Timing of endovascular repair of blunt traumatic thoracic aortic transections

Timing of endovascular repair of blunt traumatic thoracic aortic transections

From the Midwestern Vascular Surgical Society Timing of endovascular repair of blunt traumatic thoracic aortic transections Amy B. Reed, MD, J. Keith...

94KB Sizes 0 Downloads 37 Views

From the Midwestern Vascular Surgical Society

Timing of endovascular repair of blunt traumatic thoracic aortic transections Amy B. Reed, MD, J. Keith Thompson, DO, Charles J. Crafton, PA-C, Cindy Delvecchio, RN, and Joseph S. Giglia, MD, Cincinnati, Ohio Background: Patients with blunt traumatic thoracic aortic transection (BTTAT) just distal to the takeoff of the left subclavian artery typically have concomitant injuries that make open emergent surgical repair highly risky. Over the past decade, endovascular repair of the injured thoracic aorta with commercially available and custom-made covered stents has developed as a viable option, with reported decreases in short-term morbidity and mortality. If active extravasation of contrast from the injured thoracic aorta is not appreciated on chest computed tomography scan, other concurrent injuries of the head, abdomen, and extremities can often be repaired with careful control of blood pressure. The timing of endovascular repair of the traumatic thoracic aortic transection, however, often comes into question, particularly with the presence of fever, pneumonia, or bacteremia. We sought to identify a time frame during which endovascular repair of BTTAT could safely be performed. Methods: Age, concomitant injuries, time from trauma to repair, type of device, and major outcomes were recorded. Results: Over a 5-year period (January 2000 to March 2005), 51 patients presented with BTTAT. Twenty-seven (52.9%) patients with BTTAT died shortly after arrival. Of the remaining 24, 9 underwent emergent open repair, with 1 intraoperative death. Two delayed open repairs were performed. Thirteen patients with BTTAT underwent delayed endovascular repair. Successful endovascular repair of BTTAT was performed in all 13 patients, with no intraoperative deaths. Seven patients were treated with commercial devices and six with custom-made covered stents. None of the repairs was performed emergently. The timing of repair ranged from 1 day to 7 months (median, 6 days), and all patients were treated aggressively with ␤-blockade before surgery. One patient was discharged from the hospital and underwent elective repair at a later date. Three patients died in the postoperative period (30 days): two from multisystem organ failure and one from iliac artery complications encountered at the time of device deployment. The remaining 10 patients were successfully discharged to a rehabilitation facility. Conclusions: The opportunity to successfully perform endovascular repair of BTTAT may be possible many days after the initial injury in the hemodynamically stable trauma patient. ( J Vasc Surg 2006;43:684-8.)

Blunt traumatic thoracic aortic transection (BTTAT) has been considered a surgical emergency with an immediate need for repair since Parmley and colleagues’1 seminal article in 1958 that described an 85% at-the-scene death rate. Mortality over the ensuing 48 hours was assumed to be approximately 1% per hour in patients who did not receive intervention. Currently, only 10% to 20% of patients with BTTAT survive long enough to be treated in a hospital setting, and of these, 5% to 10% will die in the first 4 to 6 hours. This leaves a select subset of patients who remain hemodynamically stable enough to be considered for intervention, albeit at a high risk for mortality from associated injuries. Maintaining systolic blood pressure less than 120 mm Hg or mean arterial pressure less than 80 mm Hg has been shown to significantly reduce the risk of rupture.2-5 In keeping with the practice of rate and blood pressure conFrom the Division of Vascular Surgery, University of Cincinnati Medical Center. Competition of interest: none. Presented at the Twenty-ninth Annual Meeting of the Midwest Vascular Surgical Society, Chicago, Ill, Sept 15, 2005. Reprint requests: Amy B. Reed, MD, Division of Vascular Surgery, University of Cincinnati, 231 Albert Sabin Way, Cincinnati, OH 45267 (e-mail: [email protected]). 0741-5214/$32.00 Copyright © 2006 by The Society for Vascular Surgery. doi:10.1016/j.jvs.2005.12.006

