could not be removed, so the valve was replaced with a closed valve. Spinal fluid constituents finally normalized. Four weeks after discharge, neuroophthalmologic examination disclosed normal visual function and normal-appearing optic discs. She was then lost to follow-up. DISCUSSION This study has disclosed that VP shunt failures, including elevated ICP due to obstruction or discontinuity, overdrainage, or shunt, abdominal, or scalp infection, occurred in 6 (18.7%) of 32 patients over a minimum follow-up period of 12 months. There were no deaths and no lingering medical, neurologic, or visual consequences among the patients with complications except for the need for long term anti-epileptic medication to prevent seizures in 1 patient. Some of the shunt complications were precipitated by poorly healing abdominal wounds secondary to unrelated procedures or minor scalp trauma. The 6 patients who suffered complications went through a saga of frequent emergency room visits prompted by headache, vision loss, or abdominal pain, often eliciting a need for intensive care. Complications included meningitis and intracerebral hemorrhage. Although persisting Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution morbidity was slight, it may have been worse had the patients been managed at a facility that lacked the facilities that our institution has for providing intensive medical care. Moreover, even when there were no actual complications, shunt placement so preoccupied patients that 14 of them came to the emergency room 38 times. Indeed, Murphy et al (18) have reported that patients who had undergone a shunt procedure for IIH used emergency department resources more frequently than those who had not undergone such procedures. Yet the shunt failure rate of 18.7% in our study of 32 patients is lower than the 23%–67% failure rates reported in previously published comparable studies (3,4,6,7,11). The largest study (7) to document clearly the complication rate of VP shunting included only 19 patients. In that study, the shunt survival rate was only 48% at 36 months of follow-up, and 7 (37%) had repeat shunt surgery. The next largest comparable study, by Bynke et al (3), included only 17 patients with a 41% failure rate. Because we entered only those patients in our study who had a minimum follow-up period that would encompass the vast majority of expected complications, we avoided the inherent overestimation of a retrospective survival analysis that includes patients with shorter followup times along with documented failures. We also included only patients with papilledema, who would most merit CSF diversion to protect against vision loss. When papilledema is not present pre-operatively, the basis for a diagnosis of shunt failure and the decision to replace the shunt become more difficult. Long-term survival of a shunt, which may affect provider perception of shunt performance, does not necessarily apply in CSF diversion for IIH. Because the high ICP is often self-limited, CSF diversion is typically only needed temporarily. Thus, Patient 15 in our study developed shunt malfunction 7 years after shunt placement, but clinical manifestations of IIH did not occur. Patient 6 developed imaging evidence of shunt obstruction, but lumbar puncture showed normal ICP. Patient 5 underwent elective shunt removal 1 year after shunt placement without recurrence of high ICP. The relatively lower rate of complications we have identified for VP shunts becomes relevant in the selection of surgical options. LP shunts may require more frequent revision than VP shunts and include the risks of CNS infection and, as demonstrated in Patient 2, radiculopathy or cauda equina syndrome. Optic nerve sheath fenestration risks optic neuropathy and other cranial neuropathies (8). Venous sinus stenting may be a viable alternative, but widespread experience is lacking. Although this study is the largest of its kind, its widespread applicability may be limited because it is derived from a single institution. Even so, it includes the efforts of 13 neurosurgeons performing a relatively standardized Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 procedure. Complications of VP shunts may be less common than previous reports have suggested. STATEMENT OF AUTHORSHIP Category 1: a. Conception and design: A. J. Brune, T. Girgla, and J. D. Trobe; b. Acquisition of data: A. J. Brune, T. Girgla, and J. D. Trobe; c. Analysis and interpretation of data: A. J. Brune, T. Girgla, and J. D. Trobe. Category 2: a. Drafting the manuscript: A. J. Brune, T. Girgla, and J. D. Trobe; b. Revising it for intellectual content: A. J. Brune, T. Girgla, and J. D. Trobe. Category 3: a. Final approval of the completed manuscript: A. J. Brune, T. Girgla, and J. D. Trobe. REFERENCES 1. Smith SV, Friedman DI. The idiopathic intracranial hypertension treatment trial: a review of the outcomes. Headache. 