Ventriculostomy and Infection: A 4-year-review in a local hospital

Background: To review the complication rate of ventriculostomy-related infection in a local regional hospital, to identify risk factors of infections and suggest measures to prevent infections. Methods: Retrospective review of all cases involving ventriculostomy in one centre of a 4-year-period (Pamela Youde Nethersole Eastern Hospital, a local regional hospital in Hong Kong). 336 cases of admission involving 328 patients with a total of 368 ventriculostomy performed in the centre in a 4-year-period. Main outcome measures include Rate of infection and risk factors related to infections. Results: 10 cases of out 336 cases (2.98%) of ventriculostomy were complicated by infection. Neither the duration of ventriculostomy, revision, urokinase instillation or haemorrhage showed significance in the rate of ventriculostomy-related infection. The low infection rate is compatible with other international literatures that used strict infection control measures. Conclusion: Strict measures for prevention aid in achieving a low complication rate of ventriculostomy related infection.


INTRODUCTION
Indwelling ventriculostomy catheter or External Ventricular Drain (EVD) is a common Neurosurgical procedure. It can be defined as a surgical establishment of a communication between cerebral ventricular system and the external environment (outside of the human body). EVD is integral in the management of patients with a variety of neurosurgical conditions, including elevated intracranial pressure (ICP), a variety of intracranial haemorrhages, intracranial tumour, traumatic head injury, cerebral oedema and ICP monitoring. However, ventriculostomy, as with all surgical procedures, sustains a risk for complications. One of the complications of concern is ventriculostomy related infection (VRI). The reported incidence of ventriculostomy related infection varies from 0% to 22%. [11] A number of risk factors of ventriculostomy related infection had been identified but the significance of these factors varied among various studies. [2,7,8,11,12,14] Various prevention protocols and practice guidelines had also been suggested with an aim to reduce the incidence of VRI. [3] A previous 5-year review of ventriculostomy at the author's centre, the infection rate of EVD is <1%. [1] This is a 4-year retrospective review of all the consecutive cases requiring ventriculostomy at the author's centre (a local hospital in Hong Kong) in which the complication of ventriculostomy related infection addressed.

METHODS
A retrospective review of the records of all the patients with ventriculostomy performed at the Pamela Youde Nethersole Eastern Hospital in Hong Kong SAR during the study period. The study period is from 1 st  Outcome measure includes duration of ventriculostomy, revision rate, CSF or EVD culture positive rate, rate of ventriculostomy related infection, Outcome of cases with ventriculostomy related infection The results were compared with published international studies.

RESULTS
336 cases of admission involving 328 patients with a total of 368 external ventricular drains performed at the author's centre during the period of study.
All ventriculostomy were performed in the operation theatre. The insertions were performed in the operating theatre with the aim to decrease infection rate and improvement of surgical safety.
None the ventriculostomy catheters were antibiotics coated.
Peri-operative prophylactic antibiotics were routinely administered to all patients. For patients with the ventriculostomy procedure within 1 week of admission to hospital, a one-gram intravenous dose of Ceftriaxone was administrated at the start of the operation and every 8 hours for three doses after the operation. For patients with hospital stay of more than 1 week before the operation, peri-operative antibiotic prophylaxis involved routine administration of one-gram intravenous dose of vancomycin and one-gram dose of sulbactam/ cefoperazone once before the operation and every 8 hours for three doses after the operation.
All patients were nursed in the intensive care unit (ICU), high dependent unit (HDU) or Neurosurgical ward after the insertion of the EVD until the ventriculostomy had been removed. Continuous ICP monitoring with an external transducer was deployed for all cases until the ventriculostomy was removed. CSF samplings were only performed when indicated and performed with aseptic technique. The EVDs were removed once it had served its clinical purpose and the removal of catheter involved aseptic technique. The tips of the ventriculostomy catheters were routinely saved for bacterial culture upon removal, with exceptions only with cases of patients who succumbed to medical conditions unrelated with ventriculostomy related infection before the ventriculostomy catheters were removed.

Gender
The study revealed a higher number of male patients (190 cases) compared to female patients (145) with a male to female ratio of 1.3 to 1. [ Figure 1, Table 1.]

Age (years)
The mean age of all the patients was 54.1 years old with a standard deviation (SD) of + 18.01. The mode and median of the all the patients' age were all 55 years old. The oldest patient was 98 years on day of operation. There were 4 patients who underwent EVD   Table 4].

Manipulation
During the period in which the external ventricular drain inserted within the 1 year of birth and the youngest only 1 day old. Most number of patients had the operation during the age range 50-59 years old. A large majority of the total number of patients had the ventriculostomy during the age range of 40-79 years old. [ Figure 2 and 3, Table 2].

Indications of ventriculostomy
The indication of ventriculstomy for a majority of the patients (64.3%, 216 cases) was haemorrhage which included cases of subarachnoid haemorrhage (SAH), intraventricular haemorrhage (IVH), intracerebral hemorrhage (ICH) and cerebellar haemorrhage. The second most common indication was trauma (15.8%, 53 cases) and it includes traumatic fracture of the skull requiring craniotomy or craniectomy, traumatic acute epidural haemorrhage requiring clot evacuation and traumatic head injury requiring ICP monitoring. The indication of tumour(10.1%, 34 cases) involved cases of tumour related obstructive hydrocephalus or operative removal of intracranial tumour with ventriculostomy as post-operative ICP monitoring. Indication of infection represent cases of infected CSF shunt system, tuberculosis meningitis induced hydrocephalus and infantile ventriculitis for intraventricular antibiotics irrigation and CSF sampling. Infarct as an indication (3.9%, 13 cases) involves cases with cerebellar infraction related hydrocephalus and post infarction cerebral oedema with significant massive effect.

Duration of ventriculostomy
The duration of ventriculostomy in this review was regarded as the number of days from the operative insertion of EVD to the day of removal of ventriculostomy. As some cases involved revision of ventriculostomy or more than one ventriculostomy catheter inserted in    was inserted, a number of cases involved manipulation of the external system. Such manipulations were performed under aseptic technique and only performed if clinically indicated. However, the data collected in this study only involved 3 types of manipulations (urokinase instillation, amikicin irrigation and frequent CSF sampling). Urokinase instillation via the ventriculostomy was performed for patients with intraventricular haemorrhage (IVH) after aneurismal rupture and arteriovenous malformation were ruled out by CT cerebral angiogram or Digital subtraction cerebral arteriogram. [10] Amikicin irrigation were used in patients with central nervous system (CNS) infections which include pre-existing meningitis, post traumatic and post-operative CNS infections in the presence of positive bacterial culture in CSF via a ventriculostomy in situ. [15] CSF sampling were performed in cases with fever, raised serum white cell count, impaired mental status, duration of EVD longer than 5 days and manipulations which involved exposing the closed sterile external collection system for drained CSF. Frequent CSF sampling, which we defined as repeated CSF sampling of once every 2 to 3 days for more than 3 occasions, only involved patients with pre-existing meningitis or suspected post operative CNS infections of which frequent CSF sampling via the ventriculostomy was performed to monitor the response to treatment of the respective conditions. The table and graph represents the number of patients with urokinase instillation, amikicin irrigation and frequent sampling of CSF (without any of the 2 former procedures involved). [ Figure 7, Table 5].

Revision of ventriculostomy
A small number of cases required revision of ventriculostomy. A majority of these cases was due to blockage of the ventriculostomy diagnosed by observation of CSF flow and CT imaging. Other indications for revision include re-bleeding from previous intracranial haemorrhage, suboptimal position of the ventriculostomy and hydrocephalus (not related to the above mentioned causes). Most of the revisions were performed within 8 days of the initial insertion of ventriculosotmy with a mean duration of ventriculostomy before revision at 6.2 days. Most of the cases with a revised ventriculostomy required a longer duration before removal of the ventriculostomy. [ Figure 8 and 9.

Culture of the ventriculostomy tip
The intra-cranial end (drain tip) of the external ventricular drain would be sent for routine bacteria culture during operative revision or after removal under aseptic technique when the EVD were no longer required. A small number of ventricular drain tips were not saved for culture however, these include patients who died before the ventriculostomy were removed in which there were no clinical features suggesting a ventriculostomyrelated infection. Only 10 cases (3.5%) had a positive culture of the EVD tip, the majority bacteria group was Coagulase negative staphylococcus with bacillus as the second most common. [ Table 9, Figure 10].         Table 10. Figure 11].

Bacillus outbreak
Bacillus infection outbreak occurred during the period from September to November of 2007 involving 4 cases of postoperative bacillus CNS infection. Extensive investigation found heavy growth of bacillus in the air ventilation system of the operation theatre and intensive care unit. Major disinfection was performed, and the infection rate returned to similar levels before the outbreak. [4] Ventriculostomy-related infection A total of 23 patients had a positive culture result for CSF or EVD tip or both. Of these 23 patients, 8 patients were already diagnosed of meningitis before the procedure of ventriculostomy. Thus the remaining 15 patients had a positive culture after the procedure of ventriculostomy. For example:

Definition of Ventriculostomy-related infection
Martinez [12] and Mayhall [13] used the term for infection as "nosocomial ventriculitis" which is defined as fever (>38.5 Degrees Celsius) more than 2 days after ventriculostomy catheter insertion, and a positive CSF culture.
Alan P. Lozier [11] proposed a classification to infection in case of ventriculostomy as "Contamination", "Ventriculostomy colonization", "Suspected ventriculostomy-related infection", "Ventriculostomyrelated infection" and "Ventriculitis". The term of "Contamination" regarded as isolated positive CSF culture with normal CSF profile of protein, glucose and cell count. "Ventriculostomy colonization" being regarded as cases with repeated positive bacterial culture but no clinical symptoms of infection and normal CSF profile of protein, glucose and cell count.
Harrop JS [6] defined ventriculostomy infection as 2 positive CSF cultures from ventriculostomy catheters with a concurrent increase in cerebrospinal fluid white blood cell count. GKC Wong [16] Defined infection as a positive CSF culture or fever in the absence of other causes and either a raised white cell count, raised protein, decreased glucose in CSF, or organisms visible on CSF Gram stain.
Using the various definitions of infection in the abovementioned example with regards to the 15 cases of positive cultures in this review (pre-existing meningitis excluded) [ Table 11].      Figure 12].

Bacterial culture of ventriculostomy related infection
Of the 10 cases of ventriculostomy-related infection, 5 cases involved bacterial culture of bacillus species all resistant to penicillin or methicllin. The 4 cases grew a bacterial culture of coagulase negative staphylococcus resistant to methicillin. 1 case grew methicillin resistant staphylococcus aureus. All were sensitive to vancomycin and gentamicin. 7 cases were treated with a course of intravenous vancomycin according to sensitivity. Intravenous gentamicin was given to 4 of the cases [9] and intraventricular irrigation with amikicin via the ventriculosotmy had been used in 3 of the cases. [15] [ Figure 13].

Duration of EVD for ventriculostomy related infection
The mean duration of the ventriculostomy in the 10 cases of ventriculostomy-related infection is 9 days.

Revision of ventriculostomy and infection
27 cases in this review had revision, and among these cases, 2 (7.4%) were complicated with ventriculostomyrelated infection. Compared to the cases without revision complicated by infection (8 cases, 2.59%), the difference is statistically insignificant with a Fisher's exact test twotailed P=0.1876 [Table 14].

Urokinase instillation and infection
42 cases involved intraventricular urokinase instillation via the ventriculostomy, [10] only 1 case was complicated by ventriculostomy-related infection. The infection rate compared between the cases with urokinase instillation (2.38%) and with no urokinase instillation (3.06%) had no statistical significance with a Fisher's exact test P-value of 1.000 [ Table 15].

Haemorrhage and infection
7 out of 10 cases of ventriculostomy-related infection in this study presented with haemorrhagic events involving 5 cases subarachnoid haemorrhages (SAH) and 2 cases intra-cranial haemorrhage. The 2 cases involved head trauma with resulting in acute epidural haemorrhage, acute subdural haematoma and multiple contusion. The infection rate of haemorrhagic event was 3.24% and the rate of haemorrhage not as primary event was 2.50%, statistically insignificance (Fisher's Exact test P=1.000) [ Table 16].

Multiple catheter insertions and infection
There were 5 cases in this review involved a duration in which more than one ventriculostomy catheter were present in the patient at any time. 1 of these 5 cases was complicated by ventriculostomy-related infection. Fisher's exact two-tailed test P=0.141 (Statistically insignificant) [ Table 17].
Various factors including duration of ventriculostomy, revision, urokinase instillation, haemorrhage and multiple catheterization were considered for the association with ventriculostomy-related infection. These factors are all statistically insignificant in this review.

DISCUSSION
The ventriculostomy related infection rate in this   Table 19].
The results in this study were compared the statistics from the individual studies involved in the above mentioned literature review that included consideration of clinical features and pre-existing meningitis. The infection rate in our result was significantly lower compared to 4 out of 5 of the studies except for that of Lyke et al, (P=0.2798) which determined ventriculostomy-related infection as a positive bacterial culture after ruling out pre-existing meningitis but without considerations of clinical features. [Table 20].
2 literatures, which were not included in the abovementioned review, [11] did record a much lower infection rate. These two literatures, one by Friedman WA and Vries JK [5] and another by Bader, Littlejohns and Palmer [3] had stated an infection rate of 0%. A previous retrospective review of our centre for a 5-year-period prior the period of this study also demonstrated a low ventriculostomy-related infection rate of 0.92% (4/436). [1] The common feature among these literatures with low rate of infection was a very meticulous peri-operative management and strict nursing care aimed at reducing ventriculostomy-related infection. Bader, Littlejohns and Palmer [3] had a number of suggestions for methods to reduce infection.
At the author's centre, a number of peri-operative management and nursing practices were observed to prevent infection of ventriculostomy (listed in the table  below) [Table 21].

CONCLUSION
The infection rate of ventriculostomy at our department in the previous 4 years has been low (2.98%). No factors considered in this study showed a significant associated to infection. This overall result is comparable or more ideal compared to most other international studies.
A collective effort of meticulous peri-operative management and stringent nursing care had been proven to successfully decrease the incidence of ventriculostomyrelated infection in the author's centre and other centers that recorded a low infection rate.