Introduction
External ventricular drainage (EVD) has been one of the most frequently practiced operations in neurosurgery. EVD can be extremely useful by which to monitor and/or manage intracranial pressure. The indications for EVD insertions include monitoring and treatment of severe head injuries, hydrocephalus, ventriculo-peritoneal shunt infections, post-operative cerebrospinal fluid leakage, prophylaxis for cerebrospinal fluid leakage and ventriculitis. 
Even though EVD insertion is a commonly performed procedure, studies regarding EVD techniques, complications, or risk factors are limited. It is well-known that infections or hemorrhage frequently occurs as complications of EVD insertion.4)8) Numerous studies have been conducted as early as 1976 on EVD-related infections including preventive techniques and protocols to minimize the infection rate.3)7) In contrast, only a few studies have addressed EVD-related hemorrhage. Meyer et al.3) conducted a prospective study on EVD-related hemorrhage and reported a 1% rate of hemorrhagic complications in 200 patients (0.5% from actual ventricular puncture), but did not utilize post-procedural computerized tomography (CT) scans to analyze the risk factors. In this study, a retrospective analysis was conducted for patients who underwent EVD insertions to determine the rate of hemorrhage and possible underlying risk factors.
Materials and Methods
Between January 2003 and January 2011, 229 patients were selected from a single institution who underwent EVD procedures to manage increased intracranial pressure due to various intracranial diseases. All of the selected patients received catheter insertions for EVD at Kocher’s point (right, left, or bilateral). The catheter diameter was 3.0mm (small) or 3.8mm (large). To minimize confounding factors, only the patients who had the same operating room settings were selected for analysis. The patients who underwent EVD catheter insertion at other points, such as the parieto-occipital point, were excluded because the number of cases was too small for analysis. Patients who underwent ventriculo-peritoneal shunts were also excluded because of the different type of catheter used.
All pre-procedural and post-procedural CT scans were reviewed. Catheter-induced hemorrhage (CIH) was defined as any evidence of new hemorrhage along the tract of the catheter or newly developed intraventricular hemorrhage on the post-procedural CT scan obtained within 24 hours of EVD catheter insertion. Patients were divided into non- catheter-induced hemorrhage group (non-CIH group) and catheter-induced hemorrhage group (CIH group). The rate of hemorrhage was calculated and the possible risk factors were analyzed. Possible risk factors are gender, age, side of catheter insertion sites (right, left, or bilateral), underlying morbidity, histories of heart disease including the use of antiplatelet or anticoagulant medications and the catheter diameter. These factors were statistically analyzed using chi-square or Fisher’s exact tests. After determining statistically significant risk factors, logistic regression analysis and multiple logistic regression were performed to calculate the odds ratio. All statistical analyses were performed using SAS 9.1 (SAS Institute, Inc., Cary, NC, USA) and statistical significance was defined as a p value <0.05.
Results
The patients underwent EVD catheter insertions for a diverse range of indications, including subarachnoid hemorrhage due to aneurysm rupture, traumatic brain injuries, ventriculo-peritoneal shunt infections, intraparenchymal hemorrhage and intraventricular hemorrhage.
Among the 229 patients, 133 were males and 96 were females. In the non-CIH group, 125 (60.1%) patients were males and 83 (39.9%) patients were females. In the CIH group, 8 (38.1%) patients were males and 13 (61.9%) patients were females. The rate of hemorrhage was higher in females (31 of 96 [13.5%]) than males (6.0%, p=0.051). Overall, the review of the records revealed that 21 patients experienced CIH resulting in a hemorrhage rate of 9.17% among 229 patients. Fig. 1 depicts the case of a 78-year-old female with colon cancer and liver metastasis who developed spontaneous intraparenchymal and intraventricular hemorrhage. She was admitted and underwent EVD catheter insertion at both Kocher's points bilaterally using large-diameter-catheters (3.8mm). The post- procedural CT showed CIH along both EVD catheters.
The EVD catheters also carried a risk of causing intraventricular hemorrhage because the intended target was inside the ventricle. Fig. 2 displays the post-procedural CT scans of a 84-year-old female with a right internal carotid artery territory infarction, and she underwent decompressive craniectomy and EVD insertion at the left Kocher’s point. A small diameter catheter (3.0mm) was used for EVD insertion and caused minimal intraventricular hemorrhage in the posterior horn of the left lateral ventricle.
