- Email: [email protected]
Decreased CSF output as a clinical indicator of cerebral vasospasm following aneurysmal subarachnoid hemorrhage
Clinical Neurology and Neurosurgery 144 (2016) 101–104
Contents lists available at ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Decreased CSF output as a clinical indicator of cerebral vasospasm following aneurysmal subarachnoid hemorrhage Christine Hammer, Badih Daou, Nohra Chalouhi, Robert M. Starke, Lina Ya’qoub, Nikolaos Mouchtouris, Sravanthi Koduri, Stavropoula Tjoumakaris, Robert H. Rosenwasser, Pascal Jabbour ∗ Department of Neurological Surgery, Thomas Jefferson University and Jefferson Hospital for Neuroscience, Philadelphia, PA, USA
a r t i c l e
i n f o
Article history: Received 15 January 2016 Received in revised form 22 February 2016 Accepted 27 March 2016 Available online 28 March 2016 Keywords: CSF output Subarachnoid hemorrhage Vasospasm
a b s t r a c t Objective: Vasospasm is a significant cause of morbidity and mortality among those with aneurysmal subarachnoid hemorrhage (aSAH). Treating increased intracranial pressure by drainage of cerebral spinal fluid through an external ventriculostomy is routine practice. The objective of this study is to evaluate the trends of CSF output in patients who experience vasospasm. Methods: Electronic medical charts were reviewed to identify two groups of patients with aSAH, 75 consecutive patients who developed vasospasm and 75 matched patients who did not develop vasospasm. CSF output was recorded within 3 days before and 3 days after the occurrence of vasospasm. CSF output was recorded for the same days after SAH in matched patients with no vasospasm. Results: Total CSF output was lower in patients with vasospasm as compared to patients without vasospasm matched for the same day (p < 0.001). In patients with vasospasm, CSF output recordings were significantly higher prior to the occurrence of vasospasm (438 ml/day) than the period following vasospasm (325.7 ml/day), with a consistent decrease in CSF drainage from day 3 before vasospasm to day 3 after vasospasm (p = 0.012). Decreasing CSF output was significantly associated with the occurrence of vasospasm (p = 0.017). Youden indices demonstrated that daily CSF drainage <160 ml was significantly associated with the occurrence of vasospasm. The sensitivity of this test was 64.79% and the specificity was 55.38%. Conclusions: In addition to clinical exam findings, observation of a CSF output decline to less than 160 ml/day may be used as additional support for the diagnosis of vasospasm. © 2016 Published by Elsevier B.V.
1. Introduction A subarachnoid hemorrhage (SAH) is the release of blood into the subarachnoid space. The most common causes include hemorrhage from a ruptured aneurysm, an arteriovenous malformation (AVM) or traumatic injury such as blunt force injury to the skull. Immediate life threatening consequences of SAH occur from the clotting and hemolysis of blood in the subarachnoid space and the concomitant hydrocephalus that often occurs leading to increased
Abbreviations: aSAH, aneurysmal subarachnoid hemorrhage; AVM, arteriovenous malformation; CBF, cerebral blood flow; CSF, cerebrospinal fluid; DCI, delayed cerebral ischemia; EVD, external ventricular drain; ICP, intracranial pressure; PTA, percutaneous transluminal angioplasty; TCD, transcranial doppler. ∗ Corresponding author at: Division of Neurovascular Surgery and Endovascular Neurosurgery, Department of Neurosurgery, Thomas Jefferson University Hospital, 901Walnut Street, Philadelphia, PA 19107, USA. E-mail addresses: [email protected], [email protected] (P. Jabbour). http://dx.doi.org/10.1016/j.clineuro.2016.03.024 0303-8467/© 2016 Published by Elsevier B.V.
intracranial pressure (ICP) [1–4]. As a result of the increase in ICP, an external ventricular drain (EVD) is placed in SAH patients to allow for drainage of CSF and maintenance of a normal ICP [2]. Cerebral vasospasm is the delayed-onset narrowing of arteries following SAH and is a significant cause of morbidity and mortality in these patients [5]. It has been shown that facilitating the drainage of CSF leads to faster clot evacuation and subsequently a decreased incidence of vasospasm [2]. Vasospasm occurs approximately three days after SAH, peaks between six to eight days and is usually resolved within twelve days [6]. The onset of vasospasm can include symptoms such as increasing headache, progressive confusion and delirium, weakness or lethargy [7]. About one third of patients with aneurysmal SAH will experience clinical vasospasm and approximately one half to two thirds of aneurysmal subarachnoid hemorrhage patients will have angiographic vasospasm. The incidence of cerebral vasospasm is reported to be between 40 and 70% in patients with SAH with 30% of those patients developing delayed cerebral ischemia (DCI), making an adverse outcome more likely [8]. As a result of DCI, patients may display clinical
102
C. Hammer et al. / Clinical Neurology and Neurosurgery 144 (2016) 101–104
symptoms such as hemiparesis, aphasia, or a decrease in the level of consciousness [9]. Due to the severe complications of vasospasm, Transcranial Doppler (TCD) ultrasound is often used to monitor cerebral blood flow and velocity in SAH patients [10]. The strong correlation between flow velocities in cerebral arteries and cerebral vasospasm severity after SAH makes TCD ultrasound a valuable tool in monitoring vasospasm [10]. Currently, there are three widely-used treatments for vasospasm: percutaneous transluminal angioplasty (PTA) [11], intra-arterial infusion of vasodilators [12,13], and induced hypertension [14]. Previous studies have shown endovascular therapy to have a remarkable safety-efficacy profile with both PTA and nicardipine, a calcium channel antagonist, being equally effective [5]. Induced hypertension involves the elevation of mean arterial pressure with vasopressors and has been shown to be effective in increasing cerebral blood flow (CBF) [4]. At this time, it is difficult to predict which patients will develop vasospasm and when. However, our findings suggest that there is a correlation between CSF output trends and vasospasm. The influence of such a correlation would allow for pre-emptive treatment and decrease the adverse effects of vasospasm. In this study, we evaluated the CSF output trends before and after patients were treated for cerebral vasospasm in order to determine if CSF output can be used as a clinical indicator for impending vasospasm. 2. Materials and methods 2.1. Patient selection and outcomes The study protocol was approved by the University Institutional Review Board. A patient list of consecutive patients with aneurysmal subarachnoid hemorrhage (aSAH) was generated based on ICD-9 codes. Electronic medical charts were reviewed to identify two groups of patients with aSAH: patients who developed clinical vasospasm and patients who did not develop clinical vasospasm post aSAH. Clinical vasospasm was defined as a worsening in neurologic status that could not be attributed to any other factor including rebleeding, hydrocephalus, intracerebral hematoma and metabolic factors. The presence of arterial vasospasm was confirmed with a cerebral angiogram in all patients as evidence of >30% arterial luminal narrowing. 116 consecutive patients with clinical vasospasm post aSAH were identified between January 2007 and December 2012. All patients had an EVD. Initially, the drainage is set at 15–20 cm H2O. Hourly CSF outputs were collected and combined to determine the daily CSF output. Daily CSF output levels were evaluated within three days prior to the occurrence of clinical vasospasm and within three days after vasospasm. Patients with missing CSF output levels were excluded from the study (n = 42). 75 patients with clinical vasospasm meeting the study criteria were included. Blinded to the outcomes of the study, 75 patients with aSAH who did not develop clinical vasospasm were matched in a 1:1 fashion to patients with vasospasm based on patient age, gender, Hunt and Hess, and Fisher grades. Hunt and Hess and Fisher grades were determined by the treating neurosurgeon. The primary outcome of this study was to evaluate trends of CSF output in patients who develop vasospasm after aSAH and identify any potential associations between CSF drainage and the occurrence of vasospasm. Secondary outcomes included (1) sensitivity and specificity analysis of CSF output in predicting the occurrence of clinical vasospasm and (2) an evaluation of other factors that may be associated with the occurrence of vasospasm.
Table 1 Baseline characteristics.
Mean patient age Women Men Smoking Hypertension
Vasospasm
No vasospasm
P value
54 70.7% 29.3% 33.3% 44%
53.4 70.7% 29.3% 63.5% 49%
0.72 1 1 <0.001 0.52
Table 2 Hunt and Hess grades.
1 2 3 4 5
Vasospasm
No vasospasm
Percent
2 7 40 24 2
2 7 39 25 2
2.7 9.3 53.3 32 2.7
Vasospasm
No vasospasm
Percent
0 9 12 54
0 9 12 54
0 12 16 72
Table 3 Fisher grades.
1 2 3 4
not experience vasospasm were matched based age, gender, H&H and Fisher grades and days of CSF output measurement blinded to all other outcomes. Analysis was carried out using Wilcoxon paired rank sum test and McNemar’s test as appropriate. Univariate conditional (matched) analysis was used to test covariates predictive of vasospasm. The following factors were tested: age, sex, smoking, hypertension, Hunt and Hess grade, Fisher grade, aneurysm location and CSF output. Interaction and confounding was assessed through stratification and relevant expansion covariates. Factors predictive in univariate analysis (p < 0.20) [15] were entered into a multivariate conditional logistic regression analysis. Regression models were assessed using area under the receiver operating characteristic curve (AUC). Youden Indices were calculated to determine cutoffs for the dichotomized continuous variable CSF output that yielded the optimal discrimination of vasospasm (sensitivity, specificity, positive predictive value, negative predictive value). P-values of ≤0.05 were considered statistically significant. Statistical analysis was carried out with Stata 10.0 (College Station, TX) [15]. 3. Results 3.1. Baseline characteristics The mean patient age was 54 years. Baseline characteristics of patients with and without vasospasm are detailed in Table 1. 70.7% of patients were women and 29.3% men. The only significant difference between the two groups was a higher percentage of smokers in the no vasospasm group (p < 0.001). The majority of patients had Hunt and Hess grade 3 (53%) (Table 2). SAH was Fisher grade 2 in 12% of patients, Fisher grade 3 in 16% and Fisher grade 4 in 72% (Table 3). In patients with vasospasm, 17.3% of aneurysms were located in the posterior circulation versus 14.8% in patients without vasospasm (P = 0.68). Aneurysm locations are detailed in Table 4.
2.2. Statistical analysis
3.2. CSF output
Data are presented as mean and range for continuous variables, and as frequency for categorical variables. Patients who did and did
CSF output was recorded within 3 days before and 3 days after the occurrence of vasospasm (Table 5). CSF output was recorded for
C. Hammer et al. / Clinical Neurology and Neurosurgery 144 (2016) 101–104
103
Table 4 Aneurysm location.
