Procedure

The general procedure is illustrated in Figure 16.1. The patient is positioned depending on the location of the incision. For a frontal insertion, a curvilinear incision is made approximately overlying Kocher’s point. An additional retro-auricular or cervical incision is also often required as the frontal location requires a longer tunneling path. For a parieto-occipital placement, a single comma-shaped retro-auricular incision is made behind and above the pinna.

Figure 16.1 Ventriculoperitoneal shunt insertion. The distal catheter generally is inserted off midline, through the rectus abdominis muscle and fascia, although many other techniques can be used.

For both procedures, a 10 blade scalpel and the Bovie, in cutting and coagulator mode, are used to make the incision through the skin and galea, while preserving the pericranium. A subgaleal pocket is then made to house the valve and reservoir away from where the incision is located. A cerebellar self-retaining retractor is inserted. Next, using a perforator drill, a burr hole is made. The dura is then coagulated and cut using the Bovie knife at power of 18. The pia-arachnoid is then gently coagulated with the bipolar coagulator.

Following tunneling of the catheter from the distal site to the proximal incision, the valve is secured to the proximal and distal catheters with silk ties, and is positioned in the aforementioned subgaleal pocket. The ventricles are then again outlined using the CT-based wand and the best trajectory to enter the ventricles is again confirmed. Next, the wand, which is covered by a 9 French peel-away sheath, is advanced gently into the lateral ventricle, while confirming optimal trajectory in real time with the stereotactic system. For most patients with parieto-occipital trajectories, we cross midline with the wand to enhance communication between the lateral ventricles, maximize drainage, and minimize proximal catheter obstruction or differential ventricle collapse. Once the optimal location is found and CSF is obtained, the wand is removed, leaving the peel-away sheath in place. The appropriately sized ventricular catheter (generally 5 ± 2 cm for frontal approach and 9 ± 2cm for parieto-occipital approach), which is already connected to the valve, is then advanced into the peel-away sheath. The peel-away sheath is then stripped away. Next, 4–0 Nurolon sutures are used to anchor the new valve to the pericranium. Pumping of the valve is then performed, and the exit of CSF from the distal end is confirmed prior to insertion of the distal catheter. The incisions are then closed as described previously.

Discussion and complications

It has been demonstrated that the location of the catheter tip has a significant correlation with the rates of shunt malfunction, as a catheter tip entirely in CSF space had a failure rate of around 20%, compared to around 33% for a catheter tip only partially surrounded by CSF [10]. Wan et al. [11] studied different factors that affected optimal ventricular catheter placement in a retrospective study of 141 patients of all ages who underwent freehand catheter placement, and found that, as expected, patients with ventriculomegaly had better catheter placement. Additionally, they found that patients who underwent a frontal burr hole approach to the frontal horn had better catheter placement than patients who underwent a parieto-occipital burr hole approach to the frontal horn. Lind et al. [12] demonstrated similar results in patients over age 16 using a freehand method, as frontal catheters were found to be in the frontal horn target zone at a significantly higher rate than parieto-occipital catheters, with rates of 64% and 42%, respectively. The same group also performed a morphometric study based on imaging of normal adults and those with acute hydrocephalus to determine the margin of error of the different ventricular catheter approaches [13]. They found that the parieto-occipital placement had a significantly smaller margin of error than the frontal or parietal approaches, with mean range of successful trajectories of 8° in the sagittal plane and 11° in the coronal plane for parieto-occipital placement, compared to 42° in the sagittal plane and 30° in the coronal plane for frontal placement, and 23° in the sagittal plane and 36° in the coronal plane for parietal placement. They also found that there was no significant difference in successful trajectories with slight variations in the burr hole location. However, even with the wide range of successful trajectories with frontal placement, there was only an 86% chance of successful catheter placement.

As such, the authors advocated the use of image guidance or stereotactic navigation for catheter placement, particularly for the parieto-occipital burr hole. We agree with this recommendation at our institution and perform the majority of our adult shunts using stereotactic navigation with the Vector Vision BrainLab software (Heimstetten, Germany). All patients receive preoperative CT scans with thin (2–3 mm) slices of the brain and skull with six radiopaque fiducial markers in place. MRI can be used alternatively, but is generally more difficult and time-consuming and at a greater cost than CT. The patient is positioned in the Mayfield clamp, and a reference star is then used to register the navigation system. As described above, the wand, surrounded by a peel-away catheter, is then passed into the ventricle. This protocol was described previously by our group in the treatment of PTC, in which patients typically have small ventricles [14]. MRI guidance was used to place frontal ventricular catheters, and all catheters were successfully placed in slit ventricles with a single pass, with no shunt malfunction due to proximal obstruction found over the 24-month follow-up period. A similar setup was used by Gil et al. [15] in a small series of children with slit ventricle syndrome or small ventricles, and had no complications during surgery; additionally, none of the patients required revision of the proximal catheter during the 8-month follow-up period. A different group utilized an alternative CT-based negative system, the Stealth Station (Medtronic Sofamor-Danek), which does not require the placement of fiducial markers preoperatively. In a series of 26 adult and pediatric patients with ventricles of varying sizes, the operative time was still short at an average of 46 min, and 20 of the 21 patients had excellent catheter positioning while 1 patient had acceptable catheter positioning [16].

Although there has been success with stereotactic guidance for ventricular catheter placement, other methods have been used to improve proximal catheter location. For instance, Villavicencio et al. [17] utilized direct visualization of catheter placement using an endoscope system in which a semi-rigid endoscope was used as the catheter stylet. In a retrospective review of pediatric patients who underwent shunt placement with or without the endoscope, they demonstrated that endoscopic placement did not affect overall shunt survival, but did significantly decrease the rate of proximal catheter failure, with no increase in the infection rate. Additionally, although the use of intraoperative guidance has been used previously for insertion of ventricular catheters in infants with an open fontanelle [18], Whitehead et al. [19] adapted the procedure for use in older children with closed fontanelles. They made 1.5–2 cm burr holes that were large enough for placement of the ultrasound transducer head and catheter tip simultaneously, allowing for catheter placement under ultrasound guidance. In their series of 10 children, all catheters were placed with a single pass, postoperative CT scans confirmed good catheter placement within the ventricle, and there was no evidence of intraventricular hemorrhage postoperatively.

As mentioned above, we advocate the use of a peel-away sheath to aid with placement of the proximal catheter. This serves two functions: it prevents obstruction of the ventricular catheter perforations with brain tissue during passage into the ventricle, and it allows for passage of the catheter through the exact trajectory that the wand is passed under image guidance. The use of a similar system, though with a stylet in place of the wand in the peel-away sheath, led to a reduction in proximal ventricular catheter obstruction in the first year postoperatively from 18% to 5%, though this result was not significant due to the small sample size.

The most common cause of proximal shunt malfunction is obstruction of the catheter tip with choroid plexus and ependymal tissue [20]. This can cause additional problems during proximal shunt revision as the catheter is often difficult to remove from the ventricle, and more forceful removal will often result in intraventricular hemorrhage. Although the use of a peel-away sheath as described above can diminish the risk of occlusion of the catheter perforations with brain tissue, an alternative solution was proposed recently in which a metal guidewire inside a plastic sheathed 21-gauge needle was passed percutaneously through a Rickham reservoir and then coagulated with a monopolar in 5 s bursts to coagulate any debris or choroid plexus obstructing the catheter tip. This technique relieved the ventricular catheter blockage in their entire series of seven patients [21]. A downside to this approach is that it requires a Rickham reservoir as the reservoir access point is directly in line with the proximal catheter. We have adapted a similar approach that aids with proximal catheter removal during shunt revision or removal [22]. In our approach, intraoperatively, the proximal catheter is disconnected from the reservoir, and a stylet is passed down the proximal tubing. The stylet is then coagulated with a monopolar as described above, and the proximal catheter can then generally be removed with minimal force.

If the proximal catheter removal or placement does result in intraventricular hemorrhage, there are several approaches in management. If there is severe intraventricular hemorrhage, often placement of a ventriculostomy catheter is required. Simultaneously, if there is an adjustable shunt valve, the valve can be dialed up to the highest setting in an attempt to encourage drainage of the intraventricular blood through the ventriculostomy in an attempt to salvage the shunt. Before the ventriculostomy is removed, the shunt should be accessed to check for proximal flow, or alternatively, a radionuclide shunt patency test can be performed to verify the patency of the shunt. If a shunt revision is required, as the blood clears, the CSF can be sampled serially for protein level. We generally refrain from revising the shunt until the protein level drops below a level of 100, as levels above 100 tend to lead to shunt obstruction.

Procedure

We make a paramedian incision in the upper quadrant, just lateral to the midline. The skin and subcutaneous layers are incised, and a self-retaining retractor is inserted. Next, the fascia is incised parallel to the direction of the rectus muscle fibers. The fibers of the rectus muscle are then separated bluntly, followed by incising the inner fascia of the muscle and the parietal peritoneum. When this is done, small mosquito clamps are used to hold the edges of the incision. After inserting an Olivecrona dissector in the peritoneum to verify the proper location, a purse-string suture is applied to the incision. The distal peritoneal catheter can then be tunneled to the proximal incision site. Once the valve is in place and catheters are attached and secured, the distal catheter is advanced into the peritoneal incision, and the wound is closed as described previously (Figure 16.1).

Discussion and complications

Although we typically utilize an open approach to peritoneal shunt placement due to comfort with the procedure, other groups have utilized laparoscopic insertion of the peritoneal catheter, either alone or in conjunction with general surgeons. For instance, Stoddard and Kavic [23] described an approach in which a small incision was made for passage of a Veress needle, and a separate small incision was used to place a trocar to pass an endoscope. The distal catheter was then passed into the peritoneal cavity with a Hickman introducer kit. This offered the advantage of a less invasive incision, as well as visual confirmation of distal shunt placement and drainage of CSF from the distal catheter. Furthermore, due to the smaller fascial incision, there is theoretically a lower risk of distal catheter withdrawal and hernia formation, though this has not been studied in detail yet. Other variations of this technique using a single trocar and access hole have been described as well [24,25].

Although the most common complications result from distal catheter infection, obstruction, or fracture, and management of these is described elsewhere in this chapter, a multitude of other more obscure complications have been documented in isolated case reports. Some of these complications include volvulus [26], bowel obstruction [27], inguinal hernia or hydrocele [28], pleural effusion/hydrothorax [29], or migration and erosion into other organs such as the bladder [30] and heart [31].

Procedure

Our approach has been described in detail previously [32], and is illustrated in Figure 16.2. The patient is placed in slight Trendelenburg position to minimize the risk of air embolism. The optimal location of the incision is determined using ultrasound to visualize the carotid artery medially and the internal jugular vein laterally prior to preparation of the skin and draping. A 1 cm incision is then made over the vein using a 10 blade scalpel and the Bovie in coagulator mode through the skin and underlying soft tissues. Under ultrasound guidance, an introducer needle attached to a syringe is then passed, medial to the sternocleidomastoid muscle edge and pointing slightly laterally, into the vein while gently aspirating venous blood. The syringe is then removed and the straightened “J” tip of guidewire is then inserted into the needle and advanced until some electrocardiographic changes are noted, at which point it is pulled back slightly. The needle is then removed and the guidewire is anchored. At this point, the distal catheter is tunneled to the proximal incision and connected to the proximal valve system prior to proceeding, with confirmation of distal CSF flow obtained.

Figure 16.2 Ventriculoatrial shunt insertion. The distal catheter is introduced through the internal jugular vein (IJV), with a goal termination at the cavoatrial junction. SVC, superior vena cava.

Next, the vessel dilator and sheath introducer are passed together over the guidewire using a rotational motion. The length of the distal catheter is calculated as described previously [33], using a preoperative chest X-ray. The guidewire and vessel dilator are then removed from the peel-away introducer, and the atrial catheter is then passed through the peel-away introducer into the superior vena cava. At this point, the peel-away introducer is stripped away and gentle pressure is applied over the neck incision to stop the venous oozing. The incision is then closed as described previously.

Discussion and complications

The main complications occur from the development of a fibrous capsule and eventually occlusion of the distal catheter or vessel thrombosis if the catheter length is too short [34], or risk of intra-atrial clot formation and subsequent shunt occlusion, pulmonary embolism, or congestive heart failure if the catheter is too long. As such, we have adopted our approach as outlined above. Using this approach with preoperative chest X-ray measurement, 14 patients in the series had catheter placement within 0.4 cm of the cavoatrial junction, and no patients required distal revision for a malpositioned catheter tip [34]. Other groups have used intraoperative transesophageal echocardiography to confirm distal catheter location with similarly positive results [35], although the patients are subjected to a more invasive echocardiogram with inherent risks of its own. Typically, duplex ultrasound of the internal jugular vein is obtained preoperatively to ensure patency for passage of the distal catheter; however, if difficult placement is encountered intraoperatively due to anatomical variations, thrombosis, or large venous valves, we have previously used intraoperative venography to navigate around the obstruction [36]. There is an additional risk of pneumothorax or carotid artery cannulation, though this risk is mitigated with a postoperative chest X-ray and the use of ultrasound guidance during venipuncture.

Procedure

Our approach is demonstrated in Figure 16.3. A curvilinear incision is made laterally to the breast, centered on the anterior axillary line and following along the course of the underlying rib (often the seventh). The skin and subcutaneous layers are incised. The most anterolateral attachment of the fascia (of the serratus anterior muscle) is then incised, and the attachments of the external intercostal muscle, intermedius intercostal muscle and internal intercostal muscle are separated from the upper margin of the rib using the Bovie knife. The parietal pleura is then visualized and the lung can be observed moving underneath it with each respiratory cycle. A tunneler is then used to pass the distal catheter from the proximal incision to the chest incision. The distal end of the new catheter is inserted into the pleural cavity, with closure as described above.

Figure 16.3 Ventriculopleural shunt insertion. The distal catheter is introduced above the sixth rib in order to avoid the intercostal vessels and nerve.

Discussion and complications

Tension hydrothorax is the most severe complication following placement of a pleural distal catheter [37]. Following a shunt malfunction or decreased drainage of CSF from the pleural space, fluid accumulation can occur generally subacutely, leading to tachycardia, hypotension, desaturation, and eventually cardiopulmonary collapse if not recognized. Treatment is with emergent chest tube placement to decompress the fluid, and often shunt externalization and distal revision once the patient is stabilized. As such, a patient with a pleural distal catheter with any concern for malfunction should have an urgent chest X-ray.

Procedure

Our approach is illustrated in Figure 16.4. The patient is positioned to minimize the risk of injury during surgery on a bean bag in a lateral position, as is done to perform a spinal tap. The first incision is made in the midline in the lumbar region, and the fascia is exposed. A special Touhy needle is then advanced between two adjacent spinous processes into the subarachnoid space, until good return of CSF is obtained. After removal of the stylet, the proximal catheter is advanced in the subarachnoid space for at least 5 cm and the Touhy needle is withdrawn. Next, the catheter is tunneled to a separate incision in the flank, where a special pocket is created in the subcutaneous tissue to accept the valve. A third incision is made in the abdomen for the distal shunt. The proximal and distal catheters are then connected to the valve, and a silk tie is used to secure the connections. The catheter can then be tunneled to its intended destination, with wound closure as described above.

Figure 16.4 Lumboperitoneal shunt insertion. The proximal catheter is often introduced at the L3–4 level, although adjacent levels are also used commonly.

Discussion and complications

Lumbar shunting holds some advantages over ventricular catheterization, including a decreased risk of parenchymal injury as no catheter is passed through the brain. Additionally, since drainage occurs through the lumbar cistern for communicating hydrocephalus, drainage occurs symmetrically throughout the entire ventricular system without the risk of unilateral ventricular collapse leading to shunt malfunction that can be seen with ventricular catheters. Furthermore, there is some evidence that lumbar shunts are associated with a lower rate of infection than ventricular shunts [39]. Finally, lumbar shunts are often easier to place than ventricular shunts, particularly in PTC patients who typically have small slit ventricles. Additionally, a lower rate of subdural hematomas and hygromas has been noted with lumbar catheter placement compared to some of the series examining ventricular catheter placement [40,41]. Finally, lumbar shunts are believed to have a lower risk of siphoning as the distal catheter is typically at a similar height as the proximal catheter.

However, lumbar catheter placement is not without complications. Catheter obstruction is the predominant complication that has been noted with lumbar shunt placement, and was the cause of 65% of the revisions in one series [40]. Additionally, lumbar shunting may lead to overdrainage causing intracranial hypotension. Overdrainage was responsible for 15% of the revisions in a series [40], although this could be alleviated somewhat with the use of adjustable valves. This overdrainage has even led to an increased risk of Chiari malformations in the pediatric population [42,43], resulting in a 4% rate of suboccipital decompression for symptomatic Chiari. However, this complication does not seem to be as significant in the adult population [40,41]. Furthermore, lumbar radiculopathy and back pain has been noted in patients with lumbar shunts, and was responsible for 4.5% of the revisions in a series [40]. Also, although lumbar catheterization is generally less invasive than ventricular catheterization, patients with lumbar stenosis often require a partial laminectomy in order to pass the catheter in the subarachnoid space. There is an additional risk of spinal cord injury or extra-axial hematoma formation, though these complications are rarely seen with proper technique. Another major drawback of lumbar shunting is that the shunts are often inserted without a reservoir, or the reservoir is difficult to access given the patient’s body habitus; thus, it is much more difficult to assess shunt function with a shunt tap or radionuclide shunt patency study, and opening pressure often must be assessed with a lumbar puncture. This is especially problematic in PTC patients who frequently present with worsening headaches and concern for shunt malfunction. Finally, in a series of 76 patients who underwent lumbar shunt placement, a complication of lumbar shunt placement with a horizontal-vertical valve included malposition of the valve, as it requires the components to be in a horizontal and vertical orientation for proper functioning [41]. This complication was addressed by suturing the valve in place, although this method is not always foolproof especially in PTC patients who tend to have a larger body habitus.

There is not a consensus on which type of shunt to use for communicating hydrocephalus, although many groups advocate the use of lumbar shunting for PTC, particularly given their typically difficult-to-access slit ventricular system. Although we previously demonstrated that ventricular catheterization could be reliably performed under image guidance for PTC patients [14], we have progressed to generally using lumbar shunts in this patient population due to the greater ease of this procedure and minimization of siphoning effects. Tarnaris et al. [44] compared the outcomes of ventricular versus lumbar catheterization in PTC patients, and found similar outcomes in visual symptoms and headache improvement between the two locations, with no significant difference in the length of time of catheter patency. However, Abubaker et al. [45] found a higher revision rate of lumbar shunts compared to ventricular shunts. Similarly, McGirt et al. [46] showed that lumbar shunts had a 2.5-fold increase in shunt revision and 3-fold increase in shunt obstruction compared to ventricular shunts. Based on the conflicting data, it is difficult to make a final recommendation regarding the ideal shunt type for PTC, though if ventricular catheterization is selected, it should always be performed under image guidance.