Anatomical variation of cerebral venous drainage: the theoretical effect on jugular bulb blood samples

Transcription

1 Anatomical variation of cerebral venous drainage: the theoretical effect on jugular bulb blood samples S. C Beards, 1 S. Yule, 2 A. Kassner 3 and A. Jackson 2 1 Intensive Care Unit, Withington Hospital, South Manchester University Hospitals Trust, Nell Lane, Withington, Manchester, UK 2 Department of Diagnostic Radiology, Stopford Medical School, Victoria University of Manchester, Oxford Rd, Manchester M13 9PT, UK 3 Philips Medical Systems, Hammersmith, London, UK Summary Recent studies have demonstrated significant variation in bilateral jugular venous oxygen saturation measurements which may be of clinical significance. We have therefore measured variations in normal dural sinus venous drainage to assess the possible effects of normal anatomical variations on measured jugular venous oxygen saturation. Normal volunteers (n ¼ 25) were imaged using magnetic resonance venography to demonstrate variations in venous anatomy. Flow was measured in the superior sagittal sinus and bilaterally in the transverse sinus, sigmoid sinus proximal to the jugular bulb and proximal jugular vein using phase difference magnetic resonance imaging. Examination of magnetic resonance venogram images showed considerable variability in the symmetry of transverse sinus flow. Complete absence of one transverse sinus was seen in four cases and significant asymmetry in the size of the transverse sinuses was present in 13. Quantitative flow studies demonstrated that the ratio of superior sagittal sinus to combined jugular bulb flow showed remarkably little variation ( ). Measurements of transverse sinus flow showed significant asymmetry (< 40% of superior sagittal sinus flow in one transverse sinus) in 21 of 25 volunteers. The effect of the observed asymmetry on jugular venous oxygen saturation was modelled based on the assumption of either a supratentorial or infratentorial lesion. This model predicted significant asymmetry in jugular venous oxygen saturation measurements (> 10%) in 65% of cases with a supratentorial lesion which is in close agreement with clinical observations. This study suggests that normal variations in venous drainage may account for observed asymmetry in jugular venous oxygen saturation measurements. Keywords Anatomy; cerebral venous drainage. Measurement techniques; magnetic resonance imaging. Oxygen; blood levels. Veins; jugular.... Correspondence to: Dr S. C. Beards Accepted: 28 October 1997 Continuous monitoring of jugular venous oxygen saturation (S J O 2 ) is widely used as an indicator of cerebral oxygenation [1 6]. The use of S J O 2 relies on the assumption that changes in cerebral blood flow will be detected by fluctuations in S J O 2 to provide guidance for appropriate therapeutic intervention. In fact, S J O 2 measurements are subject to changes in several physiological variables other than cerebral blood flow [3, 6]. The S J O 2 in any individual is highly dependent on the haemoglobin content of blood and the arterial oxygen concentration. Although these are relatively constant in most patients over a short measurement period, their effect is generally obviated by calculation of the arterial venous oxygen difference (AVDO 2 ) [3, 7, 8]. Unfortunately even the measurement of AVDO 2 is subject to misinterpretation since it is affected both by systemic and by localised changes in cerebral blood flow as well as by changes in cerebral oxygen extraction ratio, which may be elevated in the presence of sepsis or decreased in patients with ischaemia or head injury [3, 6, 9]. Interpretation is further complicated by the loss of cerebral autoregulation, which is common in critically ill patients, particularly those suffering from 627

2 S. C. Beards et al. Cerebral venous drainage Anaesthesia, 1998, 53, pages head injury. In the absence of normal autoregulation, transient small changes in systemic mean arterial blood pressure and cardiac output are reflected in the cerebral blood flow giving rise to short-lived but substantial changes in S J O 2 of up to 30% [3, 8, 10, 11]. In view of these factors, it is clear that isolated measurement of S J O 2 may lead to inappropriate management decisions unless they are considered in combination with other indices such as direct measurement of cerebral blood flow and intracranial pressure. None of these indicators alone is capable of providing adequate information for treatment management. Rises in intracranial pressure may reflect brain swelling and be associated with cerebral hypovolaemia, decreased cerebral blood flow and raised AVDO 2 or may result from rises in mean arterial pressure in the absence of autoregulation where they will be associated with cerebral hyperaemia, increased cerebral blood flow and decreased AVDO 2 [9]. Similar shortcomings can be identified in use of