1:30 p.m. SIMULTANEOUS SCIENTIFIC SESSION A -ADULT CARDIAC SURGERY

Constitution Hall, Metro Toronto Convention Centre

Moderators: Verdi DiSesa, M.D.

Marko I. Turina, M.D.

7. The Porcelain Aorta at Aortic Valve Replacement: Surgical Strategies and Results

Bruce W. Lytle, A. Marc Gillinov*, Vu Hoang*, Delos M. Cosgrove, Michael K. Banbury*, Patrick M. McCarthy, Gosta B. Pettersson*, Joseph F. Sabik*, Nicholas G. Smedira* and Eugene H. Blackstone, Cleveland, Ohio

Discussant: Nicholas T. Kouchoukos, M.D.

OBJECTIVE: Aortic valve replacement (AYR) in patients (pts) with severe ascending aortic atherosclerosis poses technical challenges. A "no-touch" technique including AVR under deep hypothermic circulatory arrest (HCA) has been advocated when dealing with the porcelain or unclampable aorta. The purpose of this study was to determine operative strategies and results of AVR in pts with a severely atherosclerotic ascending aorta that could not be safely cross-clamped.

METHODS: From 1/90 to 12/98, 4983 pts had aortic valve surgery; of these, 62 pts (1.2%) had a severely atherosclerotic ascending aorta and required HCA to facilitate AVR, and they form the study group. 40% had previous cardiac surgery and 13% had history of chest irradiation. Severe aortic atherosclerosis was recognized preoperatively in 50%.

RESULTS: All pts had HCA, but several different strategies were used to manage the ascending aorta (table). Overall mortality was 14%, and 10% of pts suffered strokes. Increasing NYHA functional class and impaired left ventricular function were risk factors for hospital mortality. Choice of operative technique did not influence pt outcome; however, no pt having ascending aortic replacement suffered a stroke.

CONCLUSIONS: AVR in pts with severe ascending aortic atherosclerosis is associated with increased operative morbidity and mortality. Complete AVR under HCA requires a prolonged period of HCA. Ascending aortic replacement is a preferred technique as it requires a short period of HCA and results in comparable mortality with a low risk of stroke.

	AVR under HCA	Aortic Endarterectomy Replacement	Ascending Aortic	Inspect and Cross-Clamp	Balloon Occlusion
Number	24	16	12	6	4
HCA (min)	53±20*	13±6	17±6	4±5	5±4
Stroke	17%	12%	0	0	0
Mortality	12%	25%	17%	0	0

*P<.001 vs. other techniques

*By Invitation

8. Valvular Heart Surgery in Patients with Previous Mediastinal Radiation Therapy

Nobohiro Handa*, Christopher G. A. McGregor*, Gordon K. Danielson, Richard C. Daly*, Joseph A. Dearani*, Charles J. Mullany, Thomas A. Orszulak, Hartzell V. Schaff, Kenton J. Zehr*, Paula J. Schomberg*, Betty J. Anderson* and Francisco J. Puga, Rochester, Minnesota

Discussant: R. Scott Mitchell, AID.

OBJECTIVE: To characterize the outcome of valvular heart surgery for patients with previous mediastinal radiation therapy from January 1976 through December 1998.

METHODS: The study consists of 60 patients (37 females, 23 males) with a mean age of 62±15 years (28 to 88 years). Valvular heart surgery performed included aortic valve replacement (n=26), mitral valve procedure (n=16), tricuspid valve procedure (n=6), and multiple valve procedure (n=12). Associated procedures included coronary bypass surgery (48%), pericardiectomy (12%), myectomy (5%), chest wall reconstruction (5%) and permanent pacemaker placement (2%).

RESULTS: Early mortality was 7 cases (12%). Early mortality in patients with constrictive pericarditis was 40%(4/10) compared with 6%(3/50) in patients without constrictive pericarditis. By univariate analysis, early mortality was associated with constrictive pericarditis (P=0.011), reduced preoperative ejection fraction (P=0.015) and longer cardiopulmonary bypass times (P=0.037). A total of 14 patients (23%) required PPM before (n=7), during(n=l), or early(n=6) after valvular heart surgery. Total follow-up was 199 patient-years. There were 19 late deaths (malignancy 7, heart failure 5, other cardiac 4, other non-cardiac 3). Survival rates free of all causes of death, late cardiac death and cardiac reoperation at 5 years for hospital survivors were 66±8%, 82±7% and 93±4%, respectively. By univariate analysis, late cardiac death was associated with low ejection fraction (P=0.002), NYHA class IV(P=0.004), preoperative congestive heart failure(P=0.02), and preoperative atrial fibrillation(P=0.038). Eighty-five percent of the discharged patients were in NYHA class I or II at follow-up.

CONCLUSIONS: Early results of valve replacement after mediastinal radiation therapy were good except in the presence of constrictive pericarditis. Long-term outcome was limited by malignancy and heart failure. Early surgical intervention is recommended before the development of risk factors for late death, namely, severe symptoms, left ventricular dysfunction and atrial fibrillation.

*By Invitation

9. Late Results of Heart Valve Replacement with the Hancock II Bioprosthesis

Gideon Cohen*, Tirone E. David, Susan Armstrong* and Joan Ivanov*, Toronto, ON, Canada

Discussant: W.R. Eric Jamieson, M.D.

OBJECTIVE: The Hancock II bioprosthesis was recently approved by FDA for clinical use in the USA. This report describes the late clinical outcomes of patients who had AVR and MVR with this bioprosthesis.

METHODS: From 1982 to 1994, 670 pts had AVR and 310 had MVR with Hancock II bioprosthesis. Patients' mean age was 65±12 years for both groups. Most patients were in NYHA class III and TV and 41% of AVR group and 45% of MVR had coronary artery disease. Patients were followed prospectively at annual intervals. The mean follow-up was 87±45 months for AVR and 75+48 months for MVR, and it was 99% complete for both groups.

RESULTS: Table 1 shows the freedom from morbid events at 10 and 15 years. Patient's age and valve position were independent predictors of primary tissue failure. The freedom from primary tissue failure after AVR at 15 years was 72%±7% for patients <65 years of age and 99.6%±0.4% for pts >65 years of age whereas after MVR was 60%+9% for pts <65 years and 74%±9% for >65 years.

