Evaluation of the sorin bicarbon bileaflet valve in 488 patients (519 prostheses)

Evaluation of the Sorin Bicarbon Bileaflet Valve in 488 Patients (519 Prostheses) Ira Goldsmith, FRCS, Gregory Y.H. Lip,

MD,

and Ramesh L. Patel,

MD

The Sorin bicarbon bileaflet prosthesis was introduced in 1990. To evaluate the clinical performance of this prosthesis, we reviewed 519 prostheses that were implanted in 488 patients (275 men, 213 women; mean age 59 years; SD 10.8, range 19 to 88) from 1993 to 1997. Preoperatively, 82% of patients were in New York Heart Association (NYHA) functional class III or IV. There were 263 aortic valve replacements (AVRs) (54%), 194 mitral valve replacements (MVRs) (40%), and 31 AVRs and MVRs (both) (6%). Concomitant procedures were performed in 82 patients (17%) . Follow-up was complete in 471 (97%) with a total cumulative follow-up of 866 patient-years. The 30-day mortality for patients with AVR was 5.7% (95% confidence interval [CI] 2.9 to 8.5), MVR 17.5% (CI 9.9 to 19.7), and both 19% (CI 7.6 to 51.1), with no early valve-related deaths. Patient survival at 55 months was 76% (SE 2.27%), with patients with AVR being 90%, MVR 63%, and both 61%. This was influenced by the following: (1) valve position, which was higher for MVR (p 5 0.0001); (2) poor NYHA functional class (p 5 0.0006); (3) reoperation (p 5 0.02); and (4) age >70 years (p 5 0.0001). Valverelated complications (expressed as percentage per patient year and number of events) were major thrombo-

embolism at 0.9% per year (8), with AVR rates being 1.2% per year (6) and MVR 0.7% per year (2); major hemorrhage at 2.3% per year (20) with AVR rates being 2.4% per year (12) and MVR 2.5% per year (7); bacterial endocarditis at 0.2% per year (2); and nonstructural dysfunction rate of 0.7% per year (6). The reoperation rate was 0.9% per year (8) with AVR being 0.6% per year (3) and MVR 1.7% per year (5). At 55 months, actuarial freedom from major thromboembolism was 97% (SE 1.1%) with AVR being 96% and MVR 98%; major hemorrhage 89% (SE 3.1%) with AVR being 88.6% and MVR 91%; structural valve dysfunction 100% (SE 0.0%); and reoperation 97.1% (SE 1.1%) with AVR being 98.5% and MVR 94.6%. At follow-up, 88% of survivors were in NYHA class I or II. In this series, hospital mortality and overall survival in patients were influenced by the patients’ clinical characteristics. There were no early valve-related deaths. Valve-related complications were similar to previously reported series with no episode of structural failure. Our experience with the Sorin bicarbon bileaflet prosthesis suggests that it has a satisfactory clinical performance, with low complication rates. Q1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;83:1069 –1074)

t our center, the Sorin bicarbon bileaflet prosthesis (Sorin Biomedica Cardio S.p.A., Saluggia, A Italy) is commonly used when mechanical valve re-

bicarbon bileaflet prosthesis for both early and longterm treatment is warranted and is the objective of the present study.

placement is undertaken. This is a third-generation bileaflet prosthesis made of pyrolytic carbon-coated components.1 The housing is made of titanium and is therefore slim, providing a significantly larger orifice while maintaining structural stability and integrity. This, as well as curved leaflets, provide improved hemodynamic performance. The hinge mechanism is unique, because it allows the leaflets to move by rolling rather than by sliding, thus exposing all areas to a full washing effect at each point of the cardiac cycle. The 2 leaflets open to 80° from a 20° horizontal axis, and the housing is rotatable within the sewing ring. The prosthesis is presented in 2 models, the aortic and mitral, for the respective positions of implantation. Introduced for implantation in the United Kingdom in 1991, clinical evaluation of the Sorin From the Department of Cardiothoracic Surgery, Walsgrave Hospital, Coventry; and the University Department of Medicine, City Hospital, Birmingham, United Kingdom. Manuscript received June 4, 1998; revised manuscript received and accepted November 13, 1998. Address for reprints: Ira Goldsmith, FRCS, Department of Cardiothoracic Surgery, Walsgrave Hospital, Clifford Bridge Road, Coventry, CV2 2DX, England. ©1999 by Excerpta Medica, Inc. All rights reserved.

