ACTA NEUROCHIRURGICA
Acta Neurochirurgica 39, 1--14 (1977)
9 by Springer-Verlag 1977
The Department of Neurosurgery, Hirosaki University School of Medicine, Hirosaki, Japan
Changes in the Subarachnoid Space After Experimental Subarachnoid Haemorrhage in the Dog: Scanning Electron Microscopic Observation By
S. Suzuki, M. Ishii, M. Ottomo, and T. Iwabuchi With 10 Figures
Summary The possible changes in the subarachnoid space after subarachnoid haemorrhage were studied in animals by using a scanning electron microscope (SEM). About 1 ml/kg of autogenous blood was injected intracisternally in 36 adult mongrel dogs to investigate changes in the subarachnoid space, over periods ranging from immediately after the injection to as long as 6 months. We have come to the conclusion that the injected blood disappears in about one to two weeks; the fibrosis or thickening of the arachnoid membrane appears in one to three weeks, and then returns to normal in a month in instances of rapid recovery, but there are some cases in which fibrosis persists for a long period and becomes chronic. The fact that an increase of fibrous tissue was found in the parietal region, where the injected blood had hardly reached, appears to indicate that the fibrosis is not always limited to the site of the baemorrhage but can occur in remote regions, We also discuss the usefulness of the SEM in the observation of the subarachnoid space, and the finding that vascular specimen preparations can be made by perfusing the brain with 2-10~ phosphate-buffered formaldehyde solution.
Key words: Subarachnoid hemorrhage, hydrocephalus, subarachnoid fibrosis, scanning electron microscope.
No definite explanation has yet been given for secondary hydrocephalus following subarachnoid hemorrhage, which may be classified as high-pressure hydrocephalus (HPH), or normal pressure hydrocephalus (NPH) in which cerebrospinal fluid (CSF) pressure is normal but psychic symptoms are predominant. Studies are being made of various aspects, with attention directed to disturbances, either of the production and absorption of CSF, or of CSF flow in the ventricular system and subarachnoid space. Acta Neurochirurg/ca, Vol. 89, Fast. 1--~
1
2
S. Suzuki et al.: Subarachnoid Change After Haemorrhage
Bagley (1928) ~ has demonstrated the production of ventricular dilatation by intracisternal blood injection in pups and pointed out that an increase of fibrous tissue was noted in the subarachnoid space. Since then, a large number of studies, mainly clinical, have reported hydrocephalus following subarachnoid hemorrhage, and changes in the subarachnoid space, in support of the view that the disturbance
Fig. 1. Upper: Basal view of the brain removed immediately after the intracisternal blood injection in the dog. Lower: Sites of specimen excision: (1) Prepontine subarachnoid space; (2) Interpeduncular cistern; (3) Temporal subarachnoid space; and (4) Parietal subarachnoid space
of CSF flow in the subarachnoid space is definitely incriminated as a factor in the development of secondary hydrocephalus. We made an experimental study using dogs to explore the changes over time in the subarachnoid space after subarachnoid hemorrhage. To observe the changes we employed a scanning electron microscope (SEM).
Fig. 2. Subarachnoid space of the control dog. Upper left: Prepontine subarachnoid space containing cross section of the basilar artery. Upper right: Parietal subarachnoid space containing cross section of the vein. Lower left: Interpeduncular cistern. Lower right: Normal trabeculae in the interpeduncular cistern
Fig. 3. Basal subarachnoid space immediately after the blood injection. Left: Prepontine subarachnoid space filled with haematoma. Right: Magnified SEM photograph of the haematoma. Fibrin fibres were seen among the RBC
4
S. Suzuki et al.:
Material and Methods We used 57 adult mongrel dogs weighing 4.8-12.6 kg (6.8 kg on average), 14 of which were control cases for purposes of comparison. Forty-three were intracisternally injected with about 1 ml/kg of autogenous blood under intravenous anaesthesia with 3.5-4.0 mg/kg of Nembutal. The method was as follows: after confirming the outflow of CSF by means of cisternal puncture (the dog in lateral position with its head slightly lowered), blood collected by femoral artery puncture was immediately injected through the cisternal puncture needle. This
Fig. 4. Trabeculae and RBC in the interpeduncular cistern. Left: At lower magnification. Right: At higher magnification cisternal blood injection was given only once to each subject. They were then sacrificed at varying periods ranging from immediately after the blood injection (i.e., artificial subarachnoid hemorrhage) to six months: six dogs immediately after the injection, eight in two to five days, eight in about a week, eight in two to three weeks, six in a month, and seven in six weeks to six months. The animals were killed by vascular perfusion of the brain with 2-10% phosphate-buffered formaldehyde solution at the aortic arch after anaesthesia as described above. In view of the macroscopic findings in the brains removed immediately after the injection (Fig. 1 upper), specimens for observation were collected at four sites, selected for the following reasons (Fig. 1 lower): 1. The area surrounding the basilar artery or prepontine subarachnoid space had the most blood entrapped. 2. The interpeduncular cistern is rich in fibrous tissue in the subarachnoid space even in the normal state, and is easily preserved in the process of preparing the specimen. 3. The temporal snbarachnoid space is the furthermost area of diffusion of the injected blood. 4. The parietal subarachnoid space was an area where almost no blood reached. Each specimen was cut 2-3 mm thick, fixed by 50/0 glutaraldehyde solution, fixed again by 1~ osmic acid, dehydrated by alcohol, and gold coated; it was then observed and photographed with a SEM (Model MSM-4, made by Hitachi-Akashi). The Nissei Sangyo Company's ion coater IB 3 was used for the gold coating.
Subarachnoid Change After Haemorrhage
5
Results of Experiments
a) C o n t r o l cases. The prepontine subarachnoid space is scanty in fibrous structure and contains the basilar artery. Although the artery is apt to be deformed in specimen preparation it could be kept approximately in its original form by our vascular perfusion (Fig. 2 upper left).
Fig. 5. Subarachnoid haematoma at two to five days after blood injection. Upper left: Deformed RBC with fibrin fibres two days after the blood injection. Upper right: Deformed RBC without fibrin fibres in the same dog. Lower left: More deformed RBC with increased fibrin fibres five days after the blood injection. Lower right: Organized haematoma around the basilar artery five days after blood injection
The interpeduncular cistern abounds in fibrous structure even in normal circumstances (Fig. 2 lower left). The fibres were relatively uniform in thickness, and there was no suspicion of adhesions (Fig. 2 lower right). In the parietal and temporal regions the cerebral suM, in which
6
S. Suzuki et al.:
the structure of the subarachnoid space may be relatively well preserved, were selected for observation. Here there is also little fibrous tissue (Fig. 2 upper right). b) Sacrifice immediately after blood injection. The prepontine subarachnoid space was filled with blood round the basilar artery in the form of subarachnoid haematoma. The magnified picture of the haematoma showed a high content of fine material considered to be fibrin, together with red blood cells, which are the major element of the haematoma (Fig. 3). The interpeduncular cistern was also filled with haematoma in which were relatively thick fibres, which may be trabeculae disrupted by the blood, together with red blood cells and fibrin nets. In the temporal subarachnoid space a moderate number of blood cells were noted, in support of our macroscopic finding. There was no injected blood noted in the parietal subarachnoid space in most cases although a red blood cells were sometimes seen. Although there may be a possibility that the fibres we thought to be fibrin net were the disrupted trabeculae, our observation of the interpeduncular cistern in an area where red blood cells happened to be present togehter with the finest trabeculae, revealed that the trabeculae were obviously thicker than the fine fibres in the haematoma (Fig. 4). c) Sacrifice two to five days after the blood injection. Eight dogs were examined. It seemed to be in this period that the process of breakdown of red blood cells constituted a major change, although fibrin nets were observed to have temporarily increased in some cases. The specimens obtained from the dog sacrificed on the second day showed an increase in fibrin nets and concomitant deformation of red blood cells in one place, and only deformed red blood cells with no fibrin net in another (Fig. 5 upper). In some subjects sacrificed on the fifth day red blood cells in nearly their original form were present with a great number of fibrin nets in the prepontine subarachnoid space, while in other cases there was haematoma organization round the basilar artery (Fig. 5 lower). In all eight dogs in this group we found macroscopic evidence of haematoma, and in no case was there any indication of fibrosis on SEM examination. d) Sacrifice seven to ten days after the blood injection. In five of eight dogs in this group haematoma was macroscopically noted round the basilar artery, and in the remaining three no
Subarachnoid Change After Haemorrhage
7
blood was seen at all. In this period, thick and dense fibrous tissue thought to represent fibrosis begins to develop in the subarachnoid and subpial regions (Fig. 6 upper left). Under magnification, a number of infiltrating cells could be seen among the thickened fibres (Fig. 6 upper right). These findings were observed also in the parietal
Fig. 6. Subarachnoid fibrosis one week after blood injection. Upper left: Interpeduncular cistern. Upper right: Cell elements among the thickened fibres. Lower: Parietal subarachnoid space
subarachnoid space where almost no blood had reached (Fig. 6 lower). Incidentally, one of the dogs in this group, sacrificed on the ninth day, showed no increase at all in fibrous tissue.
e) Sacrifice fourteen to twenty-one days after the blood injection. In only two of eight dogs in this group could faint traces of haematoma be noted macroscopically round the basilar artery. In the remainder there was no evidence of blood. In all cases, fibres had increased in number (Fig. 7). This period seems to be the time when fibrosis develops most severely.
8
S. Suzuki et al.:
f) Sacrifice one month after blood injection. This group consists of six dogs which were sacrificed between the 27th and 36th days following injection. Three of the six showed fibrosis; in the remaining three findings approximately similar to those in the control dogs were obtained. The fibrosis in this period had relatively fine and not very dense fibres. The increased fibres
Fig. 7. Subarachnoid and subpial fibrosis two to three weeks after the blood injection. Upper left: Interpeduncular cistern. Upper right: Temporal subarachnoid space. Lower left: Subpial fibrosis at lower magnification. Lower right: Subpial fibrosis at higher magnification
tended to gather just below the arachnoid membrane in a condition which may be described as "arachnoid thickening". This was in agreement with the result of light microscopic observation of the hematoxylin and eosin stained specimen obtained from the same dog (Fig. 8).
