Multiple mechanical valve replacement surgery comparison of St. Jude Medical and CarboMedics prostheses

European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 Multiple mechanical valve replacement surgery comparison of St. Jude Medical and CarboMedics prostheses W.R. Eric Jamieson *, A. Ian Munro, Robert T. Miyagishima, Gary L. Grunkemeier, Lawrence H. Burr, Samuel V. Lichtenstein, G. Frank O. Tyers Department of Surgery, St. Paul s Hospital-Heart Centre, Vancou er General Hospital, Uni ersity of British Columbia, 3100-910 West 10th A enue, Vancou er V5Z 4E3, Canada Received 28 July 1997; received in revised form 8 December 1997; accepted 16 December 1997 Abstract Objective: The experience with the St. Jude Medical (SJM) and CarboMedics (CM) bileaflet mechanical prostheses was evaluated to determine thromboembolic and hemorrhagic complications and predictive risk factors. Methods: From 1989 to 1994, a total of 246 patients had multiple valve replacement (SJM, 140; CM, 106); concomitant procedures 20.3% (50) [coronary artery bypass 10.6% (26)] and 53.7% (132) previous cardiac surgery, primarily valve replacement procedures. The pre-operative variables [coronary artery disease, previous cardiovascular surgery, concomitant procedures, valve lesion (except mitral stenosis), status, atrial fibrillation, and NYHA III/IV] did not distinguish the prosthesis-type (p). Results: The prosthesis-type and the pre-operative variables, including atrial fibrillation, were not predictive of overall thromboembolism (TE). The linearized rate of total TE events for overall multiple replacements (MR) was 5.4%/patient-year (minor, 2.52; major 2.85); the total TE for CM and SJM was 5.4%/patient-year, respectively. The 30 day major TE crude rate was 0.82%, while the 30 day major event rate was 2.7%/patient-year. Of the total (major and minor) TE events 100% (3) of 30 days and 72% (29), 30 days had an INR 2.5 at or immediately prior to the event. The thrombosis rate (included in total TE events) was 0.67%/patient-year (4 events, 100% INR 2.5). Of the various TE event categories the prosthesis-types (CM and SJM) were not differentiated (p). The freedom, at 5 years, from major/fatal TE, thrombosis and hemorrhage from anticoagulation was 89.3 3.8% for CM and 87.9 3.7% for SJM and, 91.3 3.5% and 89.3 3.7%, respectively, (p) exclusive of early events. Conclusions: The performance of the CarboMedics and St. Jude Medical prostheses in multiple valve replacement surgery in this non-randomized prospective study revealed no significant differences in performance with regard to thromboembolic and hemorrhagic complications. 1998 Elsevier Science B.V. All rights reserved. Keywords: Mechanical prostheses; Thromboembolism; Predictive factors and anticoagulation 1. Introduction The first mechanical prosthesis, the Starr Edwards caged-ball prosthesis (Baxter Healthcare, Edwards CVS Division, Irvine, CA) was introduced 35 years ago. The developments in mechanical prostheses have incorporated both monoleaflet and bileaflet designs, both formulated with pyrolytic carbon in whole or in part. The * Corresponding author. Tel. +1 604 8754141; fax: +1 604 8754036. two predominant bileaflet mechanical prostheses are the St. Jude Medical (St. Jude Medical, Minneapolis, MN) and the CarboMedics (Sulzer-Carbomedics, Austin, TX). The St. Jude Medical prosthesis was introduced in 1977 and has been the most widely implanted mechanical bileaflet prosthesis. The documentation with the St. Jude Medical prosthesis primarily reveals the 10-year experience [1 12]. The CarboMedics prosthesis was introduced as an investigational prosthesis in 1986 and received approval to the marketplace in 1993. The CarboMedics prosthesis has an increasing number of 1010-7940/98/$19.00 1998 Elsevier Science B.V. All rights reserved. PII S1010-7940(97)00323-0

152 W.R.E. Jamieson et al. / European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 published clinical studies [13 18]. There has been no randomized or non-randomized trials comparing the clinical performance of bileaflet prostheses. The purpose of this communication is to document the thromboembolic rate of the bileaflet mechanical prostheses, St. Jude Medical and CarboMedics, in a prospective non-randomized study comparing the experience between the prostheses in multiple valve replacements and to determine predictors of performance. 2. Patient Population From September 1989, the time of commencement of the University of British Columbia in the United States-Food and Drug Administration CarboMedics investigational study, to June 1994 the CarboMedics (CM) and St. Jude Medical (SJM) prostheses were evaluated in a non-randomized prospective study. The prostheses were implanted, generally, by different surgeons within the teaching hospitals of the University of British Columbia. Multiple valve replacements were performed in 246 patients (male, 108; female, 138). The mean age of the population was 57.6 13.3 (S.D.) years (range 23 80 years). There were 52.4% NYHA class III and 33.3% class IV patients. The mean follow-up was 2.42 1.70 (S.D.) years, a total of 596.4 years. The median followup was 2.32 years. The completeness of follow-up was 99.6%. The overall patient population comprised 217 (CM, 98; SJM, 119) aortic valve replacements (AVR) and mitral valve replacements (MVR); two (CM, one; SJM, one) AVR and tricuspid valve replacements (TVR); 11 (CM, three; SJM, eight) MVR and TVR; two (SJM, two) AVR and pulmonary valve replacements; 12 (CM, three; SJM, nine) AVR, MVR and TVR; and two (CM, 1; SJM, 1) AVR and MVR and pulmonary valve replacements. The patient population incorporated 12 porcine bioprostheses implanted in the tricuspid and pulmonary positions. There were also three Starr-Edwards prostheses and one Duromedics mechanical prosthesis. Concomitant procedures were performed in 20.3% (50) of the patients, coronary artery bypass in 10.6% (26) of the patients. Of the multiple population, 53.7% (132) had previous cardiac procedures, primarily valve replacement procedures. In these multiple replacements (MR) the CM was implanted in 106 patients (male, 49; female, 57) and the SJM in 140 patients (male, 59; female, 81). The mean age of the MR-CM population was 57.0 14.1 (S.D.) years (range 23 78 years) and for SJM, 58.1 12.6 (S.D.) years (range 24 80 years). (p) The MR-CM patients were 57.5% NYHA class III and 31.1% class IV, while for SJM patients, 48.6% and 35.0%, respectively (p). The mean follow-up for MR-CM was 2.29 1.68 (S.D.) years (total 242.4 years) and for SJM was 2.53 1.72 (S.D.) years (total 354.0 years). The median follow-up was 2.04 years for CM and 2.48 years for SJM. The completeness of follow-up was 100 and 99.3%, respectively. The patient populations for MR for the CM and SJM groups were essentially the same considering the incidence of coronary artery disease, atrial fibrillation, previous cardiovascular surgery, concomitant procedures, valve pathology and status of procedures (p) (Table 1). The incidence of mitral stenosis (CM, 27.3%; SJM, 15.7%) did distinguish the populations (P= 0.038). 3. Methods and Statistical Analysis The Guidelines for Reporting Morbidity and Mortality After Cardiac Valvular Operations were used for definitions of valve-related complications, categorization and statistical methods [19]. The recommended definitions of thromboembolic and hemorrhagic events proposed by Bodnar et al. [20]. in the article entitled Proposal for Reporting Thrombosis, Embolism and Bleeding after Heart Replacement have been incorporated in the evaluation. The thromboembolic complications have been categorized by these authors, as follows: 1. valve thrombosis; 2. cerebrovascular accident as occlusive (ischemic) or hemorrhage in origin, in terms of transient ischemic attack ( 24 h), reversible ischemic neurological deficit (RIND-minor stroke) ( 72 h with complete resolution within 3 weeks), major stroke; 3. systemic (non-cerebral) embolism; 4. systemic (non-cerebral) bleeding. The information on valve-related complications for this study was obtained by direct contact with patients, family physicians, consultants, hospital health records, and official death registries over a closing interval of 6 months. The valve-related complications and composite indexes of these valve-related complications were evaluated in the time-related manner by actuarial life-table techniques (Cutler Ederer method). The Wilcoxon Gehan statistic was used to provide comparison of complication-free curves. Linearized occurrence rates (events per 100 patient-years or percent per patientyear) were utilized specifically for documenting thromboembolic and hemorrhagic complications. The instantaneous hazard functions were determined from combined (CM and SJM) life tables and were demonstrated as bar grafts for thromboembolic and hemorrhagic events and valve-related mortality and reoperation. The predictors of early mortality were

W.R.E. Jamieson et al. / European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 153 Table 1 Multiple valve replacementcomparison of groups (CarboMedics and St. Jude Medical prostheses Variables n (%) Carbomedics (n=6) St Jude Medical (n=140) P-value Coronary artery disease 9 (8.5%) 17 (12.1%) Previous CV surgery 51 (48.1%) 81 (57.9%) Concomitant procedures 24 (22.6%) 26 (18.6%) Atrial Fibrillation 49 (46.2%) 69 (49.3%) Valve lesion Aortic regurgitation 49 (46.2%) 66 (47.1%) Aortic stenosis 34 (32.1%) 33 (23.6%) Aortic mixed 12 (11.3%) 23 (16.4%) Mitral regurgitation 67 (63.2%) 91 (65.0%) Mitral stenosis 29 (27.3%) 22 (15.7%) 0.038* Mitral mixed 7 (6.6%) 20 (14.3%) Status Elective 86 (81.1%) 116 (82.9%) Urgent 17 (16.1%) 19 (13.6%) Emergent 3 (2.8%) 5 (3.6%) ** NYHA III/IV 94 (88.6%) 117 (83.6%) 2 -test. * Yates corrected 2 -test. ** Fisher s exact test. determined by stepwise logistic regression, and the predictors of late mortality and thromboembolism by the Cox proportional hazard regression model. The variables evaluated were prosthetic valve type, coronary artery bypass surgery, NYHA IV, age, gender, atrial fibrillation, urgency status and previous valvular surgery. SPSS for MS Windows release 6.1.3 software was used for analysis. 4. Results The early and late mortality and mortality from valve-related complications are detailed in Table 2 for the overall population (including the 95% confidence limits, CL) and by prostheses types. The early mortality for the total population was 12.2% (30) (CM, 15.1%; SJM, 10.2% p). There were three (10%) valve-related deaths while cardiac causes were 40% (12) and non-cardiac causes and surgical causes were 50% (15). The risk factors of early mortality were NYHA IV, and advancing age (Table 3). The late mortality for the overall patient population was 3.85%/patient-year (23). The valve-related rate of late mortality was 2.2%/patient-year. There were four deaths due to major thromboembolism, two due to prosthesis thrombosis and one due to hemorrhage. There were also two due to prosthetic valve endocarditis and four due to non-structural dysfunction. The patient survival, at 5 years, was 74.3 5.6%. The risk factors of late mortality were advancing age, NYHA IV and concomitant coronary artery bypass (Table 3). The valve-related complications are summarized in Table 4 for overall multiple valve replacement and by prosthesis-type (CM and SJM). The freedom from valve-related mortality, excluding early mortality and sudden unexpected deaths was, at 5 years, 92.6 2.1%. The only risk factor of valve-related mortality was combined thromboembolism and hemorrhage. The experience from valve-related reoperation for MR is shown in Table 5. There was three fatality for 15 reoperations (20%). The freedom from reoperation for the overall population was 93.5 1.8% at 3 years and 91.1 3.0% at 5 years. There were no risk factors predictive of reoperation. The major consideration of the clinical report is the assessment of thromboembolic complications. Tables 4 and 6 detail thromboembolic and hemorrhagic complications for the overall MR population but also by the prosthesis-types (CM and SJM). The incidence of total thromboembolic events was 5.4%/patient-year, a total of 32 events (Table 4). The linearized rate for minor events was 2.52%/patient-year (15) and for major events was 2.85%/patient-year (17). There were four thrombotic events or 0.67%/patient-year. There were no risk factors of overall thromboembolism, valve type was not a factor (Table 3). The overall linearized rate of hemorrhage was 1.34%/patient-year with one fatality (Table 4). The major thromboembolic events and major anticoagulant complications have been considered for two time intervals, 30 days and 30 days (Table 6). The overall major events ( 30 days) (major, reversible,

