1:30 p.m. SCIENTIFIC SESSION

Room 6, San Diego Convention Center

Moderators: D. Glenn Pennington, M.D.

Douglas J. Mathisen, M.D.

6. HETEROTOPIC CARDIAC TRANSPLANTATION IN INFANTS AND CHILDREN.

Francesco Santini, M.D.*, †Cornelius Dyke, M.D.*, Rosemary Radley-Smith, FRCP*, Asghar Khaghani, FRCS* and Magdi H. Yacoub, FRCS

Harefield, Middlesex, United Kingdom

Discussant: Charles B. Huddleston, M.D.

Children with advanced heart failure, particularly those with elevated pulmonary vascular resistance, pose a difficult management problem because of the fact that normal donor right ventricle cannot cope with the high pulmonary resistance, and the relative shortage of donor organs of an appropriate size for this age group. To address these issues, in the period between January 1, 1991 and October 31, 1995, 11 children, 6 boys and 5 girls, ranging in age between 11 months to 15 years (mean 7.2 years), and having a mean weight of 24.4 kg (range, 7.6 to 56.8 kg) underwent heterotopic heart transplantation (HHTx). Seven patients had dilated and 4 restrictive cardiomyopathy. Seven patients (63.6%) had significant elevation of pulmonary artery pressure (PAsys 68.7 ± 7.4, mean transpulmonary gradient = 20.3 ± 2.4). In all patients the donor pulmonary artery was anastomosed to the right atrium, avoiding the use of any prosthetic material. Ischemic time varied between 135 and 255 min (mean, 181.3 min). Immunosuppression regiment included cyclosporin and azathioprine. Steroids were not routinely used. One patient died in the hospital of acute rejection (p.o.d.=16). One patient developed right middle lobe collapse, treated successfully. Ten survivors (90.9%) are alive and active with a normal pattern of growth at a mean follow-up time of 20.8 months (range 2 to 52 months). Repeated cardiac catheterization performed in 4 patients showed slow progressive drop in pulmonary vascular resistance. Echocardiogram at 6 months postoperatively showed a mean donor heart shortening fraction of 38.7 ± 4.3%. There was no significant change in the function of the recipient hearts.

It is concluded that HHTx in children is feasible for a selected group of patients with good median term results, notably regression of pulmonary vascular disease, normal growth, and lack of chest complications.

†1995-96 Graham Fellow

*By invitation

7. NEUROPHYSIOLOGICAL MONITORING FOR OPTIMAL BRAIN PROTECTION DURING RETROGRADE CEREBRAL PERFUSION.

Brian L. Ganzel, M.D.*, Harvey L. Edmonds, Jr., Ph.D.* and John R. Pank, M.D.*

Louisville, Kentucky

Sponsored by: Laman A. Gray, Jr., M.D., Louisville, Kentucky

Discussant: Randall B. Griepp, M.D.

Purpose: Selective RCP has been recently advocated as a means of protecting the brain during aortic surgery requiring hypothermic circulatory arrest. To aid in the establishment of optimal conditions for producing RCP, we here report on the electrical, hemodynamic and metabolic changes occurring before, during and after effective retroperfusion of the middle cerebral arteries. The results are compared with arrest using no RCP.

Methods: Monitoring was performed on 25 patients requiring aortic surgery (7 without RCP). Cerebral electrical activity was monitored by EEG and brainstem auditory evoked potentials, while transcranial Doppler ultrasound (TCD) measured the direction and velocity of cerebral blood flow bilaterally in the middle cerebral arteries. Metabolic activity was assessed continuously by regional cerebrovenous oxygen saturation (CVOS) using transcranial near-infrared spectroscopy. In all cases, patients were cooled to electrocerebral silence (ECS) prior to circulatory arrest.

Results: There was a wide variation in the nasopharyngeal temperature required to maintain complete ECS (21-8°C). Arrest times were similar in groups with (35 ± 17 min) and without (28 ± 16) RCP. Retrograde flow was verified by TCD in 13/18 patients during RCP. In the other 5, no middle cerebral arterial flow was evident at the onset of retrograde perfusion. TCD signals indicating reversed flow subsequently appeared in 2 of these patients after adjustment of the superior vena caval snare or increasing perfusion pressure. Corrective measures were ineffective in producing a TCD signal in the other 3 patients despite the effux of dark blood from the carotid arteries. In the RCP group with TCD-documented flow reversal, the CVOS decreased by 3 ±4% during retroperfusion. In contrast, CVOS decreased by 19 ± 14% in the arrest group (P = 0.03). The magnitude of cerebral oxygen desaturation in the 3 RCP patients without TCD-documented retrograde flow likewise decreased 19 ± 14%. Cerebral electrical activity returned in 17 ± 8 min in the RCP group with adequate flow compared with 60 ± 43 min in the arrest group (P = 0.007). The EEG in RCP patients without evidence of cortical flow returned in 36 ± 21 min (P - 0.03).

Conclusions: 1) The EEG defined the degree of hypothermia required for effective cerebral cooling prior to arrest. 2) TCD documented the establishment and maintenance of retrograde cerebral, as opposed to cranial, flow. 3) Optimally-adjusted RCP prevented further cerebral oxygen desaturation and facilitated the return of cerebral electrical activity with antegrade perfusion.

*By invitation

8. HEMODYNAMIC BENEFITS OF THE TORONTO STENTLESS PORCINE AORTIC VALVE PROSTHESIS.

Dario Del Rizzo, M.D.*, Bernard S. Goldman, M.D., George T. Christakis, M.D. and Tirone E. David, M.D.

Toronto, Ontario, Canada

Discussant: Delos M. Cosgrove, M.D.

The Toronto SPV™ is a stentless porcine heterograft used for replacement of the human aortic valve. Our institutions are participating in an ongoing multicentre international Phase II clinical trial to assess the efficacy of this valve. We report herein the hemodynamic benefits of this bioprosthesis. To assess valve function, echo Doppler studies were performed after operation, at 3-6 months, at 12 months, and annually thereafter. We previously demonstrated in 118 patients operated upon between March 1992 and December 1993, that effective orifice area (EOA) increased on average by 40% over a 1-year follow-up period while mean transvalvular gradient decreased by a similar 40%. With 3 year data now available, we report that these trends are persistent. Mean transvalvular gradient has continued to decrease such that at three years the average gradient is 2 mmHg for all valve sizes (23 mm to 29 mm) implanted. By repeated measures analysis with Wilks' Lambda (MANOVA), we showed that the regression of transvalvular gradient was significant in all valve sizes at all follow-up intervals, and that the rate of regression was independent of valve size.

In the next phase of this study we investigated the relationship between changes in gradient and changes in left ventricular (LV) mass. We looked at 71 patients (50 males and 21 females; age 59.8 ± 11.9 years [range: 33-80]) between March 4, 1992 and June 30, 1995 (all operated upon at one centre). Valve pathology was stenosis in 44, insufficiency in 11, and mixed disease in 16 patients. In 90% of patients there was 0 or trace insufficiency, while there was minimal insufficiency (1+) in the remaining 10% of cases. No patient had moderate insufficiency (2+) and there was no deterioration in valve performance during the follow-up period. Again, we demonstrated an average 36% reduction in mean transvalvular gradient (10.8 ± 4.6 mmHg to 6.9 ± 3.6 mmHg) in the initial follow-up. Regression analysis revealed that Ag (change in gradient) was dependent on changes in transvalvular velocity (V₂) (R = 0.93, p < 0.0001). Mean LV mass was 363.8 ± 96.8 gm postoperative and decreased to 220.2 ± 76.8 gm at 3-6 months. Scatter plot analysis showed correlation between AV₂ and LV mass regression; we found congruency in 70% of patients (p < 0.0001) and LV mass regression in 80% of patients (p < 0.0001). The highest congruency was in aortic stenosis where LV mass regression occurred in 90% of cases. LV mass index (LV mass/body surface area) was found to decrease over time in all valve sizes. These results are consistent with the hypothesis that LV remodeling occurs following AVR with the SPV valve.

In conclusion, these observations suggest the Toronto SPV™ valve is an excellent substitute in patients deemed appropriate candidates for a tissue bioprosthesis, especially in those patients with marked LVH and in those in whom a mismatch between valve size and patient size might be anticipated. The reduction in LV mass and transvalvular gradient and the potential anticoagulation free durability may have beneficial prognostic implications.

*By invitation

9. ADENOID CYSTIC CARCINOMA OF THE AIRWAY: A THIRTY-TWO YEAR EXPERIENCE.

Donna E. Maziak, MDCM, FRCSC*, Thomas R.J. Todd, M.D., FRCSC, Shafique H. Keshavjee, M.D., FRCSC*, Timothy L. Winton, M.D., FRCSC*, Peter Van Nostrand, M.D.* and F. Griffith Pearson, M.D., FRCSC

Toronto, Ontario, Canada

Discussant: Douglas Wood, M.D.

We have reviewed our experience in 36 patients with adenoid cystic carcinoma of the upper airway seen during a 32 year period (1963-1995). The average age was 45 years (range: 15 - 80) with a male:female ratio of 1.1:1. Of these 36 patients, 30 were managed by resection and reconstruction (primary anastomosis 25; marlex prosthesis 5) of whom 25 received adjuvant radiotherapy. Six were managed with radiotherapy only.

Pathology revealed local invasion in all cases beyond the outer wall of the trachea. In most cases, microscopic extension was found in submucosal and perineural lymphatics, well beyond grossly viable or palpable tumour. Lymphatic metastases were rare: 4 of 30 (13.3%) resected cases. Subsequent haematogenous metastases occurred in 15 of 36 patients (41.6%). Twelve of 36 cases (35%) had pulmonary metastases.

Thirteen of 30 resections were complete and potentially curative. There were 2 operative deaths, and the mean survival in 11 surviving patients was 7.8 years (range: 6 months-18 years). Seventeen of 30 resections were incomplete (residual tumour at final pathology), with one operative death and a mean survival of 7.7 years (range: 9 months - 29 years). The 6 patients treated with radiation only had a mean survival of 3.3 years (range: 2 months -10 years). In the 12 patients with pulmonary metastases, mean survival was 3 years (range: 6 months - 6 years) following its identification.

Adenoid cystic carcinoma is a rare tumour, which is always locally invasive but frequently amenable to resection. Although late local recurrence following resection is common (up to 29 years) excellent long-term palliation is commonly achieved following both complete and incomplete resection. There was little difference in survival between complete and incompletely resected cases. Pulmonary metastases are common, usually slow growing, and are not necessarily a contraindication to palliative resection of the primary tumour. Long periods of control can be obtained with radiotherapy alone. The best results, however, are obtained by resection. Adjuvant radiotherapy is assumed to favourably influence survival.

*By invitation

10. CRICOPHARYNGEAL MYOTOMY FOR NEUROGENIC OROPHARYNGEAL DYSPHAGIA.

Nancy Claire Poirier, M.D.*, Luigi Bonavina, M.D.*, Raymond Taillefer, M.D.*, Attilio Nosadini, M.D.*, Alberto Peracchia, M.D. and Andre C.H. Duranceau, M.D

Montreal, Quebec, Canada and Milan, Italy

Discussant: Mark B. Orringer, M.D.

Between 1976 and 1994, 40 patients with neurological disorders associated with incapacitating upper esophageal sphincter (UES) dysfunction underwent a cricopharyngeal myotomy. Etiology for the dysphagia was a cerebrovascular accident (65%), amyotrophic lateral sclerosis (5%), trauma (5%), pseudobulbar palsy (7.5%), iatrogenic (2.5%), peripheral nerve disorders (IX-X-XII) (7.5%), Parkinson's disease (2.5%), Arnold Chiari (2.5%), and multiple sclerosis (2.5%). Patients were assessed pre- and post-operatively with a mean follow-up period of 48 months (1-255 months).

The mortality was 2.5% (1 patient) and resulted from aspiration pneumonia. Seven patients are asymptomatic, 23 patients are improved and 10 patients report no change in symptoms. Radiological evidence of functional obstruction at the pharyngoesophageal junction is reduced. UES resting and contraction pressures are significantly decreased in the UES but the coordination abnormalities are unchanged. Single liquid bolus radionuclide emptying studies of the pharynx show persistant stasis in all patients where this study was obtained.

3:15 p.m. INTERMISSION - VISIT EXHIBITS

*By invitation

4:00 p.m. SCIENTIFIC SESSION

Room 6, San Diego Convention Center

Moderators: Andrew S. Wechsler, M.D.

Bruce A. Reitz, M.D.

11. WHAT IS THE APPROPRIATE SIZE CRITERION FOR RESEARCH OF THORACIC AORTIC ANEURYSMS?

John A. Elefteriades, M.D., Michael A. Coady, M.D.*, John A. Rizzo, Ph.D.*, Kevin Johnson, M.D.*, Umer Darr, M.D.*, Ellen Frank, M.D., M.P.H.* and Gary S. Kopf, M.D.

New Haven, Connecticut

Discussant: O. Wayne Isom, M.D.

Although many papers have described techniques for resection of thoracic aortic aneurysms (TAA), limited information on the natural history is available to aid in defining criteria for surgical intervention.

Data on 225 patients with TAA at one center from 1985 to 1995 were analyzed. This computerized database included 638 imaging studies (MRI, CT, ECHO). Overall survival at 1 and 5 years was 88% and 65%, respectively. Patients having aortic dissection had markedly lower survival (83%, 1 year; 31%, 5 year) than the non-dissected cohort (90%, 1 year; 70%, 5 year).

Mean size of the thoracic aorta in these patients was 5.1 cm. The mean growth rate was 0.15 cm/year. Multivariate regression analysis to isolate risk factors for aortic expansion revealed a size of > 6.0 cm to be a significant predictor (p<0.0001).

Median size at time of rupture or dissection was 6.1 cm (5.9 cm for ascending; 7.0 cm for descending). The incidence of dissection or rupture increased with aneurysm size, as illustrated below.

Seventy-four patients underwent surgery for their TAA. For elective operations, the mortality was 9.7%; emergent operations had a mortality of 30.2%.

If the median value at the time of dissection or rupture were used as the intervention criterion, half of the patients would have suffered a devastating complication prior to surgery. Accordingly, a criterion lower than the median is appropriate. We recommend 5.5 cm as an acceptable size for elective resection of ascending aortic aneurysms, as these can be performed with relatively low mortality. For aneurysms of the descending aorta, where peri-operative complications are greater and the median size at time of complication is larger, we recommend intervention at 6.5 cm.

*By invitation

12. ONE STAGE COMPLETE UNIFOCALIZATION IN INFANTS: WHEN NOT TO CLOSE THE VENTRICULAR SEPTAL DEFECT.

Frank L. Hanley, M.D., V. Mohan Reddy, M.D.*, Edwin Petrossian, M.D.*, Phillip Moore, M.D.* and David F. Teitel, M.D.*

San Francisco, California

Discussant: Hillel Laks, M.D.

Background: The decision to close the ventricular septal defect at the time of unifocalization in patients with pulmonary atresia with ventricular septal defect and major aortopulmonary collaterals (PA.VSD.MAPCAS) may be difficult. Standard indices of true pulmonary arteries cannot be used since reconstruction using the collaterals makes the preoperative values meaningless. We have developed an index of total pulmonary vascular cross-sectional area which systematically evaluates true pulmonary arteries and all collateral vessels to be unifocalized. The purpose of this study is to determine if this new index is predictive of peak right ventricular systolic pressure following complete repair.

Patients and methods: Since July 1992, 17 infants (age: mean ± SD = 4.1 ± 3.3 months) with PA.VSD.MAPCAS have been managed at our institution. Complete unifocalization (n=16) from a midline approach and VSD closure (n=12) was the procedure of choice. In 4 patients the VSD was left open and only one patient with severe distal collateral artery stenosis was staged. The preoperative angiograms were evaluated to estimate the size of the true pulmonary arteries and size of the collaterals. The indexed cross-sectional areas of the true pulmonary arteries (PAI), of the MAPCAS (CAI), and the total bed (TI = PAI + CAI) were calculated. True pulmonary arteries were measured just proximal to the first branching point and the collaterals were measured at the narrowest point beyond the site of surgical unifocalization and proximal to any branches. The ratio of the peak systolic pressure in the right ventricle (pRV) to that in the left ventricle (pLV) was estimated from the absolute pressures obtained in the operating room.

Results: The median PAI was 40.5 mm²/M² (range = 0 to 294.4 mm²/M²), the median CAI was 72.9 mm²/M² (range = 22.5 to 422.1 mm²/M²), and the median TI was 174.7 mm²/M² (range = 50.5 to 422.1 mm²/M²). The median pRVipLV was 0.63 (range = 0.29 to 0.85). By unpaired /-Test there was a significant difference (p = 0.011) in the TI between the two groups. By linear regression there was a significant correlation (p = 0.023) between the TI and VSD closure. In addition, the pRV:pLV ratio appeared to be lower in patients with higher TI and approached significance (p = 0.07).

Conclusion: The total pulmonary vascular index appears to correlate with postrepair pRV:pLV. This index is useful in deciding when not to close the VSD.

*By invitation

13. DOES SUCCESSFUL BRIDGING WITH THE IMPLANTABLE LEFT VENTRICULAR ASSIST DEVICE AFFECT CARDIAC TRANSPLANTATION OUTCOME?

Malek G. Massad, M.D.*, Patrick M. McCarthy, M.D., Nicholas G. Smedira, M.D.*, Daniel J. Cook, Ph.D.*, Norman B. Ratliff, M.D.*, Marlene Goormastic, M.P.H.*, Jose Navia, M.D.*, James B. Young, M.D.* and Robert W. Stewart, M.D.*

Cleveland, Ohio

Discussant: Eric A. Rose, M.D.

Purpose: We sought to determine if cardiac transplant recipients who required a bridge-to-transplant with an implantable LVAD had a different outcome than patients (pts) who had primary cardiac transplants.

Methods: A retrospective study of 238 primary cardiac transplants at our institution between January 1992 and October 1995 included 44 pts who received the HeartMate^® LVAD and 194 pts who had no LVAD support. LVAD transplant rate was 77%. Comparison of demographic and clinical factors was performed and associations were analyzed using Chi square test. Distribution of continuous factors was compared using Wilcoxon Rank Sum test. Kaplan-Meier survival estimates were compared using log-rank test.

Results: LVAD transplants constituted 7.5% of all primary transplants in 1992 (n=67), 17% in 1993 (n=58) and 1994 (n=63), and 36% in 1995 (n=50, YTD). LVAD and non-LVAD pts had similar age distribution (median 52 yrs) and female gender (16% vs 26%, NS). LVAD recipients had greater weight (81 kg vs 70 kg, p = 0.003) and body surface area (1.94 m² vs 1.81 m², p = 0.05). They were more likely to have type O blood group (52% vs 34%, p = 0.07), and ischemic cardiomyopathy (70% vs 42%, p = 0.001). All LVAD pts and 41% of non-LVAD pts had previous cardiac operations (mean number per pt: 1.3 vs 0.3, p < 0.001). LVAD related sepsis occurred in 20/44 pts (45%) during support. More pts in the LVAD group had anti-HLA antibodies prior to transplant (T-cell PRA level >10% in 66% of LVAD pts vs 15% of non-LVAD pts, p < 0.0001). Blood utilization during LVAD support of more than 96 units of blood products (median) results in significantly higher anti-HLA antibody levels. Sixty-one percent of non-LVAD pts were UNOS status I. Time on the waiting list was longer for LVAD pts compared to non-LVAD status I pts (median 88 days vs 33 days, p = 0.002). There was no difference in the length of post-transplant hospital stay (median 15 days for each) and operative mortality (4.6% LVAD vs 5.2% non-LVAD). No significant difference was found in the Kaplan-Meier survival curves. One year survival was 93% in the LVAD group and 88% in the non-LVAD group (NS). Two-year survival was 82% vs 84% (NS). Comparison of post-transplant events showed no significant difference in CMV infection rates (23% vs 16%), or the incidence of vascular rejection (20% vs 15%) or moderate/severe cellular rejection (79% vs 71%) at one-year follow-up.

Conclusions: LVAD support intensified donor shortage by including recipients who otherwise would not have survived to transplant. Bridging affected transplant demographics favoring pts who are larger, have ischemic cardiomyopathy, have had previous cardiac surgery, are HLA sensitized, and have blood group O. Successfully bridged pts wait longer for transplant but have similar post-transplant hospital stay and operative mortality and achieve survival similar to that of pts not requiring LVAD support.

*By invitation

14. THE DISCHARGE OF PATIENTS WITH NEW-ONSET ATRIAL FIBRILLATION AFTER HEART SURGERY IN ATRIAL FIBRILLATION IS SAFE AND COST-EFFECTIVE.

Robert L. Hannan, M.D.*, Allen J. Solomon, M.D.*, Peter C. Kouretas, M.D.*, Richard A. Hopkins, M.D., Nevin M. Katz, M.D. and Robert B. Wallace, M.D.

Washington, DC

Discussant: Renee S. Hartz, M.D.

Atrial fibrillation is a frequent complication of open heart surgery and leads to prolonged hospitalization and increased cost. We retrospectively reviewed our experience over a one year period to determine if patients in new-onset atrial fibrillation (AF) could safely be discharged in atrial fibrillation after ventricular rate had been controlled and anticoagulation initiated. In the year ending in August 1995, 505 adult patients underwent heart surgery utilizing cardiopulmonary bypass. Of the 488 survivors (96.6%), 93 (19%) developed AF postoperatively. Of these 93 patients, 82 (88%) were discharged in sinus rhythm (NSR) and 11 (12%) were discharged in AF. Results are summarized in the table below: age is mean age in years, risk is by the modified Parsonnett score, postop LOS is length of stay postoperatively, and mean hospital cost and charge are normalized to 1.00 in patients without AF.

At mean follow-up period of 222 days, all 11 patients discharged in AF were alive and well: 9 patients were in sinus rhythm. Outpatient cardioversions were performed on 3 patients, one of whom remained in sinus rhythm. No systemic emboli or other complications were noted in any of the patients discharged in AF. Patients discharged in AF had similar cost and charges to patients without AF: patients with AF discharged in NSR had increased cost and charges. We suggest that the discharge of patients in persistent AF after heart surgery is safe and cost-effective, and hypothesize that earlier discharge of patients in AF may be appropriate.

*By invitation