- Email: [email protected]
Routine immediate direct bronchial artery revascularization for single-lung transplantation
Routine Immediate Direct Bronchial Artery Revascularization for Single-Lung Transplantation Richard C. Daly, MD, and Christopher G. A. McGregor, FRCS Section of Cardiac Surgery, Mayo Clinic and Mayo Foundation, Rochester, Minnesota
Ischemia of the donor airway remains a significant cause of morbidity after single-lung transplantation; serious manifestations may occur early (anastomoticdehiscence) or late (stricture). Direct, immediate revascularization of the donor bronchial arteries, using the recipient internal thoracic artery, was performed in 10 consecutive recipients of single-lung transplants for whom we procured the organs. Mean recipient age was 52.6 years (range, 43 to 59 years); 6 were male and 4 female. Recipient diagnoses were emphysema (6), obliterative bronchiolitis (2), pulmonary fibrosis (l),and primary pulmonary hypertension (1). Bronchial artery revascularization initially prolonged the ischemic time by only 15 to 20 minutes;
T
he techniques and indications for single-lung transplantation (SLT) have evolved considerably over the last decade [l].However, the lung remains the only solid organ transplanted without a systemic arterial blood supply. Bronchial ischemia may result in early (granulation tissue, mucosal sloughing, or bronchial dehiscence) or late (stricture) complications of airway healing. Wrapping the bronchial anastomosis with vascularized tissue can provide a systemic blood supply to the distal airway [Z,31, but this takes time to develop and is provided only via small collaterals. In spite of bronchial omentopexy, early and late complications of airway ischemia continue to be a significant cause of morbidity after SLT, affecting approximately 10% to 19% of patients [GI.In addition, evidence is emerging of an association between bronchial ischemia and an increased incidence of perioperative pneumonia [7, 81. Bronchial ischemia may also play a role in the development of obliterative bronchiolitis (OB) after SLT. The concept of direct bronchial artery revascularization (BAR) was first suggested and demonstrated experimentally by Metras in 1950 [9, 101. Others have subsequently contributed to an understanding of the anatomy of the bronchial arteries, and to the technical goal of BAR [ 11-14]. Bronchial artery revascularization has been successfully used clinically for en-bloc double-lung transplantation with considerable improvement in tracheal anastomotic healing [15, 161 over previous reports involving a tracheal anastomosis [14, 171. Presented at the Thirtieth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 31-Feb 2, 1994. Address reprint requests to Dr Daly, Section of Cardiac Surgery, Mayo Clinic, 200 First St, SW, Rochester, MN 55905.
0 1994 by The Society of Thoracic Surgeons
this improved with experience. There was one early death and two late deaths in the series. Internal thoracic arteriography was performed 7 to 10 days postoperatively in all 9 surviving patients. There was excellent perfusion of the donor bronchial arteries in 7 of these 9 patients. Bronchoscopy was performed when clinically indicated. No patient had early or late airway healing complications at a median follow-up of 13 months (range, 6 to 16 months). We conclude that direct, immediate bronchial artery revascularization is feasible on a routine basis for single-lung transplantation, and airway healing has been excellent. (Ann Thorac Surg 1994;57:1446-52)
Because BAR can be accomplished with a minimal increase in ischemic time and recipient operative time [15], we recently embarked on a program of routine, immediate BAR using the recipient internal thoracic artery (ITA) at the time of SLT. Our early experience has been encouraging, and we continue to employ this technique routinely for SLT.
Material and Methods Based on previously reported experience with BAR in patients undergoing double-lung transplantation [15], we recently began a trial of routine BAR for SLT. In 10 of 11 recent consecutive SLTs at the Mayo Clinic, BAR was attempted. In the patient who was excluded, the lung was harvested distantly by another team not familiar with the technique for bronchial artery preservation during organ procurement. The anatomy of the bronchial arteries has been well described [7, 11, 131, and various techniques for harvest of the double-lung block that preserve the bronchial arteries have been described previously [ll,13, 151. The bronchial arteries originate from the descending thoracic aorta. Usually, a right bronchial artery arises as a branch of the first or second intercostal artery. This right intercostobronchial artery passes posterior to the esophagus; the right bronchial artery then passes anteriorly between the esophagus and azygos vein. The left bronchial artery, which may be multiple, arises from the descending thoracic aorta anterior to the left intercostal branches and passes anteriorly directly to the tracheobronchial tree. Thus, the bronchial arteries pass on either side of the esophagus (Fig 1). Occasionally, both right and left bronchial arteries arise from a common trunk on the anterior 0003-4975/94/$7.00
Ann Thorac Surg 1994;571446-52
B A Fig 1. Bronchial artery anatomy. (A) Cross-sectional view with spine at the top. The left bronchial arteries usually arise anterior to the left intercostal arteries. The right bronchial artery usually arises as a branch of the first or second right intercostal artery, passing behind the esophagus. (A = aorta; E = esophagus; LB = left main bronchus; RB = right main bronchus.) ( B ) Demonstration of the right bronchial artery passing behind the esophagus; the intercostal vessels are not shown in this view.
aspect of the aorta. There is a network of collateral vessels in the mediastinal tissue around the carina.
Organ Procurement A median sternotomy is performed and the organs are examined. One hour before the estimated time of organ harvesting, the donor is given 1 g of intravenous methylprednisolone. Fifteen to 30 minutes before harvest, an intravenous infusion of prostaglandin E, is titrated to produce a 25% decrease in the systolic blood pressure. The nasogastric tube is withdrawn into the esophagus (confirmed by palpation through the left pleural space), into which 10 mL of povidone iodine solution is injected. After aortic cross-clamping, University of Wisconsin solution (60 mL/kg at 4°C) is administered directly into the pulmonary trunk while the lungs are gently ventilated with room air. The tip of the left atrial appendage is excised to decompress the left heart. Simultaneously, cardioplegia is administered into the aortic root. The heart can be excised before excision of the double-lung block, or it can be separated on the back table later. The nasogastric tube is removed, and ventilation is stopped for the dissection. Protection of the bronchial arteries during excision of the double-lung block requires en-bloc removal of the esophagus and descending thoracic aorta. The distal esophagus is mobilized bluntly and divided with a stapling device. The distal descending thoracic aorta is divided and the parietal pleural is incised, caudal to cranial, lateral to azygos vein on the right, and lateral to the aorta on the left. The aorta and esophagus are dissected off the anterior vertebral ligament with the double-
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
1447
lung block. The branches of the arch vessels are divided, and the cranial aspect of the esophagus is divided with a stapling device. The lungs are gently inflated to a normal tidal volume with room air, and the trachea is divided with the stapler. The double-lung block is transferred to a cold (4°C) saline solution bath. The esophagus is stripped subadventially from the mediastinal tissue. The proximal aortic arch is excised. The descending thoracic aorta is opened longitudinally along its pleural surface, just anterior to the left intercostal branches; this incision will pass between the left intercostal arteries and the left bronchial arteries. The bronchial artery orifices can be identified and confirmed by gently passing a probe through them to the tracheobronchial tree, and they may be marked with a suture [15]. We have stopped passing a probe and simply observe the orifices for backbleeding after the pulmonary circulation is reestablished in the recipient. Backbleeding is due to pulmonary-bronchial collateral vessels, and implies that the orifice on the donor aorta leads to a bronchial artery. The distal half of the descending thoracic aorta is removed from the organ block. To prevent injury to the bronchial arteries, the orifice is not dissected further from the opened, donor, descending thoracic aorta. The lungs are separated by dividing the left atrium and pulmonary arteries in the usual manner. The main bronchi are divided at the carina with a stapler. The remaining aortic patch is kept with the desired lung along with adjacent mediastinal tissue containing the bronchial arteries. The aortic patch can be divided in such a way that the bronchial arteries are preserved for each lung. A probe is used to identify the bronchial arteries, and the adjacent mediastinal tissue is divided without injuring the bronchial arteries. This maneuver requires additional time and will not be possible in every donor due to anatomic variability of the bronchial arteries.