154 W.R.E. Jamieson et al. / European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 Table 2 Mortality and mortality from valve-related complications CM rate (event)* SJM rate (event) Total rate (event) 95% CI Early mortality 15.1% (16) 10.0% (14) 12.2% (30) Late mortality 3.71 (9) 3.96 (14) 3.85 (23) 2.44 5.79 Death-related failure 1.24 (3) 0.57 (2) 0.84 (5) 0.27 1.96 Sudden death 0.28 (1) 0.17 (1) 0.004 0.93 Valve-related death Complications 2.06 (5) 2.26 (8) 2.18 (13) 1.16 3.73 Thromboembolism 1.13 (4) 0.67 (4) 0.18 1.72 Thrombosis 0.41 (1) 0.28 (1) 0.34 (2) Hemorrhage 0.41 (1) 0.17 (1) 0.004 0.93 Prosthetic valve endocarditis 0.82 (2) 0.34 (2) Structural valve deterioration Non structural dysfunction 0.41 (1) 0.85 (3) 0.67 (4) 0.18 1.72 Reoperation 0.14 (1) 0.57 (2) 0.50 (3) 0.10 1.47 * All rates except early mortality are calculated as % per patient-year. Table 3 Multiple valve replacementpredictors of performance Valve type Atrial fibrillation CAB NYHA III/IV Age Sex Urgency status Previous valve surgery Early mortality Late mortality TE (overall) TE (excluding early events) CAB, coronary artery bypass; TE, thromboembolism. thrombosis and fatal) was 2.7%/patient-year. Of the 246 patients the number of major events 30 days were one major TE reversible, nine major TE permanent, two systemic TE major, four thrombosed prostheses and of the total five fatalities. There were six hemorrhagic events, with one fatality, requiring hospitalization or outpatient transfusions. Within the hospital or early period 30 days there were two major thromboembolic events, a crude rate of 0.81% with one fatality and two major hemorrhagic events, 0.81%. The Figs. 1 and 2 illustrate the freedom for MR of thromboembolic and hemorrhagic events overall and exclusive of early events ( 30 days). The figures show the freedoms from major thromboembolism, and overall major/fatal thromboembolism, thrombosis and hemorrhagic events. The freedom from overall major thromboembolism was, at 5 years, 91.2 2.6% and 92.0 2.6% with exclusion of early events (CM, 94.5 2.8%; SJM, 89.9 3.5%; and exclusion of early events 94.5 2.8% and 91.3 3.4%, respectively) (p) (Fig. 1). The freedom from thrombosis, at 5 years, was 98.5 0.8%. The assessment of overall events (Fig. 2) revealed 5 year freedom of 87.7 2.9% and 88.5 3.0% with exclusion of early events (CM, 89.3 3.8%; SJM, 87.9 3.7%; and exclusion of early events 91.3 3.5% and 89.3 3.7%, respectively) (p). The instantaneous hazard rates of overall thromboembolic and hemorrhagic events are demonstrated in Fig. 3 showing six-monthly rates. The instantaneous hazard rates of valve-related mortality and reoperation, including mortality from reoperation is illustrated in Fig. 4, also showing six-monthly rates. The anticoagulant status of patients who experienced thromboembolic events at the various time intervals was determined. The recommended INR range to patients and family physicians was 3.0 3.5. There was one in-hospital event and that person was considered inadequately anticoagulated considering an INR of 2.5 as the minimal level of anticoagulant adequacy. There were 31 post-hospital events and 76% were inadequately anticoagulated. Of the 32 events, three occurred 30 days and all were inadequately anticoagulated while of 29 30 day events 72% were not on adequate anticoagulation. The rhythm at the time of anticoagulation inadequacy was atrial fibrillation, 46%, and normal sinus rhythm, 54%. There were four episodes of prosthesis thrombosis, 0.68%/patient-year (Tables 4 and 6), 1 patient had both a thrombosed tricuspid prosthesis managed by explant surgery and subsequently a fatal thrombosed mitral prosthesis identified at autopsy, both CM prostheses. All these patients occurred out of hospital beyond the 1

W.R.E. Jamieson et al. / European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 155 Table 4 Valve-related complicationsmultiple valve replacement CM rate (event)* SJM rate (event) Total rate (event) 95% CI Total TE (incl. Thrombosis) 5.36 (13) 5.37 (19) 5.37 (32) 3.67 7.57 Major TE and thrombosis (incl. Reversible)overall 2.89 (7) 2.83 (10) 2.85 (17) 1.66 4.56 Minor TE 2.47 (6) 2.54 (9) 2.52 (15) 1.41 4.15 Major TE and thrombosis (incl. Reversible) 30 days 2.89 (7) 2.26 (8) 2.52 (15) 1.41 4.15 Fatal TE (excl. thrombosis) 1.13 (4) 0.67 (4) 0.18 1.72 Thrombosisoverall 1.24 (3) 0.28 (1) 0.67 (4) 0.18 1.72 Fatal thrombosis 0.41 (1) 0.28 (1) 0.34 (2) Hemorrhageoverall 2.06 (5) 0.85 (3) 1.34 (8) 0.58 2.64 Hemorrhage 3 days 1.24 (3) 0.85 (3) 1.01 (6) 0.37 2.19 Prosthetic valve endocarditis 0.82 (2) 1.13 (4) 1.01 (6) 0.37 2.19 Non structural dysfunction 2.47 (6) 2.83 (10) 2.86 (16) 1.53 4.36 Reoperation 3.71 (9) 1.70 (6) 2.52 (15) 1.41 4.15 * Rates are calculated as % per patient-year. Table 5 Valve-related reoperationmultiple valve replacement CM rate (event)* SJM rate (event) Total rate (event) 95% CI Complication Thrombosis 0.82 (2) 0.28 (1) 0.50 (3) 0.10 1.47 Thromboembolism Hemorrhage Prosthetic valve endocarditis 0.28 (1) 0.17 (1) 0.004 0.93 Non structural dysfunction 2.47 (6) 1.41 (5) 1.84 (11) 0.92 3.30 Structural valve deterioration Rates are calculated as % per patient-year., represents no events. month period. One of the patients was managed initially with thrombolytic therapy. A total of 3 patients were managed with reoperation with one fatality. One of these patients, noted above, subsequently died with thrombosis of the mitral prosthesis. The additional patient died without reoperation, thrombosis identified at autopsy. Of the three reoperations, one had explant of the thrombosed tricuspid prosthesis without replacement and one thrombectomy and one valve replacement which resulted in the fatality. The patients were inadequately anticoagulated at the time of the event, three in atrial fibrillation and one in normal sinus rhythm. 5. Discussion Thromboembolism, including thrombosis, and anticoagulant hemorrhage remain the major valve-related complications of mechanical prostheses. The purpose of this study is to evaluate the thromboembolic and hemorrhagic complications of the St. Jude Medical and CarboMedics mechanical prostheses in multiple valve replacement. The St. Jude Medical prosthesis has received documentation of clinical performance to 10 years [1 12]. The CarboMedics prosthesis has an increasing number of publishing clinical studies [13 18]. There has been no randomized or non-randomized trials comparing the clinical performance of the two bileaflet prostheses. The University of British Columbia commenced utilization of the CarboMedics prosthesis in September 1989 as part of the multicentre investigational study. The experience with the CarboMedics and St. Jude Medical prostheses between September 1989 and June 1994 inclusive formulates the patient population of this study. The study was conducted as a non-randomized prospective evaluation. The study, although early, revealed no significant difference in the clinical performance of the two prostheses in multiple replacements. The prostheses were not differentiated with regard to the incidence of overall thromboembolism. The current knowledge of mechanical protheses has been derived primarily from recent reports on the clinical performance of individual prostheses [13,1,2,14 16,3,17,4 11,21,22,12,18]. There have been only two randomized trials, of significance, both comparing the performance of the St. Jude Medical prosthesis to the Medtronic-Hall monoleaflet prosthesis [23,24]. The major concern with randomized reports has been the wide

156 W.R.E. Jamieson et al. / European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 Table 6 Thromboembolic and hemorrhagic events 30 days and 30 daysmultiple replacement CM rate (event)* SJM rate (event) Total rate (event) 95% CI Major events 30 days CVA TE reversible 0.28 (1) 0.17 (1) 0.004 0.93 CVA TE major (non fatal) 0.82 (2) 1.13 (4) 1.01 (6) 0.37 2.19 CVA TE major (fatal) 0.85 (3) 0.50 (3) 0.10 1.47 Systemic TE major 0.82 (2) 0.34 (2) Thrombosis (non fatal) 0.82 (2) 0.34 (2) Thrombosis (fatal) 0.41 (1) 0.28 (1) 0.34 (2) Hemorrhage (non fatal) 0.82 (2) 0.85 (3) 0.84 (5) 0.27 1.96 Hemorrhage (fatal) Major events 30 days 0.41 (1) 0.17 (1) 0.004 0.93 CVA TE major (non fatal) 0.28 (1) 0.41 (1) CVA TE major (fatal) 0.28 (1) 0.41 (1) Thrombosis Hemorrhage (non fatal) 0.82 (2) 0.82 (2) * Rates for major events 30 days are calculated as % per patient year. 30 days is crude rate. variation of reported results on thromboembolic events [23,25,24,26]. There remains extensive inconsistency of clarification of definition, for most reports do not differentiate