2017;57:1303–1310. 2. Kanagalingam S, Subramanian PS. Update on idiopathic intracranial hypertension. Curr Treat Options Neurol. 2018;20:24. 3. Bynke G, Zemack G, Bynke H, Romner B. Ventriculoperitoneal shunting for idiopathic intracranial hypertension. Neurology. 2004;63:1314–1316. 4. Kandasamy J, Hayhurst C, Clark S, Jenkinson MD, Byrne P, Karabatsou K, et al. Electromagnetic stereotactic ventriculoperitoneal csf shunting for idiopathic intracranial hypertension: a successful step forward? World Neurosurg. 2011;75:155–160; discussion 32–33. 5. Maher CO, Garrity JA, Meyer FB. Refractory idiopathic intracranial hypertension treated with stereotactically planned ventriculoperitoneal shunt placement. Neurosurg Focus. 2001;10:E1. 6. Abu-Serieh B, Ghassempour K, Duprez T, Raftopoulos C. Stereotactic ventriculoperitoneal shunting for refractory idiopathic intracranial hypertension. Neurosurgery. 2007;60:1039–1043; discussion 1043–1044. 7. Huang LC, Winter TW, Herro AM, Rosa PR, Schiffman JC, Pasol J, Trombly RS, Tawfik M, Lam BL. Ventriculoperitoneal shunt as a treatment of visual loss in idiopathic intracranial hypertension. J Neuroophthalmol. 2014;34:223–228. 8. Kalyvas AV, Hughes M, Koutsarnakis C, Moris D, Liakos F, Sakas DE, Stranjalis G, Fouyas I. Efficacy, complications and cost of surgical interventions for idiopathic intracranial hypertension: a systematic review of the literature. Acta Neurochir (Wien). 2017;159:33–49. 9. Woodworth GF, McGirt MJ, Elfert P, Sciubba DM, Rigamonti D. Frameless stereotactic ventricular shunt placement for idiopathic intracranial hypertension. Stereotact Funct Neurosurg. 2005;83:12–16. 10. Tarnaris A, Toma AK, Watkins LD, Kitchen ND. Is there a difference in outcomes of patients with idiopathic intracranial hypertension with the choice of cerebrospinal fluid diversion site: a single centre experience. Clin Neurol Neurosurg. 2011;113:477–479. 11. Abubaker K, Ali Z, Raza K, Bolger C, Rawluk D, O’Brien D. Idiopathic intracranial hypertension: lumboperitoneal shunts versus ventriculoperitoneal shunts—case series and literature review. Br J Neurosurg. 2011;25:94–99. 12. McGirt MJ, Woodworth G, Thomas G, Miller N, Williams M, Rigamonti D. Cerebrospinal fluid shunt placement for pseudotumor cerebri-associated intractable headache: predictors of treatment response and an analysis of long-term outcomes. J Neurosurg. 2004;101:627–632. 13. Rosenberg ML, Corbett JJ, Smith C, Goodwin J, Sergott R, Savino P, Schatz N. Cerebrospinal fluid diversion procedures in pseudotumor cerebri. Neurology. 1993;43:1071–1072. 14. Menger RP, Connor DE, Thakur JD, Sonig A, Smith E, Guthikonda B, Nanda A. A comparison of lumboperitoneal and 231 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution ventriculoperitoneal shunting for idiopathic intracranial hypertension: an analysis of economic impact and complications using the nationwide inpatient sample. Neurosurg Focus. 2014;37:E4. 15. Azad TD, Zhang Y, Varshneya K, Veeravagu A, Ratliff JK, Li G. Lumboperitoneal and ventriculoperitoneal shunting for idiopathic intracranial hypertension demonstrate comparable failure and complication rates. Neurosurgery. 2019;86:272–280. 16. Hanauer DA, Mei Q, Law J, Khanna R, Zheng K. Supporting information retrieval from electronic health records: a report of 232 University of Michigan’s nine-year experience in developing and using the Electronic Medical Record Search Engine (EMERSE). J Biomed Inform. 2015;55:290–300. 17. Friedman DI, Liu GT, Digre KB. Revised diagnostic criteria for the pseudotumor cerebri syndrome in adults and children. Neurology. 2013;81:1159–1165. 18. Murphy S, Friesner DL, Rosenman R, Waslo CS, Au J, Tanne E. Emergency department utilization among individuals with idiopathic intracranial hypertension. Int J Health Care Qual Assur. 2019;32:152–163. Brune et al: J Neuro-Ophthalmol 2021; 41: 224-232 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.