CONCLUSIONS: The Hancock II bioprosthesis has provided good clinical outcomes and it is a durable valve in older patients, particularly in the aortic position.

Table 1: Freedom from morbid events at 10 and 15 years
	AVR		MVR
Freedom from:	10 yr.	15 yr.	10 yr.	15 yr.
Death	61%±2%	47%±3%	52%±3%	30%±5%
Valve-related death	95%±1%	92%±2%	89%±1%	86%±3%
Cardiac-related death	80%±2%	72%±3%	73%±3%	47%±7%
Thromboembolism	87%±2%	83%±3%	89%±2%	87%±3%
Endocarditis	97%±1%	96%±1%	96%±1%	91%±4%
Primary tissue failure	97%±1%	81%±5%	86%±3%	66%±6%
Reoperation	94%±1%	77%±5%	85%±3%	69%±6%

*By Invitation

10. Mitral Valve Repair and Aortic Valve Replacement Is Superior to Double Valve Replacement

A. Marc Gillinov*, Eugene H. Blackstone, Delos M. Cosgrove, Paul Kerr*, Antonino Marullo*, Patrick M. McCarthy, Nicholas G. Smedira* and Bruce W. Lytle, Cleveland, Ohio

Discussant: Cary W. Akins, M.D.

OBJECTIVE: Double valve replacement has been advocated for patients with concomitant aortic and mitral valve disease. The purpose of this study was to determine if mitral valve repair is superior to mitral replacement.

METHODS: From 1975 to 1998,984 patients underwent double valve surgery. Of these, 819 had aortic valve replacement with either mitral valve replacement (n=518) or repair (n=301). Mitral valve pathology was rheumatic in 70% and degenerative in 20%. Mitral valve repair included commissurotomy in 131 (44% of repairs), ring annuloplasty in 170 (56%), leaflet resection in 27 (9%) and chordal procedures in 14 (5%). The prevalence of mitral valve repair increased from 25% in the 1970s to 50% in the 1990s. Mitral valve replacement was more common in pts with severe mitral stenosis P<.0001, atrial fibrillation P=.0009, and patients receiving a mechanical aortic prosthesis P=.0005. Mitral valve repair was more common in patients with annular dilatation P<.0001. These differences were used for propensity-matched multivariable comparisons. Follow-up extended to 22 years, mean 6.9+5.9 years, with 5199 patient-years of follow-up available for analysis.

RESULTS: Hospital mortality was 6.4%. It was similar for mitral valve repair (5.3%) and mitral valve replacement (7.0%) P=.4. Survival at 5,10, 15, and 20 years was 80%, 63%, 46%, and 31% after mitral valve repair vs. only 72%, 52%, 34%, and 21% after mitral valve replacement P=.006. Late mortality was increased by older age P<.0001, atrial fibrillation P=.009, and mitral valve replacement rather than repair P<.0001. After repair of non-rheumatic mitral valves, 5, 10, and 15-year freedom from valve replacement was 94%, 92%, and 90%. In contrast, after repair of rheumatic mitral valves, freedom from valve replacement at these intervals was 97%, 89%, and 73%.

CONCLUSIONS: Mitral valve repair is 1) feasible in a large proportion of patients with double valve disease, 2) improves late survival in patients with double valve disease, and 3) should be considered in all patients with double valve disease, including those with rheumatic mitral stenosis.

*By Invitation

11. Cardiac Surgery Combined with the Maze-Ill Procedure

James L. Cox, Niv Ad* and Terry Palazzo*, Washington, District of Columbia

Discussant: Hartzell V. Schaff, M.D.

OBJECTIVE: This study was designed to determine the efficacy of combining the Maze procedure with other types of cardiac surgical procedures.

METHODS: Between April 1992 and October 1999, we performed 301 Maze-Hi procedures. 180 patients underwent the Maze only, and 121 patients had the Maze plus other cardiac surgery, including valve surgery in 75 patients and non-valve cardiac surgery in 46 patients. Events within the first 3 months of surgery were considered perioperative . 263 patients were followed from 3 months to 7.5 years (Mean: 3.9 + 2.7 years) ( Late ).

RESULTS: See Table. The operative mortality rate for Maze plus mitral valve surgery was 2.5 % and the arrhythmia control was 98 % (n=40).

CONCLUSIONS: Perioperative mortality and morbidity are related directly to age > 65 years but not to cardiac surgery performed concomi-tantly with the Maze procedure. Atrial fibrillation is controlled in 98-99% of patients whether or not concomitant cardiac surgery is required.

Table 1
	Maze Only	Maze + Other	P
Perioperative Stroke Rate	0.4%	0%	NS
Overall Operative Mortality Rate	1.3%	5.0%	*
Operative Mortality Rate <65	0%	0.8%	NS
Operative Mortality Rate >65	10.3%	10.6%	NS
Late Stroke Rate	0.6%	0%	NS
Late Mortality Rate	1.3%	3.0%	NS
Arrhythmia Control	99%	98%	NS

*p<0.05 by univariant analysis; p=0.06 by multivariant analysis

3:15 p.m. INTERMISSION - VISIT EXHIBITS

*By Invitation

4:00 p.m. SIMULTANEOUS SCIENTIFIC SESSION A ADULT CARDIAC SURGERY

Constitution Hall, Metro Toronto Convention Centre

Moderators: Verdi DiSesa, M.D.

Marko I. Turina, M.D.

12. Optimal Surgical Management of Mitral Regurgitation from Anterior Leaflet Prolapse

Ian A. Nicholson*, Lawrence H. Cohn, Gregory S. Couper and David H. Adams, Boston, Massachusetts

Discussant: Ottavio Alfieri, M.D.

OBJECTIVE: Anterior leaflet prolapse of the mitral valve remains a challenge in mitral valve repair for myxomatous degeneration. We reviewed 173 patients undergoing MV repair for anterior leaflet prolapse to determine the most durable operative method.

