Germinal Matrix Hemorrhage Origin in Preterm Neonates
of Venous
YASUHIRO NAKAMURA MD, TOSHIO OKUDERA MD, SEIICHI FUKUDA MD, AND TAKE0 HASHIMOTO, MD The rupturle point of germinal matrix hemorrhage in premature neonates was examined by postmortem angiography. Thirteen cases of germinal matrix hemorrhage were injected via artery (four cases), vein (five cases), and both artery and vein simultaneously (four cases). Only material injected via vein leaked into the hemorrhage, which was confirmed by stereomicroscopic and histologic examination. This study suggests that germinal matrix hemorrhage is venous in origin. HUM PATHOL 21:1059-1062. ‘B 1990 by W..B. Saunders Company.
Intraventricular hemorrhage (IVH) occurs predominantly in preterm neonates of less than 32 gestational weeks of age and less than 1,500 g birth weight.‘.‘? ?‘he most common source of IVH found in Ijremature infants is germinal matrix hemorrhage (GMH).:’ ‘IIe germinal matrix is a feature of the developing brain and much of it lies in the lateral walls of the lateral ventricles. It is a well-vascularized zone with fragile vessels, structurally considered as capillaries. formed by a single layer of endothelial cells.+ The etiology of GMH and IVH has been discussed as either venous or arterial in origin. A number of investigators have stated that the bleeding originates from the terminal vein or from smaller vein branches; the sharp angle at which the terminal veins join the internal cerebral veins at the foramen of ‘Monro has been credited with causing a point of vulnerability in the deep vein drainage system.‘.” On the other hand, the work of Hambleton and Wigglesworth’ pointed to disruption of the arterial microcirculation of the germinal matrix as the origin of GMH. Capillaries within the germinal matrix appear to be unable to withstand abrupt changes in blood pressure or flow so that GMH may result from elevation of arterial blood pressure, which is due to cerebral vasodilatation caused by hypercapnea and hypoxia. This study was undertaken to identify and characterize the venous or arterial blood vessels in the germinal matrix layer that are responsible for GMH.
From the Departments of Pathology, Neuroradiology, and Neonatolugy. St. Mary’s Hospital, Kurume. ,Japan. Accepted for publication November 14, I WY. Supported by grant 63-.4 from the National Crnter of Neurology and Ps\ch&try (NCNP) of the Ministr\ of Health and Welfare, Japan. k:q wool\: germinal matrix. hemmorhage, venous origin. Address correspondence and reprint requests to Yasuhiro Nakamura, MD. Department of Pathology. St Maw’s Hospital. 422, Tcub~lkuhon-machi. Kurume-shi, Japan 830. e 1989 b\ W.B. Saunders Company. 0046-X 177/90/? I I o-00 10$5.00/0
MATERIALS
AND METHODS
Fifty-tour autopsy neonates were selected for postmortem anglography. The internal carotid arteries and jugular veins were isolated at autopsy, and a nylon intravenous catheter was inserted into an artery, a vein, or both. Barium with 5% gelatin solution was injected into the artery or vein. White-colored barium solution was used when the injection was solely arterial or venous. In the cases of simultaneous in,jection of both artery and vein, a red solution was injected into the artery and a white one was injected into the vein. When the injection was completed. the brain was dissected and fixed in 10% buffered formalin for 2 weeks. After fixation, the injected brain was exposed to soft x-ray stereoradiography. The brain was then examined by coronal, tangential. or horizontal slicing. There was also stereoradiography of the thin slices. The slices of the injected brain (0 mm) were dehydrated in ethyl alcohol and cleared in methyl salicylate. The blocks were retained in methyl salicylate and examined under the stereomicroscope. Some blocks were returned to alcohol and taken for histology. Hematoxylin-eosin-stained sections were observed under the light microscope. Of 54 brains injected, 13 cases of GMH, with or without IVH, were found. Arterial in,jection was performed solely in four cases, venous injection solely in five cases, and simultaneous arterial and venous injection in four cases. In all 13 examined cases, autopsy was performed within 12 hours after death, usually within 4 hours after death.
TABLE 1. Case No.
Gestational Age (wk)
Solely arterial injection I 27 2 26 3 24 4 “8 Solely venous injection 5 30 6 7 8 9 Simultaneous 10 I1 12 I3
27 26 27 27 arterial 23 26 25 28
Clinical and Pathologic Birth Weight (g) group 90.5 91.5 570 1.060 group 1,450
Age Cd)
Data
Sex
Site of GMH
IVH
0
F M F M
a, b a, c c a
+ + +
2 I9
1 3
M
a, b
-
1.010 460 865 755
.i ?I I 0
F F F hl
b, c a,b,c a. b a
+ + + _
and venous 488 970 1,043 1,040
injection 2 2 9 1
group F M M M
a a, h h a,b,d
+ + + +
Abbreviations: a. caput of caudate lamic groove; c. tail of caudate nucleus: lateral ventricle.
nucleus; d. lateral
b, caudate thawall of body of
HUMAN PATHOLOGY
Volume 21, No. IO (October 1990)
FIGURE 1. (Left) Coronal section of cleared solely venous injected specimen [at the level of caput of caudate nucleus] shows GMH with leakage of injected material (black circle) drained into subependymal veins (arrowheads), and branches of inferior striate vein (INF). C, Caput of caudate nucleus. (Right) Hemotoxylin-eosin staining reveals leakage of injected material (black circle) in the GMH. [Magnification x 20.)
FIGURE 2. (Left) Coronal section of cleared solely arterial injected specimen (at the level of caput of caudate nucleus) shows GMH without leakage of injected material (arrowheads] despite repletion of the terminal arterioles with injected material, C, Caudate nucleus. (Right) Hematoxylin-eosin staining reveals no leakage of injected material into GMH [black circle) despite repletion of the terminal arterioles farrowheadsl with iniected material. [Maanification x 10.1
1060
GERMINAL MATRIX HEMORRHAGE (Nakamura et
al)
FIGURE 3. (Top left] Coronal section of cleared simultaneoLs arterial and venous injected specimen [white, veins; red, arteries), (at the level of caudate tholamic groove) shows GMH with leakage of Injected material [black circle], subependymal veins [black arrowheads), lateral striate branches of the middle cerebral artery (arrows), and medial striate branches of the anterior cerebral artery [white arrowheads). AC, Anterior cerebral artery, MC, middle cerebral artery. (Top right] Saggital section of contralaterol GMH on coudate thalamic groove shows leakage of injected material as a mass drained into subependymal veins (arrowheads). (Bottom] Hematoxylin-eo:;ln staining reveals mass of leaked Inlected material into GMH. (Magnification x IO.)
RESULTS Thv clinical and pathologic data of the 13 cases are shovcn in ‘I‘able 1. GMH was found over the caput of the caudate nucleus directly opposite the foramen of hIonro, overt the caudate thalamic groove, over the cauda of the caudate nucleus, and over the lateral wall along the body of lateral ventricles which are the sites of abundant germinal matrix. The arterial supply varied with the site: the medial striate branches ot the anterior cerebral artery. the lateral striate branches of the middle cerebral artery, the anterior and posterior choroidal arteries, and the medullary branches of’ the anterior and middle cerebral arteries were thfe vessels that supplied the corresponding sites involved. The venous drainage from all sites of GMH was into the subependymal veins, of which most branches drained into the internal cerebral vein via terminal vein or via anterior caudate vein, and a small number of’ branches drained into the basal vein. Germinal matrix was abundant with the small arterioles, cenules. and capillaries.
M’ith solely venous side injection, leakage of injected material was clearly found in the GMH of all five cases (Fig 1). With solely arterial injection, leakage of.injected material was not found in anv of four cases, despite repletion of the terminal artercoles with barium gelatin (Fig 2). With simultaneous arterial and venous injections, only white, barium gelatin solution was leaked into the
of Venous
YASUHIRO NAKAMURA MD, TOSHIO OKUDERA MD, SEIICHI FUKUDA MD, AND TAKE0 HASHIMOTO, MD The rupturle point of germinal matrix hemorrhage in premature neonates was examined by postmortem angiography. Thirteen cases of germinal matrix hemorrhage were injected via artery (four cases), vein (five cases), and both artery and vein simultaneously (four cases). Only material injected via vein leaked into the hemorrhage, which was confirmed by stereomicroscopic and histologic examination. This study suggests that germinal matrix hemorrhage is venous in origin. HUM PATHOL 21:1059-1062. ‘B 1990 by W..B. Saunders Company.
Intraventricular hemorrhage (IVH) occurs predominantly in preterm neonates of less than 32 gestational weeks of age and less than 1,500 g birth weight.‘.‘? ?‘he most common source of IVH found in Ijremature infants is germinal matrix hemorrhage (GMH).:’ ‘IIe germinal matrix is a feature of the developing brain and much of it lies in the lateral walls of the lateral ventricles. It is a well-vascularized zone with fragile vessels, structurally considered as capillaries. formed by a single layer of endothelial cells.+ The etiology of GMH and IVH has been discussed as either venous or arterial in origin. A number of investigators have stated that the bleeding originates from the terminal vein or from smaller vein branches; the sharp angle at which the terminal veins join the internal cerebral veins at the foramen of ‘Monro has been credited with causing a point of vulnerability in the deep vein drainage system.‘.” On the other hand, the work of Hambleton and Wigglesworth’ pointed to disruption of the arterial microcirculation of the germinal matrix as the origin of GMH. Capillaries within the germinal matrix appear to be unable to withstand abrupt changes in blood pressure or flow so that GMH may result from elevation of arterial blood pressure, which is due to cerebral vasodilatation caused by hypercapnea and hypoxia. This study was undertaken to identify and characterize the venous or arterial blood vessels in the germinal matrix layer that are responsible for GMH.
From the Departments of Pathology, Neuroradiology, and Neonatolugy. St. Mary’s Hospital, Kurume. ,Japan. Accepted for publication November 14, I WY. Supported by grant 63-.4 from the National Crnter of Neurology and Ps\ch&try (NCNP) of the Ministr\ of Health and Welfare, Japan. k:q wool\: germinal matrix. hemmorhage, venous origin. Address correspondence and reprint requests to Yasuhiro Nakamura, MD. Department of Pathology. St Maw’s Hospital. 422, Tcub~lkuhon-machi. Kurume-shi, Japan 830. e 1989 b\ W.B. Saunders Company. 0046-X 177/90/? I I o-00 10$5.00/0
MATERIALS
AND METHODS
Fifty-tour autopsy neonates were selected for postmortem anglography. The internal carotid arteries and jugular veins were isolated at autopsy, and a nylon intravenous catheter was inserted into an artery, a vein, or both. Barium with 5% gelatin solution was injected into the artery or vein. White-colored barium solution was used when the injection was solely arterial or venous. In the cases of simultaneous in,jection of both artery and vein, a red solution was injected into the artery and a white one was injected into the vein. When the injection was completed. the brain was dissected and fixed in 10% buffered formalin for 2 weeks. After fixation, the injected brain was exposed to soft x-ray stereoradiography. The brain was then examined by coronal, tangential. or horizontal slicing. There was also stereoradiography of the thin slices. The slices of the injected brain (0 mm) were dehydrated in ethyl alcohol and cleared in methyl salicylate. The blocks were retained in methyl salicylate and examined under the stereomicroscope. Some blocks were returned to alcohol and taken for histology. Hematoxylin-eosin-stained sections were observed under the light microscope. Of 54 brains injected, 13 cases of GMH, with or without IVH, were found. Arterial in,jection was performed solely in four cases, venous injection solely in five cases, and simultaneous arterial and venous injection in four cases. In all 13 examined cases, autopsy was performed within 12 hours after death, usually within 4 hours after death.
TABLE 1. Case No.
Gestational Age (wk)
Solely arterial injection I 27 2 26 3 24 4 “8 Solely venous injection 5 30 6 7 8 9 Simultaneous 10 I1 12 I3
27 26 27 27 arterial 23 26 25 28
Clinical and Pathologic Birth Weight (g) group 90.5 91.5 570 1.060 group 1,450
Age Cd)
Data
Sex
Site of GMH
IVH
0
F M F M
a, b a, c c a
+ + +
2 I9
1 3
M
a, b
-
1.010 460 865 755
.i ?I I 0
F F F hl
b, c a,b,c a. b a
+ + + _
and venous 488 970 1,043 1,040
injection 2 2 9 1
group F M M M
a a, h h a,b,d
+ + + +
Abbreviations: a. caput of caudate lamic groove; c. tail of caudate nucleus: lateral ventricle.
nucleus; d. lateral
b, caudate thawall of body of
HUMAN PATHOLOGY
Volume 21, No. IO (October 1990)
FIGURE 1. (Left) Coronal section of cleared solely venous injected specimen [at the level of caput of caudate nucleus] shows GMH with leakage of injected material (black circle) drained into subependymal veins (arrowheads), and branches of inferior striate vein (INF). C, Caput of caudate nucleus. (Right) Hemotoxylin-eosin staining reveals leakage of injected material (black circle) in the GMH. [Magnification x 20.)
FIGURE 2. (Left) Coronal section of cleared solely arterial injected specimen (at the level of caput of caudate nucleus) shows GMH without leakage of injected material (arrowheads] despite repletion of the terminal arterioles with injected material, C, Caudate nucleus. (Right) Hematoxylin-eosin staining reveals no leakage of injected material into GMH [black circle) despite repletion of the terminal arterioles farrowheadsl with iniected material. [Maanification x 10.1
1060
GERMINAL MATRIX HEMORRHAGE (Nakamura et
al)
FIGURE 3. (Top left] Coronal section of cleared simultaneoLs arterial and venous injected specimen [white, veins; red, arteries), (at the level of caudate tholamic groove) shows GMH with leakage of Injected material [black circle], subependymal veins [black arrowheads), lateral striate branches of the middle cerebral artery (arrows), and medial striate branches of the anterior cerebral artery [white arrowheads). AC, Anterior cerebral artery, MC, middle cerebral artery. (Top right] Saggital section of contralaterol GMH on coudate thalamic groove shows leakage of injected material as a mass drained into subependymal veins (arrowheads). (Bottom] Hematoxylin-eo:;ln staining reveals mass of leaked Inlected material into GMH. (Magnification x IO.)
RESULTS Thv clinical and pathologic data of the 13 cases are shovcn in ‘I‘able 1. GMH was found over the caput of the caudate nucleus directly opposite the foramen of hIonro, overt the caudate thalamic groove, over the cauda of the caudate nucleus, and over the lateral wall along the body of lateral ventricles which are the sites of abundant germinal matrix. The arterial supply varied with the site: the medial striate branches ot the anterior cerebral artery. the lateral striate branches of the middle cerebral artery, the anterior and posterior choroidal arteries, and the medullary branches of’ the anterior and middle cerebral arteries were thfe vessels that supplied the corresponding sites involved. The venous drainage from all sites of GMH was into the subependymal veins, of which most branches drained into the internal cerebral vein via terminal vein or via anterior caudate vein, and a small number of’ branches drained into the basal vein. Germinal matrix was abundant with the small arterioles, cenules. and capillaries.
M’ith solely venous side injection, leakage of injected material was clearly found in the GMH of all five cases (Fig 1). With solely arterial injection, leakage of.injected material was not found in anv of four cases, despite repletion of the terminal artercoles with barium gelatin (Fig 2). With simultaneous arterial and venous injections, only white, barium gelatin solution was leaked into the