Recipient Procedure The patient is placed in the lateral decubitus position. The ipsilateral groin is included in the field for cannulation should cardiopulmonary bypass be required. A standard posterolateral thoracotomy is performed. The ipsilateral ITA is dissected from its bed before anticipated arrival of the donor lung. Pneumonectomy and standard SLT are performed [ 181. The bronchial anastomosis is performed first with 4-0 polypropolene suture; the membranous portion is continuous, and the cartilagenous portion is interrupted. No specific attempt is made to “telescope” the anastomosis; the bronchial cartilages are allowed to overlap, or ”telescope,” if their respective sizes permit it while the interrupted sutures are tied. When the pulmonary circulation has been reestablished, the donor aortic cuff is inspected. Backbleeding will be noted from the orifice of the bronchial arteries due to pulmonary-bronchial collaterals. We choose the largest vessel with good backbleeding for revascularization, and avoid probing or further dissection of the bronchial arteries. Heparin, 5000 units intravenously, is administered at this time.
1448
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
The ITA is prepared by incising the distal aspect longitudinally for 0.5 to 1.0 cm and anastomosing it to the chosen orifice of the bronchial artery on the donor descending thoracic aorta with running 8-0 polypropylene suture. Bleeding from collateral branches in the adjacent mediastinal tissue and backbleeding from other orifices on the donor aorta are common. Hemostasis is carefully achieved with suture or clips, avoiding injury to adjacent vessels in the donor mediastinal tissue that may be important bronchial arteries or collaterals. Aspirin, 81 mg per day, is administered for 3 months postoperatively.
lrnrnunosuppression Standard triple immunosuppressive therapy is employed with OKT3 induction (14-day course) according to a previously reported algorithm [181. Intravenous methylprednisone is administered in the operating room (500 mg) and for three subsequent doses (125 mg) at 8-hour intervals. Steroids are then withheld for the remainder of the first 14 days, after which a standard course is administered (prednisone, 1 mg/kg/day tapered to 0.3 mglkglday maintenance by day 28 postoperatively). Azothioprine (4 mg/kg) is administered preoperatively, and administration is continued postoperatively, adjusting the dose to maintain a leukocyte count of 4,000 to 6,OOO/pL. Cyclosporin A administration is started during the first 14 days postoperatively, after renal function is stable. The dose is adjusted to achieve therapeutic levels by the end of the course of OKT3 (200 to 300 ng/mL serum by EIA); the serum levels are decreased to 75 to 150 ng/mL after 6 weeks.
Angiogruphy All surviving patients underwent angiography of the ITA 7 to 10 days after SLT. One of the 10 patients died perioperatively and an angiogram was not performed. The ITA was patent in all 9 patients studied.
Bronchoscopy Bronchoscopic examination of the anastomoses was performed in the operating room immediately after SLT and just before extubation (1 to 4 days postoperatively). Subsequent bronchoscopy was performed for clinical indications only: suspicion of rejection or infection, or unexplained changes on the chest roentgenogram.
Results We performed BAR as a routine part of SLT in 10 of 11 consecutive recipients; 1 was excluded because the donor operation was performed by another team. There were 6 male and 4 female recipients with a mean age of 52.6 years (range, 43 to 59 years). Eight patients had left SLT and 2 had right SLT. Recipient diagnoses were as follows: a,-antitrypsin deficiency, 4; emphysema, 2; obliterative bronchiolitis, 2; primary pulmonary hypertension, 1; and pulmonary fibrosis, 1. Median donor age was 36 years (mean, 28.9 years; range, 16 to 47 years). Median ischemic time was 250 minutes (mean, 236.2 minutes; range, 132 to 373 minutes). Ischemic time was initially prolonged by
Ann Thorac Surg 1994;57 1446552
about 15 to 20 minutes for BAR: 5 additional minutes for organ harvesting and 10 to 15 additional minutes to prepare the lung for implantation. This additional time was reduced by half as we gained experience with the technique. Recipient operative time (not additional ischemic time) was initially prolonged by 30 to 45 minutes (takedown of ITA, preparation for and performance of ITA to bronchial artery anastomosis, and additional hemostasis); this additional time also decreased with experience. Reperfusion injury occurred in 2 of the 9 surviving patients. Mean intubation time was 39 hours after SLT (range, 30 to 96 hours). Median hospital stay was 21 days (range, 15 to 28 days). There was one perioperative death in the series due to donor organ dysfunction. This patient underwent right SLT for emphysema; the ischemic time was 278 minutes. The same donor’s left lung was used in another recipient (“twin”), whose ischemic time was 373 minutes. Bronchial artery revascularization was performed in both patients by dividing the donor aorta and mediastinal tissue between two bronchial arteries. The surviving patient did well and had good bronchial artery perfusion by the ITA on angiography. Median follow-up of surviving patients was 13 months (range, 6 to 16 months). There have been two late deaths, both 9 months after SLT. One late death was due to posttransplantation B-cell lymphoma that did not respond to chemotherapy. The other late death was due to nonHodgkin’s lymphoma that was discovered in the early postoperative period and, we believe, predated the transplantation. Exercise capacity and pulmonary function in the survivors improved significantly with SLT, and have been stable during follow-up.
Angiogruphy Angiography of the ITA was performed in all surviving patients 7 to 10 days postoperatively. Excellent perfusion of the bronchial arteries was demonstrated in 7 of these 9 patients; there was good perfusion of the entire distal bronchial artery distribution as well as perfusion of the area of the bronchial anastomosis (Fig 2). The 2 patients who had poor perfusion of the bronchial arteries had patent ITAs, but no flow was seen beyond the ITA with the exception of a few small donor mediastinal vessels.
Bronchoscopy and Airway Healing All patients underwent bronchoscopy in the operating room and before extubation (postoperative day 1 to 4). Further bronchoscopic examinations were performed for clinical reasons only. Median time from operation to the last bronchoscopic examination was 7 months (range, 2 to 15 months). The distal bronchial mucosa of the transplanted lung looked pink and healthy at the early perioperative time and throughout the follow-up period in all patients. All patients have had excellent airway healing, with no evidence of granulation tissue, necrosis, or dehiscence. No late bronchial strictures or malacia have developed either at the anastomosis or in the distal airways. Three patients are more than 6 months from
Ann Thorac Surg 1994;57:144652
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
1449
A
C
D
Fig 2. Postoperative internal thoracic arteriograms performed 7 days after single-lung transplantation with direct, immediate bronchial artery revascularization. The recipient internal thoracic artery has been anastomosed to the origin of the bronchial artery on the donor descending thoracic aorta. Good perfusion of the bronchial arteries and the distal bronchial tree is demonstrated.
their last bronchoscopy and are doing well with stable pulmonary function tests.
Comment Impaired healing of the bronchial anastomosis has been a significant source of morbidity and mortality for patients
undergoing SLT. In the initial era after the first SLT in a human by Hardy and associates in 1963 [19] and extending through 1978, approximately 38 SLTs were performed. Of these, 12 patients survived more than 2 weeks, and 7 of these 12 died of bronchial leak [20]. The Toronto Lung Transplant Group substantially improved
1450
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
on this experience after a series of laboratory studies led them to optimize perioperative condition and nutrition, eliminate early postoperative steroids, and wrap the bronchial anastomosis with a vascularized pedicle of omentum [I, 211. Omentopexy provided delayed revascularization of the donor bronchial arteries [2], which occurred in as little as 4 days; however, ischemic changes have been observed even at this time [22]. Recently, one prospective, randomized trial found no difference in bronchial complications among 36 recipients of SLT randomized to omentopexy, ITA pedicle wrap, and no wrap [23]. Many centers have stopped using omentopexy as a routine component of SLT [5, 61. In the current era, complications of bronchial healing have been reported in 19% of patients, with 10% to 14% of these being serious complications: anastomotic dehiscence or late stricture requiring a stent [P6]. The causes of impaired airway healing after SLT may include ischemia, perioperative steroid use, preservation, rejection, and infection [I, 41. We believe that ischemia is the primary factor, particularly in the current era. The use of OKT3 has allowed us to eliminate steroids, after 24 hours, for the first 2 weeks after SLT, and the Washington University Lung Transplant Group noted no correlation between rejection and ischemic time or immunosuppressive regimen [6]. Recently, Schreinemakers and associates [24] have shown that healing of divided canine main bronchi is impaired by interrupting the bronchial arteries. The importance of restoring the systemic bronchial circulation for improved airway healing was recognized by Metras [9, 101, who implanted the bronchial artery orifice on a button of donor aorta into the descending thoracic aorta of dogs undergoing SLT in 1950. In 1973, Haglin and associates (251 revascularized the left bronchial artery in a patient undergoing sequential bilateral SLT by implanting the donor aortic button into the recipient descending thoracic aorta; bronchial healing on the left side was normal, whereas necrosis developed on the right. If BAR is to be applied routinely to SLT, several challenges must be met. First, the small size and variable anatomy of the bronchial arteries requires the use of special techniques for organ procurement (without compromising procurement of other organs) and the ability to identify the bronchial artery origins on the descending thoracic aorta. Second, the ischemic time and the recipient operative time must not be prolonged unduly. Third, an optimal technique for BAR must be selected; that is, a method that protects the small bronchial arteries and employs a proper conduit. The human bronchial circulation has proven to be predictable enough that BAR often can be accomplished [15,16]. The anatomic pattern of human bronchial arteries is reasonably predictable with a consistent left bronchial artery in 93.3% and a consistent right bronchial artery in 83.3% of cases [ ll] . We believed that we could identify a bronchial artery orifice in all cases in this series. Initially, we used a probe to identify the orifice of an intact bronchial artery; however, with experience, we relied on the presence of significant backbleeding from the vessel
Ann Thorac Surg 1994;571446-52
orifice after reestablishing the pulmonary circulation. Nevertheless, the variability of bronchial arteries is a limiting factor in universal application of this technique, and may have been a factor in the 2 cases of failed BAR in this series. We have attempted to preserve and separate the bronchial arteries to both lungs for implantation into separate recipients ("twinning") with BAR of both lungs. The variable anatomy will limit the performance of this for lungs from all donors, but it is possible occasionally. Routine BAR for SLT depends on performing the procedure in a timely manner. We found that our technique initially prolonged harvesting of the double-lung block by about 5 minutes; essentially the time needed to staple across the esophagus at its cranial and caudal ends. Our technique did not interfere with the interests of the other harvesting teams at the time of multiorgan donation. The ischemic time of the lungs will be minimally prolonged by the back table dissection of the double-lung block. Subadventitial stripping of the esophagus, opening and trimming of the donor descending thoracic aorta, inspection of the donor aorta for the bronchial artery origins, and division of the bronchi with preservation of adjacent mediastinal tissues all must be performed. This process initially required 10 to 15 minutes. Nothing additional is required until after the donor lung is implanted and the pulmonary circulation reestablished. Thus the total ischemic time was initially prolonged by about 15 to 20 minutes to accomplish BAR; with experience, this time was reduced by half. The recipient operation is prolonged to allow BAR. We usually mobilize the recipient ITA from its bed before the anticipated arrival of the donor lung. Preparation of the ITA, performing the ITA to bronchial artery anastomosis, and achieving hemostasis in the revascularized donor mediastinal tissue initially added 30 to 45 minutes to the recipient operation. This is not additional ischemic time. These estimates of prolonged ischemic and recipient operative time are generous and have improved as we have gained experience with the technique. Several techniques have been suggested for BAR at lung transplantation including direct implantation of a donor aortic button onto the recipient aorta [7, 9, 11, 221, use of a conduit of donor aorta [13, 141, use of a saphenous vein graft [16], and use of the ITA [15]. Direct implantation of the donor aortic button containing a bronchial artery orifice into the recipient aorta often requires dissection of the bronchial artery out of the donor mediastinal tissue, risking injury. Further, there is risk of morbidity from manipulating the recipient descending thoracic aorta. Conduits consisting of tailored donor aorta may be affected by stasis or embolus. The ITA is a proven conduit in coronary artery bypass grafting and has superior long-term patency even in the presence of poor run-off. Small bronchial vessels and vasoconstriction from ischemia or cold preservation may result in very slow run-off after BAR. In addition to prevention of bronchial ischemia, other advantages of BAR have been suggested. Mills and associates [7] found that pneumonia developed in 50% of
DALY AND MCGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
Ann Thorac Surg 1994;57144&52
transplanted lungs i n dogs that had bronchial mucosal ulceration, b u t i n none of those without mucosal ulceration. Aeba and colleagues [8] recently reported that interruption of t h e bronchial artery circulation contributes to increased severity of pneumonia in rats, with or witho u t l u n g transplantation. Mucociliary clearance has been shown to be compromised by interruption of the bronchial arteries and nerves [26], which m a y contribute to t h e development of pneumonia. Experimental induction of bronchial ischemia has been shown to result in transient impairment i n gas exchange [12] a n d pulmonary edema [27], suggesting a persistent bronchial ischemic compon e n t to t h e ”implantation response” after l u n g transplantation. W e have n o t noted an improvement in t h e incidence of reperfusion injury; it occurred i n 2 of 9 patients in this series. A final potential benefit of BAR with SLT is t h e possibility of favorably influencing t h e incidence of late OB. The cause of O B is likely multifactorial, a n d may be related to rejection, infection, ischemia, preservation, and other factors. Bronchial artery revascularization may have an impact on infection rate, as already noted [7, 81. It is apparent that BAR will reduce ischemia early after SLT; its late effect on bronchial oxygen saturation remains to be studied. The development of OB has been considered to represent chronic rejection [28], and may be related to repeated episodes of acute rejection [29]. Reduced mucosal blood flow has been reported t o occur with pulmonary rejection [30],and has been postulated as a mechanism for the role of rejection i n t h e development of late OB. It is conceivable that BAR could limit t h e degree of mucosal ischemia related to rejection and, thus, modify t h e contribution of rejection t o t h e development of late OB. Obliterative bronchiolitis remains a major factor limiting late functioning capacity i n a significant number of lung transplant patients. W e are hopeful about t h e potential for BAR to reduce t h e incidence of OB. In conclusion, we have demonstrated t h e feasibility of routine BAR for SLT. W e attempted BAR using t h e recipient ITA i n 10 consecutive patients undergoing SLT for whom we procured t h e organs. Excellent bronchial artery perfusion was confirmed i n 7 of 9 patients in whom angiography w a s performed. Neither organ ischemic time nor recipient operative time was u n d u l y prolonged. Bronchial healing was excellent with no short-term o r longterm airway complications i n a n y patient. W e now perform BAR a s a routine part of SLT. Variable bronchial artery anatomy will limit universal application of BAR for SLT, particularly i n some cases when both lungs are transplanted (although BAR of both lungs is feasible). In addition to improving airway healing, BAR may favorably affect posttransplantation implantation response, pneumonia, and late OB; however, the long-term benefits remain to be proven.
References 1. Cooper JD. The evolution of techniques and indications for lung transplantation. Ann Surg 1990;212:249-56. 2. Morgan E, Lima 0, Goldberg M, Ayabe H, Ferdman A,
Cooper JD. Improved bronchial healing in canine left lung
3. 4. 5. 6.
7. 8.
1451
reimplantation using omental pedicle wrap. J Thorac Cardiovasc Surg 1983;85:134-9. Fell SC, Mollenkopf FP, Montefusco CM, et al. Revascularization of ischemic bronchial anastomoses by an intercostal pedicle flap. J Thorac Cardiovasc Surg 1985;90:172-8. Schafers H-J, Haydock DA, Cooper JD. The prevalence and management of bronchial anastomotic complications in lung transplantation. J Thorac Cardiovasc Surg 1991;101:1044-52. Egan TM, Westerman JH, Lambert CJ, et al. Isolated lung transplantation for end-stage lung disease: a viable therapy. Ann Thorac Surg 1992;53:59M. Cooper JD, Patterson A, Trulock EP, Washington University Lung Transplant Group. Results of single and bilateral lung transplantation in 131 consecutive recipients. J Thorac Cardiovasc Surg 1994;107460-71. Mills NL, Boyd AD, Gheranpong C. The significance of bronchial circulation in lung transplantation. J Thorac Cardiovasc Surg 1970;60:866-74. Aeba R, Stout JE, Francalancia NA, et al. Aspects of lung transplantation that contribute to increased severity of pneumonia: an experimental study. J Thorac Cardiovasc Surg
1993;106:449-57. 9. Metras H. Note prCliminare sur la greffe totale du poumon chez le chien. Proc Acad Sci 1950;231:1176-7. 10. Metras D. Henri Metras: a pioneer in lung transplantation. J Heart Lung Transplant 1992;11:121?4. 11. Schreinemakers HHJ, Weder W, Miyoshi S, et al. Direct 12. 13. 14. 15.
revascularization of bronchial arteries for lung transplantation: an anatomical study. Ann Thorac Surg 1990;49:44-54. Pearson FG, Goldberg M, Stone RM, Colapinto RF. Bronchial arterial circulation restored after reimplantation of canine lung. Can J Surg 1970;13:243-50. Laks H, Louie HW, Haas GS, et al. New technique of vascularization of the trachea and bronchus for lung transplantation. J Heart Lung Transplant 1991;10:280-7. Raju S, Heath BJ, Warren ET, Hardy JD. Single- and doublelung transplantation. Ann Surg 1990;211:681-93. Daly RC, Tadjkarimi S, Khaghani A, Banner NR, Yacoub MH. Successful double-lung transplantation with direct bronchial artery revascularization. Ann Thorac Surg 1993;56:
885-92. 16 Couraud L, Baudet E, Martigne C, et al. Bronchial revascularization in double-lung transplantation: a series of 8 pa-. tients. Ann Thorac Surg 1992;53:8%94. 17. Patterson GA, Todd TR, Cooper JD, Pearson FG, Winton TL,
Maurer J. Airway complications after double lung transplantation. J Thorac Cardiovasc Surg 1990;99:14-21. 18. McGregor CGA, Daly RC, Peters SG, et al. Evolving strategies in lung transplantation for emphysema. Ann Thorac Surg 1994;57:1513-21. 19. Hardy JD, Webb WR, Dalton ML, Walker GR. Lung homotransplantation in man. Report of the initial case. JAMA
1963;186:1065-74. 20. Veith FJ. Lung transplantation. Surg Clin North Am 1978;58: 35744. 21. Toronto Lung Transplant Group. Unilateral lung transplantation for pulmonary fibrosis. N Engl J Med 1986;314:1140-5. 22. Morgan E, Lima 0, Goldberg M, Ferdman A, Luk SK, Cooper JD. Successful revascularization of totally ischemic bronchial autografts with omental pedicle flaps in dogs. J Thorac Cardiovasc Surg 1982;84:204-10. 23. Khaghani A, Tadjkarimi S, Daly R, et al. Influence of different types of wrap versus no wrap in single lung transplantation: a prospective, randomized trial [abstract no. 901. J Heart Lung Transplant 1992;11:213. 24. Schreinemakers HHJ, De Leyn PRJ, Marcus M, et al. Role of
the bronchial arteries in lung function. Reply. [Letter]. Ann Thorac Surg 1992;54:396-7. 25. Haglin JJ, Ruiz E, Baker RC, Anderson WR. Histologic studies of human lung allotransplantation. In: Wildevur C, ed. Morphology in lung transplantation. Basel, Switzerland: S. Karger, 1973:13-22. 26. Paul A, Marelli D, Shennib H, et al. Mucociliary function in
1452
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
ed. Morphology in lung transplantation. Basel, Switzerland: S. Karger, 1973:13-22. 26. Paul A, Marelli D, Shennib H, et al. Mucociliary function in autotransplanted, allotransplanted, and sleeve resected lungs. J Thorac Cardiovasc Surg 1989;98:523-8. 27. Ventemiglia RA, Braverman B, DiMauro J, et al. The ischemic lung: role of the bronchial arteries in lung function. Cardiovasc Dis Bull Tex Heart Inst 1981;8:48&98. 28. LoCicero J 111, Robinson PG, Fisher M. Chronic rejection in
Ann Thorac Surg 1994;57144&52
single-lung transplantation manifested by obliterative bronchiolitis. J Thorac Cardiovasc Surg 1990;99:1059-62. 29. Scott JP, Higenbottam TW, Clelland CA, et al. Natural history of chronic rejection in heart-lung transplantation. J Heart Transplant 1990;9:51&5. 30. Takao M, Katayama Y, Tanabe H, et al. Histologic changes in donor bronchi may explain the reduced mucosal blood flow seen during acute lung allograft rejection. J Heart Lung Transplant 1992;11:994-1000.
DISCUSSION DR THOMAS M. EGAN (Chapel Hill, NC): I thank Dr Daly and Dr McGregor for the opportunity of reviewing the manuscript well in advance. They are to be congratulated for an excellent description of a technique that may prove to be of value in isolated lung transplantation, but it is a technique that must be compared with other strategies. At the University of North Carolina in the last 3% years, we have performed 80 lung transplantations in 76 patients. Twentythree of these patients have had single-lung transplants and 53 patients have had double-lung transplants. Excluding early deaths in the first 3 weeks, there were 71 patients with 122 airway anastomoses that are able to be evaluated. Of these 71 patients, 67 patients underwent 115 airway anastomoses using omentopexy as described by the Toronto group initially. In these 67 patients we had one death that was related to airway dehiscence and two strictures that developed that required the insertion of endobronchial stents. There were three other contained airway dehiscences that went on to heal without any further sequelae. If we look at our results and compare them with the results reported by Dr Daly today, it is clear that we had 64 patients with acceptable airway healing compared with 9 of Dr Daly’s with internal mammary artery revascularization. Three of our patients had unacceptable airway healing, evidenced either by death or by requirement of a stent. As one can see from Table 1, if we were going to come up with a therapy that was going to be superior, we would need 85 consecutive patients with no failures to have a statistically significant difference by ,$ analysis. If we look at anastomoses instead of patients, we have 112 with an acceptable result and three that were clearly unacceptable compared with Dr Daly’s nine acceptable anastomoses, but to be significantly different statistically compared with current techniques, we would need 146 consecutive airway anastomoses with no failures.
Table 1. Comparison of University of North Carolina and Mayo Clinic Results Omentopexy (UNC)
Variable Patients Acceptable Not acceptable Anastomoses Acceptable Not acceptable a
p < 0.05 by
2 test
Direct Revascularization Better (Mayo) therapy?”
64 3
9
0
85 0
112 3
9 0
146 0
I would agree with the conclusions of Drs Daly and McGregor that they have in fact demonstrated the feasibility of this technique, but I question whether the technique is useful or better than other strategies. I have two questions for Dr Daly. How did the results in these 9 patients compare with results at your center before institution of your new protocol? How many patients had bronchoscopies in the 3 weeks after transplantation when bronchial dehiscence is most likely to be apparent bronchoscopically? DR G. ALEXANDER PATTERSON (St. Louis, MO): I congratulate Dr Daly as well. I think this is a sort of sequel to an article that he recently published describing a similar technique, revascularizing en-bloc double-lung grafts. Doctor Daly, the operative photographs you showed here demonstrated a large patch of aorta posterior to a left lung allograft, and you revascularized the left bronchial vessel. Is it possible to dissect two lung allografts from the same donor and revascularize both of those allografts in separate recipients? DR DALY I thank Dr Egan and Dr Patterson for their comments. Other than before extubation, only about half of these patients had bronchoscopy during the 3 weeks immediately after transplantation, because we were doing it only for clinical indications and at the time were not attempting to demonstrate, as you say, utility of the technique. I agree with you that that yet needs to be proven. In our previous experience we did have complications of airway healing, including the need for stent placement, and one of the things that enthuses me about this procedure is the potential for completely avoiding the need for stent placements in patients. It is frequently stated the stent placements are simple and solve the problem of stricture, but they seriously affect the patient’s quality of life; they require multiple bronchoscopic examinations and inhalation of nebulized saline solution several times per day. So I remain enthusiastic that bronchial artery revascularization will have utility, but I agree with you, it needs to be proven yet. In answering Dr Patterson’s question, in 1case we did perform single-lung transplantation in two separate patients from a single donor, that is, twins, as your group has called it, and we separated the bronchial arteries. We divided the identified separate orifices on the patch of the donor descending thoracic aorta and separated that patch and the arteries going to each side. It took us somewhere between 10 and 15 additional minutes to do that. One patient had successful bronchial artery revascularization by arteriogram, and the other patient was the one patient in the series who died perioperatively, and we did not get an autopsy so we do not know whether the bronchial artery was successfully revascularized.
Ischemia of the donor airway remains a significant cause of morbidity after single-lung transplantation; serious manifestations may occur early (anastomoticdehiscence) or late (stricture). Direct, immediate revascularization of the donor bronchial arteries, using the recipient internal thoracic artery, was performed in 10 consecutive recipients of single-lung transplants for whom we procured the organs. Mean recipient age was 52.6 years (range, 43 to 59 years); 6 were male and 4 female. Recipient diagnoses were emphysema (6), obliterative bronchiolitis (2), pulmonary fibrosis (l),and primary pulmonary hypertension (1). Bronchial artery revascularization initially prolonged the ischemic time by only 15 to 20 minutes;
T
he techniques and indications for single-lung transplantation (SLT) have evolved considerably over the last decade [l].However, the lung remains the only solid organ transplanted without a systemic arterial blood supply. Bronchial ischemia may result in early (granulation tissue, mucosal sloughing, or bronchial dehiscence) or late (stricture) complications of airway healing. Wrapping the bronchial anastomosis with vascularized tissue can provide a systemic blood supply to the distal airway [Z,31, but this takes time to develop and is provided only via small collaterals. In spite of bronchial omentopexy, early and late complications of airway ischemia continue to be a significant cause of morbidity after SLT, affecting approximately 10% to 19% of patients [GI.In addition, evidence is emerging of an association between bronchial ischemia and an increased incidence of perioperative pneumonia [7, 81. Bronchial ischemia may also play a role in the development of obliterative bronchiolitis (OB) after SLT. The concept of direct bronchial artery revascularization (BAR) was first suggested and demonstrated experimentally by Metras in 1950 [9, 101. Others have subsequently contributed to an understanding of the anatomy of the bronchial arteries, and to the technical goal of BAR [ 11-14]. Bronchial artery revascularization has been successfully used clinically for en-bloc double-lung transplantation with considerable improvement in tracheal anastomotic healing [15, 161 over previous reports involving a tracheal anastomosis [14, 171. Presented at the Thirtieth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 31-Feb 2, 1994. Address reprint requests to Dr Daly, Section of Cardiac Surgery, Mayo Clinic, 200 First St, SW, Rochester, MN 55905.
0 1994 by The Society of Thoracic Surgeons
this improved with experience. There was one early death and two late deaths in the series. Internal thoracic arteriography was performed 7 to 10 days postoperatively in all 9 surviving patients. There was excellent perfusion of the donor bronchial arteries in 7 of these 9 patients. Bronchoscopy was performed when clinically indicated. No patient had early or late airway healing complications at a median follow-up of 13 months (range, 6 to 16 months). We conclude that direct, immediate bronchial artery revascularization is feasible on a routine basis for single-lung transplantation, and airway healing has been excellent. (Ann Thorac Surg 1994;57:1446-52)
Because BAR can be accomplished with a minimal increase in ischemic time and recipient operative time [15], we recently embarked on a program of routine, immediate BAR using the recipient internal thoracic artery (ITA) at the time of SLT. Our early experience has been encouraging, and we continue to employ this technique routinely for SLT.
Material and Methods Based on previously reported experience with BAR in patients undergoing double-lung transplantation [15], we recently began a trial of routine BAR for SLT. In 10 of 11 recent consecutive SLTs at the Mayo Clinic, BAR was attempted. In the patient who was excluded, the lung was harvested distantly by another team not familiar with the technique for bronchial artery preservation during organ procurement. The anatomy of the bronchial arteries has been well described [7, 11, 131, and various techniques for harvest of the double-lung block that preserve the bronchial arteries have been described previously [ll,13, 151. The bronchial arteries originate from the descending thoracic aorta. Usually, a right bronchial artery arises as a branch of the first or second intercostal artery. This right intercostobronchial artery passes posterior to the esophagus; the right bronchial artery then passes anteriorly between the esophagus and azygos vein. The left bronchial artery, which may be multiple, arises from the descending thoracic aorta anterior to the left intercostal branches and passes anteriorly directly to the tracheobronchial tree. Thus, the bronchial arteries pass on either side of the esophagus (Fig 1). Occasionally, both right and left bronchial arteries arise from a common trunk on the anterior 0003-4975/94/$7.00
Ann Thorac Surg 1994;571446-52
B A Fig 1. Bronchial artery anatomy. (A) Cross-sectional view with spine at the top. The left bronchial arteries usually arise anterior to the left intercostal arteries. The right bronchial artery usually arises as a branch of the first or second right intercostal artery, passing behind the esophagus. (A = aorta; E = esophagus; LB = left main bronchus; RB = right main bronchus.) ( B ) Demonstration of the right bronchial artery passing behind the esophagus; the intercostal vessels are not shown in this view.
aspect of the aorta. There is a network of collateral vessels in the mediastinal tissue around the carina.
Organ Procurement A median sternotomy is performed and the organs are examined. One hour before the estimated time of organ harvesting, the donor is given 1 g of intravenous methylprednisolone. Fifteen to 30 minutes before harvest, an intravenous infusion of prostaglandin E, is titrated to produce a 25% decrease in the systolic blood pressure. The nasogastric tube is withdrawn into the esophagus (confirmed by palpation through the left pleural space), into which 10 mL of povidone iodine solution is injected. After aortic cross-clamping, University of Wisconsin solution (60 mL/kg at 4°C) is administered directly into the pulmonary trunk while the lungs are gently ventilated with room air. The tip of the left atrial appendage is excised to decompress the left heart. Simultaneously, cardioplegia is administered into the aortic root. The heart can be excised before excision of the double-lung block, or it can be separated on the back table later. The nasogastric tube is removed, and ventilation is stopped for the dissection. Protection of the bronchial arteries during excision of the double-lung block requires en-bloc removal of the esophagus and descending thoracic aorta. The distal esophagus is mobilized bluntly and divided with a stapling device. The distal descending thoracic aorta is divided and the parietal pleural is incised, caudal to cranial, lateral to azygos vein on the right, and lateral to the aorta on the left. The aorta and esophagus are dissected off the anterior vertebral ligament with the double-
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
1447
lung block. The branches of the arch vessels are divided, and the cranial aspect of the esophagus is divided with a stapling device. The lungs are gently inflated to a normal tidal volume with room air, and the trachea is divided with the stapler. The double-lung block is transferred to a cold (4°C) saline solution bath. The esophagus is stripped subadventially from the mediastinal tissue. The proximal aortic arch is excised. The descending thoracic aorta is opened longitudinally along its pleural surface, just anterior to the left intercostal branches; this incision will pass between the left intercostal arteries and the left bronchial arteries. The bronchial artery orifices can be identified and confirmed by gently passing a probe through them to the tracheobronchial tree, and they may be marked with a suture [15]. We have stopped passing a probe and simply observe the orifices for backbleeding after the pulmonary circulation is reestablished in the recipient. Backbleeding is due to pulmonary-bronchial collateral vessels, and implies that the orifice on the donor aorta leads to a bronchial artery. The distal half of the descending thoracic aorta is removed from the organ block. To prevent injury to the bronchial arteries, the orifice is not dissected further from the opened, donor, descending thoracic aorta. The lungs are separated by dividing the left atrium and pulmonary arteries in the usual manner. The main bronchi are divided at the carina with a stapler. The remaining aortic patch is kept with the desired lung along with adjacent mediastinal tissue containing the bronchial arteries. The aortic patch can be divided in such a way that the bronchial arteries are preserved for each lung. A probe is used to identify the bronchial arteries, and the adjacent mediastinal tissue is divided without injuring the bronchial arteries. This maneuver requires additional time and will not be possible in every donor due to anatomic variability of the bronchial arteries.
Recipient Procedure The patient is placed in the lateral decubitus position. The ipsilateral groin is included in the field for cannulation should cardiopulmonary bypass be required. A standard posterolateral thoracotomy is performed. The ipsilateral ITA is dissected from its bed before anticipated arrival of the donor lung. Pneumonectomy and standard SLT are performed [ 181. The bronchial anastomosis is performed first with 4-0 polypropolene suture; the membranous portion is continuous, and the cartilagenous portion is interrupted. No specific attempt is made to “telescope” the anastomosis; the bronchial cartilages are allowed to overlap, or ”telescope,” if their respective sizes permit it while the interrupted sutures are tied. When the pulmonary circulation has been reestablished, the donor aortic cuff is inspected. Backbleeding will be noted from the orifice of the bronchial arteries due to pulmonary-bronchial collaterals. We choose the largest vessel with good backbleeding for revascularization, and avoid probing or further dissection of the bronchial arteries. Heparin, 5000 units intravenously, is administered at this time.
1448
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
The ITA is prepared by incising the distal aspect longitudinally for 0.5 to 1.0 cm and anastomosing it to the chosen orifice of the bronchial artery on the donor descending thoracic aorta with running 8-0 polypropylene suture. Bleeding from collateral branches in the adjacent mediastinal tissue and backbleeding from other orifices on the donor aorta are common. Hemostasis is carefully achieved with suture or clips, avoiding injury to adjacent vessels in the donor mediastinal tissue that may be important bronchial arteries or collaterals. Aspirin, 81 mg per day, is administered for 3 months postoperatively.
lrnrnunosuppression Standard triple immunosuppressive therapy is employed with OKT3 induction (14-day course) according to a previously reported algorithm [181. Intravenous methylprednisone is administered in the operating room (500 mg) and for three subsequent doses (125 mg) at 8-hour intervals. Steroids are then withheld for the remainder of the first 14 days, after which a standard course is administered (prednisone, 1 mg/kg/day tapered to 0.3 mglkglday maintenance by day 28 postoperatively). Azothioprine (4 mg/kg) is administered preoperatively, and administration is continued postoperatively, adjusting the dose to maintain a leukocyte count of 4,000 to 6,OOO/pL. Cyclosporin A administration is started during the first 14 days postoperatively, after renal function is stable. The dose is adjusted to achieve therapeutic levels by the end of the course of OKT3 (200 to 300 ng/mL serum by EIA); the serum levels are decreased to 75 to 150 ng/mL after 6 weeks.
Angiogruphy All surviving patients underwent angiography of the ITA 7 to 10 days after SLT. One of the 10 patients died perioperatively and an angiogram was not performed. The ITA was patent in all 9 patients studied.
Bronchoscopy Bronchoscopic examination of the anastomoses was performed in the operating room immediately after SLT and just before extubation (1 to 4 days postoperatively). Subsequent bronchoscopy was performed for clinical indications only: suspicion of rejection or infection, or unexplained changes on the chest roentgenogram.
Results We performed BAR as a routine part of SLT in 10 of 11 consecutive recipients; 1 was excluded because the donor operation was performed by another team. There were 6 male and 4 female recipients with a mean age of 52.6 years (range, 43 to 59 years). Eight patients had left SLT and 2 had right SLT. Recipient diagnoses were as follows: a,-antitrypsin deficiency, 4; emphysema, 2; obliterative bronchiolitis, 2; primary pulmonary hypertension, 1; and pulmonary fibrosis, 1. Median donor age was 36 years (mean, 28.9 years; range, 16 to 47 years). Median ischemic time was 250 minutes (mean, 236.2 minutes; range, 132 to 373 minutes). Ischemic time was initially prolonged by
Ann Thorac Surg 1994;57 1446552
about 15 to 20 minutes for BAR: 5 additional minutes for organ harvesting and 10 to 15 additional minutes to prepare the lung for implantation. This additional time was reduced by half as we gained experience with the technique. Recipient operative time (not additional ischemic time) was initially prolonged by 30 to 45 minutes (takedown of ITA, preparation for and performance of ITA to bronchial artery anastomosis, and additional hemostasis); this additional time also decreased with experience. Reperfusion injury occurred in 2 of the 9 surviving patients. Mean intubation time was 39 hours after SLT (range, 30 to 96 hours). Median hospital stay was 21 days (range, 15 to 28 days). There was one perioperative death in the series due to donor organ dysfunction. This patient underwent right SLT for emphysema; the ischemic time was 278 minutes. The same donor’s left lung was used in another recipient (“twin”), whose ischemic time was 373 minutes. Bronchial artery revascularization was performed in both patients by dividing the donor aorta and mediastinal tissue between two bronchial arteries. The surviving patient did well and had good bronchial artery perfusion by the ITA on angiography. Median follow-up of surviving patients was 13 months (range, 6 to 16 months). There have been two late deaths, both 9 months after SLT. One late death was due to posttransplantation B-cell lymphoma that did not respond to chemotherapy. The other late death was due to nonHodgkin’s lymphoma that was discovered in the early postoperative period and, we believe, predated the transplantation. Exercise capacity and pulmonary function in the survivors improved significantly with SLT, and have been stable during follow-up.
Angiogruphy Angiography of the ITA was performed in all surviving patients 7 to 10 days postoperatively. Excellent perfusion of the bronchial arteries was demonstrated in 7 of these 9 patients; there was good perfusion of the entire distal bronchial artery distribution as well as perfusion of the area of the bronchial anastomosis (Fig 2). The 2 patients who had poor perfusion of the bronchial arteries had patent ITAs, but no flow was seen beyond the ITA with the exception of a few small donor mediastinal vessels.
Bronchoscopy and Airway Healing All patients underwent bronchoscopy in the operating room and before extubation (postoperative day 1 to 4). Further bronchoscopic examinations were performed for clinical reasons only. Median time from operation to the last bronchoscopic examination was 7 months (range, 2 to 15 months). The distal bronchial mucosa of the transplanted lung looked pink and healthy at the early perioperative time and throughout the follow-up period in all patients. All patients have had excellent airway healing, with no evidence of granulation tissue, necrosis, or dehiscence. No late bronchial strictures or malacia have developed either at the anastomosis or in the distal airways. Three patients are more than 6 months from
Ann Thorac Surg 1994;57:144652
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
1449
A
C
D
Fig 2. Postoperative internal thoracic arteriograms performed 7 days after single-lung transplantation with direct, immediate bronchial artery revascularization. The recipient internal thoracic artery has been anastomosed to the origin of the bronchial artery on the donor descending thoracic aorta. Good perfusion of the bronchial arteries and the distal bronchial tree is demonstrated.
their last bronchoscopy and are doing well with stable pulmonary function tests.
Comment Impaired healing of the bronchial anastomosis has been a significant source of morbidity and mortality for patients
undergoing SLT. In the initial era after the first SLT in a human by Hardy and associates in 1963 [19] and extending through 1978, approximately 38 SLTs were performed. Of these, 12 patients survived more than 2 weeks, and 7 of these 12 died of bronchial leak [20]. The Toronto Lung Transplant Group substantially improved
1450
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
on this experience after a series of laboratory studies led them to optimize perioperative condition and nutrition, eliminate early postoperative steroids, and wrap the bronchial anastomosis with a vascularized pedicle of omentum [I, 211. Omentopexy provided delayed revascularization of the donor bronchial arteries [2], which occurred in as little as 4 days; however, ischemic changes have been observed even at this time [22]. Recently, one prospective, randomized trial found no difference in bronchial complications among 36 recipients of SLT randomized to omentopexy, ITA pedicle wrap, and no wrap [23]. Many centers have stopped using omentopexy as a routine component of SLT [5, 61. In the current era, complications of bronchial healing have been reported in 19% of patients, with 10% to 14% of these being serious complications: anastomotic dehiscence or late stricture requiring a stent [P6]. The causes of impaired airway healing after SLT may include ischemia, perioperative steroid use, preservation, rejection, and infection [I, 41. We believe that ischemia is the primary factor, particularly in the current era. The use of OKT3 has allowed us to eliminate steroids, after 24 hours, for the first 2 weeks after SLT, and the Washington University Lung Transplant Group noted no correlation between rejection and ischemic time or immunosuppressive regimen [6]. Recently, Schreinemakers and associates [24] have shown that healing of divided canine main bronchi is impaired by interrupting the bronchial arteries. The importance of restoring the systemic bronchial circulation for improved airway healing was recognized by Metras [9, 101, who implanted the bronchial artery orifice on a button of donor aorta into the descending thoracic aorta of dogs undergoing SLT in 1950. In 1973, Haglin and associates (251 revascularized the left bronchial artery in a patient undergoing sequential bilateral SLT by implanting the donor aortic button into the recipient descending thoracic aorta; bronchial healing on the left side was normal, whereas necrosis developed on the right. If BAR is to be applied routinely to SLT, several challenges must be met. First, the small size and variable anatomy of the bronchial arteries requires the use of special techniques for organ procurement (without compromising procurement of other organs) and the ability to identify the bronchial artery origins on the descending thoracic aorta. Second, the ischemic time and the recipient operative time must not be prolonged unduly. Third, an optimal technique for BAR must be selected; that is, a method that protects the small bronchial arteries and employs a proper conduit. The human bronchial circulation has proven to be predictable enough that BAR often can be accomplished [15,16]. The anatomic pattern of human bronchial arteries is reasonably predictable with a consistent left bronchial artery in 93.3% and a consistent right bronchial artery in 83.3% of cases [ ll] . We believed that we could identify a bronchial artery orifice in all cases in this series. Initially, we used a probe to identify the orifice of an intact bronchial artery; however, with experience, we relied on the presence of significant backbleeding from the vessel
Ann Thorac Surg 1994;571446-52
orifice after reestablishing the pulmonary circulation. Nevertheless, the variability of bronchial arteries is a limiting factor in universal application of this technique, and may have been a factor in the 2 cases of failed BAR in this series. We have attempted to preserve and separate the bronchial arteries to both lungs for implantation into separate recipients ("twinning") with BAR of both lungs. The variable anatomy will limit the performance of this for lungs from all donors, but it is possible occasionally. Routine BAR for SLT depends on performing the procedure in a timely manner. We found that our technique initially prolonged harvesting of the double-lung block by about 5 minutes; essentially the time needed to staple across the esophagus at its cranial and caudal ends. Our technique did not interfere with the interests of the other harvesting teams at the time of multiorgan donation. The ischemic time of the lungs will be minimally prolonged by the back table dissection of the double-lung block. Subadventitial stripping of the esophagus, opening and trimming of the donor descending thoracic aorta, inspection of the donor aorta for the bronchial artery origins, and division of the bronchi with preservation of adjacent mediastinal tissues all must be performed. This process initially required 10 to 15 minutes. Nothing additional is required until after the donor lung is implanted and the pulmonary circulation reestablished. Thus the total ischemic time was initially prolonged by about 15 to 20 minutes to accomplish BAR; with experience, this time was reduced by half. The recipient operation is prolonged to allow BAR. We usually mobilize the recipient ITA from its bed before the anticipated arrival of the donor lung. Preparation of the ITA, performing the ITA to bronchial artery anastomosis, and achieving hemostasis in the revascularized donor mediastinal tissue initially added 30 to 45 minutes to the recipient operation. This is not additional ischemic time. These estimates of prolonged ischemic and recipient operative time are generous and have improved as we have gained experience with the technique. Several techniques have been suggested for BAR at lung transplantation including direct implantation of a donor aortic button onto the recipient aorta [7, 9, 11, 221, use of a conduit of donor aorta [13, 141, use of a saphenous vein graft [16], and use of the ITA [15]. Direct implantation of the donor aortic button containing a bronchial artery orifice into the recipient aorta often requires dissection of the bronchial artery out of the donor mediastinal tissue, risking injury. Further, there is risk of morbidity from manipulating the recipient descending thoracic aorta. Conduits consisting of tailored donor aorta may be affected by stasis or embolus. The ITA is a proven conduit in coronary artery bypass grafting and has superior long-term patency even in the presence of poor run-off. Small bronchial vessels and vasoconstriction from ischemia or cold preservation may result in very slow run-off after BAR. In addition to prevention of bronchial ischemia, other advantages of BAR have been suggested. Mills and associates [7] found that pneumonia developed in 50% of
DALY AND MCGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
Ann Thorac Surg 1994;57144&52
transplanted lungs i n dogs that had bronchial mucosal ulceration, b u t i n none of those without mucosal ulceration. Aeba and colleagues [8] recently reported that interruption of t h e bronchial artery circulation contributes to increased severity of pneumonia in rats, with or witho u t l u n g transplantation. Mucociliary clearance has been shown to be compromised by interruption of the bronchial arteries and nerves [26], which m a y contribute to t h e development of pneumonia. Experimental induction of bronchial ischemia has been shown to result in transient impairment i n gas exchange [12] a n d pulmonary edema [27], suggesting a persistent bronchial ischemic compon e n t to t h e ”implantation response” after l u n g transplantation. W e have n o t noted an improvement in t h e incidence of reperfusion injury; it occurred i n 2 of 9 patients in this series. A final potential benefit of BAR with SLT is t h e possibility of favorably influencing t h e incidence of late OB. The cause of O B is likely multifactorial, a n d may be related to rejection, infection, ischemia, preservation, and other factors. Bronchial artery revascularization may have an impact on infection rate, as already noted [7, 81. It is apparent that BAR will reduce ischemia early after SLT; its late effect on bronchial oxygen saturation remains to be studied. The development of OB has been considered to represent chronic rejection [28], and may be related to repeated episodes of acute rejection [29]. Reduced mucosal blood flow has been reported t o occur with pulmonary rejection [30],and has been postulated as a mechanism for the role of rejection i n t h e development of late OB. It is conceivable that BAR could limit t h e degree of mucosal ischemia related to rejection and, thus, modify t h e contribution of rejection t o t h e development of late OB. Obliterative bronchiolitis remains a major factor limiting late functioning capacity i n a significant number of lung transplant patients. W e are hopeful about t h e potential for BAR to reduce t h e incidence of OB. In conclusion, we have demonstrated t h e feasibility of routine BAR for SLT. W e attempted BAR using t h e recipient ITA i n 10 consecutive patients undergoing SLT for whom we procured t h e organs. Excellent bronchial artery perfusion was confirmed i n 7 of 9 patients in whom angiography w a s performed. Neither organ ischemic time nor recipient operative time was u n d u l y prolonged. Bronchial healing was excellent with no short-term o r longterm airway complications i n a n y patient. W e now perform BAR a s a routine part of SLT. Variable bronchial artery anatomy will limit universal application of BAR for SLT, particularly i n some cases when both lungs are transplanted (although BAR of both lungs is feasible). In addition to improving airway healing, BAR may favorably affect posttransplantation implantation response, pneumonia, and late OB; however, the long-term benefits remain to be proven.
References 1. Cooper JD. The evolution of techniques and indications for lung transplantation. Ann Surg 1990;212:249-56. 2. Morgan E, Lima 0, Goldberg M, Ayabe H, Ferdman A,
Cooper JD. Improved bronchial healing in canine left lung
3. 4. 5. 6.
7. 8.
1451
reimplantation using omental pedicle wrap. J Thorac Cardiovasc Surg 1983;85:134-9. Fell SC, Mollenkopf FP, Montefusco CM, et al. Revascularization of ischemic bronchial anastomoses by an intercostal pedicle flap. J Thorac Cardiovasc Surg 1985;90:172-8. Schafers H-J, Haydock DA, Cooper JD. The prevalence and management of bronchial anastomotic complications in lung transplantation. J Thorac Cardiovasc Surg 1991;101:1044-52. Egan TM, Westerman JH, Lambert CJ, et al. Isolated lung transplantation for end-stage lung disease: a viable therapy. Ann Thorac Surg 1992;53:59M. Cooper JD, Patterson A, Trulock EP, Washington University Lung Transplant Group. Results of single and bilateral lung transplantation in 131 consecutive recipients. J Thorac Cardiovasc Surg 1994;107460-71. Mills NL, Boyd AD, Gheranpong C. The significance of bronchial circulation in lung transplantation. J Thorac Cardiovasc Surg 1970;60:866-74. Aeba R, Stout JE, Francalancia NA, et al. Aspects of lung transplantation that contribute to increased severity of pneumonia: an experimental study. J Thorac Cardiovasc Surg
1993;106:449-57. 9. Metras H. Note prCliminare sur la greffe totale du poumon chez le chien. Proc Acad Sci 1950;231:1176-7. 10. Metras D. Henri Metras: a pioneer in lung transplantation. J Heart Lung Transplant 1992;11:121?4. 11. Schreinemakers HHJ, Weder W, Miyoshi S, et al. Direct 12. 13. 14. 15.
revascularization of bronchial arteries for lung transplantation: an anatomical study. Ann Thorac Surg 1990;49:44-54. Pearson FG, Goldberg M, Stone RM, Colapinto RF. Bronchial arterial circulation restored after reimplantation of canine lung. Can J Surg 1970;13:243-50. Laks H, Louie HW, Haas GS, et al. New technique of vascularization of the trachea and bronchus for lung transplantation. J Heart Lung Transplant 1991;10:280-7. Raju S, Heath BJ, Warren ET, Hardy JD. Single- and doublelung transplantation. Ann Surg 1990;211:681-93. Daly RC, Tadjkarimi S, Khaghani A, Banner NR, Yacoub MH. Successful double-lung transplantation with direct bronchial artery revascularization. Ann Thorac Surg 1993;56:
885-92. 16 Couraud L, Baudet E, Martigne C, et al. Bronchial revascularization in double-lung transplantation: a series of 8 pa-. tients. Ann Thorac Surg 1992;53:8%94. 17. Patterson GA, Todd TR, Cooper JD, Pearson FG, Winton TL,
Maurer J. Airway complications after double lung transplantation. J Thorac Cardiovasc Surg 1990;99:14-21. 18. McGregor CGA, Daly RC, Peters SG, et al. Evolving strategies in lung transplantation for emphysema. Ann Thorac Surg 1994;57:1513-21. 19. Hardy JD, Webb WR, Dalton ML, Walker GR. Lung homotransplantation in man. Report of the initial case. JAMA
1963;186:1065-74. 20. Veith FJ. Lung transplantation. Surg Clin North Am 1978;58: 35744. 21. Toronto Lung Transplant Group. Unilateral lung transplantation for pulmonary fibrosis. N Engl J Med 1986;314:1140-5. 22. Morgan E, Lima 0, Goldberg M, Ferdman A, Luk SK, Cooper JD. Successful revascularization of totally ischemic bronchial autografts with omental pedicle flaps in dogs. J Thorac Cardiovasc Surg 1982;84:204-10. 23. Khaghani A, Tadjkarimi S, Daly R, et al. Influence of different types of wrap versus no wrap in single lung transplantation: a prospective, randomized trial [abstract no. 901. J Heart Lung Transplant 1992;11:213. 24. Schreinemakers HHJ, De Leyn PRJ, Marcus M, et al. Role of
the bronchial arteries in lung function. Reply. [Letter]. Ann Thorac Surg 1992;54:396-7. 25. Haglin JJ, Ruiz E, Baker RC, Anderson WR. Histologic studies of human lung allotransplantation. In: Wildevur C, ed. Morphology in lung transplantation. Basel, Switzerland: S. Karger, 1973:13-22. 26. Paul A, Marelli D, Shennib H, et al. Mucociliary function in
1452
DALY AND McGREGOR BRONCHIAL ARTERY REVASCULARIZATION FOR SLT
ed. Morphology in lung transplantation. Basel, Switzerland: S. Karger, 1973:13-22. 26. Paul A, Marelli D, Shennib H, et al. Mucociliary function in autotransplanted, allotransplanted, and sleeve resected lungs. J Thorac Cardiovasc Surg 1989;98:523-8. 27. Ventemiglia RA, Braverman B, DiMauro J, et al. The ischemic lung: role of the bronchial arteries in lung function. Cardiovasc Dis Bull Tex Heart Inst 1981;8:48&98. 28. LoCicero J 111, Robinson PG, Fisher M. Chronic rejection in
Ann Thorac Surg 1994;57144&52
single-lung transplantation manifested by obliterative bronchiolitis. J Thorac Cardiovasc Surg 1990;99:1059-62. 29. Scott JP, Higenbottam TW, Clelland CA, et al. Natural history of chronic rejection in heart-lung transplantation. J Heart Transplant 1990;9:51&5. 30. Takao M, Katayama Y, Tanabe H, et al. Histologic changes in donor bronchi may explain the reduced mucosal blood flow seen during acute lung allograft rejection. J Heart Lung Transplant 1992;11:994-1000.
DISCUSSION DR THOMAS M. EGAN (Chapel Hill, NC): I thank Dr Daly and Dr McGregor for the opportunity of reviewing the manuscript well in advance. They are to be congratulated for an excellent description of a technique that may prove to be of value in isolated lung transplantation, but it is a technique that must be compared with other strategies. At the University of North Carolina in the last 3% years, we have performed 80 lung transplantations in 76 patients. Twentythree of these patients have had single-lung transplants and 53 patients have had double-lung transplants. Excluding early deaths in the first 3 weeks, there were 71 patients with 122 airway anastomoses that are able to be evaluated. Of these 71 patients, 67 patients underwent 115 airway anastomoses using omentopexy as described by the Toronto group initially. In these 67 patients we had one death that was related to airway dehiscence and two strictures that developed that required the insertion of endobronchial stents. There were three other contained airway dehiscences that went on to heal without any further sequelae. If we look at our results and compare them with the results reported by Dr Daly today, it is clear that we had 64 patients with acceptable airway healing compared with 9 of Dr Daly’s with internal mammary artery revascularization. Three of our patients had unacceptable airway healing, evidenced either by death or by requirement of a stent. As one can see from Table 1, if we were going to come up with a therapy that was going to be superior, we would need 85 consecutive patients with no failures to have a statistically significant difference by ,$ analysis. If we look at anastomoses instead of patients, we have 112 with an acceptable result and three that were clearly unacceptable compared with Dr Daly’s nine acceptable anastomoses, but to be significantly different statistically compared with current techniques, we would need 146 consecutive airway anastomoses with no failures.
Table 1. Comparison of University of North Carolina and Mayo Clinic Results Omentopexy (UNC)
Variable Patients Acceptable Not acceptable Anastomoses Acceptable Not acceptable a
p < 0.05 by
2 test
Direct Revascularization Better (Mayo) therapy?”
64 3
9
0
85 0
112 3
9 0
146 0
I would agree with the conclusions of Drs Daly and McGregor that they have in fact demonstrated the feasibility of this technique, but I question whether the technique is useful or better than other strategies. I have two questions for Dr Daly. How did the results in these 9 patients compare with results at your center before institution of your new protocol? How many patients had bronchoscopies in the 3 weeks after transplantation when bronchial dehiscence is most likely to be apparent bronchoscopically? DR G. ALEXANDER PATTERSON (St. Louis, MO): I congratulate Dr Daly as well. I think this is a sort of sequel to an article that he recently published describing a similar technique, revascularizing en-bloc double-lung grafts. Doctor Daly, the operative photographs you showed here demonstrated a large patch of aorta posterior to a left lung allograft, and you revascularized the left bronchial vessel. Is it possible to dissect two lung allografts from the same donor and revascularize both of those allografts in separate recipients? DR DALY I thank Dr Egan and Dr Patterson for their comments. Other than before extubation, only about half of these patients had bronchoscopy during the 3 weeks immediately after transplantation, because we were doing it only for clinical indications and at the time were not attempting to demonstrate, as you say, utility of the technique. I agree with you that that yet needs to be proven. In our previous experience we did have complications of airway healing, including the need for stent placement, and one of the things that enthuses me about this procedure is the potential for completely avoiding the need for stent placements in patients. It is frequently stated the stent placements are simple and solve the problem of stricture, but they seriously affect the patient’s quality of life; they require multiple bronchoscopic examinations and inhalation of nebulized saline solution several times per day. So I remain enthusiastic that bronchial artery revascularization will have utility, but I agree with you, it needs to be proven yet. In answering Dr Patterson’s question, in 1case we did perform single-lung transplantation in two separate patients from a single donor, that is, twins, as your group has called it, and we separated the bronchial arteries. We divided the identified separate orifices on the patch of the donor descending thoracic aorta and separated that patch and the arteries going to each side. It took us somewhere between 10 and 15 additional minutes to do that. One patient had successful bronchial artery revascularization by arteriogram, and the other patient was the one patient in the series who died perioperatively, and we did not get an autopsy so we do not know whether the bronchial artery was successfully revascularized.