Chronic hydrocephalus-induced changes in cerebral blood flow: mediation through cardiac effects

Transcription

1 & 2006 ISCBFM All rights reserved X/06 $30 Chronic hydrocephalus-induced changes in cerebral blood flow: mediation through cardiac effects Stephen M Dombrowski 1, Soren Schenk 2, Anna Leichliter 1, Zack Leibson 1, Kiyotaka Fukamachi 2 and Mark G Luciano 1 1 Department of Neurological Surgery, Pediatric and Congenital Neurological Surgery, The Cleveland Clinic Foundation, Cleveland, Ohio, USA; 2 Department of Biomedical Engineering, Lerner Research Institute, The Cleveland Clinic Foundation, Cleveland, Ohio, USA Decreased cerebral blood flow (CBF) in hydrocephalus is believed to be related to increased intracranial pressure (ICP), vascular compression as the result of enlarged ventricles, or impaired metabolic activity. Little attention has been given to the relationship between cardiac function and systemic blood flow in chronic hydrocephalus (). Using an experimental model of chronic obstructive hydrocephalus developed in our laboratory, we investigated the relationship between the duration and severity of hydrocephalus and cardiac output (CO), CBF, myocardial tissue perfusion (MTP), and peripheral blood flow (PBF). Blood flow measures were obtained using the microsphere injection method under controlled hemodynamic conditions in experimental (n = 23) and surgical control (n = 8) canines at baseline and at 2, 4, 8, 12, and 16 weeks. Cardiac output measures were made using the Swan Ganz thermodilution method. Intracranial compliance (ICC) via cerebrospinal fluid (CSF) bolus removal and infusion, and oxygen delivery in CSF and prefrontal cortex (PFC) were also investigated. We observed an initial surgical effect relating to 30% CO reduction and B50% decrease in CBF, MTP, and PBF in both groups 2 weeks postoperatively, which recovered in control animals but continued to decline further in animals at 16 weeks. Cerebral blood flow, which was positively correlated with CO (P = 28), showed no significant relationship with either CSF volume or pressure. Decreased CBF correlated with oxygen deprivation in PFC (P = 06). Cardiac output was inversely related with ventriculomegaly (P = 19), but did not correlate with ICP. Decreased CO corresponded to increased ICC, as measured by CSF infusion (P = 4). Our results suggest that may have more of an influence on CO and CBF in the chronic stage than in the early condition, which was dominated by surgical effect. The cause of this late deterioration of cardiac function in hydrocephalus is uncertain, but may reflect cardiac regulation secondary to physiologic response or brain injury. The relationship between cardiac function and CBF should be considered in the pathophysiology and clinical treatment of.. doi: /sj.jcbfm ; published online 22 February 2006 Keywords: aquaductal stenosis; autoregulation; canine; cerebrovascular; compliance Introduction Chronic hydrocephalus (), characterized by enlarged cerebral cerebrospinal fluid (CSF) ventricles with or without increased intracranial pressure (ICP), has often been associated with a global Correspondence: Dr SM Dombrowski, Department of Neurosurgery, S-80, Pediatric and Congenital Neurological Surgery, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA. dombros@ccf.org Support: 1 RO1 NS A1 Received 7 September 2005; revised 29 December 2005; accepted 1 January 2006; published online 22 February 2006 reduction in cerebral blood flow (CBF) (Graff- Radford et al, 1987; Mamo et al, 1987; Larsson et al, 1994; Tanaka et al, 1997; Chang et al, 1999; Mori et al, 2002; Klinge et al, 2002a, b; Mataro et al, 2003; Owler et al, 2004; Momjian et al, 2004) and a disruption of cerebral autoregulation (Czosnyka et al, 2002, 2003; Minhas et al, 2004). The reduction of CBF seen in is believed to be caused by increased CSF pressure and ventricular volume (VV) (Wozniak et al, 1975; Mabe et al, 1990; Nakada et al, 1992; Goh and Minns, 1995), which results in cortical compression, and blood vessel and white matter fiber stretching (Del Bigio, 1993;

2 Edwards et al, 2004). These structural brain changes leading to decreased CBF may be directly responsible for ischemic conditions in, and have been shown to be reversible with CSF shunt treatment (Del Bigio and Bruni, 1988a, b; Del Bigio, 1989; Fukuhara et al, 2001; Luciano et al, 2001). While in acute hydrocephalus increased ICP, decreased cerebral perfusion pressure, and blood vessel compression are most likely the primary determinants of decreased CBF, in these may not play as significant a role. Other factors that affect CBF, such as changes in metabolic activity, systemic blood flow, and autoregulation, may play a more significant role in the chronic condition. Using an experimental model of chronic obstructive hydrocephalus developed previously in our laboratory (Johnson et al, 1999; Fukuhara et al, 2001; Luciano et al, 2001), we investigated the relationship between the duration and severity of and cerebral and systemic blood flow, cardiac function, tissue oxygen delivery in the prefrontal cortex (PFC), and intracranial compliance (ICC). We hypothesize that in, cardiac and systemic factors are the primary determinants leading to decreased CBF. Materials and Methods Animals Thirty-one (n = 31) young adult, male hounds (canis familiaris), approximately 8 to 9 months of age, weighing 22 to 30 kg were used in this study. Animals were divided into two groups: obstructive (N = 23) and surgical controls (SC, N = 8). Animals were obtained from licensed suppliers and quarantined for a minimum of 7 days before entering into the study. All animals were maintained in the Cleveland Clinic Foundation fully accredited Animal Care Facility in accordance with Public Health Service (PHS) policy and the Health Research Extension Act (PL99-158) under the rules and regulations of the Guide for the Care and Use of Laboratory Animals. Surgery The surgical procedure to induce chronic obstructive hydrocephalus in an experimental animal was previously developed and investigated in our laboratory (Johnson et al, 1999; Fukuhara et al, 2001; Luciano et al, 2001). In brief, each animal received a thoracotomy for cannulation of the left atrium, and a suboccipital craniectomy for the induction/sham -induction of hydrocephalus under general anesthesia (1.0% to 1.5% isoflurane) and sterile conditions. Presurgical medications included Dilantin (5 mg/kg, intravenous) to prevent postoperative seizures, dexamethasone (0.25 to 1.0 mg/kg, intravenous) to reduce inflammation, glycopyrrolate (1 mg/kg, intravenous) to reduce respiratory secretions, and gentamicin (3 mg/kg, intravenous) and cefazolin (1 g, subcutaneous) to prevent infection. Blood flow and cardiac changes in Left Atrium Catheterization and Microsphere Injection In each animal, the left atrium was cannulated (Access Technologies, Skokie, IL, USA; #CNC-7IS) and a subcutaneous portal system (Access Technologies, #CP2) was implanted for CBF microsphere technique. Hemodynamic parameters were monitored via femoral artery (ph = 7.4; PaCO 2 (35 to 45 mm Hg); PaO 2 (150 to 300 mm Hg); O 2 saturation (95% to 100%)). Approximately 10 million (4 ml) microspheres (BioPalt, Medford, MA, USA) were injected into the left atrial (LA) portal system, while a reference blood sample was collected from the femoral artery as a comparative baseline blood flow measure. Randomly chosen, colored microspheres were selected for different time points (i.e., baseline, 2, 4, 8, 12, and 16 weeks) to measure CBF change over time. Induction of Chronic Obstructive Hydrocephalus In brief, animals were placed in a prone position in a stereotaxic headframe, and a sub occipital craniectomy was performed under sterile conditions to access the CSF fourth ventricular space. A flexible, silicon catheter (1.5 mm outer diameter (OD)) was then inserted into the fourth ventricle and approximately 0.4 to 0.6 cm 3 of cyanoacrylic gel was injected. Surgical control animals were generated using procedures identical to those described for experimental animals, with the only exception being the substitution of saline instead of the gel. Postoperative Care Immediately postoperatively, all animals (i.e., and SC) received Dexamethasone (0.5 mg/kg, four times daily (q.i.d.), intravenous), Gentamicin (3 mg/kg, q.i.d., intravenous) and Cefazolin (1 g, q.i.d., subcutaneous). The animals also received Tylenol in combination with Butorphanol (0.2 mg/kg, intravenous), buprenorphine (2 mg/kg, intravenous), or buprenex (1 to 3 mg/ kg i.v., or subcutaneous (s.q.) thrice daily (t.i.d.)) for the management of pain. Oral antibiotics (cephalexin, 1 g, twice daily (b.i.d.), per os (p.o.)) were administered to prevent infection for 5 days postoperatively. Routine postoperative care was provided for pain, hydration, and infection, and included regular examination of vitals, pupillary reflex, and leg responsiveness. Dextrose feeding (intravenous) was initiated in two animals in which prolonged food/water abstinence occurred. In animals, medical treatment for increased ICP included mannitol (1.5 mg/kg) and decadron (0.5 mg/kg). To control for drug treatment differences, all animals received the same dose of mannitol b.i.d. during the immediate postoperative period in (48 h) and SC (24 h), and discontinued when the animal recovered well enough to return to normal housing. Similarly, intravenous fluids were limited to a rate of 25 ml/h until animals ( and SC) were able to drink independently. Body temperature, respiratory rate, 1299

3 1300 heart rate (HR), urine production, activity, fluid intake, and intravenous fluid intake were recorded every hour. Magnetic Resonance Imaging and Volumetric Analysis Magnetic resonance images were collected before induction (i.e., baseline), magnetic resonance imaging (MRI) surgery, and again at killing (2, 4, 8, 12, or 16 weeks) for and SC groups. The MRI at baseline and killing were used to evaluate the anatomical severity and progression of hydrocephalus. 3D magnetic resonance images were acquired using a 1.5-T Siemens Vision Magnetom and used for volumetric analysis. Separate measures for brain and VV were obtained by manually tracing their contours on approximately 60 to 80 sections in the coronal plane of 1 mm thickness from MR digital images using a commercially available image analysis system (Microbrightfieldt, Colchester, VT, USA) (Figure 1). Cardiac Output Cardiac function measures were obtained via the thermodilution method by pulmonary artery (Swan-Ganz) catheterization. Measures included cardiac output (CO) (L/ min), ejections fraction (EF, %), end diastolic volume (EDV), HR (b.p.m.), aortic arterial pressure (AoP, mm Hg), pulmonary arterial pressure (PAP, mm Hg), central venous pressure (CVP, mm Hg), and wedge pressure (W, mm Hg). Cardiac output measures were obtained at baseline and before killing (2, 4, 8, 12, and 16 weeks). Microsphere and Blood Flow Studies Blood flow and cardiac changes in Stable isotope labeled microspheres (BioPAL TM, BioPhysics Assay Laboratory, Inc.). were injected directly into the arterial blood circulation via LA access port to determine regional blood flow. At killing, tissue samples (B0.5 g) were collected from the brain, heart, and peripheral organs. Brain tissue samples were obtained from eight regions: anterior and posterior medial frontal cortex, occipital cortex, superior temporal lobe, inferior frontal cortex, dorsolateral PFC, caudate nucleus, and cerebellar lobules I to VII from the right hemisphere in each animal. Cardiac samples included tissue collected from 13 regions of the left and right atria, and left and right ventricles. Peripheral organ tissue collection included samples from the liver, lung, kidney, and spleen. Tissue and blood samples were stored at 41C until processed by BioPhysics Assay Labs (Worcester, MA, USA). Intraoperative Intracranial Pressure and Oxygen (O 2 ) Measurement Intraoperative ICP and O 2 measures were obtained for both and SC animals at baseline and before killing as described previously (Fukuhara et al, 2001). Hemodynamics conditions were mechanically controlled via ventilative volume and frequency. Under general anesthesia and sterile conditions, two small (5 mm) burr hole craniectomies were performed using stereotaxic coordinates to: (1) allow access to the CSF lateral ventricle and monitoring of CSF O 2 saturation (O 2 -CSF) (Licox CC1-SB; Integra, Plainsboro, NJ, USA), and (2) monitor ICP (Camino, #110-4BT; Integra NeuroSciences) and tissue O 2 saturation (O 2 -TISSUE) in the cortex. Intracranial pressure and O 2 measurements were obtained online and in real time with a digital acquisition system and software (LICOX DAQ ver , GMS-Integra, Germany). Intracranial pressure measurements were obtained continuously at normal ventilation, hyperventilation, and during CSF removal and infusion. Intracranial compliance was measured using a modification of the Masserman technique. A bolus of CSF (0.2 cm 3 ) was withdrawn via the ventricular access catheter. After equilibrium was re-established, a bolus of CSF (0.2 cm 3 ) was infused. Intracranial compliance was then calculated as the change in CSF volume divided by the change in ICP. Separate values were calculated for ICC removal (ICCr) and infusion (ICCi). Killing and Tissue Preparation For killing, animals were deeply anesthetized with sodium pentobarbital in combination with inhaled A B C D652 D717 III FM CA D D652 E D717 CM 1 Cyanoacrylate Figure 1 Gross hemi-section photograph in the sagittal plane (A) showing our model for chronic obstructive hydrocephalus (reprint modification from J. Neuroscience Methods, 1999) using cyanoacrylic injection into the fourth ventricle (); confirmed on MRI for pre- and postsurgical induction of (B, D); and pre- and postsurgery for SC (C, E) (FM, foramen of Monro; III, third ventricle; CA, cerebral aqueduct).

4 isoflorane and perfused with 4.0% paraformaldehyde (PFA) through the LA portal system. Brains were removed and postfixed for 24 h in 4% PFA for study of gross pathology, frozen sectioning, and routine histology (hematoxyline and eosin (H&E), cresyl violet). All CBF tissue samples were consistently obtained from the right hemisphere, while the left hemisphere was cut at 40 mm in the coronal plane and either stored long term or subsequently use for routine immunohistochemical analyses. In addition, B0.5 g brain tissue samples (described in the section Microsphere and Blood flow Studies) were collected for microsphere CBF analyses. Statistical Analysis The results of this study are expressed as mean7the standard error of the mean. Statistical comparisons were made across six different time points (baseline, 2, 4, 8, 12, and 16 + weeks) using repeated-measures analysis of variance (ANOVA) and the Student Neuman-Keuls test. For analysis between and SC groups, two-tailed unpaired t-tests were used. Correlations between CSF volume and pressure, CBF, CO, oxygen delivery, and ICC were also made among time points and between groups. Multiple linear regression analysis was performed using CBF as the outcome (dependent) measure and CO, ventricular CSF volume and ICP, and time as independent variables. Statistical significance was accepted at the probability level less than 5. Results Confirmation of Chronic Obstructive Hydrocephalus Overall, there was no incidence of mortality or morbidity in any animal for either the or SC groups. Animals recovered from anesthesia without complication and began eating and drinking within 36 h. In four animals, postoperative care was extended and assistance in feeding was necessary. Most animals exhibited signs of lethargy, motor weakness, anorexia, and ataxia in the first 7 days immediately after surgery. Vomiting often accompanied neurologic signs (i.e., pupillary dilation and reflex impairment) reflecting increased ICP in animals. These neurologic signs were observed to be transient and resolved within the first 1 to 2 days, with the exception of a few animals which were successfully managed with mannitol to reduce the ICP. In both the and SC groups, animals showed a similar reduction in body weight 2 to 4 weeks postoperatively, which returned to normal baseline levels after 12 weeks (Figure 2C). Statistically, there was no significant difference in weight across time points for either or SC group. Cerebrospinal Fluid Volume: Changes in CSF volume that were quantitated from MRI are graphically illustrated in Figure 2A for animals undergoing induction surgery. Overall, baseline CSF volume for Blood flow and cardiac changes in both the third ventricle (4. to 0.45 cm 3, mean 0.21 cm 3 ) and the lateral ventricle (5 to 1.75 cm 3, mean 0.61 cm 3 ) increased significantly at 2 weeks after induction surgery (P < 1), then decreased slightly, but remained significantly higher than baseline from 4 to 16 + weeks. The increase in CSF volume was shown to be greatest in the lateral ventricles (three- to sixfold) compared with the third ventricle (two- to threefold). Direct comparison of the percent change (%D) in CSF VV from baseline to 16 weeks was significantly greater in (232% 755%) than SC (21%75%) (P < 1; Figure 2D). Intracranial Pressure: Baseline ICP, which ranged from 4.5 to 16.0 mm Hg, mean 8.65 mmhg (Figure 2B), gradually increased, but was not significantly different at any time points after induction (ANOVA, P = 57). In the majority of animals, the average increase in ICP ranged from 21% (2 weeks) to 41% (16 + weeks) from baseline. There was no significant difference in the percent change (%D) of CSF pressure from baseline to 16 weeks between and SC groups (Figure 2D). Gross Morphology and Histologic Analysis: Gross morphologic changes were observed in animals that were induced with. In animals with moderate to severe hydrocephalus (i.e., ventriculomegally), the brains were shown to have ventricular distension, sulcal widening, gyral flattening, cortical compression, and distortion. Immediately after termination of each animal, autopsy revealed no evidence of intracerebral or intraventricular bleeding, which corroborates with MRI. Histologic analysis using routine cresylate violet staining to identify architectonic boundaries and structures showed no signs of intracerebral bleeding or anomalous pathologic condition. Cerebral Blood Flow Overall, baseline CBF ranged from ml/min g. (anterior cingulated gyrus) to ml/min g (caudate nucleus), with an average of 0.59 ml/min g (Table 1). In general, anterior cortical regions had a lower CBF than posterior areas and deep structures. In addition, cortical regions that have almost exclusively a primary blood supply such as anterior cingulated gyrus (anterior cerebral artery (ACA)), superior temporal cortex (middle cerebral artery (MCA)), and occipital cortex (posterior cerebral artery (PCA)) had a lower baseline CBF compared with watershed areas or regions that have an anastamosis of two or more blood vessels, including cerebellum (SCA, PCA), dorsolateral PFC (MCA, ACA), and inferior frontal cortex (ACA, MCA). After induction surgery, changes in CBF for all eight areas investigated showed a consistent pattern (Figure 3A). Overall, CBF decreased significantly at 1301

5 Blood flow and cardiac changes in 1302 A 6 CSF Ventricular Volume B 16 INTRACRANIAL PRESSURE CSF Ventricular Volume (cc) Total 3rd V LV ICP (mmhg) BASE LINE BASE LINE C 30 BODY WEIGHT D % CSF Volume and pressure vs SC at 16 weeks WEIGHT (kg) SC % CSF VOLUME SC % ICP p= BASE LINE rd V LV Total ICP Figure 2 Average CSF volume in the lateral ventricles and third ventricle for the group (A); intraoperative ICP changes for group (B); and average weight (C) for experimental () and SC groups (, N = 23 and SC, N = 8). The percent change (%D) of CSF volume and pressure between and SC is shown for the latest time point, 16 weeks (D). The %DCSF volume, and not ICP, was significantly greater in the group than SC (Pr1). 2 weeks (P < 5), followed by a subsequent and relative improvement or recovery from 4 to 12 weeks, and finally a decrease again at the latest time point (16 + weeks). After 2 weeks of, CBF decreased by 22% to 60%, with the greatest effect observed in the inferior frontal cortex ( 60%) and caudate nucleus ( 59%) compared with least in the anterior and posterior cingulate cortices ( 22% and 26%, respectively). After 16 weeks of, CBF decreased by 49% to 73% compared with baseline CBF levels in which the inferior frontal and caudate nucleus remained most affected. The pattern of CBF for compared with SC was similar early after surgical induction (Figure 3B). CBF was shown to significantly decrease in both and SC groups after 2 weeks, and remained similar, while slightly increasing through 8 weeks postoperatively. However, at 12 and 16 + weeks postoperative, CBF in SC animals was significantly higher than animals (P < 1). Myocardial Tissue Perfusion Myocardial tissue perfusion (MTP) obtained from 13 unique cardiac samples at baseline ranged from ml/min g. (right atrium) to 1.87 ml/min g (posterior papillary muscle), with an average of 1.31 ml/min g. In general, the left ventricle had the greatest MTP, followed by the right ventricle, while the left and right atria had the least MTP. An average MTP for all 13 areas investigated was calculated and used in the direct comparison between and SC groups (Table 2, Figure 4A). At 2 weeks postoperative, MTP in both the and SC groups decreased significantly (P < 5). From week 4 through week 12, the average MTP for and SC groups was similar. At 16 + weeks postoperation, the average MTP for the group was significantly less than baseline and earlier time points (P < 1), and was significantly less than the SC group (P < 5). By comparison, the average

6 Blood flow and cardiac changes in Table 1 Cerebral blood flow for hydrocephalus and control animals obtained from eight different brain regions: anterior and posterior cingulate gyrus, inferior and dorsolateral prefrontal cortex, superior temporal cortex, occipital cortex, cerebellum, and caudate nucleus 1303 Region Baseline 8 12 weeks (%D) 16+weeks (%D) Posterior cingulate gyrus SC Dorsolateral prefrontal cortex SC Superior temporal cortex SC Anterior cingulate cyrus SC Interior frontal cortex SC Cerebellum (lingula) SC Occipital cortex SC Caudate nucleus SC The average percent change (%D) from baseline was also calculated for each area., N = 23; SC, N =8. Table 2 Average CBF, MTP, and PBF for and SC groups for baseline 2, 4 6, 8 12, and 16 weeks, and the percent change (%D) from baseline Baseline 2 Weeks (%D) 4 6 Weeks (%D) 8 12 Weeks (%D) 16+ Weeks (%D) CBF SC MTP SC PBF SC , N = 23; SC, N =8. MTP for the SC group was not significantly different between baseline and the last time point (16 + weeks). Peripheral Blood Flow Overall baseline peripheral blood flow (PBF) varied widely, ranging from ml/min g. (liver), 1.14 ml/min g (spleen), 3.53 ml/min g (lung), and 4.31 ml/min g (kidney). Except for the kidney which showed a significant decrease in blood flow 2 weeks postoperatively (P < 5, data not shown), PBF for liver, lung and spleen did not differ across time or between and SC groups. Peripheral blood flow for liver, kidney, lung and spleen was combined to calculate the average PBF and compared between and SC groups (Table 2, Figure 4B). By comparison, the average PBF between and SC groups was not significantly different at any time point investigated. Cardiac Output Overall, baseline CO ranged from to 8.00 L/min with an average of 4.97 L/min for and 4.10 L/min for SC (Table 3). Figure 3C illustrates the differences in CO over time for and SC groups. In general, CO in both the and SC groups significantly decreased from baseline to 2 weeks (P < 5). From 2 weeks through 12 weeks, CO for and SC groups was similar. At 16 + weeks, CO for the group was significantly less than baseline (P < 5), compared with no significant difference for the SC group (Figure 3C).

7 Blood flow and cardiac changes in 1304 A 1.4 CBF (ml/g min) C Base Line CEREBRAL BLOOD FLOW Posterior cingulate Dorsolateral prefrontal cortex Temporal lobe Anterior cingulate Inferior frontal cortex Cerebellum Occipital cortex Caudate nucleus Base Line CARDIAC OUTPUT B CBF (ml/ming) D CEREBRAL BLOOD FLOW CENTRAL VENOUS PRESSURE SC p<1 C.O. (L/min) SC p<1 0 0 Base Line Base Line Figure 3 Individual CBF for eight different brain regions for group (A); average CBF calculated from eight different brain regions (B), CO measured via the Swann Ganz thermodilution method (C); and CVP (D) comparing the hydrocephalus and SC groups at baseline, 2, 4 to 6, 8 to 12, and 16 + weeks postoperatively (, N = 23; SC, N = 8). Significance was determined at the Pr1 level. CVP (mmhg) SC Table 3 Average CO, intracranial compliance via ICCr and ICCi methods, and CVP for and SC for baseline 2, 4 6, 8 12, and 16 weeks, and the percent change (%D) from baseline Baseline 2 Weeks (%D) 4 6 Weeks (%D) 8 12 Weeks (%D) 16+ Weeks (%D) CO (ml/min) SC ICCr SC ICCI SC CVP (mm Hg) SC , N = 23; SC, N =8. Central Venous Pressure In addition to CO, CVP has been an additional indicator of cardiac function. Central venous pressure, obtained during CO measures, was similar from baseline through 8 weeks for both the and SC groups (Figure 3D). Then at 12 weeks, CVP increased for the group and decreased for the control group. Central venous pressure for the control group was significantly lower than baseline (P < 1) and the group (P < 1) after 16 weeks (Figure 3D).

8 Blood flow and cardiac changes in A MYOCARDIAL TISSUE PERFUSION 0.35 INTRACRANIAL COMPLANCE & CEREBRAL BLOOD FLOW vs TIME MTP (ml/g min) B PBF (ml/g min) Base Line Base Line Intracranial Compliance PERIPHERAL BLOOD FLOW SC SC Intracranial compliance (DV/DP) was obtained via bolus CSF removal (ICCr) and infusion (ICCi) intraoperatively for and SC groups (Table 3, Figure 5). Overall, the patterns of ICCr and ICCi were similar. There was no significant difference in ICCr and ICCi across time or between groups. Both measures showed a nonsignificant increase from baseline to 2 weeks postoperatively. Increased ICCr was directly correlated with an increase in CO (P = 4, R = 0.39). Relationship between Chronic Hydrocephalus and Cerebral Blood Flow, Cardiac Output, O 2 and Intracranial Compliance p<5 Figure 4 Average MTP calculated from 13 different cardiac regions (A), and the average PBF calculated from kidney, liver, lung, and spleen (B), comparing and SC groups at baseline, 2, 4 to 6, 8 to 12, and 16 + weeks postoperatively (, N = 23; SC, N = 8). Significance was determined at the Pr5 level. Cardiac output was shown to be directly correlated with CBF averaged from eight brain regions (P = 28, R = 0.389) (Figure 6A), and inversely ICC ( V/ P, ml/mmhg) BASE LINE related to the degree of CSF VV (P = 19, R = 0.389) in the group (Figure 6B). Cerebral blood flow did not correlate with changes in CSF volume or ICP for either or SC groups (F = 0.33, F = 0.21, respectively), except for CSF volume in of the caudate nucleus (P = 43; R = 0.307; Figure 6D). In addition, CBF obtained for dorsolateral PFC (DLPFC) was directly correlated with O 2 -TISSUE measures obtained from the same area (P = 06, R = 0.450) (Figure 6C). Multiple regression analysis showed a significant relationship between CBF and CO, CSF VV and ICP, and duration (P = 5). Figure 7 shows the inter-relationship between CBF, CO, O 2 (PFC), and CSF ventriculomegaly. Discussion ICCi ICCr CBF 16+ Figure 5 Intracranial compliance measured via bolus ICCr (closed circle) and infusion (ICCi; open circle) methods for compared with CBF (closed triangle, blue line) at baseline, 2, 4 to 6, 8 to 12, and 16 + weeks postoperatively (, N = 23, SC, N = 8). Decrease in CBF corresponds with increase in ICC at 2 weeks, inversely compared with an increase in CBF and decrease in ICC at 8 to 12 weeks. No significant difference was observed between ICCr and ICCi at any time period. This study shows parallel changes in cardiac function, systemic blood flow, and CBF after surgical induction of hydrocephalus. The degree and pattern of CBF reduction and recovery were directly related to changes in cardiac function over time and not to changes in either CSF volume or pressure. While similar reductions in cardiac function and CBF in hydrocephalic and control animals during the early period may reflect surgical effects, the further decline of CO and CBF in only hydrocephalic animals at 16 weeks suggests that the effect may be directly related to the hydrocephalus. Finally, regional CBF reduction was significantly correlated with increased ICC and decreased tissue CBF (ml/gmin)

9 Blood flow and cardiac changes in 1306 A C.O. (L/min) CEREBRAL BLOOD FLOW vs. CARDIAC OUTPUT p=28 R=0.389 B Ventricular Volume (cc) CARDIAC OUTPUT vs. CSF VENTRICULAR VOLUME p=19 R= CBF (ml/g min) Cardiac Output (L/min) C CBF vs. O 2 TISSUE (Dorsolateral Prefrontal Cortex) D CEREBRAL BLOOD FLOW (Caudate Nucleus) vs. CSF VENTRICULAR VOLUME CBF (ml/gmin) p=06 R=0.450 CBF (ml/gmin) p=43 R= O 2 Saturation (mmhg) CSF Ventricular Volume (cc) Figure 6 Scatterplots showing significant correlation between (A) CBF and CO (P = 28); (B) CO and CSF VV (P = 19); (C) CBF and tissue oxygen delivery (P = 06); and (D) CBF for the caudate nucleus and CSF VV (P = 43). p=06 CBF p=0.189 O 2 TISSUE (PFC) p=28 (VENTRICULOMEGALY) CARDIAC OUTPUT p=19 Figure 7 Summary diagram illustrating the relationship between ventriculomegaly, cardiac function, CBF, and oxygen delivery in our experimental (canine) model of chronic obstructive hydrocephalus. The degree of ventriculomegaly in was significantly correlated with CO and not directly related to CBF. In turn, CBF was directly correlated with oxygen delivery in the PFC. oxygen levels, suggesting that the observed changes in CBF were physiologically significant. Using our experimental animal model of, we confirmed and expanded earlier findings for CSF volume, pressure, and compliance from our laboratory (Johnson et al, 1999; Fukuhara et al, 2001; Luciano et al, 2001) and baseline regional CBF measures that were consistent with earlier findings using the identical microsphere technique (Reinhardt et al, 2001). Further, the induced hydrocephalus used in this study resulted in a significant global reduction in CBF, which is also consistent with global decreases reported in earlier experimental studies in cats (Higashi et al, 1986; da Silva et al, 1995) and rats (Jones et al, 1993; Klinge et al, 2003; Kawamata et al, 2003) using different animal models of hydrocephalus and different methods of blood flow measurement. A similar blood flow reduction has been observed in patients with adult (or normal pressure hydrocephalus (NPH)) using a variety of techniques such as Xe clearance with contrast CT, Kety Schmidt dilution, and SPECT, and PET imaging (for a review, see Owler and Pickard, 2001)). Similar to these earlier studies, we found the

10 greatest reduction in CBF in frontal cortical areas and caudate nucleus. Cerebral blood flow reduction in is often considered to be primarily due to changes in ICP and compression and stretch of vessels distorted by ventriculomegaly. Previous anatomical studies have characterized progressive ventriculomegaly as the distortional expansion of the anterior and inferior horns of the lateral cerebral ventricles, suggesting that hydrocephalus may result from differential ventricular expansion relating to force vectors acting on adjacent cortical structures (Greitz et al, 1992). However, while earlier reports found a relationship between CBF and CSF volume (Wozniak et al, 1975; Mabe et al, 1990; Nakada et al, 1992; Goh and Minns, 1995) and pressure (Zierski, 1987), the present study found no correlation between the severity (either CSF volume or pressure) of hydrocephalus and the average CBF of eight brain areas investigated. The exception to this was that the individual CBF measure for the caudate nucleus was correlated with VV. The particular sensitivity of the caudate to increased lateral VV may result from generally high regional flow rate and proximal location to the ventricle, and may explain motor and gait deficits found prominently in clinical. Although this study cannot differentiate between autoregulatory system damage or the breaching of its adaptive limits, the degrees of CO and CBF reduction (B30% and 50%, respectively) shown in this study match or exceed the autoregulation limits suggested in other studies (Davis and Sundt Jr., 1980; Tranmer et al, 1992) (Larsen et al, 1994; Kadas et al, 1997; Ursino and Giulioni, 2003; Van Lieshout et al, 2003; Lakin et al, 2003). The concept of an overwhelmed, but not damaged, autoregulatory system may be further suggested by recovery of response with treatment of hydrocephalus. In an early study we found that the response to hyperventilation (CO 2 or chemical autoregulation) as manifested by oxygen saturation change was absent in hydrocephalic animals, where baseline oxygen level was already low. With CSF shunting, the response to CO 2 changes was restored along with baseline oxygen levels. These experimental findings are consistent with clinical evidence that cerebral autoregulation may be simply overwhelmed, and not destroyed, in (Czosnyka et al, 2002, 2003; Minhas et al, 2004). The cause of the observed changes in cardiac function was likely different in the early versus the late time periods of this study. Cranial and cardiac procedures performed in this investigation, which were identical in both experimental and control groups, likely resulted in surgical trauma responsible for decreased cardiac function and blood flow in the early (2-week) time point. This finding was consistent with clinical and experimental studies that have shown changes in CBF relating to either cranial or cardiac injury (Gruhn et al, 2001; Williams et al, 2001; Nortje and Menon, 2004). By Blood flow and cardiac changes in comparison, differences in cardiac function and blood flow between groups in the late phase were more likely associated with hydrocephalus-induced brain distortion and injury. At this last time point studied (16 weeks), the hydrocephalic group showed a significant divergence from the SC group in blood flow, cardiac function and CVP measures. Taken together, this suggests a decrease in CBF associated with decreased cardiac function specific to the hydrocephalus. This could not be attributed to changes in weight or behavior since both groups were similar (and normal) across all time points. Unfortunately, evolution of this hydrocephalusspecific effect could not be followed past 16 weeks since this was the latest time point investigated. Finally, this study found a significant correlation between ventricular size and CO in the hydrocephalus group, indicating that ventricular enlargement might be affecting cardiac function perhaps through alteration of autonomic function. Previous studies have shown that neurogenic injury may be directly responsible for cardiac dysfunction. Neurologic studies have shown ECG abnormalities to be associated with cerebral infarction, glioma, head trauma, intraparenchmyal hemorrhage, meningitis, seizures, and headaches (Cropp and Manning, 1960; Hersch, 1964; Talman, 1985; Marion et al, 1986; Elrifai et al, 1996). Furthermore, brain injury relating to subarachnoid hemorrhage has shown pathologic changes in cardiac tissue, including contraction band necrosis, myocardial enzyme release, and elevated sympathetic activity (Greenhoot and Reichenbach, 1969; Ibayashi et al, 1986; Lieb et al, 1996; Mayer et al, 1999; Masuda et al, 2002). Though not the main focus of the current study, we did not find ECG or any microscopic pathologic changes in cardiac tissue in any animal with. While severe increases in ICP have been shown to be associated with abnormal ECG (Dicker et al, 1983; Rudehill et al, 1987a, b; Biswas et al, 2000; van Aken et al, 2003), we report a gradual, nonsignificant increase in ICP in our animal model, which would most likely have insignificant clinical implications. Evidence for a heart brain relationship relating to comes mostly from studies investigating cardiac abnormalities and disease that report secondarily changes in CSF space (Graff- Radford and Godersky, 1987; Casmiro et al, 1989; for a review, see Krauss et al, 1996). Our data support the possibility of cardiac function changes in hydrocephalus and further suggest that CNSinduced cardiac changes in turn affect CBF. Hydrocephalus decreased oxygen saturation levels in the DLPFC, and this decrease in tissue oxygen delivery was significantly correlated with decreased CBF in this same region. This finding validates and gives functional significance to our observed decrease in CBF. Our finding for the level of CBF after 16 weeks after induction (19 to 28 ml/g min, mean 24 ml/g min) is considerably less than the minimum CBF of 50 to 60 ml/g min 1307

11 1308 required for adequate oxygen delivery to the brain (Van Lieshout et al, 2003). Cortical areas with blood flow deficits of this magnitude have been shown to be associated with cellular dysfunction, including anaerobic glycolysis and decreased glucose metabolism, and have often been referred to clinically as an ischemic penumbra (Marshall, 2004). Finally, we found that ICC was inversely related to CO, CVP, and blood flow in the early phase. Intracranial compliance (DV/DP), described as the brain s ability to accommodate changes in volume with respect to pressure, is influenced by arterial and venous pressures, blood volume, cellular and extracellular matrices, and surrounding meninges. This study suggests that a decreased CBF may result in decreased brain turgor and a more compliant brain. A more compliant brain may in turn make it more susceptible to further compression by the ventricular system. This increased compliance is an apparent discrepancy with clinical evidence, which suggests a decreased brain compliance in the hydrocephalus. Decreased compliance has also been suggested in other neurologic disorders such as Alzheimer s disease and normal pressure hydrocephalus (Bateman, 2004, 2005), and migraine headaches (Rupprecht et al, 2001; de Hoon et al, 2003; Silvestrini et al, 2004), where CBF may be affected. Indeed, in our own earlier study (Fukuhara et al, 2001), we show decreased cranial compliance in in agreement with that found clinically. This apparent discrepancy is based on the measurement of the volume/pressure relationship under different conditions. We found that with fully expanded ventricles the compliance is decreased, but with CSF drainage and lower pressures the compliance is increased relative to control animals (Fukuhara et al, 2001). These findings are consistent with the hydrocephalus altering the nonlinear pressure volume curve both by increasing compliance in the earlier adaptive phase of the curve, and by decreasing compliance in the steep nonadaptive phase. Taken together, these observations are all consistent with a softer brain in hydrocephalus, which appears less compliant when studied in the compressed state. Finally, these findings are consistent with the common concept of a role for CBF and CVP in ICC and an abnormality in hydrocephalus. Summary/Conclusion Blood flow and cardiac changes in Using an experimental model of chronic obstructive hydrocephalus, this investigation found quantitative changes in cardiac function, cerebral and systemic blood flow, oxygen delivery, and ICC, as each relate to the progression and severity of the hydrocephalus. Overall, we report a close association between blood flow and cardiac function. Initially, we found significant decreases in cardiac function and blood flow for both experimental and SC animals, which later recovers in control animals and further decline in hydrocephalus animals. The results from this study suggest the following: (1) even in developing hydrocephalus, characterized by increased CSF VV and pressure, cardiac function plays a dominant role in CBF; (2) cerebral autoregulation may be overwhelmed, but not likely damaged, in ; (3) is associated with cardiac dysfunction and decreased CBF; (4) decreased CBF may be responsible for impaired oxygen delivery and increased ICC. While cerebral pressure and morphology changes in hydrocephalus undoubtedly play an important role in CBF and diminished brain function, the concurrent measurement of systemic hemodynamics performed in this study shows that the heart continues to have a strong influence over the brain during times of vulnerability. The brain also influences the heart, and alterations in cardiac function associated with the hydrocephalus may further impair CBF. The heart brain interaction in the hydrocephalus will need to be further studied to develop effective treatment strategies. References Bateman GA (2004) Pulse wave encephalopathy: a spectrum hypothesis incorporating Alzheimer s disease, vascular dementia and normal pressure hydrocephalus. Med Hypotheses 62:182 7 Bateman GA et al (2005) The pathophysiology of the aqueduct stroke volume in normal pressure hydrocephalus: can co-morbidity with other forms of dementia be excluded? Neuroradiology 47:741 8 Biswas AK et al (2000) Heart rate variability after acute traumatic brain injury in children. Crit Care Med 28: Casmiro M et al (1989) Risk factors for the syndrome of ventricular enlargement with gait apraxia (idiopathic normal pressure hydrocephalus): a case control study. J Neurol Neurosurg Psychiatry 52: Chang CC et al (1999) Cerebral blood flow in patients with normal pressure hydrocephalus. Nucl Med Commun 20:167 9 Cropp GJ, Manning GW (1960) Electrocardiographic changes simulating myocardial ischemia and infarction associated with spontaneous intracranial hemorrhage. Circulation 22:25 38 Czosnyka M et al (2003) Continuous assessment of cerebral autoregulation: clinical and laboratory experience. Acta Neurochir Suppl 86:581 5 Czosnyka ZH et al (2002) Cerebral autoregulation among patients with symptoms of hydrocephalus. Neurosurgery 50: da Silva MC et al (1995) Reduced local cerebral blood flow in periventricular white matter in experimental neonatal hydrocephalus restoration with CSF shunting. J Cereb Blood Flow Metab 15: Davis DH, Sundt TM, Jr (1980) Relationship of cerebral blood flow to cardiac output, mean arterial pressure, blood volume, and alpha and beta blockade in cats. J Neurosurg 52:745 54

12 de Hoon JN et al (2003) Cranial and peripheral interictal vascular changes in migraine patients. Cephalalgia 23: Del Bigio MR (1989) Hydrocephalus-induced changes in the composition of cerebrospinal fluid. Neurosurgery 25: Del Bigio MR (1993) Neuropathological changes caused by hydrocephalus. Acta Neuropathol (Berlin) 85: Del Bigio MR, Bruni JE (1988a) Changes in periventricular vasculature of rabbit brain following induction of hydrocephalus and after shunting. J Neurosurg 69: Del Bigio MR, Bruni JE (1988b) Periventricular pathology in hydrocephalic rabbits before and after shunting. Acta Neuropathol (Berlin) 77: Dicker D et al (1983) Effect of intracranial pressure changes on the fetal heart rate. Study of a hydrocephalic fetus. Isr J Med Sci 19:364 7 Edwards RJ et al (2004) Chronic hydrocephalus in adults. Brain Pathol 14: Elrifai AM et al (1996) Characterization of the cardiac effects of acute subarachnoid hemorrhage in dogs. Stroke 27: Fukuhara T et al (2001) Effects of ventriculoperitoneal shunt removal on cerebral oxygenation and brain compliance in chronic obstructive hydrocephalus. J Neurosurg 94: Goh D, Minns RA (1995) Intracranial pressure and cerebral arterial flow velocity indices in childhood hydrocephalus: current review. Child Nerv Syst 11:392 6 Graff-Radford NR et al (1987) Regional cerebral blood flow in normal pressure hydrocephalus. J Neurol Neurosurg Psychiatry 50: Graff-Radford NR, Godersky JC (1987) Idiopathic normal pressure hydrocephalus and systemic hypertension. Neurology 37: Greenhoot JH, Reichenbach DD (1969) Cardiac injury and subarachnoid hemorrhage. A clinical, pathological, and physiological correlation. J Neurosurg 30: Greitz D et al (1992) Pulsatile brain movement and associated hydrodynamics studied by magnetic resonance phase imaging. The Monro Kellie doctrine revisited. Neuroradiology 34: Gruhn N et al (2001) Cerebral blood flow in patients with chronic heart failure before and after heart transplantation. Stroke 32: Hersch C (1964) Electrographic changes in subarachnoid haemorrhage, meningitis, and intracranial spaceoccupying lesions. Br Heart J 26: Higashi K et al (1986) Cerebral blood flow and metabolism in experimental hydrocephalus. Neurol Res 8: Ibayashi S et al (1986) Cerebral blood flow and tissue metabolism in experimental cerebral ischemia of spontaneously hypertensive rats with hyper-, normo-, and hypoglycemia. Stroke 17:261 6 Johnson MJ et al (1999) Development and characterization of an adult model of obstructive hydrocephalus. J Neurosci Methods 91:55 65 Jones HC et al (1993) Local cerebral blood flow in rats with congenital hydrocephalus. J Cereb Blood Flow Metab 13:531 4 Kadas ZM et al (1997) A mathematical model of the intracranial system including autoregulation. Neurol Res 19: Kawamata T et al (2003) Metabolic derangements in interstitial brain edema with preserved blood flow: Blood flow and cardiac changes in selective vulnerability of the hippocampal CA3 region in rat hydrocephalus. Acta Neurochir Suppl 86:545 7 Klinge P et al (2002a) The role of cerebral blood flow and cerebrovascular reserve capacity in the diagnosis of chronic hydrocephalus a PET-study on 60 patients. Acta Neurochir Suppl 81:39 41 Klinge P et al (2002b) Regional cerebral blood flow profiles of shunt-responder in idiopathic chronic hydrocephalus a 15-O-water PET-study. Acta Neurochir Suppl 81:47 9 Klinge PM et al (2003) Cerebral hypoperfusion and delayed hippocampal response after induction of adult kaolin hydrocephalus. Stroke 34:193 9 Krauss JK et al (1996) Vascular risk factors and arteriosclerotic disease in idiopathic normal-pressure hydrocephalus of the elderly. Stroke 27:24 9 Lakin WD et al (2003) A whole-body mathematical model for intracranial pressure dynamics. J Math Biol 46: Larsen FS et al (1994) Transcranial Doppler is valid for determination of the lower limit of cerebral blood flow autoregulation. Stroke 25: Larsson A et al (1994) Regional cerebral blood flow in normal pressure hydrocephalus: diagnostic and prognostic aspects. Eur J Nucl Med 21: Lieb K et al (1996) Interleukin-1 beta and tumor necrosis factor-alpha induce expression of alpha 1-antichymotrypsin in human astrocytoma cells by activation of nuclear factor-kappa B. J Neurochem 67: Luciano MG et al (2001) Cerebrovascular adaptation in chronic hydrocephalus. J Cereb Blood Flow Metab 21: Mabe H et al (1990) Cerebral blood flow after ventriculoperitoneal shunt in children with hydrocephalus. Child Nerv Syst 6: Mamo HL et al (1987) Cerebral blood flow in normal pressure hydrocephalus. Stroke 18: Marion DW et al (1986) Subarachnoid hemorrhage and the heart. Neurosurgery 18:101 6 Marshall RS (2004) The functional relevance of cerebral hemodynamics: why blood flow matters to the injured and recovering brain. Curr Opin Neurol 17:705 9 Masuda T et al (2002) Sympathetic nervous activity and myocardial damage immediately after subarachnoid hemorrhage in a unique animal model. Stroke 33: Mataro M et al (2003) Postsurgical cerebral perfusion changes in idiopathic normal pressure hydrocephalus: a statistical parametric mapping study of SPECT images. J Nucl Med 44: Mayer SA et al (1999) Myocardial injury and left ventricular performance after subarachnoid hemorrhage. Stroke 30:780 6 Minhas PS et al (2004) Pressure autoregulation and positron emission tomography-derived cerebral blood flow acetazolamide reactivity in patients with carotid artery stenosis. Neurosurgery 55:63 7 Momjian S et al (2004) Pattern of white matter regional cerebral blood flow and autoregulation in normal pressure hydrocephalus. Brain 127: Mori K et al (2002) Quantitative local cerebral blood flow change after cerebrospinal fluid removal in patients with normal pressure hydrocephalus measured by a double injection method with N-isopropyl-p-[(123)I] iodoamphetamine. Acta Neurochir (Wien) 144: Nakada J et al (1992) Changes in the cerebral vascular bed in experimental hydrocephalus: an angio-architectural 1309