Search

Little is known about what influences cerebrospinal fluid pressure (CSFP) during anesthesia prior to aortic cross-clamping (AXC). Therefore, this study measured the effect of anesthetic induction, of various drugs administered during the course of surgery prior to AXC, and of hemodynamic changes on CSFP, and calculated spinal cord perfusion pressure (SCPP = mean arterial pressure [MAP] - CSFP) in 11 patients undergoing surgery on the descending thoracic aorta. A lumbar drainage catheter was placed to facilitate drainage of CSF and to measure CSFP. Anesthesia was induced with fentanyl, 50 micrograms/kg, and midazolam, 1 mg, using a pancuronium-metocurine mixture for neuromuscular blockade. Data were collected prior to and after (1) anesthetic induction, (2) mannitol to augment diuresis, (3) sequential use of sodium nitroprusside (SNP) and isoflurane (ISO) to lower MAP by 20%, (4) drainage of spinal fluid, (5) intrathecal injection of papaverine (IP), and (6) AXC. Statistical comparisons of recorded data were made using the least squares mean method and Friedman test. Linear regression was used to test for correlation between CSFP and hemodynamics. Anesthetic induction affected neither hemodynamics nor CSFP. Mannitol significantly increased heart rate, central venous pressure (CVP), pulmonary capillary wedge pressure (PCWP), cardiac output (CO), and CSFP (P less than 0.05). SNP or ISO altered neither CVP, PCWP, CO, nor CSFP, which remained elevated at the postmannitol infusion level. ISO, unlike SNP, caused a significant decrease in SCPP (P less than 0.005). Subsequent drainage of 20 mL of CSF improved SCPP (P less than 0.05). IP did not have any effect on hemodynamics or CSFP. CSFP showed a strong correlation with CVP (r = 0.86).(ABSTRACT TRUNCATED AT 250 WORDS)

Permanent ligation of arteries supplying blood to the spinal cord in operations for aortic aneurysm can lead to spinal cord ischemia, which can result in either paraparesis or paraplegia. This report describes a rapid method of intraoperative identification of those arteries that supply the spinal cord by use of an intrathecal platinum electrode to detect hydrogen in solution that has been injected into the aortic ostia. Preservation or perfusion of those identified arteries supplying the spinal cord may decrease the rate of postoperative neurologic complications. Of 28 porcine experiments with postoperative observation for 24 hours, there were 3 initial pilot experiments in which saline saturated with hydrogen was injected into the temporarily cross-clamped aorta. Twenty animals were then randomized to (1) preservation of only the vessels sequentially identified to supply blood to the spinal cord from T-13 to L-5 (n = 10); (2) division of the vessels supplying the spinal cord (n = 10). A further five animals underwent perfusion experiments wherein the identified cord arteries were perfused by a shunt, the other nonsupply arteries were divided, and the aorta was kept clamped for 45 minutes. Spinal motor evoked potentials were elicited with an intrathecal electrode and were highly sensitive for paralysis. Paralysis occurred in 0/3 pilot (p less than 0.013 vs division); 8/10 division; 1/10 preservation (p less than 0.0017 vs division); and perfusion 1/5 (p less than 0.025 vs division). Results of a pilot study in eight humans shows that the technique can be used to rapidly identify segmental arteries supplying the spinal cord, to determine if distal perfusion is supplying the spinal cord with blood flow, and if reattached segmental arteries are patent.

This article is concerned with the study of the effect of several variables, principally that of cerebrospinal fluid drainage, on the incidence of neurologic deficit in a prospective randomized series of patients with extensive aneurysms of the descending thoracic and abdominal aorta (thoracoabdominal type I and II). Forty-six patients had cerebrospinal fluid drainage, and 52 were controls, with a total of 98 available for study. Cerebrospinal fluid pressure was continuously monitored in the former group and pressure maintained less than or equal to 10 mm Hg in 20, less than or equal to 15 mm Hg in 20, and greater than 15 mm Hg in 6 patients during period of aortic clamping. The method of treatment including reattachment of intercostal and lumbar arteries (p = 0.2), temporary atriofemoral bypass during aortic occlusion (p = 0.3), and spinal fluid drainage (p = 0.8) were not statistically significant in reducing the incidence of neurologic deficits. Thus cerebrospinal fluid drainage as we used it, was not beneficial in preventing paraplegia. On appropriate statistical analysis we found that the only significant predictor of delayed deficits was postoperative hypotension (p = 0.006).

We report the improving surgical results in a consecutive series of 690 patients referred to one of us (E.S.C.) for aortic dissection between December 1956 and September 1989, a substantial portion of whom had dissection as a complication of either previous aortic (n = 113, 16) or previous cardiac (n = 54, 8%) operation. Our initial operation of choice in patients requiring multiple operations in this group of 690 patients was based on the most life-threatening or symptomatic aortic segment involved, which was ascending aorta and/or aortic arch (Asc/Arch) in 301 (44%) patients, descending thoracic aorta (Desc) in 195 (28%) patients, and thoracoabdominal aorta (TaA) in 194 (28%) patients. As detailed below, considerable improvement occurred in the 30-day survival rates over time, particularly for acute dissection: [table; see text] The independent determinants of both early and long-term mortality were identified. Independent determinants of late fatal rupture, reoperation, and neuromuscular dysfunction for distal dissectors were also identified. In our experience, continued aggressive surgical intervention for aortic dissection with modern operative techniques has resulted in markedly improved 30-day operative survival (approaching 95% including those patients with acute dissection) and significant improvement in late results.

The life expectancy of patients with aortic aneurysm is significantly prolonged by graft replacement therapy. Regardless, a significant predictor of late death is complications of either residual aortic aneurysmal disease or the development of additional aortic aneurysm. This paper reviews a personal experience in the treatment of 4170 patients with aneurysmal disease of either dissection or medial degenerative origin, indicating that multiple segment involvement was or became present in 1262 (30%) patients, 463 (67%) of 694 patients with dissection, and 799 (23%) of 3476 patients without dissection. Regardless of etiology, multiple involvement varied with the location of the presenting involved segment, i.e., ascending aorta (38%), ascending and arch (70%), descending thoracic aorta (73%), and abdominal aorta (26%). This study was limited in detail to 811 patients who had ascending and ascending and aortic arch replacement for aneurysm. These patients were divided into 3 subgroups: (1) 524 patients with no distal disease; (2) 135 patients with distal disease treated by subtotal replacement in 82 and total replacement in 53; and (3) 152 patients with distal disease not treated. The 5-year survival rate from the time of first operation, including early death from operation was 75% in group 1, 65% in group 2, and 39% in group 3. The causes of death in group 3 patients were aneurysmal rupture and/or associated disease. It is concluded that initial total aortic study and regular postoperative monitoring with computed tomographic scanning is indicated to detect extensive disease or recurrence of disease and that aggressive replacement is indicated except in patients with associated disease that does not permit operation.

One source of paraplegia after aortic operations is the failure to reattach the spinal cord blood supply, the origins of which are not evident at operation. This report is concerned with a rapid new method of identifying these vessels intraoperatively. In 9 pigs, a specially designed catheter with platinum and stainless steel electrodes was inserted intrathecally. Saline solution saturated with hydrogen was injected sequentially into arterial ostia at T-15 to L-4 inclusive, and the generated current impulses from the conditioned platinum electrode were recorded. Of 90 potential segmental arteries supplying the spinal cord, 28 gave rise to spinal radicular arteries. Hydrogen-induced current impulses correctly located 25 of the radicular arteries and all those larger than 180 microns in diameter. When injected with indigo carmine, the vessels localized by the hydrogen-induced current impulses filled the entire anterior spinal artery from the low thoracic to the sacral region, whereas injection of the other vessels did not show filling. After refinement and testing for safety, this method has been employed clinically to rapidly localize and reattach routes of critical cord circulation.

We have previously described a technique for intrathecal administration of papaverine and cerebrospinal fluid drainage to prevent paraplegia after aortic surgery. Herein we report the cerebrospinal fluid and hemodynamic alterations that occurred in 11 patients who had 30 mg of a specially prepared papaverine hydrochloride 10% dextrose solution injected before aortic cross-clamping and also had cerebrospinal fluid drainage. A mean of 26.6 ml (SD +/- 7.1 ml) was drained before and 34.6 ml (SD +/- 24.1 ml) was drained during aortic cross-clamping. The cerebrospinal fluid pressure increased significantly with anesthetic induction (p less than 0.03), during the period between anesthetic induction and cerebrospinal fluid drainage (p less than 0.005), and with aortic cross-clamping (p less than 0.05). These cerebrospinal fluid pressure alterations were similar to central venous pressure increases with a significant linear correlation between cerebral spinal fluid pressure and central venous pressure before anesthetic induction (r2 = 0.81, p less than 0.005), and both before (r2 = 0.94, p less than 0.005) and after (r2 = 0.74, p less than 0.005) aortic cross-clamping. As expected, cerebrospinal fluid pressure was significantly reduced by cerebrospinal fluid drainage before aortic cross-clamping (p less than 0.001). The administration of intrathecal papaverine had no significant effect on mean arterial pressure, systemic vascular resistance, cerebrospinal fluid pressure, nor the pH of cerebrospinal fluid. Neither were there any complications noted related to the technique. All the patients survived, and no new immediate postoperative paraparesis or paraplegia occurred.(ABSTRACT TRUNCATED AT 250 WORDS)

There is a high incidence of paraplegia associated with thoracic aortic cross-clamping, even when cardiopulmonary bypass or shunts are used. In 56 adult baboons, spinal cord blood flow (SCBF), vascular anatomy, and paraplegia rates were evaluated. Tissue blood flow was measured by radioactive microspheres. Various procedures were used to increase SCBF and to prevent ischemia-reperfusion injury. It was found that the rate of paraplegia was inversely correlated with neural tissue ischemia (SCBF) and directly correlated with reperfusion hyperemia. Two methods completely prevented paraplegia. These two methods were a thoracic shunt with occlusion of the infrarenal aorta or cerebrospinal fluid drainage plus intrathecal papaverine injection, both of which were associated with an increased SCBF. Furthermore, papaverine dilated the anterior spinal artery (ASA) (p = 0.007) and increased the blood flow through the lower ASA. Whereas procedures utilizing a calcium channel blocker (flunarizine), allopurinol, superoxide dismutase (SOD), laminectomy alone, and a thoracoabdominal shunt not perfusing the arteria radicularis magna (ARM) all failed to prevent paraplegia, allopurinol (p = 0.026) and SOD (p = 0.004) did prevent gastric stress lesions, indicating that their failure to prevent paraplegia was not due to a lack of activity. Of great clinical interest is that, if a shunt is used and the ARM is perfused, infrarenal aortic cross-clamping increases SCBF, thus preventing paraplegia. Intrathecal application of papaverine proved to be even more effective in increasing SCBF and also completely prevented paraplegia. As this is a safer procedure than the insertion of shunts, this is the method of choice for the prevention of paraplegia associated with thoracic aortic cross-clamping. The preliminary trial using intrathecal papaverine in human beings has thus far shown no adverse side effects from the drug, and no paraplegia has occurred.

Eleven patients undergoing operation on the descending or thoracoabdominal aorta were administered papaverine intrathecally in an attempt to protect the spinal cord from ischemic damage. Concurrently, 19 patients, also undergoing operation on the thoracic or thoracoabdominal aorta, were operated on with a variety of conventional techniques, including distal aortic perfusion, but were not given intrathecal papaverine. No signs of early neurologic injury developed in any of the patients in the intrathecal papaverine group, although delayed paraparesis developed in one of the patients (9%; 70% confidence limits = 1% to 28%). On the other hand, eight of 19 patients undergoing operation with conventional techniques had either lower extremity paraparesis or paraplegia postoperatively (42%; 70% confidence limits = 29% to 57%; p = 0.058). Intrathecal papaverine appeared to provide spinal cord protection during thoracic aortic operations, particularly during prolonged periods of aortic cross-clamping. Papaverine was not associated with increased risk and may be superior to other conventionally used modalities. We conclude that continued evaluation of this technique is justified.

A case of malignant insulinoma in a Black patient is presented. An investigational protocol is suggested with a discussion of the newer techniques of pancreatic adenoma localization and the therapeutic measures available for the management of malignant islet cell tumours.

This report is concerned with the surgical treatment of cardiovascular manifestations of Marfan syndrome in 151 patients by 280 operations, 70 before referral, 203 by the authors during a 16-year period, and seven by others in the follow-up interval. This therapy resulted in mitral valve replacement in 13, aortic valve replacement in 135, aortic segment replacement in 151 (including total aortic replacement in 10), nearly total aortic replacement in 26, entire thoracic aortic replacement in seven, and replacement of the entire descending and abdominal aorta in 20 patients. The early (30-day) survival was 94%, and the 5- and 10-year Kaplan-Meier survival was 75% and 56%, respectively. Fusiform aneurysms were present in the ascending aorta, aortic arch, or both in 135, the descending aorta in two, the thoracoabdominal aorta in four, and the abdominal aorta in seven. Aortic dissection occurred in 102 (67%) patients, and 34 (22%) patients with ascending aortic operations required reoperation. This study suggests that the results of current techniques warrant a more aggressive approach to these problems to permit better long-term results.

Aortocaval fistulisation is a rare condition which results in a high mortality rate if misdiagnosed before surgery. The presentation can be either acute or insidious. The paucity of definitive clinical symptoms, the haemodynamic consequences of a large arteriovenous fistula and the technical difficulties of surgery all contribute to this high mortality rate. We describe 6 patients whose case histories illustrate these points and discuss the management of patients with aortocaval fistulas.

Four patients are presented who sustained blunt trauma and later developed necrotizing fasciitis in the areas of contusion, associated with the distant locus of sepsis. Three of the cases developed acute renal failure and later died. Early recognition of this condition allows prompt debridement beyond the margins of involved necrotic fascia, followed by wound inspection 8-hourly, and treatment with antibiotics, most recently third-generation cephalosporins, piperacillin, amikacin, and vancomycin.

Stress ulceration is frequently encountered after cardiovascular surgery. In this study of 32 male baboons, severe gastric ischaemia was used to produce gastric stress lesions. The occurrence of these lesions was reduced by allopurinol (P = 0.03) and completely prevented by the combination of allopurinol with superoxide dismutase (P = 0.004). A shorter ischaemic period also reduced the number of lesions (P = 0.02). Concurrent with the stress lesion formation, there was a fall in mucosal glutathione and oxidized glutathione levels (P less than 0.05).

In acute experiments carried out in 27 baboons under general anesthesia, the regional gastric mucosal and muscle layer blood flow and gastric acid secretion were measured during 4 hr. Baboons were allocated to each of the following six groups: control, gastric acid stimulation with histamine 40 micrograms/kg/hr intravenous, inhibition of basal or stimulated acid secretion with either ranitidine 50 mg intravenous or omeprazole 1 microgram/kg/hr. There were no significant cardiovascular alterations during the experiments. Histamine stimulation produced a concomitant rise in acid secretion and increase in blood flow only to the mucosal layer of the parietal-cell-bearing area of the stomach. Neither the underlying muscle layer nor the non-parietal-cell-bearing fundic or antral mucosa took part in this response, suggesting that a mechanism controlling blood flow is present in close proximity to the parietal cells. It was also established that the increase in blood flow occurs in response to parietal cell activity and not as a result of the action of histamine on the vascular system. Suppression of both basal and stimulated acid secretion did not cause a fall of mucosal blood flow below basal levels in any part of the stomach, indicating that drugs that inhibit parietal cell activity can be used in conditions where gastric mucosal ischemia should be avoided.

Ascending aorta and/or aortic arch reconstruction by composite valve graft (281, 39%), separate valve graft (117, 16%), graft only (256, 36%), and other procedures (63, 9%) was used for aneurysm or dissection caused by trauma (6), infection (20), aortitis (46), dissection (261: acute 72, chronic 189), and medial degeneration (384) in 717 patients during the 9-year period between Jan. 11, 1980, and Jan. 16, 1989. Of these, 150 had 173 previous heart or aortic operations and needed reoperation for progression or recurrence of aneurysm, rupture, valvular insufficiency, aortocutaneous or aorta-heart chamber fistulas, great vein or airway obstruction, and infection. Concurrent distal aneurysmal disease was present or developed in 267 (37%) patients, being most prevalent in patients with arch involvement (211/395, 53%). These patients were treated either simultaneously or later. The ages ranged from 10 to 88 years, median 61. Aneurysm symptoms were mild or absent in 593 (83%) and severe in 124 (17%). The 30-day survival rate was 91%. The independent determinants predictive of 30-day death were increasing age, severe aneurysm symptoms, diabetes, previous proximal aortic operation, need for cardiac support, postoperative tracheostomy, postoperative heart dysfunction, and stroke. Of the 319 patients who had none of the four preoperative factors, 308 (97%) survived. Survival decreased to 74% in those with two or more factors. After a total of 1193 operations, the entire aorta was replaced in 53, near total in 35, total thoracic replacement in 78, and total aorta except arch in 27. Late survival rates (Kaplan-Meier) were 66% and 57% at 5 and 7 years. Independent predictors of death were severe aneurysm symptoms, preoperative angina, extent of proximal replacement, associated residual distal aneurysm, balloon pump, renal dysfunction, cardiac dysfunction, and stroke. Five-year survival rates varied with the incidence of the four preoperative variables and age in a single patient: 78% in 413 patients with up to one variables, 57% in 193 patients with two or three, and 39% in 111 patients with three or four (p less than 0.0001).

We evaluated 1525 consecutive patients who had undergone thoracic or thoracoabdominal aortic surgery to ascertain the factors associated with the development of acute renal failure. Complete data were available in 1233 patients who were treated recently, and these were analyzed. Acute renal failure, severe enough to require dialysis, developed in 5.5% of this group (68/1233): 2.3% and 7%, respectively, for descending (9/391) and thoracoabdominal repairs (59/842). Of interest, on multivariate analysis, both renal artery endarterectomy for occlusive disease (p = 0.0006) and chronic dissection (p = 0.03) were associated with significantly less acute renal failure. On multivariate analysis, the significant independent predictors (p less than 0.05) of acute renal failure were preexistent renal dysfunction, evidence of diffuse atherosclerosis, the use of the pump bypass, and markers of hemodynamic instability. Contrary to earlier reports based on a smaller number of patients, we found that neither the use of pump bypass (7% acute renal failure), atriorenal bypass (8% acute renal failure), nor cold Ringer's lactate (3% acute renal failure) appeared to significantly avert the complication of acute renal failure. Indeed, pump bypass appeared to be deleterious (p = 0.0146) and perfusion with cold Ringer's lactate was not without risk. Furthermore, in a prospective evaluation of angiotensin converting enzyme blockers, we were unable to show that they afforded renal protection after transient renal ischemia. This study has clarified the clinical problems associated with acute renal failure and lays the foundation for future research.

No satisfactory explanation exists as to why paraplegia occurs despite distal aortic perfusion during thoracic aortic operations. We studied the hemodynamics, paraplegia rate, and spinal cord blood flow with radioactive microspheres in 17 male adult baboons, with particular reference to the arteria radicularis magna. The groups consisted of control animals, subjected to cross-clamping for 60 minutes, and animals with aorto-aortic shunts operational for 60 minutes. There were no significant left ventricular hemodynamic advantages with shunting. Shunting significantly increased lumbar spinal cord blood flow (p = 0.0009), which correlated with the distal aortic mean pressure (r = 0.59, p = 0.008). However, lower thoracic spinal cord blood flow did not increase during shunting (p = 0.2) and did not correlate with the distal aortic pressure (r = 0.11, p = 0.64). This is due to the vascular anatomy of the anterior spinal artery, which was, as in man, smaller above (0.278 mm) than below (0.744 mm) the entry of the arteria radicularis magna. Resistance to flow, as calculated by Poiseuille's equation, was 51.7 times greater up the anterior spinal artery as compared with down this artery. The vascular anatomy explains the absence of paraplegia in one baboon in the cross-clamp group and paraplegia in one baboon in the shunt group. Thus, distal aortic perfusion protects the spinal cord below the arteria radicularis magna but not above it.

Operation was employed in the treatment of 546 patients for complications of aortic dissection during the 32-year period of 1956-1988. Current concepts and operative techniques evolved during this period. Fortunately, about half the patients were treated during the latter 4 years, as modern therapy became standardized. The cumulative survival rate was 86% for all patients and 94% for those treated during recent years. Pathologic processes and requirements of operation became clearer by treating 174 patients who had had 198 previous operations by the time of referral. Reoperation was required for complications of operations now considered outdated, heart operations in patients with ascending aortic dilatation, and progressive dilatation of residual segments of the aorta. The 546 patients were followed, and a total of 838 operations were finally employed, resulting in total aortic replacement in 18, near total replacement in 41, entire thoracic aorta in 22, near total thoracic aorta in 33, and the entire thoracoabdominal aorta in 148 patients. Long-term survival in 439 patients after final operation was 66% and 44% at 5 and 10 years, respectively, despite the fact that the median age at first admission was 59. Operative treatment appears to be well-established for this disease.

The major controversy in the surgical repair of acute traumatic rupture of the aorta (TRA) is about the need for distal perfusion and whether or not the benefits of clamping justify the risks involved. To resolve this issue, our experience and 596 cases of acute TRA reported in the English-language literature were reviewed. The operative mortality and incidence of paraplegia among patients who underwent total or partial cardiopulmonary bypass, local temporary aorto-aortic shunting or simple cross-clamping were analysed statistically. The respective mortality rates after the above three procedures were 16,7%, 11,4% and 5,8%, and mortality was significantly increased when distal perfusion techniques were used (P less than 0,01). The incidences of paraplegia among patients who underwent partial left heart bypass, temporary shunting and simple aortic cross-clamping were 2,2%, 2,3% and 5,8% respectively (the differences did not reach statistical significance). Methods of management are discussed. Simple aortic cross-clamping remains the method of choice in uncomplicated cases of acute TRA.

In this study we sought to define the role of oxygen-derived free radicals during ischemia and reperfusion in the production of acute damage to the gastric mucosa of baboons. The protective effect of the xanthine oxidase inhibitor, allopurinol, the superoxide scavenger, superoxide dismutase (SOD), and a long-acting SOD-albumin was determined. Mucosal damage was evaluated using light and scanning electron microscopy. Evidence for oxidative insult to the gastric mucosa was sought by measuring tissue concentrations of reduced (GSH) and oxidized (GSSG) glutathione. Gastric mucosal blood flow was estimated using the microsphere technique. A similar pattern of tissue damage was found at the end of ischemia in all three groups. Thirty minutes after reperfusion, severe mucosal damage (grade 3) increased only in the untreated control. In the two treated groups, grade 3 damage remained unchanged during reperfusion and a decrease in the percentage of moderate damage (grade 2) was seen. Both GSH and GSSG tissue concentrations were lower in the untreated controls as compared to the scavenger-treated groups, making it questionable whether GSH/GSSG tissue levels adequately reflect oxidative stress. We conclude that in our ischemia-reperfusion model the generation of oxygen-derived free radicals produces mucosal damage and prevents the restitution of moderate mucosal damage during reperfusion. In ischemia, factors other than free radicals seem to be responsible for mucosal damage. The protective effect of allopurinol and SOD was not mediated by changes in gastric mucosal blood flow.

General surgical complications after cardiopulmonary bypass (CPB) are infrequent but serious. No prospective studies have evaluated their incidence. We analyzed in such a study 135 patients who were to undergo CPB. Among these 135 patients, an abdominal complication developed in 6.6%; it contributed to 2 of the 5 deaths. Postoperative hyperamylasemia was found in 36% of patients, but only 2 had overt pancreatitis. The hyperamylasemia was not due to the salivary component, pulmonary complications, or prolonged CPB (r = 0.22). A gastrointestinal hemorrhage occurred in 7 patients. No patients had acute cholecystitis. We conclude that abdominal complications are more frequent than reported in retrospective studies, and result in significant morbidity and mortality. Postoperative hyperamylasemia is common but usually is not associated with untoward morbidity or mortality.

The cytologic diagnosis by intraoperative fine-needle aspiration of a case of partially cystic adenosquamous carcinoma of the pancreas is reported. Tissue for histologic confirmation was not obtained. Confirmatory ultrastructural and immunochemical findings are thus presented.

The purpose of this study was to investigate whether the anatomic origin of micro-emboli influences their intracranial distribution. In twenty-two baboons, we examined the distribution of 99-Technetium labelled albumin aggregates (5 to 40 microns in size) after injection into the circulation at the left atrium (LA), carotid trifurcation (CA), and anterior and posterior common carotid artery (CCI). In a further subgroup, the emboli were introduced at the carotid trifurcation with the contralateral carotid artery ligated (CA + L). The results of this study demonstrated that aggregates introduced at the carotid artery lodged preferentially in the ophthalmic (p = 0.032) and middle cerebral artery territories (p = 0.016). If the contralateral common carotid artery was ligated, however, more aggregates were found in the ipsi- and contralateral anterior cerebral artery territories (p = 0.01, p = 0.003). Aggregates introduced into the cardiac circulation were equally distributed throughout the brain. This experimental model determined patterns of flow that might be analogous to the human situation where unilateral or bilateral carotid stenosis or stenosis with contralateral occlusion has occurred or embolus from cardiac source has occurred. The results do not imply that the 40 micron microaggregates do cause TIA. These experimental findings support clinical observations that cardiac lesions may cause transient ischemic attacks (TIA) anywhere in the brain. In contrast, those of carotid artery origin cause predominantly middle cerebral or ophthalmic artery territory TIAs unless the contralateral carotid artery is severely stenosed or occluded.