METHODS Walsgrave Hospital, Coventry, is a tertiary referral cardiothoracic unit serving the population of West Midlands, United Kingdom. For patients admitted to this unit, information for the study was obtained and entered into a computerized database. There are 4 cardiac surgeons, all of whom perform valvular surgery, and the operative techniques were at the discretion of the individual surgeon. After surgery, antithrombotic therapy with warfarin was initiated as soon as oral intake was possible, and efficacy of therapy was monitored by routine measurement of the international normalized ratio (INR). Follow-up of the patients was conducted by clinical review and telephone contact with the patients and their general practitioners according to a set protocol and proforma. The valve-related events were ascertained, and causes of death were determined from the official death register, general practitioner, and, where possible, postmortem reports. Performance of the prosthesis was studied according to the “Guidelines for Reporting Morbidity and Mortality After Cardiac 0002-9149/99/$–see front matter PII S0002-9149(99)00017-X

1069

FIGURE 1. NYHA functional class. TABLE I Diagnosis at the Time of Surgery Diagnosis

No. (%)

Aortic regurgitation 55 (11) Aortic stenosis 133 (27) Aortic stenosis 1 aortic regurgitation 67 (14) Mitral regurgitation 93 (19) Mitral stenosis 26 (5) Mitral stenosis 1 mitral regurgitation 62 (13) Aortic regurgitation 1 mitral regurgitation 14 (3) Aortic regurgitation 1 mitral stenosis/regurgitation 6 (1) Mixed aortic valve disease 1 mitral regurgitation 7 (2) Mixed aortic and mitral valve disease 9 (2) Aortic stenosis 1 mitral stenosis/regurgitation 6 (1) Mitral stenosis/regurgitation 1 tricuspid regurgitation 10 (2) Total 488 (100)

Valvular Operations” approved by the Society of Thoracic Surgeons and the American Association for Thoracic Surgery.2 The valve-related complications were defined as thromboembolism, hemorrhage, structural valve deterioration, nonstructural dysfunction, and prosthetic valve endocarditis. Mortality was defined as death within 30 days of operation or in the same hospital admission as operation, regardless of cause. Statistical analysis: Data are expressed as mean (SD) and median (interquartile range), as appropriate. The valve-related complications were evaluated in the time-related manner with Kaplan-Meyer survival curves. Linearized rates, expressed as percentage per patient year, were calculated by dividing the number of events by the total follow-up time in years and multiplying by 100. Events were also assessed by the site of implantation; by patient age groups stratified to those aged ,60 years, 60 to 70 years, and .70 years; and by gender. Survival was also assessed by the valve lesion, namely, pure aortic stenosis, aortic regurgitation, mixed aortic lesion, mitral stenosis, mitral regurgitation, and mixed mitral valve lesion. A p value of ,0.05 was considered statistically significant.

RESULTS

Patient demographics: Five hundred nineteen Sorin bicarbon prostheses were implanted in 488 patients in TABLE II Early and Late Mortality According to Valve Position 488 operations between 1993 and 1997. Aortic valve AVR MVR Both replacement (AVR) was performed in 263 patients Timing (n 5 263) (n 5 194) (n 5 31) Total (54%), mitral valve replacement (MVR) in 194 patients (40%), and both aortic and mitral valve replaceHospital mortality (,30 d) 15 (5) 34 (17) 6 (19) 55 (11) .30 d mortality 8 (3) 28 (14) 3 (9) 39 (8) ment (both) in 31 patients (6%). There were 275 men Total 23 (8) 62 (32) 9 (29) 94 (19) (56%) and 213 women (44%). The mean patient age for the whole cohort was 59 years (SD 10.8, range 19 Values are expressed as number of episodes (percent). to 88): the mean age of patients in years undergoing AVR was 58 (SD 10.9); MVR 61 (SD 10.3); and both 57 (SD 12.2), respectively. Preoperatively, 82% of patients were in New York Heart Association (NYHA) functional class III or IV (Figure 1). The different valve lesions at the time of surgery are detailed in Table I. Seven percent (n 5 14) of patients had a previous cardiac operation while concomitant surgical procedures, primarily coronary artery bypass grafts, were performed in 17% (n 5 82). Hypothermic crystalloid cardioplegia was used in 94% (n 5 458) of the operations for myocardial protection, whereas blood cardioplegia was used in the remainder. Follow-up: By the date of census, 471 patients (97%) had been folFIGURE 2. Patient survival at 4 years and age groups. DF 5 degrees of freedom; lowed up. Total cumulative folProb 5 probability. 1070 THE AMERICAN JOURNAL OF CARDIOLOGYT

VOL. 83

APRIL 1, 1999

low-up was 866 patient-years (for AVR 505, MVR 316, and both 45, respectively). The mean follow-up period was 22.4 months (range 1 to 55). At discharge, the median INR of patients was 2 (interquartile range [IQR] 2 to 3), and at follow-up, the median INR was 2 (IQR 2 to 3) for AVR and 3 (IQR 3 to 4) for MVR. Mortality: The 30-day early mortality for the whole cohort, which included 16 operative deaths, was 11% (confidence interval [CI] 7.6 to 12.9). The 30-day mortality for patients with AVR was 5.7% CI 2.9 to 8.5), MVR 17.5% (CI 9.9 to 19.7), FIGURE 3. Patient survival at 4 years according to valve position. DVR 5 doubleand both 19% (CI 7.6 to 51.1), with valve replacement; other abbreviations as in Figure 2. no early valve-related deaths. The hospital mortality (Table II) and long-term patient survival (Figures 2, 3, and Table III) were influenced by (1) the patients’ TABLE III Influence of the Patients’ New York Heart age, being higher in the patients aged .70 years (p 5 Association (NYHA) Functional Class and Previous Cardiac 0.0001); (2) the NYHA functional class (p 5 0.0006); Surgery on Patient Survival at Four Years (3) previous cardiac surgery (p 5 0.02); and (4) valve % (SE) p Value position, which was poor for those undergoing MVR (p 5 0.0001). The overall patient survival at 5 years NYHA functional class was 77% (SE 2.3%). The late mortality expressed as a I 95 (5.1) II 94 (4.4) Df 3 linearized occurrence rate was 2.7% per patient-year. III 86 (3.4) 0.0006 VALVE-RELATED MORTALITY: There were 6 late IV 73 (4.1) valve-related deaths; 2 were because of stroke, 1 bePrevious cardiac cause of cerebral hemorrhage, and 3 occurred after surgery operation for paraprosthetic leakage. The valve-reNo 79 (2.4) Yes 63 (7.3) 0.02 lated mortality rate was 0.7% per patient-year (Table IV), and overall freedom from valve-related mortality Df 5 degrees of freedom. at 5 years was 97% (SE 1.1%; Table V). SURVIVAL ACCORDING TO VALVE LESION: Outcome for aortic and TABLE IV Valve-Related Complications AVR Valve-Related Complications Thromboembolism Major*/minor†/fatal Transient‡ Hemorrhage Major§ Minor\ Structural failure Nonstructural dysfunction Prosthetic valve endocarditis Reoperation Mortality

MVR

Both

No.

Rate

No.

Rate

No.

Rate

6 8

(1.2) (1.6)

2 1

(0.7) (0.3)

0 0

(0) (0)

12 23 0 2 1 3 2

(2.4) (4.6) (0) (0.4) (0.2) (0.6) (0.4)

7 14 0 4 1 5 4

(2.5) (4.9) (0) (1.4) (0.2) (1.7) (1.4)

1 2 0 0 0 0 0

(2.2) (4.5) (0) (0) (0) (0) (0)

Total No.

Rate

8 9

(0.9) (1.0)

20 39 0 6 2 8 6

(2.3) (4.5) (0) (0.7) (0.2) (0.9) (0.7)

*Major thromboembolic events defined as any neurologic deficit or peripheral ischemic phenomenon with residual disability. † Minor thromboembolic events defined as any neurologic or peripheral embolic event with no residual disability. ‡ Transient thromboembolic events defined as any neurologic or peripheral embolic event of ,24 hours duration and included clinically suspected episodes. § Major bleeding events defined as any bleeding episode that required hospital admission and intervention or blood transfusion. \ Minor bleeding events defined as any bleeding event that did not require hospital admission and intervention or blood transfusion. Number of events (n) and linearized rates expressed as percentage per patient year.

mitral valve lesions are detailed in Figures 4 and 5. Although the overall survival of patients who underwent surgery for aortic valve disease was 90% (SE 2.1%), the survival was significantly higher for those with pure aortic stenosis compared with those with pure aortic regurgitation and mixed aortic valve disease (Figure 4; p ,0.0001). However, significant survival difference between pure and mixed mitral valve disease groups was not observed (Figure 5). Morbidity: IMPROVEMENT IN PATIENT SYMPTOMS: There was a signif-

icant impact in the NYHA functional class after surgery (Figure 1). At follow-up, 88% of survivors were in NYHA functional class I or II. Interestingly, 16% (n 5 58) of survivors reported that the clicking of the valve was disturbing, especially during the night.

VALVULAR HEART DISEASE/SORIN BICARBON BILEAFLET VALVE

1071

FIGURE 4. Patient survival at 4 years according to aortic valve lesion. Abbreviations as before.

FIGURE 5. Patient survival at 4 years according to mitral valve lesion. Abbreviations as before.

4.2% per year, whereas the overall freedom from all valve-related complications at 55 months was 85% (SE 3.3) (Figure 6). THROMBOEMBOLISM: There were 17 thromboembolic events, despite the patients being on anticoagulant therapy. Of these, 9 (53%) were transient episodes and only clinically suspected (Table IV). The linearized occurrence rate for major, minor, and fatal thromboembolic events was 1.2% per year for AVR and 0.7% per year for MVR. The overall freedom from major, minor, and fatal thromboembolic episodes at 5 years was 97.6% (SE 1.0). Although this was lower for AVR compared with MVR (96% [SE 1.7] vs 98% [SE 1.3]), the difference was not statistically significant (p 5 NS; Table V). There were no episodes of obstructive or nonobstructive valve thrombosis leading to valve failure or fatality noted. BLEEDING EVENTS: Twenty patients were admitted to hospital with bleeding episodes due to anticoagulant therapy (Table IV), of which 1 was fatal due to an intracranial hemorrhage; 12 (60%) of the bleeding episodes occurred after AVR and 7 (35%) after MVR. The median INR at the time of the bleeding events was 4.5 (range 2.7 to 7.1). Overall freedom from bleeding events at 5 years was 89% (SE 2.3%); although the freedom rate was lower for AVR compared with MVR (88.6% [SE 4.3] vs 90.8% [SE 3.4]), this was not significantly different (p 5 NS; Table V). The major hemorrhage rate was 2.4% per year (Table IV). PROSTHETIC VALVE ENDOCARDITIS:

There were 2 episodes of prosthetic valve endocarditis (PVE), which occurred in patients in whom the prosthesis was inserted for infective endocarditis in the first instance (Table IV). The overall freedom from PVE and freedom from PVE after AVR, MVR, and both AVR and MVR are shown in Table V. FIGURE 6. Freedom from all valve-related complications and valve position. Abbreviations as before.

VALVE-RELATED COMPLICATIONS: The valve-related complications were thromboembolism, anticoagulantrelated hemorrhage, bacterial endocarditis, and nonstructural valve dysfunction. The linearized occurrence rate for all valve-related complications was

1072 THE AMERICAN JOURNAL OF CARDIOLOGYT

VOL. 83

NONSTRUCTURAL

DYSFUNCTION:

There were 6 episodes of nonstructural valve dysfunction, 2 occurring at the aortic site and 4 at the mitral site, where all 6 were because of paraprosthetic leaks. The overall freedom from nonstructural dysfunction and after AVR, MVR, and both AVR and MVR at 55 months are shown in Table V. The nonstructural dysfunction rate was 0.7% per patient-year (Table IV). APRIL 1, 1999

TABLE V Freedom from Valve-Related Complications at Four Years AVR Valve-Related Complications †

Thromboembolism (major*/minor /fatal) Major‡ hemorrhage Structural failure Nonstructural dysfunction Prosthetic valve endocarditis Reoperation Mortality

MVR

Both

Overall

%

SE

%

SE

%

SE

%

SE

96.0 88.6 100 99.0 99.5 98.5 98.8

1.7 4.3 0 0.7 0.5 0.9 0.9

98.2 90.8 100 95.1 99.3 94.5 94.6

1.3 3.4 0 2.6 0.6 2.6 2.7

100 95.8 100 100 100 100 100

0 4.1 0 0 0 0 0

97.1 89.2 100 96.7 99.5 97.1 97.4

1.1 3.1 0 1.0 0.4 1.1 1.1

*Major thromboembolic events defined as any neurologic deficit or peripheral ischemic phenomenon with residual disability. † Minor thromboembolic events defined as any neurologic or peripheral embolic event with no residual disability. ‡ Major bleeding events defined as any bleeding episode that required hospital admission and intervention or blood transfusion.