METHODS: One hundred and seventy three patients (114 males , 59 females) underwent mitral valve repair between 1984 and 1999. Mean age at operation was 59 years. Patients underwent either chordal shortening and/or anterior leaflet resection (Group 1, N= 100) or Gortex chordoplasty {2-4 mattress sutures of C5 Gortex to anterior leaflet}(Group 2 , N= 73). The mean follow-up was 3.1 years in Group 1 and 2.25 years in group 2 .

RESULTS: Cardiopulmonary bypass and aortic cross clamp times were similar in the two groups. The incidence of concomitant CAD requiring C ABG was 26% in Group 1 and 33% in Group 2 . Operative death rate was 3% for Group 1 vs 1.4% for Group 2 [P = N.S.]. Late deaths were 4 (4%) in Groupl and 3 (4.1%) in Group 2 [P= N.S] Re-operation for structural valve degeneration occurred in 15/100 (15%) in Group 1 and only 4/73 (5.4%) in Group 2 [ P< 0.04].

CONCLUSIONS: Gortex chordoplasty is a more reproducible technique for anterior leaflet prolapse repair with a much lower reoperation rate for failed repair.

*By Invitation

13. Increased Mortality of Aortic Valve Re-Replacement Is Not Due to Aortic Valve Reoperation

Terrence M. Yau*, Joan Ivanov* and Tirone E. David, Toronto, ON, Canada

Discussant: Thomas Orszulak, M.D.

OBJECTIVE: We quantified the contribution of redo AV surgery itself to the mortality of AV re-replacement.

METHODS: Predictors of early outcomes and the effect of reoperation were determined by logistic regression in 1881 patients undergoing AV surgery from 1990-1998.

RESULTS: Patients undergoing redo AV surgery (N=205, 11%) were younger, more likely to require urgent surgery, to have heart failure, endocarditis, and AI or mixed AS/AI than primary patients (all p=0.001), but less likely to have diabetes (p=0.003) or coronary disease (p=0.001). NYHA class, LV function, BSA, valve size and crossclamp times were not different. Annular enlargements were more common in redo procedures (23% vs. 34%, p=0.0002). Mortality (2.3% vs. 4.4%, p=0.07) and stroke (2.2% vs. 4.9%, p=0.02) were greater in redos, but MI, low output syndrome and IABP use were similar. Redo AV surgery itself carried only a slightly increased odds ratio for mortality (Table) compared to other risk factors; the mortality of elective re-replacement (1.7%) was similar to that of primary surgery (1.5%) (p=0.8).

CONCLUSIONS: The risk of AV re-replacement is due mostly to endocarditis or shock, annular enlargement, and comorbidity, rather than the requirement for AV re-replacement itself. This data supports primary implantation of bioprosthetic AVs in young patients to avoid anticoagulation and its complications, as elective reoperation for primary tissue failure is associated with low risk.

Independent Predictors of Mortality	Odds Ratio	95% Cl
Age	1.04	1.01-1.07
PVD	3.76	1.77-7.99
Shock	4.37	1.71-11.2
Active endocarditis	4.87	1.67-14.2
Annular enlargement	2.18	1.17-4.05
Redo AV surgery	1.55	1.02-2.33

*By Invitation

14. Heparinless Cardiopulmonary Bypass for Repair of Aortic Trauma

Stephen W. Downing*, Marcelo G. Cardarelli*, Safuh Attar, Douglas C. Wallace*, Aurelio Rodriguez*, Joseph S. McLaughlin, Jamie Brown* and Glenn J. R. Whitman, Baltimore, Maryland

Discussant: Irving L. Kron, M.D.

OBJECTIVE: Distal circulatory support for the repair of traumatic rupture (TR) of the aorta reduces paraplegia. However, standard cardiopul-monary bypass (CPB) requires heparin and may increase bleeding and death. Left atrial to aortic bypass eliminates heparin, but cannot heat, cool, oxygenate or rapidly add volume. We hypothesized that a heparin-bonded CPB system would be simple, effective, and free of these shortcomings.

METHODS: A retrospective review over a 5 year period at a regional level I trauma center. A heparin-bonded bypass system was utilized consisting of a 19 or 21 French femoral vein (right atrial) line, an oxygenator-heater/ cooler and a centrifugal pump flowing at 3-5 L per minute. Arterial return was to the femoral artery or distal aorta. No systemic heparin was given.

RESULTS: From 7/6/94 to 9/8/99, 54 patients underwent repair of a TR. Two patients repaired with simple clamping, 2 patients already on ECMO and 1 patient who exsanguinated at thoracotomy were excluded. The mean age was 43 ± 17 years. 14% were hypotensive, 16% had intracranial injuries, 37% had pelvic injuries, 63% had abdominal injuries and 24% had pulmonary contusions. The cross clamp time was 32 ± 11 minutes and bypass time was 64 ± 44 minutes. In the first 15 patients the femoral artery and vein were cannulated in radiology after angiography. There was one femoral artery and one femoral vein injury with one limb loss and this procedure was discontinued. The subsequent 34 patients had percutaneous femoral vein and direct distal aortic cannulation without event. The mortality rate was 10%. One death was intraoperarive due to arrythmia, the remainder were due to other injuries. There was no new paraplegia and no worsening of neurologic or pulmonary injuries.

CONCLUSIONS: This approach has advantages over standard CPB and left atrial to aortic bypass including simple cannulation without intrapericardial or hilar dissection, avoidance of anticoagulation; and the ability to easily treat hypothermia, hypoxia and hypovolemia. The mortality rate was below published averages and paraplegia effectively prevented.

*By Invitation

15. Extracorporeal Membrane Oxygenation in 242 Adults: Survival at 1 Year

Nader Moazami*, Nicholas G. Smedira*, Patrick M. McCarthy, Camille M. Golding*, Bruce W. Lytle, Eugene H. Blackstone and Delos M. Cosgrove, Cleveland, Ohio

Discussant: Robert H. Bartlett, M.D.

OBJECTIVE: To define the survival and the changing role of ECMO in a diverse population of patients in the modern era of LVAD support and thoracic transplantation.

METHODS: Retrospective review of 242 adult patients with a mean age of 53±14 years who were placed on ECMO support from 1992-1999. Indications were post-cardiotomy (119), myocardial infarction (35), ARDS (23), cardiac arrest (11), decompensated heart failure (31), deterioration during cardiac cath-eterization (6), and after cardiac (10) and lung (7) transplantation.

RESULTS: Veno-arterial support was employed in 209 patients. In this group, 54 (26%) were bridged to LVAD or heart transplantation, 80 (38%) were weaned, and 75 (36%) died on ECMO. Overall, 68 (33%) were discharged; 30 (55%) in the LVAD versus 38 (48%) in the weaned group (p£0.2). At 1 year, 51 (75%) patients were alive. Veno-venous support was used in 33 patients, 18 with ARDS, 5 post-lung transplantation, 8 post-cardiotomy, and 1 after heart transplantation and acute MI. Overall, 17 patients died on ECMO, 16 (48%) were weaned and 11 (69%) were discharged home; 8 (72%) were alive at 1 year. In specific subgroups, survival to discharge varied from 66% in the post-catheterization to £20% in the heart transplant and cardiac arrest groups (p£0.001). Mortality was associated with severe neurologic deficit in 19 (11%), irreversible myocardial damage in 63 (39%) non-transplant candidates, and multi-system organ failure in 81 patients (50%).

CONCLUSIONS: ECMO can be used in a large and diversified setting of cardiopulmonary collapse. Although overall mortality remains high, 1-year survival after discharge is excellent. Use of veno-arterial ECMO support as bridge to LVAD implantation allows survival of a large number of patients who would otherwise die.

*By Invitation

1:45 p.m. SIMULTANEOUS SCIENTIFIC SESSION B

GENERAL THORACIC SURGERYRoom 205

Moderators: Thomas W. Rice, M.D.

Joseph I. Miller, Jr., M.D.

16 Detection of Early Lung Cancer. CT Scan or Chest X-Ray? Survival Implications

Nasser K. Altorki, Michael S. Kent*, David Yankelevitz*, Claudia Henschke*, Daniel Libby*, Mark Pasmantier* and James P. Smith, New York, New York

Discussant: Joe Friedberg, M.D.

OBJECTIVE: It has been recently proposed that chest CT scans may be a useful method for early detection of lung cancer. In this study we determined the stage distribution of lung cancers detected by a screening CT scan. This was compared with the stage distribution of patients whose lung cancers were detected by a routine chest X-ray (CXR).

METHODS: Two groups of patients were reviewed. Twenty patients had biopsy-proven non-small cell lung cancer detected through a CT scan screening program. A second group of patients (n=103)had their lung cancers detected on routine CXR. Patients with pulmonary symptoms or prior history of cancer were excluded.

RESULTS: There was no difference in age, gender or cell-type distribution between the two groups. Stage distribution is shown in the following table. There was no difference between the groups in the overall prevalence of Stage I disease versus more advanced disease. However, a significantly greater number of patients were stage IA in the CT group compared to the CXR group (p=.004). Of 15 patients with Stage I disease in the CT group, 7 had tumors 1 cm. or less versus 8 out of 74 stage I patients detected by CXR.

CONCLUSIONS: As a screening modality for lung cancer CT scan yields a higher incidence of Stage LA disease than that achievable by a CXR. This may result in significant imporoval of survival in patients with Stage I disease.

TNM stage	CXR(n=103)	CT scan (n=20)
IA	41 (40%)	15(75%)
IB	33 (32%)	1 (6%)
IIA	6 (6%)	1 (6%)
IIB	15(15%)	0
IIIA	1 (1%)	2(12%)
IIIB		1 (6%)

*By Invitation

17. Subcentimeter Non-Small Cell Lung Cancer a Program for Detection and Resection Is Warranted

Scott J. Swanson*, Raphael Bueno*, Michael T. Jaklitsch*, Steven J. Mentzer, Jeanne M. Lukanich* and David J. Sugarbaker, Boston, Massachusetts

Discussant: Joel D. Cooper, M.D.

OBJECTIVE: For lung cancer screening to have a favorable impact, survival of patients whose tumor is detected when relatively small should be superior to that of patients with larger tumors. To look at this, we examined the survival of patients who had a resection of non-small cell lung cancer that was less than or equal to 1 centimeter.

METHODS: From 1990-1998, 182 patients had malignant solitary lung nodules less than or equal to 1 centimeter resected at the Brigham and Women s Hospital. Of these, 40 patients had primary non-small cell lung cancer (node-negative or indeterminate). Preoperative, perioperative and follow up data were recorded in our prospective thoracic database. Survival was performed by Kaplan-Meier lifetable analysis.

RESULTS: 27 women and 13 men (37/40 with smoking history), median age 64 years (46-86) underwent 9 anatomical (lobes/segments), 8 wedge resection with node sampling and 23 wedge resection without node sampling. Median tumor size = 0.8 cm (0.2-1.0). Histologically, there were 34 adenocarcinomas, 5 squamous cell and 1 undifferentiated carcinoma. There was no perioperative mortality. Median length of stay was 4 days (1-15). Follow up is complete. Five-year survival is 88%. Median survival has not been reached at a median follow up of 3.3 years. Type of resection was not statistically significant (logrank, p = 0.43) although there were no recurrences or late deaths in the anatomic resection subgroup. This may reflect more accurate staging.

CONCLUSIONS: Long-term survival following resection for subcentimeter non-small lung cancers appears better than that for overall stage I lung cancer. These data support aggressive screening and surgical strategies for small non-small cell lung cancers. The use of helical CT scanning for early lung cancer, as recently reported, may be the screening method of choice.

*By Invitation

18. CALGB 9335: a Multi-Center Phase-II Prospective Study of Video-Assisted Wedge Resection Followed by Radiotherapy for T1NO NSCLC in High-Risk Patients; Preliminary Analysis of Technical Outcome

Hani Shennib, Leslie Kohman, James E. Herndon*, Jeffrey Bogart*, David J. Sugarbaker, Mark Green* and Robert Keenan, Montreal, Canada; Syracuse, New York; Durham, North Carolina; Boston, Massachusetts; Charleston, South Carolina; Pittsburgh, Pennsylvania

Discussant: Robert J. Ginsberg, M.D.

OBJECTIVE: Video-assisted technology may offer advantages, not yet proven, in cardiothoracic surgery. The objective of this NIH sponsored phase-II prospective multicenter trial was to determine the feasibility of treating patients with cardiopulmonary dysfunction and T1 peripheral non-small cell lung cancer by video-assisted wedge resection and local (56Gy)radiotherapy. High-risk patients had one or more of the following: FEV₁ <40%predicted, DLCO <50% Vo₂max<15ml/Kg/min, use of Supplemental oxygen,and Pa Co₂ >45mmHg.

METHODS: Between September 1995 and September 1999, 65 patients were accrued of which 60 were eligible [50%male, median age 69YJ. Technical failure occured in 15 patients [25%]. These included conversion to open thoracotomy in 9 patients, abortion of the operation in 2 patients, 1 postoperative death and 3 patients with postoperative positive resection margins. Postoperative staging was raised to T2 in 6 patients [10%] and benign in another 6 patients [10%]. Other complciations included prolonged air leak 10%, pneumonia 6%,respiratory failure 4% arrhythmia 6%. Resection was by staplers except 6 patients by cautery and 1 patient by laser. Adhesions were absent in 48%, minimal in 21% and moderated to extensive in 28%. 39/48 patients had VATS accessible intrathoracic lymph nodes. Margins were >lcm in [45%] and l patients had microscopic positive resection margins. Minimal intraoperative bleeding occured in 22 patients and moderate in 1 patient. Median duration of the procedure was 160 min [40-255min]. Only 22 patients proceeded to radiotherapy.

CONCLUSIONS: We conculde that VATS wedge resection is feasible and relatively safe in the majority of patients with poor cardiopulmonary status but there is a substantial incidence of conversion to thoracotomy and positive resection margins. Long-term local control with this method is as yet unknown, but the low incidence of successful completion of radiotherapy indicates that this approach may not be feasible.

*By Invitation

19. Factors Affecting Early Morbidity and Mortality After Pneumonectomy for Malignant Disease

Alain Bernard*, Claude Deschamps, Mark S. Allen, Daniel L. Miller*, Victor F. Trastek, Greg D. Jenkins*, and Peter C. Pairolero, Rochester, Minnesota

Discussant: Malcolm M. DeCamp, Jr., M.D.

OBJECTIVE: Pneumonectomy may be associated with significant morbidity and mortality with little information existing as to the factors involved.

METHODS: From January 1985 to September 1998, 639 consecutive pts (469 males and 170 females) underwent pneumonectomy for malignancy. Median age was 64 years (range, 20 to 86 yrs). Indication for resection was primary malignancy in 607 pts (95.0%) and metastatic disease in 32 (5.0%). Forty-nine pts (7.7%) underwent completion pneumonectomy. Factors affecting in-hospital morbidity and mortality were analyzed using univariate and multivariate analysis.

RESULTS: Cardiopulmonary complications occurred in 245 pts (morbidity, 38.3%; 95% CI, 34.6 to 42.2%). Univariate analysis demonstrated that factors adversely affecting morbidity included increasing age (p < 0.01), male gender (p = 0.04), associated respiratory (p = 0.02) or cardiovascular disease (p < 0.01), amount of cigarette smoking (p = 0.02), preoperative radiation (p = 0.02), muscle reinforcement of bronchial stump (p < 0.001), and amount of blood transfused (p = 0.01). Factors adversely affecting morbidity with multivariate analysis included increasing age (p < 0.001), associated cardiovascular disease (p = 0.001) and muscle reinforcement of bronchial stump (p < 0.001). There were 43 deaths (mortality, 6.7%; 95% CI, 4.9 to 9.0%). Mortality was 6.6% (n = 40) for primary malignancy and 9.4% (n = 3) for metastatic disease. Factors adversely affecting mortality with univariate analysis included associated cardiovascular (p = 0.05) or hematologic disease (p = 0.03), preoperative chemotherapy (p = 0.01) or radiation (p = 0.04), muscle reinforcement of bronchial stump (p = 0.03), extended resection, (p = 0.02), and decreased DLCO (p < 0.01, N = 388). Factors affecting mortality with multivariate analysis included associated cardiovascular (p = 0.04) hematologic disease (p = 0.01), and lower body mass index (p = 0.01).

CONCLUSIONS: Multiple factors adversely affect morbidity and mortality after pneumonectomy. Appropriate selection and meticulous perioperative care are paramount to minimize risks in pts who require pneumonectomy for cure.

3:05 p.m. INTERMISSION - VISIT EXHIBITS

*By Invitation

3:50 p.m. SIMULTANEOUS SCIENTIFIC SESSION B

GENERAL THORACIC SURGERYRoom 205

Moderators: Thomas W. Rice, M.D.

Joseph I. Miller, Jr., M.D.

20. Experience with Pulmonary Resection from Gynecologic Malignancies

John J. McMahon*, Chukwumere E. Nwogu*, Mathew W. Pombo*, M. Steven Fiver*, Shashikant B. Lele*, Deborah L. Driscoll* and Timothy M. Anderson*, Buffalo, New York

Discussant: Michael Maddaus, M.D.

OBJECTIVE: Pulmonary metastases from cervical and uterine primaries are uncommon. Although thoracotomy for removal of isolated pulmonary metastasis is well documented in a wide variety of solid rumors, there is a paucity of data regarding the optimal managment of patients with gynecologic malignancies metastatic to lung. We have analyzed a single institution experience in an attempt to clarify the role of metastasectomy for uterine and cervical cancers.

METHODS: We retrospectively reviewed the Roswell Park Cancer Institute experience between 1982 and 1999 of eighty-two patients with gynecologic tumors metastatic to lung, including 25 who underwent pulmonary resection.

RESULTS: Among 82 patients there were 60 uterine and 22 cervical primaries. Nineteen patients with uterine and 6 with cervical origin underwent pulmonary resection for lung metastases. Median survival for the combined surgery group (n=25) was 65 months compared to 32 months for the combined non-surgical group (n=57, p=0.04). Median time from lung metastasis until death or last follow-up was 30 months in the surgical group compared to 10 months in the non-surgical group (p=0.01). Among patients with uterine primaries undergoing metastasectomy (n=19) median survival was 67 months compared to 37 months for the non-surgical uterine group. Median time from lung metastases until death or last follow-up was 26 months in the uterine surgical group compared to 13 months in the non-surgical group. Uterine leiomyosarcomas tended to have a worse prognosis than other uterine pathologies. Among patients with cervical primaries undergoing surgery (n=6), median survival was 65 months compared to 23 months in the non-surgical group (n=16, p=0.03). Median time from lung metastases until death or last follow-up in the surgical group was 36 months, compared to 6 months in the non-surgical group (p=0.003).

CONCLUSIONS: Pulmonary resection provides a survival advantage in patients with uterine and cervical metastases isolated to lung. Furthermore, there appears to be a greater disease-specific survival advantage in patients undergoing lung resections for metastases from cervical origin compared to those of uterine derivation.

*By Invitatio

21. Early Results of Isolated Single Lung Perfusion for Treatment of Unresectable Sarcomatous Metastases

Joe B. Putnam, Jr., Robert S. Benjamin*, Soo J. Rha*, Garrett L. Walsh*, Stephen G. Swisher*, Ara A. Vaporciyan*, W. Roy Smythe* and Jack A. Roth, Houston, Texas

Discussant: Robert Downey, M.D.

OBJECTIVE: Despite resection and chemotherapy, patients with sarcoma-tous pulmonary metastases (PM) frequently progress to respiratory insufficiency and death. We examined the role of isolated single lung perfusion (ISLP) with adriamycin for patients with (1) unresectable sarcomatous PM, (2) absence of other more effective chemotherapy, and FEV1 of 0.8 liters in the con-tralateral lung.

METHODS: 15 patients, who were entered onto a Phase I study, were treated with ISLP between January, 1995 and April, 1999. Ipsilateral pulmonary artery and veins were isolated, clamped, cannulated, and perfused. ISLP was performed over 20 minutes with adriamycin in a buffered crystalloid solution at 60mg/m²,200 mg/1 (Group 1, n=7); 75 mg/m²' 250 mg/1 (Group 2, n=4), or 60mg/m2,100 mg/1 (Group 3, n=4). Adriamycin levels were determined for lung, tumor, and serum. Actuarial survival was calculated.

RESULTS: No intraoperative complications occurred. Higher drug levels were obtained in lung tissues (median 125 mcg/g tissue, range 9.4 -193 mcg/g) compared to tumor (median 58 mcg/g tissue, range 9.5 -117 meg/ g). Serum drug levels were negligible. Two patients developed Grade IV pulmonary toxicity (Group II). Operative mortality was 20% (3/15): 1, paradoxical tumor embolus (Group 1); 1, acute lung injury (Group II), and 1, pneumonia 3 weeks postop (Group 3). Late toxicity included 40% decrease in ventilation and perfusion to the treated lung. Two international patients were lost to follow-up. Five of ten evaluable patients had regression or stabilization of PM compared to PM in the untreated lung. All other patients had continuous growth of PM. Actuarial median survival was 19.1 months. Four patients remain alive greater than 2 years after ISLP.

CONCLUSIONS: ISLP may be performed safely at a dose of 60 mg/m2 (200mg/l or 100 mg/1). ISLP minimizes systemic chemotherapy toxicity, achieves high drug levels in tissue, and is associated with prolonged survival in patients with isolated unresectable sarcomatous PM.

*By Invitation

22. Is There Ever a Role for Salvage Operation in Malignant Pleural Mesothelioma

Tarek M. Aziz*, Scott Queen*, Hosney Yosef* and Dhurv Prakash*, Glasgow, United Kingdom

Discussant: L. Penfield Faber, M.D.

OBJECTIVE: We analyzed our experience in the period January 1989-December 1998 aiming to confirm the role of surgery in the multimodality treatment of malignant pleural mesothelioma

METHODS: 109 patients were diagnosed as malignant pleural mesothelioma. The median age was 62 years (range 46-73). Apre-operative tissuse diagnosis was confirmed in all patients by open-pleural biopsy. The surgical procedures included palliative pleurectomy in 18 patients, radical pleuropneumonectomies in 63 patients. Radical surgical treatment was only considered if the patient is generally fit, and the tumour was confined to the hemithorax and there was no mediastinal invasion. Post operative chemotherapy (carboplatin + epirubicin) was used the majority of patients who underwent radical surgery ( except the first 13 patients)

RESULTS: The operative mortality was 8.9%. The median follow up is 42 months ( range 2-87). The median survival for palliative therapy was 8 months compared to a median survival of 38 months for patients who underwent radical surgery + post-operative chemotherapy (p=0.02). However, the median survival for those who did not have post operative chemotherapy following their radical surgery was poor (13 months).Thirty four patients were still alive at 30 months following their radical surgery + chemotherapy and 21 of them being disease-free. The main factors affecting the results is the number and development of metastasis following surgery.

CONCLUSIONS: Radical surgery and adjuvant chemotherapy might represent an effective form of treatment in selected malignant pleural mesothelioma. We advocate general radical pleuro-pneumopnectomy for malignant mesothelioma if it is part of multi-modal theraputic protocol.

*By Invitation

23. A Single-Institution, Multidisciplinary Approach to Primary Sarcomas Involving the Chest Wall Requiring Full Thickness Resections

Garrett L. Walsh*, Bryan M. Davis*, Stephen G. Swisher*, Ara A. Vaporciyan*, W. Roy Smythe*, Jack A. Roth and Joe B. Putnam, Jr.*, Houston, Texas

Discussant: Mark S. Allen, M.D.

OBJECTIVE: Primary sarcomas involving the chest wall (PSCW) requiring full thickness excision are rare. We reviewed our experience with these lesions in a tertiary referral cancer center using multidisciplinary approaches.

METHODS: A 10 year retrospective study identified 51 patients referred with PSCW; 38 for initial treatment (I) and 13 after previous failed surgical excisions elsewhere (Recur). Presenting symptoms were pain alone 23/51 (45%), pain with an associated mass 8/51 (16%) and an asymptomatic mass alone 12/51 (24%). Median symptom duration was 258 days in the primary group and 184 days in the recurrent group. Tumor locations were sternal (n=11), rib alone (n=36) and posterior rib with extension into vertebral bodies (n=4). Histologies included: chondrosarcomas (15), malignant fibrous histiocytomas (7), osteosarcoma (4), Swing's (1), desmoids (7) and other histologies (17). The median tumor volume of those presenting initially were 509 cm³ compared to 131 cm³ in patients with recurrent lesions.