Subarachnoid Change After Haemorrhage
9
g) Sacrifice six weeks to six months after the blood injection. One subject in this group was sacrificed on the 46th day, one each in the second, third, and fourth months, and three in the sixth month, making a total of seven. Four of these showed normal findings similar to those of the control dogs, while the dogs sacrificed in the second and fourth months showed obvious increases in fibres.
Fig. 8. Subarachnoid fibrosis one month after the blood injection. Left: Fibrosis marked at just under the arachnoid, presumably coinciding with arachnoid thickening. Right: Arachnoid thickening in the same dog. (HE staining and light microscopic finding)
In the dog examined in the fourth month, thick fibres were present round the basilar artery. Fibres were dense and thick in the interpeduncular cistern, and a remarkable thickening of the arachnoid membrane was observed in the parietal region, although there was a subarachnoid space here (Fig. 9). Discussion
To summarize the changes over 4 period of time in the subarachnoid space after the injection of blood, the experimental results are given as follows (Table 1). The injected blood was present mainly in the basal subarachnoid space in the form of a haematoma which seemed to disappear within one or two weeks. The increase of fibres in the subarachnoid space, or fibrosis, began to develop in about a week, as if it replaced the blood, reaching its peak in two to three weeks. It has been learned that the fibrosis can develop not only in the area where the blood was present, but also in the parietal region, a remote place.
S, Suzuki et al,:
10
To conclude, the fibrosis begins to subside after passing through the peak period, and the fibre material seems to return to normal in a month in cases of rapid recovery. In some cases, however, fibrosis persists for a long time, developing into chronic subarachnoid fibrosis. Although the factors contributing to the persistence of this fibrosis are obscure, it may be in such cases that shunt operations are necessary. Table I. Number of Dogs and Changes in Subarachnoid Space at Each Period After CisternaI Blood Injections
Control Immediately after the injection 2-5 days after the injection 7-10 days after the injection 2-3 weeks after the injection 1 month after the injection 1.5-6 months after the injection
Number of dogs
Subarachnoid hzematoma
14 6 8 8 8 6 7
/ 6 8 5 2 0 0
Subarachnoid fibrosis
With regard to the time at which fibrosis appears after hemorrhage, Kibler and Crompton (1961) 11 stated that fibrosis of the meninges requires approximately ten days to develop after subarachnoid hemorrhage, and Hammes (1944)10 who studied 53 autopsy cases of aneurysm, was of the same opinion. These statements are in accord with our results, in that we observed fibrosis in the one to two week period after blood injection. Bagley (1928) ~ mentioned meningeal cellularity but after several weeks the cells are less numerous in the meninges and a large amount of fibrous tissue is present. Hammes (1944) l0 also observed that the meningeal cellular reaction after subarachnoid hemorrhage is transient, and permanent effects occur in the form of fibrosis of the pia-arachnoid space which blocks the fluid pathways. Our experiment also revealed that substances considered to be cell elements appeared before the fibrosis peak. It is clear that severe fibrosis develops after subarachnoid hemorrhage. Hammes (1944) 10 rated the incidence of fibrosis at 55%. Judging from the results of our experiment, however, it may be said that its occurrence although inevitable, is variable in severity. The increased fibrous tissue is not always permanent as Hammes suggested, but returns to normal spontaneously in one to two months in many cases; in only certain cases is there chronic fibrosis which persists.
Subarachnoid Change After Haemorrhage
11
As to the cause of hydrocephalus resulting from subarachnoid haemorrhage, many authors have pointed out that the obstruction of CSF flow due to fibrosis or arachnoid-pia adhesions plays a role. Bagley (1928) ~ presented data on clinical cases where the ventricular dilatation and thickening of the arachnoid were concomitant. Moritz and Wartman (1938) 14 described four autopsy cases with post-traumatic hydrocephalus caused by the partial or complete
Fig. 9. Subarachnoid fibrosis and arachnoid thickening four months after blood injection. Upper left: Prepontine subara&noid space. Upper right: Interpeduncular cistern. Lower left: Temporal subarachnoid space. Lower right: Parietal arachnoid thi&ening fusion of the pia-arachnoid in the region of the fourth ventricle by fibrous adhesions. Krayenbiihl and Liithy (1948) 12 and Askenasy et al. (1953) ~ also reported cases in which ventricular dilatations could be attributed to the thickening, adhesion, and fibrosis of the arachnoid. On the other hand, there are some opinions that obstruction of
12
S. Suzuki et al.:
the arachnoid villi, which are the terminals of the CSF pathway and should be described as the absorbing or filtrating organs of the CSF, is principally responsible for the production of hydrocephalus. Ellington and Margolis (1969) 8 made a postmortem study of six brains and attributed the obstruction of CSF flow in the acute stage of subaraehnoid haemorrhage to the arachnoid villi having filled with blood cells, and the obstruction of CSF flow in the chronic stage to the production of fibrosis or siderosis in the arachnoid villi. DeLand et al. (1972) 7 were more deeply impressed, in autopsy cases of NPH, by obliteration of normal architecture of the arachnoid villi by connective tissue and chronic inflammatory cells in the vicinity of the superior sagittal sinus, the area of greatest CSF absorption, than by the fibrosis demonstrated in several parts of the brain surface. Foltz and Ward (1956) 9 found communicating hydrocephalus resulting from massive bleeding into the basal cisterns among ten cases of hydrocephalus following subarachnoid haemorrhage. We also demonstrated the haematoma filling the whole basal subarachnoid space in dogs sacrificed immediately after blood injection. Taking all of these findings into account, there is also be claim by Ellington et al. 8 that the blood cells per se are responsible for the obstruction of CSF flow even before the development of fibrosis. In their view, ventricular drainage 25 in the acute stage of subarachnoid haemorrhage due to aneurysmal rupture is useful. We have often encountered cases in which there was no ventricular dilatation despite the presence of subarachnoid haemorrhage or fibrosis. In our experiment, the ventricles were investigated in four dogs with fibrosis, two each on the ninth day and in the first month after blood injection, and they were found to be approximately normal. Hammes (1944) s0 also failed to demonstrate hydrocephalus in any of his 53 autopsy cases of subarachnoid haemorrhage, although there was some fibrosis. This, according to Hammes, is presumably due to patchy distribution of fibrosis, and the occurrence of hydrocephalus may depend upon the site and amount of the bleeding. Lastly, let us refer to the application of SEM to observation of the subarachnoid space. Cloyd and Low (1974) ~, and Malloy and Low (1974) 1~, made intravenous injections of silicone rubber (MICROFIL) to prevent the collapse of thin-walled veins when they made close SEM observations of the spinal subarachnoid space in dogs. The method we employed, perfusion of the brain by 2-10~ phosphatebuffered formaldehyde solution, also allows the specimens to be prepared without significant deformation of the vessels, if they are cut carefully. This method offered the additional merit of allowing the internal observation of the vessels (Fig. 10 left), if necessary.
Subarachnoid Change After Haemorrhage
i3
For cutting the fixed specimens we used a disposable scalpel as advised by Cloyd 6. This seemed to cause less damage to the cut section than a double-edged safety razor. Moreover, SEM is particularly useful for observing three-dimentional structure, and may give a clear picture also in the observation of the connection between the vessel walls and the trabeculae (Fig. 10 right), as practised by Arutiunov et aI. (1974) 1.
Fig. 10. Left: Inner surface of the artery. Aperture of the branch was well observed. Right: Trabeculae connecting leptomeninges and vessel wall References
1. Arutiunov, A. I., Baron, M. A., Majorova, N. A., The role of mechanical factors in the pathogenesis of short term and prolonged spasm of the cerebral arteries. J. Neurosurg. 40 (1974), 459--472. 2. Askenasy, H. M., Herzberger, E. E., Wijsenbeek, H. S., Hydrocephalus with vascular malformations of brain; preliminary report. Neurology 3 (1953), 213--220. 3. Bagley, C. Jr., Blood in the cerebrospinal fluid. Resultant functional and organic alterations in the central nervous system. A. Experimental data. Arch. Surg. (Chicago) 17 (1928), 18--38. 4. Bagley, C. Jr., Blood in the cerebrospinal fluid. Resultant functional and organic alterations in the central nervous system. B. Clinical data. Arch. Surg. (Chicago) 17 (1928), 39--81. 5. Cloyd, M. W., Low, F. N., Scanning electron microscopy of the subarachnoid space in the dog. I. Spinal cord levels. J. comp. neurol. 153 (1974), 325--368. 6. Cloyd, M. W., Personal communication. 1975. 7. DeLand F. H., James, A. E. Jr., Ladd, D. J., Konigsmark, B. W., Normal pressure hydrocephalus. A histologic study. Amer. J. din. Path. 58 (1972), 58--63. 8. Ellington, E., Margolis, G., Block of arachnoid villus by subarachnoid hemorrhage. J. Neurosurg. 30 (1969), 651--657.
14
S. Suzuki et aI.: Subarachnoid Change After Haemorrhage
9. Fortz, E. L., Ward, A. A., Communicating hydrocephalus from subarachnoid bleeding. J. Neurosurg. 13 (1956), 546--556. 10. Hammes, E. M. Jr., Reaction of the meninges to blood. Arch. Neurol. Psychiat. 52 (1944), 505--514. 11. Kibler, R. F., Crompton, M. R., Hydrocephalus in the adult following spontaneous subarachnoid hemorrhage. Brain 84 (1961), 45--61. 12. Krayenbiihl, H., Lilthy, F., Hydrocephalus als Sp~itfolge geplatzter basaler Hirnaneurysmen. Schweiz. Arch. Neurol. Psychiat. 61 (1948), 7--21. 13. Malloy, J. J., Low, F. N., Scanning electron microscopy of the subarachnoid space in the dog. II. Spinal nerve exists. J. comp. neurol. 157 (1974), 87--108. 14. Moritz, A. R., Wartman, W. B., Post-traumatic internal hydrocephalus. Amer. J. Med. Sci. 195 (1938), 65--70. 15. Suzuki, J., Yoshimoto, T., Hori, S., Continuous ventricular drainage to lessen surgical risk in ruptured intracranial aneurysm. Surg. Neurol. 2 (1974), 87--90. Author's address: Shigeharu Suzuki, M.D., Department of Neurosurgery, Hirosaki University School of Medicine, 5 Zaifu-cho, Hirosaki, 036 Japan.
Acta Neurochirurgica 39, 1--14 (1977)
9 by Springer-Verlag 1977
The Department of Neurosurgery, Hirosaki University School of Medicine, Hirosaki, Japan
Changes in the Subarachnoid Space After Experimental Subarachnoid Haemorrhage in the Dog: Scanning Electron Microscopic Observation By
S. Suzuki, M. Ishii, M. Ottomo, and T. Iwabuchi With 10 Figures
Summary The possible changes in the subarachnoid space after subarachnoid haemorrhage were studied in animals by using a scanning electron microscope (SEM). About 1 ml/kg of autogenous blood was injected intracisternally in 36 adult mongrel dogs to investigate changes in the subarachnoid space, over periods ranging from immediately after the injection to as long as 6 months. We have come to the conclusion that the injected blood disappears in about one to two weeks; the fibrosis or thickening of the arachnoid membrane appears in one to three weeks, and then returns to normal in a month in instances of rapid recovery, but there are some cases in which fibrosis persists for a long period and becomes chronic. The fact that an increase of fibrous tissue was found in the parietal region, where the injected blood had hardly reached, appears to indicate that the fibrosis is not always limited to the site of the baemorrhage but can occur in remote regions, We also discuss the usefulness of the SEM in the observation of the subarachnoid space, and the finding that vascular specimen preparations can be made by perfusing the brain with 2-10~ phosphate-buffered formaldehyde solution.
Key words: Subarachnoid hemorrhage, hydrocephalus, subarachnoid fibrosis, scanning electron microscope.
No definite explanation has yet been given for secondary hydrocephalus following subarachnoid hemorrhage, which may be classified as high-pressure hydrocephalus (HPH), or normal pressure hydrocephalus (NPH) in which cerebrospinal fluid (CSF) pressure is normal but psychic symptoms are predominant. Studies are being made of various aspects, with attention directed to disturbances, either of the production and absorption of CSF, or of CSF flow in the ventricular system and subarachnoid space. Acta Neurochirurg/ca, Vol. 89, Fast. 1--~
1
2
S. Suzuki et al.: Subarachnoid Change After Haemorrhage
Bagley (1928) ~ has demonstrated the production of ventricular dilatation by intracisternal blood injection in pups and pointed out that an increase of fibrous tissue was noted in the subarachnoid space. Since then, a large number of studies, mainly clinical, have reported hydrocephalus following subarachnoid hemorrhage, and changes in the subarachnoid space, in support of the view that the disturbance
Fig. 1. Upper: Basal view of the brain removed immediately after the intracisternal blood injection in the dog. Lower: Sites of specimen excision: (1) Prepontine subarachnoid space; (2) Interpeduncular cistern; (3) Temporal subarachnoid space; and (4) Parietal subarachnoid space
of CSF flow in the subarachnoid space is definitely incriminated as a factor in the development of secondary hydrocephalus. We made an experimental study using dogs to explore the changes over time in the subarachnoid space after subarachnoid hemorrhage. To observe the changes we employed a scanning electron microscope (SEM).
Fig. 2. Subarachnoid space of the control dog. Upper left: Prepontine subarachnoid space containing cross section of the basilar artery. Upper right: Parietal subarachnoid space containing cross section of the vein. Lower left: Interpeduncular cistern. Lower right: Normal trabeculae in the interpeduncular cistern
Fig. 3. Basal subarachnoid space immediately after the blood injection. Left: Prepontine subarachnoid space filled with haematoma. Right: Magnified SEM photograph of the haematoma. Fibrin fibres were seen among the RBC
4
S. Suzuki et al.:
Material and Methods We used 57 adult mongrel dogs weighing 4.8-12.6 kg (6.8 kg on average), 14 of which were control cases for purposes of comparison. Forty-three were intracisternally injected with about 1 ml/kg of autogenous blood under intravenous anaesthesia with 3.5-4.0 mg/kg of Nembutal. The method was as follows: after confirming the outflow of CSF by means of cisternal puncture (the dog in lateral position with its head slightly lowered), blood collected by femoral artery puncture was immediately injected through the cisternal puncture needle. This
Fig. 4. Trabeculae and RBC in the interpeduncular cistern. Left: At lower magnification. Right: At higher magnification cisternal blood injection was given only once to each subject. They were then sacrificed at varying periods ranging from immediately after the blood injection (i.e., artificial subarachnoid hemorrhage) to six months: six dogs immediately after the injection, eight in two to five days, eight in about a week, eight in two to three weeks, six in a month, and seven in six weeks to six months. The animals were killed by vascular perfusion of the brain with 2-10% phosphate-buffered formaldehyde solution at the aortic arch after anaesthesia as described above. In view of the macroscopic findings in the brains removed immediately after the injection (Fig. 1 upper), specimens for observation were collected at four sites, selected for the following reasons (Fig. 1 lower): 1. The area surrounding the basilar artery or prepontine subarachnoid space had the most blood entrapped. 2. The interpeduncular cistern is rich in fibrous tissue in the subarachnoid space even in the normal state, and is easily preserved in the process of preparing the specimen. 3. The temporal snbarachnoid space is the furthermost area of diffusion of the injected blood. 4. The parietal subarachnoid space was an area where almost no blood reached. Each specimen was cut 2-3 mm thick, fixed by 50/0 glutaraldehyde solution, fixed again by 1~ osmic acid, dehydrated by alcohol, and gold coated; it was then observed and photographed with a SEM (Model MSM-4, made by Hitachi-Akashi). The Nissei Sangyo Company's ion coater IB 3 was used for the gold coating.
Subarachnoid Change After Haemorrhage
5
Results of Experiments
a) C o n t r o l cases. The prepontine subarachnoid space is scanty in fibrous structure and contains the basilar artery. Although the artery is apt to be deformed in specimen preparation it could be kept approximately in its original form by our vascular perfusion (Fig. 2 upper left).
Fig. 5. Subarachnoid haematoma at two to five days after blood injection. Upper left: Deformed RBC with fibrin fibres two days after the blood injection. Upper right: Deformed RBC without fibrin fibres in the same dog. Lower left: More deformed RBC with increased fibrin fibres five days after the blood injection. Lower right: Organized haematoma around the basilar artery five days after blood injection
The interpeduncular cistern abounds in fibrous structure even in normal circumstances (Fig. 2 lower left). The fibres were relatively uniform in thickness, and there was no suspicion of adhesions (Fig. 2 lower right). In the parietal and temporal regions the cerebral suM, in which
6
S. Suzuki et al.:
the structure of the subarachnoid space may be relatively well preserved, were selected for observation. Here there is also little fibrous tissue (Fig. 2 upper right). b) Sacrifice immediately after blood injection. The prepontine subarachnoid space was filled with blood round the basilar artery in the form of subarachnoid haematoma. The magnified picture of the haematoma showed a high content of fine material considered to be fibrin, together with red blood cells, which are the major element of the haematoma (Fig. 3). The interpeduncular cistern was also filled with haematoma in which were relatively thick fibres, which may be trabeculae disrupted by the blood, together with red blood cells and fibrin nets. In the temporal subarachnoid space a moderate number of blood cells were noted, in support of our macroscopic finding. There was no injected blood noted in the parietal subarachnoid space in most cases although a red blood cells were sometimes seen. Although there may be a possibility that the fibres we thought to be fibrin net were the disrupted trabeculae, our observation of the interpeduncular cistern in an area where red blood cells happened to be present togehter with the finest trabeculae, revealed that the trabeculae were obviously thicker than the fine fibres in the haematoma (Fig. 4). c) Sacrifice two to five days after the blood injection. Eight dogs were examined. It seemed to be in this period that the process of breakdown of red blood cells constituted a major change, although fibrin nets were observed to have temporarily increased in some cases. The specimens obtained from the dog sacrificed on the second day showed an increase in fibrin nets and concomitant deformation of red blood cells in one place, and only deformed red blood cells with no fibrin net in another (Fig. 5 upper). In some subjects sacrificed on the fifth day red blood cells in nearly their original form were present with a great number of fibrin nets in the prepontine subarachnoid space, while in other cases there was haematoma organization round the basilar artery (Fig. 5 lower). In all eight dogs in this group we found macroscopic evidence of haematoma, and in no case was there any indication of fibrosis on SEM examination. d) Sacrifice seven to ten days after the blood injection. In five of eight dogs in this group haematoma was macroscopically noted round the basilar artery, and in the remaining three no
Subarachnoid Change After Haemorrhage
7
blood was seen at all. In this period, thick and dense fibrous tissue thought to represent fibrosis begins to develop in the subarachnoid and subpial regions (Fig. 6 upper left). Under magnification, a number of infiltrating cells could be seen among the thickened fibres (Fig. 6 upper right). These findings were observed also in the parietal
Fig. 6. Subarachnoid fibrosis one week after blood injection. Upper left: Interpeduncular cistern. Upper right: Cell elements among the thickened fibres. Lower: Parietal subarachnoid space
subarachnoid space where almost no blood had reached (Fig. 6 lower). Incidentally, one of the dogs in this group, sacrificed on the ninth day, showed no increase at all in fibrous tissue.
e) Sacrifice fourteen to twenty-one days after the blood injection. In only two of eight dogs in this group could faint traces of haematoma be noted macroscopically round the basilar artery. In the remainder there was no evidence of blood. In all cases, fibres had increased in number (Fig. 7). This period seems to be the time when fibrosis develops most severely.
8
S. Suzuki et al.:
f) Sacrifice one month after blood injection. This group consists of six dogs which were sacrificed between the 27th and 36th days following injection. Three of the six showed fibrosis; in the remaining three findings approximately similar to those in the control dogs were obtained. The fibrosis in this period had relatively fine and not very dense fibres. The increased fibres
Fig. 7. Subarachnoid and subpial fibrosis two to three weeks after the blood injection. Upper left: Interpeduncular cistern. Upper right: Temporal subarachnoid space. Lower left: Subpial fibrosis at lower magnification. Lower right: Subpial fibrosis at higher magnification
tended to gather just below the arachnoid membrane in a condition which may be described as "arachnoid thickening". This was in agreement with the result of light microscopic observation of the hematoxylin and eosin stained specimen obtained from the same dog (Fig. 8).
Subarachnoid Change After Haemorrhage
9
g) Sacrifice six weeks to six months after the blood injection. One subject in this group was sacrificed on the 46th day, one each in the second, third, and fourth months, and three in the sixth month, making a total of seven. Four of these showed normal findings similar to those of the control dogs, while the dogs sacrificed in the second and fourth months showed obvious increases in fibres.
Fig. 8. Subarachnoid fibrosis one month after the blood injection. Left: Fibrosis marked at just under the arachnoid, presumably coinciding with arachnoid thickening. Right: Arachnoid thickening in the same dog. (HE staining and light microscopic finding)
In the dog examined in the fourth month, thick fibres were present round the basilar artery. Fibres were dense and thick in the interpeduncular cistern, and a remarkable thickening of the arachnoid membrane was observed in the parietal region, although there was a subarachnoid space here (Fig. 9). Discussion
To summarize the changes over 4 period of time in the subarachnoid space after the injection of blood, the experimental results are given as follows (Table 1). The injected blood was present mainly in the basal subarachnoid space in the form of a haematoma which seemed to disappear within one or two weeks. The increase of fibres in the subarachnoid space, or fibrosis, began to develop in about a week, as if it replaced the blood, reaching its peak in two to three weeks. It has been learned that the fibrosis can develop not only in the area where the blood was present, but also in the parietal region, a remote place.
S, Suzuki et al,:
10
To conclude, the fibrosis begins to subside after passing through the peak period, and the fibre material seems to return to normal in a month in cases of rapid recovery. In some cases, however, fibrosis persists for a long time, developing into chronic subarachnoid fibrosis. Although the factors contributing to the persistence of this fibrosis are obscure, it may be in such cases that shunt operations are necessary. Table I. Number of Dogs and Changes in Subarachnoid Space at Each Period After CisternaI Blood Injections
Control Immediately after the injection 2-5 days after the injection 7-10 days after the injection 2-3 weeks after the injection 1 month after the injection 1.5-6 months after the injection
Number of dogs
Subarachnoid hzematoma
14 6 8 8 8 6 7
/ 6 8 5 2 0 0
Subarachnoid fibrosis
With regard to the time at which fibrosis appears after hemorrhage, Kibler and Crompton (1961) 11 stated that fibrosis of the meninges requires approximately ten days to develop after subarachnoid hemorrhage, and Hammes (1944)10 who studied 53 autopsy cases of aneurysm, was of the same opinion. These statements are in accord with our results, in that we observed fibrosis in the one to two week period after blood injection. Bagley (1928) ~ mentioned meningeal cellularity but after several weeks the cells are less numerous in the meninges and a large amount of fibrous tissue is present. Hammes (1944) l0 also observed that the meningeal cellular reaction after subarachnoid hemorrhage is transient, and permanent effects occur in the form of fibrosis of the pia-arachnoid space which blocks the fluid pathways. Our experiment also revealed that substances considered to be cell elements appeared before the fibrosis peak. It is clear that severe fibrosis develops after subarachnoid hemorrhage. Hammes (1944) 10 rated the incidence of fibrosis at 55%. Judging from the results of our experiment, however, it may be said that its occurrence although inevitable, is variable in severity. The increased fibrous tissue is not always permanent as Hammes suggested, but returns to normal spontaneously in one to two months in many cases; in only certain cases is there chronic fibrosis which persists.
Subarachnoid Change After Haemorrhage
11
As to the cause of hydrocephalus resulting from subarachnoid haemorrhage, many authors have pointed out that the obstruction of CSF flow due to fibrosis or arachnoid-pia adhesions plays a role. Bagley (1928) ~ presented data on clinical cases where the ventricular dilatation and thickening of the arachnoid were concomitant. Moritz and Wartman (1938) 14 described four autopsy cases with post-traumatic hydrocephalus caused by the partial or complete
Fig. 9. Subarachnoid fibrosis and arachnoid thickening four months after blood injection. Upper left: Prepontine subara&noid space. Upper right: Interpeduncular cistern. Lower left: Temporal subarachnoid space. Lower right: Parietal arachnoid thi&ening fusion of the pia-arachnoid in the region of the fourth ventricle by fibrous adhesions. Krayenbiihl and Liithy (1948) 12 and Askenasy et al. (1953) ~ also reported cases in which ventricular dilatations could be attributed to the thickening, adhesion, and fibrosis of the arachnoid. On the other hand, there are some opinions that obstruction of
12
S. Suzuki et al.:
the arachnoid villi, which are the terminals of the CSF pathway and should be described as the absorbing or filtrating organs of the CSF, is principally responsible for the production of hydrocephalus. Ellington and Margolis (1969) 8 made a postmortem study of six brains and attributed the obstruction of CSF flow in the acute stage of subaraehnoid haemorrhage to the arachnoid villi having filled with blood cells, and the obstruction of CSF flow in the chronic stage to the production of fibrosis or siderosis in the arachnoid villi. DeLand et al. (1972) 7 were more deeply impressed, in autopsy cases of NPH, by obliteration of normal architecture of the arachnoid villi by connective tissue and chronic inflammatory cells in the vicinity of the superior sagittal sinus, the area of greatest CSF absorption, than by the fibrosis demonstrated in several parts of the brain surface. Foltz and Ward (1956) 9 found communicating hydrocephalus resulting from massive bleeding into the basal cisterns among ten cases of hydrocephalus following subarachnoid haemorrhage. We also demonstrated the haematoma filling the whole basal subarachnoid space in dogs sacrificed immediately after blood injection. Taking all of these findings into account, there is also be claim by Ellington et al. 8 that the blood cells per se are responsible for the obstruction of CSF flow even before the development of fibrosis. In their view, ventricular drainage 25 in the acute stage of subarachnoid haemorrhage due to aneurysmal rupture is useful. We have often encountered cases in which there was no ventricular dilatation despite the presence of subarachnoid haemorrhage or fibrosis. In our experiment, the ventricles were investigated in four dogs with fibrosis, two each on the ninth day and in the first month after blood injection, and they were found to be approximately normal. Hammes (1944) s0 also failed to demonstrate hydrocephalus in any of his 53 autopsy cases of subarachnoid haemorrhage, although there was some fibrosis. This, according to Hammes, is presumably due to patchy distribution of fibrosis, and the occurrence of hydrocephalus may depend upon the site and amount of the bleeding. Lastly, let us refer to the application of SEM to observation of the subarachnoid space. Cloyd and Low (1974) ~, and Malloy and Low (1974) 1~, made intravenous injections of silicone rubber (MICROFIL) to prevent the collapse of thin-walled veins when they made close SEM observations of the spinal subarachnoid space in dogs. The method we employed, perfusion of the brain by 2-10~ phosphatebuffered formaldehyde solution, also allows the specimens to be prepared without significant deformation of the vessels, if they are cut carefully. This method offered the additional merit of allowing the internal observation of the vessels (Fig. 10 left), if necessary.
Subarachnoid Change After Haemorrhage
i3
For cutting the fixed specimens we used a disposable scalpel as advised by Cloyd 6. This seemed to cause less damage to the cut section than a double-edged safety razor. Moreover, SEM is particularly useful for observing three-dimentional structure, and may give a clear picture also in the observation of the connection between the vessel walls and the trabeculae (Fig. 10 right), as practised by Arutiunov et aI. (1974) 1.
Fig. 10. Left: Inner surface of the artery. Aperture of the branch was well observed. Right: Trabeculae connecting leptomeninges and vessel wall References
1. Arutiunov, A. I., Baron, M. A., Majorova, N. A., The role of mechanical factors in the pathogenesis of short term and prolonged spasm of the cerebral arteries. J. Neurosurg. 40 (1974), 459--472. 2. Askenasy, H. M., Herzberger, E. E., Wijsenbeek, H. S., Hydrocephalus with vascular malformations of brain; preliminary report. Neurology 3 (1953), 213--220. 3. Bagley, C. Jr., Blood in the cerebrospinal fluid. Resultant functional and organic alterations in the central nervous system. A. Experimental data. Arch. Surg. (Chicago) 17 (1928), 18--38. 4. Bagley, C. Jr., Blood in the cerebrospinal fluid. Resultant functional and organic alterations in the central nervous system. B. Clinical data. Arch. Surg. (Chicago) 17 (1928), 39--81. 5. Cloyd, M. W., Low, F. N., Scanning electron microscopy of the subarachnoid space in the dog. I. Spinal cord levels. J. comp. neurol. 153 (1974), 325--368. 6. Cloyd, M. W., Personal communication. 1975. 7. DeLand F. H., James, A. E. Jr., Ladd, D. J., Konigsmark, B. W., Normal pressure hydrocephalus. A histologic study. Amer. J. din. Path. 58 (1972), 58--63. 8. Ellington, E., Margolis, G., Block of arachnoid villus by subarachnoid hemorrhage. J. Neurosurg. 30 (1969), 651--657.
14
S. Suzuki et aI.: Subarachnoid Change After Haemorrhage
9. Fortz, E. L., Ward, A. A., Communicating hydrocephalus from subarachnoid bleeding. J. Neurosurg. 13 (1956), 546--556. 10. Hammes, E. M. Jr., Reaction of the meninges to blood. Arch. Neurol. Psychiat. 52 (1944), 505--514. 11. Kibler, R. F., Crompton, M. R., Hydrocephalus in the adult following spontaneous subarachnoid hemorrhage. Brain 84 (1961), 45--61. 12. Krayenbiihl, H., Lilthy, F., Hydrocephalus als Sp~itfolge geplatzter basaler Hirnaneurysmen. Schweiz. Arch. Neurol. Psychiat. 61 (1948), 7--21. 13. Malloy, J. J., Low, F. N., Scanning electron microscopy of the subarachnoid space in the dog. II. Spinal nerve exists. J. comp. neurol. 157 (1974), 87--108. 14. Moritz, A. R., Wartman, W. B., Post-traumatic internal hydrocephalus. Amer. J. Med. Sci. 195 (1938), 65--70. 15. Suzuki, J., Yoshimoto, T., Hori, S., Continuous ventricular drainage to lessen surgical risk in ruptured intracranial aneurysm. Surg. Neurol. 2 (1974), 87--90. Author's address: Shigeharu Suzuki, M.D., Department of Neurosurgery, Hirosaki University School of Medicine, 5 Zaifu-cho, Hirosaki, 036 Japan.
