Section of tissue or prosthetic valve. A five-year prospective, randomized comparison.

After considerable experience, controversy persists between the selection of a tissue or prosthetic valve. In order to provide a nonbiased comparison, we designed a prospective, randomized study in September of 1970. Ninety-nine consecutive patients with isolated primary single valve replacement were randomized to tissue or prosthetic valves. The Tissue Bank supplied homografts which were either mounted for mitral replacement or used as free grafts for patients with small aortic roots. Starr-Edwards clot-covered prostheses were used for comparison. There was an even distribution of patients by age, sex, valve lesion, and New York Heart Classification attesting to the accuracy of the random selection. Anticoagulants were used in 33 patients who had operative or embolic evidence of atrial cloth or else history of thromboembolism. Thirty-five patients have undergone postoperative catheterization. Causes of death and valve failure are presented on an actuarial basis. Death in the Starr-Edwards patient group was sudden or due to fabric wear with subsequent thromboembolism, hemorrhage, or infection. We conclude that the tissue valve is a better choice for valve replacement. While there is no significant functional or hemodynamic apparent difference between the tissue and prosthetic valve, there is improvement in patient morbidity and mortality rate with the use of homografts, particularly in the aortic position. The primary cause of complications in the homograft is tissue deterioration, and this problem has been markedly reduced with advent of glutaraldehyde fixation. No such advance in solving the problem of host incompatability with the prosthesis has been forthcoming.     

Early Results of Valve Replacement with the Björk-Shiley Convexoconcave Prosthesis

The Björk-Shiley convexoconcave prosthetic valve has design characteristics that may result in a lower incidence of thromboembolic complications than the conventional spherical Björk-Shiley prosthesis. We evaluated the results of valve replacement with the convexoconcave prosthesis in 248 patients receiving 301 prosthetic valves between March, 1979, and June, 1981. One hundred thirteen patients had aortic valve replacement (AVR), 73 had mitral valve replacement (MVR), and 62 had multiple valve replacement. Two hundred nine (84%) were in New York Heart Association Class III or IV. The median duration of follow-up was 13 months, and follow-up information was available for 246 (99%) of the patients. The actuarial incidence of freedom from thromboembolism at two years was 98% in the AVR group, 97% in the MVR group, and 87% in the group having multiple valve replacement. There were no documented episodes of valve thrombosis or mechanical failure and no fatal thromboembolic complications. The absence of valve thrombosis is in marked contrast to the results reported with the spherical disc valve. Although longer follow-up is necessary, it appears that the convexoconcave design represents a major improvement in the Björk-Shiley prosthesis.     

Clinical assessment of the St. Jude medical cardiac prosthesis. A 5-year experience

Mechanical prosthetic valves have established their long-term durability, but the incidence of thromboembolism has been less satisfying. Therefore, the search for the perfect valve continues. The choice of replacing a cardiac valve with a prosthetic valve is based upon its hemodynamics, durability, thromboembolic and hemolytic characteristics. The hemodynamic assessment of the St. Jude Valve demonstrates that it performs well in the aortic, mitral, and tricuspid positions. The central bileaflet flow design leads to minimal gradients, even in the small valves which make it ideal for its use in children. The purpose of its all-pyrolite construction is to decrease the incidence of thromboembolism. Replacement of heart valves with the St. Jude Medical prosthesis was initiated in April, 1979. An analysis of all patients undergoing valve replacement between April 1, 1979, and June 1, 1984, was carried out to assess the durability, incidence of thrombosis, and thromboembolic as well as other complications.     