TABLE VI Thromboembolic Complications Associated With Mechanical Prosthesis Overall

Single tilting disc Medtronic-Hall Omniscience Caged-ball valve Starr-Edwards

MVR

% Free at 5 Years

Rate (%/patient-year)

% Free at 5 Years

Rate (%/patient-year)

% Free at 5 Years

Wang et al 19967 Antunes 19908 Wang et al 19967 Bernal et al 19989

1.2 4.3 2.7 —

94.8 92 89.9 —

— — — —

— — — 87.1

— — — —

— — — 89.1

Antunes 19908 Keenan et al 199010 Rabago et al 198411 Carrier et al 198712

3.3 4.6

89 90

— 4.4 4.7

— 70

— 6.3 2.0

— 67

Miller et al 198313

2.7

—

—

—

—

Valve Model Bileaflet valve St Jude Medical CarboMedics

AVR

Rate (%/patient-year)

Authors and Year

3.8

87.9 at 4 years 86

STRUCTURAL VALVE DETERIORATION: There were no episodes of structural valve deterioration. Hence, freedom from structural valve failure at 55 months was 100%. REOPERATION: There were 8 patients who underwent reoperation, 3 at the aortic and 5 at the mitral position. Six reoperations were undertaken for paravalvular leaks and 2 for PVE. Overall freedom from reoperation and after AVR, MVR, and both AVR and MVR at 55 months are shown in Table V. The reoperation rate was 0.9% per patient year (Table IV).

DISCUSSION The present study represents the largest reported single-center study assessment of the Sorin bicarbon bileaflet prosthesis.1,3 It is limited by being the experience of a single center, and the results may not be generalizable to the experience of other centers. Nevertheless, the study permits us to draw some perceptions regarding the early clinical performance of the Sorin bicarbon bileaflet prosthesis. In the present study, the overall risks of thromboembolic complications with the Sorin bicarbon prostheses were very similar to other series of mechanical prostheses (Table VI).4 –16 However, unlike the other series, thromboembolic episodes that did occur were higher after AVR when compared with MVR. The reason for this was not entirely clear, although the

presence of arterial disease in the aorta or carotid arteries may have been a possible source of emboli, especially because 8 of the 14 episodes noted after AVR were transient ischemic attacks involving the carotid artery territory. Furthermore, when transcranial Doppler is used to quantify microembolism in patients with prosthetic aortic valves, almost 50% of patients with mechanical aortic valves have detectable cerebral microemboli.17 Another potential source of microemboli may be from microthrombus deposited on atheromatous plaques in the ascending aorta that are disturbed by blood flow above the aortic prosthesis.18,19 In the present series, the overall major bleeding events were similar to other mechanical valve series.4 –16 However, the bleeding episodes that did occur were higher in patients following AVR when compared with MVR. Most instances in which bleeding events occurred were in patients maintained on high INR levels, which reflects the need for stringent anticoagulation intensity control and possibly a lower target INR range, especially because the thromboembolic complications with this valve were generally low. The hospital mortality and long-term patient survival after valve surgery varies with the reported age groups20 and valve positions.6 The 30-day hospital

VALVULAR HEART DISEASE/SORIN BICARBON BILEAFLET VALVE

1073

mortality and 55-month patient survival in our series of mainly elderly patients was influenced by age, the NYHA functional class, previous cardiac surgery, and position of the implant. We have also noticed symptomatic improvement, because 88% of surviving patients were in NYHA functional class I or II at followup, in keeping with the significant morbidity associated with hemodynamically significant valve disease. 1. Di Salvo C, Walesby K. Early single centre experience with 192 sorin bicarbon valves. J Cardiovasc Surg 1996;37:1–3. 2. Clark RE, Edmunds LH, Cohn LH, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Eur J Cardiothorac Surg 1988;2:293–295. 3. Casselman F, Herijgers P, Meyns B, Flameng W, Daenen W. The bicarbon heart valve prosthesis: short term results. J Heart Valve Dis 1997;6:410 – 415. 4. Ibrahim M, O’Kane H, Cleland J, Gladstone D, Sarsam M, Patterson C. St Jude Medical prosthesis: a thirteen-year experience. J Thorac Cardiovasc Surg 1994; 108:221–230. 5. Jageden O, Eker A, Delahaye F, Montagna P, Osette J, de Gevigney GD, Mikaeloff PH. Thromboembolic risk and late survival after mitral valve replacement with the St. Jude Medical valve. Ann Thorac Surg 1994;58:1721–1728. 6. Akins CW. Results with mechanical cardiac valvular prostheses. Ann Thorac Surg 1995;60:1836 –1844. 7. Wang SS, Chu SH, Tsai CH, Lin FY. Clinical use of CarboMedics and St Jude Medical valves. Artificial Organs 1996;20:1299 –1303. 8. Antunes MJ. Clinical performance of St Jude and Metronic-Hall prostheses: a randomized comparative study. Ann Thorac Surg 1990;50:743–747. 9. Bernal JM, Rabasa JM, Gutierrez-Garcia F, Morales C, Nistal JF, Revuelta JM. The CarboMedics valve: experience with 1,049 implants. Ann Thorac Surg 1998;65:137–143.

1074 THE AMERICAN JOURNAL OF CARDIOLOGYT

VOL. 83

10. Keenan RJ, Armitage JM, Trento A, Siewers RD, Hardesty RL, Balmson HT, Griffith BP. Clinical experience with the Medtronic-Hall valve prosthesis. Ann Thorac Surg 1990;50:748 –753. 11. Rabago G, Martinell J, Fraile J, Andrede IG, Montenegro R. Results and complications with the Omniscience prosthesis. J Thorac Cardiovasc Surg 1984; 87:136 –140. 12. Carrier M, Martinean JP, Bonan R, Pelletier LC. Clinical and haemodynamic assessment of the Omniscience prosthetic heart valve. J Thorac Cardiovasc Surg 1987;93:300 –307. 13. Miller DC, Oyer PE, Mitchell S, Stinson EB, Jamieson SW, Baldwin JC, Shumway NE. Performance characteristics of the Starr-Edwards Model 1260 aortic valve prosthesis beyond ten years. J Thorac Cardiovasc Surg 1984;88: 193–207. 14. Foire AC, Naunheim KS, D’Orazio S, Kaiser GC, McBride LR, Pennington DG, Peigh PS, Willman VL, Labovitz AJ, Barner HB. Mitral valve replacement: randomized trial of St Jude and Medtronic-Hall prostheses. Ann Thorac Surg 1992;54:68 –73. 15. Cobanoglu A, Fessler CL, Guvendik L, Grukemeier C, Starr A. Aortic valve replacement with the Starr-Edwards prosthesis: a comparison of the first and second decades of follow-up. Ann Thorac Surg 1988;45:248 –252. 16. Van der Meulen JHP, Steyerberg EW, vand der Graaf Y, van Herwerden LA, Verbaun CJ, Defauw JJAMT, Habbema JDF. Age thresholds for prophylactic replacement of Bjork-Shiley convexo-concave heart valves: a clinical and economic evaluation. Circulation 1993;88:156 –164. 17. Rams JJ, Davis DA, Lolley DM, Berger MP, Spencer M. Detection of microemboli in patients with artificial heart valves using transcranial Doppler: preliminary observations. J Heart Valve Dis 1993;2:37– 41. 18. Stein PD, Sabbah HN, Pitha JV. Continuing disease process of calcific aortic stenosis: role of microthrombi and turbulent flow. Am J Cardiol 1977;30:159 – 163. 19. Cerebral Embolism Task Force. Cardiogenic brain embolism. Arch Neurol 1989;46:727–743. 20. Tsai TP, Matloff J, Chaux A, Kass RM, Lee ME, Czer LSC, De Robertis MA, Gray RJ. Combined valve and coronary artery bypass procedures in septuagenarians and octogenarians: results in 120 patients. Ann Thorac Surg 1986;42:681– 684.

APRIL 1, 1999