Effect of Omental, Intercostal, and Internal Mammary Artery Pedicle Wraps on Bronchial Healing Mark W. Turrentine, MD, Kenneth A. Kesler, MD, Cameron D. Wright, MD, Keith E. McEwen, MD, Philip R. Faught, MD, Michael E. Miller, PhD, Yousuf Mahomed, MD, Harold King, MD, and John W. Brown, MD Departments of Surgery (Cardiothoracic Division), Pathology, and Biostatistics, Indiana University School of Medicine, Indianapolis, Indiana

Bronchial transection and devascularization is necessary in the course of sleeve resection or lung transplantation, leaving distal bronchial segments ischemic and subject to stricture or dehiscence. Thirty mongrel dogs under- went left lung autotransplantation. The bronchial anas- tomosis was wrapped with omentum (n = 9), intercostal muscle pedicle (n = 9), or internal mammary artery pedicle grafts (n = 6). Six control animals underwent bronchial anastomosis without an external wrap. Bron- chial revascularization by capillary ingrowth from the pedicle to the bronchial submucosal plexus was demon- strated with all three types of vascular pedicle grafts; however, more consistent and confluent vascular in- growth was provided by internal mammary artery pedi- cle grafts. Additionally, the bronchial anastomotic cross-

he bronchial circulation accepts only 1% of the total T cardiac output. It is, however, critical in maintaining the integrity of the bronchial wall and mucosa [l]. Radical bronchial dissection with bronchial arterial disruption is often necessary in the course of sleeve resections, repair of traumatic tracheobronchial injuries, and lung trans- plantations. This leaves the distal bronchial segment relatively ischemic, thus subject to mucosal necrosis, luminal stricture, or even frank dehiscence. Spontaneous revascularization of the distal bronchial circulation has been shown to occur between 15 and 30 postoperative days [2-51. It is generally believed, however, that earlier revascularization will promote uncomplicated bronchial healing [5].

Various techniques have been described to protect bronchial suture lines and anastomoses. These include pleural or pericardial flaps [6, 71, telescoping of the distal bronchial end into the proximal segment [8], and transec- tion of the bronchus close to the pulmonary hilus [9]. These techniques, although mechanically supporting bronchial suture lines, have not been found to provide early revascularization to ischemic tissue. In contrast, vascularized pedicle grafts have been shown to provide a

Presented at the Twenty-fifth Anniversary Meeting of The Society of Thoracic Surgeons, Baltimore, MD, Sep 11-13, 1989.

Address reprint requests to Dr Kesler, Department of Surgery, Cardiotho- rack Division, 545 Barnhill Dr, EM #212, Indianapolis, IN 46202-5124.

sectional area was significantly better in the internal mammary artery group (84.5 f 3.3) as compared with that of the omental (68.4 f 8.31, intercostal muscle (66.9 f 10.9), or control groups (70.2 f 7.6). An internal mam- mary artery pedicle graft and the presence of dense confluent submucosal vascular ingrowth from any pedi- cle graft were independently predictive ( p < 0.05) of minimizing bronchial anastomotic narrowing. These data are consistent with previous findings suggesting that omental and intercostal muscle pedicle grafts pro- mote early bronchial revascularization; moreover, the data demonstrate the superiority of an internal mammary artery pedicle graft to provide submucosal vascular in- growth and to minimize anastomotic stenosis.

(Ann Thorac Surg 1990;49:574-9)

source of rapid capillary ingrowth into ischemic bronchial tissue. Cooper and associates [5] have demonstrated initial mucosal vascular communications when canine bronchial suture lines were wrapped with greater omen- tum. These experimental results using the greater omen- tum are in part responsible for the recent success of human lung allografts. Intercostal muscle grafts have also been used as a source of capillary ingrowth [lo, 111. This study investigates the use of an internal mammary artery pedicle graft as a source for indirectly revascularizing ischemic bronchial segments as compared with previously described omental and intercostal muscle pedicle grafts.

