Cerebral (brain) arteriovenous malformations (BA VMs) are a tangle of disorganized vessels that are a rare cause of hemorrhagic stroke in the general population. Although clinical presentation of hemorrhage may be related to the structure of BAVM vessels, there has been no systematic quantitative analysis of BAVM vessel morphology. Histological sections of excised BAVM lesions were prepared from patients who presented with hemorrhage (n = 14) and from patients with no history of hemorrhage (n = 22). Mean values of radius and wall thickness in each section were determined. BA VM radii were 422 +/- 136 (mu)m (mean +/- SD), minimum wall thickness (thinnest portion of the wall) was 54 +/- 14 (mu)m; and the minimum thickness/radius ratio was 0.23 +/- 0.07. Greater vessel wall thickness was associated with hemorrhagic presentation (OR= 1.1; p = 0.046) after adjusting for feeding artery pressure. Because BAVM vessels from patients presenting with hemorrhage had thicker vessel walls, the search for structural properties predisposing BAVM rupture should be expanded beyond the morphological properties analyzed here. [Neurol Res 2002; 24: 495-500]
Keywords: Cerebral arteriovenous malformation; risk stratification
INTRODUCTION
Brain arteriovenous malformations (BAVMs) are a tangle of abnormal vessels having a number of gross structural deficiencies. BAVM lesions are characterized by a lack of capillaries connecting arteries and veins1,2. Large arteriovenous fistulae3 yield abnormal hemodynamics, including high nonnutritive shunt flows, arterial hypotension, and venous hypertension. Besides these gross structural and hemodynamic abnormalities, the individual vessels that constitute the lesion have dysfunctional smooth muscle and deranged periendothelial support structure4.
BAVMs are of clinical interest, because they are a cause of hemorrhagic stroke typically affecting otherwise healthy individuals1. Several risk factors for BAVM hemorrhage have been identified. Patients initially presenting with hemorrhage have higher feeding artery pressures and exclusively deep venous drainage . The role of BAVM vessel morphology in risk of hemorrhage has not been explored.
The only study quantifying morphology from BAVM histological sections was reported by Isoda et al.6, who described the distribution of vessel sizes within a single BAVM histological section. Although some ultrastructural features of BAVM vessels have been characterized, there have been no systematic attempts to determine how the morphology of BAVM vessels from patients presenting with hemorrhage differs from BAVM vessels from patients with no history of hemorrhage. This lack of information is surprising, since stress borne by the vessel wall is related to the radius and wall thickness8.
The purpose of this work is therefore to quantify the morphology of BAVM vessels, and to test the hypothesis that BAVMs that hemorrhage have structural deficiencies as evidenced by vessel radius and wall thickness.
MATERIALS AND METHODS
Sample selection
As part of a previous study , the records of 475 patients who had undergone surgical excision of a BAVM lesion at Columbia-Presbyterian Medical Center were reviewed, and the specimens of their BAVMs were available in our pathology archives. Fifty-seven patients were identified who met the following criteria:
1. Primarily middle cerebral artery supply.
2. Documented BAVM feeding artery pressure.
3. Documented BAVM size.
4. Surgery for BAVM excision within the previous nine years.
5. Superficial/hemispheric convexity location.
6. Lack of radiosurgery treatment.
From this set, patients with either history of hemorrhage or those that did not present with any hemorrhage on history, clinical exam, or imaging were identified. Thirty-six such patients were identified - 14 patients with a history of hemorrhage as an initial presentation and 22 with no history of hemorrhage. All but one had undergone embolization therapy before excision of the lesion. Table 1 lists relevant clinical data.
Sample preparation
The BAVM specimens were fixed in formalin and preserved as paraffin blocks with no transmural pressure. Sections of the tissue 4 (mu)m thick (one section per patient) were obtained and stained with Mason's trichrome to better visualize the vessel wall. This stain (for collagen and smooth muscle) aided identification of the media and adventitia. Therefore, measurement from the luminal surface outwards represented endothelium, media, and adventitia. Sections were inspected for evidence of damage or artifact from cauterization during surgical excision.
