Original Article

1. ¹Department of Imaging and Biophysics, Amiens University Hospital, Amiens Cedex, France


2. ²Department of Neurology, Amiens University Hospital, Amiens Cedex, France


3. ³Department of Radiology, Amiens University Hospital, Amiens Cedex, France


4. ⁴Academic Neurosurgical Unit, Department of Clinical Neurosciences, University of Cambridge, Addenbrooke's Hospital, Cambridge, UK

Correspondence: Dr S Stoquart-ElSankari, Department of Imaging and Biophysics, Centre Hospitalier Universitaire Nord, Place Victor Pauchet, Amiens cedex 80054, France. E-mail: sorayaelsankari560@hotmail.com; Dr O Balédent, Department of Imaging and Biophysics, Centre Hospitalier Universitaire Nord, Place Victor Pauchet, Amiens cedex 80054, France. E-mail: olivier.baledent@chu-amiens.fr

Received 4 December 2008; Revised 24 February 2009; Accepted 3 March 2009; Published online 8 April 2009.

Abstract

Although crucial in regulating intracranial hydrodynamics, the cerebral venous system has been rarely studied because of its structural complexity and individual variations. The purpose of our study was to evaluate the organization of cerebral venous system in healthy adults. Phase-contrast magnetic resonance imaging (PC-MRI) was performed in 18 healthy volunteers, in the supine position. Venous, arterial, and cerebrospinal fluid (CSF) flows were calculated. We found heterogeneous individual venous flows and variable side dominance in paired veins and sinuses. In some participants, the accessory epidural drainage preponderated over the habitually dominant jugular outflow. The PC-MRI enabled measurements of venous flows in superior sagittal (SSS), SRS (straight), and TS (transverse) sinuses with excellent detection rates. Pulsatility index for both intracranial (SSS) and cervical (mainly jugular) levels showed a significant increase in pulsatile blood flow in jugular veins as compared with that in SSS. Mean cervical and cerebral arterial blood flows were 714±124 and 649±178?mL/min, respectively. Cerebrospinal fluid aqueductal and cervical stroke volumes were 41±22 and 460±149?µL, respectively. Our results emphasize the variability of venous drainage for side dominance and jugular/epidural organization. The pulsatility of venous outflow and the role it plays in the regulation of intracranial pressure require further investigation.

Introduction

In adults, the cranium is considered as a rigid closed box, characterized with a dynamic interaction between its four main compartments, namely brain tissue, arterial blood, venous blood, and cerebrospinal fluid (CSF). This dynamic interplay is ruled by the Monro–Kellie doctrine, which states that the total intracranial content should remain constant all through the cardiac cycle (CC) to control the intracranial pressure (ICP) equilibrium. Greitz et al, 1992 have suggested the concept of a ‘modified Monro–Kellie doctrine,’ which takes into account brain expansion in the regulation of ICP. For these authors, the arterial systolic expansion results in an increase in intracranial blood volume. This results in an increase in ICP, which is partially compensated by CSF venting through the subarachnoid spaces. Net brain expansion is responsible for the compression of the ventricular system (the ‘piston-like action’), and thus for aqueductal CSF flow. The coordinated temporal succession between time curves of volumetric changes in arterial, venous, CSF, and brain tissue compartments is responsible for their mechanical coupling throughout the CC in physiologic states (Alperin et al, 1996). An imbalance in this mechanical coupling has been hypothesized to be responsible for pathologic states, such as NPH (normal pressure hydrocephalus)(Greitz, 2004) as well as for morphologic changes reported in normal aging (Greitz, 1969).

Since Greitz (Greitz et al, 1992), several authors (Alperin et al, 1996; Baledent et al, 2001; Greitz, 2004) have considered that the intracranial blood and CSF flush and fill that flows through the CC are initiated by the systolic intracerebral arterial inflow. The arterial cerebral vasculature is characterized by a stable anatomic pattern, which makes it reproducible in anatomic and functional studies. Besides, as it plays a crucial role in brain perfusion and in dynamic intracranial regulation, the arterial cerebral system has been widely studied. The cerebral total and regional blood flows have been evaluated using TCCS (transcranial color-coded sonography) (Stolz et al, 1999) and magnetic resonance imaging (MRI) (Baledent et al, 2001; Baledent et al, 2006) techniques in healthy young and elderly populations (Stoquart-ElSankari et al, 2007), and in pathologic states (Stolz et al, 1999; Baledent et al, 2004).

Conversely, the role played by the venous system is still controversial. Some authors believe that cerebral veins are collapsible, and thus passively change their cross-sectional configuration depending on applied pressure, and that their pulsations are mainly consequent of concomitant arterial pulsations (Schaller, 2004). On the contrary, Bateman, 2000 stressed the role played by the venous system in the mechanical exchanges among intracranial compartments.

Few studies have been devoted to the physiologic and pathologic assessment of cerebral venous system, because vein structures and function remain partly unknown. Anatomic systematization is difficult in veins, especially in the superficial system, because of the multiple individual and hemispheric variations (Schaller, 2004). More recent publications have shown physiologic variations in the dural sinuses drainage owing to respiratory conditions (Valsalva's maneuver) (Mehta et al, 2000) or posture (Gisolf et al, 2004; Alperin et al, 2005). Furthermore, the cervical epidural venous system, which is an accessory venous drainage pathway from the intracranial compartment, may preponderate over the internal jugular system in some physiologic (upright position) or pathologic states, as intracranial hypotension (Clarot et al, 2000). Thus, venous circulation has been rarely studied, and has been mainly evaluated with TCCS in healthy participants. Discrepancies in the detection rate of veins and sinuses have been observed and attributed to age difference in the studied populations, and to heterogeneity of the examination protocols.

Phase-contrast MRI (PC-MRI) enables reliable (Barkhof et al, 1994) noninvasive and rapid measurements of CSF and blood flows, and have been used for venous evaluation in jugular veins, intracerebral veins, and major sinuses (Mehta et al, 2000; Baledent et al, 2004), but physiologic quantitative flow parameters are still lacking.

The purpose of our study was to evaluate the organization of the venous intracranial and internal jugular systems in healthy participants, to investigate the physiologic role of cerebral veins in the regulation of ICP.

Material and methods

Participants

The studied population consisted of 18 healthy young volunteers who underwent cine PC-MRI. There were 9 women and 9 men and their mean age was 31±7 years. The exclusion criteria were any neurologic, psychiatric or severe general disease, alcoholism, or abnormalities detected with clinical MRI exams.

Protocol was accepted by the Institutional Review Board. All patients gave their informed agreement before participation in the study.

Data Acquisition

All MRI exams were performed using a 3 Tesla machine (Signa, General Electric Medical System, Milwaukee, WI, USA). The participants were supine.

Flow images were acquired with a two-dimensional fast cine PC-MRI pulse sequence with retrospective peripheral gating, so that the 32 frames analyzed covered the entire CC. The MRI parameters were as follows: echo time, TE=11 to 17?msecs; repetition time, TR=29 to 43?msecs; 4 views per segment, flip angle: 25° for vascular flows, 20° for CSF flows; field-of-view, FOV=14 × 14?mm²; matrix 256 × 128; and slice thickness 5?mm. Velocity (encoding) sensitization was set at 80?cm/sec for the vessels, 15?cm/sec for the aqueduct, and 5?cm/sec for the cervical subarachnoid space. Sagittal scout view and three-dimensional TOF (time of flight) sequences were used as localizers to select the anatomic levels for flow quantification. The acquisition planes were selected perpendicular to the presumed direction of the flow, and sections through these different levels for each flow series are represented in Figure 1. The acquisition time for each flow series was ~40?secs, with a slight fluctuation that depended on the participant's heart rate, resulting in a total additional examination time of 5?mins.

Figure 1.

Data acquisition. Sagittal scout view sequences were used as a localizer to select the anatomic levels for flow quantification. The acquisition planes were selected perpendicular to the presumed direction of the flows. Sections through the C2–C3 subarachnoid space levels (A1), and the Sylvius aqueduct (B) were used for CSF flow measurement. By varying the velocity encoding, the same cervical section level (A2) was used to measure vascular flows in the left and right internal carotids (ICA), vertebral arteries (VA), and internal jugular veins (IJV). Using the three-dimensional TOF sequence as a localizer, and with adjusted velocity encoding, the intracerebral vascular plane (C) was used to measure arterial flows in the basilar artery (BA), and venous flows in the superior sagittal sinus (SSS) and in the straight sinus (SRS). Finally, a coronal section (D) enabled measurement of the venous flows in the SSS, and in the left and right lateral transverse sinuses (TS). Black pixels represent the flows entering into the section plane, white pixels represent flows directed out of the section plane, and gray pixels correspond to immobile tissues. EV, epidural vein.

Full figure and legend (277K)

Data Analysis

Data were analyzed using an in-house image processing software (Baledent et al, 2001) with an optimized CSF and blood flow segmentation algorithm, which automatically extracts the region of interest at each level and calculates its flow curves over the 32 segments of the CC (for more details, refer to the description of the processing protocol in earlier studies (Baledent et al, 2001; Stoquart-ElSankari et al, 2007)).

Thereafter, the venous, arterial, and CSF flow curves were generated versus time during one CC.

For arterial flows, we calculated the mean arterial total blood flow (CBF) in milliliters per minute at cervical (Figure 1A2) and cerebral (Figure 1C) levels. The cervical cerebral blood inflow consisted of the sum of arterial flows in the left and right internal carotids (ICA) and in VA (vertebral arteries), and the cranial CBF of the sum of arterial flows in the left and right intrapetrous ICA and in the basilar artery.

Similarly, the cervical cerebral venous flow (Figure 1A2) was calculated by summing the venous flows in the left and right internal jugular vein (IJV) and in the epidural vein (EV). The cerebral venous flow (Figure 1C) consisted of the sum of mean venous flows in the superior sagittal sinus (SSS) and straight sinus (SRS).

In addition, we compared the left to right flows in the arterial and venous even vessels (ICA, VA, IJV, and EV), to determine the lateral dominance. A side difference of vascular blood flow of >50% was defined as dominant.

Thereafter, in each participant, we calculated the venous pulsatility index (VPI) at both cervical and cerebral levels. At the cerebral level, this index corresponded to the flow oscillations in the SSS, whereas at the cervical level, it corresponded to the flow oscillations in the dominant vessel (right or left IJV or EV). When generating the venous flow curve (Figure 2A), we calculated the maximum and minimum flow amplitudes (respectively, Fmax and Fmin).

Figure 2.

Venous flow curves. (A) Represents calculation of venous pulsatility index (VPI): The venous flow in an internal jugular vein is plotted during the cardiac cycle and expressed in millimeters per second. The maximum (Fmax) and the minimum (Fmin) amplitudes of the venous flow are represented by crosses. VPI was calculated as follows: VPI=(Fmax-Fmin/Fmax) × 100. (B) Compares the flow curves in the jugular vein and in the superior sagittal sinus: Venous flows are represented across the cardiac cycle at both cervical (jugular vein, in continuous line) and cerebral levels (superior sagittal sinus, in dashed line). This figure represents the difference in pulsatility between the two curves, with more oscillating venous flows in the jugular vein. Although represented in one participant, this finding was present all through the studied population.

Full figure and legend (61K)

The VPI was calculated as follows: VPI=((Fmax-Fmin)/Fmax) × 100.

The VPI represents the pulsatility, and thus indirectly the compliance of the venous intracranial (SSS) or cervical (IJV or EV) compartments under the assumption that the pulse pressure in all sections of the venous pool is the same. The mean cervical and cerebral VPIs were calculated in this population.

In addition, we calculated the cervical arteriovenous delay representing the temporal shift between the arterial systolic maximum flow peak (in the ICA) and the venous maximum peak (in the JV), and expressed it in milliseconds or in percentage of CC. This ratio represents the compliance of the intracerebral compartments.

Finally, CSF flow curves at aqueductal (Figure 1B) and cervical (Figure 1A1) levels were integrated, providing the CSF stroke volumes, which represent the CSF volumes displaced in both directions through the considered region of interest at each level (Nitz et al, 1992; Enzmann and Pelc, 1993). These volumes represent the ‘mobile compliance’ of each compartment and contribute to the rapid regulation of ICP throughout the CC.

Statistical Analysis

Flow amplitude parameters were compared between right and left sides, jugular and EVs, and cerebral and cervical levels, using a univariate analysis with a nonparametric Mann–Whitney test. Similarly, mean cervical and cerebral VPIs were compared using the Mann–Whitney test. The level for statistical significance (P-value) was set at 0.05.

Results

Venous Flows

Table 1 shows the individual mean flow values in each vein and sinus, which illustrate their heterogeneity and dispersion in a normal control population.

Table 1 - Venous blood flows in healthy volunteers.

Full table

When comparing flows in even vessels, we found significant higher right-sided venous flows, for the IJV (right versus left: 399±143 versus 161±150, P<0.001) and for the TS (right versus left: 434±204 versus 195±152, P<0.001).

We found heterogeneity in side dominant veins, with right dominance for IJV in 14 participants (78%), for EV in 4 participants, (22%) and for TS in 12 participants (67%). Left dominance was found in 4 cases (22%) for IJV, 7 (39%) for EV, and 3 (17%) for TS. An interesting finding was that eight volunteers had a strictly unilateral IJV collecting the cerebral venous blood flow, without any contralateral IVJ flow (six right IJV and two left IJV). For 6 of these participants, the TS flow was significantly reduced (< 20%) or null (V3, V7, V9, V10, V13, V14) on the same side as that of the nonexistant IVJ flow.

The jugular venous drainage (left+right IJV mean flow: 519±205?mL/min) preponderated over the epidural drainage (52±54?mL/min) in a significant manner (P<0.001).

Venous flows in EV were peculiar in three participants. V11 and V14 had a unilateral IJV responsible for the totality of cerebral venous drainage. On the contrary, V18 had an increased and dominant EV drainage (flow was null in the right IJV and significantly reduced in the left).

The mean cerebral and cervical venous CBFs were comparable in the young healthy volunteers' population (respectively, 457±80 and 506±137?m/min, P=0.2). However, interestingly, the VPI comparison between cervical and cerebral levels showed a significant difference (P<0.0001), suggesting that the cervical vessel (generally the right IJV) was highly more pulsatile (VPI=38±15) than the cerebral principal venous drainage sinus, represented by the SSS (VPI=21±10). This difference is illustrated by SSS and IJV flow curve representation in one participant in Figure 2B.

Finally, the measured arteriovenous delay was 72±24?msecs (representing 8±2% of the CC duration).

Arterial and Cerebrospinal Fluid Flows

The mean arterial blood flows (expressed in mL/min) were calculated at the cervical and cerebral levels. At the cervical level, the mean flows were 269±49 and 265±59?mL/min, respectively, for the right and left ICA, and 77±32 and 104±41?mL/min for the VA. At the cerebral level, the mean arterial blood flows were 238±71, 251±89, and 161±62?mL/min, respectively, for the right and left intrapetrous ICA, and for the basilar artery. The comparison indicated no side difference, except for VAs that were predominant at the left side for 33% of our population (6 of 18 participants).

The mean cervical and cranial arterial CBF flows were calculated, and no significant difference was found (714±124, 649±178?mL/min; P=0.2).

The CSF flow curve analysis provided the CSF stroke volumes at the aqueductal (41±22?µL) and cervical (460±149?µL) levels.

Discussion

Limits

Only few studies devoted to cerebral venous vasculature can be cited. Anatomic descriptions are based on embryological data. Rare physiologic examinations of the venous vasculature clearly distinguish superficial and deep venous systems (Meder et al, 1994), but many individual and hemispheric variations have been described (Schaller, 2004), which make venous systematic description of venous outflow more difficult than that of arterial inflow. Furthermore, the venous system is more frequently studied in pathologic conditions (cerebral venous thrombosis (CVT), for instance), with incidental findings difficult to correlate with normal function. Finally, cerebral venous hemodynamics are difficult to study as structural and dynamic characteristics of cerebral veins are different from those of blood vessels in other parts of the human body (Plaschke et al, 2003).

The results of our study confirmed the heterogeneity in venous flows in a population of healthy young adults as shown in Table 1. These results are relevant as they express difficulties that may be encountered in clinical situations. For instance, in suspicions of CVT, morphologic MRI explorations aim to visualize the venous obstruction and, in some patients with significant venous flow asymmetry, it is sometimes difficult to settle between physiologic side dominance and partial venous thrombosis.

Side Dominance

Our results showed high levels of lateralized venous flows in TS in our healthy volunteers' population (right: 67% and left: 17%). Furthermore, three participants had drainage only through the right transverse sinus. Rare, earlier published papers have studied side dominance in even venous pathways. An angiographic retrospective evaluation of dural sinuses in 189 cases (Durgun et al, 1993) has found lower levels of venous side dominance (right: 41% and left: 18%), and rare, exclusively one-sided drainage (right: 2.1% and left: 0.53%). These discrepancies are probably related to the difference in the used technique (invasive angiography versus PC-MRI), as our results were comparable with those obtained in a prospective analysis of 100 MRI examinations with normal results (Ayanzen et al, 2000).

We did not find earlier reports of lateral dominance in IJV. In our study, side dominance in IJV was found in similar proportions to that observed in intracranial venous sinuses. This result suggests reliability of PC-MRI in the evaluation of venous flows. These results have relevant clinical implications; vein ligature is indicated in some cervical tumoral syndromes, and may have different clinical effects (intracranial hypertension, neurologic deficits) depending on the side dominant vein. Similarly, some surgical indications need dorsal decubitus, with lateral position of the neck, which can create jugular compression and induce modifications of ICP. Additionally, in patients with suspicion of CVT, it can be uneasy to differentiate the side dominance of venous flow, and partial obstruction, even with appropriate techniques, such as magnetic resonance venography. Actually, in a prospective study, transverse sinus flow gaps have been observed in 31% of patients with normal MRI findings, even with standardized angiographic magnetic resonance protocols (Ayanzen et al, 2000).

Detection Rates

Several earlier publications (Becker et al, 1995; Baumgartner et al, 1997; Stolz et al, 1999) have evaluated intracranial venous hemodynamics in sinuses and in deep cerebral veins, in both healthy adults and in patients with CVT, using TCCS, as it is safe, noninvading, easily realizable in reanimation, and can provide useful information about flow velocity increase or flow direction reversal (Stolz et al, 1999). However, TCCS is operator-dependent, frequently needs arterial landmarks to localize the venous structures and intravenous administration of echo-contrast enhancing agent to visualize dural sinuses (Stolz et al, 1999). Moreover, these studies have found variable detection rates, which were 73% for SS (Becker et al, 1995), 44% (Baumgartner et al, 1997) to 71% (Stolz et al, 1999) for TS, and only 52% for SSS, with unreliable angle-corrected measurements (Stolz et al, 1999). On the contrary, PC-MRI is a noninvasive, sensitive-to-flow, validated technique, which has enabled the quantification of venous flows in SSS in severalearlier publications (Mattle et al, 1990; Mehta et al, 2000), but our study is the first to enable flow quantification in SSS, SRS, and TS, with excellent detection rates.

Principal and Accessory Venous Drainage Pathways

Since the first anatomic descriptions (Breschet, 1829), it has been considered that the vertebral and cervical epidural venous drainages are accessory, and that, physiologically, the cephalic venous drainage is dominantly achieved by the jugular system. It has now been admitted that EVs can enlarge in some pathologic states, such as in arteriovenous malformations, jugular venous thrombosis, achondroplasia, spinal cord disease, or intracranial hypotension (Clarot et al, 2000). Some authors have suggested that the meningorachidian plexus may play a crucial role in physiologic cerebral venous drainage, depending on posture (Alperin et al, 2005) or intrathoracic pressure (Epstein et al, 1970; Dilenge and Perey, 1973). Animal (Epstein et al, 1970; Zippel et al, 2001) and human (Cirovic et al, 2003) studies, as well as mathematical models (Gisolf et al, 2004), have shown that jugular veins collapse in the standing position, which shunts the cephalic blood flow into the vertebral plexus. Conversely, collapsed IJV can be reopened by positive pressure breathing (Toung et al, 2000; Cirovic et al, 2003). In our study, the jugular venous flow was significantly dominant in all healthy volunteers, and we even found inexistent epidural drainage in two participants (V11 and V14 in Table 1). Nevertheless, the cephalic venous drainage was dominantly (V2, V3, and V12) or exclusively (V18) shunted into the epidural system in four participants. Retrospectively, we did not find any pathologic explanation or abnormal magnetic resonance result, and postural factors were not implicated, as all patients were supine during the scan. These various results in physiologic and no-posture-related situations are illustrated in Figure 3.

Figure 3.

Physiologic variation of venous drainage pathways. Axial views at the cervical level show the vascular, arterial, and venous vessels in four healthy volunteers. In the first participant (1), we can observe a habitual venous pattern, with predominant jugular drainage, right dominant (arrow a). Accessory venous drainage is also obvious in epidural veins (arrow b) and in the vertebral plexus (arrow c). In the second (2), this jugular drainage is exclusive, and strictly unilateral (unique right jugular vein) in the third (3). On the contrary, the fourth participant (4) has venous drainage completely shunted from the jugular veins to the epidural and vertebral pathways.

Full figure and legend (113K)

The Role of Veins in Regulating Intracerebral Hydrodynamics

Furthermore, PC-MRI provides a dynamic characterization of the intracranial and cervical vascular and CSF flows and of their interaction during the CC. In the last decade, the increasing use of this technique has helped in understanding the temporal coordinated succession of these flows in healthy young participants, in physiologic aging (Greitz, 1969; Uftring et al, 2000; Stoquart-ElSankari et al, 2007), and in main neurodegenerative diseases (Alzheimer disease, NPH) (Greitz, 1969; Barkhof et al, 1994; Bateman, 2000). The arterial intracerebral systolic inflow is commonly considered now as the ‘driving force’ (Greitz et al, 1992) that initiates a mechanical coupling between intracerebral mobile compartments, resulting in aqueductal, subarachnoid CSF, and venous venting during the systole, to preserve a stable ICP throughout the CC.

Traditionally, veins have been considered as passive, and their cross-sectional changes, as well as their pulsations have been considered to passively copy that of the arterial, in accordance with the Monroe–Kellie doctrine. On the contrary, Bateman, 2000 stressed the role played by the venous system in the mechanical exchanges between intracranial blood and CSF compartments, and suggested that venous dysfunction or insufficiency may be a crucial initiating factor in some cerebral diseases (such as idiopathic intracranial hypertension or NPH) (Bateman et al, 2007; Bateman, 2008), leading to the recently proposed ‘venous theory of hydrocephalus’ (Williams, 2008). Further studies of intracranial flows in these patients, and a comparison with our results in healthy volunteers, may help to understand the pathophysiology of these diseases.

Furthermore, our results suggest that the intracranial and the extracranial venous drainage systems act differently. For each, the calculation of the VPI represented the venous flow oscillation through the CC. This index was significantly reduced in the intracranial SSS. As suggested by earlier papers (Schaller, 2004), although sinuses are supposed to be rigid, they do have some kind of compliance, represented by a pulsatile venous flow in the SSS across the CC. However, we found that the extracranial compartment, mainly represented by the jugular veins, has a higher elasticity, and thus an increased VPI. These results are relevant, for they suggest the role of cervical venous drainage outflow in the ICP regulation. For instance, in achondroplasia, narrowing of the jugular foramen is responsible for jugular veins compression, and induces restriction of jugular venous outflows. An induced reduction of pulsatility in jugular veins has been suggested to cause increased pressure in the SSS, which is initially compensated by increased flows in the SSS (Hirabuki et al, 2000). In a second phase, these compensatory mechanisms are exceeded, which results in reduced flows and velocities in the SSS and in the hydrocephalus (Sainte-Rose et al, 1984). Therefore, it seems that this discrepancy in VPIs between SSS and jugular veins plays a crucial role in the regulation of ICP. An imbalance in this pulsatility equilibrium (achondroplasia, venous thrombosis in the SSS) may be the initial factor responsible for ICP increase and hydrocephalus.

Phase-contrast MRI enables a noninvasive and reliable evaluation of vascular and CSF flows, and of their temporal coordination. Furthermore, this study is the first to use this interesting technique in the physiologic evaluation of cerebral venous flows, with results arguing for heterogeneity and lateralization different from those observed in arteries. Moreover, these results should be used as reference values to study the venous drainage modifications in patients with venous dysfunction (CVT), or even in patients with hydrocephalus, to investigate the probably preponderant role of the venous system in the regulation of ICP equilibrium and cerebral compliance.

http://www.nature.com/jcbfm/journal/v29/n6/full/jcbfm200929a.html