Material and Methods Thirty conditioned mongrel dogs weighing from 18 to 23 kg were selected for this study. All animals were hu- manely cared for in accordance with the ”Principles of Laboratory Animal Care” and the “Guide for the Care and Use of Laboratory Animals” (NIH Publication No. 85-23, revised 1985). The dogs were anesthetized and main- tained on a combination of halothane and oxygen.

Thirty dogs received an internal mammary artery (IMA, n = 6), omental (OM, n = 9), intercostal muscle (ICM, n = 9), or no pedicle wrap (control, n = 6 ) around the ischemic “donor” bronchus after single-lung autotrans- plantation. All vascular pedicle grafts were mobilized before pneumonectomy. After upper midline laparotomy,

0 1990 by The Society of Thoracic Surgeons 0003-4975/90/$3.50

Ann Thorac Surg 1990;49574-9

TURRENTINE ET AL 575 BRONCHIAL HEALING WITH VASCULAR WRAPS

the greater omentum, based on the right gastroepiploic artery, was advanced through a small anterior incision in the left hemidiaphragm beneath the xiphoid process. Both IMA and ICM pedicle grafts were directly mobilized through a left posterolateral thoracotomy in the fifth intercostal space. The ICM pedicle was based on the fifth intercostal artery and vein. The IMA and vein, along with a 1-cm segment of intrathoracic muscle medial and lateral to the vessels, were dissected from the chest wall using electrocautery beginning at their origin and extending to the terminal bifurcation (Fig 1A).

Left lung autotransplantation was performed using a modification of the technique as described by Cooper and associates [5]. After dissection and isolation of the pulmo- nary artery and vein, as well as the left main bronchus, an 8F balloon-tip catheter (American Edwards Laboratories, Santa Ana, CA) was used to occlude the left main bron- chus. Left pneumonectomy was accomplished by divid- ing the pulmonary artery and the left main bronchus at their respective midpoints. The pulmonary veins were divided and left with a small cuff of left atrium. The left atrial cuff was then anastomosed with a running 5-0 Prolene suture (Ethicon, Somerville, NJ) using a hori- zontal mattress suture technique. The bronchial anasto- mosis was carried out with interrupted horizontal mat- tress sutures of 5-0 Vicryl (Ethicon) and the bronchial occluding catheter was removed. The pulmonary artery anastomosis was performed with a continuous running 6-0 Prolene suture. Before reestablishing pulmonary ar- tery flow, retrograde flushing was accomplished by brief release of the left atrial clamp. Except for the control animals, the bronchial anastomosis was then circumfer- entially wrapped with the OM or with the pleural surface of either the ICM or distal IMA pedicle grafts (Fig 1B). The surfaces of the pedicle grafts were closely coapted and gently sutured to the bronchial adventitia with inter- rupted 5-0 Vicryl sutures in four-quadrant fashion while avoiding tension on the pedicle vascular supply. The animals were maintained on prophylactic antibiotics con- sisting of 1 g of intravenous Kefzol (Eli Lilly, Indianapolis, IN) intraoperatively, followed by 1.2 x lo7 U of intramus- cular Distrycillin (Solvay Veterinary Inc, Princeton, NJ) daily for five postoperative days.

The animals in each study group were equally divided and killed at 4, 8, and 12 postoperative days. The dogs were killed in accordance with the laboratory animal guidelines. In control animals, the thoracic aorta was clamped distal to the left subclavian artery and proximal to the diaphragm. Intercostal artery branches from the thoracic aorta were then ligated and 20 mL of Microfil latex (Canton Biomedical, Boulder, CO) was injected into the thoracic aorta, filling the bronchial arteries and their mediastinal branches. In bronchial wrap animals, the pedicle arteries were examined for patency, cannulated with a 20F angiocatheter, flushed with heparinized saline solution, and then injected with 5 mL of Microfil. The heart and lungs were then removed from all animals en bloc while the integrity of the pedicle wraps was main- tained. Transverse (A) and sagittal (B) diameters of the left main bronchus proximal to and at the anastomotic site were measured by a blinded investigator. The corre-

A

B Fig 1 . (A) Left main bronchial anastomosis with mobilized internal mamma y artery pedicle graft after distal vessel ligation and pedicle division. ( B ) Orientation of internal mammary artery pedicle after circumferential wrapping of bronchial anastomosis.

sponding cross-sectional areas (CSAs) were calculated by the formula A/2 x B/2 x 7r. The degree of luminal reduction was determined by dividing the anastomotic CSA by the proximal CSA. The amount of submucosal latex present in the donor bronchus was graded either 0 (no latex), 1 (sparse or spotty latex), 2 (latex streaks), or 3 (confluent latex) (Fig 2). The latex was allowed to set for three hours, and the specimen was placed in formalin.

Histological analysis was performed by a single pathol- ogist who was blinded as to both the study group and

576 TURRENTINE ET AL BRONCHIAL HEALING WITH VASCULAR WRAPS

Ann Thorac Surg 1990;49:5749

Fig 2. Submucosal vascular ingrowth grading system. (A) Grade 0-no latex (control, day 12). ( B ) Grade I-sparse or spotty latex (control, day 8). (C) Grade 2-latex streaks (intercostal muscle, day 4), (D) Grade 3 4 o n f l u e n t latex (omenturn, day 12). (B = bronchial lumen; large arrow = bronchial suture line; small arrow = submucosal vascular latex.)

time interval of animal death. Hematoxylin and eosin- stained specimens were examined for the presence of Microfil within the submucosal vascular plexus and for mucosal integrity.

All statistical analyses were accomplished using the SAS software system. A Kruskal-Wallis test and Dunn's post hoc multiple comparison procedure were used to investigate the relationship between treatment group and revascularization category (grade 0-3). One-way analysis of variance and Duncan's multiple comparison procedure [12] were used to compare cross-sectional donor bronchial area between treatment groups, whereas simple linear regression was used to investigate the relationship be- tween cross-sectional donor bronchial area and revascu- larization category (grade 0-3). Multivariate relationships between anastomotic cross-sectional area and the degree of submucosal revascularization, day of death, and treat- ment group were investigated using analysis of covariance.

Results There was one premature death, which was unrelated to the operative procedure itself, and the animal was ex- cluded from the study. All pedicle arteries remained

patent. There was no bronchial dehiscence in any wrapped anastomosis; however, there was one frank anastomotic disruption (16.7%) in the control group.

Microfil injection into the collateral mediastinal arteries of control animals failed to reveal any capillary bridging across the anastomosis at postoperative day 4 (Fig 3) . Sparse submucosal latex streaking (grade 0-1) was found in the ischemic bronchus of only 1 control animal killed on postoperative day 8. In contrast, at least 1 animal in every group with a wrapped anastomosis demonstrated grade 1 submucosal latex at 4 postoperative days. Moreover, 2

'1 n

4 DAY 8 DAY 12 DAY

Fig 3 . Mean submucosal vascular ingrowth by latex grading at time of death. (solid = control, crossed hatched = intercostal muscle; diagonal = omenturn; open = internal mammary artery.)

Ann Thorac Surg 1990;49574-9

TURRENTINE ET AL 577 BRONCHIAL HEALING WITH VASCULAR WRAPS

K 100, I-

Histological evaluation of the bronchial mucosa in the control group demonstrated ulceration and necrosis in 3 animals, with only 3 dogs showing either regenerating epithelium or intermediate changes (50%). Random histo- logical sections failed to detect submucosal latex in 5 of the 6 control animals. In contrast, bronchi supported by IMA external wraps demonstrated 5 animals with regenerating mucosa (83%), and only 1 with minimal focal ulceration. Similarly, 78% (7/9) of OM-wrapped and 89% (8/9) of IQM- wrapped bronchi had epithelial regeneration.

DECREE OF B R O N C H I A L R E V A S C U L A R I Z A T I O N

Fig 4. Scattergram plot of “donor“ bronchus diameter with respect to the degree of submucosal vascular ingrowth including simple linear regression line with 95% confidence limits.

animals in every group with wrapped anastomoses had circumferential and confluent submucosal vascular in- growth into the ischemic donor bronchus at or after postoperative day 8. Bronchi wrapped with the IMA pedicle appeared to have more consistent and progressive confluent vascular ingrowth than the other two pedicle grafts. Results obtained from the Kruskal-Wallis test indi- cated the presence of an overall significant difference in vascular ingrowth between the treatment groups ( p = 0.02). Furthermore, Dunn’s multiple comparison proce- dure demonstrated that there existed a trend for the vascular ingrowth observed using IMA, OM, and ICM pedicle grafts to be greater than that observed in the control group ( p = 0.09).

The extent of submucosal vascular ingrowth influenced the degree of bronchial luminal narrowing at the anasto- mosis. All donor bronchi with grade 3 submucosal latex retained 81.2% ? 4.8% of original bronchial cross- sectional area as compared with bronchi with grade 2 (76.5% & 6.6%), grade 1 (68.4% & 10.9%), and grade 0 (65.0% * 8.4%) (Fig 4). Simple linear regression demon- strated a significant trend of increasing bronchial anasto- motic CSA with increasing grade of submucosal latex ( p = 0.002). Application of analysis of covariance to these data, controlling for treatment group and day of death, indi- cated that in the presence of these other variables, the amount of submucosal latex remained significantly pre- dictive of an increase in anastomotic CSA ( p = 0.04).

Results obtained from one-way analysis of variance demonstrated a statistically significant difference ( p = 0.003) between the observed CSA for IMA (84.5% * 3.3%), ICM (66.9% ? 20.9%), OM (68.4% ? 8.3%), and control groups (70.2% & 7.6%) (Fig 5). Furthermore, Duncan’s multiple comparison procedure indicated that bronchi supported with IMA pedicle grafts had signifi- cantly greater CSA than those in the ICM, OM, and control groups, and that no significant differences existed between these latter three treatment groups. The treat- ment group remained a significant predictor ( p = 0.008) of anastomotic stricture using analysis of covariance to con- trol for day of death and amount of submucosal latex. Duncan’s test again indicated that bronchi wrapped with IMA pedicle grafts had significantly greater CSA than bronchi in the other three treatment groups.

Comment Bronchial healing after radical tracheobronchial operation is not infrequently complicated by anastomotic stenosis or even frank dehiscence. This has been attributed to isch- emic injury secondary to interruption of the bronchial arterial circulation. Bronchial arteries originate from the aorta or intercostal arteries. They initially branch at the carina, then provide blood supply to main and segmental bronchi through a peribronchial adventitial plexus [ 11. The distal main and first segmental bronchi, unlike the proximal main bronchi and carina, have bronchopulmo- nary arterial collaterals that provide an important source of bronchial blood supply if the bronchial arteries are divided. Without this dual blood supply, proximal main bronchial segments appear to be more susceptible to ischemic injury.

Both omental and intercostal artery pedicle wraps have previously been found to provide early capillary ingrowth into ischemic bronchial tissue and improve healing [5, 10, 111. The current success of human lung transplantation has been attributed, to a large degree, to the technique of wrapping the bronchial anastomosis with these vascu- lar tissue pedicles, thereby providing improved blood supply to the anastomosis and reducing the incidence of bronchial dehiscence. By necessity, omental wraps use celiotomy for mobilization with graft passage through a diaphragmatic defect resulting in a pleuroperitoneal com- munication. Additionally, cachectic, malnourished pa- tients or patients with previous upper abdominal opera- tion may not have sufficient omentum for bronchial coverage. Intercostal muscle pedicle grafts, based on an intercostal artery, have not previously demonstrated mu-

LT W c Y I 5 n VI 3 I

z- 0

u 9

e! m

e! 0

0 z

9

100

70

60

€

::I , , , , 30

C O N T R O L OMENTUM I C M IMA

Fig 5. “Donor“ bronchus diameter expressed as mean percentage of proximal bronchial diameter observed between treatment groups. (Bar = one standard deviation; ICM = intercostal muscle; IMA = inter- nal mammary artery.)

578 TURRENTINE ET AL BRONCHIAL HEALING WITH VASCULAR WRAPS

Ann Thorac Surg 1990;49:574-9

cosal capillary ingrowth until the seventh postoperative day [lo]. Moreover, this pedicle graft is based on a smaller, more tenuous vessel that may be easily damaged during mobilization or kinked when the pedicle is rotated into position around the bronchus. Although direct revas- cularization by bronchial arterial anastomosis has great theoretical appeal, this requires microvascular techniques without uniform reliability.

Diaphragmatic muscle wraps based on branches of the IMA have been described [13], but our purpose is to describe use of the IMA pedicle for bronchial revascular- ization. The IMA pedicle graft has several theoretical advantages. Its dissection and mobilization have become a routine part of most coronary artery bypass procedures, which may be done through either a sternotomy or posterior lateral thoracotomy incision. Because of its ana- tomical location and length, the internal mammary artery pedicle graft can circumferentially wrap any mediastinal tracheal or main bronchial suture line without tension, kinking, or distortion. In our experiment, the IMA pedicle graft demonstrated more consistent and progressive re- vascularization of the ischemic bronchus as compared with the other study treatments or a control (no wrap). This early revascularization resulted in improved preser- vation of bronchial mucosa, promoted normal healing, and minimized anastomotic stricture.

Confluent submucosal vascular ingrowth, provided by any of the three types of arterial based wraps, was found to be independently predictive of reducing bronchial anastomotic stenosis. Several animals with intercostal artery or omental wraps, however, had no capillary in- growth that resulted in severe anastomotic strictures similar to controls. This provides further evidence that revascularization, even by indirect methods, promotes more normal bronchial healing.

Before assuming that these results will extrapolate clinically, one must be mindful of several important anatomical variations that exist between the human and canine IMA. First, the canine IMA pedicle has a well- developed muscle with a transparent pleural surface as opposed to that of humans, which is mainly fibrofatty tissue with a thick fascia1 layer lining the pleural surface that could potentially limit vascular ingrowth. Sparse muscle is found only on the inferior one third of its intrathoracic course. Additionally, the canine IMA diam- eter and flow is significantly greater per body weight than that of humans. Possibly more important are functional differences between the IMA pedicle graft and omental wraps with respect to reticuloendothelial function. Exten- sive phagocytic properties of the omentum are not present in the IMA pedicle graft, which theoretically may be important for local bacteriologic control, especially in immunosuppressed patients.

This study supports earlier evidence regarding the ability of omental and intercostal artery pedicle grafts to promote early indirect revascularization of ischemic bron- chial tissue. Based on the findings of this study, the canine IMA pedicle graft appears to be an excellent intrathoracic source for early submucosal revasculariza- tion of the bronchus, providing a more consistent and progressive ingrowth pattern as compared with wraps currently used. It would seem prudent, however, to clinically evaluate IMA vascular pedicle ingrowth charac- teristics in nonimmunosuppressed patients undergoing radical tracheobronchial operation before the IMA's appli- cation in lung transplantation.

We thank Randy Bills and Stacy Wesemann for assistance in the operative phase of this study, as well as Myrtle Parsley for secretarial support.

References 1. Deffebach ME, Charan NB, Lakshminarayan S, Butler J. The

bronchial circulation. Am Rev Respir Dis 1987;135:463-81. 2. Stone RM, Ginsberg RJ, Colapinto RF, Pearson FG. Bronchial

artery regeneration after radical hilar stripping. Surg Forum

3. Pearson FG, Goldberg M, Stone RM, Colapinto RF. Bronchial artery circulation restored after reimplantation of canine lung. Can J Surg 1970;13:243-50.

4. Siegelman SS, Hagstrom JWC, Koerner SK, Veith FJ. Resto- ration of bronchial artery circulation after canine lung allo- transplantation. J Thorac Cardiovasc Surg 1977;73:7925.

5. Lima 0, Goldberg M, Peters WJ, Ayabe H, Townsend E, Cooper JD. Bronchial omentopexy in canine lung transplan- tation. J Thorac Cardiovasc Surg 1983;83:418-21.

6. Blumenstock DA, Kahn DR. Replantation and transplanta- tion of the canine lung. J Surg Res 1961;1:40-7.

7. Largiader F, Manax WG, Lyons GW, Lillehei RC. Technical aspects of transplantation of preserved lungs. Dis Chest

8. Veith FJ, Richards K. Improved technique for canine lung transplantation. Ann Surg 1970;171:55M.

9. Huggins CE. Reimplantation of lobes of the lung: an exper- imental technique. Lancet 1959;2:1059-62.

10. LeGial YM, Chittal SM, Wright ES. Early bronchial revascu- larization with an intercostal pedicle graft following canine lung autotransplantation. Can J Surg 1985;28:518-22.

11. Fell SC, Mollenkopf FD, Montefusco CM, et al. Revascular- ization of ischemic bronchial anastomosis by an intercostal pedicle flap. J Thorac Cardiovasc Surg 1985;90:172-8.

12. Milliken GA, Johnson VE. Analysis of messy data; Vol 1: Designed experiments. New York: Van Nostrand Reinhold, 1984.

13. Westby S, Shepherd ME', Nohl-Oser HC. The use of pedicle grafts for reconstructive procedures in the esophagus and tracheobronchial tree. Ann Thorac Surg 1982;33:486-90.

1966; 17: 109-10.

1966;49:1-7.

Ann Thorac Surg 1990;49:574-9

TURRENTINE ET AL 579 BRONCHIAL HEALING WITH VASCULAR WRAPS

DISCUSSION

DR G. ALEXANDER PATTERSON (Toronto, Ont, Canada): I would like to compliment Dr Turrentine and associates for a very well-designed and executed study of a problem that I think is becoming increasingly important as the number of pulmonary transplant centers increases. In our own center, 3 of 27 single- lung and 8 of 17 double-lung recipients have had major airway complications. Three of these were fatal.

Oriani Lima, working with Dr Joel Cooper, has demonstrated the remarkable capacity of omentum to revascularize ischemic donor bronchi in a very short time. Dr Veith’s group subse- quently demonstrated reliable revascularization with intercostal muscle.

We have used omentopexy routinely, but in recent cases, we have employed a pedicle of anterior mediastinal fat and thymus on a superior-based pericardial flap. We continue to use omen- tum when this flap is of questionable substance or viability. The Pittsburgh group has employed IMA pedicles to encircle tracheal anastomoses after heart-lung and double-lung transplantation.

We are also learning about angiogenesis. Recent reports have described a lipid extract of omentum with remarkable angiogenic properties. We have recent experimental data suggesting that basic fibroblastic growth factor, a potent angiogenic agent, does not augment omental revascularization of transplanted tracheal segments. We have also found that improved pulmonary pres- ervation appears to be associated with improved bronchial via- bility after lung allotransplantation. This is presumably due to increased preservation of microcirculation and collateral flow between the pulmonary artery and bronchial artery so well described by the Pittsburgh group.

I would like to make one further comment and ask a couple of questions. I hate to expose my statistical ignorance, but I would be a little cautious about assuming a superiority of the IMA pedicle in such a small number of animals killed at variable times.

I do not have any question that bronchial wrapping is impor- tant. I am not so certain that we can be categorical about what is the best flap at the moment.

I wonder if Dr Turrentine can tell me whether or not these animals were all heparinized and whether the lungs were flushed before extraction.

I am also rather in admiration of 31 left lung allografts done without any evidence of atrial thrombosis, and I wonder if you could comment on whether you had an incidence of atrial thrombosis in your transplanted animals. You mentioned differences in cross-sectional area from group

to group, but was there a difference in cross-sectional area from time to time? In other words, did cross-sectional area at four days differ from that observed at eight or 12 days, and if so, was there a difference between groups?

DR STANLEY C. FELL (Chappaqua, NY): I congratulate Turren- tine and associates on demonstrating yet another method of revascularizing the bronchus. However, I do not believe their test was sufficiently rigorous. Only one of their control animals had a disruption. We have used the same experimental model that the Toronto group used, namely, the postpneumonectomy bronchial autograft, when we did a study on intercostal muscle flaps [Fell SC, Mollenkopf FP, Montefusco CM, et al. Revascularization of ischemic bronchial anastomoses by an intercostal pedicle flap. J

Thorac Cardiovasc Surg 1985;90:72-8], and the control is abso- lutely lethal. So I would suggest that this study be repeated using the proper control model.

DR JOEL D. COOPER (St. Louis, MO): I was not sure whether steroids were used. Just to recap, in the original work that Dr Lima and co-workers in Toronto did with us, in the absence of steroids, canine reimplanted bronchi and transplanted bronchi did appear to heal well. It was the combination of postoperative steroids and ischemia that led to poor healing.

We evaluate stenosis at various times, but stenosis is a late manifestation of the ischemic injury. By 12 days it is not usually apparent. The stenosis occurred without steroids, even in the presence of good wound healing. When observed at 3 weeks or later, the cross-sectional area began to narrow, and in that group, the omental wrap prevented stenosis compared with no omental wrap. So a couple of different factors were mixed in.

Finally, science has crept up with clinical practice. We now know why the omentum, as Dr Patterson commented, works well. It is the angiogenesis factor, of which it has a very potent supply, though this is not unique to the omentum. Angiogenesis factor is known to be in all fatty tissues. One of the contributions of the late Lyman Brewer was the use of the pericardial fat pad, also rich in angiogenesis factor and blood supply. So I do not think the omentum is unique. It happens to be bulky and easy to use, but other tissues throughout the body may also be used as long as they have good blood supply.

Dr Turrentine, can you tell me about the effects of steroids? Also, did you keep some of the dogs longer than 12 days so that you might perhaps begin to see some of the stenotic complica- tions that, we think, are late manifestations of ischemia?

DR TURRENTINE: I thank all the discussants for their com- ments. It is certainly an honor to have the pioneers of clinical lung transplantation discuss this work.

First, Dr Patterson, our technique was to heparinize, explant, and directly reimplant the lung soon after with ischemic times less than 45 minutes. We did experience problems with left atrial clot early on while developing this autotransplant model. After we changed the left atrial suture technique to a running horizon- tal mattress, which everts the muscular atrial wall, we found no incidence of left atrial thrombosis in any study animal.

Dr Cooper, we hesitated to use the term bronchial stenosis, and perhaps we should rather use the term reduction in luminul ureu. Histologically, ulcerating mucosa in these “stenotic” areas ap- peared heaped up, which we feel represented an ischemic insult rather than a true scarified stenosis.

We did not use steroids in this experimental protocol. Perhaps this would have made a more challenging model by placing the anastomosis at greater risk for dehiscence; however, the primary purpose of the study was to evaluate vascular ingrowth in the ischemic bronchial segment with a comparative analysis between the vascularized pedicles suitable for such use.

We have had 4 animals that have undergone IMA-wrapped bronchial anastomosis after left lung autotransplantation and were studied up to 3 months postoperatively. In these longer term experiments, we have observed that bronchial lumen diam- eters compare equally with those of 12-day animals whose bronchi were wrapped with similar grafts.