ACOM PCOM Basilar PICA ICA MCA CO Vertebral Anterior choroidal AICA ACA
Vasospasm
No vasospasm
34 11 3 4 6 7 1 5 1 1 2
30 18 7 3 5 4 2 1 0 0 5
the same days after SAH in matched patients with no vasospasm. The average CSF output day 3 pre-vasospasm was 159 ml/day and in patients who did not have vasospasm matched for the same day it was 184.5 ml/day (p = 0.14). Mean CSF output day 2 pre-vasospasm was 140.8 ml/day and 189.5 ml/day in patients without vasospasm (p = 0.005). Average CSF output day 1 prevasospasm was 128.4 ml/day versus 175.3 ml/day in patients with no vasospasm matched for same days of CSF recordings (p = 0.003). Mean CSF output on the day of vasospasm was 121.6 ml/day. In patients who were matched for the same day after SAH, mean CSF output was 179.3 ml/day (p = 0.0001). Total CSF output was lower in patients with vasospasm as compared to patients without vasospasm matched for the same day (p < 0.001). Furthermore, in patients with vasospasm, CSF output recordings were significantly higher prior to the occurrence of vasospasm (438 ml/day) than the period following vasospasm (325.7 ml/day), with a consistent decrease in CSF drainage from day 3 before the occurrence of vasospasm to day 3 after vasospasm (p = 0.012). In univariate analysis, the only significant factor associated with the occurrence of vasospasm was decreasing CSF output, with the significance increasing from day 3 to day 1 pre-vasospasm (D3 pre-vasospasm: OR = 1.002, p = 0.1, [0.999–1.006], D2 pre-vasospasm: 1.004, p = 0.007, [1.001–1.008], D1 pre-vasospasm: OR = 1.006, p = 0.004, [1.002–1.009], D1–D3 pre-vasospasm: OR = 1.002, p = 0.017, [1.001–1.003]). Similarly in multivariate analysis, CSF outputs at D1 and D2 pre-vasospasm were significantly associated with the occurrence of vasospasm after controlling for age, gender, Fisher grade, Hunt and Hess grade and smoking history (OR d1 = 1.005, p = 0.017, [1.001–1.009], OR d2 = OR = 1.004, p = 0.04, [1.001–1.007]). CSF output D3 prevasospasm was less predictive (OR = 1.001, p = 0.35, [1.001–1.005]). Sensitivity of monitoring CSF output in predicting clinical vasospasm was 64.79%, specificity was 55.38%, positive predictive value was 61.33% and negative predictive value was 59.02% (Fig. 1). Youden indices demonstrated that CSF output <160 ml/day and ≥160 provided the best dichotomized breakpoint (p = 0.018), with daily CSF drainage less than 160 being significantly associated with the occurrence of vasospasm.
Fig 1. Sensitivity and specificity of CSF output <160 ml/day in predicting vasospasm.
4. Discussion We have presented several unique results that suggest that the CSF output trends may be significant in the indication if not the prediction of vasospasm in patients with aSAH. First, the decline in the average CSF output on day 3 pre-vasospasm to less than 160 ml/day may be used as additional support for the diagnosis of vasospasm (P = < 0.05). In addition, in patients with vasospasm, CSF output recordings were significantly higher prior to the occurrence of vasospasm (438 ml/day) than the period following vasospasm (325.7 ml/day), with a consistent decrease in CSF drainage from day 3 before the occurrence of vasospasm to day 3 after vasospasm (p = 0.012). This may be due to the fact that cerebral blood flow is decreased during vasospasm which can lead to hypoperfusion and decreased pressure in the choroid plexus resulting in decreased CSF production. Furthermore, the autonomic dysfunction and excess sympathetic activity that may occur in the setting of vasospasm can have an inhibitory effect on the choroid plexus and result in a reduction of CSF formation [16]. In addition, the occurrence of cerebral edema may alter the rate of CSF formation [17]. This may be related to clearance of brain edema fluid by the CSF or as a result of a decrease in the extracellular perfusion pressure resulting in a decreased CSF formation rate [17–19]. Disruption of the blood-brain barrier in the setting of SAH and vasospasm results in increased permeability of the damaged endothelium and may contribute to further brain swelling and edema formation with a decrease in CSF production [20,21]. Characteristics studied during this project included age, gender, Fisher grade, Hunt and Hess grade and smoking history. Interestingly, between the two patient groups, only smoking was a significant factor between the group with vasospasm and the group without vasospasm with 33.3% and 63.5, respectively (P < 0.001). Smoking affects the brain vasculature, with increased incidence of plaques and hardening of the vessels, causing the vessels to become less compliant less stretchable and subsequently maybe resistant
Table 5 Average CSF output (ml/day).
D3 pre-vasospasm D2 pre-vasospasm D1 pre-vasospasm Day of vasospasm D1 post-vasospasm D2 post-vasospasm D3 post-vasospasm D1–D3 pre-vasospasm D1–D3 post vasospasm
Vasospasm
No vasospasm (with same day CSF output recordings)
159 140.8 128.4 121.6 117.8 101 90 438 325.7
184.5 189.5 175.3 179.3 198.7 186.7 160.5 551.5 546.7
P value 0.14 0.005 0.003 0.0001 <0.001 <0.001 <0.001 0.014 <0.001
104
C. Hammer et al. / Clinical Neurology and Neurosurgery 144 (2016) 101–104
to the vasospastic effect of the blood degradation product in the CSF. As anyone familiar with the workings of an EVD could attest, there are some challenges that may present limitations to the data presented. Such challenges include sluggish or clogged EVDs that may have been changed or flushed or changes in the draining requested. Other limitations of the study include the retrospective single center design and the relatively small patient number. Considering these limitations, it is hard to draw definitive clinical implications at this point. Rather, the reported findings may stimulate future experimental and clinical studies. 5. Conclusion In patients being monitored with external ventriculostomy in the setting of aneurysmal subarachnoid hemorrhage, a CSF output decline to less than 160 ml/day may be predictive of vasospasm. The sensitivity and specificity of this data is not such that one could rule out or in vasospasm with the decline in CSF alone, however it can be an additional tool and perhaps an early predictor of vasospasm. Our study presents interesting findings that can set the stage for larger and prospective studies. Acknowledgement None. References [1] K. Karnchanapandh, Effect of increased intracranial pressure on cerebral vasospasm in SAH, Acta Neurochirurgica Suppl. 102 (2008) 307–310. [2] Y. Maeda, S. Shirao, H. Yoneda, et al., Comparison of lumbar drainage and external ventricular drainage for clearance of subarachnoid clots after Guglielmi detachable coil embolization for aneurysmal subarachnoid hemorrhage, Clin. Neurol. Neurosurg. 115 (7) (2013) 965–970. [3] K. Karnchanapandh, Effect of increased ICP and decreased CPP on DND and outcome in ASAH, Acta Neurochirurgica Suppl. 114 (2012) 339–342. [4] E. Muench, P. Horn, C. Bauhuf, et al., Effects of hypervolemia and hypertension on regional cerebral blood flow, intracranial pressure, and brain tissue oxygenation after subarachnoid hemorrhage, Crit. Care Med. 35 (8) (2007) 1844–1851, quiz 1852. [5] N. Chalouhi, S. Tjoumakaris, V. Thakkar, et al., Endovascular management of cerebral vasospasm following aneurysm rupture: outcomes and predictors in 116 patients, Clin. Neurol. Neurosurg. 118 (2014) 26–31.
[6] G.J. Kaptain, G. Lanzino, N.F. Kassell, Subarachnoid haemorrhage: epidemiology, risk factors, and treatment options, Drugs Aging 17 (September 3) (2000) 183–199. [7] R. Loch Macdonald, Management of cerebral vasospasm, Neurosurg. Rev. 29 (July 3) (2006) 179–193. [8] N.W. Dorsch, M.T. King, A review of cerebral vasospasm in aneurysmal subarachnoid haemorrhage Part I: incidence and effects, J. Clin. Neurosci. 1 (1) (1994) 19–26. [9] M.D. Vergouwen, M. Vermeulen, J. van Gijn, et al., Definition of delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage as an outcome event in clinical trials and observational studies: proposal of a multidisciplinary research group, Stroke J. Cerebral Circ. 41 (10) (2010) 2391–2395. [10] S. Purkayastha, F. Sorond, Transcranial doppler ultrasound: technique and application, Semin. Neurol. 32 (September 4) (2012) 411–420. [11] R. Khatri, M.Z. Memon, H. Zacharatos, et al., Impact of percutaneous transluminal angioplasty for treatment of cerebral vasospasm on subarachnoid hemorrhage patient outcomes, Neurocrit. Care 15 (August 1) (2011) 28–33. [12] N.F. Kassell, G. Helm, N. Simmons, C.D. Phillips, W.S. Cail, Treatment of cerebral vasospasm with intra-arterial papaverine, J. Neurosurg. 77 (December 6) (1992) 848–852. [13] J.G. Tejada, R.A. Taylor, M.S. Ugurel, M. Hayakawa, S.K. Lee, J.C. Chaloupka, Safety and feasibility of intra-arterial nicardipine for the treatment of subarachnoid hemorrhage-associated vasospasm: initial clinical experience with high-dose infusions, Am. J. Neuroradiol. 28 (May 5) (2007) 844–848. [14] M.M. Kimball, G.J. Velat, B.L. Hoh, Participants in the international multi-disciplinary consensus conference on the critical care management of subarachnoid H. critical care guidelines on the endovascular management of cerebral vasospasm, Neurocritical Care 15 (September 2) (2011) 336–341. [15] D.G. Altman, Practical Statistics for Medical Research, Chapman & Hall/CRC, Boca Raton, FL, 1999. [16] M. Lindvall, C. Owman, Autonomic nerves in the mammalian choroid plexus and their influence on the formation of cerebrospinal fluid, J. Cereb. Blood Flow Metab. 1 (3) (1981) 245–266. [17] K.G. Go, G.M. Hochwald, L. Koster-Otte, A.K. van Zanten, M. Gandhi, The effect of cold-induced brain edema on cerebrospinal fluid formation rate, J. Neurosurg. 53 (November 5) (1980) 652–655. [18] H.J. Reulen, M. Tsuyumu, A. Tack, A.R. Fenske, G.R. Prioleau, Clearance of edema fluid into cerebrospinal fluid A mechanism for resolution of vasogenic brain edema, J. Neurosurg. 48 (May 5) (1978) 754–764. [19] K. Ohata, A. Marmarou, Clearance of brain edema and macromolecules through the cortical extracellular space, J. Neurosurg. 77 (September 3) (1992) 387–396. [20] D.L. Penn, S.R. Witte, R.J. Komotar, E. Sander Connolly Jr., Pathological mechanisms underlying aneurysmal subarachnoid haemorrhage and vasospasm, J. Clin. Neurosci. 22 (January 1) (2015) 1–5. [21] J. Ivanidze, K. Kesavabhotla, O.N. Kallas, et al., Evaluating blood-brain barrier permeability in delayed cerebral infarction after aneurysmal subarachnoid hemorrhage, Am. J. Neuroradiol. 36 (May 5) (2015) 850–854.
Contents lists available at ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Decreased CSF output as a clinical indicator of cerebral vasospasm following aneurysmal subarachnoid hemorrhage Christine Hammer, Badih Daou, Nohra Chalouhi, Robert M. Starke, Lina Ya’qoub, Nikolaos Mouchtouris, Sravanthi Koduri, Stavropoula Tjoumakaris, Robert H. Rosenwasser, Pascal Jabbour ∗ Department of Neurological Surgery, Thomas Jefferson University and Jefferson Hospital for Neuroscience, Philadelphia, PA, USA
a r t i c l e
i n f o
Article history: Received 15 January 2016 Received in revised form 22 February 2016 Accepted 27 March 2016 Available online 28 March 2016 Keywords: CSF output Subarachnoid hemorrhage Vasospasm
a b s t r a c t Objective: Vasospasm is a significant cause of morbidity and mortality among those with aneurysmal subarachnoid hemorrhage (aSAH). Treating increased intracranial pressure by drainage of cerebral spinal fluid through an external ventriculostomy is routine practice. The objective of this study is to evaluate the trends of CSF output in patients who experience vasospasm. Methods: Electronic medical charts were reviewed to identify two groups of patients with aSAH, 75 consecutive patients who developed vasospasm and 75 matched patients who did not develop vasospasm. CSF output was recorded within 3 days before and 3 days after the occurrence of vasospasm. CSF output was recorded for the same days after SAH in matched patients with no vasospasm. Results: Total CSF output was lower in patients with vasospasm as compared to patients without vasospasm matched for the same day (p < 0.001). In patients with vasospasm, CSF output recordings were significantly higher prior to the occurrence of vasospasm (438 ml/day) than the period following vasospasm (325.7 ml/day), with a consistent decrease in CSF drainage from day 3 before vasospasm to day 3 after vasospasm (p = 0.012). Decreasing CSF output was significantly associated with the occurrence of vasospasm (p = 0.017). Youden indices demonstrated that daily CSF drainage <160 ml was significantly associated with the occurrence of vasospasm. The sensitivity of this test was 64.79% and the specificity was 55.38%. Conclusions: In addition to clinical exam findings, observation of a CSF output decline to less than 160 ml/day may be used as additional support for the diagnosis of vasospasm. © 2016 Published by Elsevier B.V.
1. Introduction A subarachnoid hemorrhage (SAH) is the release of blood into the subarachnoid space. The most common causes include hemorrhage from a ruptured aneurysm, an arteriovenous malformation (AVM) or traumatic injury such as blunt force injury to the skull. Immediate life threatening consequences of SAH occur from the clotting and hemolysis of blood in the subarachnoid space and the concomitant hydrocephalus that often occurs leading to increased
Abbreviations: aSAH, aneurysmal subarachnoid hemorrhage; AVM, arteriovenous malformation; CBF, cerebral blood flow; CSF, cerebrospinal fluid; DCI, delayed cerebral ischemia; EVD, external ventricular drain; ICP, intracranial pressure; PTA, percutaneous transluminal angioplasty; TCD, transcranial doppler. ∗ Corresponding author at: Division of Neurovascular Surgery and Endovascular Neurosurgery, Department of Neurosurgery, Thomas Jefferson University Hospital, 901Walnut Street, Philadelphia, PA 19107, USA. E-mail addresses: [email protected], [email protected] (P. Jabbour). http://dx.doi.org/10.1016/j.clineuro.2016.03.024 0303-8467/© 2016 Published by Elsevier B.V.
intracranial pressure (ICP) [1–4]. As a result of the increase in ICP, an external ventricular drain (EVD) is placed in SAH patients to allow for drainage of CSF and maintenance of a normal ICP [2]. Cerebral vasospasm is the delayed-onset narrowing of arteries following SAH and is a significant cause of morbidity and mortality in these patients [5]. It has been shown that facilitating the drainage of CSF leads to faster clot evacuation and subsequently a decreased incidence of vasospasm [2]. Vasospasm occurs approximately three days after SAH, peaks between six to eight days and is usually resolved within twelve days [6]. The onset of vasospasm can include symptoms such as increasing headache, progressive confusion and delirium, weakness or lethargy [7]. About one third of patients with aneurysmal SAH will experience clinical vasospasm and approximately one half to two thirds of aneurysmal subarachnoid hemorrhage patients will have angiographic vasospasm. The incidence of cerebral vasospasm is reported to be between 40 and 70% in patients with SAH with 30% of those patients developing delayed cerebral ischemia (DCI), making an adverse outcome more likely [8]. As a result of DCI, patients may display clinical
102
C. Hammer et al. / Clinical Neurology and Neurosurgery 144 (2016) 101–104
symptoms such as hemiparesis, aphasia, or a decrease in the level of consciousness [9]. Due to the severe complications of vasospasm, Transcranial Doppler (TCD) ultrasound is often used to monitor cerebral blood flow and velocity in SAH patients [10]. The strong correlation between flow velocities in cerebral arteries and cerebral vasospasm severity after SAH makes TCD ultrasound a valuable tool in monitoring vasospasm [10]. Currently, there are three widely-used treatments for vasospasm: percutaneous transluminal angioplasty (PTA) [11], intra-arterial infusion of vasodilators [12,13], and induced hypertension [14]. Previous studies have shown endovascular therapy to have a remarkable safety-efficacy profile with both PTA and nicardipine, a calcium channel antagonist, being equally effective [5]. Induced hypertension involves the elevation of mean arterial pressure with vasopressors and has been shown to be effective in increasing cerebral blood flow (CBF) [4]. At this time, it is difficult to predict which patients will develop vasospasm and when. However, our findings suggest that there is a correlation between CSF output trends and vasospasm. The influence of such a correlation would allow for pre-emptive treatment and decrease the adverse effects of vasospasm. In this study, we evaluated the CSF output trends before and after patients were treated for cerebral vasospasm in order to determine if CSF output can be used as a clinical indicator for impending vasospasm. 2. Materials and methods 2.1. Patient selection and outcomes The study protocol was approved by the University Institutional Review Board. A patient list of consecutive patients with aneurysmal subarachnoid hemorrhage (aSAH) was generated based on ICD-9 codes. Electronic medical charts were reviewed to identify two groups of patients with aSAH: patients who developed clinical vasospasm and patients who did not develop clinical vasospasm post aSAH. Clinical vasospasm was defined as a worsening in neurologic status that could not be attributed to any other factor including rebleeding, hydrocephalus, intracerebral hematoma and metabolic factors. The presence of arterial vasospasm was confirmed with a cerebral angiogram in all patients as evidence of >30% arterial luminal narrowing. 116 consecutive patients with clinical vasospasm post aSAH were identified between January 2007 and December 2012. All patients had an EVD. Initially, the drainage is set at 15–20 cm H2O. Hourly CSF outputs were collected and combined to determine the daily CSF output. Daily CSF output levels were evaluated within three days prior to the occurrence of clinical vasospasm and within three days after vasospasm. Patients with missing CSF output levels were excluded from the study (n = 42). 75 patients with clinical vasospasm meeting the study criteria were included. Blinded to the outcomes of the study, 75 patients with aSAH who did not develop clinical vasospasm were matched in a 1:1 fashion to patients with vasospasm based on patient age, gender, Hunt and Hess, and Fisher grades. Hunt and Hess and Fisher grades were determined by the treating neurosurgeon. The primary outcome of this study was to evaluate trends of CSF output in patients who develop vasospasm after aSAH and identify any potential associations between CSF drainage and the occurrence of vasospasm. Secondary outcomes included (1) sensitivity and specificity analysis of CSF output in predicting the occurrence of clinical vasospasm and (2) an evaluation of other factors that may be associated with the occurrence of vasospasm.
Table 1 Baseline characteristics.
Mean patient age Women Men Smoking Hypertension
Vasospasm
No vasospasm
P value
54 70.7% 29.3% 33.3% 44%
53.4 70.7% 29.3% 63.5% 49%
0.72 1 1 <0.001 0.52
Table 2 Hunt and Hess grades.
1 2 3 4 5
Vasospasm
No vasospasm
Percent
2 7 40 24 2
2 7 39 25 2
2.7 9.3 53.3 32 2.7
Vasospasm
No vasospasm
Percent
0 9 12 54
0 9 12 54
0 12 16 72
Table 3 Fisher grades.
1 2 3 4
not experience vasospasm were matched based age, gender, H&H and Fisher grades and days of CSF output measurement blinded to all other outcomes. Analysis was carried out using Wilcoxon paired rank sum test and McNemar’s test as appropriate. Univariate conditional (matched) analysis was used to test covariates predictive of vasospasm. The following factors were tested: age, sex, smoking, hypertension, Hunt and Hess grade, Fisher grade, aneurysm location and CSF output. Interaction and confounding was assessed through stratification and relevant expansion covariates. Factors predictive in univariate analysis (p < 0.20) [15] were entered into a multivariate conditional logistic regression analysis. Regression models were assessed using area under the receiver operating characteristic curve (AUC). Youden Indices were calculated to determine cutoffs for the dichotomized continuous variable CSF output that yielded the optimal discrimination of vasospasm (sensitivity, specificity, positive predictive value, negative predictive value). P-values of ≤0.05 were considered statistically significant. Statistical analysis was carried out with Stata 10.0 (College Station, TX) [15]. 3. Results 3.1. Baseline characteristics The mean patient age was 54 years. Baseline characteristics of patients with and without vasospasm are detailed in Table 1. 70.7% of patients were women and 29.3% men. The only significant difference between the two groups was a higher percentage of smokers in the no vasospasm group (p < 0.001). The majority of patients had Hunt and Hess grade 3 (53%) (Table 2). SAH was Fisher grade 2 in 12% of patients, Fisher grade 3 in 16% and Fisher grade 4 in 72% (Table 3). In patients with vasospasm, 17.3% of aneurysms were located in the posterior circulation versus 14.8% in patients without vasospasm (P = 0.68). Aneurysm locations are detailed in Table 4.
2.2. Statistical analysis
3.2. CSF output
Data are presented as mean and range for continuous variables, and as frequency for categorical variables. Patients who did and did
CSF output was recorded within 3 days before and 3 days after the occurrence of vasospasm (Table 5). CSF output was recorded for
C. Hammer et al. / Clinical Neurology and Neurosurgery 144 (2016) 101–104
103
Table 4 Aneurysm location.
ACOM PCOM Basilar PICA ICA MCA CO Vertebral Anterior choroidal AICA ACA
Vasospasm
No vasospasm
34 11 3 4 6 7 1 5 1 1 2
30 18 7 3 5 4 2 1 0 0 5
the same days after SAH in matched patients with no vasospasm. The average CSF output day 3 pre-vasospasm was 159 ml/day and in patients who did not have vasospasm matched for the same day it was 184.5 ml/day (p = 0.14). Mean CSF output day 2 pre-vasospasm was 140.8 ml/day and 189.5 ml/day in patients without vasospasm (p = 0.005). Average CSF output day 1 prevasospasm was 128.4 ml/day versus 175.3 ml/day in patients with no vasospasm matched for same days of CSF recordings (p = 0.003). Mean CSF output on the day of vasospasm was 121.6 ml/day. In patients who were matched for the same day after SAH, mean CSF output was 179.3 ml/day (p = 0.0001). Total CSF output was lower in patients with vasospasm as compared to patients without vasospasm matched for the same day (p < 0.001). Furthermore, in patients with vasospasm, CSF output recordings were significantly higher prior to the occurrence of vasospasm (438 ml/day) than the period following vasospasm (325.7 ml/day), with a consistent decrease in CSF drainage from day 3 before the occurrence of vasospasm to day 3 after vasospasm (p = 0.012). In univariate analysis, the only significant factor associated with the occurrence of vasospasm was decreasing CSF output, with the significance increasing from day 3 to day 1 pre-vasospasm (D3 pre-vasospasm: OR = 1.002, p = 0.1, [0.999–1.006], D2 pre-vasospasm: 1.004, p = 0.007, [1.001–1.008], D1 pre-vasospasm: OR = 1.006, p = 0.004, [1.002–1.009], D1–D3 pre-vasospasm: OR = 1.002, p = 0.017, [1.001–1.003]). Similarly in multivariate analysis, CSF outputs at D1 and D2 pre-vasospasm were significantly associated with the occurrence of vasospasm after controlling for age, gender, Fisher grade, Hunt and Hess grade and smoking history (OR d1 = 1.005, p = 0.017, [1.001–1.009], OR d2 = OR = 1.004, p = 0.04, [1.001–1.007]). CSF output D3 prevasospasm was less predictive (OR = 1.001, p = 0.35, [1.001–1.005]). Sensitivity of monitoring CSF output in predicting clinical vasospasm was 64.79%, specificity was 55.38%, positive predictive value was 61.33% and negative predictive value was 59.02% (Fig. 1). Youden indices demonstrated that CSF output <160 ml/day and ≥160 provided the best dichotomized breakpoint (p = 0.018), with daily CSF drainage less than 160 being significantly associated with the occurrence of vasospasm.
Fig 1. Sensitivity and specificity of CSF output <160 ml/day in predicting vasospasm.
4. Discussion We have presented several unique results that suggest that the CSF output trends may be significant in the indication if not the prediction of vasospasm in patients with aSAH. First, the decline in the average CSF output on day 3 pre-vasospasm to less than 160 ml/day may be used as additional support for the diagnosis of vasospasm (P = < 0.05). In addition, in patients with vasospasm, CSF output recordings were significantly higher prior to the occurrence of vasospasm (438 ml/day) than the period following vasospasm (325.7 ml/day), with a consistent decrease in CSF drainage from day 3 before the occurrence of vasospasm to day 3 after vasospasm (p = 0.012). This may be due to the fact that cerebral blood flow is decreased during vasospasm which can lead to hypoperfusion and decreased pressure in the choroid plexus resulting in decreased CSF production. Furthermore, the autonomic dysfunction and excess sympathetic activity that may occur in the setting of vasospasm can have an inhibitory effect on the choroid plexus and result in a reduction of CSF formation [16]. In addition, the occurrence of cerebral edema may alter the rate of CSF formation [17]. This may be related to clearance of brain edema fluid by the CSF or as a result of a decrease in the extracellular perfusion pressure resulting in a decreased CSF formation rate [17–19]. Disruption of the blood-brain barrier in the setting of SAH and vasospasm results in increased permeability of the damaged endothelium and may contribute to further brain swelling and edema formation with a decrease in CSF production [20,21]. Characteristics studied during this project included age, gender, Fisher grade, Hunt and Hess grade and smoking history. Interestingly, between the two patient groups, only smoking was a significant factor between the group with vasospasm and the group without vasospasm with 33.3% and 63.5, respectively (P < 0.001). Smoking affects the brain vasculature, with increased incidence of plaques and hardening of the vessels, causing the vessels to become less compliant less stretchable and subsequently maybe resistant
Table 5 Average CSF output (ml/day).
D3 pre-vasospasm D2 pre-vasospasm D1 pre-vasospasm Day of vasospasm D1 post-vasospasm D2 post-vasospasm D3 post-vasospasm D1–D3 pre-vasospasm D1–D3 post vasospasm
Vasospasm
No vasospasm (with same day CSF output recordings)
159 140.8 128.4 121.6 117.8 101 90 438 325.7
184.5 189.5 175.3 179.3 198.7 186.7 160.5 551.5 546.7
P value 0.14 0.005 0.003 0.0001 <0.001 <0.001 <0.001 0.014 <0.001
104
C. Hammer et al. / Clinical Neurology and Neurosurgery 144 (2016) 101–104
to the vasospastic effect of the blood degradation product in the CSF. As anyone familiar with the workings of an EVD could attest, there are some challenges that may present limitations to the data presented. Such challenges include sluggish or clogged EVDs that may have been changed or flushed or changes in the draining requested. Other limitations of the study include the retrospective single center design and the relatively small patient number. Considering these limitations, it is hard to draw definitive clinical implications at this point. Rather, the reported findings may stimulate future experimental and clinical studies. 5. Conclusion In patients being monitored with external ventriculostomy in the setting of aneurysmal subarachnoid hemorrhage, a CSF output decline to less than 160 ml/day may be predictive of vasospasm. The sensitivity and specificity of this data is not such that one could rule out or in vasospasm with the decline in CSF alone, however it can be an additional tool and perhaps an early predictor of vasospasm. Our study presents interesting findings that can set the stage for larger and prospective studies. Acknowledgement None. References [1] K. Karnchanapandh, Effect of increased intracranial pressure on cerebral vasospasm in SAH, Acta Neurochirurgica Suppl. 102 (2008) 307–310. [2] Y. Maeda, S. Shirao, H. Yoneda, et al., Comparison of lumbar drainage and external ventricular drainage for clearance of subarachnoid clots after Guglielmi detachable coil embolization for aneurysmal subarachnoid hemorrhage, Clin. Neurol. Neurosurg. 115 (7) (2013) 965–970. [3] K. Karnchanapandh, Effect of increased ICP and decreased CPP on DND and outcome in ASAH, Acta Neurochirurgica Suppl. 114 (2012) 339–342. [4] E. Muench, P. Horn, C. Bauhuf, et al., Effects of hypervolemia and hypertension on regional cerebral blood flow, intracranial pressure, and brain tissue oxygenation after subarachnoid hemorrhage, Crit. Care Med. 35 (8) (2007) 1844–1851, quiz 1852. [5] N. Chalouhi, S. Tjoumakaris, V. Thakkar, et al., Endovascular management of cerebral vasospasm following aneurysm rupture: outcomes and predictors in 116 patients, Clin. Neurol. Neurosurg. 118 (2014) 26–31.
[6] G.J. Kaptain, G. Lanzino, N.F. Kassell, Subarachnoid haemorrhage: epidemiology, risk factors, and treatment options, Drugs Aging 17 (September 3) (2000) 183–199. [7] R. Loch Macdonald, Management of cerebral vasospasm, Neurosurg. Rev. 29 (July 3) (2006) 179–193. [8] N.W. Dorsch, M.T. King, A review of cerebral vasospasm in aneurysmal subarachnoid haemorrhage Part I: incidence and effects, J. Clin. Neurosci. 1 (1) (1994) 19–26. [9] M.D. Vergouwen, M. Vermeulen, J. van Gijn, et al., Definition of delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage as an outcome event in clinical trials and observational studies: proposal of a multidisciplinary research group, Stroke J. Cerebral Circ. 41 (10) (2010) 2391–2395. [10] S. Purkayastha, F. Sorond, Transcranial doppler ultrasound: technique and application, Semin. Neurol. 32 (September 4) (2012) 411–420. [11] R. Khatri, M.Z. Memon, H. Zacharatos, et al., Impact of percutaneous transluminal angioplasty for treatment of cerebral vasospasm on subarachnoid hemorrhage patient outcomes, Neurocrit. Care 15 (August 1) (2011) 28–33. [12] N.F. Kassell, G. Helm, N. Simmons, C.D. Phillips, W.S. Cail, Treatment of cerebral vasospasm with intra-arterial papaverine, J. Neurosurg. 77 (December 6) (1992) 848–852. [13] J.G. Tejada, R.A. Taylor, M.S. Ugurel, M. Hayakawa, S.K. Lee, J.C. Chaloupka, Safety and feasibility of intra-arterial nicardipine for the treatment of subarachnoid hemorrhage-associated vasospasm: initial clinical experience with high-dose infusions, Am. J. Neuroradiol. 28 (May 5) (2007) 844–848. [14] M.M. Kimball, G.J. Velat, B.L. Hoh, Participants in the international multi-disciplinary consensus conference on the critical care management of subarachnoid H. critical care guidelines on the endovascular management of cerebral vasospasm, Neurocritical Care 15 (September 2) (2011) 336–341. [15] D.G. Altman, Practical Statistics for Medical Research, Chapman & Hall/CRC, Boca Raton, FL, 1999. [16] M. Lindvall, C. Owman, Autonomic nerves in the mammalian choroid plexus and their influence on the formation of cerebrospinal fluid, J. Cereb. Blood Flow Metab. 1 (3) (1981) 245–266. [17] K.G. Go, G.M. Hochwald, L. Koster-Otte, A.K. van Zanten, M. Gandhi, The effect of cold-induced brain edema on cerebrospinal fluid formation rate, J. Neurosurg. 53 (November 5) (1980) 652–655. [18] H.J. Reulen, M. Tsuyumu, A. Tack, A.R. Fenske, G.R. Prioleau, Clearance of edema fluid into cerebrospinal fluid A mechanism for resolution of vasogenic brain edema, J. Neurosurg. 48 (May 5) (1978) 754–764. [19] K. Ohata, A. Marmarou, Clearance of brain edema and macromolecules through the cortical extracellular space, J. Neurosurg. 77 (September 3) (1992) 387–396. [20] D.L. Penn, S.R. Witte, R.J. Komotar, E. Sander Connolly Jr., Pathological mechanisms underlying aneurysmal subarachnoid haemorrhage and vasospasm, J. Clin. Neurosci. 22 (January 1) (2015) 1–5. [21] J. Ivanidze, K. Kesavabhotla, O.N. Kallas, et al., Evaluating blood-brain barrier permeability in delayed cerebral infarction after aneurysmal subarachnoid hemorrhage, Am. J. Neuroradiol. 36 (May 5) (2015) 850–854.