There were patients with liver diseases such as liver cirrhosis or chronic hepatitis (n=5) whose blood chemistry revealed mild prolongation of prothrombin time (PT); International Normalized Ratio (INR) <2.0, or activated partial thromboplastin time (aPTT) <40 seconds. However, that did not affect or cause any complication after operations. In fact, patients with severe hematologic disorder (platelet count <50,000) or compromised their coagulation system were excluded from surgical indication due to clear evidence of bleeding tendency.
Only eight patients were <20 years of age and the rate of hemorrhage increased as the age increased (p=0.34, Table 1). Table 2 indicates that elderly patients ≥60 years of age had an increased probability of CIH, and age <60 years showed a decreased rate of hemorrhage (p=0.023). In most cases, the amount of hemorrhage was minimal (approximately 1 cc along the catheter tract) and none of the patients in the CIH group had any detectable neurologic deficits from the hemorrhage. 
The general protocol was that the catheter was inserted at the right Kocher’s point in an attempt to avoid the dominant hemisphere, unless the condition of the patient required otherwise. The data showed that the right Kocher’s point had a lower rate of hemorrhage than left or bilateral insertion (p<0.001). Pre-existing high blood pressure appears to increase the risk of hemorrhage, but showed no significance. However, the patients who had a history of heart disease had a higher risk of CIH (p<0.001). A large diameter catheter was used for patients diagnosed with intraventricular hemorrhage on the pre-procedural CT scan to evacuate the hemorrhage, but the diameter did not affect the risk of complications.
After calculating p values, a number of potential risk factors showed some significance, so the odds ratios were evaluated (Table 3). The risk of CIH was 2.45-fold higher in females than males. Patients ≥60 years of age had a 2.8-fold increased risk of CIH compared to patients <60 years of age. The patients who underwent bilateral EVD catheter insertions had a higher risk of CIH than patients who had a unilateral EVD insertion. A history of heart disease increased the risk of CIH 20 times higher than patients without any heart problems (p<0.001). To manage their heart problems, taking antiplatelet or anticoagulant medication is very common. We found that 15 patients from non-CIH group and 4 from CIH group were taking either aspirin or warfarin. Statistically antiplatelet or anticoagulant medication increased hemorrhagic complication risk by 3.26 times (p=0.057).
Based on the data analysis, significant risk factors for CIH were age, bilateral catheter insertion and a history of heart disease; these risk factors were subjected to multiple logistic regression analysis (Table 4). Assuming that a patient has undergone bilateral EVD insertion and has a history of heart disease, a patient ≥60years of age would have a 3-fold increased risk of CIH than a patient <60years of age. Bilateral EVD insertion in an elderly patient (age ≥60years) with heart disease would increase the risk of CIH by 6 times. Lastly, a history of heart disease would increase the risk of CIH 18-fold for elderly patients (age ≥60years) who underwent bilateral EVD insertion.
Discussion
The most common complication of EVD insertion is infection, ranging from 5%~30%;2) however, hemorrhages induced by catheters are frequently occurring complications that cannot be overlooked.1) North and Reilly5) compared complication rates between a ventricular catheter, Richmond screws and a subdural catheter and they concluded that a ventricular catheter caused the most infection and hemorrhages. One study was conducted on 648 patients who underwent EVD insertion and acknowledged only one symptomatic CIH, but provided no data for asymptomatic hemorrhage since their study focus was on infection.8) Wiesmann and Mayer9) reported a hemorrhage rate of 6.5% related to EVD catheter insertion during typical catheter ventriculostomies rather than needle trephination (12.2%). They simply compared two methods of EVD and concluded that hemorrhagic complications occur far more frequently than anticipated. In this study, we evaluated data of a single institution and tried to correlate a number of variables to identify statistically significant risk factors for CIH. According to the retrospective analysis, the incidence rate of CIH was 9.17%. Gardner et al. conducted a study comparing rates of hemorrhage between operating room and intensive care unit (ICU) settings; the ICU setting had a much higher rate of CIH (43.4%).1) They demonstrated that proper monitoring and equipment in the operating room, including necessary staff members, is essential to minimize complications associated with EVD catheter insertion. A potential confounding factor is that catheter-induced intraventricular hemorrhage may have been occurred spontaneously with delayed onset and inadvertently misinterpreted as hemorrhages induced by catheters. Unfortunately, the number of puncture trials for each patient or procedure was not included in the records, thus it cannot be determined whether or not the hemorrhage was induced due to multiple punctures through the brain parenchyma until the desired target was reached. 
The data analysis concluded that female patients were more prone to CIH after EVD insertions. Thirteen of 96 females developed CIH, which is higher than males (8 of 133), but gender was not significant as a risk factor. In contrast, the data indicate that the rate of hemorrhage increased as the age of the patient increased. Most cases of CIH occurred in patients ≥60 years of age, confirming the age is a significant risk factor. The most significant risk factor for CIH after EVD insertions was a history of heart disease. Six of 21 patients with CIH had a history of heart disease, such as atrial fibrillation (n=4), angina pectoris (n=1), or myocardial infarction (n=1). Three of them were using aspirin and one patient was using warfarin to manage the heart conditions. Other two patients were not aware of their atrial fibrillation and it was incidentally found out during evaluations. Among the patients without any history of heart disease, only 15 patients (6.85%) developed CIH and none of them were on any antiplatelet or anticoagulation medication. In the non-CIH group (n=208), 10 patients were taking aspirin as prophylactic agent with their hypertension medication. Additional 5 patients were on aspirin due to previous cerebral infarction (n=4) or atrial fibrillation (n=1). Overall, 19.05% (n=4) of CIH group and 6.73% (n=15) of non-CIH group were taking either aspirin (n=18) or warfarin (n=1). That led to CIH rate of 22.22% among patients with antithrombotic therapy (p=0.068), and taking antiplatelet or anticoagulant medication increases hemorrhagic complication risk by 3.26 times (p=0.057). Statistical values may not have proven significance (p<0.05); however, we believe that more sufficient number of data would correlate such medication with CIH.
According to our series, age ≥60years, heart diseases, and bilateral catheter insertions were proven to be the most significant risk factors for CIH. Multiple logistic analysis (Fig. 3) suggested that the combination of all three risk factors dramatically increased the likelihood of CIH after EVD procedures. In this study, a history of heart disease was statistically proven to be the most influential risk factor for CIH for EVD procedures.
Conclusion
Although the age, bilateral catheter insertions and heart problems were statistically proven to have clinical significance as potential risk factors, there are limits to which they are managed to reduce complications. Therefore, when catheter insertion for EVD is required to manage intracranial pressure or a hematoma removal, all presenting risk factors should be considered and managed. Furthermore, more careful and skillful maneuvers and techniques are imperative to reduce the chance of hemorrhagic complications for high risk patients.
REFERENCES
1)Gardner PA, Engh J, Atteberry D, Moossy JJ. Hemorrhage rates after external ventricular drain placement. J Neurosurg 110: 1021-5, 2009
2)Kim DK, Uttley D, Bell BA, Marsh HT, Moore AJ. Comparison of rates of infection of two methods of emergency ventricular drainage. J Neurol Neurosurg Psychiatry 58:444-6, 1995
3)Meyer B, Schaller K, Rohde V, Hassler W. Percutaneous needle trephination. Experience in 200 cases. Acta Neurochir (Wien) 127:232-5, 1994
4)Narayan RK, Kishore PR, Becker DP, Ward JD, Enas GG, Greenberg RP et al. Intracranial pressure: to monitor or not to monitor?: a review of our experience with severe head injury. J Neurosurg 56:650-9, 1982
5)North B, Reilly P. Comparison among three methods of intracranial pressure recording. Neurosurgery 18:730-2, 1986
6)Pfausler B, Spiss H, Beer R, Kampl A, Engelhardt K, Schober M et al. Treatment of staphylococcal ventriculitis associated with external cerebrospinal fluid drains: a prospective randomized trial of intravenous compared with intraventricular vancomycin therapy. J Neurosurg 98:1040-4, 2003
7)Rosner MJ, Becker DP. ICP monitoring: complications and associated factors. Clin Neurosurg 23:494-519, 1976
8)Sundbarg G, Nordstrom CH, Soderstrom S. Complications due to prolonged ventricular fluid pressure recording. Br J Neurosurg 2:485-95, 1988
9)Wiesmann M, Mayer TE. Intracranial bleeding rates associated with two methods of external ventricular drainage. J Clin Neurosci 8:126-8, 2001