Department of Neurological Surgery, Columbia University, New York

Transcription

1 J Neurosurg 111: , 2009 Efficacy of lamina terminalis fenestration in reducing shunt-dependent hydrocephalus following aneurysmal subarachnoid hemorrhage: a systematic review Clinical article Ri c a r d o J. Ko m o t a r, M.D., Dav i d K. Ha h n, M.D., Gr a c e H. Kim, M.D., Ro b e r t M. Sta r k e, B.A., Ma t t h e w C. Ga r r e t t, B.S., Max w e l l B. Me r k o w, B.S., Mar c L. Ot t e n, M.D., Ro b e r t R. Sc i a c c a, En g.sc.d., a n d E. Sa n d e r Co n n o l ly Jr., M.D. Department of Neurological Surgery, Columbia University, New York Object. Chronic hydrocephalus requiring shunt placement is a common complication following aneurysmal subarachnoid hemorrhage (SAH). Controversy exists over whether microsurgical fenestration of the lamina terminalis during aneurysm surgery affords a reduction in the development of shunt-dependent hydrocephalus. To resolve this debate, the authors performed a systematic review and quantitative analysis of the literature to determine the efficacy of lamina terminalis fenestration in reducing aneurysmal SAH associated shunt-dependent hydrocephalus. Methods. A MEDLINE ( ) database search was performed using the following keywords, singly and in combination: ventriculoperitoneal shunt, hydrocephalus, subarachnoid hemorrhage, aneurysm, fenestration, and lamina terminalis. Additional studies were manually singled out by scrutinizing references from identified manuscripts, major neurosurgical journals and texts, and personal files. A recent study from the authors institution was also incorporated into the review. Data from included studies were analyzed using the chi-square analysis and Student t-test. The Cochran-Mantel-Haenszel test was used to compare overall incidence of shunt-dependent hydrocephalus. Results. The literature search revealed 19 studies, but only 11 were included in this review, involving 1973 patients. The fenestrated and nonfenestrated cohorts (combined from the various studies) differed significantly with regard to patient sex, age, and clinical grade as well as aneurysm location (p = , , , and 0.017, respectively). The overall incidence of shunt-dependent hydrocephalus in the fenestrated cohort was 10%, as compared with 14% in the nonfenestrated cohort (p = 0.089). The relative risk of shunt-dependent hydrocephalus in the fenestrated cohort was 0.88 (95% CI ). Conclusions. This systematic review revealed no significant association between lamina terminalis fenestration and a reduced incidence of shunt-dependent hydrocephalus. The interpretation of these results, however, is restricted by unmatched cohort differences as well as other inherent study limitations. Although the overall literature supports lamina terminalis fenestration, a number of authors have questioned the technique s benefits, thus rendering its efficacy in reducing shunt-dependent hydrocephalus unclear. A well-designed, multicenter, randomized controlled trial is needed to definitively address the efficacy of this microsurgical technique. (DOI: / JNS0821) Ke y Wo r d s aneurysm lamina terminalis microsurgical fenestration shunt-dependent hydrocephalus subarachnoid hemorrhage Ch r o n i c hydrocephalus requiring shunt placement is a common complication following aneurysmal SAH. More specifically, aneurysmal SAH is the leading cause of chronic hydrocephalus requiring shunt placement in adults, accounting for ~ 35% of all cases. 13 Although the precise mechanism is unknown, aneurysmal SAH appears to cause fibrosis of the leptomeninges and arachnoid granulations, 19,24 leading to impaired CSF Abbreviations used in this paper: ACoA = anterior communicating artery; SAH = subarachnoid hemorrhage. J Neurosurg / Volume 111 / July 2009 dynamics, reduced CSF absorption, and shunt-dependent hydrocephalus in > 20% of patients with aneurysmal SAH. 4,9,10,12 However, CSF shunts have a historically high risk of failure, with some reports quoting a shunt failure rate as high as 43% at 1 year and 85% at 10 years after placement. 5,20,25 Thus, in light of these treatment limitations, chronic hydrocephalus remains a significant contributor to morbidity and death following aneurysmal SAH. Authors of recent publications have investigated the possibility that microsurgical fenestration of the 147

2 R. J. Komotar et al. lamina terminalis during aneurysm surgery reduces the incidence of shunt-dependent hydrocephalus. 1,3,6,14,17,21 23 Opening the lamina terminalis creates an anterior ventriculostomy that may facilitate CSF dynamics. Improved CSF flow may reduce subarachnoid fibrosis and vascular inflammation, thereby leading to decreased hydrocephalus. However, the results of these investigations have been conflicting, rendering the efficacy of this surgical maneuver controversial. A better understanding of the relationship between microsurgical fenestration of the lamina terminalis and shunt-dependent hydrocephalus is critical to refining the operative treatment of aneurysmal SAH. To this end, we performed a systematic review of the literature to determine the efficacy of this surgical technique. Methods Study Selection We performed a literature search using the Ovid gateway of the MEDLINE database for the years between 1950 and The following keywords were queried singly and in combination: ventriculoperitoneal shunt, hydrocephalus, subarachnoid hemorrhage, aneurysm, fenestration, and lamina terminalis. The search was limited to studies published in English, and human was specified as the study category. All papers focusing on the incidence, origin, or prognosis of shunt-dependent hydrocephalus following aneurysmal SAH were selected, whereas editorials, commentaries, case reports, and reviews were not. Papers were carefully examined, and only those studies directed at the effect of lamina terminalis fenestration on the incidence of shunt-dependent hydrocephalus were included in our review. A manual search for reports was also conducted by scrutinizing the references from identified papers, major neurosurgical journals and texts, and personal files. Data Extraction Included studies were reviewed and carefully scrutinized for study design, methodology, patient characteristics, and primary findings. The numbers of patients in each study were extracted and divided into cohorts according to treatment, that is, fenestration versus no fenestration. Data on cohort characteristics were recorded when available, including mean age, sex, admission hydrocephalus, clinical grade, Fisher grade, and aneurysm location. The incidence of shunt-dependent hydrocephalus was also recorded: the number of patients who received shunts postoperatively divided by the number of patients evaluated. Data from the individual studies were combined by cohort and then compared. The chi-square test and Student t-test were used to compare cohort characteristics. The Cochran-Mantel-Haenszel test was used to compare the overall incidence of shunt-dependent hydrocephalus between the 2 cohorts and to calculate a pooled relative risk and its corresponding 95% CI. The Breslow-Day test was also performed to test for significant heterogeneity in ORs among the studies. A probability value < 0.05 was considered significant. Results Study Selection Fifteen published studies were identified through our initial MEDLINE database search. After careful scrutiny of these studies, 8 were rejected from our review because they did not specifically address the effect of lamina terminalis fenestration on postoperative shunt placement. An additional 3 studies were gleaned from a manual review of the references of identified papers, major neurosurgical journals and texts, and personal files. A recent study from our institution was also included. 15 A total of 11 studies were included in this review. Study Characteristics Characteristics, primary findings, and limitations of the studies included in this review are summarized in Table 1. Regarding study design, 6 analyses were retrospective 2,14,15,17,21,26 and 5 were prospective. 1,3,6,22,23 There were no randomized controlled trials. Data on patient age were provided in 10 studies; the weighted mean age for the overall study population was 51 years. Data on patient sex were also provided in 10 studies; 66% of the overall study population for whom data were available was female. All studies had data on patient clinical grade, and all studies except 1 26 recorded Fisher grade data. Nine (90%) of 10 studies were composed primarily (> 50%) of patients with Fisher Grade 3 or 4 SAH. Data regarding hydrocephalus on admission were provided in 7 studies; 1,3,6,14,15,17, patients (33%) of those for whom data were available had been admitted with hydrocephalus. Three studies 14,15,17 adjusted their primary outcome measure for this variable (Table 1). Aneurysm site data were available in all except 2 studies. 6,22 The majority of aneurysms (1461 [92%]) in those patients for whom data were available were located in the anterior circulation: carotid artery or its branches, anterior and middle cerebral arteries, or ACoA. Fewer aneurysms (118 [8%]) were located in the posterior circulation: basilar, vertebral, or posterior cerebral arteries (Table 1). The limitations of each study are summarized in Table 1. In all analyses, the decision to fenestrate the lamina terminalis was based on surgeon preference and aneurysm location; patients were not randomized to the fenestration arm in any of the studies. Other study limitations included poor or no control groups, small cohort sizes, and the participation of multiple surgeons resulting in variable management paradigms within a study. In addition, information on surgeon experience or the criteria for shunt placement was never made explicit in any study. Patient Characteristics In total, 1973 patients with aneurysmal SAH were included in this review. Patient demographic and clinical characteristics were organized according to fenestration versus no fenestration when the distinction was made clear and are summarized in Table 2. Among the 11 studies, 7 3,6,14,15,17,21,22 had proper control groups, whereas 4 1,2,23,26 had historical controls. 148 J Neurosurg / Volume 111 / July 2009

3 Microsurgical fenestration of the lamina terminalis TABLE 1: Characteristics of studies on the effect of lamina terminalis fenestration on shunt-dependent hydrocephalus* Authors & Year Study Design FLT No. of Patients no FLT Total No. of Females (%) Mean Age (yrs) Aneurysm Location Clinical Grade (% patients) % Patients w/ Fisher Grade 3 4 SAH % Patients w/ HCP on Admission Main Findings Study Limitations Sindou, 1994 pros NA NA NA 14% H&H IV V 34 NA n o significant difference in shunt rates: 15% FLT vs 11% non-flt Yonekawa et al., 1998 Tomasello et al., 1999 Schmieder et al., 1999 Komotar et al., 2002 A ndaluz & Zuccarello, 2003 Andaluz et al., 2004 Dehdashti et al., 2004 A kyuz & Tuncer, 2006 Kim et al., 2006 F LT & non-flt cohorts not completely divided, patient characteristics not provided retro (69) 50 94% ant circ 43% WFNS III V NA NA overall shunt rate 9.0% re tro, patients not randomized, no control group, effect of FLT not well isolated pros (71) % ant circ 100% H&H IV V overall shunt rate 4.2% sm all cohort, patients not randomized, no control group retro (70) 51 93% ant circ 33% H&H IV V 65 NA overall shunt rate 11% re tro, patients not randomized, effect of FLT not well isolated, small benefit (15 17% vs 11%) retro (78) 52 83% ant circ 17% H&H IV V sh unt rate reduced by 80%: 2.3% FLT vs 12.6% non-flt re tro, multiple surgeons, small control group retro (53) % ACoA 18% H&H IV V 68 NA overall shunt rate 6.9% re tro, small cohort, patients not randomized, no control group, effect of FLT not well isolated pros (15) % ACoA 16% H&H IV V si gnificantly reduced shunt rate: 4.3% FLT vs 13.9% non-flt pros (46) 49 NA 16% WFNS IV V n o significant difference in shunt rate: 14% vs 19% coiled pa tient selection, difficult to extrapolate outside of ACoA, patients w/ Fisher Grade 3 SAH, multiple surgeons pa tients not randomized, poor control group (patients w/ coiling), effect of FLT not well isolated pros (58) % ACoA 31% H&H IV V o verall shunt rate 9.8% no control group, effect of FLT not well isolated retro (51) % ACoA 54% H&H IV V n o significant difference in shunt rate: 22% FLT vs 29% non-flt retro, small cohort, multiple surgeons Komotar et al., 2008 retro (75) % ant circ 22% H&H IV V n o significant effect: 25% vs 16% shunt rate retro * ant circ = anterior circulation; FLT = fenestration of lamina terminalis; H&H = Hunt and Hess; HCP = hydrocephalus; NA = not available; pros = prospective; retro = retrospective; WFNS = World Federation of Neurosurgical Societies. All Grade 3 SAH. J Neurosurg / Volume 111 / July

4 R. J. Komotar et al. TABLE 2: Patient cohort characteristics in studies examining the effect of lamina terminalis fenestration on shunt-dependent hydrocephalus* Authors & Year Total No. of Patients No. of Females Age (yrs) Patients w/ Adm HCP Patients w/ H & H Grade Good (I III) Poor (IV V) Patients w/ Fisher Grade SAH Good (1 2) Poor (3 4) Patients w/ Ant Lesion Location Patients w/ Shunt-Dependent HCP Sindou, /126 NA NA NA NA NA NA NA NA 11/14 Y onekawa et 150/NA 104/NA 50/NA NA 85/NA 65/NA NA NA 141/NA 13/NA al., 1998 T omasello et 52/NA 37/NA 55/NA 5/NA 0/NA 52/NA 0/NA 52/NA 52/NA 2/NA al., 1999 S chmieder et 112/26 NA NA NA NA NA NA NA NA 11/4 al., 1999 K omotar et 139/395 NA NA NA NA NA NA NA NA 3/50 al., 2002 A ndaluz & 40/NA 21/NA 52/NA NA 33/NA 7/NA 13/NA 27/NA 40/NA 2/NA Zu ccarello, 2003 A ndaluz et 53/53 7/9 52/51 23/27 46/43 7/10 0/0 53/53 53/53 2/5 al., 2004 D ehdashti et 180/65 112/45 49/52 34/30 151/45 29/20 63/20 117/45 NA 25/13 al., 2004 A kyuz & 71/NA 41/NA 48/NA 16/NA 49/NA 22/NA 32/NA 39/NA 71/NA 7/NA Tuncer, 2006 K im et al., 36/35 17/19 49/55 27/17 19/14 17/21 13/8 23/27 36/35 8/ K omotar et 71/298 48/230 51/53 16/41 47/242 24/56 19/113 52/185 71/298 18/48 al., 2008 total (%) 975/ (59)/ 303 (67) 50.2/ (26)/ 115 (26) 430 (66)/ 344 (76) 223 (34)/ 107 (24) )/ 141 (31) 363 (72)/ 310 (69) 464 (98)/ 386 (100) 102 (10)/144 (14) p Value * Values are expressed for FLT/no FLT group. Abbreviation: Adm = admission. Four studies did not have control groups. Data from primary analysis only. Control group consists of patients who underwent coil embolization. Only anterior circulation aneurysms were fenestrated, but this number was not reported. Percentages refer to studies in which patient data were available. Authors in 8 studies distinguished patient sex and age between fenestrated and nonfenestrated cohorts: 59 and 67% of patients were female (p = ) and the weighted mean age was 50.2 and 52.8 years (p = ), respectively. Clinical grade was distinguished between fenestrated and nonfenestrated groups in 8 studies, with 34 and 24% of patients having a poor grade (Hunt and Hess Grades IV or V), respectively (p = ). Seven studies documented aneurysm location in fenestrated and nonfenestrated groups, with 98 and 100% of aneurysms located in the anterior circulation, respectively (p = 0.017; Table 2). Six studies differentiated the number of patients with acute hydrocephalus on admission, showing 26% of patients with acute hydrocephalus in both the fenestrated and nonfenestrated cohorts. Seven studies distinguished Fisher grades, showing 72 and 69% of Fisher Grade 3 or 4 SAHs in the fenestrated and nonfenestrated cohorts, respectively. This difference in Fisher grades was not statistically significant (p = 0.31). Incidence of Shunt-Dependent Hydrocephalus The overall incidence of shunt-dependent hydrocephalus in the fenestrated cohort was 10%, as compared with 14% in the nonfenestrated cohort (Table 2). A pooled analysis included only the 7 studies 3,6,14,15,17,21,22 with a nonfenestrated cohort, and the other 4 studies 1,2,23,26 were excluded. The overall association between fenestration and the occurrence of shunt-dependent hydrocephalus was not statistically significant (p = 0.089). The pooled relative risk of shunt-dependent hydrocephalus in the fenestrated 150 J Neurosurg / Volume 111 / July 2009

5 Microsurgical fenestration of the lamina terminalis cohort was 0.88 (95% CI ). The Breslow-Day test revealed significant heterogeneity between studies (p < 0.05). Seven studies supported the benefit of lamina terminalis fenestration and were analyzed separately; their characteristics are summarized in Table 3. Three of these studies 3,17,21 included proper control groups, whereas 4 1,2,23,26 used only historical controls. All studies included patient populations in which > 50% of patients had Fisher Grade 3 or 4 SAHs. The pooled estimate of effect size was 6 versus 12% for the incidence of shunt-dependent hydrocephalus in fenestrated versus nonfenestrated cohorts (p = 0.001), respectively. The relative risk of shuntdependent hydrocephalus in the fenestrated cohort was 0.69 (95% CI ). The data in 4 studies refuted the benefit of lamina terminalis fenestration and were analyzed separately (Table 4). All 4 studies included proper control groups. Three (75%) of the 4 studies included populations in which > 50% of patients had Fisher Grade 3 or 4 SAHs. The pooled estimate of effect size was 17 versus 16% for the incidence of shunt-dependent hydrocephalus in fenestrated versus nonfenestrated cohorts (p = 0.74), respectively. The relative risk of shunt-dependent hydrocephalus in the fenestrated cohort was 1.05 (95% CI ). Discussion Chronic hydrocephalus is a well-known complication after aneurysmal SAH, occurring in > 20% of patients. 11 On diagnosis of this disorder, shunt-placement is necessary to avoid cerebral injury due to ventricular dilation. 7,8 Note, however, that CSF shunts have high rates of malfunction, with > 40 and 80% of shunts failing at 1 and 10 years after placement, respectively. 5,20,25 Moreover, shunt placement and revision can lead to intracerebral hemorrhage, infection, CSF leakage, and bowel or lung perforation, with mortality rates as high as 7 9% in select populations. 20,25 In this context, it is critical to understand the impact of fenestration of the lamina terminalis on the incidence of shunt-dependent hydrocephalus to better guide the surgical treatment of patients with aneurysmal SAH. Several authors 1 3,6,14,15,17,21 23,26 have investigated the effect of this maneuver. Tomasello et al. 23 and Sindou 22 have reported on small series in which fenestration of the lamina terminalis during aneurysm surgery reduced the incidence of chronic hydrocephalus and improved clinical outcomes, respectively. In 1998 Yonekawa and colleagues 26 cited a shunt placement rate of only 9% in their structured approach to aneurysm surgery, including routine fenestration of the lamina terminalis. In 1999 Schmieder et al. 21 retrospectively reviewed the records of 138 patients with aneurysmal SAH and found that fenestration of the lamina terminalis was potentially beneficial in selected patients; the maneuver had been performed in 73% of patients with shunt-dependent hydrocephalus compared with 82% of patients without. In 2002 Komotar and colleagues 17 recommended fenestration of the lamina terminalis whenever possible during aneurysm surgery after having performed a retrospective review of 582 patients in whom fenestration was associated with a > 80% reduction in the incidence of shunt-dependent hydrocephalus. In 2003 Andaluz et al. 2 reported a postaneurysmal SAH shunting rate of only 6.9% in patients who had undergone lamina terminalis fenestration. In 2004 Andaluz and Zuccarello 3 prospectively studied the effect of lamina terminalis fenestration on shunt-dependent hydrocephalus in 106 patients with isolated ACoA aneurysms and Fisher Grade 3 SAH. These authors found the need for a shunt to be significantly reduced in the patients who had undergone fenestration compared with those who did not (4.25 vs 13.9%, respectively, p < 0.001). In 2004 Dehdashti and colleagues 6 prospectively investigated the influence of an aneurysmal SAH treatment modality, either surgical clip application or endovascular coil embolization, on the incidence of shunt-dependent hydrocephalus. In patients who had undergone clip application of anterior circulation aneurysms, the lamina terminalis was systematically fenestrated. Shunt-dependent hydrocephalus occurred in 14% of the surgical cases and 19% of the endovascular cases TABLE 3: Summary of characteristics in studies supporting the benefit of FLT Parameter Yonekawa et al., 1998 Tomasello et al., 1999 Schmieder et al., 1999 Komotar et al., 2002 Andaluz & Zuccarello, 2003 Andaluz et al., 2004 Akyuz & Tuncer, 2006 no. of patients * % female mean age (yrs) % patients w/ HCP on adm NA 10 NA 45 NA % patients w/ ant lesion location % patients w/ H&H Grade IV V % patients w/ Fisher Grade 3 4 SAH NA % patients w/ FLT overall shunt rate (FLT vs no FLT) vs 15 3 vs vs * Entire study population included. All ACoA location. No control group in study. J Neurosurg / Volume 111 / July

6 R. J. Komotar et al. TABLE 4: Summary of characteristics in studies refuting the benefit of FLT Parameter Sindou, 1994 Dehdashti et al., 2004 Kim et al., 2006 Komotar et al., 2008 no. of patients % female NA mean age (yrs) NA % patients w/ HCP on adm NA % patients w/ ant lesion location NA NA 100* 100 % patients w/ H&H Grade IV V % patients w/ Fisher Grade 3 4 SAH % patients w/ FLT 30 NA shunt rate (FLT vs no FLT) 11 vs vs vs vs 20 * All ACoA location. (no statistical significance at p = 0.53). In 2006 Akyuz and Tuncer 1 conducted a prospective study in which they fenestrated the Liliequist membrane as well as the lamina terminalis during surgery for ruptured ACoA aneurysms. These authors found no significant difference in either the occurrence of chronic hydrocephalus or a favorable outcome between the 2 groups (that is, those who underwent the procedures and those who did not), leading them to conclude that opening the Liliequist membrane should be pursued only in patients with cisternal overflow, ventricular dilation, and fourth ventricular clots. In 2006 Kim and colleagues 14 retrospectively analyzed the efficacy of lamina terminalis fenestration on the incidence of shuntdependent hydrocephalus following aneurysmal SAH due to ruptured ACoA aneurysms. They found no significant association between fenestration and the rate of shunt placement. Most recently, we completed a retrospective analysis in which we compared postsurgical outcomes in patients who had undergone fenestration of the lamina terminalis with those in patients who did not. 15 At our institution, all craniotomies for the surgical management of aneurysmal SAH are performed by a single neurosurgeon (E.S.C.), inherently controlling for the differences in surgical technique and postoperative management that can exist between surgeons. Our single-surgeon series revealed similar rates of shunt-dependent hydrocephalus in fenestrated and nonfenestrated cohorts, suggesting that microsurgical fenestration of the lamina terminalis may not reduce the incidence of shunt-dependent hydrocephalus following aneurysmal SAH. Note that none of the literature studies refuting the benefit of lamina terminalis fenestration was adequately powered to detect a difference in outcome (power = 80%, a = 0.05). To evaluate appropriate patient selection for lamina terminalis fenestration, we performed secondary analyses of patients at our institution who had undergone shunt placement for hydrocephalus after aneurysmal SAH. Not unexpectedly, multivariable analysis revealed the following risk factors for shunt placement: an age 60 years, a Hunt and Hess Grade IV or V, a Fisher Grade 3 or 4 SAH, and acute hydrocephalus on admission (Table 5). Patients that met all these criteria had a shunt placement rate of nearly 60%, whereas young patients with a good clinical grade, minimal hemorrhage, and no hydrocephalus on admission had a shunt placement rate of only 4% (Table 6). These high-risk individuals represent the cohort that would most benefit from lamina terminalis fenestration, if indeed the procedure is efficacious. Unfortunately, subgroup analysis of this cohort failed to show a diminished shunting rate in patients who had undergone lamina terminalis fenestration (Table 7). In short, the role of lamina terminalis fenestration in the operative management of aneurysmal SAH remains controversial. Thus, we performed the first systematic review and quantitative analysis of studies in the modern literature to determine the efficacy of this technique in reducing the incidence of aneurysmal SAH associated shunt-dependent hydrocephalus. In our review, we found 7 studies that supported the benefit of lamina terminalis fenestration and 4 studies that did not, with the overall analysis remaining inconclusive. An analysis of pooled data from studies supporting the benefit of lamina terminalis fenestration showed the incidence of shunt-depen- TABLE 5: Risk factors associated with the development of shuntdependent hydrocephalus after aneurysmal SAH, according to multivariable analysis* Variable p Value OR 95% CI patient age 60 yrs < H&H Grade IV or V < Fisher Grade 3 or acute HCP female sex NS lesion location NS lesion size NS FLT NS * Based on data from Komotar et al., Abbreviation: NS = not significant. 152 J Neurosurg / Volume 111 / July 2009

7 Microsurgical fenestration of the lamina terminalis TABLE 6: Patient characteristics associated with the ocurrence of shunt-dependent hydrocephalus after aneurysmal SAH* Variable Best Case Scenario Worst Case Scenario patient age < 60 yrs 60 yrs H&H grade I III IV or V Fisher grade 1 or 2 3 or 4 acute HCP absent present patient sex no difference no difference lesion location no difference no difference lesion size no difference no difference shunt rate 4% 58% * Analysis based on data from Komotar et al., dent hydrocephalus to be 6% compared with 12% in fenestrated versus nonfenestrated groups, respectively (1091 patients, p = 0.001). In contrast, an analysis of pooled data from studies refuting the benefit of fenestration showed the incidence of shunt-dependent hydrocephalus to be 17% compared with 16% in fenestrated versus nonfenestrated groups, respectively (882 patients, p = 0.74). Taken together (1973 patients), our results trend toward a lower incidence of shunt placement in patients who undergo lamina terminalis fenestration (10 vs 14%, p = 0.089). The interpretation of our results is limited in part TABLE 7: Multivariable analysis of factors predictive of shuntdependent hydrocephalus in patient subgroups* Patient Subgroup & Factor p Value OR patients w/ Fisher Grade 3 or 4 SAH patient age H&H grade < acute HCP FLT patients w/ H&H Grade IV or V patient age Fisher grade acute HCP FLT patients w/ age > 60 yrs H&H grade Fisher grade acute HCP FLT patients w/ adm HCP H&H grade Fisher grade patient age > 60 yrs FLT * Based on data from Komotar et al., J Neurosurg / Volume 111 / July 2009 by significant differences in patient demographic and clinical characteristics between the combined (from the various studies) fenestrated and nonfenestrated cohorts. Importantly, in the fenestrated cohort there were significantly more patients in poor clinical condition (Hunt and Hess Grade IV or V). However, given that poor clinical grade is a risk factor for the development of chronic hydrocephalus, 21 this cohort difference may in fact reinforce the efficacy of lamina terminalis fenestration. Other inherent limitations preclude more definitive conclusions. First, the existing literature describes only case series and case-control studies. Second, parent papers were underpowered and included poor or no control groups. Most notably, patients were never randomized to fenestration and multiple surgeons were involved, leading to selection bias and additional confounders. Third, study heterogeneity was considerable, as multiple centers presented various study designs, methodologies, management paradigms, and patient populations. In general, fenestration of the lamina terminalis is a relatively straightforward and safe procedure without associated long-term morbidity. At our institution, however, fenestration of the lamina terminalis in conjunction with intraoperative spinal drainage is believed in certain cases to be related to critical CSF hypovolemia and a dramatic herniation syndrome. 16,18 Third ventriculostomy, particularly in combination with lumbar drain placement, is a risk factor for the development of CSF hypovolemia by creating a direct conduit for CSF loss. Fortunately, rapid intervention obviates excessive long-term morbidity and death in these patients. 18 Conclusions Our systematic review showed no significant association between fenestration of the lamina terminalis and the incidence of shunt-dependent hydrocephalus. Note, however, that the interpretation of these results is restricted by unmatched cohort differences as well as other inherent study limitations. Although the overall literature supports this technique, a number of studies question its benefit, rendering the efficacy of lamina terminalis fenestration in reducing shunt-dependent hydrocephalus unclear. A well-designed, multicenter, randomized controlled trial is needed to resolve this controversy. Disclosure David K. Hahn and Christopher P. Kellner were supported in part by a Doris Duke Clinical Research Fellowship. References 1. Akyuz M, Tuncer R: The effects of fenestration of the interpeduncular cistern membrane arousted to the opening of lamina terminalis in patients with ruptured ACoA aneurysms: a prospective, comparative study. Acta Neurochir (Wien) 148: , Andaluz N, Van Loveren HR, Keller JT, Zuccarello M: Anatomic and clinical study of the orbitopterional approach to anterior communicating artery aneurysms. Neurosurgery 52: ,

8 R. J. Komotar et al. 3. Andaluz N, Zuccarello M: Fenestration of the lamina terminalis as a valuable adjunct in aneurysm surgery. Neurosurgery 55: , Auer LM, Mokry M: Disturbed cerebrospinal fluid circulation after subarachnoid hemorrhage and acute aneurysm surgery. Neurosurgery 26: , Borgbjerg BM, Gjerris F, Albeck MJ, Hauerberg J, Borgesen SV: A comparison between ventriculo-peritoneal and ventriculo-atrial cerebrospinal fluid shunts in relation to rate of revision and durability. Acta Neurochir (Wien) 140: , Dehdashti AR, Rilliet B, Rufenacht DA, de Tribolet N: Shuntdependent hydrocephalus after rupture of intracranial aneurysms: a prospective study of the influence of treatment modality. J Neurosurg 101: , Del Bigio MR, Bruni JE: Periventricular pathology in hydrocephalic rabbits before and after shunting. Acta Neuropathol 77: , Fode NC, Laws ER Jr, Sundt TM Jr: Communicating hydrocephalus after subarachnoid hemorrhage: results of shunt procedures. J Neurosurg Nurs 11: , Gjerris F, Borgesen SE, Sorensen PS, Boesen F, Schmidt K, Harmsen A, et al: Resistance to cerebrospinal fluid outflow and intracranial pressure in patients with hydrocephalus after subarachnoid hemorrhage. Acta Neurochir (Wien) 88:79 86, Grant JA, McLone DG: Third ventriculostomy: a review. Surg Neurol 47: , Gruber A, Reinprecht A, Bavinzski G, Czech T, Richling B: Chronic shunt-dependent hydrocephalus after early surgical and early endovascular treatment of ruptured intracranial aneurysms. Neurosurgery 44: , Joakimsen O, Mathiesen EB, Monstad P, Selseth B: CSF hydrodynamics after subarachnoid hemorrhage. Acta Neurol Scand 75: , Katzman R: Normal pressure hydrocephalus, in Wells CE (ed) Dementia, ed 2. Philadelphia: FA Davis, 1977, pp Kim JM, Jeon JY, Kim JH, Cheong JH, Bak KH, Kim CH, et al: Influence of lamina terminalis fenestration on the occurrence of the shunt-dependent hydrocephalus in anterior communicating artery aneurysmal subarachnoid hemorrhage. J Korean Med Sci 21: , Komotar RJ, Hahn DK, Kim GH, Khandji J, Mocco J, Mayer SA, et al: The impact of microsurgical fenestration of the lamina terminalis on shunt-dependent hydrocephalus and vasospasm following aneurysmal subarachnoid hemorrhage. Neurosurgery 62: , Komotar RJ, Mocco J, Ransom ER, Mack WJ, Zacharia BE, Wilson DA, et al: Herniation secondary to critical postcraniotomy cerebrospinal fluid hypovolemia. Neurosurgery 57: , Komotar RJ, Olivi A, Rigamonti D, Tamargo RJ: Microsurgical fenestration of the lamina terminalis reduces the incidence of shunt-dependent hydrocephalus after aneurysmal subarachnoid hemorrhage. Neurosurgery 51: , Komotar RJ, Ransom ER, Mocco J, Zacharia BE, McKhann GM II, Mayer SA, et al: Critical postcraniotomy cerebrospinal fluid hypovolemia: risk factors and outcome analysis. Neurosurgery 59: , Kosteljanetz M: CSF dynamics in patients with subarachnoid and/or intraventricular hemorrhage. J Neurosurg 60: , Lam CH, Villemure JG: Comparison between ventriculoatrial and ventriculoperitoneal shunting in the adult population. Br J Neurosurg 11:43 48, Schmieder K, Koch R, Lucke S, Harders A: Factors influencing shunt dependency after aneurysmal subarachnoid haemorrhage. Zentralbl Neurochir 60: , Sindou M: Favourable influence of opening the lamina terminalis and Lilliequist's membrane on the outcome of ruptured intracranial aneurysms. A study of 197 consecutive cases. Acta Neurochir (Wien) 127:15 16, Tomasello F, d'avella D, de Divitiis O: Does lamina terminalis fenestration reduce the incidence of chronic hydrocephalus after subarachnoid hemorrhage? Neurosurgery 45: , Torvik A, Bhatia R, Murthy VS: Transitory block of the arachnoid granulations following subarachnoid haemorrhage. A postmortem study. Acta Neurochir (Wien) 41: , Udvarhelyi GB, Wood JH, James AE Jr, Bartelt D: Results and complications in 55 shunted patients with normal pressure hydrocephalus. Surg Neurol 3: , Yonekawa Y, Imhof HG, Ogata N, Bernays R, Kaku Y, Fandino J, et al: Aneurysm surgery in the acute stage: results of structured treatment. Neurol Med Chir (Tokyo) 38:45 49, 1998 Manuscript submitted March 4, Accepted January 1, Please include this information when citing this paper: published online March 13, 2009; DOI: / JNS0821. Address correspondence to: Ricardo J. Komotar, M.D., De partment of Neurosurgery, Columbia University, 710 West 168th Street, Room 431, New York, New York rjk2103@ columbia.edu. 154 J Neurosurg / Volume 111 / July 2009