The CarboMedics bileaflet prosthetic heart was introduced

The CarboMedics Valve: Experience With 1,049 Implants José M. Bernal, MD, José M. Rabasa, MD, Francisco Gutierrez-Garcia, MD, Carlos Morales, MD, J. Francisco Nistal, MD, and José M. Revuelta, MD Department of Cardiovascular Surgery, Hospital Universitario Valdecilla, Universidad de Cantabria, Santander, Spain Background. The lack of valve rotatability, the structural deterioration, and the rate of valve-related complications with the standard mechanical bileaflet prosthesis led to the development of a new second-generation bileaflet valve in 1986. Methods. Between January 1989 and March 1994, 1,049 CarboMedics valves were implanted in 859 patients. The rotatability was used in 109 mitral prostheses (21.5%) and in 61 aortic prostheses (11.6%). Follow-up was 97.1% complete, with 3,049 patient-years. Results. The hospital mortality was 6.9% for the mitral group, 3.4% for the aortic group, and 10.7% for the double-valve group (p < 0.005). The actuarial survival curve at 5 years was 77.3% 3.6%, 90.1% 2.5%, and 79.2% 3.7% (p 0.0003), freedom from thromboembolism was 89.1% 3.6%, 87.1% 3.8%, and 68.8% 8.2%, freedom from reoperation was 95.9% 1.4%, 98.9% 0.6%, and 94.9% 2.4%, and freedom from valve-related complications was 68.8% 4.1%, 79.5% 3.5%, and 55.3% 5.9% after mitral, aortic, and mitral and aortic valve replacement, respectively. There were five episodes of valve thrombosis, but no structural deterioration occurred. Conclusions. The clinical performance of the Carbo- Medics valve is quite satisfactory, with a low incidence of valve-related mortality and morbidity. The rotatability feature was useful when the native valve was preserved or for repeat valve replacement. (Ann Thorac Surg 1998;65:137 43) 1998 by The Society of Thoracic Surgeons The CarboMedics bileaflet prosthetic heart was introduced for clinical use in 1986. The CarboMedics heart valve is a bileaflet pyrolitic carbon heart valve, different from the St. Jude Medical prosthesis in several characteristics including the Biolite carbon-covered blood contacting surface on the sewing ring, the valve pivot design with absence of pivot guards, the presence of a titanium stiffening ring, and particularly the rotatability of the valve after implantation. This mechanical prosthesis has two leaflets that sit at an angle of 25 degrees in the closed position and open to 53 degrees from the closed to the open position, or 78 degrees from the plane of the ring. The hemodynamic performance both in vitro and in vivo has been reported earlier, with satisfactory results similar to those of other widely used bileaflet prostheses [1 4]. The largest clinical experience with this prosthesis was described by Copeland [5] and the investigators of the international, multicenter study, with an average follow-up of 30.2 months. The CarboMedics valve has been our mechanical prosthesis of choice since 1989, when the first valve was implanted in our institution. We were encouraged by the mentioned characteristics, particularly its rotatability in our patient population with a high rate of valve reoperations. Since then, 1,049 prostheses have been implanted in 859 patients with a total follow-up of 3,049 patientyears. The object of the present study is to analyze the Accepted for publication Aug 25, 1997. Address reprint requests to Dr Bernal, Cirugía Cardiovascular, Hospital Universitario Valdecilla, 39008-Santander, Spain. clinical experience with this second-generation bileaflet valve. Material and Methods Between January 1989 and March 1994, 859 consecutive patients underwent valve replacement with a CarboMedics valve (CarboMedics, Inc, Austin, TX) in our institution and have had at least 12 months of follow-up. This group comprised 386 women, (44.9%) and 473 men (55.1%), with a mean age of 55.2 10.6 years (range, 15 to 78 years). Most of the patients were in the sixth and seventh decade of life (n 594, 69.2%). Mitral valve replacement (MVR) was performed in 330 patients (38.4%), aortic valve replacement (AVR) in 349 patients (40.6%), double-valve replacement (DVR) in 177 