RESULTS: 24/51 patients (47%) received treatment prior to resection including: chemotherapy alone (20), radiation alone (3) and combined chemo/XRT (1). The complete sternum was removed in 6/11 and the average rib resections required was 3.9. Four patients had vertebral body resections. Prosthetic meshes were required in 16/51 and mesh with methylmethacrylate in 18/51. Muscle flap reconstructions by plastic surgery were required in 24 patients. Negative margins were obtained in 47/51. There were no perioperative deaths with morbidities occurring in 12/51 (24%) [wound (3), prolonged air leak (1), prolonged ventilator requirement (1), arrhythmias (2), Adriamycin induced cardiomyopathy (1) and other (4)]. Post-operative treatment was administered to 13 patients [chemo alone (9) and chemo/XRT (4)]. The cumulative five-year survival of all patients was 65% [67.4% (I) and 55.4% (Recur)]. The average follow-up is 35.3 months.

CONCLUSIONS: A combined aggressive multidisciplinary approach to PSCW resulted in no treatment-related deaths and a prolonged survival in both (I) and (Recur) patient subsets.

*By Invitation

24. Chest Wall Invasion in Non-Small Cell Lung Carcinoma.

Microscopically Negative Margins Represent the Rationale for En-Bloc Resection.

Francesco Facciolo*, Giuseppe Cardillo*, Michele Lopergolo*, Guido Pallone*, and Massimo Martelli*, Rome, Italy

Discussant: Valerie W. Rusch, M.D.

OBJECTIVE: Intraoperative assessment of chest wall invasion represents a challenge either for thoracic surgeon or for pathologist. Most surgeons do extrapleural dissection until they do no find clear evidence of chest wall invasion. According to such criteria the number of incomplete resection is high and the prognosis of these patients is very poor. The aim of the present study is to evaluate the need for en-bloc resection in NSCLC invading the chest wall.

METHODS: Beteween January 1990 and December 1998, out of 1621 major pulmonary resections for lung carcinoma performed at our Institution, 97 (6%) patients with NSCLC invading parietal pleura or chest wall underwent en-bloc resection of the chest wall and lung plus radical mediastinal lymphadenectomy. Indications for chest wall resection were: CT or MRI evidence of chest wall invasion intraoperatively confirmed by parietal pleura attachment. No attempt to extrapleural dissection has been performed in our series. Five of our patients underwent preoperative induction therapy because of an N2 status. Seventy-nine patients underwent adjuvant therapy .

RESULTS: All patients underwent RO radical resections with microscopically negative margins. The pathologic depth of invasion was into the pleura alone in 28 (28.9%), into the pleura and soft tissue in 31 (32%), and into the pleura, soft tissue and bone in 38(39.1%). No 30-day mortality was reported. Major complications occurred in 12 (12.4%)patients. Eighty-nine of our 97 patients were included in the follow-up program (median: 27 months; range:9-96 months). The overall 5-year Kaplan-Meier estimated survival was 52%. The 5-year survival of patients with T3N0M0 disease was 46.8%(71 cases), T3N1M0 disease 100%(6 cases), and T3N1M0 disease 18.8%(12 cases).No locoregional recurrence was reported.

CONCLUSIONS: In patients with NSCLC invading chest wall, a complete (R0) resection can only be achieved with en-bloc resection of the chest wall and lung. The impressive 0% of locoregional recurrence justifies our aggressive approach.The long-term survival appears to be very appealing in T3N0 and T3N1 patients.

*By Invitation

1:45 p.m. SIMULTANEOUS SCIENTIFIC SESSION C

CONGENITAL HEART DISEASE Room 201

Moderators: John E. Mayer, M.D.

Roger B. B. Mee, M.D.

25. Is Modified Ultrafiltration Truly Superior to Conventional Ultrafiltration for Hemoconcentration After Pediatric Cardiac Surgery?

LeNardo D. Thompson*, Doff B. McElhinney*, Pauline Findlay*, Wanda Miller-Hance*, Mark J. Chen*, Maiko Minami*, V. Mohan Reddy*, Andrew J. Parry* and Frank L. Hanley, San Francisco, California

Discussant: William Gaynor, M.D.

OBJECTIVE: Although several studies have shown MUF to be better than CUF for minimizing the consequences of hemodilution after cardiac surgery with cardiopulmonary bypass (CPB) in children, any such benefit may be due to the volume of fluid removed. We conducted a randomized study to test the hypothesis that MUF and CUF have similar efficacy when a standardized volume of fluid is removed.

METHODS: From 10/98-9/99,110 children £15 kg were randomized to MUF (43) or CUF (67) for hemoconcentration after cardiac surgery with CPB. MUF was administered after CPB and CUF during rewarming, using a Hemocor HPH 400 filtration system. UF flow and suction rates were equal, and the volume of fluid removed was standardized as a percentage of effective volume (EV) added (the sum of prime volume and volume added during and after CPB, less urine output). In pts <10 kg, 50% of EV was removed, while 60% of EV was removed in pts between 10-15 kg. Hemoglobin, hemodynamics, and shortening fraction were measured before CPB, and 10 min and 1 hr after UF.

RESULTS: Median age was 6 mo (1 d-5 yr) and median weight was 6 kg (2-15 kg). Median duration of CPB was 109 min (32-313 min). Median pre-and postoperative hematocrit levels were 35% (20-49%) and 36% (25-53%), respectively. There were no significant differences between pts assigned to MUF or CUF in age, weight, or duration of CPB. Median UF duration was 10 min (3-25 min) and did not differ between groups. Median volume of UF effluent was greater in pts receiving CUF than MUF (95163 vs 68128 mL/kg, p=0.01). The total volume of blood products received during and after CPB was greater in CUF pts (129178 vs 102148 mL/kg, p=0.05). By repeated measures ANOVA, pts receiving MUF and CUF did not differ with respect to hematocrit (p=0.87), mean arterial pressure (p=0.85), heart rate (p=0.43), or left ventricular shortening fraction (p=0.21) from pre-CPB to 10 min and 1 hr post-UF.

CONCLUSIONS: When a standardized volume of fluid is removed based on weight and EV added, hematocrit, hemodynamics, and ventricular function do not differ between pediatric pts receiving MUF and CUF for hemofilrration after cardiac surgery.

*By Invitation

26. Hypothennic Cardiopulmonary Bypass Alters Oxygen/Glucose Uptake by the Pediatric Brain.

Frank A. Pigula*, Edwin M. Nemoto*, Ira S. Landsman* and Ralph D. Siewers, Pittsburgh, Pennsylvania

Discussant: Richard A. Jonas, M.D.

OBJECTIVE: The effects of hypothermic cardiopulmonary bypass (CPB) on the pediatric brain remain ill defined and may contribute to brain injury. Uptake of oxygen and glucose by the brain is a critical, tightly coupled process that may be expressed as the oxygen-glucose index(OGI). We hypothesize that CPB alters OGI in the pediatric brain.

METHODS: Cerebral arteriovenous (A-V) oxygen, glucose, and lactate differences were compared in 11 children during CPB. Five paired arterial and jugular bulb samples were obtained (preCPB, CPB_cooling', CPB_nadir', CPB_rewarm, postCPB). OGI was calculated: OGI(%)=[A-V_O2(ųmol/ml)/6 * A-V_{glu(ųmol/ml)} * 100 Dissolved O₂ was included.

RESULTS: On CPB_coolmg and CPB_nadir, OGI decreased significantly as A-V_GLUC remained stable with lower A-V₀₂ At CPB_rewarm both A-V _O2 and A-V_GLUC increased, and OGI remained depressed. A-V_LACT increased at re-warming.

CONCLUSIONS: We conclude that CPB alters oxygen and glucose uptake by the pediatric brain. On CPB, OGI decreased as a result of excessive cerebral glucose uptake relative to oxygen. The resulting substrate imbalance (excess glucose) may lead to osmotic cerebral edema. Also, excess glucose availability at rewarming may induce anaerobic metabolism, reflected by increased lactate production. Thus, this phenomenon may contribute to CPB related brain injury in children.

	temp(0c)	A-V02	A-VGLUC	OGI(%)	A-VLACT
preCPB	35±.4	2.5±.9	.4±.9	117±70	-1.3±.2
CPBcooling	28±1	1.1 ±.5*	.4±.4	53±19†	-2.5±1.7
CPBnadir	24±4	1.4±.6*	.6±.8	54±25†	-2.8±1.9
CPBrewarm	26±6	2.8±.9	.8±.4	62±16†	-3.7±3.2‡
postCPB	36±.5	2.6±.8	.4±.2	149+83	-3.4±3.1‡

values are mean±SD. *p<.01 compared to preCPB, CPB_rewarm™, postCPB. ^†p^<-01compared to preCPB, postCPB. ^‡p<.05 compared to preCPB, ANOVA.

*By Invitation

27. Percutaneous Arteriovenous Carbon Dioxide Removal Improves Survival in Acute Respiratory Distress Syndrome: a Prospective Randomized Outcomes Study in Adult Sheep

Joseph B. Zwischenberger, Scott K. Alpard*, Weike Tao*, Donald J. Deyo* and Akhil Bidani*, Galveston, Texas

Discussant: Robert H. Bartlett, M.D.

OBJECTIVE: AVCO₂R is a simple arteriovenous shunt for CO₂ removal to minimize baro/volutrauma secondary to mechanical ventilation. We performed a prospective randomized outcomes study in our clinically relevant model of ARDS.

METHODS: Our LD40 model of ARDS requires smoke inhalation (36 breaths) and a 40% TBSA 3rd degree burn followed by protocol driven volume-controlled mechanical ventilation. All animals developed ARDS (PaO₂/FiO₂ < 200) 48-52 hours after injury. 18 animals randomized to AVCO₂R (n=9) or SHAM (n=9). One in each group died of technical complications (statistics based on 8 per group). AVCO₂R animals were anesthetized, systemically heparinized, then the common carotid artery and jugular vein cannulated with percutaneous 10F arterial and 14F venous cannulas connected to a commercially available 2.5 cm2 low resistance gas exchanger. SHAM received identical operative exposure without can-nulation. Both groups received identical, algorithm-directed pressure-controlled ventilation to normal blood gases.

RESULTS: The study involved 2,946 hours of cage-side critical care and 696 hours of AVCO₂R without significant complications. 8/8 AVCO₂R and only 3/8 SHAM survived the 7 day study. AVCO₂R survivors averaged 2.4 days of mechanical ventilation versus 6.2 days for SHAM. The circuit pressure gradient was less than 10 mmHg and CO₂ removal averaged 103 mL/min (97% of total CO₂ production). AVCO₂R blood flow ranged from 820 to 968 mL/min (11-14% of cardiac output). Cardiac output, heart rate, mean arterial pressure, and pulmonary artery wedge pressure did not significantly change despite AVCO2R. At 48 hours of ARDS, AVCO₂R achieved significant reductions compared to SHAM in TV (420 to 270 mL/ min), PIP (25 to 14 cmH₂O), MV (13 to 5 L/min), RR (25 to 16 breaths/ min), and FiO₂ (.88 to .35).

CONCLUSIONS: Percutaneous AVCO₂R is a simple arteriovenous shunt capable of near-total CO2 removal, which, in this model of ARDS, allowed a significant reduction in minute ventilation, significantly decreased ventilator dependent days, and significantly improved survival.

*By Invitation

28. Surgically Created Double Orifice Left Atrioventricular Valve: a Valve-Sparing Repair in Selected Atrioventricular Septal Defects.

Loc Mac*, Patrice Dervanian*, Virginie Lambert*, Jean Losay* and Jean-Yves Neveux*, Paris, France

Discussant: Constantine Mavroudis, M.D.

OBJECTIVE: Reconstruction of a competent left atrioventricular valve (LAW) is the cornerstone of the repair of atrioventricular septal defects (AVSD). Regardless of used techniques, some structural features of LAW (large mural leaflet, dysplastic tissue valve) represent a challenge for repair without a postoperative regurgitation. A retrospective study was conducted to evaluate the results of a surgically created double orifice LAW performed in such circumstances.

METHODS: