History of Cerebrospinal Fluid Shunt Devices

Before 1900, treatment of hydrocephalus was often uninformed and usually ineffective. Head bandaging, intraventricular injection of a strong iodine solution, exposure of the head to bright sunlight, and irradiation of the choroid plexus were among the more extreme procedures advocated.¹ Direct ventricular puncture and repeated lumbar puncture were rarely sufficient to control hydrocephalus, and efforts were directed toward internal diversion of CSF. In the 1890s, Miculicz developed a gold, flanged hollow tube that diverted CSF from the ventricle to the subgaleal space, but this valveless device was only rarely effective.² Other valveless CSF diversion techniques involving the use of glass, rubber, or silver tubes and linen threads in the subdural or subgaleal space were all equally unsuccessful and carried high mortality. Attention then turned to the choroid plexus, with Walter Dandy promoting the technique of extirpation of the choroid plexus in an attempt to reduce production of CSF.³ Before 1950 this was probably the most common procedure undertaken for infantile hydrocephalus,⁴ but success remained limited,^5–7 and it was largely abandoned by the 1970s after reports of high failure rates.^8,9 Interestingly, there has been a recent resurgence of interest in ablation of the choroid plexus, the efficacy of which may have been underestimated.^10–12 In 1914, Heile described the first diversion of CSF from the lumbar subarachnoid space to the peritoneum with the use of a valveless rubber tube, but this too was unsuccessful.¹³ In 1939, Torkilsden described a shunt from the lateral ventricles to the cisterna magna for obstructive hydrocephalus that was modestly successful.¹⁴ In 1949, Matson described a shunt from the lumbar subarachnoid space to the ureter, which became popular in the 1950s.¹⁵ The procedure necessitated nephrectomy and was often complicated by hypochloremic alkalosis, which was frequently fatal, and with the introduction of valved shunt systems, the technique was abandoned.

The advent of the “modern” era of CSF shunt devices was heralded by the publication of Nulsen and Spitz’s ventriculojugular shunt with a ball and spring differential pressure valve.¹⁶ The first shunt to use silicone was the Spitz-Holter valve, a slit valve designed by engineer John Holter for his son who had hydrocephalus.¹⁷ At around the same time, Robert Pudenz also designed both a distal-slit and a sleeve valve, both differential pressure silicone valves for use in ventriculoatrial shunts.¹⁸ Silicone has since become the material of choice for implanted shunts. Although the initial preferred site for shunt placement was the vascular system, the risks, particularly infection and associated shunt nephritis or pulmonary hypertension,^19,20 and identification of the peritoneal cavity as a suitable site for CSF absorption²¹ led to the peritoneum becoming the preferred site for distal catheter placement. Despite several new shunt and catheter designs over the past 50 years, many of the problems associated with shunts, such as blockage, overdrainage, and infection, still persist. The search for the “ideal” shunt system or an alternative, more efficacious treatment continues.

Cerebrospinal Fluid Shunt Hydrodynamics

To understand the mechanisms behind the multitude of shunt designs available, a basic knowledge of hydrodynamics, or the physics of fluid flow, is necessary. Three important physical concepts must be understood: pressure, flow, and resistance.

Flow and Resistance

Flow (Q) in a tube is defined as the volume of fluid (V) passing a point during a given time (t) (e.g., mL/min). Flow from one end of the shunt system to the other is defined by the equation Q = ΔP/R_T + R_V, where ΔP is the difference in pressure between the ventricle and distal catheter location and R_T and R_V are the resistance of the tube and valve, respectively.

Resistance to the flow of fluid through a shunt system (R_T + R_V) depends on a number of factors. Because flow of fluid through catheters is laminar (smooth), resistance of catheters (R_T) is defined by Poiseuille’s law:

where r is the radius of the tube, L is the length of the tube, µ is the viscosity of the fluid (CSF), and r the diameter of the tube.

Laboratory studies have demonstrated that a 90-cm-long distal catheter provides an additional resistance to flow that is roughly equivalent to that provided by a differential pressure valve.^22,23

The increase in CSF viscosity (e.g., proteinaceous CSF) seen in patients with optic pathway gliomas does not have a great impact in that even the most proteinaceous CSF reduces CSF flow by only around 7%.^24,25 More importantly, CSF viscosity decreases with increasing temperature, with flow rates at body temperature some 30% higher than at room temperature, which has important implications for in vitro testing of new shunt designs, particularly those in which CSF flow occurs through a very small orifice, such as in the flow-controlled Orbis Sigma II valve.

Shunt catheter resistance rises as a fourth power of the radius, and this has been exploited in designing valveless shunt systems such as the “Mexican shunt,” which has an internal diameter of 0.51 mm as opposed to a standard catheter diameter of 1.0 to 1.6 mm.²⁶ The relationship of shunt catheter resistance to radius has also been exploited as a means of reducing excessive drainage through lumbar shunt catheters.

Although R_T has linear pressure-flow characteristics, flow through the narrow orifices of valves, despite being laminar at low flow rates, may become turbulent at higher rates of flow. Therefore, the resistance of the shunt valve (R_V) is not constant in the range of physiologic flow rates, and a nonlinear pressure-versus-flow relationship is seen. Debris and air bubbles in the shunt valve or catheter will significantly increase turbulence and restrict the diameter of the lumen, both of which will significantly increase resistance to flow; although this does not necessarily occlude the shunt, it may have a major impact on shunt performance.²³

The pressure gradient driving CSF flow in a ventriculoperitoneal shunt system is determined by the formula ΔP = IVP + ρhg − OPV − IAP, where IVP is intraventricular pressure, ρhg (h being the difference in vertical height between the head and distal drainage site) is hydrostatic pressure, OPV is the opening pressure of the valve, and IAP is intra-abdominal pressure. Thus, in the upright position, the predominant influence on the pressure gradient (and therefore CSF flow) is hydrostatic pressure, not OPV (Fig. 192-1).

FIGURE 192-1 Illustration of compartment pressures and the effect of position. Hydrostatic pressure predominates in the upright position.

(From Drake JM, Sainte-Rose C. The Shunt Book. New York: Blackwell Science; 1995:23.)

Siphoning

With differential pressure valves (see later), once the patient moves to the upright position and the valve opens, the hydrostatic forces acting on the shunt system will predominate and result in excessively high flow rates despite negative intracranial pressure (ICP). In a valveless system, ICP would continue to fall until IVP equals negative ρhg to balance the siphon effect (Fig. 192-2). Such a drop in ICP does not occur in a normal brain because there is no posture-related change in the CSF–sagittal sinus pressure gradient.²³ The exception occurs in patients who have undergone a wide decompressive craniectomy, in which the head is essentially exposed to atmospheric pressure. In these patients in the upright position, CSF will continue to siphon until the ventricles are emptied and the craniectomy is maximally sunken, which may result in marked shift and deformation of the underlying brain tissue and have significant neurological sequelae, the “sinking skin-flap syndrome.”²⁷

FIGURE 192-2 Changes in intraventricular pressure that occur with siphoning. Pressure (P) and flow (F) in horizontal (A) and vertical (B) position are shown.

(From Drake JM, Sainte-Rose C. The Shunt Book. New York: Blackwell Science; 1995:21.)

The excessively negative pressures generated by siphoning are surprisingly well tolerated by the majority of patients, but around 10% of patients will experience low-pressure symptoms.^28,29 Other serious sequelae include ventricular collapse with tearing of bridging veins and subdural hematoma formation, premature suture closure, acquired aqueductal stenosis, Parinaud’s syndrome, and slit ventricle syndrome.^28,30–33

It should be remembered that raising the opening pressure of the valve will decrease the magnitude of the negative IVP generated by siphoning but will not prevent siphoning or its clinical sequelae from occurring because the dominant influence of hydrostatic forces will persist.

Proximal and Distal Shunt Catheters

All shunt catheters (both proximal and distal) are made of artificial silicone rubber or polyurethane. The catheters are stiff enough to resist kinking but compliant enough to minimize the risk of brain injury as the ventricles reduce in size and the catheter comes in contact with the ependyma. Catheters are available in a range of internal and external diameters, the smaller internal diameters being used in neonates or valveless shunt systems to add further resistance to CSF flow. Most modern catheter designs are impregnated with tantalum or barium to facilitate radiologic identification. The latter is associated with an increased rate of distal shunt catheter deterioration and host reaction leading to calcification and loss of elasticity and strength of the catheter tubing.^34,35 The catheters then become tethered, typically in the neck, and are prone to fracture, particularly in growing children.³⁶ To mitigate this complication, some manufacturers now produce catheters with a barium strip or coating of pure silicone to reduce host reaction to the tubing.

Packaged catheters carry a static charge and, when opened, can attract airborne dust particles carrying microorganisms; accordingly, non–antibiotic-impregnated catheters should be soaked in sterile saline solution immediately on opening to reduce the risk of contamination. To reduce the risk for shunt infection, manufacturers have introduced specialized catheters, some of which are impregnated with antibiotics, such as the Bactiseal catheter system, which is impregnated with clindamycin and rifampicin (Codman, Johnson & Johnson, Inc., Raynham, MA) and releases the antibiotics in the weeks after implantation to potentially reduce the risk for shunt infection by preventing biofilm-forming organisms from colonizing the catheter. Other manufacturers have developed catheters that are impregnated with silver nanoparticles (Silverline, Speigelberg, Hamburg), which have antibacterial properties,³⁷ or coated with antibiotics to reduce the risk for shunt infection. It should be noted, however, that to date, no prospective multicenter randomized controlled trials have been completed that demonstrate an overall reduction in infection rates with any of these catheters. Some retrospective studies have shown promising results in high-risk populations,^38–40 but other series have not demonstrated a benefit,⁴¹ and a relative increase in more severe gram-negative infections may offset any benefit from an overall reduction in infection rates. Other measures such as the intraventricular administration of antibiotics at the time of shunt implantation may be of similar efficacy.⁴²

The proximal shunt catheter is usually placed in the lateral ventricle. Occasionally, catheters are placed within the subarachnoid space, arachnoid cysts, syrinx cavities, and subdural hygromas. The most common cause of shunt malfunction is blockage of the proximal catheter, which is usually secondary to ingrowth of choroid plexus. Attempts to identify a preferred site for catheter placement remote from the choroid plexus have been unsuccessful.⁴³ Even endoscopically assisted placement of proximal catheters does not reduce blockage rates,⁴⁴ although post hoc analysis of the data suggested that if catheters could be accurately positioned away from the choroid plexus, shunt failure could be reduced. A variety of proximal catheter designs with baskets, flanges, or recessed holes, as well as the “J”-shaped Hakim catheter with holes on the inside curve of the “J,” have been produced in an effort to reduce mechanical obstruction by the choroid plexus, but none have been successful in reducing ventricular catheter blockage rates. Endoscopic coagulation of the choroid plexus itself on the side of the shunt may be the most effective means of reducing proximal catheter obstruction (Fig. 192-3).⁴⁵

FIGURE 192-3 Proximal catheter blockage-free survival in shunted patients who had previously undergone endoscopic choroid plexus coagulation (CPC) versus those with no previous choroid plexus coagulation (VPS).

The size and number of drainage holes in the tip of the catheter vary, but most holes are redundant. Laboratory studies have shown that even if only a single hole remains patent, there is no significant increase in the total resistance of the shunt system.⁴⁶ Ventricular catheters come with a stylet to facilitate passage of the catheter through the brain parenchyma. Hollow stylets to allow visualization of CSF when the ventricle is cannulated are now available. A number of devices can be used to facilitate proximal catheter placement, including the Ghajar guide, a tripod designed to ensure a perpendicular trajectory to the ventricle from a coronal approach,⁴⁷ as well as ultrasound probes and intraluminal ventriculoscopes.^44,48 All these devices can assist in either accessing the ventricle or confirming position within the ventricular system (or both). Recently, frameless, image-guided neuronavigation has been used to facilitate catheter placement,⁴⁹ and the advent of electromagnetic navigation technology has enabled the use of such neuronavigation in infants.⁵⁰ When contoured or cylinder valves are used, external right-angle guides are used to reduce the risk of shunt migration, but in very small infants these guides may be unduly prominent and increase the risk for scalp breakdown. A variety of rigid connectors (either polyethylene or titanium) are available, either straight, right angled, or “Y,” “X,” or “T” shaped to facilitate the assembly of complex shunt systems. Bur hole reservoirs that redirect the flow of CSF 90 degrees can also be used in conjunction with contoured valves, particularly if the valve has no integral reservoir of its own.

Distal shunt catheters may have a distal slit or a single open-ended lumen. The latter is associated with a significantly lower rate of distal catheter occlusion,⁵¹ and we advocate removal of any distal slits before intraperitoneal placement. When the distal catheter is placed in the vascular system, a distal slit valve is required. Recently, distal catheters have been placed with laparoscopic assistance. This may be useful when cosmesis is a major consideration or when coexistent intra-abdominal pathology such as adhesions or obesity may compromise optimal placement, and it allows confirmation of the implanted functioning shunt system.^52–57

Shunt Valves

There are an enormous number of shunt valves available on the market today. This is a reflection of the failure of any single device to demonstrate functional superiority in terms of valve performance, longevity, and complications. A new valve design may solve one problem, but only at the expense of another and with no net reduction in shunt-related morbidity. Valve types may be categorized by their mechanism of action: differential pressure valves, which open when the differential pressure of the fluid across the valve exceeds the opening pressure of the valve; flow-controlled valves; and gravitational (gravity-actuated) valves. Valves may be fixed pressure or “programmable” (i.e., have adjustable valve opening pressure). Devices intended to reduce siphoning are also available as either separate components or integrated into the valve design itself.