cerebral blood flow monitoring when used in isolation [6]. If intracranial pressure and cerebral blood flow monitoring are available, the measurement of AVDO 2 still supplies important information by allowing estimation of apparent oxygen extraction which may indicate either systemic changes in oxygen extraction ratio or, more commonly, regional abnormalities in oxygen extraction ratio or cerebral blood flow which may not be detected by measurement of overall cerebral blood flow. Unfortunately all of these predictions rely on the assumption that S J O 2 provides a true representative indication of admixed cerebral venous blood. There are many possible errors in this assumption which are highlighted by the observations of Lam et al. [12] who showed clinically significant differences between oxygen saturation measured at the confluence of the sinuses and S J O 2 in 30% of patients. Waltz and colleagues [13] demonstrated that cerebral ischaemia can occur without significant changes in AVDO 2 and suggested that this may indicate that venous blood from the area of ischaemia may not pass into the jugular vein. The concept that measurements of S J O 2 may be affected by regional variations in venous drainage patterns is further supported by two studies demonstrating significant variation in bilateral S J O 2 measurements [14, 15]. This concept relies on the assumptions that changes in venous oxygen saturation and/ or blood flow occur regionally within the brain and that these regions have distinct venous drainage patterns which partially or completely bypass the jugular bulb being monitored. The compartmentalisation of cerebral injury effects is easily substantiated by variations in regional intracranial pressure measurements observed in head injury and stroke patients and by the imaging findings in these cases [6, 9, 16]. Variation in cerebral venous drainage is well recognised and common [16, 17]. The aims of the current study were to document the variation in cerebral venous drainage patterns in a group of normal subjects in order to model the effects these might have on measured S J O 2 in the presence of disease. Methods The study was approved by the local ethics committee. In 25 healthy volunteers, imaging of the cerebral venous system was performed using a 1.5 T Philips ACS NT scanner. Initial localisation images consisted of sagittal, coronal and transverse T1-weighted localisers. A midline sagittal 2D magnetic resonance phase contrast venogram (MRV) and a coronal MRV extending from the torcula posteriorly to the anterior border of the jugular foramina anteriorly were then performed [18]. These initial venogram images were used to locate the sites for measurement of dural venous sinus flow (Fig. 1). Dural venous sinus flow was measured using phase velocity mapping [19]. Flow measurements were taken from single slice images perpendicular to the vessel. Flow was subsequently measured at seven points in the dural venous sinus system: (1) the superior sagittal sinus immediately above the confluence of the sinuses, (2) the right and (3) the left transverse sinus immediately distal to the confluence, (4) the right and (5) the left sigmoid sinus immediately above the jugular bulb bilaterally and (6) the right and (7) the left proximal jugular vein immediately below the skull base (Fig. 1). The velocity encoding was performed in each of the three orthogonal directions and the collections were ECG-gated to produce 15 images per cardiac cycle. Gradient echo, modulus flow and phase difference images were reconstructed for each of the 15 cardiac phases (Fig. 2) and transferred to an independent work station (Sun Micro Systems) for analysis. Measurement of venous flow at each site was made using automatic measurement software developed locally using interactive data language (IDL, Floating Point Systems Ltd). Vessel localisation was performed from modulus flow images using an edge thresholding and seed growing algorithm. Measurements of blood flow (ml.min ¹1 ) were then taken from phase difference images by averaging the flow volume (average flow rate per cross-sectional area) through the regions of interest over the cardiac cycle and multiplying by the heart rate. Using these observations, a simplified model was constructed to examine the potential effects of asymmetric venous drainage on measured S J O 2 values. For the purpose of the model we have assumed that abnormal brain has a normal regional blood flow, that the superior sagittal sinus flow represents all forebrain venous drainage, that the oxygen saturation of venous blood draining normal brain is 70% and that the measured S J O 2 has fallen to 50%. 628