patients (20.6%), and isolated tricuspid, pulmonary, or tricuspid and pulmonary valve replacement in 3 patients (0.4%). There were 526 aortic, 507 mitral, 14 tricuspid, and 2 pulmonary valves implanted, for a total of 1,049 CarboMedics prostheses. A primary valve replacement was performed in 569 patients (66.2%) and a repeat valve replacement in the remaining 290 patients (33.8%). A bioprosthesis was explanted in 227 patients, with a mean interval of 94.2 36.9 months. In our early series [6, 7] a bioprosthesis was mainly indicated when valve repair was not feasible, due to our patient population characteristics, not after an adequate anticoagulation regimen after valve replacement. Among the patients having primary valve replacement, 20 patients had a previous closed mitral commissurotomy and 6 had a percutaneous mitral valvuloplasty. 1998 by The Society of Thoracic Surgeons 0003-4975/98/$19.00 Published by Elsevier Science Inc PII S0003-4975(97)01238-1

138 BERNAL ET AL Ann Thorac Surg THE CARBOMEDICS VALVE 1998;65:137 43 Table 1. Cause of Valve Disease a Cause MVR AVR DVR Rheumatic 161 (48.8) 115 (33.0) 85 (48.0) Degenerative/calcification 14 (4.2) 95 (27.2) 2 (1.1) Congenital 2 (0.6) 23 (6.6)... Ischemic 16 (4.8)...... Infective endocarditis 5 (1.5) 25 (7.2) 9 (5.1) Previous valve repair 28 (8.5) 10 (2.9) 5 (2.8) Previous valve replacement 94 (28.5) 77 (22.1) 76 (42.9) Unknown/not reported 10 (3.0) 4 (1.1)... a Numbers in parentheses are percentages. AVR aortic valve replacement; DVR double-valve replacement; MVR mitral valve replacement. According to the New York Heart Association, 32 patients were in functional class I (3.7%), 172 in class II (20.0%), 530 in class III (61.7%), and 125 in class IV (14.6%) preoperatively. Patients in class I were operated on for severe aortic stenosis, progressive left ventricular dysfunction, or significant structural deterioration of a previous bioprosthesis. The cause of valve disease in the primary valve replacement group was predominantly rheumatic heart valve disease. Those 3 patients who had isolated tricuspid, pulmonary, or tricuspid and pulmonary valve replacement were excluded from this study. From the repeat valve replacement group, 43 patients had a previous open valve repair (14.8%) and 247 a previous valve replacement (85.2%). The causes of valve disease in the AVR, MVR, and DVR groups are shown in Table 1. Preoperative clinical data are summarized in Table 2. All patients were operated on with standard cardiopulmonary bypass with moderate hypothermia (28 to 30 C). Myocardial protection was assured with either crystalloid Table 2. Preoperative Clinical Data Variable MVR AVR DVR Age (y) Mean SD 55.4 10.4 55.3 10.9 55.1 9.8 Range 22 75 15 78 24 75 Male/female 125/205 259/90 88/89 NYHA class I & II 15.6% 35.9% 15.7% III 67.8% 55.8% 62.0% IV 16.6% 8.3% 22.3% Previous 71 (21.5%) 25 (7.2%) 31 (17.5%) thromboembolism Sinus rhythm 26.9% 85.3% 30.4% Atrial fibrillation 73.1% 14.7% 69.6% Primary valve 206 (62.4%) 275 (78.8%) 88 (49.7%) replacement Repeat valve replacement 124 (37.6%) 74 (21.2%) 89 (50.3%) DVR double-valve replacement; NYHA New York Heart Associ- AVR aortic valve replacement; MVR mitral valve replacement; ation; SD standard deviation. Table 3. Associated Cardiac Operations Operation MVR AVR DVR Tricuspid commissurotomy 3 1 3 Tricuspid annuloplasty 54 3 20 Mechanical TVR 9... 4 Bioprosthetic TVR... 1 3 CABG 17 29 5 Left atrial thrombectomy 40... 13 ASD a 5...... Congenital disease 1 1 2 Mitral commissurotomy... 10... Mitral valve repair... 4... Aortic root enlargement... 5... Ascending aorta... 14... aneurysmectomy Total 129 (39.1%) 68 (19.5%) 50 (28.2%) a Iatrogenic after percutaneous mitral valvuloplasty. ASD atrial septal defect; AVR aortic valve replacement; CABG coronary artery bypass grafting; DVR double-valve replacement; MVR mitral valve replacement; TVR tricuspid valve replacement. or blood cardioplegia delivered antegrade, retrograde, or both ways. Surgical technique for valve implantation in all positions consisted in 2-0 pledget-supported U- Tycron (Ethicon Ltd, Edinburgh, UK) everting sutures (12 to 15 sutures per valve). For mitral valves, sizes 27 and 29 were most common, and for aortic valves, sizes 21 and 23 were most common. The rotation feature of the CarboMedics valve was used in 109 mitral prostheses (21.5%) and in 61 aortic prostheses (11.6%). Orientation of the mitral valve was most commonly in the anatomic position, whereas the orientation of the aortic prosthesis was most commonly transverse with pivots at the midpoint of the noncoronary cusp and in the commissure between the right and left cusps. In the MVR group, the posterior mitral leaflet was preserved in 84 patients (25.5%), the anterior in 4 (1.2%), and both (transvalvular implantation) in 19 (5.8%). Both leaflets were resected in 172 patients (52.1%), and this information was not reported in 51 instances (15.5%). In the DVR group, the posterior mitral leaflet was preserved in 38 patients (21.5%) and both leaflets were preserved in 3 (1.7%). Nonleaflet preservation or previous valve replacement occurred in 102 patients (57.6%), and leaflet preservation was not reported in 34 patients (19.2%). Associated cardiac procedures are summarized in Table 3. One hundred twenty-nine patients from the MVR group required an associated cardiac procedure (39.1%). Thromboembolic Prophylaxis Anticoagulation with acenocoumarol was started 48 hours after the CarboMedics valve implant. No heparin was given before the procedure. The international normalized ratio was maintained within a range of 2.5 to 3.5 in those patients in sinus rhythm and between 3.0 and 4.5 when atrial fibrillation was present. Patients undergoing coronary artery bypass grafting received aspirin (125 to

Ann Thorac Surg BERNAL ET AL 1998;65:137 43 THE CARBOMEDICS VALVE 139 200 mg per day). However, this thromboembolic prophylaxis was modified according to patient requirement during follow-up. Follow-up The follow-up data were procured over a 6-month period between October 1994 and April 1995. From 802 patients discharged, follow-up was obtained in 779 patients. Those patients were followed up directly in our outpatient clinic, mailed questionnaires, or contacted directly by telephone. The completeness of follow-up during the closing interval was 97.1%. The mean follow-up was 46.2 months, ranging from 12 to 76 months. Cumulative follow-up was 3,049 patient-years. Statistical Analysis All valve-related mortality and complication definitions follow the guidelines approved by the Ad Hoc Liaison Committee for Standardizing Definitions of Prosthetic Heart Valve Morbidity [8]. All continuous variables are presented as mean standard deviation. Basic methods of univariate analysis included the 2 and Student s t test. Actuarial curves were obtained by the life-table method. The Patient Analysis and Tracking System database, version 06.02.03 (Dendrite Clinical Systems, Inc, Portland, OR), was used for collection and analysis of data. Table 4. Causes of Hospital and Late Mortality Cause MVR AVR DVR Hospital mortality 23 (6.96%) 12 (3.44%) 19 (10.7%) Cardiac 12 7 8 Respiratory 1... 3 Hemorrhage 5 3 7 Sepsis 2 2 1 Sudden 3...... Late mortality 24 (7.3%) 10 (2.9%) 10 (5.6%) Cardiac 13 3 5 Embolism 1 1 1 Valve thrombosis 1...... Sepsis 2...... Sudden 3 5 4 Unknown 2...... Hemorrhage... 1... Noncardiac 2...... AVR aortic valve replacement; MVR mitral valve replacement. DVR double-valve replacement; Fig 1. Actuarial survival curve for all patients. Results Hospital and Late Mortality There were 54 hospital deaths, with a total mortality of 6.3%. Mortality for the MVR group was 6.97% (23/330), for the AVR group 3.44%, (12/349), and for the DVR group 10.73% (19/177) (p 0.005). The causes of hospital mortality were most often related to low cardiac output syndrome (n 27) or hemorrhage (n 15). Other causes were sepsis (n 5), respiratory insufficiency (n 4), and sudden unknown death (n 3) (postmortem examination was not allowed by the family) (Table 4). Late death occurred in 44 patients. Late mortality was 7.3% in the MVR group, 2.9% in the AVR group, and 5.6% in the DVR group (p 0.025). The most common cause of late death was cardiac failure (21/44). Twenty-one patients died of valve-related causes (thromboembolism in 3, sudden/ unknown death in 14, mitral valve thrombosis in 1, sepsis in 2, and hemorrhage in 1). Other causes are summarized in Table 4. The actuarial survival curve for CarboMedics valve replacement was 82.8% 1.9% at 5 years (Fig 1). Fiveyear survival was 77.3% 3.6%, 90.1% 2.5%, and 79.2% 3.7% for MVR, AVR, and DVR, respectively (p 0.0003) (Fig 2). There were 27 hospital deaths in the primary valve replacement group (27/567; 4.76%) and 27 deaths in the repeat valve replacement group (27/289; 9.34%) (p 0.025). Late mortality was 4.06% in the primary valve replacement group and 7.27% in the repeat valve replacement group (p 0.005). The actuarial survival at 5 years for primary valve replacement and repeat valve replacement groups was 84.8% 2.5% and 78.4% 2.9%, respectively (p 0.0007) (Fig 3). Anterior leaflet, posterior leaflet, or the entire mitral valve was preserved in 107 patients of the MVR group, with a hospital mortality of 5.6%. The mitral valve was completely resected in 48 patients, with a hospital mortality of 6.3% (p not significant). Valve preservation was done in 41 patients in the DVR group and complete valve Fig 2. Actuarial survival curves for mitral valve replacement (MVR), aortic valve replacement (AVR), and double-valve replacement (DVR).

140 BERNAL ET AL Ann Thorac Surg THE CARBOMEDICS VALVE 1998;65:137 43 Fig 3. Actuarial survival curves for primary valve replacement and repeat valve replacement. resection in 13 patients, with a hospital mortality of 9.8% and 15.1%, respectively (p not significant). Thromboembolism Thromboembolism occurred in a total of 47 patients. There were 10 peripheral episodes and 37 central nervous system occurrences (21 patients with permanent neurologic impairment and 16 with total recovery). Three patients died due to direct thromboembolic occurrences. The actuarial freedom from thromboembolism at 5 years was 89.1% 3.6% for the MVR group, 87.1% 3.8% for the AVR group, and 68.8% 8.2% for the DVR group, none of which displayed statistical significance. Figure 4 displays the actuarial freedom from thromboembolism and Figure 5 the respective hazard function. Anticoagulant-Related Complications Major hemorrhagic episodes requiring hospitalization were reported in 9 patients and were more frequent in the DVR group (2.6%) than in the MVR group (0.3%) or the AVR group (1.1%) (p not significant). All patients with major hemorrhagic episodes required hospitalization with transfusion. Minor anticoagulant-related hemorrhage occurred in 51 patients, with no difference among the three valve groups. These minor anticoagulant-related hemorrhages did not require transfusion. Reoperation Twenty-one patients required a valve reoperation. Reoperation was needed in 9 patients from the MVR group, in Fig 5. Hazard function for thromboembolism for aortic valve replacement (AVR), mitral valve replacement (MVR), and doublevalve replacement (DVR). 4 patients from the AVR group, and in 8 patients from the DVR group. Nonstructural deterioration (paravalvular leak) was the most common cause of reoperation and was found in 10 patients (3 MVR, 2 AVR, and 5 DVR). Reoperation due to valve thrombosis occurred in 5 patients (2 MVR with one death, 1 AVR, and 2 DVR). Three patients with reoperation suffered from early prosthetic valve endocarditis, with one death. Another patient died of prosthetic endocarditis before reoperation. No late prosthetic endocarditis requiring reoperation was observed. In three instances in the MVR group of patients, reoperation was required due to progression of aortic (n 1) or tricuspid (n 1) valve disease between 10 and 36 months after the initial procedure. The third patient underwent cardiac transplantation. Of the 21 patients requiring a reoperation, two late deaths occurred, one in the MVR group and one in the DVR group. Freedom from valve reoperation was 97.1% 0.8% at 5 years for the whole series (95.9% 1.4% for the MVR group, 98.9% 0.6% for the AVR group, and 94.9% 2.4% for the DVR group) (Fig 6). The hazard function for reoperation is shown in Figure 7. Valve-Related Complications Figure 8 shows the actuarial freedom from all valverelated complications. Freedom at 5 years was 68.8% 4.1% for the MVR group, 79.5% 3.5% for the AVR group, and 55.3% 5.9% for the DVR group (p 0.0001). Fig 4. Freedom from thromboembolism curves for aortic valve replacement (AVR), mitral valve replacement (MVR), and doublevalve replacement (DVR). Fig 6. Freedom from reoperation curves for aortic valve replacement (AVR), mitral valve replacement (MVR), and double-valve replacement (DVR).

Ann Thorac Surg BERNAL ET AL 1998;65:137 43 THE CARBOMEDICS VALVE 141 Fig 7. Hazard function for reoperation for aortic valve replacement (AVR), mitral valve replacement (MVR), and double-valve replacement (DVR). Fig 9. Hazard function for valve-related complications for aortic valve replacement (AVR), mitral valve replacement (MVR), and double-valve replacement (DVR). Figure 9 shows the hazard function for valve-related complications. Functional Class At the end of this study, follow-up information had been obtained for 779 patients, 299 from the MVR group, 327 from the AVR group, and 153 from the DVR group. In the MVR group, there were 151 patients in New York Heart Association class I (50.5%), 111 in class II (37.1%), 23 in class III (7.7%), and 14 in class IV (4.7%). In the AVR group, 221 patients were in New York Heart Association class I (67.6%), 88 in class II (26.9%), 12 in class III (3.7%), and 6 in class IV (1.8%). For the DVR group there were 88 patients in class I (57.5%), 46 in class II (30.1%), 11 in class III (7.2%), and 8 in class IV (5.2%). Comment Recent studies indicate that survival and freedom from all valve-related complications are similar when comparing mechanical prostheses and porcine bioprostheses [9, 10]. However, the latter have a greater incidence of structural deterioration, making mechanical prostheses the current valve replacement of choice for most patients. The main disadvantage of mechanical valves lies in their thromboembolic tendencies and, thus, the need to use anticoagulants for life, with the related risk of hemorrhage. In vitro and in vivo studies, both invasive and Fig 8. Freedom from valve-related complications curves for aortic valve replacement (AVR), mitral valve replacement (MVR), and double-valve replacement (DVR). noninvasive, show that the CarboMedics prosthesis has an excellent hemodynamic performance, similar to that of the St. Jude Medical prosthesis, the first-generation bileaflet valve [3, 4]. Most clinical studies of valve prostheses analyze all the data pertaining to valve-related mortality and morbidity, and according to some authors [11, 12], the risks of hemorrhage, hemolysis, endocarditis, nonstructural deterioration, and reoperation do not appear to differ significantly between different mechanical prostheses. The risks of hospital death and late mortality, including the risk of endocarditis and thromboembolism, may be affected more by the different study populations than by the prosthesis itself. However, the incidences of valve thrombosis and structural deterioration are more intrinsically related to each type of mechanical prosthesis. In our experience, hospital mortality was 6.3% and was significantly higher in the DVR group (10.7%) as compared with the MVR group (6.96%) and the AVR group (3.4%). This ratio is similar to that reported by Nitter- Hauge and Abdelnoor [13] for the Medtronic-Hall prosthesis. Consequently, the actuarial survival at 5 years is also significantly better for isolated AVR, compared with isolated MVR and DVR. In our study, primary valve replacement produced better results than repeat valve replacement, in terms of mortality and long-term survival. The proportion of repeat valve replacements in the series analyzed, together with other preoperative risk factors, may justify the different results in terms of hospital and late mortality and actuarial survival among patients receiving different types of prostheses. Although no statistically significant differences were found when the mitral valve was totally or partially preserved or completely resected, lower hospital mortality were observed in the group of patients who had mitral valve preservation. Linearized rates are frequently used to describe the occurrence of valve-related complications. According to the Guidelines for Reporting Morbidity and Mortality After Cardiac Valvular Operations [8], in those events that do not occur on a constant basis in time, the complications must be analyzed as linearized rates for each study period (the hazard function). The results of the largest published series of patients implanted with the CarboMedics valve were presented in 1994 and 1995

142 BERNAL ET AL Ann Thorac Surg THE CARBOMEDICS VALVE 1998;65:137 43 [5, 12]. These groups differ in their mean follow-up (16.1 versus 30.2 months). A significant decrease of the linearized rates is observable between them, even with the small difference in follow-up time. The incidence of anticoagulant-related hemorrhage decreased from 2.36%/patient-year in 1994 to 1.51%/patient-year in 1995, for example. The same is true for other incidence rates, that is, valve thrombosis, hemolysis, and reoperation. In our initial experience with the CarboMedics prosthesis [14], we found that valve-related complications were more frequent during the first years of follow-up. Furthermore, as observed in the different hazard functions for thromboembolism, reoperation, and all valve-related complications, the linearized rates for each year of follow-up do not remain constant. Thromboembolism has been the most frequent valverelated complication in our series, especially in the DVR group. These data contrast significantly with those reported by Smith and associates [15] with the St. Jude prosthesis in a group of 1,184 patients averaging 59 years of age. This group presented actuarial freedom-fromthromboembolism curves at 10 years with values greater than 96% for the three patient groups. The curves reported by Smith and associates are probably similar to those of the human population without prosthetic heart valves. The results of other authors are more logical. Ibrahim and colleagues [16], in a study of 1,184 patients implanted with a St. Jude prosthesis, show actuarial freedom-from-thromboembolism rates of 76% at 13 years for the MVR and DVR groups. These data also correspond to those published by Fernandez and co-workers [17] in a study of 1,200 patients with a St. Jude prosthesis, who had an actuarial freedom-from-thromboembolism curve of 81% at 9 years. Furthermore, the linearized rate of thromboembolism with the St. Jude prosthesis ranges from 0.5%/patient-year, as reported by Smith and coworkers [15], to 6.62%/patient-year, published by Horstkotte and associates [18]. The disparity of the results for the same prosthesis, for example, Smith and associates [15] versus Horstkotte and associates [18], indicates that the specific characteristics of each patient and the method of anticoagulation may play a greater role in thromboembolism than the type of mechanical prosthesis. Anticoagulation-related hemorrhage is closely linked to thromboembolism. As mentioned before [18], those patients maintained at a higher international normalized ratio (3.0 to 4.5) have a greater incidence of anticoagulantrelated hemorrhage, whereas the thromboembolism rate is not significantly higher in groups of patients maintained with an international normalized ratio less than 3.0. In accordance with results reported by Copeland and associates [12], the risks of endocarditis and nonstructural deterioration, such as paravalvular leak, encountered in our series do not appear to be different from those mentioned for other types of prostheses, whether mechanical or biological. Conversely, valve thrombosis requiring emergency reoperation or fibrinolytic treatment is considered to be one of the most serious complications of mechanical prostheses. Although the role of well-managed anticoagulation seems to be essential, some models of mechanical prostheses may be more thrombotic than others. The North American experience with the CarboMedics valve reported 10 cases of thrombosis in 1,228 patients, with a cumulative follow-up of 3,082 patient-years. In all such cases, a level of anticoagulation lower than that recommended was documented. In our experience, only 5 cases of thrombosis were diagnosed in 859 patients with a cumulative follow-up of 3,049 patient-years, also in the presence of lowered levels of anticoagulation. Although Smith and colleagues [15] reported just a single case of thrombosis with a St. Jude prosthesis in 3,075 patient-years, 4 cases of thrombosis with a St. Jude prosthesis were reported by Horstkotte and associates [18] in 4,080 patient-years. Thirty events of thrombosis are discussed in an article by Baudet and colleagues [19] in a series of 1,112 patients. Except for the most optimistic data [15], most of the series of mechanical bileaflet valves present similar results. The study reported by Baudet and colleagues [19] mentioned above is one of the largest series of patients implanted with a St. Jude bileaflet prosthetic valve. With a follow-up of 8,988 patient-years, it showed actuarial curves of freedom from each of the valve-related complications similar to those of our study. Based on published results of mortality, endocarditis, thromboembolism, anticoagulant-related hemorrhage, nonstructural deterioration, and prosthetic thrombosis, we conclude that our outcomes were generally comparable with those reported by other authors for the St. Jude prosthesis. There are two main differences between the two types of mechanical bileaflet prosthetic valves. Although cases of structural deterioration have been described for the St. Jude prosthesis [20 23], none have been reported for the CarboMedics prosthesis. However, the worldwide experience with the St. Jude Medical valve is significantly higher and has longer follow-up than with the Carbo- Medics valve, so a longer experience and follow-up is mandatory to find any significant differences in the behavior of the two bileaflet prosthesis. The largest CarboMedics valve series [5, 12], plus ours, have shown a total absence of structural deterioration. In the St. Jude prosthesis, two types of structural deterioration were reported. Leaflet fracture was reported by Orsinelli and colleagues [20], and leaflet dislocation causing leaflet embolization was described on several occasions [21 23], both for aortic and mitral prostheses. In contrast to the St. Jude prosthesis, the use of a titanium stiffening ring in the valve housing minimizes the chance for orifice distortion and likely explains the absence of structural deterioration in the CarboMedics prosthesis. Another remarkable difference is valve rotatability, which is absent in the St. Jude prosthesis. In our experiment, the rotatability feature of the CarboMedics valve was used in at least 21.5% of the mitral prostheses and in 11.6% of the aortic prostheses. These data correspond to those reported in the CarboMedics North American experience [12]. The rotation was used in many cases out of the need to improve leaflet mobility or to properly orient them, which was particularly useful when the subvalvular ap-

Ann Thorac Surg BERNAL ET AL 1998;65:137 43 THE CARBOMEDICS VALVE 143 paratus was totally or partially preserved. This also applied to situations where the valve was used in patients with prior valve replacements. Not everyone agrees on the need for mechanical prostheses to have the rotation feature, but the recent introduction for clinical use of a new prosthesis marketed by St. Jude Medical, Inc (St. Paul, MN), which has a rotation feature, appears to give significant importance to this feature. The midterm results with the CarboMedics mechanical valve prostheses are similar to those reported by most authors for other bileaflet prostheses, in terms of hospital and late mortality and other valve-related complications, which are frequently due more to the patient than the prosthesis itself. The rate of valve thrombosis is low and always related to inadequate anticoagulation. The great advantages of this second-generation, bileaflet, mechanical valve over its predecessor are the total absence of structural deterioration and, in our opinion, the rotation feature, which is very useful when a decision is made to preserve the mitral leaflets or when a previous prosthesis is explanted. References 1. Chambers J, Cross J, Deverall P, Sowton E. Echocardiographic description of the CarboMedics bileaflet prosthetic heart valve. J Am Coll Cardiol 1993;21:398 405. 2. Ihlen H, Molstad P, Simonsen S, et al. Hemodynamic evaluation of the CarboMedics prosthetic heart valve in the aortic position: comparison of noninvasive and invasive techniques. Am Heart J 1992;123:151 9. 3. Nygaard H, Paulsen PK, Hasenkarn JM, Pedersen EM, Rovsing PE. Turbulent stress downstream of three mechanical aortic valve prostheses in human beings. J Thorac Cardiovasc Surg 1994;107:438 46. 4. Butterfield M, Fisher J, Davies GA, Spyt TJ. Comparative study of the hydrodynamic function of the CarboMedics valve. Ann Thorac Surg 1991;52:815 20. 5. Copeland JG. An international experience with the Carbo- Medics prosthetic heart valve. J Heart Valve Dis 1995;4: 56 62. 6. Bernal JM, Rabasa JM, Vilchez FG, Cagigas JC, Revuelta JM. Mitral valve repair for rheumatic disease: the flexible solution. Circulation 1993;88:1746 53. 7. Bernal JM, Rabasa JM, Cagigas JC, Echevarria JR, Carrion MF, Revuelta JM. Valve-related complications with the Hancock I porcine bioprosthesis: a twelve- to fourteen-year follow-up study. J Thorac Cardiovasc Surg 1991;101:871 80. 8. Edmunds LH Jr, Clark RE, Cohn LH, Grunkemeier GL, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Ann Thorac Surg 1996;62:932 5. 9. Bloomfield P, Wheatley DJ, Prescott RJ, Miller HC. Twelveyear comparison of a Björk-Shiley mechanical heart valve with porcine bioprostheses. N Engl J Med 1991;324:573 9. 10. Hammermeister KE, Sethi GK, Henderson WG, Oprian C, Kim T, Rahimtoola S. A comparison of outcomes in men 11 years after heart-valve replacement with a mechanical valve or bioprosthesis. N Engl J Med 1993;328:1289 96. 11. Blackstone EH, Kirklin JW. Death and other time-related events after valve replacement. Circulation 1985;72:753 67. 12. Copeland JG, Sethi GK, North American team of clinical investigators for the CarboMedics prosthetic heart valve. Four-year experience with the CarboMedics valve: the North American experience. Ann Thorac Surg 1994;58:630 8. 13. Nitter-Hauge S, Abdelnoor M. Ten-year experience with the Medtronic Hall valvular prosthesis. A study of 1,104 patients. Circulation 1989;80(Suppl 1):43 8. 14. Nistal JF, Hurle A, Revuelta JM, Gandarillas M. Clinical experience with the CarboMedics valve: early results with a new bileaflet mechanical prosthesis. J Thorac Cardiovasc Surg 1996;112:59 68. 15. Smith JA, Westlake GW, Mullerworth MH, Skillington PD, Tatoulis J. Excellent long-term results of cardiac valve replacement with the St. Jude Medical valve prosthesis. Circulation 1993;88(Suppl 2):49 54. 16. Ibrahim M, O Kane H, Cleland I, Gladstone D, Sarsam M, Patterson C. The St. Jude Medical prosthesis. A thirteen-year experience. J Thorac Cardiovasc Surg 1994;108:221 30. 17. Fernandez J, Laub GW, Adkins MS, et al. Early and latephase events after valve replacement with the St. Jude Medical prosthesis in 1200 patients. J Thorac Cardiovasc Surg 1994;107:394 407. 18. Horstkotte D, Schulte H, Bircks W, Strauer B. Unexpected findings concerning thromboembotic complications and anticoagulation after complete 10 year follow up of patients with St. Jude Medical prosthesis. J Heart Valve Dis 1993;2: 291 301. 19. Baudet EM, Puel V, McBride JT, et al. Long-term results of valve replacement with the St. Jude Medical prosthesis. J Thorac Cardiovasc Surg 1995;109:858 70. 20. Orsinelli DA, Becker RC, Cuenoud HF, Moran JM. Mechanical failure of a St. Jude Medical prosthesis. Am J Cardiol 1991;67:906 8. 21. Antunes MJ. Mechanical failure of a St. Jude Medical prosthesis. Am J Cardiol 1991;68:841 2. 22. Odell JA, Durandt J, Shama DM, Vythilingum S. Spontaneous embolization of a St. Jude prosthetic mitral leaflet. Ann Thorac Surg 1985;39:569 72. 23. Burckhardt D, Striebel D, Vogt S, et al. Heart valve replacement with the St. Jude medical valve prosthesis: long-term experience in 743 patients in Switzerland. Circulation 1988; 78(Suppl 1):18 24.