Chapter 22: External Ventricular Drain Management and Ventriculoperitoneal Shunt

This patient's presentation is consistent with Hunt and Hess (HH) grade III and Fisher grade 3 SAH (see Chapter 1 for information on SAH grading). Radiographic evidence of acute hydrocephalus along with neurologic decline (failure to follow commands) call for emergent placement of an external ventricular drain (EVD) to alleviate intracranial hypertension. EVDs serve three primary functions in SAH: to monitor intracranial pressure (ICP), to drain cerebrospinal fluid (CSF) for treatment of hydrocephalus, and/or to acutely reduce ICP. EVD placement is therefore indicated when a patient is thought to have symptomatic hydrocephalus and/or elevated ICP, based on neurologic examination and radiographic findings.

Ventriculostomy is considered standard-of-care for treating SAH-associated hydrocephalus and has been shown to improve both short- and long-term outcomes.¹ However, there is no standard, evidence-based guideline for EVD placement in SAH patients. Although the Glasgow Coma Scale (GCS) score (eg, 12 or less)² and HH grade (eg, III or greater)³ have been used to establish an objective threshold for ventriculostomy, the procedure should be generally considered in patients who demonstrate clinical or radiologic deterioration or have an unreliable neurologic examination. Patients who present comatose or severely lethargic are almost universally considered for emergent ventriculostomy.⁴ At our institution, all SAH patients not following commands are strong candidates for EVD insertion (Figure 22-2). Minimal improvement in neurologic status despite normalization of the ICP in these patients may point to other etiologies such as seizure, medication effect, or metabolic derangement and prompts immediate investigation. As for those with fluctuating levels of consciousness, the impact of ventriculostomy on outcome remains unclear and, therefore, careful risk-benefit analysis is warranted.² Figure 22-3 outlines our ICP management protocol for patients admitted to our NICU with aneurysmal SAH (aSAH).

Similarly, EVD placement is recommended in a variety of other neurologic conditions causing symptomatic acute hydrocephalus related to either noncommunicating or communicating hydrocephalus or evidence of elevated ICP.¹ Common causes of obstructive or noncommunicating hydrocephalus include posterior fossa tumors, intraventricular hemorrhage, and intraventricular cysts or tumors; communicating hydrocephalus may be caused by SAH, intraventricular hemorrhage, and meningitis; elevated ICP without significant hydrocephalus is commonly caused by closed-head trauma.⁵ Normalization of ICP with EVDs can prevent secondary cortical injury, improve neurologic status, and enhance operative exposure during surgery.^6-9 Of note, normal ICP ranges for adults and young children are considered less than 10 to 15 and 3 to 7, respectively.⁵

Another indication for EVD placement is to aid wound healing in the postoperative neurosurgical patient. Those who undergo operations of the posterior fossa, for example, via a suboccipital or a far lateral approach, are at particularly high risk for a postoperative CSF leak through their incision. Therefore, the neurosurgeon often leaves an EVD in the patient for a few days postoperatively, which allows for drainage of CSF through the EVD and alleviation of CSF pressure on the fresh incision. Once the incision has begun to heal and there is no evidence of CSF leak, the EVD can be removed.⁵

The EVD is placed by making a frontal incision and single burr hole in the skull, so that the catheter can be passed through the frontal lobe and into the lateral ventricle, with the goal of placing the catheter tip at the foramen of Monro. The nondominant frontal lobe is preferred as an entry point because it minimizes the risk of significant neurologic deficit in the event of a procedural complication such as subdural, epidural, or intraparenchymal hemorrhage.⁵ The burr hole entry point should be lateral to midline (in the mid-pupillary line) to avoid the sagittal sinus and its tributaries, and anterior to the coronal suture in order to avoid the motor strip, which is usually located 4 to 5 cm behind the coronal suture. This well-known entry point is called the Kocher point, and is typically located 12 to 13 cm posterior to the nasion in the anteroposterior (A-P) direction, and 3 cm lateral to midline (Figure 22-4). These measurements should always be correlated with anatomic landmarks by palpating the coronal suture through the scalp and ensuring the entry point is anterior, as well as visually ensuring the entry point lies in the mid-pupillary line. Although a right-sided (nondominant frontal lobe) placement is preferred, the EVD may be placed through the left side if the surgeon needs to avoid a lesion or hemorrhage in the right frontal lobe or lateral ventricle. In trauma patients, the surgeon may prefer to pass the EVD through whichever frontal lobe has already been injured in order to minimize risk to the remaining normal parenchyma. When the Kocher point is used as an entry point, the trajectory of the EVD pass through the parenchyma should be aimed at the medial canthus in a medial-lateral direction and the tragus in an A-P direction. This trajectory is usually achieved if the EVD is aimed exactly perpendicular to the skull surrounding the burr hole. Depending on the size of the lateral ventricles, the surgeon may pass the catheter 4 to 5 cm into the parenchyma before feeling a "pop" sound indicating entry into the lateral ventricle, with CSF return through the catheter. In order to advance the catheter tip to the foramen of Monro, the EVD is usually "soft passed" (advancing the catheter without the stylet) to a total of 6.5 to 7.0 cm from the outer table of the skull. The catheter is then tunneled a short distance under the scalp, in a direction that avoids a potential future shunt pathway.

The rate of EVD catheter misplacement can be as high as 60%, morbidity rate as high as 28%, and hemorrhagic complications 17.5%.^8,10 Therefore, a routine postplacement CT scan is recommended to ensure correct placement and rule out hemorrhagic complications.

EVDs can be placed either in the operating room (OR) or at the bedside using local anesthesia.^11,12 The OR provides a more sterile, controlled environment where acute complications can be efficiently managed. However, waiting for an OR room may delay treatment when emergent EVD is needed, and transporting critically ill patients leaves physicians without resources during transport. Although EVD placement in an ICU setting may increase the risk of severe infection, there are no conclusive data to determine how the environment of EVD placement affects the overall complication rate and outcome.^11,13

EVDs have traditionally been placed by neurosurgeons to ensure that the person who places the device can also manage procedure-associated complications, such as subdural or intracerebral hemorrhage. Recently, an increasing number of EVDs and ICP monitors have been placed by nonneurosurgeons, including neurointensivists, nurse practitioners, physician assistants, trauma surgeons, and general surgery residents.¹⁴ The impact of EVD placement by nonneurosurgeons on procedural success, complication rates, and outcomes is yet to be determined. Nevertheless, the increasing trend toward nonneurosurgeon EVD and ICP monitor placement, if proven equally safe and effective, may provide patients with more expeditious care in select environments.

Although an EVD is thought to offer the most accurate ICP measurements, the surgeon may be unable to place an EVD in patients with collapsed ventricles, slit ventricles, or significant mass effect.⁵ In such cases, ICP can be monitored via placement of a fiberoptic parenchymal ICP bolt, which is a small fiberoptic cable passed through a burr hole and into the parenchymal edge, usually at the Kocher point. If an ICP bolt is placed and the team believes intracranial CSF volume is contributing to critically elevated ICP, a spinal drain can be placed in patients with communicating hydrocephalus. Just like an EVD, the spinal drain allows for lowering of ICP and treatment of hydrocephalus through CSF drainage. In effect, a parenchymal ICP bolt in conjunction with a spinal drain can serve the same functions as an EVD.¹⁵ Although intraparenchymal monitors are easily placed and can be disconnected during patient transport without the need for recalibration, they cannot accurately detect pressure alterations in deeper parts of the brain.^15,16 The devices are also associated with mechanical failure, fragility, and monitor malfunction. Table 22-1 describes advantages and disadvantages of ICP-monitoring devices.

EVD placement has been suggested to increase the risk of rebleeding in aSAH patients.^2,17,18 An abrupt drop in ICP and resultant increase in the transmural pressure on the aneurysm are thought to cause the rebleeding.¹⁹ However, elevated risk of rebleeding in aSAH patients after EVD placement remains controversial, and no study to date has conclusively established a causal relationship.^20,21 Of note, patients requiring an EVD tend to present with worse HH grade, which in turn is associated with independent risk factors for rebleeding, such as larger aneurysm and dense SAH. Furthermore, conflicting results may be attributed to the failure to account for the following confounding variables: clinical grade, aneurysm treatment timing, duration of EVD placement during which the aneurysm remained untreated, and the interval between onset of SAH and EVD placement.^21,22 In the absence of conclusive data, the presence of an aneurysm should not be a contraindication for EVD placement in the setting of SAH-associated hydrocephalus. Nevertheless, ICP should be normalized gradually to minimize the theoretical risk of rebleeding.

After the EVD is placed, the patient's head is elevated 30 degrees to reduce ICP and increase venous return. The drainage bag is fixed to a pole next to the bed, with "0" set exactly at the level of the external auditory meatus, which is approximately at the level of the foramen of Monro and, thus, the intraventricular catheter tip⁵ (Figure 22-6). The drainage port can then be set at a certain level above or below this reference point ("0") with the valve open. Whenever the intraventricular pressure exceeds the level of the open drainage port, CSF should drain. For example, if the drainage port is set at 15 cm H₂O (15 cm above the "0" reference point), CSF should only drain when the intraventricular pressure exceeds 15 cm H₂O. The level of the drain can be titrated to the desired spillover pressure.

At our institution, we set the EVD open at 20 cm H₂O immediately after placement in SAH patients whose aneurysms remain unprotected to prevent overdrainage and the aforementioned theoretical creation of a transmural pressure gradient across the aneurysm wall. Postoperatively, once the aneurysm has been protected with clipping or coiling, we may lower the EVD to 5, 10, or 15 cm H₂O, depending on the patient's clinical status and degree of radiographic hydrocephalus. The lower the EVD is set, the more CSF is drained. An alternative approach is to drain off a fixed amount of CSF per unit of time up to a maximum of 20 mL per hour, which is roughly the rate of normal CSF production.⁵ Overdrainage should be avoided as it can lead to acute subdural hemorrhage or even theoretical herniation.

In the setting of a posterior fossa tumor, EVDs are often placed intra-operatively.⁵ CSF can be drained prior to opening the dura to relax the brain away from the dura. During the postoperative period, the EVD is usually set at a low height (eg, 10 mm H₂O) for 24 hours, and is progressively elevated over a 48-hour period and discontinued by 72 hours. Of note, both mm Hg and mm H₂O are commonly used for pressure measurements. Conversion factors between the two units are as following:

Soft tissue infection and ventriculitis are the most common ventriculostomy-related infections (VRI).^23,24 Less common are intraparenchymal abscesses, subdural empyema, and osteomyelitis. Steroid use and duration of catheter placement are known risk factors for EVD-associated infection,²⁵ and there have been no conclusive studies regarding the effects of systemic infection on the risk of neurologic infection.

While surgeons universally give one dose of antibiotic (ie, Ancef [cefazolin] 1 to 2 g) prior to skin incision for EVD placement, as in any operation, prophylactic antibiotic use for the entire duration the EVD is in place remains institution- and physician-dependent. As presented in Table 22-2, data for prophylactic antibiotic use remain conflicting and inconclusive. There is currently an ongoing clinical trial comparing perioperative prophylactic systemic antibiotics (PSA) (ampicillin/sulbactam and ceftriaxone) with EVDs coated with antibiotics versus prolonged PSA with a standard EVD. An interim report including analysis of 110 patients reported no viral respiratory infections.

Once an EVD is placed, it should be checked regularly (eg, 2 to 4 hours) to ensure that it is functioning properly. The system should be assessed for patency, position, and reliability of ICP measurements. Any changes in ICP and neurologic examination also warrant close examination of the system. The drip chamber should be emptied into the drainage bag on a regular basis to keep the chamber from becoming full. Drain position should also be checked to make sure that it is at the prescribed level. Whenever the level of a patient's head is changed in relation to the drip chamber, the stopcock should be closed to drainage until the prescribed level is reestablished. Patients undergoing magnetic resonance imaging (MRI) or CT scan should have the stopcock closed for the duration of the scan because the drainage system generally cannot be set up on a pole at the appropriate level during a scan. Otherwise, if the physician intends for the EVD to be open and continuously draining CSF, it should only be clamped when measuring ICP, obtaining CSF samples, or following administration of intrathecal medications. Efforts should be made to minimize clamping time in EVD-dependent patients. If the physician suspects the monitor does not accurately reflect the ICP, he or she can test for a change in ICP by lowering the head of the bed or applying gentle pressure on the jugular veins bilaterally; either of these procedures should increase ICP.

ICP waveforms reflect transmission of systemic blood pressure to the intracranial compartment.⁵ The waveforms consist of the arterial systolic pressure waves and the central venous A waves, which are, in turn, superimposed on slower respiratory variations.⁵ ICP rises during cardiac systole due to distention of the intracranial arteriolar vasculatures, and during expiration because the concomitant pressure elevation in the superior vena cava reduces venous outflow from the cranium.^5,31

Under normal circumstances, ICP elevation plateaus at a level at which further increase in the intracranial volume is balanced by CSF absorption. However, volume expansion beyond a critical point leads to a rapid and persistent increase in the ICP along with decreasing cerebral compliance.³² In a normal, compliant brain, the first peak (P1 or the percussive wave), which represents the arterial pressure, is taller than the second peak (P2 or the tidal wave) which reflects brain rigidity. As the ICP rises to the level of arterial pressure, the waveforms demonstrate prominent arterial pressure waves, with the tidal wave becoming taller than the percussive wave as the brain loses compliance.^5,33

In 1960, Lundberg described the three ICP wave patterns outlined in Table 22-3. Like ICP measurements, assessment of the waveforms should be evaluated with the stopcock turned off to the drain but open to the pressure transducer.

If the EVD is open to drainage, the nursing staff should monitor and record CSF output every 1 to 2 hours. The amount of CSF drained in a day is typically about 250 mL, and although uncommon, can reach up to 450 to 700 mL if the patient fails to absorb any CSF.⁵

If no CSF has drained, the waveform is dampened or lost, or blockage is suspected, one should consider the algorithm in Figure 22-7.

CSF should be drawn to screen for infection in a febrile patient, to assess CSF protein level prior to ventriculoperitoneal shunt placement, to measure CSF levels of intraventricularly administered antibiotics (every 24 hours), and to check antibiotic response in patients with CSF infection (every 24 hours).

Generally, routine prospective sampling of CSF is not advised as it can increase the risk of infection.²⁵ Routine analysis of CSF may not predict EVD-associated infection, and no specific or reproducible changes in CSF parameters, including leukocytes, pleocytosis, and protein/glucose levels, are definitively predictive of infection.³⁴ Thus, CSF analysis should be performed only when there is a clinical suspicion of infection.³⁵ Figure 22-8 outlines how to sample CSF using EVD.

As for intraventricular antibiotic administration, no randomized controlled trial has been performed to determine its efficacy in treating central nervous system (CNS) device–associated infection. Owing to potential toxicity and difficult delivery, intraventricular administration of antibiotics is generally considered when conventional intravenous therapy has failed to sterilize the CSF, infection is caused by highly resistant organisms that are only sensitive to antibiotics with poor CSF penetration, and the device cannot be removed.^36,37

No antibiotic has been approved by the Food and Drug Administration (FDA) for intraventricular use. The use of vancomycin and gentamicin has been reported more than any other antibiotic in the literature, and others are mostly limited to case reports.^38-40 Cephalosporins and penicillins should not be given intraventricularly because of significant neurotoxicity.^36,41 Intraventricular antibiotic dosing has not been standardized, and dosing has been guided by calculation of the "inhibitory quotient," which can be determined by dividing the trough CSF antibiotic concentration by the minimum inhibitory concentration (MIC) of the drug for the isolate; it should not exceed 10 to 20.⁴²

Administration of intraventricular thrombolysis using tissue plasminogen activator (tPA), for example, is considered in patients with a significant amount of intraventricular hemorrhage (IVH) that will either clot the EVD repeatedly or impair the patient's own CSF outflow for a prolonged period.⁴³ Although unproven by any well-designed study, more rapid resolution of IVH through thrombolytic administration should theoretically reduce the complications associated with prolonged EVD usage in the ICU and aim to reduce the incidence of shunt-requiring hydrocephalus.⁴³ However, theoretical risks associated with intraventricular tPA include increased incidence of infectious ventriculitis and hemorrhage.⁴⁴ General contraindications to IV tPA include unprotected vascular malformations and coagulopathy. A large, randomized controlled study, Clot Lysis: Evaluating Accelerated Resolution of Intraventricular Hemorrhage (CLEAR-IVH), testing intraventricular recombinant tPA versus placebo in IVH, is currently under way.

There is no standard guideline for EVD removal and it remains a controversial topic among neurointensivists and neurosurgeons. EVDs are often discontinued gradually over days toward the end of a patient's ICU stay. However, it is not clear whether gradual removal of an EVD provides any benefit over more rapid discontinuation.⁴⁵ In addition, which subgroup of patients benefits from prolonged EVD placement remains inconclusive. Figure 22-9 shows a sample protocol.

Aneurysmal SAH (aSAH) is the leading cause of chronic hydrocephalus requiring VPS placement in adults, affecting approximately 35% of all cases.⁴⁶ Approximately 10% to 20% of these patients will require permanent CSF diversion with a shunt.^47,48 Obstruction of the arachnoid granulations by blood products, fibrosis of the leptomeninges and arachnoid granulations, adhesion formation within the ventricular system, and resulting alterations of CSF dynamics have all been suggested as possible mechanisms. Factors associated with the development of chronic hydrocephalus following aSAH include old age, high HH grade, female sex, acute hydrocephalus, posterior circulation aneurysm, and aneurysm size greater than 2.5 cm.^46,49,50 A high CSF protein level and large third ventricle diameter at EVD challenge may also predict challenge failure.⁵¹ A number of retrospective studies reported that prolonged or continuous CSF drainage is associated with an increased risk of chronic hydrocephalus.^52,53 However, causality of the association remains inconclusive.

EVD to VPS conversion is generally not performed in the presence of signs of infection (ie, fever), at the peak of vasospasm or elevated transcranial Doppler readings, and on the same day as percutaneous endoscopic gastrostomy (PEG) tube placement. Significantly increased CSF protein (greater than 200) and red blood cell (RBC) counts are thought to increase CSF viscosity and thus the likelihood of shunt blockage and malfunction.^51,52 However, proteinaceous and hemorrhagic CSF are common among SAH, IVH, and trauma patients requiring a VPS, and the decision to convert a temporary EVD to permanent VPS is left to the discretion of the neurointensivist and neurosurgeon.

There are no clear guidelines on technique and placement site when converting an EVD to a permanent shunt in aSAH patients. A new site is routinely established for shunt placement in order to avoid the theoretical risk of infection from the previous incision and EVD. Although further investigation is warranted, there are data to suggest an EVD can be safely converted to a permanent shunt using the existing EVD site, particularly with the use of prophylactic antibiotics and antibiotic-coated catheters.⁵²