View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector PRESERVATION OF INTERCOSTAL ARTERIES DURING THORACOABDOMINAL AORTIC ANEURYSM SURGERY: A RETROSPECTIVE STUDY Scott D. Ross, MD Objective: The purpose of this article was to examine the influence of Irving L. Kron, MD reimplantation of patent intercostal and lumbar arteries on the inci- Patrick E. Parrino, MD dence of postoperative paraplegia/paraparesis in patients undergoing Kimberly S. Shockey, MS clamp-and-sew surgical repair of thoracoabdominal aortic aneurysms. John A. Kern, MD Curtis G. Tribble, MD Methods: Data from January 1987 through December 1997 were retro- spectively collected on 132 patients. Ninety-one patients in group I underwent aneurysm repairs before January 1995 and did not undergo intercostal artery reimplantation. Group II included the more recent 41 patients who had vessels between the eighth thoracic intercostal and the second lumbar arteries reimplanted to the graft or preserved at the aor- tic anastomoses. Results: The operative mortality rate was 13.2% (12/91) in group I and 4.9% (2/41) in group II (P = .22). The incidence of post- operative paraplegia was significantly lower in the more recent cohort of patients (8.8% [8/91] in group I vs 0% [0/41] in group II, P = .05). The overall rate of spinal cord dysfunction was lowered from 9.9% (9/91) in group I to 2.4% (1/41) in group II (P = .17). However, a multivariable logistic regression analysis identified only aneurysm extent (Crawford type I and type II) as a predictor of less postoperative spinal cord injury (P = .08). The average aortic crossclamp time in group I was 30.3 ± 11.5 (SD) minutes, and the time of aortic occlusion in group II was not sig- nificantly prolonged, with an average crossclamp time of 31.0 ± 21.0 (SD) minutes (P = .88). Conclusions: An aggressive approach to maintain intercostal artery patency during clamp-and-sew repair of thoracoab- dominal aortic aneurysms may effectively lower the incidence of spinal cord injury without prolonging aortic crossclamp time. (J Thorac Cardiovasc Surg 1999;118:17-25) he surgical repair of aneurysms of the thoracoab- in resection of these aneurysms. The incidence of spinal T dominal aorta has undergone considerable advance- cord dysfunction varies, but the largest series to date ments over the past several decades. Despite many sur- reported the development of postoperative paraplegia or gical successes, the threat of spinal cord ischemia and paraparesis in 16% of their patients.1 Spinal cord injury postoperative paraplegia remains an unsolved problem during the repair of thoracoabdominal aneurysms is the result of ischemia caused by aortic crossclamping and ischemia caused by devascularization from interruption From the Department of Surgery, Division of Thoracic and of the critical blood supply of the spinal cord. However, Cardiovascular Surgery, University of Virginia Health Sciences Center, Charlottesville, Va. owing to the variability and complexity of the blood Read at the Twenty-fourth Annual Meeting of The Western Thoracic supply of the spinal cord, the management of this criti- Surgical Association, Whistler, British Columbia, June 24-27, cal circulation remains controversial. The most impor- 1998. tant artery of the thoracolumbar region of the spinal Received for publication July 15, 1998; revisions requested Sept 4, 1998; revisions received Feb 10, 1999; accepted for publication cord may be the artery of Adamkiewicz, arising most March 8, 1999. commonly in the region from the ninth thoracic inter- Address for reprints: Curtis G. Tribble, MD, Department of Surgery, costal to the first lumbar artery. The reattachment of Division of Thoracic and Cardiovascular Surgery, Box 3111, MR4 these intercostal and lumbar arteries when replacing Building, University of Virginia Health Sciences Center, Charlottesville, VA 22908. thoracoabdominal aneurysms has been postulated to Copyright © 1999 by Mosby, Inc. preserve the artery of Adamkiewicz and play an impor- 0022-5223/99 $8.00 + 0 12/6/98432 tant role in preserving spinal cord function.2-8 However, 17 18 Ross et al The Journal of Thoracic and Cardiovascular Surgery July 1999 Table I. Preoperative patient characteristics arrest were excluded. During the more recent era of the study Group I Group II since January 1995, 6 patients required surgical repair by Characteristic (n = 91) (n = 41) P value means of these various techniques. These techniques were generally applied in cases in which the proximal aorta was Age (y) 60.8 65.5 .29 not able to be clamped or in cases in which there was preop- Male (%) 52/91 (57.1) 27/41 (65.9) .34 erative evidence of severe left ventricular dysfunction. Hypertension (%) 61/91 (67.0) 32/41 (78.0) .20 Medical records of all patients were reviewed and preoper- Myocardial infarction 20/91 (22.0) 12/41 (29.3) .37 ative variables of age, sex, thoracoabdominal aneurysm (%) Prior aortic surgery 26/91 (28.6) 12/41 (29.3) .93 extent, symptomatic status, the presence of dissection, the (%) presence of rupture, history of hypertension, history of Aneurysm extent .80 myocardial infarction or congestive heart failure, and previ- Type I (%) 53/91 (58.2) 21/41 (51.2) ous aortic resection were obtained. Thoracoabdominal Type II (%) 14/91 (15.4) 6/41 (14.6) aneurysm extent was classified according to the scheme pro- Type III (%) 10/91 (11.0) 5/41 (12.2) posed by Crawford: type I, entire descending and upper Type IV (%) 14/91 (15.4) 9/41 (22.0) abdominal aorta; type II, entire descending and abdominal Dissection .47 aorta; type III, variable extents of descending and abdominal Chronic (%) 15/91 (16.5) 5/41 (12.2) aorta; and type IV, entire abdominal aorta with the graft Acute (%) 9/91 (9.9) 2/41 (4.9) extending proximal to the celiac axis. The aortic segment that Rupture (%) 18/91 (19.8) 3/41 (7.3) .07 was resected was used to determine the classification of aneurysm extent. The duration of aortic occlusion was obtained from the operative record for each patient. The post- operative outcomes of death and spinal cord injury manifest- others have questioned the significance of reimplanting ed by paraplegia or paraparesis were determined for each these vessels for spinal protection.9 patient. Operative mortality included deaths within the We10 have previously reported our experience with patients’ entire hospital length of stay or within 30 days after the clamp-and-sew technique for the repair of thora- the operation if they were discharged from the hospital. All coabdominal aortic aneurysms. Recently, we have patients who died after the operation had thorough evalua- modified our technique of these repairs by preserving tions for neurologic injury recorded before death. We defined all patent vessels between the eighth thoracic inter- paraplegia as a permanent neurologic deficit and paraparesis as temporary weakness of a lower extremity. costal and the second lumbar arteries in an effort to The two groups of patients were not significantly different reduce the incidence of spinal cord ischemia. The cur- with respect to a number of preoperative characteristics that rent study examines the addition of intercostal artery are believed to have an impact on outcome in patients with reimplantation to the clamp-and-sew technique and its aneurysmal disease, as shown in Table I. Specifically, the influence on postoperative paraplegia without addition- extent of the aortic lesions did not differ significantly al combinations of adjuvant techniques for spinal cord between the two groups of patients. In particular, 66% of the protection. more recent group and 73% of the earlier group of patients had thoracoabdominal aneurysms classified as Crawford type I and type II aneurysms, indicating that a majority of patients Patients and methods in each group had lesions posing a relatively high risk for Patient population. All patients undergoing clamp-and- paraplegia. In addition, we used a multivariable logistic sew repair of aneurysms of the thoracoabdominal aorta at the regression analysis to create a propensity score for each University of Virginia Health Sciences Center from January patient in the data set. In the analysis, we found the logistic 1987 through December 1997 were included in the study. A models with and without the propensity scores as a covariate total of 132 patients were identified for review. Ninety-one to have almost identical estimates of the baseline characteris- patients in group I underwent aneurysm repairs before tics of the two patient groups. January 1995 and did not undergo intercostal artery reim- Operative technique. All thoracoabdominal repairs were plantation. In these patients, potentially critical intercostal carried out completely retroperitoneally through a thoracoab- vessels were oversewn. Group II included 41 patients who dominal incision. All patients were intubated with a double- underwent aneurysm operations from January 1995 through lumen endotracheal tube that permitted deflation of the left December 1997 and had critical vessels between the eighth lung, facilitating exposure and repair. Intraoperatively, arteri- thoracic intercostal and the second lumbar arteries reimplant- al blood pressure, central venous pressure, pulmonary artery ed to the graft or preserved at the aortic anastomoses. Patients pressure, continuous electrocardiogram, urine output, tem- with isolated aneurysms of the descending thoracic aorta perature, arterial blood gases, electrolytes, and hemoglobin were not