Hemorrhage In Rabbits Research Articles

Effect of calcitonin gene-related peptide on delayed cerebral vasospasm after experimental subarachnoid hemorrhage in rabbits

Calcitonin gene-related peptide (CGRP) is an intrinsic vasodilatory substance contained in perivascular nerve fibers of intracranial arteries. It is suggested that CGRP plays a role in cerebral vasospasm after subarachnoid hemorrhage (SAH). An experimental SAH was produced by intracisternal injection of arterial blood in rabbits. The animals were treated with intrathecal administration of CGRP solution 3 days after SAH. The degree of vasospasm and the effect of CGRP were evaluated angiographically by measuring the basilar artery diameter. The basilar artery constricted to 73.0% of the pre-SAH values 3 days after SAH. Fifteen minutes after injection of 10(-10) mol/kg CGRP, the basilar artery dilated to 117.1% (n = 8), which was significantly larger than 67.1% in the vehicle group (n = 8) (p < 0.01). The significant vasodilatory effect of CGRP, compared with the vehicle group, lasted for 6 hours. Intrathecal administration of CGRP has therapeutic potential for treating cerebral vasospasm.

Prevention of cerebrovasospasm following subarachnoid hemorrhage in rabbits by the platelet-activating factor antagonist, E5880.

Recently, an important role of platelet-activating factor (PAF), an inflammation mediator, has been demonstrated in the genesis of cerebral vasospasm following subarachnoid hemorrhage (SAH). In the current study, the authors examined whether intravenous administration of the novel PAF antagonist, E5880, can prevent vasospasm following SAH in rabbits. A vasospasm model was produced in three groups of rabbits using two subarachnoid injections of autologous arterial blood, followed by intravenous administration of distilled water (control), a low dose of E5880 (0.1 mg/kg in distilled water), or a high dose of E5880 (0.5 mg/kg in distilled water). Neurological deterioration was largely prevented in the rabbits that received E5880. Basilar artery constriction was also reduced by both doses of E5880. Histological examination at autopsy predominantly showed ischemic changes in the brain. Animals in each E5880-treated group exhibited ischemic changes less frequently than those in the control group. Plasma thromboxane B2 concentrations were reduced in rabbits treated with E5880. Platelet-activating factor was immunolocalized in the intima and media of the basilar artery in the control group. The PAF immunoreactivity demonstrated in the basilar artery was decreased in the E5880 groups in a dose-dependent manner. Thus, this study provides evidence that PAF may play a role in the pathogenesis of vasospasm after SAH and that intravenous administration of E5880 is a promising approach in preventing vasospasm.

Reflex control of vasopressin during hemorrhage in hypertensive rabbits.

Hypertensive (HT) rabbits have impaired reflex control of heart rate and vascular tone, which is, at least in part, related to dysfunctional baroreceptors. We hypothesized that reflex control of vasopressin (AVP) would also be impaired in the HT rabbit. To test this, we compared hemorrhage-induced increases in AVP between conscious normotensive (NT) and HT rabbits. The hemorrhage-induced rise in AVP was found to be significantly (P < 0.05) attenuated in the HT rabbits. We tested a second hypothesis, that the observed impairment was related to arterial baroreceptor function, by hemorrhaging the same rabbits after reversible cardiac denervation. Under these conditions, only the arterial baroreceptors would be expected to contribute to reflex control of AVP. Impairment was still evident after cardiac denervation; that is, the hemorrhage-induced rise in AVP was significantly (P < 0.01) attenuated in the cardiac-denervated HT rabbits compared with NT rabbits. Thus impairment was, at least in part, related to arterial baroreceptors. Previously, we showed in NT rabbits that AVP only contributed to maintenance of arterial pressure (during hemorrhage), after the autonomic nervous system (ANS) had been blocked. Thus, in the present study, we compared the maintenance of arterial pressure during hemorrhage between NT and HT rabbits after ANS blockade. Blood pressure maintenance was significantly attenuated in the HT rabbits (P < 0.05). In addition, for a given fall in pressure, significantly less AVP (P < 0.05) was released in the ANS-blocked HT rabbits as compared with NT rabbits.(ABSTRACT TRUNCATED AT 250 WORDS)

Influence of nitric oxide on the hemodynamic response to hemorrhage in conscious rabbits

We investigated the role of nitric oxide, an endothelium-derived relaxing factor, in the hemodynamic response to acute hemorrhage in conscious rabbits. Chronically instrumented rabbits were treated with the nitric oxide synthase inhibitor N-nitro-L-arginine methyl ester (L-NAME) or vehicle and hemorrhaged until mean arterial pressure fell below 40 mmHg. Control animals were treated with L-NAME or vehicle but not subjected to hemorrhage. L-NAME increased mean arterial pressure and decreased heart rate in control animals. Hindquarters and mesenteric blood flow velocity and conductance were reduced by L-NAME. Nitric oxide synthase inhibition also produced significant changes in the hemodynamic response to hypotensive hemorrhage. Mean arterial pressure was higher and regional vascular conductances were lower throughout hemorrhage and during recovery. L-NAME treatment significantly (but in some cases, subtly) altered the characteristic pattern of changes in vascular conductance associated with acute hypotensive hemorrhage and recovery. Similar experiments with other arginine analogues or phenylephrine infusion showed that L-NAME's effects during hemorrhage were due to nitric oxide synthase inhibition. We conclude that nitric oxide plays a role in the hemodynamic response to acute hemorrhage in the rabbit and is essential for the full expression of the vasodilation associated with hypotensive hemorrhage.

Effects of protease inhibitor and immunosuppressant on cerebral vasospasm after subarachnoid hemorrhage in rabbits

The possible role of the immune-defense system in the development of cerebral vasospasm after subarachnoid hemorrhage (SAH) was investigated in rabbits. We used a synthetic serine protease inhibitor, gabexate mesilate (GM), a glucocorticoid, betamethasone sodium phosphate (B-P), and an immunosuppressant, ciclosporin (Cyclosporin A, CYA), to prevent cerebral vasospasm. These agents were administered intra-venously every 12 hours for three injections, starting 20 minutes after SAH. In the group treated with GM, B-P, or CYA, there were no statistically significant differences in arterial calibers between treated and untreated controls on day 2. The synthetic serine protease inhibitor, FUT-175 has been reported to prevent cerebral vasospasm when the treatment is started 20 minutes after SAH in rabbits [38]. In rabbits treated with FUT-175 at different starting times from 3 to 6 hours, reductions in arterial caliber on day 2 were significantly prevented in each group. The contrasting effects of the two serine protease inhibitors, GM and FUT-175, are discussed.

Splanchnic Release of Potassium After Hemorrhage and Succinylcholine in Rabbits

Recent clinical reports document that hyperkalemia may occur after succinylcholine (SCh) administration in hemorrhagic, acidotic humans. Hemorrhagic, acidotic rabbits are a useful model for study of this phenomenon because of profound hyperkalemia after SCh. To determine the etiology of this hyperkalemia, we anesthetized rabbits (n = 5, Group H) with halothane/N2O, and after endotracheal intubation, placed catheters in the femoral artery, femoral vein (FV), and inferior vena cava (IVC), the latter at or above the hepatic veins. After measurement of baseline K+ (from each catheter) and arterial blood gases, approximately 30 mL/kg of blood was withdrawn, and the animal was allowed to become acidotic until pH approximately 7.00-7.10. Plasma K+ was determined again, and SCh 1 mg/kg administered intravenously. Blood from each catheter was withdrawn for K+ analysis at 1, 3, 5, and 7 min after SCh. Control rabbits (Group C, n = 5) were similarly anesthetized and catheterized but were not hemorrhaged. Arterial blood gas and K+ analyses were performed before and after SCh 1 mg/kg, similar to Group H. In Group H, arterial K+ concentration increased after hemorrhage (3.2 +/- 0.3 mmol/L to 6.4 +/- 2.0 mmol/L, P < 0.05). SCh caused a further increase, peaking at 8.3 +/- 1.5 mmol/L at 7 min (P < 0.05). In Group H the IVC-FV K+ difference was 1.6 +/- 1.9 mmol/L, 1.5 +/- 2.1 mmol/L, 0.9 +/- 1.6 mmol/L and 1.6 +/- 1.4 mmol/L, respectively, for the 1-, 3-, 5-, and 7-min periods after SCh.(ABSTRACT TRUNCATED AT 250 WORDS)

Reduced cerebral blood flow but intact reactivity to hypercarbia and hypoxia following subarachnoid hemorrhage in rabbits.

Subarachnoid hemorrhage (SAH) was produced in rabbits by four subarachnoid injections of blood (n = 7) or saline (n = 6); a control group (n = 6) had no injections. Basilar artery vasospasm was assessed by serial angiograms. Resting CBF (microspheres) and CBF reactivity to hypercapnia (65 and 85 mm Hg) and hypoxia (fractions of inspired oxygen of 0.15 and 0.10) were determined. Basilar artery vasospasm was seen with SAH. Resting CBF was reduced by 31% (SAH 43 +/- 12, saline 65 +/- 17, control 60 +/- 21 ml 100 g-1 min-1), and resting cerebrovascular resistance was increased (SAH 1.84 +/- 0.30, saline 1.31 +/- 0.49, control 1.39 +/- 0.25 mm Hg ml-1 100 g-1 min-1) after SAH. CBF rose to a similar degree in all three groups in response to hypercarbia and hypoxia. We conclude that resting CBF is reduced in this model of SAH, but vascular reactivity remains intact.

Volume compensation during cyclic hemorrhage in conscious rabbits

Volume compensation during cyclic hemorrhage in conscious rabbits

DOES THE HAEMODYNAMIC RESPONSE TO ACUTE CENTRAL HYPOVOLAEMIA DEPEND ON THE RATE OF FALL OF CARDIAC OUTPUT?

1. In published studies of the effects of acute blood loss in conscious rabbits, the rates of haemorrhage have ranged for 3-9% of blood volume/min. This is potentially a confounding factor when it comes to comparing the results of different studies. We have therefore tested whether the haemodynamic response to acute central hypovolaemia depends on the rate of fall of cardiac output. 2. Cardiac output in six conscious rabbits was reduced by 4, 8 and 12% of baseline levels per min by gradual inflation of a cuff around the thoracic inferior vena cava. These rates correspond approximately to blood loss at rates of 3, 6 and 9% of blood volume/min. 3. The haemodynamic responses were biphasic. In Phase I (compensatory) there was progressive systemic vasoconstriction and tachycardia, and only a small fall in blood pressure. In Phase II (decompensatory), systemic vasoconstriction failed abruptly, arterial pressure plummeted and heart rate declined. 4. We could detect no effect of rate of fall of cardiac output on the pattern of the haemodynamic responses in either Phase I or Phase II. 5. We conclude that the rate of blood loss in different studies of haemorrhage in conscious rabbits, within the range 3 to 9 per cent of blood volume per minute, need not be regarded as a confounding factor when it comes to interpreting the results. It is likely that this conclusion can be generalized to studies of haemorrhage in other mammalian species.

Tissue Plasminogen Activator Treatment of Experimental Subretinal Hemorrhage

To determine if tissue plasminogen activator, a clot-specific fibrinolytic agent, could eventually be used to assist in the clearance or removal of subretinal hemorrhage, we studied the effect of subretinal injections of tissue plasminogen activator, autologous blood, balanced salt solution, and the combination of either tissue plasminogen activator or balanced salt solution after subretinal injection of autologous blood on retinal morphologic characteristics and clearance of subretinal hemorrhage in the albino rabbit. No morphologic evidence of tissue plasminogen activator toxicity was found in the rabbit retina at a dose of 25 to 50 μg/0.1 ml. Subretinal hemorrhage cleared faster after subretinal injection of tissue plasminogen activator when compared to balanced salt solution (P = .0005) but did not completely prevent overlying retinal degeneration. Both tissue plasminogen activator and balanced salt solution were found to decrease the toxic effects of subretinal blood on the morphologic characteristics of the rabbit retina, and this effect can be explained at least partly by dilution of the subretinal blood.

Does vasospasm occur in small pial arteries and arterioles of rabbits?

Vasospasm is a serious complication associated with subarachnoid hemorrhage. Successful management of vasospasm will ultimately depend on a clear understanding of the scope of this phenomenon, including whether arterial elements of different calibers are equally affected. We therefore examined the responses to subarachnoid hemorrhage in rabbit basilar arteries, small pial arteries, and arterioles. We compared the brain stem pial arteries of 10 perfusion-fixed male New Zealand White rabbits after experimental subarachnoid hemorrhage to those of five control rabbits using morphological analysis of cross-sections of plastic-embedded vessels. After subarachnoid hemorrhage, the internal elastic lamina was highly corrugated in all basilar arteries (mean diameter 319 +/- 51 microns). These arteries were severely constricted in comparison with the control group, in which the mean diameter was 691 +/- 17 microns, and corrugation of the internal elastic lamina was not present. In contrast, small pial arteries and arterioles very rarely demonstrated a vasoconstrictive configuration after subarachnoid hemorrhage. The contractility of the smaller vessels was confirmed by injecting 2 mg/kg BaCl2 intracisternally. Following BaCl2 injection, corrugation of the internal elastic lamina was detected in the small arteries and arterioles as well as the basilar arteries. We conclude that experimental chronic vasospasm after subarachnoid hemorrhage in rabbits tends to occur in large conducting arteries rather than in smaller pial arteries and arterioles.

Neurohumoral mechanisms and the role of arterial baroreceptors in the reno-vascular response to haemorrhage in rabbits.

1. Conscious rabbits, with implanted renal artery Doppler flow probes were bled at a constant rate (4 ml min-1). We assessed the contribution of autonomic, hormonal and local factors to the renal vasoconstrictor response to 20% loss of blood volume (BV) and the role of the sinoaortic baroreceptors in the neurohumoral response. 2. With intact autonomic effectors, 20% BV loss was associated with a small fall in vascular conductance, which was completely unaffected by inhibition or blockade of the combined effects of the two major pressor hormones angiotensin II (AII) and arginine vasopressin (AVP). Combined blockade of the autonomic effects plus those of the two pressor hormones resulted in marked elevation of vascular conductance, considered to be due to the local effects of haemorrhage. This response provided the baseline for assessing the constrictor response in the intact animal which, during 20% BV loss, was entirely due to reflex activity through the sympatho-adrenal system. 3. In contrast to the early phase of haemorrhage (less than 20% BV removal) both hormones played a role in the maintenance of mean arterial pressure immediately after haemorrhage and in the maintenance of renal vascular tone. This suggested that the contribution by hormones occurs only after more pronounced blood loss and hypotension. 4. In the presence of autonomic blockade with mecamylamine plus methscopolamine (plus a constant infusion of noradrenaline to maintain resting blood pressure) the renal vasoconstrictor response was similar to that of the intact animal. We have previously found that this regime is associated with greatly enhanced release of AVP and plasma renin activity. Sinoaortic denervation had no effect on this hormonally mediated vasoconstriction. 5. When the autonomic nervous system was intact but the effects of AII and AVP were blocked to prevent the accentuated hormonally mediated vasoconstriction, sinoaortic denervation completely abolished the normal autonomic renal constrictor response, which is thus largely under control of the arterial baroreceptors.

Immunomodulators interfere with angiopathy but not vasospasm after subarachnoid haemorrhage in rabbits

The present study deals with the effects of immunomodulators on the morphology of intracerebral arterial walls in rabbits with experimental subarachnoid haemorrhage (SAH). Immunostimulators:thymostimuline and inosine dimethylamino-isopropanol-p-acetamido-benzoate were found to aggravate the angiopathic changes, whereas immunosuppressive drugs-cyclosporine A and azathioprine appeared to prevent the damage. The authors consider the possibility of using immunosuppressive drugs in patients with ruptured intracranial aneurysms.

Interaction of vasopressin and opioids during rapid hemorrhage in conscious rabbits

We investigated possible interactions between arginine vasopressin (AVP) and endogenous opioid peptides during rapid hypotensive hemorrhage and subsequent opioid receptor blockade in conscious rabbits. Plasma AVP concentration did not change after normotensive hemorrhage but increased after hypotensive hemorrhage. Blockade of V1-AVP receptors (AVPX) did not affect prehemorrhage arterial pressure, heart rate, or hindquarter blood flow and vascular resistance. AVPX did not alter the hemodynamic response to hemorrhage or the blood loss required to reduce mean arterial pressure to less than 40 mmHg. However, hindquarter blood flow was higher and mean arterial pressure and hindquarter resistance lower after hypotensive hemorrhage in AVPX-treated animals. These differences were maintained after naloxone or saline injection. Naloxone increased mean arterial pressure and hindquarter resistance and decreased heart rate with or without AVPX. At 2 min postinjection, plasma AVP values were greater after saline than after naloxone. When naloxone's pressor response was reduced by alpha-adrenergic blockade, plasma AVP values were higher after naloxone than after saline. Thus AVP was not vital to maintenance of blood pressure during rapid normotensive hemorrhage or to the abrupt decrease in arterial blood pressure and resistance after rapid hypotensive hemorrhage. AVP release was important to spontaneous recovery from acute hypotensive hemorrhage but only of minor importance to naloxone's pressor response. Finally, AVP release appeared to be inhibited by endogenous opioids during acute hemorrhagic hypotension.

Arterial disease of the iris predicts visceral arterial necrosis in rabbits with acute one-kidney, one clip hypertension: comparisons with acute one-kidney, one wrapped hypertension

Tortuosity, dilations, aneurysms and satellite hemorrhages consistently develop in the circular arteries of the iris of rabbits with acute one-kidney, one clip (1K1C) malignant hypertension. These lesions are easily visualized, quantitated and monitored during life. Graded iridoarteriopathy correlates directly with visceral arterial necrosis (r15 = 0.843; P less than 0.001), final indirect blood pressure (r15 = 0.591; P less than 0.02) and cardiac hypertrophy (r15 = 0.565; P less than 0.02). Arterial necrosis in the irises and other viscera in acute 1K1C hypertension in rabbits is qualitatively similar to that occurring in rabbits with acute one-kidney, one wrapped (1K1W) hypertension with regard to both the histological features and organ distribution. However, over three times as many arterial lesions occur in rabbits with acute 1K1W hypertension as occur in those with acute 1K1C hypertension (P less than 0.001). Since blood pressure elevation does not correlate with graded iridoarteriopathy or with visceral arterial necrosis in 1K1W rabbits, factors other than blood pressure elevation appear to be especially important in the pathogenesis of arterial necrosis in the 1K1W model. On the other hand, the present study indicates that high grade iridoarteriopathy appears to be indicative of a high risk of cerebral hemorrhage in 1K1C rabbits, as earlier studies have shown for 1K1W rabbits.

Blood-brain barrier disturbance following subarachnoid hemorrhage in rabbits.

We studied disruption of the blood-brain barrier after experimental subarachnoid hemorrhage induced by an injection of 4 ml autologous arterial blood into the cisterna magna of rabbits. The animals were killed at 40 minutes, 6 hours, 1 day, 2 days, 4 days, or 6 days after subarachnoid hemorrhage. We assessed the integrity of the barrier function of intraparenchymal vessels in the ventral brain stem and cerebral hemispheres morphologically with transmission electron microscopy, using horseradish peroxidase as a tracer. In the ventral brain stem, which was in direct contact with the cisternal clots, markedly increased peroxidase staining toward the core of the brain stem was observed on the first day after subarachnoid hemorrhage. In an area of the cerebral hemispheres distant from the clots, barrier disturbance was prominent in the 6-hour specimens, and permeation of the tracer was spotty. From the distribution and morphological findings of these lesions, permeability changes in the ventral brain stem may have been caused by a direct effect of the cisternal clots; in the cerebral hemispheres, hemodynamic factors and changes in intracranial pressure associated with the elderly stages of subarachnoid hemorrhage seemed to be responsible. These results suggest that barrier disturbance associated with subarachnoid hemorrhage may be multifactorial in time course and location.

Renin-angiotensin system and opioids during acute hemorrhage in conscious rabbits

We measured changes in plasma renin activity (PRA) and used angiotensin-converting enzyme blockade with captopril to evaluate the role of the renin-angiotensin system during hemorrhage and after opioid receptor blockade in conscious rabbits. The increase in PRA after nonhypotensive hemorrhage was not statistically significant. PRA increased sixfold after a hypotensive hemorrhage to a mean arterial pressure less than 40 mmHg. This increase was statistically significant. Captopril altered the hemodynamic response to hemorrhage. The normal increase in vascular resistance early in hemorrhage was reduced by captopril pretreatment. After a critical blood loss, arterial pressure and heart rate decreased in both groups. The blood loss required to decrease mean arterial pressure to less than 40 mmHg was approximately 25% less in animals pretreated with captopril. The characteristic decrease in vascular resistance coincident with the onset of hypotension was still present after captopril pretreatment. Injection of naloxone or saline during acute hemorrhagic hypotension did not affect PRA. However, recovery of blood pressure after naloxone or saline was attenuated by converting-enzyme blockade. This attenuation was due primarily to a reduction in spontaneous recovery (i.e., recovery after the control saline injection) and not to a reduction in the response to naloxone. We tested whether this effect of captopril might be due to an interaction of ANG II and catecholamines. The plasma norepinephrine (NE) response to naloxone was statistically similar with and without captopril. In contrast, the response to exogenous NE after hypotensive hemorrhage was significantly reduced by captopril pretreatment. Captopril apparently did not alter baroreflex sensitivity but did reset the baroreflex to lower pressure levels during naloxone's pressor response.(ABSTRACT TRUNCATED AT 250 WORDS)

Preparation of antibodies to king cobra (Ophiophagus hannah) venom hemorrhagin and investigation of their cross-reactivity

N.-H. Tan, M. N. Saifuddin and M. I. Nik Jaafar. Preparation of antibodies to king cobra ( Ophiophagus hannah) venom hemorrhagin and investigation of their cross-reactivity. Toxicon 28, 1355–1359, 1990.—Hannahtoxin, the major hemorrhagin purified from king cobra ( Ophiophagus hannah) venom, elicits hemorrhages in rabbits but not in mice. Two antisera against hannahtoxin were prepared: one raised against purified hannahtoxin, while the other was raised against glutaraldehyde cross-linked and detoxified hannahtoxin. The antisera were refined by pepsin digestion and ammonium sulfate precipitation. They are of approximately equal potency in their ability to neutralize the hemorrhagic activity of king cobra venom in rabbits. The antisera did not form a precipitin line with venom of snakes of the Viperidae family nor neutralize hemorrhages elicited in mice by any of these venoms. However, when the hemorrhagic activity was assayed in rabbits, both antisera were able to abolish the hemorrhages elicited by all of the venoms tested. These results suggest that hannahtoxin displays few epitopes in common with hemorrhagins of viperid venoms, except those involved in the neutralization of hemorrhagic activity in rabbits. The epitopes of viperid venom hemorrhagins involved in the neutralization reaction in rabbits are different from those in mice.

Cerebral vasospasm in a double-injection model in rabbit

The present study was designed to assess the occurrence of cerebral vasospasm following an experimental subarachnoid hemorrhage model in rabbits. Sixty-nine New Zealand albino rabbits were used in this study. One milliliter of fresh arterial blood was injected through the surgically exposed atlanto-occipital membrane over a period of 20 seconds. The procedure was then repeated 24 hours later. Fifty animals underwent digital subtraction angiography at one of the following prefixed intervals: 1, 3, or 8 days after the second injection hemorrhage. Nineteen animals underwent one angiographic examination prior to the instillation of the intracisternal blood. This procedure was followed by a repeated angiography 3 days after the second experimental subarachnoid hemorrhage. For the purpose of evaluation, the films were magnified 10-fold and the diameter of the basilar artery as well as that of the extracranial vertebral artery at three different levels were measured. We assessed the diameter of the basilar artery as well as the mean ratio extracranial vertebral artery/basilar artery diameters. This ratio was considered to minimize anatomical and technical variabilities. The results in the first 50 animals showed a trend suggesting that spasmogenic activity reaches a peak at about the third day after subarachnoid hemorrhage. These results were confirmed in the latter 19 animals. However, mortality in this group was high: 50%. This double-injection model of subarachnoid hemorrhage in rabbits consistently reproduced cerebral vascular spasm 3 days after repeated subarachnoid hemorrhage. However, its usefulness as an experimental model for subarachnoid hemorrhage is limited practically by the high animal mortality in the protocols where repeated angiographic studies are necessary.

Expulsive choroidal hemorrhage in rabbits. A histopathologic study.

Expulsive hemorrhage is a catastrophic complication of intraocular surgery that can result in total loss of vision. Suprachoroidal effusion and hemorrhage may precede the development of expulsive hemorrhage; however, the relationship remains unclear. After sedation with intravenous pentobarbital sodium, 11 rabbits were given lactated Ringer's solution and heparin sodium intravenously. The right eyes were proptosed, and the central cornea, lens, and anterior vitreous were removed. After surgery, all 11 eyes (100%) developed choroidal effusion, choroidal hemorrhage, or expulsive hemorrhage. The rabbits were killed at various intervals after surgery so that the eyes could be enucleated and processed for light microscopy. Histologic examination revealed four sequential stages of expulsive hemorrhage as follows: (1) There was engorgement of the choriocapillaris. (2) Suprachoroidal effusion occurred mainly near the posterior pole. (3) As the effusion enlarged, stretching and tearing of choroidal vessels as well as tearing of the vessels and attachments at the base of the ciliary body occurred. (4) Massive extravasation of blood, primarily from the torn vessels at the ciliary body base, resulted in suprachoroidal hemorrhage and expulsion of blood through the surgical wound. This experimental model may provide new information relating to the cause and prevention of expulsive choroidal hemorrhage.