The complication rates associated with shunting CSF for the treatment of hydrocephalus remain unacceptably high despite many attempted technical improvements since shunts were first invented in the 1950s. Modifications have been made in recent decades with features intended to compensate for the changes between intracranial pressure (ICP) and receiving sites such as adjustable valves or antigravitational features, but even recent reviews fail to register the impact of these enhancements on shunt failure rates. Current technology does not address a fundamental issue: the choice of drainage sites (most commonly the peritoneum and atrium) is "unphysiological" and the fluctuating pressure differences between these sites and ICP will therefore always be a challenge to manage mechanically.

Shunting implies intervention in a complex, multidimensional system involving physiological variables such as CSF production, CSF outflow, pressure, cerebral compliance, cardiac output, body position, and physical activity. When shunting CSF to the intraperitoneal cavity, for example, which is standard practice for most hydrocephalus shunt operations, the flow in the shunt is entirely dependent on the pressure difference between the intracranial and intraabdominal compartments. The intraabdominal pressure fluctuates greatly in the short term as well as the long term and depends on posture and physical activity such as breathing, walking, lifting, etc.,¹ as well as obesity, among other factors.² When taking all these factors and their delicate balance into consideration, the diversion of CSF to the peritoneal cavity or right atrium of the heart by means of a drain with a predetermined flow or pressure restriction is highly artificial; the known risks of complications and side effects are therefore considered somewhat inevitable. However, because this intervention is the standard of care for managing the condition, the inherent complications are reluctantly tolerated as an unfortunate necessity.

Manufacturers have attempted to overcome the limitations of shunts with the introduction of antisiphoning devices (more accurately, devices with valves to compensate for siphoning), programmable shunts (adjustable valves), or self-adjusting CSF flow-regulating shunt devices. This has not, however, resulted in significant reductions of reoperations or in extended revision-free periods after shunt implantation.³ In their Cochrane review, Garegnani et al.⁴ concluded: "Standard shunt valves for hydrocephalus compared to anti-siphon or self-adjusting CSF flow-regulating valves may cause little to no difference on the main outcomes."

In fact, the frequency of reoperations within the first 6 months following shunt implantation is generally reported to be 25%–30%.^4–6 A recent report cites a particularly poor outcome of 32.6% shunt failures 30 days after insertion in patients older than 50 years in the US.⁵ The 5-year survival rate for any type of CSF shunt investigated is only approximately 50%, inevitably resulting in many repeated surgeries.⁶ There is a lack of accurate comprehensive data to be found on shunt revision surgeries across different geographical regions, but the annual figures available from the Danish national health database that captures all procedures in the country show a 58% revision and removal rate of hydrocephalus shunts (2021).⁷

Unwaveringly high failure and complication rates and the absence of breakthrough devices are compelling reasons for exploring new methods for diverting CSF. This study builds on an earlier 2002 Danish study⁸ in which 150 patients were operated on using ventriculosinus (VS) shunts and showed immediate clinical benefit until issues were discovered with the outlet. The outlet, which at the time was a simple silicone catheter, had drifted to the side of the vein and had become enveloped by endothelial growth within 3 months. The invention of a new outlet device specifically designed to overcome these earlier failings and intended for placement in the jugular foramen provided the opportunity to examine the method of VS shunting. The primary objective of the study was to test the new experimental outlet device for patency for 6 months.

Methods

Surgical Procedure

A standard ventricular drain was inserted in the right or left frontal horn of the cerebral ventricles via precoronal burr hole and connected to the subcutaneously placed control reservoir and one-way valve. The newly designed protective nitinol frame on the silicone catheter of the investigational device collapses into a standard introducer. The Seldinger technique was then used to insert the guidewire, dilator, and Peel-Away sheath (Cook Medical) into the IJV (Fig. 1) The introducer containing the device was inserted through the Peel-Away sheath and then radiographically guided as far as the junction between the sigmoid sinus and top of the IJV before the nitinol frame was released and expanded in the jugular foramen. The silicone catheter of the device was then connected to the unidirectional fixed pressure valve before being led subcutaneously to the unidirectional valve control reservoir. The correct position of the investigational device was confirmed by a final radiograph (Fig. 1). The device and accessories can also be seen on the CT scan in Fig. 2.

Step-by-step procedure for the VAP. The patient is draped and prepared in the usual fashion in a supine position with the head turned slightly to the left to expose the right neck region. An ultrasound-guided marking for entry to the IJV (A). A small skin incision was made above the IJV. The platysma muscle was bluntly dissected. Using ultrasound, a needle was inserted into the IJV and a guidewire was introduced via the needle and led up to the jugular foramen under radiographic guidance (B). A dilator (trochard) with a peel-away sheath was placed over the guidewire (C) and pushed into the IJV. The dilator was then removed and the introducer containing the investigational device (the VAP inserted using a guidewire) was led up to the jugular foramen via the peel-away sheath under radiographic guidance (D). While holding the device in place, the introducer was retracted (E); the investigational device will then unfold in the foramen. To prevent backflow, the investigational device and tube were temporally closed off with a clamp. A ventricular drain was inserted in the lateral ventricle in a standard fashion. Before the assembly of the shunt system (control reservoir and unidirectional valve), the VAP was thoroughly flushed with Ringer’s solution. Correct placement of the VAP at the venous sinus in the jugular foramen was confirmed at the end of surgery via radiographic guidance (F).

A 3D reconstructed CT scan showing the position of the VAP.

Aspirin Use

Antiplatelet therapy (aspirin, 75 mg/day) was administered for 8 weeks after shunt implantation due to the possibility of the distancer provoking an endothelial reaction leading to platelet aggregation when it was released in the jugular foramen. Platelet inhibition using aspirin reduces endothelial dysfunction and platelet aggregation and thereby diminishes thrombus formation.⁹ Once the endothelium had accommodated the nitinol frame and the inflammatory reaction had ceased, the risk of thrombus formation was considered minimal.

Water Column Test

The water column test was performed at 3 and 6 months postoperatively by penetrating the silicone dome of the control reservoir with a slim cannula (0.6-mm diameter) coupled to a transparent tube containing sterile physiological saline solution. This cannula was inserted percutaneously into the control reservoir. Keeping the tube in a vertical position, the fluid flows into the drain and passes through the device if it is patent. Free in-flow is defined as continuous steady flow of saline solution from an approximately 20-cm height to an approximately 5-cm height. The pressure in the control reservoir reflects the pressure in the ventricles (ICP). The control reservoir has a one-way valve with zero opening pressure, therefore the pressure measured by the water column test is, in this way, the same as that of the ventricles.

Investigational Device and Shunt Accessories

Since the purpose of the investigational device was to support drainage of CSF from the ventricles into the intracranial venous sinus as part of a shunt system, complementary off-the-shelf shunt accessories were required. Both the investigational device and the CE-marked accessories were provided by CSF-Dynamics A/S.

The investigational device is called the venous access port (VAP) and is manufactured from only three components, each with critical functions (Fig. 3): the silicone catheter, resistance tube, and nitinol frame. The resistance tube and nitinol frame are at the distal end of the catheter, which diverts fluid from the complementary valve (see accessories description below). The function of the frame is to support and center the outlet, ensuring that it does not contact the endothelium on the inside of the vein. The function of the polyetheretherketone (PEEK) tube, fitted at the end of the catheter, is to provide built-in flow resistance. The flow resistance in the assembled VS shunt is critical to maintain normal ICP and avoid overdrainage. The resistance tube is proportioned to achieve resistance of 10 mm Hg/ml/min following the Davson equation^10,11 (see Discussion below).

Technical drawing of the VAP showing the nitinol frame, PEEK resistance tube, and silicone catheter.

The accessories comprised a standard ventricular drain, control reservoir with unidirectional valve, and fixed-pressure unidirectional valve. The function of the main unidirectional valve with built-in control reservoir was to allow verification of patency, and the purpose of the secondary unidirectional fixed opening pressure valve was to ensure backflow did not occur. The accessories can be seen assembled with the investigational device in Fig. 4.

Schematic drawing of the shunt system with the control reservoir, unidirectional drain, ventricular drain, and VAP. © CSF-Dynamics A/S, published with permission.

Results

Twelve patients diagnosed with normal pressure hydrocephalus (NPH) aged between 66 and 82 years were treated with a standard ventricular drain, standard control reservoir with unidirectional valve, standard fixed-pressure unidirectional valve, and the investigational outlet device positioned at the junction of the jugular foramen/sigmoid sinus. Two patients did not reach the endpoints. Patient 4 (Table 2) died from pneumonia just before the 6-month follow-up; the VAP had been patent until at least the 3-month follow-up. Patient 12 (Table 2) was found to experience no clinical effect from the shunt at the 3-month follow-up, and the water column test demonstrated that the device was not patent. Shunt revision was planned; during the reoperation it was observed that the device had not been correctly positioned in the jugular foramen during surgery as intended, but low in the IJV where it was not protected by the fixed structure of the jugular foramen. In the remaining 10 patients, the shunt was patent and working at the 3- and 6-month follow-up evaluations.

While all adult types of hydrocephalus patients were eligible for enrollment, the study was conducted during the COVID-19 pandemic, when waiting times for MR venograms and lung scintigraphy were extended significantly. It was not possible to coordinate a fast turnaround of these examinations needed for patients with (for example) subarachnoid hemorrhage, which meant that the department could only recruit patients with NPH to conduct the study.

ICP pressure measured in the control reservoir via the water column tests was within normal limits in all 10 remaining patients at 3- and 6-month follow-up evaluations showing free flow through the outlet. No sign of overdrainage was observed on follow-up MRI. No thrombus formation or endothelial proliferation was noted in any of the patients on MR or CT venograms. Lung perfusion scintigraphy showed no signs of lung emboli. The results are summarized in Table 2. Apart from the 2 patients mentioned above who were omitted from these final results, in 10 patients the endpoint was achieved, with the outlet remaining patent for at least 6 months.

The clinical effect on gait disturbance, dementia, and incontinence showed improvement in general and was recorded as the secondary endpoint (Table 2). These data were based on the patients’ and/or relatives’ own observations and the sample size is too small to allow for any comparisons to historical data from the literature.

Discussion

The study was designed to demonstrate that the investigational device can remain patent when placed in the intracranial venous sinus and that it can function effectively for at least 6 months. This criterion is based on the experience from the initial study in 2002 in which more than half of the shunts became blocked at the end of the outlet in the vein (overgrown by endothelium) within approximately 3 months. Because the general survival curve of shunts decreases exponentially, we opted to double the time frame to demonstrate significant improvement. In addition, 6 months is often used in the literature to determine the general survival rate of standard shunts and is reported to be between 30%⁶ and 50%,⁵ the latter based on specific data for adults and the elderly.

The initial study on intracranial venous shunting was led by Børgesen at the University Hospital Copenhagen in 1999 using the first generation of the SinuShunt (CSF-Dynamics A/S). This shunt system worked as expected in the initial trials and resulted in immediate relief of hydrocephalus symptoms and did not cause symptoms of overdrainage. The results on the initial 46 patients were reported in 2004.⁸ Further data collected from the neurosurgical departments participating in the trial (150 total patients) confirmed these observations on the immediate clinical effect of shunting to the intracranial sinus. However, the study at the University Hospital Copenhagen was unable to demonstrate long-term success since after the preliminary clinical results were published, it was found that the majority of outlets had ceased to function after approximately 3 months and had become embedded in endothelium, with the pouch formed obliterating the distal end of the outlet. Explantations revealed an important discovery: there was no evidence of blood clots inside or around the outlet nor was there any sinus obstruction by thrombosis. Similar results of shunting to the sagittal sinus were reported by Baert et al.¹² in 2018 in which clinical benefit was also observed but where the devices rapidly became occluded. Likewise, Toma et al.’s¹³ review of 7 case series comprising 265 VS shunts noted that the clinical effect was satisfactory with no cases of thrombosis or occlusion found.

While the investigational device is, stated simply, a drainage catheter, the design addresses the challenges of ICP control as well as providing a solution to tackle overdrainage that occurs when too much fluid is drawn out of the ventricles (the siphon effect, described below). Despite antisiphon features being added to many shunt designs, overdrainage is still a common issue and causes a miserable quality of life for patients who experience it as well as being the cause of malfunction.

The device tested builds on the concept of the first generation of the SinuShunt, the functionality of which was based on resistance to outflow. The principal design change since the first generation is the introduction of the nitinol frame at the end of the outlet. In the 20 years that passed since the first trial was halted, vascular nitinol stents had become widely used in the transverse sinus for treatment of intracranial hypertension. Since the safety of nitinol had thus been established for long-term use in the vasculature, the decision was made to borrow from stent technology and use a stent-like frame to overcome endothelial encasement of the silicone drain found in the initial 2002 study.

The resistance of the assembled shunt remains critical to the balance of ICP and overdrainage and is set following the Davson equation^10,11 ICP = FR × Rout + Pss, in which FR is the formation rate of CSF, Rout is the resistance to CSF outflow (R_out), and Pss is the pressure in the sagittal sinus. Opening pressure of the one-way valve is 5 cm H₂O, and the resistance of the entire system is designed and dimensioned to 10 mm Hg/ml/min. The setting of the flow resistance is based on the R_out as measured in healthy patients by Ekstedt¹⁴ and by Albeck et al.¹⁵ with normal values ranging from 6 to 12 mm Hg/ml/min.

The device is designed to be placed in the jugular foramen to avoid the siphon effect. The siphon effect, or hydrostatic force, is commonly experienced with standard ventriculoperitoneal and ventriculoatrial shunting due to the height difference between ventricles and the shunt outlet and is extremely difficult to resolve mechanically because the force changes in tandem with the constantly changing position of the patient (sitting/standing/lying down). Placing the outlet in the jugular foramen ensures there is almost no height difference regardless of the individual’s physical position.

Different placements have been investigated for the VS shunt outlets in the past,^12,16,17 but the jugular foramen is particularly well-suited as an outlet site for several additional reasons. Primarily, it is in close proximity to the intracranial cavity while at the same time offering a fixed placement for a centralizing, supporting frame such as the one designed. The effect of the location on ICP is important: instead of ICP being determined by the shunt itself (e.g., in shunts with adjustable pressure) in combination with a distant drainage site with its own pressure variables (e.g., the peritoneum), the ICP is determined by the pressure in the venous sinus and the built-in flow resistance of the outlet. The jugular foramen is also surrounded by bone, apart from a small part of its anteromedial wall,^18,19 and will therefore not collapse; the end of the vein, the jugular bulb, is made of dura mater, a strong fibrous wall. Branches of the vagus and hypoglossal nerves are located in the anteromedial part of the foramen, separated from the vein by a fibrous membrane, a continuation of the dura. The tympanic nerve (Jacobson’s nerve) is also situated in its canal in the medial part of the foramen, however, because all the nerves are covered by bone or a fibrous membrane, they are protected from the slight mechanical force exerted by the expanded frame, thereby avoiding possible side effects such as hoarseness or dry mouth. The mean diameter of the approximately round IJV in the foramen is 9 mm with a standard deviation of 2.5 mm, and the median is 8.5 mm.²⁰ The nitinol frame of the investigational device has an unfolded diameter of 12 mm and is able to adapt to a noncircular form, and will thus be in contact with the whole circumference of the wall and able to stay in position on a long-term basis while keeping the outlet centrally placed within the frame. In addition to the advantages, the insertion procedure developed in the current study is intended to be easy to perform. The outlet device is inserted with an introducer sheath via the IJV in the neck using a standard intravascular technique and standard imaging technology (fluoroscopy).

In pursuit of the ideal shunt, the method of shunting to the intracranial venous sinus merits serious consideration because it offers several advantages over alternative diversion sites. The drainage of CSF is controlled only by the pressure differences between the ICP and the pressure in the sinus, therefore there is no need for antisiphon devices because the siphon effect is eliminated. Overdrainage will not occur from a change of position between the horizontal and vertical posture. When overdrainage is avoided, so are complications of subdural fluid collections or subdural hematomas. Fluctuations of ICP caused by fluctuations of the peritoneal pressure are avoided. The common complications related to the peritoneal drain, e.g., cyst formation, bowel perforation, peritoneal irritation, and drain displacement which will lead to shunt revision are avoided. And long periods of low ICP are avoided. Intracranial venous sinus drainage could also reduce the incidence of normal tension glaucoma; Hamarat et al.²¹ found that the risk of developing normal tension glaucoma after shunting for NPH was over 25%. It is also important to note that the placement of a device in the jugular foramen will not hinder the Starling resistor effect of the collapse of the IJV in the neck (shown by Holmlund et al.²²). The IJV collapse maintains ICP when in the upright position.²³ The IJV collapse is still possible with the device located at the point of the jugular foramen as the collapse takes place below the jugular foramen and therefore its regulating effect on ICP is not affected.

One drawback of shunting to the intracranial sinus is that the growth of the cranium in infants and children cannot be compensated through an extended drainage catheter length as it can in ventriculoperitoneal shunting. A device that can accommodate this growth will need to be developed for this group of patients. In summary, the potential exists for mimicking physiological drainage providing the outlet device placed in the intracranial venous sinus does not occlude, and it is possible to manage ICP and imitate the balance between CSF production, reabsorption resistance, and pressure in the receiving compartment by diverting CSF into the same cranial space as normalcy. In so doing, overdrainage due to hydrostatic pressure differences and complications from drains and/or the drainage site common to standard shunts can be avoided.

Conclusions

The current study represents the first attempt to place a specially designed outlet device in the jugular foramen for the purposes of diverting CSF while mimicking physiological principles. The patency of the outlet at 6 months indicates that a shunt outlet can remain in the vein without occlusion by endothelial hyperproliferation or thrombus formation and that the technological improvements made are effective. Shunting to the intracranial sinus with this device can therefore be considered a realistic future possibility for diversion of CSF with relatively few adaptations to existing practices of shunt surgery, but with significant gains by avoiding the siphon effect and lowering shunt failure rates. These findings will be explored in a larger study to further demonstrate longevity and measure clinical benefit.

Acknowledgments

We thank Katrina Gillard of CSF-Dynamics A/S for contributions to the manuscript and proofreading. This study was funded by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement no. 849502.

Disclosures

Dr. Andersen reported nonfinancial support from CSF-Dynamics A/S (travel reimbursement for a conference) during the conduct of the study, grants from the Danish 3R-Center, and nonfinancial support (fundoscopy equipment and travel for a conference) from StatuManu Aps outside the submitted work. Dr. Børgesen has received personal fees from CSF-Dynamics A/S during the conduct of the study, is the inventor of the investigational device, and is Chief Clinical Officer of CSF-Dynamics A/S.

Author Contributions

Conception and design: Munthe, Pedersen, Poulsen, Børgesen. Acquisition of data: Munthe. Analysis and interpretation of data: Munthe, Andersen, Børgesen. Drafting the article: Munthe, Børgesen. Critically revising the article: all authors. Reviewed submitted version of manuscript: Munthe, Poulsen, Andersen, Børgesen. Approved the final version of the manuscript on behalf of all authors: Munthe. Statistical analysis: Munthe, Børgesen. Study supervision: Munthe, Poulsen, Børgesen.

Supplemental Information