RESEARCHHUMANCLINICAL STUDIES

TOPIC ResearchHumanClinical Studies

Risk Factors for Conversion to Permanent Ventricular Shunt in Patients Receiving Therapeutic Ventriculostomy for Traumatic Brain Injury
David F. Bauer, MD* Gerald McGwin, Jr, MS, PhD Sherry M. Melton, MD Richard L. George, MD, MSPH James M. Markert, MD, MPH*
*Division of Neurosurgery, University of Alabama at Birmingham, Birmingham, Alabama; Section of Trauma, Burns, and Surgical Critical Care, Department of Surgery, University of Alabama at Birmingham, Birmingham, Alabama; Department of Epidemiology, School of Public Health; University of Alabama at Birmingham, Birmingham, Alabama Correspondence: James M. Markert, MD, MPH Division Director, Neurosurgery, James Garber Galbraith Professor, Neurosurgery, Physiology and Pediatrics, and Cell Biology, University of Alabama, Birmingham, FOT No. 1060, 1530 3rd Ave S, Birmingham, AL 35294-3410. E-mail: markert@uab.edu Received, September 16, 2009. Accepted, June 7, 2010. Copyright 2010 by the Congress of Neurological Surgeons

BACKGROUND: Intracranial pressure is routinely monitored in patients with severe traumatic brain injury (TBI). Patients with TBI sometimes develop hydrocephalus, requiring permanent cerebrospinal fluid (CSF) diversion. OBJECTIVE: To quantify the need for permanent CSF diversion in patients with TBI. METHODS: Patients who received a ventriculostomy after TBI between June 2007 and July 2008 were identified, and their medical records were abstracted to a database. RESULTS: Sixteen of 71 patients (22.5%) receiving a ventriculostomy required a ventriculoperitoneal or ventriculoatrial shunt before discharge from the hospital. The average number of days between ventriculostomy and shunt was 18.3. Characteristics that predispose these patients to require permanent CSF diversion include the need for craniotomy within 48 hours of admission (odds ratio, 5.20; 95% confidence interval, 1.48-18.35) and history of culture-positive CSF (odds ratio, 5.52; 95% confidence interval, 1.19-25.52). Length of stay was increased in patients receiving permanent CSF diversion (average length of stay, 61 vs 31 days; P = .04). Patient discharge disposition was similar between shunted and nonshunted patients. CONCLUSION: In this retrospective study, 22% of TBI patients who required a ventriculostomy eventually needed permanent CSF diversion. Patients with TBI should be assessed for the need for permanent CSF diversion before discharge from the hospital. Care must be taken to prevent ventriculitis. Future studies are needed to evaluate more thoroughly the risk factors for the need for permanent CSF diversion in this patient population.
KEY WORDS: Hemorrhage, Shunt, Trauma, Ventriculostomy
Neurosurgery 68:8588, 2011
DOI: 10.1227/NEU.0b013e3181fd85f4

www.neurosurgery-online.com

ntracranial pressure (ICP) is routinely monitored in patients presenting with a poor neurological examination after traumatic brain injury (TBI).1,2 Current guidelines provide Level 2 evidence that ICP monitoring is beneficial in patients with imaging-evident TBI and a Glasgow Coma Scale between 3 and 8.3-5 Devices such as a fiberoptic intraparenchymal monitor and a ventriculostomy drain are routinely used. Ventriculostomy offers the added advantage
ABBREVIATIONS: ICP, intracranial pressure; NICU, neurological intensive care unit; TBI, traumatic brain injury

of therapeutic cerebrospinal fluid (CSF) diversion to decrease ICP. Patients with TBI sometimes develop hydrocephalus with increased ICP requiring permanent CSF diversion. Previous retrospective studies and case series have found the overall rate of posttraumatic hydrocephalus in patients with TBI to be between 0.7 and 45%.6-12 No study has looked at the rate of posttraumatic hydrocephalus in patients with severe TBI requiring ICP monitoring. This study sought to quantify the absolute risk for the need for permanent CSF diversion in this patient population, in addition to determining the risk factors that may predispose these patients to the need for permanent CSF diversion.

NEUROSURGERY

VOLUME 68 | NUMBER 1 | JANUARY 2011 | 85

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

BAUER ET AL

PATIENTS AND METHODS

This study was undertaken with the approval of the University of Alabama at Birmingham Institutional Review Board. Our hospital billing database was queried for all patients billed for a ventriculostomy at the University of Alabama Hospital between June 2007 and July 2008. Patients were included if they received the ventriculostomy for a Glasgow Coma Scale # 8 with radiographically confirmed TBI. Of the 82 patients identified, complete hospital records and preprocedural and postprocedural imaging studies were available for 71 patients. The charts of these 71 patients were retrospectively reviewed. A database was compiled containing the following information: time of admission, time of ventriculostomy placement, time of postventriculostomy cranial imaging, evidence of hemorrhage on postprocedural imaging, evidence of culture-positive CSF (defined as having at least 1 positive CSF culture found on routine specimens sampled 3 times weekly in all patients with ventriculostomies), type of injury on initial head computed tomography (CT; subarachnoid hemorrhage, contusion, intraparenchymal hemorrhage, intraventricular hemorrhage, or skull fracture), number of passes needed to place the ventriculostomy, need for permanent CSF diversion, time between the ventriculostomy and permanent CSF diversion, mechanism of injury, laboratory data on admission and at ventriculostomy placement, patient medical history, combined acute injury scores, place of patient discharge, and total length of hospital stay. Patients were categorized according to their need for permanent CSF diversion placed at some time during their initial hospital admission. These groups were compared with respect to demographic and clinical characteristics using analysis of variance and x2 tests for continuous and categorical variables, respectively. A stepwise logistic regression model was used to determine which demographic and clinical characteristics demonstrated significant, independent associations with the need for permanent CSF diversion. Values of P # .05 (2 sided) were considered statistically significant.

FIGURE . A noncontrast head computed tomography obtained at hospital admission of a patient with a traumatic brain injury requiring ventriculostomy placement. This patient eventually required permanent cerebrospinal fluid diversion.

RESULTS
Of the 71 patients, 81 ventriculostomies were placed with an average of 1.5 passes attempted per successful placement. Ten patients required placement of a second ventriculostomy. Three of these patients had unintended early removal of their ventriculostomy as a result of patient manipulation of catheter; 3 patients had initially functioning ventriculostomies that stopped functioning; and 4 patients had unsuccessful initial placement of ventriculostomy. These 4 patients received noncontrast head CT scans with the initial unsuccessful ventriculostomy in place. After analysis of the CT scan, all 4 patients received successful placement of a subsequent ventriculostomy. Stereotaxis was not used (Figure). Sixteen of 71 patients (22.5%) receiving a ventriculostomy required a ventriculoperitoneal or ventriculoatrial shunt before discharge from the hospital. The average time between placement of the ventriculostomy and permanent CSF diversion was 18.3 days. Of the patients receiving permanent CSF diversion, 11 were male and 5 were female. One patient had a small hemorrhage with the placement of ventriculostomy; 3 had intraventricular hemorrhage on admission; 9 required craniotomy within 48 hours of admission to evacuate a hematoma or for cerebral

decompression; and 3 had epidural hematomas. All patients had either a contusion or a discrete intracerebral hemorrhage on admission CT scan. Twelve patients had subarachnoid hemorrhage, and 14 had a skull fracture. Nine of 16 patients (56%) who underwent craniotomy eventually required a shunt before discharge from the hospital. Of these patients, 3 (33%) had a ventriculostomy placed in the operating room and 6 patients (66%) had a ventriculostomy placed in the neurological intensive care unit (NICU). Six of 21 patients who had a craniotomy within 48 hours of admission had removal of the bone flap at the time of surgery, and 3 of these 6 patients (50%) required a shunt. Six of the 15 patients who had the bone flap replaced at surgery (40%) required a shunt. In univariate analysis, increased age, need for craniotomy within 48 hours of admission, and history of culture-positive CSF were associated with the need for permanent CSF diversion (Table). There were no differences with respect to sex, mechanism of injury, mean injury severity score, mean Glasgow Coma Score, presence of other intracranial pathology (intraventricular hemorrhage, skull fracture, intraparenchymal hemorrhage), time to ventriculostomy placement, hemorrhage from ventriculostomy placement, and time to postprocedural imaging. In a multivariable logistic regression model, only craniotomy (odds ratio, 5.20; 95% confidence interval, 1.48-18.35) and culture-positive CSF (odds ratio, 5.52; 95% confidence interval, 1.19-25.52) demonstrated significant independent relationships with the need for permanent CSF diversion. With respect to outcomes, patients receiving permanent CSF

86 | VOLUME 68 | NUMBER 1 | JANUARY 2011

www.neurosurgery-online.com

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

SHUNT AFTER VENTRICULOSTOMY

TABLE. Demographic, Clinical, and Outcome Characteristics According to Need for Permanent Cerebrospinal Fluid Diversiona Need for Permanent CSF Diversion Yes Demographic Age, mean, y Male, % Mechanism of injury, % MVC Fall Other Clinical Glasgow Coma Scale score, mean Injury severity score, mean Intraventricular hemorrhage, % Intraparenchymal hemorrhage, % Skull fracture, % Craniotomy, % Time to ventriculostomy, mean, h Time to postprocedural imaging, % # 24 h . 24 h Ventriculostomy complication, % Infection Hemorrhage Outcome Length of stay, mean, d Discharge disposition, % Inpatient rehabilitation hospital Home Dead Nursing home Other
a

No 32.8 83.6 78.2 14.6 7.3 5.9 38.7 18.2 32.7 61.8 21.8 7.6 72.7 27.3 9.1 7.3 31.0 58.2 7.3 20.0 7.3 7.3

P .03 .19 .40

41.2 68.8 68.8 12.5 18.8 6.4 37.6 18.8 37.5 87.5 56.3 7.8 68.8 31.3 31.3 6.3 61.0 50.0 12.5 6.3 25.0 6.3

.64 .74 .96 .72 .05 .008 .40 .76

.03 1.00 .04 .20

CSF, cerebrospinal fluid; MVC, motor vehicle collision.

diversion had an increased mean length of hospitalization (61 vs 31 days) but no difference in discharge disposition.

DISCUSSION
Ventriculostomy placement is an important diagnostic and therapeutic tool in the treatment of TBI.2-5,13,14 According to procedural data published by the American Association of Neurological Surgeons, 42 446 ICP monitoring procedures were performed during 2006.15 It is routine practice for neurosurgeons to place an ICP monitoring device in a patient with a TBI if the patient has a poor neurological examination. At our institution, we weigh the infectious risk of keeping a ventriculostomy in for extended periods against the risks inherent in shunt placement. In general, we try to decide to shunt or to remove the ventriculostomy after the ICP has stabilized at , 20 cm water. Our practice is to drain CSF from patients

acutely at a pop-off of 10 cm water; then, once the ICP normalizes, an attempt is made to wean the ventriculostomy by requiring increased ICP before CSF drainage occurs. If the ICP climbs or if CSF output remains substantial, a shunt is placed. Because practice patterns vary for conversion from ventriculostomy to a shunt, these data may not be applicable in all instances. In this series, the average duration of external ventricular drainage was 18.3 days (range, 728 days), excluding 2 outliers with positive CSF cultures who had ventriculostomies for 54 and 100 days. We also sought to determine whether placing the ventriculostomy in the operating room made a difference in the need for permanent CSF diversion. Intuitively, conditions in the operating room may offer superior sterility than the bedside in the NICU. Nine of 16 patients who underwent craniotomy eventually required a shunt, and of these 9 patients, only 3 had a ventriculostomy placed in the operating room. The other 6 patients had ventriculostomy placed in the NICU. Subgroup analysis found no correlation between placement of the ventriculostomy in the operating room and the need for permanent CSF diversion. The number of patients who received a ventriculostomy in the operating room was so small that no conclusion can be made from this analysis, and further study of this question is warranted because infection plays a role in the eventual need for permanent CSF diversion. We also looked at patients who received a craniotomy within 48 hours of admission to see if leaving the bone flap off at the time of surgery increased the need for permanent CSF diversion. Of the 21 patients who had a craniotomy, 6 had the bone flap removed at the time of surgery. Three of these 6 patients required a shunt, whereas 6 of the 15 patients who had replacement of the bone flap required a shunt. There was no statistically significant difference between these 2 groups. Other groups have found a correlation between decompressive craniectomy and need for permanent CSF diversion. This study did not specifically address this question because of the relatively low number of decompressive craniotomies performed at our institution. The neurological rehabilitation literature is flush with articles detailing patient outcome after mild or severe TBI.11 A few studies have looked at the rate of hydrocephalus after TBI in an attempt to differentiate between hydrocephalus ex vacuo and communicating hydrocephalus.10,16 There is no published literature regarding the need for permanent CSF diversion in patients initially requiring ventriculostomy for poor neurological examination after TBI. Many articles have been written regarding the rate of conversion from ventriculostomy to permanent shunt in patients with aneurysmal subarachnoid hemorrhage, but we believe that this information cannot be extrapolated to the TBI population.17-19 A future study is planned to evaluate more thoroughly the risk factors for permanent CSF diversion after TBI requiring ventriculostomy. Interestingly, intraventricular hemorrhage was not a risk factor for permanent CSF diversion. Some authors have postulated that intraventricular hemorrhage increases the need for permanent

NEUROSURGERY

VOLUME 68 | NUMBER 1 | JANUARY 2011 | 87

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

BAUER ET AL

CSF diversion, but others have found no link between these 2 variables.17 Although there may be a causal relationship, we were not able to find one in our retrospective study.

CONCLUSIONS
In this retrospective study, 22% of TBI patients who required ventriculostomy eventually needed permanent CSF diversion. Patients with TBI should be assessed for the need for permanent CSF diversion before discharge from the hospital. Care must be taken to prevent ventriculitis. Disclosure
The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.

REFERENCES
1. Okor MO, Markert JM. Neurologic physiology: the brain and its response to injury. In: Bland KI, M.G.Sarr MA, Buchler W, Csendes A, Garden OJ, Wong J, eds. General Surgery, Principles and Practice. 2nd ed. Springer; 2008. 2. Kelly DF, McBride DQ, Becker DP. Surgical management of severe closed head injury in adults. In: Schmidek HH, Sweet WH, eds. Operative Neurosurgical Techniques: Indications, Methods, and Results. 3rd ed. Philadelphia, PA: WB Saunders; 1995;47-67. 3. Brain Trauma Foundation, American Association of Neurological Surgeons, Congress of Neurological Surgeons, Joint Section on Neurotrauma and Critical Care. Guidelines for the management of severe traumatic brain injury, 3rd edition. J Neurotrauma. 2007;24(1):S37S44. 4. Brain Trauma Foundation, American Association of Neurological Surgeons, Joint Section on Neurotrauma and Critical Care. Guidelines for the management of severe traumatic brain injury. J Neurotrauma. 2000;17:457-554. 5. Bullock R, Chestnut RM, Clifton G, et al. Guidelines for the management of severe head injury. J Neurotrauma. 1996;13:641-734.

6. Cardoso ER, Galbraith S. Posttraumatic hydrocephalus: a retrospective review. Surg Neurol. 1985;23(3):261-264. 7. Kan P, Amini A, Hansen K, et al. Outcomes after decompressive craniectomy for severe traumatic brain injury in children. J Neurosurg. 2006;105(5)(suppl):337-342. 8. Kishore PR, Lipper MH, Miller JD, Girevendulis AK, Becker DP, Vines FS. Posttraumatic hydrocephalus in patients with severe head injury. Neuroradiology. 1978;16:261-265. 9. Licata C, Cristofori L, Gambin R, Vivenza C, Turazzi S. Post-traumatic hydrocephalus. J Neurosurg Sci. 2001;45(3):141-149. 10. Marmarou A, Abd-Elfattah Foda MA, Bandoh K, et al. Posttraumatic ventriculomegaly: hydrocephalus or atrophy? A new approach for diagnosis using CSF dynamics. J Neurosurg. 1996;85:10261035. 11. Mazzini L, Campini R, Angelino E, Rognone F, Pastore I, Oliveri G. Posttraumatic hydrocephalus: a clinical, neuroradiologic, and neuropsychologic assessment of long-term outcome. Arch Phys Med Rehabil. 2003;84(11):1637-1641. 12. Phuenpathom N, Ratanalert S, Saeheng S, Sripairojkul B. Post-traumatic hydrocephalus: experience in 17 consecutive cases. J Med Assoc Thai. 1999;82(1):46-53. 13. Vespa PM, Bergsneider M, Becker DP. Intracranial pressure monitoring: operative procedure and technical considerations. In: Kaye AH, Black P, eds. Operative Neurosurgery. London, UK: Harcourt; 2000;1205-1213. 14. Roitberg BZ, Khan N, Alp MS, Hersonskey T, Charbel FT, Ausman JI. Bedside external ventricular drain placement for the treatment of acute hydrocephalus. Br J Neurosurg. 2001;15(4):324-327. 15. National Neurosurgical Procedural Statistics. American Association of Neurological Surgeons Survey, 2006. 16. Tribl G, Oder W. Outcome after shunt implantation in severe head injury with post-traumatic hydrocephalus. Brain Injury. 2000;14(4):345-354. 17. Chan M, Alaraj A, Calderon M, et al. Prediction of ventriculoperitoneal shunt dependency in patients with aneurysmal subarachnoid hemorrhage. J Neurosurg. 2009;110(1):44-49. 18. Dorai Z, Hynan LS, Kopitnik TA, Samson D. Factors related to hydrocephalus after aneurysmal subarachnoid hemorrhage. Neurosurgery. 2003;52(4):763-769. 19. Rammos S, Klopfenstein J, Augsburger L, et al. Conversion of extra ventricular drains to ventriculoperitoneal shunts after aneurysmal subarachnoid hemorrhage: effects of site and protein/red blood cell counts on shunt infection and malfunction. J Neurosurg. 2008;109(6):1001-1004.

88 | VOLUME 68 | NUMBER 1 | JANUARY 2011

www.neurosurgery-online.com

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.