Microscopic images of the BAVM histological sections were acquired with a digital camera with image resolution of 1,600x1,200 pixels. A stage micrometer was used to calibrate image size. While scanning the images, magnification was kept constant. A representative sample of a prepared slide is shown in Figure 1.
Because histological sections were analyzed, information regarding the complex three-dimensional structure of the BAVMs was irretrievable. Shunting arterioles and connecting venules10 were not considered because we examined presumably vessels within the BAVM core and excluded vessels at the margin of the sections. No attempt was made to categorize the BAVM vessels.
Estimating radius and wall thickness
As is apparent from Figure 1, histological sections of BAVMs present a particular challenge to the estimation of vessel radius". Because the vessels are in a collapsed state, lumens have an irregular shape. Since the circumference of a vessel is 2 pi times the radius, the radius can be estimated from the measured perimeter divided by 2 pi. Overestimation of radius resulting from sectioning vessels at oblique angles is addressed in the Appendix.
Similarly, wall thickness was irregular in each of the vessels studied. Three values of wall thickness were therefore determined. First, minimum and maximum thickness were recorded, an established practice in the literature12,13. Second, vessel thickness was estimated by measuring eight radii, equally spaced, and finding the arithmetic mean. Wall thickness and the wall thickness/ radius ratios are plotted as a function of radius. All values are reported as mean +/- SD.
Comparison between patients presenting with and without hemorrhage
Morphology data for each patient were derived from a single histological section, each containing a disparate number of BAVM vessels. To obtain a representative measure for each patient, the mean of minimum, average, and maximum vessel wall thickness, as well as radius, were calculated from each section. The mean of these mean estimates was compared between patients with and without hemorrhage using the Mann-Whitney U test for independent samples.
Multivariate logistic regression was used to examine the relationship between morphology and other potential risk factors of BAVM hemorrhage. We used a regression model with hemorrhage as the dependent variable, and maximum BAVM lesion dimension (mm), feeding artery pressure (mmHg), and venous drainage (exclusively deep versus any superficial drainage) as independent variables. The measure of effect was given by the Odds Ratio (OR). The OR measures the association between hemorrhagic presentation and the independent variables in our model. For continuous measures, such as size and pressure, the OR represents the increase in risk of hemorrhage associated with a unit increase in BAVM lesion size (per mm) or feeding artery pressure (per mmHg). For categorical (dichotomous) variables, the OR represents the ratio of the risk of BAVM hemorrhage between the two categories (i.e., the exclusively deep venous drainage versus any superficial drainage).
RESULTS
General properties of brain arteriovenous malformations
Morphometric measurements were undertaken on single histological sections of all BAVM lesions (n = 36). The number of vessels measured for each lesion averaged 56 (range 7 to 157). For each lesion, mean values of radius, as well as the mean values of minimum, average, and maximum wall thickness were calculated. When grouping hemorrhage and nonhemorrhage groups, the mean value of radius was 422 +/- 136 (mu)m; the mean value of wall thickness at the thinnest portion of the wall was 54 +/- 14 (mu)m; the mean value of the averaged wall thickness was 97 +/- 28 (mu)m; and the mean value of the thickest portion of the wall was 167 +/- 52 (mu)m. The mean value of the minimum thickness/radius ratio was 0.23 +/- 0.07.
Characterizing vessel thickness-radius relationship
The relationship of wall thickness and radius were plotted for all vessels of all specimens in Figure 2A. Plotted is the mean wall thickness of each vessel. In general, vessels with a larger radius have a greater wall thickness. However, the ratio of wall thickness/radius was not consistent. The wall thickness/radius ratio was a strong function of radius, with smaller vessels having a greater relative thickness (Figure 28).
Comparing vessels from BAVMs with and without initial presentation of hemorrhage
Of the 36 BAVMs specimens available for this study, 14 came from patients presenting with intracranial hemorrhage. The distribution of vessel radii in hemorrhage and nonhemorrhage groups is illustrated in Figure 4. There were no significant differences in vessel radii, wall thickness, and thickness to radius ratio between hemorrhage (HG) and nonhemorrhage groups (NHG) (Table 1). BAVMs that presented with hemorrhage had a trend of greater minimal wall thickness (HG: n=14, 60 +/- 18, NHG: n=22, 50 +/- 9, p=0.08) and average thickness (HG: n=14, 107 +/- 37, NHG: n=22, 90 +/- 17, p=0.22). There was no significant difference in radius or wall thickness to radius ratio in hemorrhage and nonhemorrhage groups. However, higher minimum vessel wall thickness was associated with initial hemorrhagic presentation (OR= 1.1; p=0.046) when adjusted for feeding artery pressure (i.e., when accounting for the confounding effects of different pressures in HG and NHG). Feeding artery pressure itself was significantly associated with hemorrhagic presentation (OR= 1.1; p=0.041). There was a trend for smaller maximum BAVM lesion dimension (OR=0.98; p=0.43) and exclusively deep venous drainage (OR=3.5; p=0.33) to be associated with hemorrhagic presentation, although the relationship was not statistically significant in univariate or multivariate analyses.
DISCUSSION
The present study is the first to systematically quantify the wall thickness and radii of brain arteriovenous malformation (BAVM) vessels from histopathological sections, and to relate BAVM vessel morphology to the risk of hemorrhagic presentation. Contrary to expectation, BAVM vessel walls from the hemorrhage group were thicker than those from the nonhemorrhage group. The risk of BAVM rupture does not appear to be directly related to the wall thickness of vessels within the BAVM nidus11,14,15.
Histological sections of complex BAVM structure
This study utilized histological sections of BAVM lesions to characterize the morphology of BAVM vessels. While this can account for vessels within the nidus, it does not account for the feeding arteries or draining veins. Furthermore, shunting arterioles or connecting venules10, linking the BAVM to the normal circulation, cannot be analyzed by this method.
Measuring radius and wall thickness
This study emphasizes the difficulty in determining BAVM vessel radii and wall thicknesses. Morphometric measurements in sections were obtained from paraffin blocks of BAVMs. The absence of intranidal pressure at the time of fixation could have influenced measurements. Unloading the specimens may cause the vessel walls to recoil, influencing vessels with different elastic properties in different ways. Furthermore, formaldehyde fixation itself affects morphometric characteristics16. Heating effects from cautery during excision may also cause changes to both wall thickness and radius10.
Another difficulty arose in estimating in vivo radius from histological sections because of vessel orientation (see Appendix). Methods employed by Tornig et al.13 and Short12 to eliminate the influence of vessel orientation presumed a circular shape of in vivo lumen. They estimated radius from the shortest axis of ovular lumens in the histological sections. This correction method may work well for arteries, which maintain a nearly circular shape at zero pressure, but not for collapsible vessels such as veins. Isoda et al.6 described morphometric measurements in the BAVM specimens by smoothing the shape of the vessel wall and treating the lumen as an ellipse. The lack of information in the literature for the radii of veins may be a direct result of a similar difficulty faced in the present work. Although our simple method overestimates vessel radius, it does so in a predictable manner (Figure 3). Because radius and wall thickness are overestimated by the same factor, the ratio of wall thickness to radius is insensitive to vessel orientation. This may explain the good agreement between the wall thickness/radius ratio of 0.23 +/- 0.07 found from our vessels and the 0.2 (=0.1 thickness/ diameter ratio multiplied by 2) reported by Yamada et al.10.
Wall thickness
Wall thickness in normal vessels is influenced by transmural pressure and endothelial shear stress. With chronic hypertension, the walls of normal vessels grow thicker17,18 in order to maintain circumferential wall stress at acceptable levels8,19. This normal control mechanism helps resistance arteries to maintain vascular tone18,20 and may help prevent vessel rupture14,15. The thick BAVM vessel walls in patients presenting with hemorrhage may therefore be a normal response to increased pressure.
The present work illustrated that the wall thickness/ radius ratio of BAVM vessels is relatively constant for larger vessels, but increases significantly in smaller vessels (Figure 2). This pattern is similar to that described by Bevan et al.20 for normal vessels of many dissimilar types. The teleological argument for this finding in normal vessels is that the larger conductance vessels have a similar wall thickness/radius ratio in order to maintain a consistent circumferential wall stress. Smaller resistance vessels require a much thicker media to control radius. However, although smaller BAVM vessels have a relatively thick wall, they do not control flow2.
Hemorrhage
The fact that BAVMs with thinner vessels tend not to rupture helps to focus clinical research related to hemorrhage. The strength of a vessel wall depends on not only wall thickness but also elastance11,14. The elastic properties of isolated vessels or vascular beds21 are critical to pulsatile volume and pressure22. Because venous hypertension23, venous thrombosis24 , and venous derangement25 may precede the development of cerebral arteriovenous malformations, the particular topology of BAVMs should be explored.
Furthermore, because only the vessels within the BAVM core were analyzed, the primary site of hemorrhage may be outside this region. For instance, hemorrhage may occur at shunting arterioles and venules that connect the BAVM lesion to the normal circulation. Likewise, neither the major feeding arteries or draining veins have been addressed in this study, and their morphology have not been characterized. Further studies should focus on relating these morphological measurements in AVM vasculature to clinical and physiological variables in order to gain a better understanding of the underlying mechanism of BAVM hemorrhage.
ACKNOWLEDGEMENTS
This work was supported in part by National Institutes of Health grants PHS R01 NS 37921 and K24 NS02091 (WILY) and 5T32GM08440 (CMQ). We thank the Columbia AVM study group and the UCSF BAVM study project for their support and assistance. Portions of this work were presented at the 1998 meeting of the Society of Neurosurgical Anesthesia and Critical Care, Orlando, FL, USA. The specimen collection and initial data analysis were performed at Columbia-Presbyterian Medical Center, New York, NY, USA. Final data analysis and manuscript preparation were performed at the University of California, San Francisco, CA, USA.
REFERENCES
1 Arteriovenous Malformation Study Group. Current Concepts: Arteriovenous malformations of the brain in adults. N Engl J Med 1999; 340:1812-1818
2 Young WL. Intracranial arteriovenous malformations: Pathophysiology and hemodynamics (Chapter 6). In: Jafar Jj, Awad IA, Rosenwasser RH, eds. Vascular Malformations of the Central Nervous System, New York: Lippincott Williams & Wilkins, 1999: pp.95-126
3 Tanaka R, Miyasaka Y, Fujii K, Kan S, Yagishita S. Vascular structure of arteriovenous malformations. J Clin Neurosci 2000; 7: 24-28
4 Mandybur TI, Nazek M. Cerebral arteriovenous malformations. A detailed morphological and immunohistochemical study using actin. Arch Pathol Lab Med 1990; 114: 970-973
5 Duong DH, Young WL, Vang MC, Sciacca RR, Mast H, Koennecke HC, Hartmann A, Joshi S, Mohr JP, Pile-Spellman J. Feeding artery pressure and venous drainage pattern are primary determinants of hemorrhage from cerebral arteriovenous malformations. Stroke 1998; 29: 1167-1176
6 Isoda K, Fukuda H, Takamura N, Hamamoto Y. Arteriovenous malformation of the brain - histological study and micrometric measurement of abnormal vessels. Acta Pathol Jpn 1981; 31: 883-893
7 Wong JH, Awad IA, Kim JH. Ultrastructural pathological features of cerebrovascular malformations: A preliminary report. Neurosurgery 2000; 46:1454-1459
8 Quick CM, Baldick HL, Safabakhsh N, Lenihan TJ, Li JK, Weizsacker HW, Noordergraaf A. Unstable radii in muscular blood vessels. Am J Physiol 1996; 271: H2669-H2676
9 Hashimoto T, Mesa-Tejada R, Quick CM, et al. Evidence of increased endothelial cell turnover in brain arteriovenous malformations. Neurosurgery 2001; 49: 124-131
10 Yamada S, Liwnicz B, Lonser RR, Knierim D. Scanning electron microscopy of arteriovenous malformations. Neurol Res 1999; 21: 541-544
11 Hamby WB. The pathology of supra-tentorial angiomas. J Neurosurg 1958; 15: 65-75
12 Short D. Morphology of the intestinal arterioles in chronic human hypertension. Br Heart J 1966; 28: 184-192
13 Tornig J, Gross ML, Simonaviciene A, Mall G, Ritz E, Amann K. Hypertrophy of intramyocardial arteriolar smooth muscle cells in experimental renal failure. j Am Soc Nephro1 1999; 10: 77-83
14 Gao E, Young WL, Hademenos GJ, Massoud TF, Sciacca RR, Ma Q, Joshi S, Mast H, Mohr JP, Vulliemoz S, Pile-Spellman J. Theoretical modelling of arteriovenous malformation rupture risk: A feasibility and validation study. Med Eng Phys 1998; 20: 489-501
15 Massoud TF, Hademenos GJ, Young WL, Gao E, Pile-Spellman J. Can induction of systemic hypotension help prevent nidus rupture complicating arteriovenous malformation embolization?: Analysis of underlying mechanism achieved using a theoretical model. Am/ Neuroradiol 2000; 21: 1255-1267
16 Hart MN, O'Donnell SL. Effects of formaldehyde fixation on basilar artery caliber. Stroke 1980; 11: 99-100
17 Tanoi Y, Okeda R, Budka H. Binswanger's encephalopathy: Serial sections and morphometry of the cerebral arteries. Acta Neuropathol (Bert) 2000; 100: 347-355
18 Aalkjaer C, Heagerty AM, Petersen KK, Swales JD, Mulvany MJ. Evidence for increased media thickness, increased neuronal amine uptake, and depressed excitation-contraction coupling in isolated resistance vessels from essential hypertensives. Circ Res 1987; 61: 181-186
19 Baumbach GL, Siems JE, Heistad DD. Effects of local reduction in pressure on distensibility and composition of cerebral arterioles. Circ Res 1991; 68: 338-351
20 Bevan JA, Dodge J, Walters CL, Wellman T, Bevan RD. As human pial arteries (internal diameter 200-1000 micron) get smaller, their wall thickness and capacity to develop tension relative to their diameter increase. Life Sci 1999; 65: 1153-1161
21 Quick CM, Berger DS, Hettrick DA, Noordergraaf A. True arterial system compliance estimated from apparent arterial compliance. Ann Biomed Eng 2000; 28: 291-301
22 Quick CM, Berger DS, Noordergraaf A. Apparent arterial
compliance. Am J Physiol 1998; 274: H1 393-Hl 403
23 Lawton MT, Jacobowitz R, Spetzler RF. Redefined role of angiogenesis in the pathogenesis of dural arteriovenous malformations. J Neurosurg 1997; 87: 267-274
24 Phatouros CC, Halbach VV, Dowd CF, Lempert TE, Malek AM, Meyers PM, Higashida RT. Acquired pial arteriovenous fistula following cerebral vein thrombosis. Stroke 1999; 30: 2487-2490
25 Nussbaum ES, Heros RC, Madison MT, Awasthi D, Truwit CL. The pathogenesis of arteriovenous malformations: Insights provided by a case of multiple arteriovenous malformations developing in relation to a developmental venous anomaly. Neurosurgery 1998; 43: 347-351
Christopher M. Quick*, David J. James*, Kelvin Ning*, Shailandra Joshi^, Alexander X. Halim*, Tomoki Hashimoto* and William L. Young*^^
*Center for Cerebrovascular Research, Department of Anesthesia and Perioperative Care, University of California, San Francisco, CA
^Department of Anesthesiology, Columbia University, New York, NY
^^Departments of Neurosurgery and Neurology, University of California, San Francisco, CA, USA
Correspondence and reprint requests to: Christopher M. Quick, PhD, Center for Cerebrovascular Research, Department of Anesthesia and Perioperative Care, University of California, San Francisco, 1001 Potrero Ave., Room 3C-38, San Francisco, CA 94110, USA. [quickc@anesthesia.ucsf.edu] Accepted for publication March 2002.
Copyright Forefront Publishing Group Jul 2002
Provided by ProQuest Information and Learning Company. All rights Reserved