Laparoscopic management of celiac artery compression syndrome.

BACKGROUND: Celiac artery compression syndrome (CACS) remains a controversial diagnosis, despite several reported series documenting therapeutic efficacy of CA decompression. Traditional therapy consists of open surgical decompression, but since 2000, five isolated case reports have been published in which CACS has been successfully treated with laparoscopic techniques. This approach was adopted as the sole initial therapy for CACS at the Johns Hopkins Hospital in 2002. This article reports the results of a unique surgical series that triples the reported worldwide experience with this therapy. METHODS: Fifteen patients (median age, 40.6 years) diagnosed with CACS underwent laparoscopic decompression by a single vascular surgeon. CACS was diagnosed by digital subtraction angiography in 14 patients and computed tomography (CT) angiography in one patient, with images acquired in both expiratory and inspiratory phases of respiration. CA decompression was offered after the results of a thorough workup for other pathology were negative, including upper and lower endoscopy, CT scanning, gastric and gallbladder emptying studies, upper gastrointestinal series, and small-bowel follow-through studies. Indications in all patients were abdominal pain and weight loss (average, 9 lbs). The procedure consisted of laparoscopic division of the median arcuate ligament and complete lysis of the CA from its origin on the aorta to its trifurcation. RESULTS: Between November 2002 and September 2007, 15 consecutive patients underwent laparoscopic CA decompression. Median length of follow-up was 44.2 months. There were no operative deaths. Four patients were converted intraoperatively to an open decompression, all for intraoperative bleeding; only one required a blood transfusion. Average operating time was 189 minutes, and the average length of stay was 3.5 days. CA intervention was required in six patients, including three intraoperative procedures (1 patch angioplasty, 1 celiac bypass, 1 percutaneous angioplasty) and six late procedures (2 percutaneous angioplasties, 3 percutaneous stents, 1 celiac bypass). One complication occurred, a severe case of pancreatitis that developed 1 week after discharge. On follow-up, 14 of 15 patients subjectively reported significant improvement, and one patient remains symptomatic with no diagnosis. CONCLUSION: Laparoscopic decompression of the CA may be a useful therapy for CACS, but there is potential for vascular injury, and adjunctive CA intervention is often required. Surgeons should consider laparoscopic CA decompression as a therapeutic alternative for CACS and should participate in the care of patients with this diagnosis.

Multiple treatment algorithms for successful outcomes in venous thoracic outlet syndrome.

BACKGROUND: We sought to determine the outcomes in patients presenting with venous thoracic outlet syndrome. METHODS: Prospectively collected data from 67 patients between October 2003 and December 2007. The average age was 31 years (range, 16-54); the 37 males and 30 females presented on average 9.2 months (range, 1 month-6 years) after acute thrombosis. Four treatment algorithms were utilized. RESULTS: In group 1, 3 patients presented with acute occlusion and received tissue plasminogen activator (tPA) and immediate first rib resection with scalenectomy (FRRS). One vein rethrombosed and was treated by intravenous tPA postoperatively. In group 2, 39 patients presented with stenotic subclavian veins an average of 22 weeks after their initial thrombosis, all of whom underwent FRRS followed by a venogram 2 weeks postoperatively: 25 had a tight stenosis and underwent venoplasty with anticoagulation; 13 had patent, nonstenotic subclavian veins, and 1 patient required tPA and venoplasty owing to rethrombosis. Two patients had their subclavian vein thrombose after venoplasty and were treated with anticoagulation, tPA, and venoplasty. In group 3, 11 patients presented with intermittent venous obstruction without thrombosis and underwent FRRS; 3 underwent venograms because of concerns of residual stenosis, 2 of whom required venoplasty postoperatively. Finally, in group 4, 14 patients presented with occluded subclavian veins and underwent FRRS with long-term anticoagulation. Eleven have recanalized at an average of 6 months (range, 2-12). CONCLUSION: Overall, 64 of 67 patients have patent subclavian veins after a median follow-up of 10 months, and all patients are asymptomatic for a success rate of 96%. Tailored treatment algorithms including FRRS, postoperative venograms with or without intervention, and the use of long-term anticoagulation seems to be required in this complicated group of young patients to achieve optimal results.

Thoracic aortic stent-grafts: utility of multidetector CT for pre- and postprocedure evaluation.

Indications for and experience with placement of endovascular stent-grafts in the thoracic aorta are still evolving. Common pathologic conditions of the thoracic aorta that are amenable to stent placement include penetrating ulcers, aortic dissection, aortic aneurysms, aortic rupture, and congenital abnormalities. Advances in multidetector computed tomography (CT) permit high-quality two-dimensional multiplanar reformation and three-dimensional rendering, which are essential for comprehensive assessment of the thoracic aorta. The ability of multidetector CT to allow detailed evaluation in any plane or perspective enables detection of thoracic aortic disease and assessment of its relationship to normal vessels. Potential complications of endovascular stent placement include endoleaks, stent migration, pseudoaneurysms, dissection, aortic perforation, kinking, thrombosis, and coverage of vital branch vessels. It is important for the radiologist to not only detect pathologic conditions of the thoracic aorta but also to provide the referring clinician with the necessary pre- and postprocedure information to determine appropriate clinical care.

Preoperative statin and diuretic use influence the presentation of patients undergoing carotid endarterectomy: results of a large single-institution case-control study.

OBJECTIVE: Patients who present for carotid endarterectomy (CEA) with symptoms of stroke or transient ischemic attack (TIA) have worse postoperative outcomes than patients with asymptomatic carotid disease. We undertook this study to see whether preoperative medication use or patient characteristics were associated with the presence of symptomatic cerebrovascular disease at the time of operation. METHODS: A retrospective case-control study was performed among patients presenting for elective CEA at a single academic institution between 1994 and 2004. A total of 660 (42%) symptomatic patients were identified from an institutional database and compared with 901 (58%) control patients who were asymptomatic at the time of CEA. The independent association of cerebrovascular symptoms with patient variables was assessed by using multivariate logistic regression analysis after propensity score adjustment. RESULTS: The mean age and sex distribution were similar between cases and controls, although symptomatic patients were more likely to have an ulcerative plaque (18% symptomatic vs 11% asymptomatic; P < .01). Compared with asymptomatic controls, patients presenting for CEA with symptoms of stroke or TIA were less likely to have hyperlipidemia (43% vs 55%; P < .01) or a history of coronary artery disease (43% vs 54%; P < .01) and were less likely to be receiving statins (35% vs 47%; P < .01), beta-blockers (34% vs 44%; P < .01), and diuretics (22% vs 31%; P < .01). After controlling for potential interaction and confounding by using propensity score adjustment and logistic regression analysis, preoperative use of statins (adjusted odds ratio, 0.72; 95% confidence interval, 0.56-0.92; P = .01) and diuretics (adjusted odds ratio, 0.74; 95% confidence interval, 0.58-0.95; P = .02) were independently associated with a lower likelihood of having cerebrovascular symptoms at the time of CEA. CONCLUSIONS: We observed that patients receiving statins or diuretics were less likely to present for CEA with symptoms of stroke or TIA. These data raise the question of whether the preoperative use of these medications protects patients with carotid stenosis from developing symptomatic disease and contributes to improved outcomes among patients undergoing CEA. Further research is needed to assess the optimal medical management of patients before vascular surgery.

Correction of symptomatic cerebral malperfusion due to acute type I aortic dissection by transcarotid stenting of the innominate and carotid arteries.

INTRODUCTION: Acute proximal aortic dissection may be complicated by stroke due to malperfusion of the arch vessels. We report a novel case of successful endovascular treatment of acute cerebral malperfusion due to a dissection involving the aortic arch. CASE REPORT: A 66 year old man was transferred from another hospital with an acute type I aortic dissection and underwent emergent repair of the aortic valve and ascending aorta with a composite graft. Left hemiplegia and altered cognitive function were noted on postoperative day 1. A carotid duplex scan showed partial thrombosis of the right carotid artery with very slow flow and reversal of flow in the right vertebral artery. A head CT was normal, while a head MRI and MR angiogram showed intraluminal defects in the inominate and right carotid arteries and perfusion abnormality of the entire right middle cerebral artery territory, but only small infarcts of watershed areas. The patient underwent stenting of the right carotid and inominate arteries through the right carotid artery with complete resolution of a large pressure gradient that was noted prior to stenting. The patient's left hemiplegia and cognitive impairment subsequently resolved during his inpatient hospitalization. On follow up five months later, he had a normal neurologic exam and MRI showed old watershed infarcts but no perfusion abnormality. On most recent follow-up 2.5 years after treatment, he remains well and a CT angiogram shows that his stented vessels remain patent. CONCLUSION: Endovascular techniques may be safely applied to correct cerebral malperfusion that results from type I aortic dissection.

Trends in the in-hospital stroke rate following carotid endarterectomy in California and Maryland.

OBJECTIVE: We examined the outcome of carotid endarterectomy (CEA) in the state of Maryland during the last decade to identify any trends in the incidence of in-hospital stroke and mortality and compared these results with the outcome of the operation throughout the state of California as a control population. METHOD: We performed a retrospective analysis of 10 years (1994 to 2003) of the Maryland and 5 years (1999 to 2003) of the California hospital discharge databases. The following patients were included in the analysis: (1) International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) procedure code 38.12 (endarterectomy of the vessels of the head and neck other than intracranial vessels) in the primary coding position but not in any secondary position, or (2) the diagnosis code 433.00 to 433.91 (occlusion/stenosis, precerebral artery), or (3) the diagnosis-related group (DRG) 5 (extracranial vascular procedure). Symptomatic patients were identified by history of previous stroke (ICD-9 codes 342 or 438), transient ischemic attack (435 or 781.4), or amaurosis fugax (362.34 or 368.12). In-hospital strokes were identified by ICD-9 codes 997.0, 997.00, 997.01, and 997.09. Low-, moderate-, and high-volume surgeons were defined as performing <15, 15 to 74 and >or=75 CEAs annually. Hospital volumes were similarly classified as low for those performing 100 annually. RESULTS: In the Maryland data, 23,237 CEA cases were identified with 169 in-hospital strokes over 10 years (0.73%), whereas the 51,331 California CEAs had 232 in-hospital strokes over 5 years (0.45%). The stroke rate in Maryland was 2.12% in 1994, 1.47% in 1995, and 0.29% to 0.65% from 1996 to 2003. The decrease in strokes was more pronounced among symptomatic patients, where the rate was 3.82% in 1994, 4.44% in 1995, and 0.90% to 2.29% from 1996 to 2003. A similar decrease was identified in the asymptomatic patient population but was less pronounced: 1.64% in 1994, 0.81% in 1995, and 0.15% to 0.44% from 1996 to 2003. The low recent stroke rates were confirmed by the California data (0.44% to 0.48% from 1999 to 2003). Changes in the death rate for CEA during this time frame have not been as pronounced, from 0.33% to 0.58% for Maryland and 0.78% to 0.91% for California. CONCLUSIONS: A dramatic decrease in the in-hospital stroke rates in Maryland occurred around 1995. The stroke rates in Maryland in the past 5 years are similar to those in California during the same period. An analysis of data from the two states shows that the in-hospital stroke rate now for carotid endarterectomy is approximately 0.54%.

Preoperative and intraoperative determinants of incisional bulge following retroperitoneal aortic repair.

Although the left flank retroperitoneal incision is a useful approach for many patients undergoing major aortic reconstruction for aneurysmal and occlusive disease, it has been associated with weakening of the flank muscles, resulting in bulges varying from slight asymmetry to huge hernias. The purpose of this study was to determine if the incidence of this complication correlated with identifiable preoperative or intraoperative factors. Fifty consecutive patients undergoing aortic reconstruction via the retroperitoneal approach were followed for 1 year postoperatively for evidence of disfiguring bulges. Bulges were scored as follows: normal/mild, <1-inch protrusion; moderate, protrusion 1-2 inches; severe, protrusion >2 inches and/or pain or true herniation. Preoperatively, patients were administered a questionnaire to elicit demographic and comorbidity data. Fifty-six percent of patients developed a bulge at 1 year. In 43% of these, the bulge was deemed mild and in 54% moderate. One patient developed a severe bulge. Among preoperative comorbidities, no statistically significant correlations were found on bivariate analysis. However, likelihood ratios for bulge development of 5.5 for renal disease and 3.1 for cancer were demonstrated. Conversely, peripheral vascular disease had a likelihood ratio of 0.21 for bulge formation and emphysema, 0.28. On logistic analysis, incision >15 cm and body mass index (BMI) >23 mg/kg(2) were found to correlate strongly with bulge formation (p=0.003, odds ratio=9.1, and p=0.018, odds ratio=16.9, respectively). Together, these yielded a pseudo r (2) of 0.32. BMI >23 mg/kg(2 )was found to yield the greatest explanatory power. These same two variables were found to correlate with severity of bulge: p=0.02 for incision>5 cm and p=0.006 for BMI >23. Of note, gender, age, and extension of the incision into the interspace were not significant on logistic analysis. Preoperatively, surgeons should warn obese patients and those requiring large incisions for extensive disease of their increased risk for poor healing. Intraoperatively, surgeons should aim to minimize incision length.

3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors reduce the risk of perioperative stroke and mortality after carotid endarterectomy.

OBJECTIVE: There is increasing evidence that 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors (statins) reduce cardiovascular and cerebrovascular events through anti-inflammatory, plaque stabilization, and neuroprotective effects independent of lipid lowering. This study was designed to investigate whether statin use reduces the incidence of perioperative stroke and mortality among patients undergoing carotid endarterectomy (CEA). METHODS: All patients undergoing CEA from 1994 to 2004 at a large academic medical center were retrospectively reviewed. The independent association of statin use and perioperative morbidity was assessed via multivariate logistic regression analysis. RESULTS: CEA was performed by 13 surgeons on 1566 patients (987 men and 579 women; mean age, 72 +/- 10 years), including 1440 (92%) isolated and 126 (8%) combined CEA/coronary artery bypass grafting procedures. The indication for CEA was symptomatic disease in 660 (42%) cases. Six hundred fifty-seven (42%) patients received a statin medication for at least 1 week before surgery. Statin use was associated with a reduction in perioperative strokes (1.2% vs 4.5%; P < .01), transient ischemic attacks (1.5% vs 3.6%; P < .01), all-cause mortality (0.3% vs 2.1%; P < .01), and median (interquartile range) length of hospitalization (2 days [2-5 days] vs 3 days [2-7 days]; P < .05). Adjusting for all demographics and comorbidities in multivariate analysis, statin use independently reduced the odds of stroke threefold (odds ratio [95% confidence interval], 0.35 [0.15-0.85]; P < .05) and death fivefold (odds ratio [95% confidence interval], 0.20 [0.04-0.99]; P < .05). CONCLUSIONS: These data suggest that perioperative statin use may reduce the incidence of cerebrovascular events and mortality among patients undergoing CEA.

Mandatory versus selective preoperative carotid screening: a retrospective analysis.

BACKGROUND: Extracranial internal carotid artery stenosis is a risk factor for perioperative stroke in coronary artery bypass (CAB) surgery. Although both selective and nonselective methods of preoperative carotid screening have been advocated, it is unclear which approach is most clinically efficacious. METHODS: Hospital records for 1421 consecutive CAB patients from January 2000 through April 2002 were reviewed. Univariate and multivariate analyses were performed across selected parameters to identify risk factors for significant carotid stenosis (> or = 70%). Patients were retrospectively stratified into high- or low-risk groups based on risk factors common to carotid stenosis and perioperative stroke. The prevalence of carotid stenosis, surgical management, and perioperative stroke rates were determined for each group. RESULTS: One-thousand one-hundred thirty-eight patients out of 1421 patients (80.1%) underwent preoperative carotid screening. The prevalence of significant carotid stenosis was 13.4%. Univariate risk factors for stenosis included an age of more than 65 years, peripheral vascular disease, prior cerebrovascular accident, history of cerebrovascular disease, left main coronary disease, carotid bruit, female gender, and hypertension. Carotid stenosis was a risk factor for stroke, neurologic injury, in-hospital mortality, and longer hospitalization. Prevalence of carotid stenosis was greater in high-risk patients (17.8%, N = 708) versus low-risk patients (6.1%, N = 426). Concomitant or staged carotid endarterectomy (CEA)/CAB was more commonly performed in the high-risk group (5.8% vs. 1%, p < 0.001). All nine patients with significant carotid stenosis who suffered perioperative strokes were in the high-risk group (9 out of 708 vs 0 out of 426, p = 0.016). CONCLUSIONS: In our cohort, selectively screening only patients with either an age of more than 65, carotid bruit, or cerebrovascular disease would have reduced the screening load by nearly 40% with negligible impact on surgical management or neurologic outcomes.

Preoperative selective intercostal angiography in patients undergoing thoracoabdominal aneurysm repair.

OBJECTIVE: This study was designed to test the hypothesis that detection of the location of the major artery supplying the spinal cord, that is, the artery of Adamkiewicz or the great radicular artery (GRA), with angiography would help prevent paraplegia. Knowing which intercostal artery provides this important branch would enable prompt, focused revascularization. METHOD: The surgical outcome in 131 patients with Crawford extent 1 and 2 degenerative aneurysms and 69 patients with descending thoracic aortic dissection was correlated with findings on selective intercostal arteriograms. Angiographic maneuvers were done with care, and the procedures were aborted if there was loose or "shaggy" mural thrombus, significant tortuosity, or difficulty entering each dissection channel. No attempts were made to find major contributions proximal to T6. Subarachnoid drains were placed in all patients, and all but five patients underwent distal aortic perfusion with controlled cooling to 32 degrees F. Five patients underwent cold circulatory arrest, enabling replacement of the distal aortic arch. We defined paraplegia simply as the inability to walk at hospital discharge, paraparesis as impaired ambulation, and both as having spinal cord dysfunction (SCD). RESULTS: A GRA was found in 65 (43%)of the 151 patients studied. Of the 65 patients with the GRA identified, SCD developed in 3 (4.6%) patients. Thirteen of 135 (9.6%) patients in whom the GRA was not identified, either because they were not studied or were studied and the GRA was not found, developed SCD (P =.35) However, when the GRA was identified, SCD occurred only in the group with aortic dissection. None of the 45 patients with degenerative aneurysms with the GRA identified had SCD, compared with 9 of 55 (16%) patients studied but without a GRA found (P =.01). CONCLUSION: The approach with selective intercostal angiography did not improve overall results. One third of our patients were not studied, and they fared as well as patients who were studied and the GRA was localized (not studied, 4 of 49, 8% with SCD; GRA localized, 3 of 65, 5% with SCD; P =.8). However, when the GRA was found, SCD occurred only in patients with aortic dissection. The studies confirmed the concept that the existence of mural thrombus in degenerative aneurysms results in the occlusion of many intercostal arteries, leaving those remaining patent to supply rich vascular watersheds through acquired collateral channels. As a result, in the group of patients with degenerative aneurysms, the identification of the critical intercostal artery allows focused reimplantation with uniform success. This is not the case in patients with aortic dissection. In those patients, most intercostal vessels remain patent, such that the insertion of one pair is insufficient to supply the paravertebral plexus and the spinal cord. Finally, failure to identify the GRA angiographically with our methods does not provide assurance that the GRA does not exist. Therefore negative findings did not provide license to ligate all intercostal arteries.

The distribution of carotid endarterectomy procedures among surgeons and hospitals in New York state: is regionalization of specialized vascular care occurring?

INTRODUCTION: In a published analysis of all carotid endarterectomies (CEAs) performed in New York state from 1990 to 1995, perioperative mortality rate was inversely correlated with surgeon and hospital CEA volume, was significantly higher when CEAs were performed by surgeons who performed less than five CEAs annually, and was significantly lower in hospitals where surgeons performed more than 100 CEAs annually. The purpose of this study was to determine whether this information has influenced practice patterns in New York state. METHODS: The database of the Center for Medical Consumers was queried to determine the volume distribution among surgeons and hospitals of all CEAs performed in New York state in 1999 and 2000. RESULTS: During 1999, 695 surgeons in 169 hospitals performed 9458 CEAs (mean, 13.6 per surgeon). Three hundred fifty-three surgeons (51%) performed less than five CEAs, and 180 (26%) performed only one CEA during the year. Only 41 surgeons (6%) performed more than 50 CEAs. Likewise, in only 28 of the hospitals (17%) were more than 100 CEAs performed during 1999, whereas in 73 of the hospitals (43%) 20 or less CEAs were carried out during the year. During 2000, 684 surgeons performed 8196 CEAs in 165 hospitals. Three hundred fifty-three (52%) performed less than five CEAs, and 229 (33%) performed only one CEA during the year. Only 33 surgeons (5%) performed more than 50 CEAs during 2000. In only 26 hospitals (16%) were more than 100 CEAs performed during 2000, whereas in 71 hospitals (43%) 20 or less CEAs were carried out. CONCLUSION: It appears that published compelling evidence that operator and institutional volume influence outcome has not influenced referral patterns or led to a regionalization of CEA care in New York state. Robust educational programs directed to patients and referring physicians appear indicated.

Durability of thoracoabdominal aortic aneurysm repair in patients with connective tissue disorders.

OBJECTIVE: Thoracoabdominal aortic aneurysm (TAAA) repair is a durable procedure performed with reasonable perioperative mortality and morbidity in patients with atherosclerotic aortic disease. However, the long-term outcome and durability of TAAA repair performed in patients with a connective tissue disorder (CTD) is not well known. METHODS: The records of 257 patients who underwent TAAA repair at the Johns Hopkins Hospital between January 1992 and December 2001 were reviewed. Survival analysis was performed with Kaplan-Meier analysis, and subgroups were compared with the log-rank test. Multivariable analysis was performed with the Cox proportional hazards model and logistic regression. RESULTS: Patients with CTD (n = 31) were seen earlier (mean age, 48.6 +/- 2.9 years) than patients without CTD (mean age, 69.1 +/- 0.6 years; P <.0001, Mann-Whitney U test) and had a greater incidence rate of aortic dissection (52% versus 19%; P <.0001, chi(2) test) and extent I or II aneurysm (77% versus 64%; P =.04). The perioperative (30-day) mortality rate was 6.5% in patients with CTD, which was similar to the rest of the cohort (P =.39, Fisher exact test). The incidence rate of paraparesis/paraplegia was 12.9%/6.5% in patients with CTD, and CTD was the only factor predictive of paraparesis (P =.03; odds ratio, 9.3; logistic regression). The cumulative survival rate among the entire cohort was 53.4% +/- 4.4% at 5 years (Kaplan-Meier), and no difference was seen among patients with or without CTD (P =.16, log-rank test) or among different Crawford extents (P =.29). Of the two late (>6 months) deaths in patients with CTD, none were from aortic rupture or dissection, compared with two of 31 late deaths in patients without CTD. Multivariable analysis confirmed that postoperative renal failure (P =.03) predicted mortality but neither CTD (P =.93), nor Crawford extent (P =.21, Cox regression) predicted mortality. Among survivors, no mean difference was found in largest aortic diameter on follow-up imaging in patients with or without CTD (4.7 +/- 0.3 cm versus 4.4 +/- 0.3 cm; P =.47, Mann-Whitney U test). The cumulative graft patency rate, representing long-term graft stability and with death, rupture, dissection, or recurrent aneurysm as endpoints, was 47.5% +/- 4.6% at 5 years (Kaplan-Meier) and was similar in patients with or without CTD (P =.10, log-rank test). CONCLUSION: TAAA repair appears to be a durable operation, with a reasonable 5-year patient survival rate and a low risk of postoperative paraplegia or additional aortic events. Patients with CTD can expect their outcome, including long-term survival and aortic stability, to be similar to patients without CTD.

Subdural hematoma after thoracoabdominal aortic aneurysm repair: an underreported complication of spinal fluid drainage?

OBJECTIVE: Cerebrospinal fluid (CSF) drainage is a commonly used adjunct to thoracoabdominal aortic aneurysm (TAAA) repair that improves perioperative spinal cord perfusion and thereby decreases the incidence of paraplegia. To date, little data exist on possible complications, such as subdural hematoma caused by stretching and tearing of dural veins, should CSF drainage be excessive. We reviewed our experience with patients in whom postoperative subdural hematomas were detected. METHODS: The records of 230 patients who underwent TAAA repair at the Johns Hopkins Hospital between January 1992 and February 2001 were reviewed. RESULTS: Eight patients had subdural hematomas (3.5%). The four men and four women had a mean age of 60.6 years; two of these patients had a connective tissue disorder. All patients had lumbar drains placed before surgery, including one patient who underwent an emergency operation for rupture. Drains were set to allow drainage for CSF pressure greater than 5 cm H(2)O in all but one patient set for 10 cm H(2)O; spinal cooling was not performed in any patient. All drains were removed on the third postoperative day. In patients in whom subdural hematomas developed, the mean amount of CSF removed after surgery was 690 +/- 79 mL, which was significantly greater than the amount drained from patients in whom subdural hematomas did not develop (359 +/- 24 mL; P =.0013, Mann-Whitney U test). Six patients had postoperative subdural hematomas detected during hospitalization (mean postoperative day, 9.3; range, 2 to 16), and two patients were seen in delayed fashion after discharge from the hospital at 1.5 and 5 months. Four patients died of the subdural hematoma (50%); only one of these patients had neurosurgical intervention. All four survivors responded to neurosurgical intervention and are neurologically healthy. Two patients, both of whom were seen in delayed fashion, needed a lumbar blood patch. Multivariate logistic regression identified the volume of CSF drained as the only variable predictive of occurrence of subdural hematoma (P =.01). CONCLUSION: Subdural hematoma is an unusual and potentially catastrophic complication after TAAA repair. Prompt recognition and neurosurgical intervention is necessary for survival and recovery after acute presentation. Epidural placement of a blood patch is recommended if a chronic subdural hematoma is detected. Care should be taken to ensure that excessive CSF is not drained perioperatively, and higher (10 cm H(2)O) lumbar drain popoff pressures may be necessary together with meticulous monitoring of patient position and neurologic status.