3 S. C. Beards et al. Cerebral venous drainage Figure 1 Magnetic resonance venogram showing flow measurement positions. Numbered slices correspond to 1: superior sagittal sinus; 2: right and 3: left transverse sinus; 4: right and 5: left sigmoid sinus immediately proximal to jugular bulb and 6: right and 7: left proximal internal jugular vein. Results Of the 25 volunteers, 16 were male and nine were female. Their age range was years (mean 27 years). Examination of MRV images showed considerable variability in the symmetry of transverse sinus flow in keeping with previous reports (Fig. 3). No case showed significant drainage through the great vein of Labbe on either side. Maximal flows within the superior sagittal sinus ranged from 8.6 to 16.2 cm s ¹1. Complete absence of Figure 2 Images from quantitative MRV of the superior sagittal sinus. Triplets of images are produced at 15 time intervals equally spaced through the R R interval. (A) Gradient echo image demonstrating anatomy; (B) modulus flow image showing location but not magnitude or direction of flow; (C) phase difference image showing magnitude of flow. 629

4 S. C. Beards et al. Cerebral venous drainage Anaesthesia, 1998, 53, pages Figure 3 MRV images demonstrating asymmetry in transverse sinus flow. (A) Atretic left transverse sinus; (B) absent right transverse sinus. SSS: superior sagittal sinus; RTS: right transverse sinus; LTS: left transverse sinus; RSS: right sigmoid sinus; LSS: left sigmoid sinus; RJB: right jugular bulb; LJB: left jugular bulb; RJV: right jugular vein; LJV: left jugular vein; VA: vertebral arteries. one transverse sinus was seen in four cases (confirmed by gradient echo magnetic resonance images) and significant asymmetry in the size of the transverse sinuses was present in 13 (Fig. 3). The overall results of quantitative flow studies are shown in Table 1. Although there was some expected variation in total cerebral venous flow, the ratio of superior sagittal sinus to combined jugular bulb flow showed remarkably little variation ( ). Figure 4 shows the proportion of sagittal sinus flow passing into the smaller transverse sinus in each case. Using our findings in the model of S J O 2 values, we constructed theoretical lesions to either the supratentorial (Fig. 5) or infratentorial (Fig. 6) compartments in the extremes of symmetrical and unilateral transverse sinus patency. Figure 5(A) depicts a lesion above the tentorium with symmetrical venous drainage. Desaturation of the jugular bulb blood will be symmetrical and, owing to dilution by normal hindbrain venous drainage, an S J O 2 of 50% will be seen when superior sagittal sinus venous saturation falls to 30%. Figure 5(B) depicts the effect of unilateral atresia of the transverse sinus. Since jugular bulb blood on the side of the absent sinus arises entirely from the hindbrain its saturation will remain normal whilst the increased proportion of superior sagittal sinus blood draining into the patent transverse sinus means that an S J O 2 of 50% will be seen when superior sagittal sinus venous saturation falls to 40%. In the case of a supra-tentorial lesion with venous asymmetry, our model would therefore predict that S J O 2 would most sensitively reflect changes in venous saturation if the catheter is placed on the side of the larger transverse sinus whilst readings taken from the contralateral side may be misleading. This is the commonest pattern of cerebral damage in clinical practice and supports the common recommendation that jugular catheterisation be performed on the side of the dominant jugular vein [6]. Figure 6(A) depicts a lesion below the tentorium with symmetrical venous drainage. Desaturation of the jugular bulb blood will be symmetrical and, owing to dilution by normal forebrain venous drainage, an Transverse sinuses (2 þ 3) Jugular bulbs (4 þ 5) Jugular veins (6 þ 7) SSS (1) (1) RTS (2) LTS (3) RJB (4) LJB (5) RJV (6) LJV (7) Mean SD Total (SD 95) 507 (SD 139) 512 (SD 106) Table 1 Measured dural venous flows (ml.min ¹1 ) in 25 normal volunteers. Numbers in parentheses refer to measurement positions shown in Fig. 1. Abbreviations as for Fig

5 S. C. Beards et al. Cerebral venous drainage Number of subjects Percentage of SSS flow Figure 4 Histogram demonstrating asymmetry of transverse sinus flow. y-axis shows number of volunteers. x-axis shows flow in smallest TS as percentage of measured SSS flow. S J O 2 of 50% will represent a hindbrain venous saturation of 30%. Figure 6(B) depicts the effect of unilateral atresia of the transverse sinus. Since jugular bulb blood on the side of the absent sinus arises entirely from the hindbrain, its saturation will accurately reflect the desaturated hindbrain venous drainage whilst the increased proportion of superior sagittal sinus blood draining into the patent transverse sinus means that the S J O 2 on this side will remain elevated at 63%. In these circumstances, our model predicts that S J O 2 would most sensitively reflect changes in venous saturation if the catheter is placed on the side of the smaller transverse sinus. Discussion The venous drainage of the brain is complex and highly variable [16, 17]. The forebrain drains via two principal pathways, the superficial and deep venous systems. The deep cerebral veins carry blood from the upper diencephalon and periventricular cerebral structures to drain into the great cerebral vein of Galen and thence into the straight sinus and the confluence of the sinuses. The superficial veins drain the cortex and follow three main drainage pathways. The majority of forebrain superficial veins drain into the superior sagittal sinus. A variable proportion of the superficial veins draining the anterior part of the temporal cortex drain into the cavernous sinus or the petrosal sinuses into which it feeds to pass directly to the jugular bulb. This anterobasal venous system is usually very small in comparison with the superior sagittal sinus flow. A further small amount of venous drainage takes place directly into the terminal transverse sinuses via the great vein of Labbe. This superficial vein is extremely variable and, although it is usually small, it may be highly prominent and may even replace the transverse sinuses in rare cases (< 0.5%). The drainage of the hindbrain is equally complex with the upper midbrain and superior cerebellum draining into the vein of Galen while the remainder of the infratentorial structures drain into the petrosal or transverse sinuses. Figure 5 See text for full explanation. (A) Model depicting S J O 2 desaturation (50%) due to supratentorial lesion: symmetrical venous drainage; (B) model depicting S J O 2 desaturation (50%) due to supratentorial lesion: asymmetrical venous drainage. 631

6 S. C. Beards et al. Cerebral venous drainage Anaesthesia, 1998, 53, pages Figure 6 See text for full explanation. (A) Model depicting S J O 2 desaturation (50%) due to infratentorial lesion: symmetrical venous drainage. (B) Model depicting S J O 2 desaturation (50%) due to infratentorial lesion: asymmetrical venous drainage. On the basis of these recognised variations in venous anatomy, the measurement of flow in the terminal portion of the superior sagittal sinus in the present study can be assumed to consist entirely of supratentorial forebrain venous drainage. The combined flow into the proximal jugular vein will represent the posterior fossa drainage plus small, variable contributions from the deep venous system of the forebrain, the veins of Labbe and the petrosal venous systems. Thus the mean superior sagittal sinus flow ( ml) represents blood arising entirely from the supratentorial compartment whilst all blood from the infratentorial compartment will be contained in the venous flow through the jugular bulbs ( ml) and proximal jugular veins ( ml). Not only is the venous drainage of the brain variable but asymmetry of the dural venous sinuses is common with significant asymmetry of the transverse sinuses reported in 50 80% and complete atresia in 5 12% of angiographic studies [16, 17]. In patients with significant asymmetry, the right transverse sinus is usually the larger as was demonstrated in the present study. These figures suggest that the prevalence of transverse sinus asymmetry and the variation in flow rates observed in this study fairly represent the variation seen in the population as a whole. It is clear that the mathematical model used to predict S J O 2 is oversimplistic owing to errors in the basic assumptions on which it is based. Focal changes in cerebral flow are a common result of rises in intracranial pressure and represent one of the postulated mechanisms for reduction in S J O 2 on which the concept of jugular bulb monitoring is based [9]. Reduction in flow is the most commonly associated change and would result in even greater asymmetry in S J O 2 readings in cases of sinus asymmetry. The dilution of hindbrain venous flow by prosencephalic blood via the petrosal sinuses, veins of Labbe and vein of Galen will have the opposite effect diminishing apparent asymmetry in measured S J O 2. The normal variability in the prominence of these venous systems makes accurate prediction of changes in S J O 2 impossible. Lastly, although significant asymmetry of the transverse sinuses is normal, the complete absence of one sinus is seen in only 5 12% so that dilutional effects would again be less severe than in the model we have described [16]. Attempts to estimate these errors can be made based on our measurements of asymmetry. If we estimate that 50% of blood joining the venous drainage system below the superior sagittal sinus is prosencephalic then, in our population, a 10% or greater discrepancy in S J O 2 measurements would be seen in 65% of cases by the time the S J O 2 drops to 50% on one side. This is remarkably close to the 56% and 62% observed in clinical studies [14, 15]. In conclusion, this study has supported the previously described asymmetries in dural venous sinus drainage and has provided approximations of regional flow to allow the construction of a mathematical model to describe the intracranial events which would result in jugular bulb desaturation. The predictions of this model correspond closely to the clinical findings of previous workers suggesting that venous asymmetry explains much, if not all, of the observed discrepancy in clinical measurements. 632

7 S. C. Beards et al. Cerebral venous drainage The clinical implications of this study are less clear, in the presence of a predominantly supratentorial lesion, which is the commonest situation, insertion of a catheter on the side of the largest transverse sinus will produce the most sensitive indicator of changes in venous saturation. This may be assessed by measurement of the jugular vein at the skull base using the jugular occlusion test [5] or ultrasound prior to catheter insertion. If these are not available, the catheter should be placed on the right which will be the largest transverse sinus in the majority of cases. It is clear that the commonly suggested protocol that blood should be sampled from the ipsilateral side in the presence of unilateral pathology [6] has no scientific basis. However, it must be realised that in up to 35% of patients with significant asymmetry ( 15 20% overall) the largest transverse sinus will be on the left. Furthermore, if changes in venous saturation are limited to the infratentorial compartment, the S J O 2 measured on the side of the smaller transverse sinus is likely to be the most sensitive to change. Acknowledgments This work was supported by Philips Medical Systems. References 1 Chan KH, Miller JD, Dearden NM, Andrews PJ, Midgley S. The effect of changes in cerebral perfusion pressure upon middle cerebral artery blood flow velocity and jugular bulb venous oxygen saturation after severe brain injury. Journal of Neurosurgery 1992; 77: Souter MJ, Andrews PJ. Validation of the Edslab dual lumen oximetry catheter for continuous monitoring of jugular bulb oxygen saturation after severe head injury. British Journal of Anaesthesia 1996; 76: Garlick R, Bihari D. The use of intermittent and continuous recordings of jugular venous bulb oxygen saturation in the unconscious patient. Scandinavian Journal of Clinical Laboratory Investigation 1987; 188: Kiening KL, Unterberg AW, Bardt TF, Schneider GH, Lanksch WR. Monitoring of cerebral oxygenation in patients with severe head injuries: brain tissue PO2 versus jugular vein oxygen saturation. Journal of Neurosurgery 1996; 85: Dearden NM, Midgley S. Technical considerations in continuous jugular venous oxygen saturation measurement. Acta Neurochirurgica (Wien) 1993; 59: Fitch W. Cerebral blood flow. In: Sebel P, Fitch W, eds. Monitoring the Central Nervous System. Oxford: Blackwell Science Ltd, 1994; de Deyne C, Vandekerckhove T, Decruyenaere J, Colardyn F. Analysis of abnormal jugular bulb oxygen saturation data in patients with severe head injury. Acta Neurochirurgica (Wien) 1996; 138: Fortune JB, Feustel PJ, Weigle CG, Popp AJ. Continuous measurement of jugular venous oxygen saturation in response to transient elevations of blood pressure in headinjured patients. Journal of Neurosurgery 1994; 80: Bullock R. Intracranial pressure measurement. In: Sebel P, Fitch W, eds. Monitoring the Central Nervous System. Oxford: Blackwell Science Ltd, 1994; Schneider GH, von Helden A, Lanksch WR, Unterberg A. Continuous monitoring of jugular bulb oxygen saturation in comatose patients therapeutic implications. Acta Neurochirurgica (Wien) 1995; 134: Sheinberg M, Kanter MJ, Robertson CS, et al. Continuous monitoring of jugular venous oxygen saturation in headinjured patients. Journal of Neurosurgery 1992; 76: Lam JM, Chan MS, Poon WS. Cerebral venous oxygen saturation monitoring: is dominant jugular bulb cannulation good enough? British Journal of Neurosurgery 1996; 10: Waltz A, Sundt T, Michenfelder J. Cerebral blood flow, jugular venous PO2 and lactate concentration and arterial venous oxygen content during carotid endarterectomy. European Neurology 1971; 6: Stocchetti N, Paparella A, Bridelli F, et al. Cerebral venous oxygen saturation studied with bilateral samples in the internal jugular veins. Neurosurgery 1994; 34: Triginer C, Robles A, Baguena M, et al. Differences in bilateral jugular bulb oxygen saturation values in severe head injured patients. Intensive Care Medicine 1995; 21 (S1): Meder J, Chiras J, Roland J, et al. Venous territories of the brain. Journal of Neuroradiology 1994; 21: Simonds G, Truwit C. Anatomy of the cerebral vasculature. Neuroimaging Clinics of North America 1994; 4: Applegate GR, Talagala SL, Applegate LJ. MR angiography of the head and neck: value of twodimensional phase-contrast projection technique. American Journal of Roentgenology 1992; 159: Szolar DH, Sakuma H, Higgins CB. Cardiovascular applications of magnetic resonance flow and velocity measurements. Journal of Magnetic Resonance Imaging 1996; 6: