TUESDAY AFTERNOON, MAY 5, 1998

1:45 p.m. SIMULTANEOUS SCIENTIFIC SESSION A CONGENITAL HEART DISEASE

Ballroom A, Hynes Convention Center

Moderators: Frank L. Hanley, M.D.

Thomas L. Spray, M.D.

19. MULTIPLE VENTRICULAR SEPTAL DEFECTS: HOW AND WHEN SHOULD WE REPAIR?

Francesco Seddio, M.D.*, Doff B. McElhinney, M.S.*, V. Mohan Reddy, M.D.*, Wayne Tworetzky, M.D.*, Norman H. Silverman, M.D.* and Frank L. Hanley, M.D.

San Francisco, California

Discussant: Constantine Mavroudis, M.D., Chicago, Illinois

Congenital heart lesions with multiple ventricular septal defects (VSDs) remain a surgical challenge. Difficult exposure of VSDs in the trabecular component of the septum, especially the apical and anterior portions, complicates closure of such defects. Traditional approaches often rely on either ventriculotomy for exposure or palliation with pulmonary artery banding (PAB). Both of these techniques have potential drawbacks, so it is important to limit their use to pts in whom they will provide optimal benefit. However, indications for closure vs palliation and for various approaches to surgical exposure are not clearly defined. Since 1992, 32 pts with multiple VSDs have undergone surgery for this condition. Median age was 86 days and all but 4 pts were infants. Associated lesions were present in 27 pts, including transposition of the great arteries (9), aortic coarctation (4), atrioventricular septal defect (3), interrupted arch (2), tetralogy of Fallot (1), patent ductus arteriosus (3), atrial septal defect (3), and other defects (3). Prior operations had been performed elsewhere in 8 pts, including PAB in 6. Each pt had ≥ 2 VSDs (median 3). VSDs were perimembranous in 13 pts, inlet in 6, outlet in 2, anterior muscular in 13, posterior muscular in 10, apical in 21, and mid-muscular in 15. "Swiss cheese" VSDs were present in 6 pts, 5 of whom had prior PAB. Initially, 25 pts underwent complete repair, all through right atriotomy except 1 who also required aortotomy. Seven pts were palliated with PAB (5) or other procedures. No pt received a ventriculotomy. In the pts who underwent complete repair, a combination of patch and suture closure was used in 16, patch only in 6, and sutures only in 3. In pts with muscular VSDs, the left side of the septum was explored through an aortotomy or a larger VSD when possible, which aids in visualization and closure of defects in the trabecular septum. There was 1 early death in a pt who had staged repair of coarctation and multiple VSDs. 3 palliated pts underwent early reoperation for PAB revision due to failure to thrive (1), PAB revision and then VSD closure for failure to thrive (1), or PAB removal after spontaneous closure of the VSDs led to decompensation (1).
Thus, 3 of 7 palliated pts required early reoperation. Pacemakers were placed for atrioventricular dissociation in 2 pts, but sinus rhythm returned in both. At median followup of 22 mos (4.49 mos), there was 1 death in a palliated pt and 1 other palliated pt had a heart transplant for cardiomyopathy. Another palliated pt had the PAB removed and multiple apical VSDs closed 8 mos later after outgrowing the band. No pts have hemodynamically significant residual VSDs or persistent rhythm abnormalities. In our experience, palliation of multiple VSDs is associated with greater morbidity than primary repair. Almost all cases of multiple VSDs can be repaired adequately in early infancy without ventriculotomy, though "swiss cheese" VSDs are still an indication for palliation.

*By invitation

20. TRANSATRIAL-TRANSPULMONARY TETRALOGY REPAIR IS EFFECTIVE IN THE PRESENCE OF ANOMALOUS CORONARIES.

Christian P.R. Brizard, M.D.*, Andrew D. Cochrane, F.R.A.C.S.*and Tom R. Karl, M.D.

Melbourne, Australia

Discussant: John W. Brown, M.D., Indianapolis, Indiana

Objectives: To analyze the outcome of transatrial-transpulmonary tetralogy repair in children with an anomalous coronary crossing the RV outflow tract. Methods: 34/548 (6.2%) of our tetralogy repairs were associated with surgically relevant coronary artery anomalies. The median age and weight at repair were 25.8 months (4.4-88) and 10 kg (5-20). All patients had a major branch crossing the right ventricular outflow tract. Anomalies included LAD from RCA (n = 19), RCA from LCA or LAD (n = 10), large RCA conus branch (n = 4), and single RCA (n = 1). Diagnosis was established preoperatively in 24/34 with angiography (n = 23) or echo (n=1). Transatrial-transpulmonary repair was successfully used in 31 patients, and 24 of whom required a limited transannular patch in the RVOT. Three additional patients had a RV-PA conduit due to proximity of the coronary branch to the PA annulus, and inability to relieve the right ventricular outflow obstruction adequately.

Results: There has been no early or late mortality. Mean RV-PA gradient at last follow-up was 18mm (SD=10.6), compared to 15mm (SD = 24) for patients with normal coronaries operated by a similar approach. Actuarial freedom from reoperation at 105 months was 95% (± 4.9%), which was also similar to results of transatrial-transpulmonary repair in children with TOP and normal coronaries (p > 1.0).

Conclusions: Surgically important coronary anomalies in TOP can be dealt with using the transatrial-transpulmonary approach in most cases, without major alterations in technique. Outcome is similar to that for other TOP patients. Thus, the presence of anomalous coronary has not imparted incremental risk following this surgical strategy.

*By invitation

21. SINGLE STAGE REPAIR OF AORTIC ARCH OBSTRUCTION AND ASSOCIATED INTRACARDIAC DEFECTS USING PULMONARY HOMOGRAFT PATCH AORTOPLASTY.

Christo I. Tchervenkov, M.D.*, Stephen A. Tahta, M.D.*, Marie J. Beland* and Luc C. Jutras*

Montreal, Quebec, Canada

Sponsored by: Ray C.-J. Chiu, M.D., Ph.D. Montreal, Quebec, Canada

Discussant: Claude Planche, M.D., Paris, France

Intracardiac malformations associated with coarctation and aortic arch hypoplasia have traditionally been repaired in two stages, with a high mortality. Recently, improved survival has been reported with a single-stage approach predominantly using an extended end-to-end anastomosis to repair the aortic arch. Herein, we review our experience with a single-stage biventricular repair of intracardiac defects associated with aortic arch hypoplasia using a pulmonary homograft patch aortoplasty. Between October 1988 and October 1997, 39 of 40 consecutive patients underwent a single-stage biventricular repair for aortic arch obstruction and associated intracardiac defects. The median age at operation was 17 days and the mean weight was 3.71 ± 1.09 kg. Nineteen patients had either d-transposition of the great arteries or the Taussig-Bing anomaly. Sixteen patients had multiple left-sided obstructive lesions (2 critical aortic stenosis, 3 subaortic stenosis and VSD, 11 class III hypoplastic left heart syndrome as defined by Kirklin). One patient had an associated complete atrioventricular septal defect. Four patients had only an associated ventricular septal defect. By median sternotomy, the hypoplastic aortic arch was enlarged with a pulmonary homograft patch in 36 patients. In 4 patients, an extended end-to-end anastomosis was performed.

There were 2 early deaths (5%) and 21 late deaths (5%). One late death was non-cardiac related. The median followup time was 25 months (range 1 month to 8 years). The recoarctation rate was 11% but excluding those patients with associated left-sided obstructive lesions, this decreases to 0%. No aneurysm formation in the aorta has occurred. The actuarial survival rate at 8 years is 89 ± 10%.

One-stage biventricular repair of aortic arch obstruction and ‘associated intracardiac defects can achieve excellent survival. We recommend the pulmonary homograft patch aortoplasty of the hypoplastic arch because it achieves complete relief of anatomical afterload with a tension-free anastomosis and low incidence of recoarctation.

2:45 p.m. INTERMISSION - VISIT EXHIBITS

*By invitation

3:15 p.m. SIMULTANEOUS SCIENTIFIC SESSION A CONGENITAL HEART DISEASE

Ballroom A, Hynes Convention Center

Moderators: Frank L. Hanley, M.D.

Thomas L. Spray, M.D.

22. PRIMARY REPAIR IS SUPERIOR TO INITIAL PALLIATION IN CHILDREN WITH ATRIOVENTRICULAR SEPTAL DEFECT AND TETRALOGY OF FALLOT.

Hani K. Najm, M.D.*, Glen S. Van Arsdell, M.D.*, Stefan Watzka*, Lisa Hornberger, M.D.*, John G. Coles, M.D.* and William G. Williams, M.D.

Toronto, Ontario, Canada

Discussant: Hillel Laks, M.D., Los Angeles, California

Children presenting with atrioventricular septal defect in conjunction with tetralogy of Fallot (AVSD/TOF) represent an uncommon complex lesion. To explore the best management algorithm we undertook a review of this institution's experience. From March 1981 to August 1997, 38 children were referred to our division with the diagnosis of AVSD/TOF. Down's syndrome was present in 32 (84%). Twenty one children were initially palliated with a systemic to pulmonary artery shunt. Two (9.5%) died prior to repair. Complete repair has been performed on 31 children. Of these 14 underwent initial palliation with systemic to pulmonary artery shunt (mean age at shunt 20 ± 24 months). During complete repair the endocardial cushion defect was repaired with two patches in 29 children. Relief of the right ventricular outflow obstruction was by a transannular patch in 22 (70%), 12 (54%) of those had a monocusp inserted, and 4 required an infundibular patch. Operative mortality occurred in 2 (6.4%) children, one was previously palliated. Reoperations were performed in 11 children, 7 (58%) were for pulmonary arterioplasty. The incidence of reoperation was higher in the palliated versus the non palliated groups (64% vs. 12%, P= 0.007). The palliated group were repaired at an older age (78 ± 52 vs. 36 +19 months, P= 0.005), had longer ventilatory support (8 ± 11 rs. 4 ± 3 days, P= 0.05), longer inotropic support (8 ± 13 vs. 3.6 ± 1.5 days, P= 0.028), and longer hospital stay (23 ± 22 vs. 14 ± 5 days, P= 0.004). There was one late death related to a reoperation for pulmonary arterioplasty and residual ventricular septal defect in the palliated group. We conclude that repair of AVSD/TOF can be achieved with low mortality. In this series initial palliation with a shunt resulted in a more complex post-operative course and carried a higher reoperative rate, mainly due to pulmonary artery complications. Primary repair of AVSD/TOF at the time of symptomatic presentation is superior to initial palliation followed by repair.

*By invitation

23. THE MODIFIED NORWOOD PROCEDURE FOR HYPOPLASTIC LEFT HEART SYNDROME: EARLY TO INTERMEDIATE RESULTS OF 120 PATIENTS.

William J. Brawn, F.R.C.S.*Kozo Ishino, M.D.*, Oliver Stumper, M.D.*, Joseph V. De Giovanni, F.R.C.P.*, Eric D. Silove, F.R.C.P.*, John G.C. Wright, F.R.C.P.* and Babulal Sethia, F.R.C.S.*

Birmingham, England

Sponsored by: †Marc de Leval, M.D., London, England

Discussant: Roger B.B. Mee, M.B., Ch.B., Cleveland, Ohio

Background: Classical first-stage Norwood repair of hypoplastic left heart syndrome (HLHS) utilizes homograft patch enlargement to obtain an unobstructed aorta and coronary arteries. Because of possible disadvantages of the homograft such as lack of growth, degeneration and calcification, and homograft availability, we have tried to repair the aorta without patch supplementation.

Method and Results: Between February 1993 and September 1997, 120 patients, aged 0 to 47 days (median, 4 days) and weighing 1.7 to 4.4 kg (median, 3.1 kg) underwent first-stage palliation for HLHS. The ascending aortic diameter ranged from 1.5 to 8 mm (median, 3.0 mm). Eight patients had an anomalous right subclavian artery (ARSCA) arising from the descending aorta In 95 patients (Group I), all ductal tissue was excised, the descending aorta anastomosed to the aortic arch which had been opened back to the ascending aorta, then to this confluence was anastomosed the proximal main pulmonary artery. In the remaining 25 patients (Group II), continuity of the aortic arch was maintained and the repair was performed with a Damus anastomosis. The size of the systemic-to-pulmonary shunt was 3.5 mm (70), 3 nun (48), and 4 mm (2). Circulatory arrest time ranged from 19 to 105 Min (median, 54 Min). A homograft patch was necessary for the arch reconstruction in a total of 18 patients (15.0%); 9 Group I patients (9.5%) and 9 Group II (36.0%) (p< 0.01). There were 82 hospital survivors (68.3%); 69 Group I (72.6%) and 13 Group II (52.0%) (p<0.05); 71 patients without patch (69.6%) and 11 with patch (61.1%) (p = NS). By multiple logistic regression, risk factor for hospital death included ascending aortic diameter ≤ 2 MM (p = 0.1018) and the ARSCA (p = 0.039). There were 6 late deaths. Sixteen of 71 patients (22.5%) who underwent second-stage palliation had developed neo-aortic arch obstruction with a peak gradient > 10 mmHg; 14 Group I (19.4%) and 2 Group II (22.2%) (p = NS); 15 without patch (23.1%) and one with patch (16.7%) (p = NS). Overall survivals were 56.7% at 1 year and 55.0% at 2 years.

Conclusion: The modified Norwood procedure without the use of exogeneous material does not result in an increased incidence of neo-aortic arch obstruction and may allow better long-term growth of the neo-aorta.

†1973-74 AATS Graham Fellow

*By invitation

24. POST-ISCHEMIC HYPERTHERMIA EXACERBATES NEUROLOGICAL INJURY FOLLOWING DEEP HYPOTHERMIC CIRCULATORY ARREST.

Dominique Shum-Tim, M.D.*, Mitsugi Nagashima, M.D.*, Toshiharu Shin'oka, M.D.*, Jan Bucerius, M.D.*, Georg Nollert, M.D.*, Hart G.W. Lidov, M.D.*, Adre du Plessis, M.D.*, Peter C. Laussen, M.D.* and Richard A. Jonas, M.D.*

Boston, Massachusetts

Discussant: Julie A. Swain, M.D., Lexington, Kentucky

Background: Aggressive surface warming is common practice in the pediatric intensive care unit; also, fever is commonly seen after cardiopulmonary bypass (CPB). Recent rodent data emphasizes the protective effect of mild (2-3°C) hypothermia after cerebral ischemia. The purpose of this study was to evaluate different temperature regulation strategies after deep hypothermic circulatory arrest (DHCA) using a survival piglet model. Methods: Fifteen piglets (9.1 ± 1.1 kg) were randomized into 3 groups (n = 5/group) and a small temperature probe was inserted in the superficial temporal cortex. All groups underwent 100 min. of DHCA at 15°C and 40 min. of reperfusion with a Hct of 25% during CPB. Brain temperature was maintained at 34°C for 24 hr after coming off CPB in Gp I, 37°C in Gp II, and 39°C in Gp III, respectively. The cerebral redox state was determined by near infrared spectroscopy (NIRS) during and 3 hr after CPB. Neurobehavioral recovery was evaluated daily for 3 days after extubation in a blinded fashion by neurological deficit score (NDS; 0 = normal, 500 = brain death) and overall performance capacities (OPC; 1 = normal, 5 = brain death). Body weight gain and total body water content indicated by electrobioimpedance were evaluated at 3 hr, 24 hr, and 4 days post-operatively. Brain was fixed in situ on day 4 and the hemisphere contralateral to the probe was examined and scored for ischemic injury (0 = normal, 5 = necrosis) in a blinded fashion. Results: All results are expressed as mean or percentage ± SD. P-value <0.05 is considered significant. One animal in Gp III died on day 3 due to severe neurological impairment. There were no differences in % body weight gain and bioimpedance after CPB. At 24 hr, Gp III animals had significant changes in weight line break (Gp 1= 109.3± 1.7%, Gp II = 113.6 ±6.3%, and Gp III = 117.9 ± 3.4%; p<0.05 Gp I vs III) and bioimpedance (Gp I = 83.9 ± 12.5%, Gp II = 74.0 ± 9.6%, and Gp III = 62.9 ± 10.8%; p<0.05 Gp I vs III) suggesting total body edema. There were no significant differences in NIRS signals throughout CPB. Recoveries of NDS, OPC, histology and NIRS were significantly worse in hyperthermic Gp III. There was a trend towards reduced injury in Gp I (see table).

Table	NDS	OPC	Histology (POD 4)			Oxy-Hb	Deoxy-Hb	Cyt. a,«3
			cortex	hippocampus	caudate	2hr off CPB	2hr off CPB	2hr off CPB
GpI	30.0±28.3	1.0±0.0	0.0 ± 0.0	0.0 ±0.0	0.8 ± 1.0	22.7±12.5	-28.7±9.0	3.9±0.6
GpII	87.0±59.4	1.4±06	1.0±1.2	0.0 ± 0.0§	0.8 ± 1.0	9.9±8.4	-10.1±20.6	05±2.1§
GplII	186.0±131*	2.8±1.3#	2.8 ± 1 8#	2.4 ± 0.9#	2.2 ±1 8	-13.5±12.1*	25.3±16.4*	-3.9±2.3*
p < 0.05 Gp I vs III by ANOVA and Bonferroni, #p< 0.05 Gp I vs III by Mann-Whitney u tests § p < 0.05 Gp II vs III

Conclusions: Mild post-ischemic hyperthermia significantly exacerbates functional and structural neurological injury after DHCA. Fever and active warming above 37°C should be avoided after DHCA. A mild degree of hypothermia may reduce cerebral injury after prolonged DHCA.

*By invitation

25. IS IT NECESSARY ROUTINELY TO FENESTRATE AN EXTRACARDIAC FONTAN?

LeNardo D. Thompson, M.D.*, Edwin Petrossian, M.D.*, Doff B. McElhinney, M.S.*, Natalia A. Abrikosova, M.S.*, Andrew J. Parry, M.D.*, V. Mohan Reddy, M.D.* and Frank L. Hanley, M.D.

San Francisco, California

Discussant: John E. Mayer, M.D., Boston, Massachusetts

Fenestration (FEN) of a Fontan connection has been proposed as a means of improving outcomes of single ventricle palliation by allowing partial decompression of the Fontan pathway. The benefit of FEN is likely to be greatest in the early postoperative period, when pts may experience increased pulmonary vascular resistance and decreased systolic and diastolic cardiac function due to the effects of cardiopulmonary bypass (CPB), aortic cross-clamping (ACC), and positive pressure ventilation. However, there are potential drawbacks to FEN, including arterial desaturation, complications of a right-left shunt, and the need for another procedure to close the communication, suggesting that routine use may not be in the best interest of all pts. The utility of FEN with the extracardiac conduit Fontan operation has not been determined. Since 1992, 60 pts have undergone a modification of the Fontan procedure in which an extracardiac inferior cavopulmonary conduit is used in combination with a previously staged bidirectional Glenn anastomosis. FEN was performed selectively in 32 pts (54%). Among the last 25 pts, FEN was used in only 6 (24%). In 7 pts, a FEN was placed or clipped in the early postoperative period without CPB. CPB time did not differ between the 2 groups (108 vs 122 rain), but significantly fewer non-fenestrated than fenestrated pts had ACC (3 vs 8; p<0.05). There were no operative deaths. Prolonged (>2 wks) pleural drainage occurred in 12 pts, 8 with FEN and 4 without. Fontan and common atrial pressures 1 day after surgery did not differ between non-fenestrated and fenestrated pts (13.1 vs 12.6 mmHg and 4.8 vs 5.8 mmHg, respectively). At followup ranging to 5 yrs, there were 2 late deaths and no pts developed protein losing enteropathy. These data suggest that there is no difference in outcome between pts who undergo the extracardiac conduit Fontan procedure with and without FEN. The extracardiac Fortran operation has a number of advantages that minimize the need for and utility of FEN, including avoidance of ACC, shorter duration of CPB, and a more streamlined hydrodynamic connection. These factors help optimize ventricular and pulmonary vascular function in the early postoperative period, which contributes to improved hemodynamics. Moreover, when necessary, a FEN can be placed or revised without return to CPB, which allows for more accurate and practical assessment of the need for FEN after the Fontan connection has been completed. We conclude that FEN is of no benefit in the majority of Fontan pts when an extracardiac conduit technique as described is used, and therefore FEN should not be performed routinely.

4:35 p.m. EXECUTIVE SESSION (Limited to Members)

Ballroom B, Hynes Convention Center

6:30 p.m. MEMBER RECEPTION

Museum of Science

*By invitation

1:45 p.m. SIMULTANEOUS SCIENTIFIC SESSION B ADULT CARDIAC SURGERY

Ballroom B, Hynes Convention Center

Moderators: Bruce W. Lytle, M.D.

Marko I. Turina, M.D.

26. DISENTANGLING PENETRATING ULCER AND INTRAMURAL HEMATOMA OF THE THORACIC AORTA FROM CLASSIC TYPE A AND B DISSECTIONS.

Michael A. Coady, M.D.*, John A. Rizzo, Ph.D.*, Graeme L. Hammond, M.D., Lee J. Goldstein, B.A.*, Gary S. Kopf, M.D. and John A. Elefteriades, M.D.

New Haven, Connecticut

Discussant: Nicholas T. Kouchoukos, M.D., St. Louis, Missouri

Classic type A and B aortic dissections have been well described, however, less is known about the natural history of penetrating atherosclerotic ulcers (PU) and intramural hematomas (IMH) of the thoracic aorta. PU are atherosclerotic plaques which disrupt the internal elastic lamina, burrow deeply through the intima into the aortic media, and cause a localized intramedial dissection. The diagnosis is made on CT scan by demonstration of a contrast-filled outpouching in the aorta in the absence of a dissection flap or a false lumen. Patients with IMH have a hematoma in the media but have no evidence of an atheromatous ulcer on imaging studies, and an intimal tear is not visualized. The objective of this study is to determine the natural history of PU and IMH and assess appropriate therapeutic correlates of the natural history. Data on 198 patients with aortic dissection (86 type A and 112 type B) treated at our institution from 1985 to 1997 were analyzed. The aortic dissection database includes 431 imaging studies (MRI, CT, and ECHO). Of the 198 patients, 27 individuals (13.6%) were found to have either a PU or IMH by CT/MRI imaging studies, intraoperative findings, or pathology reports. For PU and IMH, 5 cases (18.5%) were located in the ascending aorta, and the remaining 22 (81.5%) were in the descending aorta.

A chart review was conducted in all patients with type A and type B dissections, PU, and IMH to evaluate rates of rupture and further define the natural history. The follow-up period ranged from 2 to 107 months, with a mean of 26.2 months. Sixteen of the 27 patients with PU or IMH (59.3%) had malignant courses (7 patients (25.9%) went on to require aortic replacement and another 9 (33.3%) died of aortic rupture prior to surgery). Results of univariate analyses are summarized in the table and figure below. Patients with PU and IMH, in comparison to type A or B thoracic aortic dissections, affect an older age population, present with larger aortic diameters, occur primarily in the descending aorta, are associated more often with a prior diagnosed or treated AAA. The risk of aortic rupture is higher in patients with PU and IMH (9/27, 33.3%) than in patients with type A (6/86, 6.97%) (p<0.0003) or type B (4/1 12, 3.57%) aortic dissection (p<0.0001). One year survival is less for patients with penetrating ulcers (71%), as opposed to patients with type A (90%) or B dissections (90%) (p<0.05).

	Penetrating Ulcer/ Intramural Hematoma	Type Dissection	Type B Dissection
Patients	27 (15m, 12 f)	86 (61m, 25 f)	1 12 (70 m, 42 f)
Age	73.9 years ± 8.2†*	52.7 years ±17.6†	67.1 years ± 13.3*
Initial Aortic Size	6.45 cm ± 1.9	5.63 cm ± 1.5	5.4 cm ± 1.6
Prior AAA	8 (29.6%)	8 (9.3%)**	35 (31.2%)**
	Table reports means ± standard deviation, *p = 0.01, †p < 0.0001, **p = 0.001

We conclude that accurate recognition at initial presentation is critical for optimal clinical management of these patients. PU and IMH presenting acutely must be differentiated from classic acute dissections. For PU and IMH, the prognosis is more serious than with classic acute aortic dissection. Surgical treatment should be considered strongly at presentation.

*By invitation

27. RUPTURE OF CHRONIC TYPE B DISSECTING ANEURYSMS: A NATURAL HISTORY STUDY.

Tatu Juvonen, M.D.*, M. Arisan Erigan, M.D., Jan D. Galla, M.D.*, Steven L. Lansman, M.D., Jock N. McCullough, M.D.*, Khanh H. Nguyen, M.D.*, David Spielvogel, M.D.*, James J. Klein, M.D.*, Carol A. Bodian, Dr.PH.* and Randall B. Griepp, M.D.

New York, New York

Discussant: D. Craig Miller, M.D., Stanford, California

In an attempt to identify risk factors for rupture and to improve management of patients with type B dissection who survive the acute phase without operation, we have studied 59 patients using serial computer-generated 3-dimensional CT scans. Patients were included if they did not undergo surgery during the acute phase of a documented type B dissection, and had at least two CT scans a minimum of three months apart thereafter. The median duration of followup was 4.4 years (range 0.3-9.1), and the median interval between scans was 204 days (98-655) in the unoperated group and 258 (61-443) days in those who ruptured.

During followup, four patients died of causes unrelated to the aneurysm, and nine patients underwent elective aneurysm resection because of large aneurysm size, rapid expansion, or development of symptoms: they are not considered further. Eleven patients (19%) experienced rupture, and 35 remain alive without operation or rupture. Possible risk factors for rupture and various dimensional parameters are shown below: the data for patients with rupture are from their last CT scans, and the data for patients still in nonoperative followup are from their penultimate studies. Data in the table are medians or percents; p values were derived using the Wilcoxon rank sum test for quantitative data, and Fisher's exact test for categorical data.

Risk Factors	No Operation or Rupture	Rupture	p value
	(n = 35)	(n = 11)
Age (years)	67	73	0.05
Years since dissection	3.0	3.4	NS
Male(%)	63	46	NS
Pain(%)	26	55	0.14
COPD(%)	14	46	0.04
Smoking (%)	54	64	NS
Hypertension (%)	69	73	NS
False lumen open (%)	59	71	NS
Aortic Dimensions (3dCT)
Descending diameter (cm)	4.7	5.4	0.05
Descending volume (cm3)	268	285	NS
Abdominal diameter (cm)	3.7	4.7	0.02
Abdominal volume (cm3)	97	171	0.03

Patients in whom aortic rupture occurred were significantly older, and they were significantly more apt to have a history of COPD when compared with those without rupture; they were also somewhat more likely to have complained of pain at the time of their last visit. In addition, the size and extent of the aneurysm appears to have significant predictive value: patients who subsequently experienced rupture had significantly larger descending aortic and abdominal aortic diameters and a significantly larger abdominal aortic volume than were observed in patients whose aneurysms remained intact. These data indicate that older patients with larger and more extensive aneurysms following acute type B dissections, especially if they have a history of COPD, are significantly more likely to experience rupture than younger patients with smaller chronic type B aneurysms. The presence of a persistently patent false lumen does not appear to increase the risk of rupture. Overall, our data from patients who have entered the chronic phase of type B dissection reveal a pattern which bears a striking resemblance to the natural history of patients with nondissecting aneurysms, suggesting that operative guidelines for patients with chronic type B dissections should be similar to those for patients with other types of chronic descending thoracic and thoracoabdominal aneurysms.

*By invitation

28. SURGERY FOR ASCENDING AORTIC AND ARCH ANEURYSM: ANTEGRADE AORTIC PERFUSION AVOID NEUROLOGICAL COMPLICATIONS.

Francesco Musumeci, M.D.*, Giovanni Casali, M.D.*, Frank Dunstan, Ph.D.*, Antonino Marullo, M.D.* and William J. Penny, M.D.*

Cardiff, Wales

Sponsored by: Lawrence H. Cohn, M.D., Boston, Massachusetts

Discussant: G. Michael Deeb, M.D., Ann Arbor, Michigan

Background. Femoral arterial cannulation is commonly used for cardiopulmonary bypass (CPB) with ascending aortic (AAo) and arch aneurysm. Retrograde cerebral embolization, retrograde aortic dissection and lower limb ischaemia are well recognized complications. In order to minimize these events, since 1991 we have used direct cannulation of the aneurysmal AAo for CPB. This paper reviews our experience and the effectiveness of antegrade aortic perfusion in reducing postoperative neurological events.

Methods. Between December 1991 and June 1997, 84 consecutive pts (50 males; age range 6 to 83 (median 54) yrs) were operated on. Twenty-four (26.4%) were older than 70 yrs. All pts were evaluated by preoperative CT or MRI and perioperative echocardiography. The aneurysm involved the AAo only in 39 pts and in 45 extended also to the aortic arch. When replacing the AAo, the aortic root was also replaced in 71 pts, 9 had root remodelling and 4 aortic valve replacement. In addition, 45 pts had arch replacement (partial 19, total 24, elephant trunk 2). Direct cannulation of the aneurysmal AAo and bicaval venous return was used for CPB. Cerebral protection was achieved with hypothermic circulatory arrest to a temperature of 15 to 28°C. Continuous retrograde cerebral perfusion (RCP) through the superior vena cava was used in 24 pts when circulatory arrest time was longer than 30 min. All pts had a short period of RCP for de-airing prior to re-establishing CPB. The arterial cannula was repositioned within the distal sutureline, proximal to the innominate artery. Median circulatory arrest, cross clamp and CPB time were 19 (range, 8 to 53) min, 89 (range, 48 to 1 64) min and 122 (range, 82 to 206) min, respectively.

Results. Two pts aged 74 and 76 died in hospital (2.5%) due to respiratory (1) and renal (1) failure. Post-operative complications were bleeding (3) and prolonged ventilatory support (4). No cerebral or peripheral embolic complications occurred. At a mean follow up of 29 (3 to 69) months, there were 3 late deaths. The estimated 4 yr survival rate is 93% (SE 7%).

Conclusion. Direct cannulation of the aneurysmal AAo is a simple and safe alternative to femoral artery cannulation. Antegrade flow throughout the procedure associated with a period of RCP at the end of circulatory arrest was not associated with neurological embolic events. These excellent results suggest that this technique is worthy of further evaluation in a prospective manner.

2.45 p.m. INTERMISSION - VISIT EXHIBITS

*By invitation

3:15 p.m. SIMULTANEOUS SCIENTIFIC SESSION B ADULT CARDIAC SURGERY

Ballroom B, Hynes Convention Center

Moderators: Bruce W. Lytle, M.D.

Marko I. Turina, M.D.

29. MITRAL VALVE REPAIR FOR DEGENERATIVE DISEASE.

A. Marc Gillinov, M.D.*, Delos M. Cosgrove, M.D., John H. Arnold, M.D.*, Nicholas G. Smedira, M.D.*, Joseph F. Sabik, M.D.*, Patrick M. McCarthy, M.D., Craig R. Saunders, M.D.*, Bruce W. Lytle, M.D., Robert W. Stewart, M.D.* and Floyd D. Loop, M.D.

Cleveland, Ohio

Discussant: Lawrence H. Cohn, M.D., Boston, Massachusetts

Objective Degenerative mitral valve disease has become the most common cause of mitral insufficiency in the United States. Mitral valve repair is applicable to the majority of these patients and has become the procedure of choice. This study was undertaken to define the operative risk and to determine risk factors for late mortality and reoperation.

Patients and Methods From 1985 to 1996, 1,101 consecutive patients (pts) had primary isolated mitral valve repair for valvular insufficiency caused by degenerative disease. Mean age was 58 ±12 years; 64% were male. Preoperative degree of mitral regurgitation was 3.6 ± 0.9. Results There were 4 hospital deaths (0.4%). Long-term followup was available in 1,089 pts (99%), with 4,319 pt-years of follow-up available for analysis. Eight-year actuarial survival was 84% and 8-year freedom from cardiac death was 94%. Univariate analysis identified moderate or severe left ventricular dysfunction (p<0.001), older age (p<0.001), chordal shortening (p<0.001) and earlier year of surgery (p<0.001) and annulopasty alone (p = 0.001) as risk factors for late mortality. At 8 years, freedom from thromboembolism was 92%, from endocarditis 99%, from anticoagulant-related hemorrhage 99%, and from reoperation 94%. Of the 30 pts who required reoperation for late mitral valve dysfunction, 15 (50%) repairs failed secondary to progressive degenerative disease. Univariate analysis identified chordal shortening (p<0.02), earlier year of surgery (p<0.001), and annuloplasty alone (p<0.001) as risk factors for reoperation.

Conclusions For patients with degenerative mitral valve disease, valve repair is associated with (1) very low operative mortality rates, and (2) low rates of valve related morbidity and mortality. (3) Isolated annuloplasty and chordal shortening jeopardize late results.

*By invitation

30. IMPACT OF CORONARY ARTERY DISEASE AND AGE ON LATE SURVIVAL IN PATIENTS UNDERGOING BIOPROSTHETIC AORTIC VALVE REPLACEMENT.

Gideon Cohen, M.D.*, Tirone E. David, M.D., Christopher M. Feindel, M.D., Joan Ivano, MSc.*, Michael Borger, M.D.* and Susan Armstrong, MSc.*

Toronto, Ontario, Canada

Discussant: Robert W. Emery, M.D., Minneapolis, Minnesota

Prolonged lifespan in a progressively aging population will likely increase the incidence of aortic valve disease and the demand for aortic valve replacement (AYR) well into the new millennium. To ensure maximal event-free survival, age and lifestyle are routinely considered upon prosthetic valve selection (tissue vs. mechanical). The presence of concomitant coronary artery disease (CAD), however, may be overlooked in the decision-making process. METHODS: Data were prospectively collected on 670 patients undergoing bioprosthetic Hancock II AYR between 1982 and 1994. Follow-up was conducted during a four month period in 1996 by clinic visit or telephone interview. All events were verified. Longitudinal survival and freedom from reoperation were evaluated univariately by Kaplan-Meier analysis and multivariately by Cox regression.

RESULTS: Follow-up was 99.7% complete at 69 ± 40 mos (median 66 mos; range 0.1-168 mos) with 75% of patients having follow-up of at least 93 mos. Mean patient age was 65 ± 12 yrs (median: 68 yrs; range: 18-86 yrs). Survival was significantly different by Kaplan-Meier analysis for both age <65 (71 ± 4% at 12 yrs) vs age > 65 (36 ± 7%, p = 0.0004) and no CAD (65 ± 4% at 12 yrs) vs CAD (35 ± 8%, p<0.0001). When both variables were combined, patients <65 years of age with no CAD demonstrated the best long-term survival, patients <65 years with CAD or patients ≥ 65 years without CAD demonstrated intermediate long-term survival, and patients ≥ 65 years with CAD demonstrated the poorest long-term survival (p<0.0001). After adjusting for gender, the Cox regression coefficient for age ≥ 65 was 0.62 ±0.17 (risk ratio 1.86) and for CAD, 0.50 ±0.16 (risk ratio 1.86). Only 24 patients required reoperation for primary tissue failure (PTF). CAD did not univariately nor multivariately influence the need for reoperation. At 12 yrs, freedom from reoperation for PTF was 83 ± 5% for those <65 yrs and 98 ± 1% for those ≥ 65 yrs (p = 0.006). Age ≥ 65 was associated with a reduced risk for reoperation: Cox regression coefficient -1.24 ± 0.5 (risk ratio 0.29).

CONCLUSIONS: Long-term survival following AYR is highly dependent upon age and the presence of concomitant CAD. Thus, patients greater than 65 years of age with CAD are ideal candidates for a bioprosthetic valve. In patients greater than 65 years without CAD, the benefit of a bioprosthesis is less clear. Age has an independent effect on the durability of the porcine bioprosthesis. Although older patients with CAD are unlikely to require reoperation for bioprosthetic failure, older patients without CAD and all younger patients may require reoperation for this purpose.

*By invitation

31. CRYOPRESERVED HOMOGRAFT VALVES IN THE PULMONARY POSITION - RISK ANALYSIS FOR LATE FAILURE.

†Christopher J. Knott-Craig, M.D.*, Kazuo Niwaya, M.D.*, Ronald C. Elkins, M.D, Mary M. Lane, Ph.D.*, K. Chandrasekaran, M.D.* and Kent E. Ward, M.D.*

Oklahoma City, Oklahoma

Discussant: Gordon K. Danielson, Jr., M.D., Rochester, Minnesota

To examine their durability, and factors associated with late failure, we reviewed our entire experience (1986-1997) with 369 hospital survivors in whom cryopreserved homografts (aortic n = 37, pulmonary n = 332) were used in the "pulmonary" position. Mean age was 18.5 ± 15.7 years (range 2 days -62 years). Operations included Ross operation (n = 273), Tetralogy of Fallot (n = 48), truncus arteriosus (n = 22), transposition of great arteries (n = 16), others (n=10). Median follow-up was 3.3 years (range 0.1-10.8 yrs) and recent echocardiographic evaluation was complete for 97% pts. Reoperation occurred in 7.3% (27/369), and failure (reoperation, gradient ≥ 40mmHg) in 12.7% (47/369) pts. Freedom from reoperation at 8 years was 83 ± 4%, and freedom from failure was 72±4%; for aortic versus pulmonary homografts, this was 38 ± 11% vs 77 ± 14% (p = 0.001). For pts <3 yrs at operation (n = 41) this was 34 ± 12% vs 75±5% for older pts (p = 0.0002). Multivariate analysis identified (i) younger age of homograft donors, (ii) aortic homografts, (iii) non-Ross operation, and (iv) later year of operation as risk factors for homograft failure (all p<0.005).

Homografts are the replacement conduit of choice for the pulmonary position. Pulmonary homografts are more durable than aortic homografts, and both fail earlier in young children. Residual pulmonary vascular abnormalities adversely impact the long-term durability of the conduit.

†1989-90 AATS Graham Fellow

*By invitation

32. A 20-YEAR EXPERIENCE WITH CARDIAC RETRANSPLANTATION.

Ranjit John, M.D.*, Jonathan Chen, M.D.*, Alan Weinberg, M.S.*, ‡Mehmet C. Oz, M.D., Silviu Itescu, M.D.*, Eric A. Rose, M.D. and Niloo M. Edwards, M.D.*

New York, New York

Discussant: William A. Baumgartner, M.D., Baltimore, Maryland

Retransplantation is the only currently available treatment for severe myocardial dysfunction in cardiac allograft recipients. However, the allocation of scarce organs to patients who have previously received a transplant is a controversial issue. We, retrospectively, reviewed our experience with patients undergoing cardiac retransplantation. Between 1977 - October 1997, 952 patients underwent cardiac transplantation for the treatment of end stage heart disease. Of these, 43 patients underwent retransplantation for cardiac failure resulting from transplant coronary artery disease, rejection and early graft failure. The mean age of these patients at primary transplantation was 37.2 years (range 2 -58 years, SD ± 16.7) and mean age at retransplantation was 41.6 years (range 5-64 years, SD ± 16.6). The interval between primary and retransplantation ranged from < 1 day to 11.4 years (mean 4.4 years). There was no significant difference in actuarial survival by Kaplan Meier analysis at 1, 2 and 5 years between patients undergoing primary and retransplantation -76%, 71% and 60% versus 66%, 66% and 45% respectively (p = 0.213). However, the actuarial survival of retransplant patients from the time of the first transplant was 91%, 88% and 74% at 1, 2 and 5 years respectively. A univariate analysis of various risk factors affecting outcome after cardiac retransplantation was performed. The variables evaluated ‘included the decade of transplantation, ages at primary and retransplantation, indications for primary and retransplantation, interval between transplants, HLA matching, recipient and donor demographic variables and the presence of pretransplant anti-HLA IgG and IgM antibodies. A shorter duration between transplants (p = 0.003) and older age at transplantation (p = 0.04) were identified as risk factors for poorer outcome after retransplantation. In conclusion, cardiac retransplantation can be performed with low morbidity and mortality and satisfactory long term outcome. However, patient characteristics and preoperative variables should assist in the rational application of retransplantation to ensure optimal utilization of donor organs.

4:35 p.m. EXECUTIVE SESSION (Limited to Members)

Ballroom B, Hynes Convention Center

6:30 p.m. MEMBER RECEPTION

Museum of Science

‡1994-96 AATS Robert E. Gross Research Scholar

*By invitation

1:45 p.m. SIMULTANEOUS SCIENTIFIC SESSION C GENERAL THORACIC SURGERY

Ballroom C, Hynes Convention Center

Moderators: Andre C.H. Duranceau, M.D.

Victor F. Trastek, M.D.

33. RESECTION MARGINS, N2 STATUS AND CELL TYPE DETERMINE SURVIVAL IN TRIMODALITY THERAPY OF MALIGNANT PLEURAL MESOTHELIOMA (MPM): RESULTS IN 183 PATIENTS.

David J. Sugarbaker, M.D., Raja M. Flores, M.D.*, Michael T. Jaklitsch, M.D.*, William G. Richards, Ph.D.*, Malcolm M. DeCamp, M.D.*, Scott J. Swanson, M.D.*, Raphael Bueno, M.D.* and Steven J. Mentzer, M.D.

Boston, Massachusetts

Discussant: Larry R. Kaiser, M.D., Philadelphia, Pennsylvania

Objectives: We review our experience with extrapleural pneumonectomy (EPP) in the multimodality management of MPM. Methods: From 1980 to 1997, 183 consecutive patients underwent trimodality therapy involving EPP followed by adjuvant chemotherapy and radiotherapy. Standardized systematic pathologic analysis was undertaken. Results: The cohort included 43 women and 140 men with a mean age of 51yrs (range 31- 76) and a median followup interval of 13 months. Overall survival was 36% at 2 yrs and 14% at 5 yrs (median 17 mo). There were seven perioperative deaths (3.8% mortality). Factors affecting long-term survival were evaluated in 176 patients surviving surgery (among these, survival was 38% at 2 yrs and 15% at 5 yrs; median 19 mo). As indicated in the table below, lack of N2 nodal involvement, negative resection margins and epithelial histology were associated with improved survival. Factorial grouping by N2 status and resection margins significantly stratified survival among all patients surviving surgery (p<0.02), and among those with epithelial histology (p<0.02). Thirty-one patients with epithelial tumors, negative resection margins and without N2 involvement had a 51 mo median survival (68% 2-yr. 46% 5-yr). A clinico-pathologic staging system previously published stratified survival (p< 0.05; see table).

Epithelial Mixed/Sarcom.	103 73	52% 16%	21% 0%	26 13	0.0001
Margins (+) Margins (-)	110 66	33% 44%	9% 25%	15 23	0.02
N2(+) N2(-)	40 136	23% 42%	0% 17%	14 21	0.004
Epithel. N2(+) Epithel. N2(-)	21 82	38% 56%	0% 24%	20 34	0.052
Epithel. N2(+) Marg(+) Epithel. N2(+) Marg(-) Epithel. N2(-) Marg(-) Epithel. N2(-) Marg(-)	12 9 51 31	29% 43% 49% 68%	0% 0% 14% 46%	14 22 22 51	0.013
Stage I Stage II Stage III	66 41 69	53% 44% 17%	20% 14% 14%	25 20 16	0.048

These data support the following conclusions: 1) Multimodality therapy including EPP is feasible in selected patients with MPM, 2) Microscopic resection margins affect long-term survival, pointing to the need for further investigation of local-regional control strategies, 3) Mediastinoscopy to evaluate N2 nodes is recommended, and 4) Patients with epithelial, margin-negative, N2-negative resection enjoy long-term survival.

*By invitation

34. INDUCTION CHEMOTHERAPY FOR T4 CENTRALLY LOCATED NON SMALL CELL LUNG CANCER.

Erino A. Rendina, M.D.*, Federico Venuta, M.D.*, Tiziano De Giacomo, M.D.*, A. Maria Ciccone, M.D.* Giorgio Furio Coloni, M.D.* and Costante Ricci, M.D.*

Rome, Italy

Sponsored by: G. Alexander Patterson, M.D., St. Louis, Missouri

Discussant: L. Penfield Faber, M.D., Chicago, Illinois

We employed induction chemotherapy in a prospective, single institution clinical trial intended to achieve resectability in patients with centrally located, irresectable T4 NSCLC. Other types of T4 (pleural effusion, N3) were excluded. Between January 1990 and April 1996 we enrolled 57 patients with histologically confirmed NSCLC. They all underwent Computed Tomography (CT), bronchoscopy, bone scan and in selected cases Magnetic Resonance (MR). Eligibility criteria for T4 were: Clinical [Superior Vena Cava Syndrome (9 patients), vocal cord paralysis (6 patients), dysphagia from esophageal involvement (1 patient)]; Radiological (CT and MR evidence of infiltration - 10 patients); Bronchoscopic (tracheal infiltration - 11 patients); Thoracoscopic (histologically proved mediastinal infiltration - 20 patients). Mediastinoscopy was employed in 38 patients. N2 was histologically confirmed in 40 patients. After 3 cycles of cisplatin (120 mg/m²), vinblastine (4 mg/m²), mitomycin (2 mg/m²) patients were reevaluated: 42 (73%) (36 males. 6 females: age range 42-75 years; mean 58 years) responded to therapy and underwent thoracotomy; 11 did not respond and 4 had major toxicity. Thirty-six patients (85%) had complete resection. We performed 4 exploratory thoracotomies, 6 pneumo-nectomies, 32 lobectomies (20 associated with reconstruction of hilar-mediastinal structures). Pathologically, 4 patients had T4, 6 had T3, 24 had T2, 3 had T1 and 5 had T0. Thirteen patients had N2, 15 had N1 and 14 had N0. Overall, 4 patients had no histological evidence of disease. We had 2 bronchopleural fistulas with 1 death, 5 other major complications and 9 cases of delayed lung reexpansion. Adjuvant chemo and/or radiotherapy was administered to N2 and N1 patients. After a follow up of 15 to 76 months (mean 26 months), 25 patients died of cancer and 2 died of unrelated causes: 14 are alive and free of disease and 1 is alive with disease. Survival at 1 and 4 years is 69% and 26% (median survival 25 months). In conclusion, of the initial group of 57 patients, 42 (73%) underwent exploration with a 4 year survival of 26%, and 36 (63%) had complete resection. Our data indicate that induction chemotherapy is effective for downstaging and surgical reconvertion of centrally located, irresectable T4 NSCLC. Survival is promising but it remains to be confirmed in larger series of patients and phase III trials.

*By invitation

35. THE PROGRESSION OF INTESTINAL METAPLASIA IN GASTROESOPHAGEAL REFLUX DISEASE.

Stefan Oberg, M.D.*, Jeffrey H. Peters, M.D.*, Peter F. Crookes, M.D.*, John J. Nigro, M.D.*, Jorg Theisen, M.D.*, Jeffrey A. Hagen, M.D.* and Tom R. DeMeester, M.D.

Los Angeles, California

Discussant: Victor F. Trastek, M.D., Rochester, Minnesota

It is currently suggested that Barrett's esophagus appears abruptly and once present does not increase in length. This widely quoted but unproven hypothesis is based on data from patients with traditional, long segment Barrett's esophagus with endstage gastroesophageal reflux disease (GERD). Increasingly it is recognized that shorter segments of intestinal metaplasia are part of the same underlying disease process. The aim of the present study was to relate physiologic abnormalities of GERD to increasing lengths of intestinal metaplasia.

One hundred thirty consecutive patients with symptoms suggestive of GERD and intestinal metaplasia of the cardia or esophagus on biopsy during upper endoscopy, were studied by esophageal motility and 24 hour pH and bilirubin monitoring. Duration of symptoms was documented. Standard measures of lower esophageal sphincter (LES) function, esophageal acid and bilirubin exposure, and lower esophageal peristaltic amplitude were assessed. Patients were divided into three groups based on the extent of intestinal metaplasia: Cardiac intestinal metaplasia (CIM) where the intestinal metaplasia was detected on biopsy in an endoscopically normal appearing cardia, short segment Barrett's Esophagus (SSBE) group with intestinal metaplasia in the tubular esophagus less than 3cm, and patients with long segment Barrett's Esophagus (LSBE).

	CIM (n = 30)	SSBE (n = 32)	LSBE (n = 68)
Duration of GERD symptoms (years)	10.0 (5-15)	10.0 (3-15)	17(10-23)*
% time esophageal pH < 4	6.5 (3.3-8.4)	10.0 (5.5-16.4)*	21.3(13.2-41.1)*
% time esophageal bilirubin abs. > 0.2	1.6 (0.0-12.8)	14.8 (1.7-25.4)	16.9 (4.3-30.5)*
LES resting pressure (mmHg)	6.6 (3.5-11.8)	5.2 (2.3-10.2)	3.2(1.2-7.4)†
LES abdominal length (cm)	1.0 (0.4-1.4)	0.6 (0.2-1.2)	0.2 (0.0-0.8)†
Prevalence of defective LES (%)	58.6	65.9	87.5*
Esophageal contraction amplitude (mmHg)	64.5 (39.2-98.5)	56.5 (38.8-83.7)	51.0 (35.0-74.0)*
Values expressed as medians (IQR) * p < 0.05 vs all other groups, †P < 0.05 vs CIM

Length of intestinal metaplasia was strongly correlated with the degree of esophageal acid exposure (r = 0.71, p = 0.0001), and inversely with the LES pressure (r = -0.26, p = 0.003). Patients with long segment Barrett's esophagus had a longer duration of symptoms than those with short segment or cardia metaplasia only. The LSBE group demonstrated the greatest degree of physiologic derangement and the CIM group the least.

The results of this study suggests that longer segments of IM are associated with longer duration of reflux symptoms and increased severity of reflux disease. Intestinal metaplasia at the cardia is characterized by the same physiologic abnormalities as traditional Barrett's esophagus, although of lesser degree. This raises the possibility that shorter segments of intestinal metaplasia may progress to longer segments as the antireflux mechanism deteriorates.

2:45 p.m. INTERMISSION - VISIT EXHIBITS

*By invitation

3:15 p.m. SIMULTANEOUS SCIENTIFIC SESSION C GENERAL THORACIC SURGERY

Ballroom C, Hynes Convention Center

Moderators: Andre C.H. Duranceau, M.D.

Victor F. Trastek, M.D.

36. PLEURAL SPACE IRRIGATION AND MODIFIED CLAGGETT PROCEDURE FOR THE TREATMENT OF EARLY POSTPNEUMONECTOMY EMPYEMA.

Farid Gharagozloo, M.D.*, Gregory D. Trachiotis, M.D.*, Andrew J. Wolfe, M.D.*, Kevin M. DuBree, P.A.C.* and James L. Cox, M.D.

Washington, DC

Discussant: Peter C. Pairolero, M.D., Rochester, Minnesota

The incidence of postpneumonectomy empyema is 5 - 10%. Approximately half of postpneumonectomy empyemas occur within 2 to 4 weeks of pneumonectomy. A bronchopleural fistula is found in more than 80% of the patients. The classic treatment of postpneumonectomy empyema includes parenteral antibiotics, drainage of the pleural space, removal of necrotic tissue, and open pleural packing for many weeks followed by obliteration of the empyema space with antibiotic fluid or muscle. This approach results in prolonged hospitalization, repeated operations, and significant morbidity. In a five year period we treated 22 patients with early postpneumonectomy empyema. All patients had a bronchopleural fistula. The patient profile and predisposing factors were: