Case Report
An 80-year-old female was referred to our hospital. The patient presented with hypertension and osteoporosis. She had suffered from progressive cognitive dysfunction and gait disturbance for 9 years and had started experiencing urge urinary incontinence 4 years earlier. Based on the improvement in gait on the tap test8) and imaging findings, including a disproportionately enlarged subarachnoid space,6) the patient was diagnosed with probable iNPH9) at the age of 77 years; in the same year, she underwent VPS surgery with the Codman Hakim programmable valve (Integra LifeSciences, Princeton, NJ, USA). Her cognitive function, gait symptoms, and urinary incontinence improved after the surgery, and a follow-up examination 1 year after the surgery confirmed no problems with the shunt condition and her postoperative course. Correspondingly, head computed tomography (CT) revealed shrinkage of the ventricular volume. She fell 1 year after the surgery, broke her right distal radius, fell again 1 year later, and broke her left distal radius, both of which were treated with orthopedic surgery and healed without significant sequelae. Her family doctor referred her to our department because she had developed gait, cognitive impairment, and urinary incontinence again over the last 3 months. When admitted to the hospital, she exhibited a broad-based, unsteady, and magnetic gait. The average time of the Timed Up and Go tests8) showed deterioration from 12.7 sec (1 year after VPS) to 21.6 sec. The longitudinal changes in this test are illustrated in Fig. 1. The reservoir of her shunt valve did not collapse, and the adjustment of the valve-opening pressure did not improve her symptoms. Fig. 2 shows X-ray photographs (A, B) and parasagittal CT (C, D) 1 year after the first surgery and at the time of admission, respectively. These images suggest that the ventricular catheter had disconnected from the shunt valve and migrated into the cranium. Imaging studies also showed that the ventricles were enlarged.
Since shunt failure was believed to have caused deterioration in cognitive function and gait disturbance, a reoperation was performed. Upon removal of the shunt system, it was found that the ligature thread holding the proximal catheter in place ruptured. As the distal end of the proximal catheter was facing the burr hole (Fig. 3), it could be carefully pulled and removed. The shunt system implanted in the first surgery was entirely removed, including the peripheral catheter, and the VPS was reconstructed using new equipment. The postoperative course was favorable, with improvements in gait and cognition. Two years after the surgery, she could still walk on her own and live independently.
Discussion
CSF shunt efficacy and risk of failure in iNPH
The importance of proper diagnosis and management of iNPH has been emphasized in recent decades6,9,10) as the reversibility of iNPH symptoms after CSF shunt surgery has been recognized. The outcomes of CSF shunts for iNPH have improved over time11) through advances in both diagnosis6,8,12-17) and treatment.18-23) For example, a study reported in 1978 involving 37 cases of iNPH with CSF shunts showed up to 19 complications and a limited improvement rate of 33%, and the authors suggested that the expected benefits of CSF shunt were not reasonable compared to the risks involved in the shunt surgery in many cases,24) while a recent review concluded that more than 75% of patients experienced improvement with shunt surgery.5) Hung et al. reported that 28.9% of iNPH cases treated with VPS and 10.7% of ventriculoatrial shunts required revision surgery.25) The authors reported that the most frequent complication was over-drainage. Recent studies have shown that programmable valves significantly reduce this complication.22,23) Considering these advances, it is advisable to discuss the surgical outcomes of iNPH based on recent data.
Although the rate of shunt failure has decreased over time, some patients who undergo shunt surgery have experienced it. A recent meta-analysis reported a rate of reoperation after VPS of approximately 18%,15) while studies employing programmable valves reported rates of 8.2%26) and 15%.27) Recent Japanese prospective cohort studies of iNPH showed more favorable outcomes; the response rates 1 year after surgery were 77% for VPS and 75% for lumboperitoneal shunt, while the reoperation rates were 1% for VPS and 6% for lumboperitoneal shunt.6,19) Most shunt failures occur within 1 month after surgery, and failures occurring more than 1 year after surgery are quite exceptional.27)
Although programmable valves and antibiotic administration protocols have effectively reduced the complications of inappropriate drainage and infections,20,28) there are mechanical shunt problems. The incidences of proximal occlusion, distal occlusion, and distal catheter dislocation have been reported to be 1.7%-12%, 1%-5.2%, and 1.4%-8%, respectively.6,22,26,29) None of these mass studies reported proximal tube migration as a cause for shunt revision. In this sense, shunt failures in iNPH differ from those in pediatric hydrocephalus.
Proximal catheter migration is a relatively frequent complication in pediatric hydrocephalus cases,30) and there are some case reports of proximal tubes and shunt systems migrating into the cranium31-34) or into the subgaleal space.35,36) As Shimizu et al. pointed out, proximal shunt failure is more likely to occur in infants with an occipital burr hole due to excessive head movements leading to shunt tube dislocation and use of valveless shunt systems.37) In addition, the skull of the infant is large relative to the body and the skull is flexible. This makes it relatively easy for the shunt system to migrate intracranially. Reported cases of migration of proximal tubes in adult VPS are exceptional, such as a case in which the head was repeatedly subjected to strong mechanical impacts37) or a case with a valveless shunt system.38)
Possible mechanism for shunt failure
In this case, 3 years following VPS surgery, the proximal catheter detached from the shunt valve and entered the cranium. The shunt system required a manual connection between the proximal tube and valve, and herein lies the potential for an accident to occur. The absence of abnormal findings on imaging studies at the 1-year postoperative follow-up examination ruled out the possibility that the proximal tube and valve were not connected during the initial surgery. Therefore, subsequent events should be considered the cause of this failure.
The first hypothetical factor is the tractive force from the distal tube, which can be caused by a patient's body movements. Although the shunt valve was unlikely to have been subjected to a large force at one time as no noticeable change was evident in its position, it is possible that repeated minor traction may have caused the peripheral tube to slowly and gradually separate from the shunt valve. The second hypothetical factor is that after the proximal shunt tube detached from the shunt valve and intruded the cranial cavity to some degree, the elasticity of the proximal tube might have induced further intrusion. Fig. 4A shows a cross-section of the shunt tube and skull in a 3D reconstructed image made from CT images taken 1 year after the first VPS surgery. Once the distal end of the proximal tube approaches the burr hole (white line in Fig. 4B), the elasticity of the tube exerts a force, which acts toward maneuvering the tube from the position of the white line to that of the black line. Subsequently, the force required for its transposition from the position of the black line to that of the dashed black line acts next. The finding that the distal end of the proximal tube was located just at the burr hole may be explained by this hypothesis.
There are several possible approaches to prevent shunt-related problems, as has been outlined here. A fundamental approach is to adopt a shunt system in which the proximal catheter is not separated from the shunt valve. However, shunt devices are usually selected based on several other factors. When using a separate shunt system, the connection of the proximal tube to the shunt valve is important. Regarding the connection, we have standardized double ligation with a single nylon thread at our hospital. Alternative threads may be less likely to slip than nylon threads; however, silk threads are inappropriate in terms of infection risk. Moreover, fixation of the shunt valve to the skull may be effective, which could prevent our first hypothetical factor from occurring. However, a disadvantage is that it hinders subsequent reoperations. Another approach to improve the fixation of the shunt valve is to make a skin incision on the cephalic side, create a subcutaneous pocket toward the body side, and insert the shunt valve into this pocket. This technique may reduce the room for the shunt valve to move distally when pulled by the distal tube. Finally, our second hypothetical factor was related to the bending angle of the proximal tube near the burr hole. Thus, it would be better to make the angle obtuse.
All prevention approaches listed here are hypothetical; however, some may be worth considering in clinical practice.
Impacts of prompt operation on functional prognosis
The importance of prompt CSF shunt reconstruction must also be discussed. Some neurosurgeons hesitate to reoperate on elderly patients, not only because of the high risk of complications but also because they anticipate less favorable outcomes. Although nationwide surveys in the United States indicate that older age is a risk factor for non-routine discharge39) and readmission40) after shunt surgery for iNPH, these results reflect the confounding effects of various background factors. Some controlled studies have shown that age does not predict postoperative improvement in iNPH.25,41) Kobayashi et al. noted that the duration of untreated iNPH, but not age, predicts unfavorable postoperative improvement.42)
Recent studies suggest that shunt therapy for iNPH can benefit patients over 5 years if properly managed.26,29,43) Although no group studies focusing on improvement after shunt revision have been conducted so far, Rinaldo et al. showed that of 18 patients who underwent shunt revision, 7 (38.9%) improved, 6 (33.3%) remained the same, and 5 (27.8%) worsened in symptoms.22) In the management of iNPH, prompt treatment may be desirable regardless of age.
In this case, the patient and her family doctor noticed the recurrence of cognitive dysfunction, gait disturbance, and urinary incontinence, which led to relatively early detection of shunt failure and reoperation. After the shunt revision, the patient showed significant improvement compared with the preoperative levels (Fig. 1), although the levels observed 1 year after the first shunt surgery were not reached. If a patient with iNPH whose symptoms improve after surgery experiences another exacerbation and shows no improvement even after lowering the set pressure, shunt failure should be suspected, and a thorough examination should be performed. The cause of failure can only be identified by imaging studies, especially in cases of failure, such as in this case. Since the site of failure, in this case, was between the burr hole and shunt valve, the site of failure could have been overlooked if attention was paid only to the intracranial region on the head CT.