between major and minor events and do not indicate the standard of reporting. For multiple valve replacement with mechanical prostheses the incidence of thromboembolic rates have ranged between 0.3 and 6.6%/patient-year [13 16,4,6,10 12]. Smith and coauthors [12] have reported the incidence of thromboembolic events to be 0.3%/patient-year for double valve replacement for the St. Jude Medical prosthesis. On the other hand, Horstkotte and colleagues [4] have reported unexpected findings with the St. Jude Medical prosthesis with an incidence of 6.62%/patient-year for double valve replacements. These authors when considering only first or most serious of several events, documented incidence of 5.40%/patient-year for double valve replacement. Ibrahim et al. [6] have also reported high total thromboembolic rates with the St. Jude Medical prosthesis, 5.0%/patient-year for multiple valve replacements. Several reports document only events that occur beyond the early 30 day interval and consequently provide an unfair assessment of patient morbidity and mortality [2,14,15,3,4,7]. Jegaden and investigators [8], on the other hand, have expressed concern over the occurrence of major thromboembolism during the early postoperative period, often when anticoagulation has yet to be adequately controlled. There also is limited information provided on the level of anticoagulation at the time of both thromboembolic and hemorrhagic events. The extent of patient follow-up and depth of evaluation, as well as the adequacy of anticoagulation management are likely contributing factors to the wide range of reported events. This report attempts to alleviate some of these concerns by reporting total overall events, events occurring 30 days, events occurring 30 days and classifying events as major (reversible and permanent), minor, fatal and thrombotic occlusion. The study has documented the adequacy of anticoagulation at the time of event occurrence. An INR level of 2.5 was considered as the minimal level of adequacy of anticoagulant management. The study did not determine the anticoagulant status of patients who did not have a thromboembolic event, or at times remote of an event. The overall thromboembolic rate, considering both major and minor events, for multiple replacements was 5.4%/patient-year. These results are comparable to the reported results by Horstkotte et al. [4] and Ibrahim et al. [6] with the St. Jude Medical prosthesis. These results are at variance to the literature for the St. Jude Medical prosthesis, the range from other reports was 0.3 3.2%/patient-year for multiple replacements [9 12]. The results with the CarboMedics prosthesis was 0.3 3.10%/patient-year for multiple replacements [13,15,16]. This wide ranging reporting of thromboembolic rates must be related to the detailed nature of evaluation and possibly the extent of anticoagulation. The most important aspect of thromboembolism is the rates of major events, as well as, hemorrhagic events. The major event rate in the present series of combined St. Jude Medical and CarboMedics prostheses is 2.85%/patient-year for multiple replacements. The 30 day minor event rate was 2.52%/patient-year. The reports by Horstkotte and Ibrahim and colleagues [4,6] range from 1.5 1.6%/patient-year for multiple replacements. The incidence of anticoagulant hemorrhage in these reports has shown less variation. The major hemorrhagic rate in the authors combined series is 1.34%/ patient-year for multiple valve replacement. The hemorrhagic rates for the series of St. Jude Medical prostheses by Horstkotte and Ibrahim and colleagues

W.R.E. Jamieson et al. / European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 157 Fig. 1. Multiple valve replacementfreedom from major thromboembolism overall and with exclusion of early events. [4,6] ranges from 1.9 2.1%/patient-year for multiple replacements. These apparent differences may explain the higher major event rates in the authors series. The inadequacy of anticoagulation in our series of overall thromboembolic events may also be a contributing factor; 100% and 72% for multiple replacements 30 days and 30 days, respectively. The prostheses thrombosis rates for our combined series was 0.68%/patient-year for multiple replacements. The thrombosis rate, reported by Orszulak and coauthors [22] for Björk Shiley standard and concave-convex prostheses was 0.27%/patient-year. Aagaard et al. [13]. reported a CarboMedics rate of 0.3%/patient-year. Nitter-Hauge et al. [21]. documented the thrombosis rate for the Medtronic-Hall monoleaflet prosthesis at 0.2%/patient-year for multiple replacements. The rate of prosthesis thrombosis with the St. Jude Medical and the Björk Shiley Monostrut has been reported as low [1,24,6,8 11,22,12]. In the authors series of the four multiple replacement events in 3 patients, 1 patient had both tricuspid and mitral thrombosis. The tricuspid thrombosis was managed with explantation and no replacement while the mitral thrombosis resulted in mortality. There were three successful reoperations and one additional fatality. Of the total four thrombosed prostheses, all four had INR levels of 2.5 at or immediately prior to the event. The fatality rate from thromboembolism has been infrequently reported. The rate in the authors combined series was 0.50%/patient-year for multiple replacements. The reported fatality rates for multiple replacements is 0.2 0.7%/patient-year [1,6,10 12]. Sudden unexpected death has been considered to be a valve-related mortality [27]. In our series the incidence of sudden unexpected death was 0.17%/patient-year for multiple valve replacement. The incidence of sudden death with the St. Jude Medical prosthesis ranges from 0.06 0.9%/patient-year for multiple replacement [1,8,10,11]. Butchart [27] reported it was illogical to attribute all sudden deaths to the prosthesis in the absence of an autopsy. The authors study revealed no influence of prosthesis type on overall thromboembolism but late major TE rates were high (2.7%/patient-year for multiple replacements) and higher including early events 30 days. The thrombosis rates were high, as well as in other reported series [1,6,8,11,12]. The commencement of anticoagulation, at our hospitals, on the 3rd to 4th day without heparinization may be less than optimal. The Fig. 2. Multiple valve replacementfreedom from major/fatal TE, thrombosis and hemorrhage overall and with exclusion of early events. Fig. 3. Multiple valve replacementhazard function of thromboembolism and hemorrhagic events.

158 W.R.E. Jamieson et al. / European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 Fig. 4. Multiple valve replacementhazard function of valve-related mortality and reoperation. long-term management of anticoagulation by numerous family physicians and internists may require reconsideration and possibly home anticoagulation management [28 30]. Irrespective of these concerns our results parallel the embolic rates confirmed with well conducted studies of Horstkotte and Ibrahim and colleagues [4,6]. The concluding comment is that mechanical prostheses in multiple replacement carry a high risk of embolic and hemorrhagic complications but the performance is not different with the CarboMedics and St. Jude Medical prostheses. References [1] Aoyagi S, Oryoji A, Nishi Y, Tanaka K, Kosuga K, Oishi K. Long-term results of valve replacement with the St. Jude Medical valve. J Thorac Cardiovasc Surg 1994;108:1021 9. [2] Baudet EM, Puel V, McBride JT, Grimaud JP, Roques F, Clerc F, Roques X, Laborde N. Surgery for acquired heart disease: long-term results of valve replacement with the St. Jude Medical prosthesis. J Thorac Cardiovasc Surg 1995;109:858 70. [3] Fernandez J, Laub GW, Adkins MS, Anderson WA, Chen C, Bailey BM, Nealon LM, McGrath LB. Early and late-phase events after valve replacement with the St. Jude Medical prosthesis in 1200 patients. J Thorac Cardiovasc Surg 1994;107:394 407. [4] Horstkotte D, Schulte H, Bircks W, Strauer B. Unexpected findings concerning thromboembolic complications and anticoagulation after complete 10 year follow up of patients with St. Jude Medical prostheses. J Heart Valve Dis 1993;2:291 301. [5] Horstkotte D, Schulte HD, Bircks W, Strauer BE. Lower intensity anticoagulation therapy results in lower complication rates with the St. Jude Medical prosthesis. J Thorac Cardiovasc Surg 1994;107:1136 45. [6] Ibrahim M, O Kank H, Cleland J, Gladstone D, Sarsam M, Patterson C. The St. Jude Medical prosthesis. A 13-year experience. J Thorac Cardiovasc Surg 1994;108:221 30. [7] Isomura T, Hisatomi K, Hirano A, Kosuga K, Ohishi K. The St. Jude medical prosthesis in the mitral position. Eur J Cardiothorac Surg 1994;8:11 4. [8] Jegaden O, Eker A, Delahaye F, Montagna P, Ossette J, durand de Gevigney G, Mikaeloff PH. Thromboembolic risk and late survival after mitral valve replacement with the St. Jude Medical valve. Ann Thorac Surg 1994;58:1721 8. [9] Khan S, Chaux A, Matloff J, Blanche C, DeRobertis M, Kass RR, Tsai TP, Trento A, Nessim S, Gray R, et al. The St. Jude Medical valve: experience with 1000 cases. J Thorac Cardiovasc Surg 1994;108:1010 20. [10] Kratz JM, Crawford FA Jr., Sade RM, Crumbley AJ, Stroud MR. St. Jude prosthesis for aortic and mitral valve replacement: a 10-year experience. Ann Thorac Surg 1993;56:462 8. [11] Nakano K, Koyanagi H, Hashimoto A, Kitamura M, Endo M, Nagashima M, Tokunaga H. Twelve years experience: 12 years experience with the St. Jude Medical valve prosthesis. Ann Thorac Surg 1994;57:697 703. [12] 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(5 Pt 2):II49 54. [13] Aagaard J, Hansen CN, Tingleff J, Rygg I. Seven-and-a-half years clinical experience with the CarboMedics prosthetic heart valve. J Heart Valve Dis 1995;4:628 33. [14] Copeland JG 3rd. An international experience with the CarboMedics prosthetic heart valve. J Heart Valve Dis 1995;4:56 62. [15] Copeland JG 3rd, Sethi GK, Bassett JS, North Amercian Team of Clinical Investigators. Four year experience with the CarboMedics valve: the North American experience. Ann Thorac Surg 1994;58:630 8. [16] de Luca L, Vitale N, Giannolo B, Cafarella G, Piazza L, Cotrufo M. Mid-term follow-up after heart valve replacement with Carbomedics bileaflet prostheses. J Thorac Cardiovasc Surg 1993;106:1158 65. [17] Fiane AE, Saatvedt K, Svennevig JL, Geiran O, Nordstrand K, Frfysaker T. The CarboMedics valve: midterm follow-up with analysis of risk factors. Ann Thorac Surg 1995;60:1053 8. [18] Subotic SL, O Grenovic B, Kusic A, et al. CarboMedics heart valve prosthesis. 4 years experience. J Cardiovasc Surg 1992;33:7. [19] Edmunds LH, Clark RE, Cohn LH, Grunkemeir GH, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Ann Thorac Surg 1996;62:932 5. [20] Bodnar E, Butchart EG, Bamford J, Besselaar AMPH, Grunkemeier GL, Frater RW. Proposal for reporting thrombosis, embolism and bleeding after heart valve replacement. J Heart Valve Dis 1994;3:120 3. [21] Nitter-Hauge S, Abdelnoor M. Ten-year experience with the Medtronic-Hall valvular prosthesis. Circulation 1989:80I 43. [22] Orszulak TA, Schaff HV, De Smet J-M, Danielson GK, Pluth JR, Puga FJ. Late results of valve replacement with the Björk Shiley valve (1973 to 1982). J Thorac Cardiovasc Surg 1993;105(2):302 311. [23] Antunes MJ. Clinical performance of St. Jude and Medtronic- Hall prostheses: a randomized comparative study. Ann Thorac Surg 1990;50:743 7. [24] Fiore 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. [25] Aramendi JI, Castellanos E, Serano A, Madiedo JB, Ortiz De Salazar A, Agosti J. A comparative study of the St. Jude Medical and Björk Shiley convexo-concave prostheses in isolated valve replacement. J Thorac Cardiovasc Surg 1991;32(5):557 63. [26] Kim YI, Lesaffre E, Scheys I, Stalpaert G, Flameng WJ, Daenen WJ. The Monostrut versus Medtronic Hall prosthesis: a prospective randomized study. J Heart Valve Dis 1994;3:254 9. [27] Butchart EG. The significance of sudden and unwitnessed death after heart valve replacement. J Heart Valve Dis 1994;3:1 4.

W.R.E. Jamieson et al. / European Journal of Cardio-thoracic Surgery 13 (1998) 151 159 159 [28] Butchart EG, Lewis PA, Grunkemeier GL, Kulatilake N, Breckenridge IM. Low risk of thrombosis and serious emboli events despite low-intensity anticoagulation. Experience with 1004 Medtronic Hall valves. Circulation 1988;78:166 77. [29] Cortelazzo S, Finazzi G, Viero P, Galli M, Remuzzi A, Parenzan L, Barbui T. Thrombotic and hemorrhagic complications in patients with mechanical heart valve prosthesis attending an anticoagulation clinic. Thromb Haemost 1993;69:316 20. [30] White RH, McCurdy SA, Von Marensdorf H, Woodruff DE Jr., Letgoff L. Home prothrombin time monitoring after the initiation of warfarin therapy: a randomized, prospective study. Ann Intern Med 1989;III(9):730 7..