684

trol, many centers have reported the ability to delay operative repair for extended periods of time until associated injuries have been treated or stabilized.2,4-8 Despite delays in repair, an operative approach for the injured aorta can still result in significant morbidity and mortality. Since Dake et al9 first reported successful endovascular repair of thoracic aortic aneurysms, numerous studies have been published regarding their use in the trauma patient10-12; however, few have focused on the timing of stent graft repair in this high-risk patient population. We sought to identify a time frame during which endovascular repair of BTTAT could safely be performed, by reviewing our experience over the past 5 years. METHODS The University of Cincinnati Vascular Surgery data registry was queried for patients who underwent endovascular repair of a BTTAT at the University of Cincinnati Medical Center from January 1, 2000, to March 31, 2005. The Trauma Registry was also queried during this time period to identify the outcomes of all patients who presented with BTTAT. All patients underwent chest radiography and computed tomography (CT). Arch aortography with a marker catheter was performed on each patient in the angiography suite by vascular surgery staff to determine the distance from the left subclavian artery and the length of the injured descending thoracic aorta before endovascular

JOURNAL OF VASCULAR SURGERY Volume 43, Number 4

Reed et al 685

Table I. Associated injuries in patients with BTTAT Patient No. 1 2 3 4 5 6 7 8 9 10 11 12 13

Age (y)

Sex

Time to repair (d)

Associated injuries*

GCS at presentation

ISS

72 77 75 34 66 64 19 77 78 47 23 56 24

M M F M M F M M F M F M M

1 1 8 25 19 7 4 8 210 2 2 1 33

CHI; chest; major abdominal; ortho Major abdominal Ortho CHI; chest; ortho Major abdominal; ortho Chest; ortho Chest; major abdominal; ortho CHI; ortho CHI; chest; ortho Chest; major abdominal Major abdominal; ortho CHI; major abdominal; ortho CHI; chest; ortho

13 15 11 6 14 15 15 15 15 15 10 15 4

50 29 25 36 29 75 29 42 41 33 41 36 54

BTTAT, Blunt traumatic thoracic aortic transection; GCS, Glasgow Coma Score; ISS, Injury Severity Score; CHI, closed head injury. *Major abdominal: significant liver, splenic, or bowel injury necessitating laparotomy; chest: pulmonary contusion, flail chest, or significant bony injury; ortho: any orthopedic injuries.

repair. The aortic neck diameter was measured on the CT scan proximal and distal to the zone of injury. Custommade grafts, compassionate-use AneuRx (Medtronic Inc.) or Excluder (W.L. Gore & Associates, Flagstaff, AZ) aortic extender cuffs, were oversized 10% to 20%. A neck length of 15 mm was considered adequate for stent-graft use. Beta blockade and antihypertensives were used liberally before surgery to keep the systolic blood pressure at or less than 120 mm Hg. All patients underwent endovascular repair in the operating room by staff vascular surgeons. Femoral cut-down was routinely used, although one patient required thoracotomy for distal thoracic aortic exposure secondary to extensive aortic and peripheral vascular occlusive disease early in the experience. Retroperitoneal exposure of the iliac artery was required in select patients to allow adequate sheath length for delivery and deployment of the stent. Brachial artery access was often used for device tracking or because of a need for more precise imaging. Intraoperative pharmacologic hypotension to 90 mm Hg was often used at time of stent deployment. Completion aortograms were performed after all endovascular stent deployments. A follow-up chest CT scan was obtained within 1 week of placement, and serial CT scans were obtained at 6 and 12 months and then annually thereafter. Medical, operative, and radiologic reports were reviewed. Age, concomitant traumatic injuries and operations, use of ␤-blockade and antihypertensives, time from initial injury to repair, type of device, and major outcomes were analyzed. Closed head injury was determined by an abnormal head CT scan (hemorrhage or edema), increased intracranial pressure, or a low Glasgow Coma Score at the time of the trauma. Pulmonary injury was defined as pulmonary contusion identified on chest imaging with one or more of the following to maintain satisfactory oxygenation: PaO2/fraction of inspired oxygen less than 300 mm Hg or positive end-expiratory pressure requirements of at least 7.5 cm H2O. The institutional review board granted approval for evaluation of this patient population.

RESULTS Fifty-one patients presented to the University of Cincinnati Medical Center with BTTAT over the 5-year study period from January 1, 2000, to March 31, 2005. Twentyseven (52.9%) patients died shortly after their arrival to the medical center. Of the surviving 24 patients, 9 (37.5%) underwent emergent open operation after extravasation of contrast was noted on CT scan. One of the nine patients undergoing emergent repair died during surgery. Two patients underwent delayed open repair at 12 days and 3 months, respectively, without complication. Successful endovascular repair of BTTAT was performed in 13 (54.2%) of 24 patients (aged 19-78 years). None of the repairs was performed emergently, thus allowing time for arch aortography and appropriate measuring for an endovascular stent. All patients who presented at the time of injury had additional injuries, with an average Injury Severity Score of 40 (Table I). During this period, all study patients were maintained on ␤-blockade and antihypertensives to keep the systolic blood pressure less than 140 mm Hg. One patient with significant head injuries was discharged to a rehabilitation facility with an untreated BTTAT given her poor prognosis. She had slow, steady neurologic improvement and returned for aortic endovascular repair 7 months after injury; no enlargement was noted on her interval CT scan. The timing of intervention ranged from 1 day to 7 months (median, 6 days), and most patients (64%) underwent repair four or more days beyond the date of injury (Fig). All patients remained hemodynamically stable from the time of injury to the time of repair. Fifty-three percent (7/13) of patients with BTTAT underwent operation for chest, abdominal, or orthopedic injuries in the interval before endovascular stent repair without obvious compromise to the injured aortic segment. Five custom endovascular stents were constructed early in the experience by stretching a 40-mm polytetrafluoroethylene graft over 35-mm Gianturco (Cook, Bloomington, IN)

JOURNAL OF VASCULAR SURGERY April 2006

686 Reed et al

Fig. Timing of repair of blunt traumatic thoracic aortic transection.

stents. Custom grafts were typically deployed via a 22F KellyTimmerman sheath. Access for this particular type of stent graft often required a through-and-through wire technique from the right brachial to the left common femoral artery for ease of trackability and deployment. Commercial devices were made available for treatment of infrarenal abdominal aortic aneurysms, and off-label compassionate use (Talent graft) was sought in other areas of aortic pathology, such as BTTAT.13 The remaining eight patients in the study underwent exclusion of the BTTAT with Excluder aortic extender cuffs. Three patients had large Wallstents (Boston Scientific Inc.) used as a scaffolding, followed by placement of covered endovascular stents. The Wallstent allowed for precise deployment with regard to the origin of the left subclavian artery, in addition to preventing short covered aortic extender cuffs from angling into the aorta, should the traumatic defect prove to be as long as or longer than the covered stent. The ease of deployment with this 18F device required cut-down of either the common femoral artery (10 patients) or the external iliac artery (2 patients) in select patients with a long chest and torso to allow adequate length for the 61-cm delivery system of the Excluder aortic extender cuffs. Brachial artery access for precise imaging of the left subclavian artery origin was used in 11 cases. One patient required thoracotomy for distal thoracic aortic exposure secondary to extensive aortic and peripheral vascular occlusive disease early on in the experience. A completion aortogram after stent deployment revealed functioning grafts with complete exclusion of BTTAT in all cases without evidence of endoleak. Estimated blood loss ranged from 50 to 500 mL except for the patient who had an iliac arterial injury. No procedure-related complications of distal limb ischemia, wound infection, or paraplegia were noted in the postoperative period. The 30-day procedure-related mortality of the study group was 7.1%: one death occurred on postoperative day 1 in a 64-year-old woman who experienced an iliac injury after receiving a custom stent graft. There were no deaths in the subset of patients who underwent BTTAT repair with Excluder aortic extender cuffs. The overall mortality was 30% (4 of 13 patients) in this patient population. Two patients died during the initial hospitalization from multisystem organ failure secondary to sepsis from pneumonia, and one died from widespread

mesenteric ischemia secondary to an avulsion injury, which was attended to on trauma day 1, 5 days before the thoracic aortic intervention. One patient who underwent initial successful treatment of BTTAT in 2000 returned with an endoleak 4 months after initial stent graft implantation but died as a result of multisystem organ failure 5 days later. His intraoperative course was complicated by hemorrhage secondary to an iliac artery injury sustained during removal of the sheath for the custom stent (Table II). CT scans obtained before discharge from the hospital revealed exclusion in all cases. Of the 10 patients available for follow-up (6-24 months; mean, 12 months) upon discharge, 1 (the first patient in the series in 2000) returned with evidence of endoleak at 4 months after the initial endovascular repair, 2 died of unknown causes, and 2 have been lost to follow-up. No patient since our first study patient in 2000 has required reintervention or open surgical repair of the previously stented thoracic aorta. DISCUSSION Patients who present with BTTAT distal to the left subclavian artery typically have concomitant injuries that make open emergent surgical repair highly risky. The natural history of aortic transection is relatively self-selective; the vast majority (80%-90%) of patients exsanguinate at the scene. Of the 10% to 20% who make it to a hospital alive, 15% to 38% die later, largely as a result of associated injuries2; this suggests that patients who do not die in the first 24 hours may remain hemodynamically stable with aggressive control of systolic blood pressure and heart rate. Consequently, delayed aortic repair has been advocated by some to help decrease the morbidity of open surgical repair in patients with associated injuries.6 Several groups have proposed waiting until the patient is resuscitated—free of active infection or sepsis and stable from a cardiopulmonary standpoint—and found little change with regard to morbidity and mortality.14 This approach seems feasible with the heart rate and blood pressure well controlled as demonstrated by Pierangeli et al,15 who noted successful delayed repair in 72% of patients with BTTAT at a mean of 8.6 months. Maggisano et al6 reported successful delayed repair in 53% of patients with BTTAT ranging from 1 day to 7 months to allow resolution of concomitant severe injuries. There has been a paradigm shift in the management of BTTAT in the past decade, with emphasis on blood pressure control and weighing the need for emergent repair against the risks of operation due to associated traumatic comorbidities.16 Clearly there exist patients, such as those with contrast extravasation on CT scan, who are at high risk for rupture.17 Alternatively, there are associated injuries, particularly intra-abdominal solid-organ bleeding, that often take precedence over immediate repair of an aortic injury.18 Several reports in the literature indicate delay or complete avoidance of repair in patients with severe head, cardiac, or pulmonary injury, as well as advanced age.2,4,6,7,19,20 Careful blood pressure control plays a central role in the nonoperative management of stable patients with BTTAT. It has been argued that the use of antihypertensive and

JOURNAL OF VASCULAR SURGERY Volume 43, Number 4

Reed et al 687

Table II. Repair of blunt traumatic thoracic aortic transection Patient No.

Age (y)

Sex

Time to repair (d)

Thoracic aortic injury

1 2

72 77

M M

1 1

PSA Tear with PSA

Talent Custom

No No

3 4 5 6

75 34 66 64

F M M F

8 25 19 7

Custom Custom Custom Custom

No No No No

7 8 9 10

19 77 78 47

M M F M

4 8 210 2

PSA Tear PSA Full-thickness injury PSA Tear PSA PSA

None* Ischemic bowel from traumatic bowel injury preop; support withdrawn POD 1 None None None Iliac artery injury; death POD 1

No No No No

None None None None

11

23

F

2

PSA

No

None

12

56

M

1

Tear

No

MSOF; death POD 6

13

24

M

33

3 Excluder cuffs Custom 3 Excluder cuffs 3 Excluder cuffs; Wallstent 3 Excluder cuffs; Wallstent 3 Excluder cuffs; Wallstent 3 Excluder cuffs

No

None

Full-thickness injury

Endoleak at completion?

Device

Complications

PSA, Pseudoaneurysm; preop, before surgery; POD, postoperative day; MSOF, multisystem organ failure. *Endoleak at 4 months; death on POD 5 after reintervention, 4 months after an initial uncomplicated procedure.

rate-controlling agents during any period of delay in stable patients with BTTAT to prevent extension of injury has become the standard of care.21 Mattox and Wall22 summarized the surgical literature in 2000 and found more than 500 patients managed with deliberate delay of surgical repair of BTTAT, 10% of whom had not yet undergone repair yet at the time of publication. Approximately 2% of these patients experienced rupture, all of whom had increased blood pressures. No patient in our study became unstable or died of a ruptured BTTAT while awaiting endovascular repair. All patients had CT scans on admission that demonstrated BTTAT without active extravasation of contrast. Emergent open repair was recommended for patients with evidence of contrast leak from the injured thoracic aorta. By delaying intervention, patients were allowed to stabilize from a hemodynamic standpoint, and attention could be directed toward associated life-threatening injuries. This period of waiting may allow for optimization and stabilization of the patient, as well as appropriate device acquisition, particularly in the case of smaller hospitals, where stent supplies may be limited. Performing these cases on stable, optimized trauma patients during peak operating room hours may help optimize patient outcome and safety. One patient with significant head injuries was discharged to a rehabilitation facility with an untreated BTTAT, given her poor prognosis. She had slow, steady neurologic improvement and returned for aortic endovascular repair 7 months after injury; no enlargement was noted on her interval CT scan. It is interesting to note that a similar case was noted in the trauma database, where a patient with BTTAT and profound neurologic impairment was evaluated by the cardiothoracic service and discharged untreated. The patient returned 3 months later, neurolog-

ically improved, and underwent successful open repair of a stable BTTAT. Reports in the open surgical literature are appearing that describe improved outcomes with delaying operation.15,23 The use of ␤-blockade and antihypertensives has allowed patients to stabilize and undergo delayed open repair with good results, as demonstrated by several authors. With a minimally invasive endovascular approach to the same problem, it would seem reasonable that acceptable outcomes with delayed repair could be achieved in the stable patient treated with ␤-blockade and antihypertensives. Presumably, repair with aortic endovascular stent grafts can be accomplished with decreased morbidity and mortality. No patient in our study required conversion to open surgery, experienced paraplegia, or developed renal failure secondary to the procedure. This is especially critical for the polytrauma patient, who often cannot tolerate the additional injuries that may come about from an open surgical approach. Access for large delivery sheaths can make traversal of diseased or diminutive vessels a challenge, particularly in the elderly female population. Alternative access sites, such as the iliac arteries and temporary prosthetic aortic conduits, may be necessary to allow safe passage of the stent to its intended location. Curvature of the aortic arch can require stiffer wires, such as the Lunderquist, or a through-and-through approach from the right axillary or brachial access to the left common femoral artery to help facilitate delivery. All of the patients in our study underwent preoperative endovascular planning with aortography with a marker catheter. Future technological advancements in noninvasive imaging, such as CT angiography with three-dimensional reconstruction, will likely make a separate diagnostic arterio-

688 Reed et al

gram unnecessary in most institutions that deal with trauma patients. Our study on timing of endovascular repair concurs with those of open surgical repair. Over the past decade, endovascular repair of the injured thoracic aorta with commercially available and custom-made covered stents has developed as a viable option with decreased short-term morbidity and mortality. Although this technique conceivably could result in definitive repair in the elderly patient, it remains unclear how this minimally invasive therapy will hold up in the more typical young trauma patient. Given that the adjacent aorta for attachment is often of normal strength and caliber, presumably this segment will not be as susceptible to dilation and potential stent migration in the long term. Poor follow-up is often the norm in the trauma population, and this makes it difficult to detect endoleaks and migration. Careful discussion with family members regarding its importance at the time of consent may help facilitate compliance. Input into local, regional, or national data registries may also help facilitate follow-up in addition to ascertaining long-term outcomes. Our study is limited by its retrospective nature and small size and the heterogeneity of patients. Selection bias undoubtedly occurs in the patient population that survives the first 4 to 6 hours of hospitalization, because this is often the group that will be able to wait many days for BTTAT repair. CONCLUSION BTTAT is a serious life-threatening injury that the vast majority of patients do not survive. For select patients who survive more than 3 to 4 hours after arriving in the hospital and remain hemodynamically stable without evidence of contrast extravasation on CT, careful control of heart rate and blood pressure may allow time for adequate resuscitation and attention to other associated injuries before endovascular repair. AUTHOR CONTRIBUTIONS Analysis and interpretation: AR, JG Data collection: KT, CC, CD Writing the article: AR Critical revision of the article: AR, JG Final approval of the article: AR Statistical analysis: AR Obtained funding: N/A Overall responsibility: AR REFERENCES 1. Parmley LF, Mattingly TW, Manion WC, Jahnke EJ Jr. Nonpenetrating traumatic injury of the aorta. Circulation 1958;17:1086-101. 2. Camp PC, Shackford SR. Outcome after blunt traumatic thoracic aortic laceration: identification of a high-risk cohort. Western Trauma Association Multicenter Study Group. J Trauma 1997;43:413-22.

JOURNAL OF VASCULAR SURGERY April 2006

3. Fabian TC, Davis KA, Gavant ML, Croce MA, Melton SM, Patton JH Jr, et al. Prospective study of blunt aortic injury: helical CT is diagnostic and antihypertensive therapy reduces rupture. Ann Surg 1998;227:66676; discussion 676-7. 4. Pate JW, Gavant ML, Weiman DS, Fabian TC. Traumatic rupture of the aortic isthmus: program of selective management. World J Surg 1999; 23:59-63. 5. Wahl WL, Michaels AJ, Wang SC, Dries DJ, Taheri PA. Blunt thoracic aortic injury: delayed or early repair? J Trauma 1999;47:254-9; discussion 259-60. 6. Maggisano R, Nathens A, Alexandrova NA, Cina C, Boulanger B, McKenzie R, et al. Traumatic rupture of the thoracic aorta: should one always operate immediately? Ann Vasc Surg 1995;9:44-52. 7. Akins CW, Buckley MJ, Daggett W, McIlduff JB, Austen WG. Acute traumatic disruption of the thoracic aorta: a ten-year experience. Ann Thorac Surg 1981;31:305-9. 8. Fabian TC, Richardson JD, Croce MA, Smith JS Jr, Rodman G Jr, Kearney PA, et al. Prospective study of blunt aortic injury: Multicenter Trial of the American Association for the Surgery of Trauma. J Trauma 1997;42:374-80; discussion 380-3. 9. Dake MD, Miller DC, Semba CP, Mitchell RS, Walker PJ, Liddell RP. Transluminal placement of endovascular stent-grafts for the treatment of descending thoracic aortic aneurysms. N Engl J Med 1994;331: 1729-34. 10. Orend KH, Pamler R, Kapfer X, Liewald F, Gorich J, SunderPlassmann L. Endovascular repair of traumatic descending aortic transection. J Endovasc Ther 2002;9:573-8. 11. Haulon S, Koussa M, Beregi JP, Decoene C, Lions C, Warembourg H. Stent-graft repair of the thoracic aorta: short-term results. Ann Vasc Surg 2002;16:700-7. 12. Wellons ED, Milner R, Solis M, Levitt A, Rosenthal D. Stent-graft repair of traumatic thoracic aortic disruptions. J Vasc Surg 2004;40: 1095-100. 13. Sam A II, Kibbe M, Matsumura J, Eskandari MK. Blunt traumatic aortic transection: endoluminal repair with commercially available aortic cuffs. J Vasc Surg 2003;38:1132-5. 14. Holmes JH IV, Bloch RD, Hall RA, Carter YM, Karmy-Jones RC. Natural history of traumatic rupture of the thoracic aorta managed nonoperatively: a longitudinal analysis. Ann Thorac Surg 2002;73: 1149-54. 15. Pierangeli A, Turinetto B, Galli R, Caldarera L, Fattori R, Gavelli G. Delayed treatment of isthmic aortic rupture. Cardiovasc Surg 2000;8: 280-3. 16. Mattox KL. Red River anthology. J Trauma 1997;42:353-68. 17. Clark DE, Zeiger MA, Wallace KL, Packard AB, Nowicki ER. Blunt aortic trauma: signs of high risk. J Trauma 1990;30:701-5. 18. Borman KR, Aurbakken CM, Weigelt JA. Treatment priorities in combined blunt abdominal and aortic trauma. Am J Surg 1982;144: 728-32. 19. Karmy-Jones R, Carter YM, Nathens A, Brundage S, Meissner MH, Borsa J, et al. Impact of presenting physiology and associated injuries on outcome following traumatic rupture of the thoracic aorta. Am Surg 2001;67:61-6. 20. Rousseau H, Soula P, Perreault P, Bui B, Janne d’Othee B, Massabuau P, et al. Delayed treatment of traumatic rupture of the thoracic aorta with endoluminal covered stent. Circulation 1999;99:498-504. 21. Feliciano DV, Rozycki GS. Advances in the diagnosis and treatment of thoracic trauma. Surg Clin North Am 1999;79:1417-29. 22. Mattox KL, Wall MJ Jr. Historical review of blunt injury to the thoracic aorta. Chest Surg Clin North Am 2000;10:167-82. 23. Langanay T, Verhoye JP, Corbineau H, Agnino A, Derieux T, Menestret P, et al. Surgical treatment of acute traumatic rupture of the thoracic aorta: a timing reappraisal? Eur J Cardiothorac Surg 2002;21: 282-7. Submitted Sep 15, 2005; accepted Dec 6, 2005.