Neurocrit Care DOI 10.1007/s12028-015-0152-7
ORIGINAL ARTICLE
Factors Associated with Acute and Chronic Hydrocephalus in Nonaneurysmal Subarachnoid Hemorrhage Peter Kang1 • Amanda Raya1 • Gregory J. Zipfel2 • Rajat Dhar1
Ó Springer Science+Business Media New York 2015
Abstract Background Hydrocephalus requiring external ventricular drain (EVD) or shunt placement commonly complicates aneurysmal subarachnoid hemorrhage (SAH), but its frequency is not as well known for nonaneurysmal SAH (NASAH). Those with diffuse bleeding may have greater risk of hydrocephalus compared to those with a perimesencephalic pattern. We evaluated the frequency of hydrocephalus in NA-SAH and whether imaging factors could predict the need for EVD and shunting. Methods We collected admission clinical and imaging variables for 105 NA-SAH patients, including bicaudate index (BI), Hijdra sum score (HSS), intraventricular hemorrhage (IVH) score, modified Fisher scale (mFS), and bleeding pattern. Hydrocephalus was categorized as acute (need for EVD) or chronic (shunt). We applied logistic regression to determine whether hydrocephalus risk was independently related to bleeding pattern or mediated through blood volume or ventriculomegaly. Results Acute hydrocephalus was seen in 26 (25 %) patients but was more common with diffuse (15/28, 54 %) versus perimesencephalic (10/59, 17 %, p < 0.001) bleeding. Patients developing acute hydrocephalus had worse clinical grade and higher BI, HSS, and IVH scores. Adjusting the relationship between hydrocephalus and diffuse bleeding
& Rajat Dhar [email protected] 1
Neurocritical Care Section, Department of Neurology, Washington University School of Medicine, 660 S. Euclid Ave, Campus Box 8111, St. Louis, MO 63110, USA
2
Department of Neurological Surgery, Washington University School of Medicine, 660 S. Euclid Ave, Campus Box 8057, St. Louis, MO 63110, USA
for HSS (but not BI) nullified this association. Nine (35 %) patients requiring EVD eventually required shunting for chronic hydrocephalus, which was associated with greater blood burden but not poor clinical grade. Conclusion Acute hydrocephalus occurs in one-quarter of NA-SAH patients. The greater risk in diffuse bleeding appears to be mediated by greater cisternal blood volume but not by greater ventriculomegaly. Imaging characteristics may aid in anticipatory management of hydrocephalus in NA-SAH. Keywords Subarachnoid hemorrhage Nonaneurysmal Angiogram negative Hydrocephalus External ventricular drain Ventriculoperitoneal shunt
Introduction Spontaneous, nontraumatic subarachnoid hemorrhage (SAH) is most often due to rupture of an intracranial aneurysm. Despite advances in modern neurovascular imaging, nonaneurysmal SAH (NA-SAH) still accounts for roughly 15 % of cases [1]. In contrast to aneurysmal SAH, it is thought that NA-SAH represents a more benign entity with less morbidity and mortality, especially in the context of a perimesencephalic SAH (PM-SAH) pattern of bleeding [2–6]. Recent studies, however, have questioned whether risk of complications such as vasospasm or hydrocephalus after NA-SAH is actually minimal, especially for those with a diffuse, aneurysmal-like pattern of bleeding [7–10]. It remains unclear whether this variability in clinical course is explained by differing pathophysiologic processes underlying these two types of bleeding, or simply mediated by the fact that those with diffuse bleeding tend to harbor more cisternal and intraventricular blood.
123
Neurocrit Care
One significant early complication of SAH is the development of acute hydrocephalus, which may occur in as many as half of those with aneurysmal SAH [11]. This may be associated with increased intracranial pressure (ICP) and neurological deterioration, requiring external ventricular drain (EVD) placement. A small subset of these patients, roughly 11–20 % [12, 13] will progress to develop chronic hydrocephalus requiring placement of a shunt for long-term cerebrospinal fluid (CSF) diversion [14]. EVD and shunt placement are both invasive interventions that carry morbidity and have important implications for the acute and long-term monitoring and management of SAH patients. A number of studies have investigated the frequency and factors associated with hydrocephalus in aneurysmal SAH [12, 14, 15]. However, there is a relative dearth of knowledge on risk factors for acute and chronic hydrocephalus in NA-SAH. It is unclear whether the incidence of hydrocephalus is truly lower in NA-SAH compared to SAH with aneurysmal etiologies, especially NA-SAH in the context of diffuse (aneurysmal-like) bleeding pattern. Furthermore, it unclear whether there is something intrinsic to the pathophysiology of diffuse bleeding that confers a greater risk of hydrocephalus in comparison to PM-SAH [9, 16, 17]. However, it is not clear whether such a relationship is mediated through the increased burden of cisternal and intraventricular blood that may be associated with diffuse bleeding. The purpose of this study is firstly to characterize the incidence of both acute and chronic hydrocephalus in NASAH patients. Second, we sought to determine risk factors for hydrocephalus; specifically whether pattern of bleeding independently predicts need for EVD and ventriculoperitoneal shunt (VPS) placement distinct from associations with greater blood volume. We hypothesized that the previously reported link between diffuse bleeding pattern and hydrocephalus in NA-SAH may be mediated primarily through greater blood volume.
Methods Participants We prospectively identified all patients who presented to our institution with nontraumatic SAH between 2005 and 2013 who had negative angiographic evaluation. All such patients underwent computed tomography angiography (CTA) within 48 h and repeat digital subtraction angiography (DSA) at 1-week post-ictus. Those who were subsequently found to have an aneurysm or other etiology of bleeding were excluded from the current analysis. The Washington University Human Studies
123
Committee approved the collection of data for this analysis.
Management All patients presenting to our institution with the diagnosis of SAH, confirmed by head CT or lumbar puncture, were admitted to the Neurology/Neurosurgery Intensive Care Unit (NNICU) for stabilization and management, including monitoring with hourly neurologic evaluations. A repeat head CT was obtained for clinical deterioration to evaluate for complications such as rebleeding or hydrocephalus. Serial neuroimaging was not routinely done in asymptomatic patients. Hydrocephalus was identified by symptoms of elevated ICP (e.g., headache, altered sensorium) along with consistent findings on head CT and was managed with placement of an EVD. Weaning of the EVD occurred only after radiographic intraventricular blood and concern for vasospasm had resolved. This involved raising the height of the EVD chamber and then clamping the drainage for a period of no less than 24 h, during which ICP and clinical status was closely followed. A repeat head CT was obtained at 24 h to evaluate for increase in ventricular size. If serial attempts to wean the EVD were unsuccessful, VPS was placed for long-term CSF diversion, at the discretion of the attending neurosurgeon.
Measures For each participant, demographic and baseline health data were collected including age, gender, race, smoking status, and the presence of premorbid coronary artery disease, diabetes mellitus, and hypertension. The severity of the presenting SAH was determined using the World Federation of Neurosurgical Societies (WFNS) grading system and dichotomized into low grade (WFNS grade I–III) and high grade (WFNS grade IV–V) [18]. Bicaudate index (BI) was used as a measure of radiographic ventriculomegaly and represents the ratio of the distance between the bilateral caudate heads to the distance between the tables of the skull [19]. BI was analyzed both as a continuous variable as well as dichotomized into a cutoff of BI 4 provided 93 % sensitivity and 48 % specificity. In contrast, BI at a cutoff of 0.2 was more specific (90 %) and less sensitive (63 %). In diffuse pattern bleeding only, AUC was 0.7 (p = 0.06) with an optimal HSS threshold of 11 (80 % sensitive, 71 % specific). When all significant univariate predictors from our primary analyses (pattern, HSS, BI, mFS, IVH score) were entered into logistic regression model with acute hydrocephalus as the outcome variable, only HSS (OR 1.25 per point, 95 % CI 1.11–1.42, p < 0.0001) and BI C0.20 (OR 11.2, 95 % CI 3.4–36.8, p < 0.0001) remained independent contributors to the model (R2 = 0.51, correctly classified 86 % of cases). All those with poor WFNS grade required acute EVD placement and so this variable could not be entered as a discriminating predictor in our final model. Progression of Acute to Chronic Hydrocephalus Within the patients who had an EVD placed, those with poor WFNS grade were less likely to require shunt placement (1 of 8 vs. 8 of 18, p = 0.19) though this did not reach statistical significance (see Table 1). Notably, this was not related to higher mortality in this subset of patients as none of these patients expired prior to discharge from hospitalization. Those with diffuse bleeding who had
123
Neurocrit Care Table 1 Clinical and neuroimaging characteristics in our cohort of NA-SAH patients by EVD and shunt status (within EVD group) No EVD (79/105)
EVD (26/105)
p value
No shunt (17/26)
Shunt (9/26)
p value
Demographic Age
51.7 (12.6)
55.7 (11.2)
0.17
57.2 (12)
52.8 (9.34)
0.34
Gender (female)
49 (62 %)
17 (65.4 %)
0.76
11 (64.7 %)
6 (66.7 %)
0.92
Caucasian
64 (81 %)
21 (80.8 %)
14 (82.4 %)
7 (77.8 %)
African-American
14 (17.7 %)
5 (19.2 %)
3 (17.6 %)
2 (22.2 %)
Asian
1 (1.3 %)
0 (0 %)
0 (0 %)
0 (0 %)
Nonsmoker
46 (58.2 %)
9 (34.6 %)
6 (35.3 %)
3 (33.3 %)
Current smoker
16 (20.3 %)
9 (34.6 %)
6 (35.3 %)
3 (33.3 %)
17 (21.5 %)
8 (30.8 %)
5 (29.4 %)
3 (33.3 %)
Coronary artery disease
8 (10.1 %)
1 (3.8 %)
0.32
1 (5.9 %)
0 (0 %)
0.46
Diabetes mellitus
11 (13.9 %)
4 (15.4 %)
0.85
4 (23.5 %)
0 (0 %)
0.11
Hypertension
37 (46.8 %)
15 (57.7 %)
0.38
9 (52.9 %)
6 (66.7 %)
Race
0.84
Smoking Status
Ex-smoker Comorbidities
0.78
0.11
Clinical
0.98
ORIGINAL ARTICLE
Factors Associated with Acute and Chronic Hydrocephalus in Nonaneurysmal Subarachnoid Hemorrhage Peter Kang1 • Amanda Raya1 • Gregory J. Zipfel2 • Rajat Dhar1
Ó Springer Science+Business Media New York 2015
Abstract Background Hydrocephalus requiring external ventricular drain (EVD) or shunt placement commonly complicates aneurysmal subarachnoid hemorrhage (SAH), but its frequency is not as well known for nonaneurysmal SAH (NASAH). Those with diffuse bleeding may have greater risk of hydrocephalus compared to those with a perimesencephalic pattern. We evaluated the frequency of hydrocephalus in NA-SAH and whether imaging factors could predict the need for EVD and shunting. Methods We collected admission clinical and imaging variables for 105 NA-SAH patients, including bicaudate index (BI), Hijdra sum score (HSS), intraventricular hemorrhage (IVH) score, modified Fisher scale (mFS), and bleeding pattern. Hydrocephalus was categorized as acute (need for EVD) or chronic (shunt). We applied logistic regression to determine whether hydrocephalus risk was independently related to bleeding pattern or mediated through blood volume or ventriculomegaly. Results Acute hydrocephalus was seen in 26 (25 %) patients but was more common with diffuse (15/28, 54 %) versus perimesencephalic (10/59, 17 %, p < 0.001) bleeding. Patients developing acute hydrocephalus had worse clinical grade and higher BI, HSS, and IVH scores. Adjusting the relationship between hydrocephalus and diffuse bleeding
& Rajat Dhar [email protected] 1
Neurocritical Care Section, Department of Neurology, Washington University School of Medicine, 660 S. Euclid Ave, Campus Box 8111, St. Louis, MO 63110, USA
2
Department of Neurological Surgery, Washington University School of Medicine, 660 S. Euclid Ave, Campus Box 8057, St. Louis, MO 63110, USA
for HSS (but not BI) nullified this association. Nine (35 %) patients requiring EVD eventually required shunting for chronic hydrocephalus, which was associated with greater blood burden but not poor clinical grade. Conclusion Acute hydrocephalus occurs in one-quarter of NA-SAH patients. The greater risk in diffuse bleeding appears to be mediated by greater cisternal blood volume but not by greater ventriculomegaly. Imaging characteristics may aid in anticipatory management of hydrocephalus in NA-SAH. Keywords Subarachnoid hemorrhage Nonaneurysmal Angiogram negative Hydrocephalus External ventricular drain Ventriculoperitoneal shunt
Introduction Spontaneous, nontraumatic subarachnoid hemorrhage (SAH) is most often due to rupture of an intracranial aneurysm. Despite advances in modern neurovascular imaging, nonaneurysmal SAH (NA-SAH) still accounts for roughly 15 % of cases [1]. In contrast to aneurysmal SAH, it is thought that NA-SAH represents a more benign entity with less morbidity and mortality, especially in the context of a perimesencephalic SAH (PM-SAH) pattern of bleeding [2–6]. Recent studies, however, have questioned whether risk of complications such as vasospasm or hydrocephalus after NA-SAH is actually minimal, especially for those with a diffuse, aneurysmal-like pattern of bleeding [7–10]. It remains unclear whether this variability in clinical course is explained by differing pathophysiologic processes underlying these two types of bleeding, or simply mediated by the fact that those with diffuse bleeding tend to harbor more cisternal and intraventricular blood.
123
Neurocrit Care
One significant early complication of SAH is the development of acute hydrocephalus, which may occur in as many as half of those with aneurysmal SAH [11]. This may be associated with increased intracranial pressure (ICP) and neurological deterioration, requiring external ventricular drain (EVD) placement. A small subset of these patients, roughly 11–20 % [12, 13] will progress to develop chronic hydrocephalus requiring placement of a shunt for long-term cerebrospinal fluid (CSF) diversion [14]. EVD and shunt placement are both invasive interventions that carry morbidity and have important implications for the acute and long-term monitoring and management of SAH patients. A number of studies have investigated the frequency and factors associated with hydrocephalus in aneurysmal SAH [12, 14, 15]. However, there is a relative dearth of knowledge on risk factors for acute and chronic hydrocephalus in NA-SAH. It is unclear whether the incidence of hydrocephalus is truly lower in NA-SAH compared to SAH with aneurysmal etiologies, especially NA-SAH in the context of diffuse (aneurysmal-like) bleeding pattern. Furthermore, it unclear whether there is something intrinsic to the pathophysiology of diffuse bleeding that confers a greater risk of hydrocephalus in comparison to PM-SAH [9, 16, 17]. However, it is not clear whether such a relationship is mediated through the increased burden of cisternal and intraventricular blood that may be associated with diffuse bleeding. The purpose of this study is firstly to characterize the incidence of both acute and chronic hydrocephalus in NASAH patients. Second, we sought to determine risk factors for hydrocephalus; specifically whether pattern of bleeding independently predicts need for EVD and ventriculoperitoneal shunt (VPS) placement distinct from associations with greater blood volume. We hypothesized that the previously reported link between diffuse bleeding pattern and hydrocephalus in NA-SAH may be mediated primarily through greater blood volume.
Methods Participants We prospectively identified all patients who presented to our institution with nontraumatic SAH between 2005 and 2013 who had negative angiographic evaluation. All such patients underwent computed tomography angiography (CTA) within 48 h and repeat digital subtraction angiography (DSA) at 1-week post-ictus. Those who were subsequently found to have an aneurysm or other etiology of bleeding were excluded from the current analysis. The Washington University Human Studies
123
Committee approved the collection of data for this analysis.
Management All patients presenting to our institution with the diagnosis of SAH, confirmed by head CT or lumbar puncture, were admitted to the Neurology/Neurosurgery Intensive Care Unit (NNICU) for stabilization and management, including monitoring with hourly neurologic evaluations. A repeat head CT was obtained for clinical deterioration to evaluate for complications such as rebleeding or hydrocephalus. Serial neuroimaging was not routinely done in asymptomatic patients. Hydrocephalus was identified by symptoms of elevated ICP (e.g., headache, altered sensorium) along with consistent findings on head CT and was managed with placement of an EVD. Weaning of the EVD occurred only after radiographic intraventricular blood and concern for vasospasm had resolved. This involved raising the height of the EVD chamber and then clamping the drainage for a period of no less than 24 h, during which ICP and clinical status was closely followed. A repeat head CT was obtained at 24 h to evaluate for increase in ventricular size. If serial attempts to wean the EVD were unsuccessful, VPS was placed for long-term CSF diversion, at the discretion of the attending neurosurgeon.
Measures For each participant, demographic and baseline health data were collected including age, gender, race, smoking status, and the presence of premorbid coronary artery disease, diabetes mellitus, and hypertension. The severity of the presenting SAH was determined using the World Federation of Neurosurgical Societies (WFNS) grading system and dichotomized into low grade (WFNS grade I–III) and high grade (WFNS grade IV–V) [18]. Bicaudate index (BI) was used as a measure of radiographic ventriculomegaly and represents the ratio of the distance between the bilateral caudate heads to the distance between the tables of the skull [19]. BI was analyzed both as a continuous variable as well as dichotomized into a cutoff of BI 4 provided 93 % sensitivity and 48 % specificity. In contrast, BI at a cutoff of 0.2 was more specific (90 %) and less sensitive (63 %). In diffuse pattern bleeding only, AUC was 0.7 (p = 0.06) with an optimal HSS threshold of 11 (80 % sensitive, 71 % specific). When all significant univariate predictors from our primary analyses (pattern, HSS, BI, mFS, IVH score) were entered into logistic regression model with acute hydrocephalus as the outcome variable, only HSS (OR 1.25 per point, 95 % CI 1.11–1.42, p < 0.0001) and BI C0.20 (OR 11.2, 95 % CI 3.4–36.8, p < 0.0001) remained independent contributors to the model (R2 = 0.51, correctly classified 86 % of cases). All those with poor WFNS grade required acute EVD placement and so this variable could not be entered as a discriminating predictor in our final model. Progression of Acute to Chronic Hydrocephalus Within the patients who had an EVD placed, those with poor WFNS grade were less likely to require shunt placement (1 of 8 vs. 8 of 18, p = 0.19) though this did not reach statistical significance (see Table 1). Notably, this was not related to higher mortality in this subset of patients as none of these patients expired prior to discharge from hospitalization. Those with diffuse bleeding who had
123
Neurocrit Care Table 1 Clinical and neuroimaging characteristics in our cohort of NA-SAH patients by EVD and shunt status (within EVD group) No EVD (79/105)
EVD (26/105)
p value
No shunt (17/26)
Shunt (9/26)
p value
Demographic Age
51.7 (12.6)
55.7 (11.2)
0.17
57.2 (12)
52.8 (9.34)
0.34
Gender (female)
49 (62 %)
17 (65.4 %)
0.76
11 (64.7 %)
6 (66.7 %)
0.92
Caucasian
64 (81 %)
21 (80.8 %)
14 (82.4 %)
7 (77.8 %)
African-American
14 (17.7 %)
5 (19.2 %)
3 (17.6 %)
2 (22.2 %)
Asian
1 (1.3 %)
0 (0 %)
0 (0 %)
0 (0 %)
Nonsmoker
46 (58.2 %)
9 (34.6 %)
6 (35.3 %)
3 (33.3 %)
Current smoker
16 (20.3 %)
9 (34.6 %)
6 (35.3 %)
3 (33.3 %)
17 (21.5 %)
8 (30.8 %)
5 (29.4 %)
3 (33.3 %)
Coronary artery disease
8 (10.1 %)
1 (3.8 %)
0.32
1 (5.9 %)
0 (0 %)
0.46
Diabetes mellitus
11 (13.9 %)
4 (15.4 %)
0.85
4 (23.5 %)
0 (0 %)
0.11
Hypertension
37 (46.8 %)
15 (57.7 %)
0.38
9 (52.9 %)
6 (66.7 %)
Race
0.84
Smoking Status
Ex-smoker Comorbidities
0.78
0.11
Clinical
0.98
