External ventricular drainage: indications and outcome among Sudanese Children

Abstract

This study reflects our experience in managing Sudanese children with different cranial conditions through external ventricular drainage (EVD): indications for EVD, pathologies faced, and early outcome. A prospective review of cases operated at the National Center for Neurological Sciences was carried out during the period from February 2014 to February 2016. The patients were closely followed up till EVD removal and discharge. Thirty-five Sudanese children were included in the study (age range between 6 days and 7 years). Majority of the cases had posterior fossa tumor with obstructive hydrocephalus (n = 19, 54.3%). Twenty (57.1%) patients presented with a decreased level of consciousness, while 28 (80%) patients presented with symptoms and signs of raised intracranial pressure (ICP). The decision for EVD was made preoperatively based on positive cranial computed tomography/magnetic resonance imaging findings in 10 (28.6%) patients. Additionally, 28 (80%) patients responded to single injectable antibiotic therapy with an average duration of 22 days. Subsequently, 25 (71.4%) patients improved or got cured, 5 deteriorated, and 11 died. We conclude that EVD can be used for many indications, including obstructive, postinfectious, and postmeningitic hydrocephalus as well as intraventricular hemorrhage. Most patients may present with either deteriorating levels of consciousness or symptoms and signs of raised ICP, but few of them may have positive brain imaging findings and therefore the decision for EVD was made intraoperatively. The average duration for EVD use was 3 weeks with single antibiotic therapy use, which was found as effective as when combined with intraventricular therapy.

INTRODUCTION

External ventricular drainage (EVD) is a common neurosurgical procedure that can be performed as a bedside procedure in the intensive care unit (ICU) or inside the operative room. The optimal setting for EVD insertion in regard to safety and accuracy of placement is still poorly defined. External ventricular drainage was first documented by Claude-Nicolas Le Cat (1700-1768) in October 1744 [1]. He performed a ventricular puncture and left a wick in place for some time [2].

Keen [3] published the next report on ventricular drainage in his book ‘Surgery of the Lateral Ventricles of the Brain’. External drainage of the ventricles was also recommended by Keen for hemorrhage and abscess treatment. The place of the entry point to the ventricular system cavity, which was advocated by Keen, bears his name and it is still commonly used nowadays: 3 cm superior and posterior to the pinna, ‘Keen’s point’. He later listed 12 conclusions on the external ventricular drainage procedure, 2 of which remain very applicable in our daily practice [3-5]. They are as follows:

1. Tapping the ventricles to relieve the hydrocephalus is a simple procedure that does not represent a danger for patients,

2. In acute hydrocephalus, this procedure may save the patient’s life, and in chronic hydrocephalus, slow ventricular tapping is successful in some cases and is worth trying [3].

The next major significant advancement in the EVD technique was introduced by Tillmanns in 1908. He reported a new technique at that time, which closely resembled that of our current practice [5]. This new technique included the use of local anesthesia and both anterior and posterior approaches to the ventricular cavity. His description was very close to that of modern neurosurgical practices. The use of subcutaneous tunneling is the only major difference between modern practice and Tillmanns’ method. The former was introduced in 1979 by Saunders and Lyons [6,7].

The introduction of a flexible silastic catheter in 1969 facilitated subcutaneous tunneling. Saunders and Lyons described and reported 174 ventriculostomy procedures performed using silastic catheters and subcutaneous tunneling. They found that the infection rate was as low as 2% in comparison to the other traditional techniques used at that time, despite the prolonged period of EVD use [6].

Fedor Krause (1857-1937) was one of the pioneering German neurosurgeons who performed EVD in patients with hydrocephalus for as long as 8 weeks without an infection. He developed his new technique for perioperative use of EVD in posterior fossa surgery, which made a major contribution in terms of reducing the overall mortality rates in posterior fossa surgeries [8].

The 1940s brought on an era of small changes in the materials and EVD drainage systems that aided the ease of use, stability, and control of EVDs. Ingraham and Campbell [9], in 1941, published a paper in the Annals of Surgery describing the first closed external drainage system. This system improved sterility compared with catheterization, cerebrospinal (CSF) egress into the patient’s bed, and their device included a stopcock system to offer slow, controlled drainage of the CSF, avoiding rapid pressure fluctuations and over drainage. Ingraham’s [9,10] system overcame the major causes of death reported in the former techniques, but they advised against long-term drainage for more than 72 hours.

In pediatric brain tumors, mainly the infratentorial lesions, external ventricular drainage may be used for the temporary relief of raised intracranial pressure without the need for permanent cerebrospinal fluid diversion. In a review regarding the management of posterior fossa tumors and hydrocephalus in children, it was found that 10%-40% of the cases demonstrated persistent hydrocephalus after posterior fossa tumor resection. It appears that young age, moderate-to-severe hydrocephalus, transependymal edema, the presence of cerebral metastases, and tumor pathology (medulloblastoma and ependymoma) on presentation can predict the possibility of postresection-persistent hydrocephalus [10].

The complications arising from EVDs are many, but the major complications reported in the literature include displacement, slippage, blockage, hemorrhage, and infection. The latter may be complicated by ventriculitis, meningitis, brain abscess, or subdural empyemas, which are unfortunately deadly complications [11].

External ventricular drainage-related infection is a serious complication occurring in 5%-23% of the patients, according to the literature. Preventive measures for this infection include prolonged prophylactic systemic antibiotics, and the use of antibiotic-coated external ventricular drains which are also known as impregnated ventricular drains [12].

According to a study published by Kitchen et al. [13], in 2011, Infection was the primary EVD complication ranging between 0% and 45%, and this was associated with significant morbidity and mortality, prolonged hospital stay, and increased financial burden on the patient’s family and on the hospital. However, it was found that prophylactic antibiotics use alone did not significantly reduce the rate of ventriculitis in patients with EVD [14].

Gram-negative microorganisms were the most frequently reported causative agents of infection. Besides that, the length of time the drainage catheter was kept in place is directly proportional to the infection rate [15]. In addition to that, patients with EVD-related infection had a significantly longer intensive care unit (ICU) and hospital stay than patients without.

Drake et al. [16] stated that EVD output decreases with Gram-negative or multiple-organism infections and with the elevation of the drip chamber; whereas resolution of the infection, sex of the patient, and method of establishing the EVD had no effect on the output. Drake et al.’s [16] work predicted that CSF production increases in the human growing brain. Patients with an EVD-associated infection had significantly more often concurrent healthcare-associated infections, which might be a consequence of the prolonged length of hospital stay, especially in the intensive care unit. Wong et al. [17] postulated that routinely changing external ventricular drainage catheters at 5-day intervals did not reduce the risk of CSF infection. An increased risk of infection has been observed in patients with subarachnoid or intraventricular hemorrhage, concurrent systemic infections, longer duration of CSF leakage, and frequent manipulation of the EVD system [18]. The current theory suggested that EVD-related infections result from either introduction of pathogens during EVD placement and/or contamination and colonization of the EVD system during the postoperative period [19].

A study published by Hader and Steinbok [20] aimed to detect the value of cerebrospinal fluid cultures among patients with external ventricular drains in a tertiary care pediatric center over a 13-year period including 157 patients. The results of the study suggested that routine CSF culture in children with EVD is not necessary, and it should be preserved for patients with new onset fever (>38.5°C), peripheral leukocytosis, neurological deterioration, or with changes in the CSF’s gross appearance. In situations where these clinical indicators might be masked, routine cultures may be valuable [20].

According to a large retrospective cohort study conducted by Simon et al. [21], surgical approaches to the treatment of shunt infection included shunt removal or new shunt placement (59.2%), externalization (12.2%), nonsurgical management (13.3%), and removal with no shunt replacement (15.3%). It has been noticed that shunt removal, new shunt placement, and externalization were all not associated with re-infection, while nonsurgical management demonstrated surprisingly higher re-infection rates [21].

EVD is also used for the initial treatment of posthemorrhagic hydrocephalus. Rhodes et al. [22] used the EVD in 37 infants with posthemorrhagic hydrocephalus and noticed that the worst outcomes were in those with parenchymal or large intraventricular hemorrhages.

According to a study conducted by Maniker et al. [23], hemorrhagic complications of EVD placement are more common than what was previously thought and catheter gauge has a direct effect on the hematoma volume. Most of the hemorrhages seen on postinsertion computed tomography (CT) scans did not cause detectable changes in the clinical examination [23].

MATERIALS AND METHODS

The present study is a prospective review of patients who have been operated at the National Center for Neurological Sciences during the period from February 2014 to February 2016. The patients were closely followed up till EVD removal and discharge. All patients with deficient operative or postoperative data or those operated outside the center or adult patients who underwent EVD operation were excluded from the study. Data were collected, analyzed, and interpreted by the authors in a designed SPSS sheet.

RESULTS

From February 2014 to February 2016, 35 patients with different conditions were operated upon with EVD. The youngest patient was 6 days old and the oldest was 7 years old, with average presentation at 2 months. Most of the patients were less than 3 years old (n = 20, 57.1%) and the majority of them in our series were male (n = 22, 62.9%). Twenty-five cases (71.4%) were were operated upon were from Khartoum, the center and capital of Sudan.

Eighteen patients (51.4%) were fully conscious before placing the EVD (Table 1). Almost one-third of the patients were comatose at the time of presentation, i.e., Glasgow Coma Scale (GCS) 8 or below (n = 10, 28.6%). The second third presented with irritability with or without fever and poor oral intake (n = 13, 37.1%). The last third presented with different conditions, such as convulsions, bulging fontanelle, and an increase in the head size. Only one-quarter of the patients in our series were associated with myelomeningocele (n = 6, 17.1%).

Table 1.

The preoperative level of consciousness of the patients.

Preoperative GCS	Frequency	Percent
15	18	51.4
14	5	14.3
8	3	8.6
4	2	5.7
6	2	5.7
7	2	5.7
13	1	2.9
11	1	2.9
5	1	2.9
Total	35	100

GCS, Glasgow Coma Scale.

All patients in this study were investigated using either cranial CT with contrast (n = 32 91.4, %) or brain magnetic resonance imaging with contrast (n = 3, 8.6%). The imaging was suggestive of infected CSF in only one-quarter of the patients (n = 8/35, 22.9%) in a form of enhanced ventricular wall (n = 6/8, 75%) or brain abscess in close vicinity to the ventricular wall (n = 1/8, 12.5%), or in a form of intraventricular fluid level (n = 1/8, 12.5%).

The primary reason for EVD insertion was for posterior fossa tumor with obstructive hydrocephalus (n = 19, 54.3%), followed by intraventricular hemorrhage (n = 10, 28.6%) and postmeningitic hydrocephalus (n = 6/35, 17.1%). Most of those with postmeningitic hydrocephalus had an infected ventriculoperitoneal (VP) shunt (n = 3/6, 50%), while only one of them had a ruptured brain abscess into the ventricle (n = 1/6, 16.7%). The remaining two had an infection of unknown source (n = 2/6, 33.3%). Some patients with postmeningitic hydrocephalus were also found to have hemorrhagic CSF (n = 2/6, 33.3%).

Considering ventricular tapping as one of the risk factors for having CSF infection, we found that 10 patients (n = 10, 28.6%) had a history of ventricular tapping, but it was done once in all of them. Only 11 patients (31.4%) had history of neurosurgical operations. Ten of them had VP shunt insertion (n = 10/11, 90.9%), while the remaining one had an aspiration of a cerebral brain abscess (n = 1/11, 9.1%).

EVD insertion points were selected to be either at Kocher’s point (n = 16/35, 45.7%) or Keen’s point (n = 19/35, 54.3%), with approximately equal number of patients for each. Average recordings for CSF pressure readings during the hospital stay were obtained. In 19 patients (54.3%) with infectious etiology, the readings ranged between 5 and 7 mmHg, followed by 7-10 mmHg in 14 patients (40%). Additionally, in 2 patients (5.7%) the readings were >10 mmHg as they had massive intraventricular hemorrhage (Table 2).

Table 2.

The average cerebrospinal fluid pressure readings in the external ventricular drainage during hospital stay.

EVD pressure readings	Frequency	Percent
5-7 mmHg	19	54.3
7-10 mmHg	14	40
More than 10 mmHg	2	5.7
Total	35	100

CSF, cerebrospinal fluid; EVD, external ventricular drainage.

CSF routine analysis was carried out in all cases, but the results were inconclusive and are therefore not displayed in this section. CSF culture results were positive in only 3 patients (8.6%), while no organisms were isolated in most of the cases (Table 3).

Table 3.

The types of microorganisms isolated from CSF cultures of the patients.

Types of isolated organisms	Frequency	Percent
None	32	91.4
Klebsiella	1	2.9
Pseudomonas aeruginosa	1	2.9
Staphylococcus aureus	1	2.9
Total	35	100

Two-thirds of the patients were admitted postoperatively to the ward (n = 22/35, 62.9%), while one-third were admitted to the ICU. Monotherapy, multitherapies, and intraventricular antibiotics administration protocols were all tried. However, the majority of patients (n = 26/35, 74.3%) were treated with single antibiotic regimen (Table 4).

Table 4.

Antibiotics used postoperatively.

Types of antibiotics used	Frequency	Percent
Ceftriaxone	22	62.8
Ceftriaxone + vancomycin	5	14.2
Ceftazidime	3	8.5
Ceftazidime + vancomycin	1	2.9
Ceftriaxone and intraventricular vancomycin	1	2.9
Ceftriaxone, gentamicin, and metronidazole	1	2.9
Ceftriaxone + gentamicin	1	2.9
Vancomycin	1	2.9
Total	35	100

The shortest duration of hospital stay (Table 5) was 2 days (n = 2, 5.7%) and both patients died, while the longest duration of hospital stay was 61 days with the average duration of 20 days (1-3 weeks).

Table 5.

The duration of hospital stay in weeks.

Duration of hospital stay	Frequency	Percent
Less than 1 week	6	17.1
1–3 weeks	19	54.3
4–6 weeks	8	22.9
More than 6 weeks	2	5.7
Total	35	100

Complications of EVD were encountered in only nine patients (25.7%). These included five cases with overdrainage, three cases with catheter blockage due to thick material, and one case with slippage of the EVD ventricular catheter. Some patients had an exposed VP shunt that required shunt removal and EVD insertion until the infection completely cured (Figure 1).

Figure 1.

Exposed ventriculoperitoneal shunt that required removal and external ventricular drainage insertion.

More than half of the patients (n = 20/35, 57.1%) improved or were cured postoperatively. Many clinical parameters were used to assess patient improvement (Table 6). Sepsis and infection sequels were the main causes for patients’ deterioration and deaths (Table 7). Ten patients died; five secondary to cardiac arrest; four secondary to sepsis; and one secondary to respiratory arrest due to severe respiratory acidosis.

Table 6.

The improvement parameters used to assess the patients postoperatively.

Improvement parameters	Frequency	Percent
No irritability + good oral intake	6	30
Ventriculoperitoneal shunt inserted	4	20
No irritability	3	15
Increased level of consciousness	3	15
Clear CSF	2	10
Convulsions stopped	1	5
ETV done	1	5
Headache resolved	1	5

CSF, cerebrospinal fluid; EVD, external ventricular drainage.

^a Percentage was calculated out of 20 (improved + cured), and some patients may have more than one improvement parameter.

Table 7.

Deterioration parameters encountered postoperatively.

Deterioration parameters	Frequency	Percent
Abscess formation	1	20
Low GCS with ongoing infections	1	20
Low level of consciousness with no improvement	1	20
No improvement with uncontrolled convulsions	1	20
Sepsis and GCS becoming 13/15	1	20
Total	5	100

GCS, Glasgow Coma Scale.

Illustrative cases from our study are shown in Figures 1-4. Figure 5 shows abnormal images findings in all eight patients.

Figure 4.

A 2-month-old male baby with dysmorphic features of the face (A) presented with seizures and sudden deterioration in the level of consciousness. (B) Cranial computed tomography (CT) revealed dilated left lateral ventricle with intraventricular hemorrhage (C) and right chronic subdural hematoma with banana-shaped brain mantle compressed in between near to the falx. Follow-up CT brain performed thereafter (D) showed complete resolution of the intraventricular hemorrhage, expansion of the brain tissue, and decrease in the hematoma size.

Figure 5.

The findings of eight patients with abnormal cranial computed tomography.

DISCUSSION

The use of EVD systems in children has many indications. It can be used as preoperative management to control raised intracranial pressure in children with posterior fossa tumors, hence reducing the perioperative mortality associated with these lesions [24]. Besides that, the use of a ventriculoperotineal (VP) shunt is not usually the best option as a large number of these children might not require VP shunt insertion after tumor removal [24,25].

Infantile intraventricular hemorrhage (IVH) is another indication for EVD insertion. The earlier the intervention, the better the outcome in terms of cognitive function and developmental milestones. In this study, it has been noticed that EVD opening pressure in patients with IVH is usually high and EVD insertion may be complicated with overdrainage if not closely monitored in the ICU.

One of the commonest indications for EVD is the management of an infected VP shunt where removal of the shunt device is indicated, followed up with the insertion of an EVD as well as administrating intraventricular and systemic antibiotics [26]. This is found to be a safe and effective way of managing children with shunt infection [27]. Interestingly, in our study we found that the children, who had their EVD inserted due to infectious etiology, mostly had an EVD for managing infected VP shunts. Therefore, EVD is considered a useful management option for both early and late shunt infections.

Figure 2.

A 6-month-old male baby (A) with postmeningitic hydrocephalus revealed by (B) cranial CT underwent ventriculoperitoneal shunt operation. Few months later, he presented with features of shunt dysfunction, high grade fever, and opisthotonos posture (C). Follow-up cranial CT revealed grossly dilated ventricles with enhancement of the ventricular wall (D). External ventricular drainage (EVD) was inserted and cerebrospinal fluid was thick, pussy, and yellow (E). The patient completed 4 weeks of intravenous antibiotics (F) and follow-up CT brain was done just before removal of the EVD (G). The CT showed resolution of the hydrocephalus, features of healed infection (calcification on the ventricular wall) but with toxic damage to the brain mantle from the previous infection (G). The general condition of the patient improved and the infection completely resolved.

It worth mentioning that CSF production rate is crucial for adjusting the CSF pressure in the EVD. However, it must be considered that the CSF production rate increases with the increase in patient age and weight and decreases in cases of Gram-negative infection or multiple-organism infections. We measured the average readings of CSF pressure in pediatric patients with EVD and found that all patients with infectious etiology had normal CSF pressure readings in contrast to patients with hemorrhagic etiology who had higher readings, which is consistent with the abovementioned theory.

Figure 3.

A 6-month-old female baby presented with features of meningitis (fever, arched back, and nuchal rigidity) with active hydrocephalus (A). When external ventricular drain was inserted, it was hemorrhagic (B), although the preoperative cranial CT was not suggestive of that (C).

The reported complications of EVD include infection, hemorrhage, increase intracranial pressure, and brain injury. All of these complications can be prevented and easily managed with good nursing care [28]. In the present study, we found that the complications were less common in patients monitored in the ICU in comparison to patients monitored in the ward.

Repetitive CSF cultures from the EVD may not be useful and should be only done for selective indications, including fever >38.5°C, peripheral leukocytosis, change in the CSF color, and neurologic deterioration [29,30]. We did not find CSF cultures useful as most of the patients had negative cultures. Additionally, brain CT, even with contrast, was not found as useful as MRI and the drawback in this study is that most of the patients were investigated preoperatively using brain CT with contrast.

CONCLUSION

EVD can be used for many indications including obstructive, postinfectious, and postmeningitic hydrocephalus, as well as IVH.

Most patients may present with either deteriorated levels of consciousness, fever, shunt dysfunction, or symptoms and signs of raised intracranial pressure; but few of them may have positive brain imaging findings, especially with brain CT, and therefore the decision for EVD may be made intraoperatively.

The average duration for EVD use in this study was 3 weeks with single antibiotic therapy use which was found to be as effective as that combined with intraventricular therapy.

Patients with EVD are preferably nursed in the ICU to avoid the device-related complications, such as slippage, overdrainage, and other medical non-neurosurgical complications.

FUNDING

None.

CONFLICT OF INTEREST

The authors declare that there is no conflict of interest regarding the publication of this article.

ETHICAL APPROVAL

Ethical approval for conducting this study was obtained from the National Center for Neurological Sciences, Khartoum, Sudan. All parents of the patients included in this study had consented to share in this study verbally and orally after explaining the objectives of the study. Data confidentiality was maintained using the known indexing methods.