Comparison of outcomes after mitral valve replacement with a mechanical versus a bioprosthetic valve in patients between forty and sixty years of age

Background  The American College of Cardiology/American Heart Association (ACC/AHA), guidelines for choice of prosthetic valve based on patients’ age are difficult to apply to the developing world because of a lower life expectancy and difficulty in maintaining correct levels of anticoagulation for a variety of reasons. While there is general agreement on the choice of prosthetic valves for patients below 40 years of age (mechanical) and above 60 years of age (biologic), the 40 to 60 age group remains a grey zone. The goal of our study was to compare outcomes after mitral valve replacement with a mechanical versus a bioprosthetic valve in patients between forty and sixty years of age. Methods  From Jan 2003 to July 2008, 250 patients between the ages of 40 and 60, undergoing mitral valve replacement at our institution were randomized to receive either a mechanical or a bioprosthetic valve. Outcomes in the form of incidence of valve thrombosis and thromboembolism, bleeding complications, incidence of prosthetic valve endocarditis and survival were compared in the two groups. Results  Out of 250 patients, 135 patients received mechanical valve and 115 patients were implanted with a bioprosthetic valve. Patients were followed up for a mean period of 3 years (range 6 months to 4.8 years). The incidence of valve thrombosis was higher in mechanical valve as compared to bioprosthetic valve (6% vs. 0.9%, p= 0.04). Similarly there was a higher incidence of thromboembolism in mechanical valves as compared to bioprosthetic valves (4.5% vs. 0%, p=0.03). Bleeding complications occurred more frequently in mechanical than bioprosthetic valve (6% vs. 0.9%, p=0.04). There was no significant difference in the incidence of prosthetic valve endocarditis (2.2% vs. 2.7%, p >0.05) or survival at three years (96.2% vs. 97.2%, p > 0.05) in the two groups. Conclusions  Patients in the age group of 40 to 60 years undergoing mitral valve replacement with a mechanical valve have a higher incidence of thrombotic and bleeding complications as compared to bioprosthetic valve, even though short term survival is similar. This favours implantation of a bioprosthetic valve in this age group.     

Mitral Valve Replacement with the Modified University of Cape Town (UCT) Prosthesis: Clinical and Hemodynamic Results

Mitral valve replacement with the modified University of Cape Town prosthesis was performed in 42 patients. In 35 the procedure was an isolated one, and the hospital mortality was 6%. The late survival rate was 60%, half of the late deaths being the result of thromboembolism or complications of anticoagulant therapy. The incidence of hemolysis was low, and hemodynamic results demonstrated improvement in cardiac index and lowering of pulmonary artery pressure, pulmonary artery wedge pressure, pulmonary arteriolar resistance, and transvalvular mean gradients. However, the calculated prosthetic valve orifice area was lower than the measured area. Because of complications of thromboembolism, the high incidence of late deaths, and high transvalvular gradients, this prosthetic valve is no longer used in patients requiring mitral valve replacement.     

Mid-term surgical results after valve replacement with the carbomedics valve prosthesis

We present herein a review of 95 patients who underwent valve replacement with the CarboMedics (CM) valve prosthesis between April 1990 and December 1992. The operative mortality for the entire group was 4.2%: 0% following aortic valve replacement (AVR), 2.7% following mitral valve replacement (MVR), and 12.5% following double valve replacement (DVR). All patients were prescribed warfarin and bucolome for anticoagulation, and were followed up for a mean period of 29.9 months. Late mortality was 8.4%: 3.1% following AVR, 10.8% following MVR, and 12.5% following DVR. There were no cases of mechanical prosthetic valve failure, significant hemolysis, infective prosthetic valve endocarditis, or bleeding complications. After 44 months of follow-up, the actuarial freedom from complications was calculated as: thromboembolism, 97.8±1.6%; valve thrombosis, 97.8±1.1%; paravalvular leak, 96.7±1.9%; and reoperation, 98.9±1.1%. The overall survival rate was 84.3±6.3% and all survivors showed a significant improvement in NYHA functional class, from 81% in classes III and IV preoperatively to 99% in classes I and II postoperatively. The CM valve exhibited no significant differences in hemolytic parameters or hemodynamic performance after isolated AVR or MVR compared with the similar type of St. Jude Medical bileaflet valve. The evidence provided by the present study therefore suggests that the CM valve prosthesis can achieve excellent mid-term clinical results and hemodynamic performance with a low incidence of thromboembolism and valve thrombosis.     

St. Jude Medical prosthesis: valve-related deaths and complications

Despite the hemodynamic improvement in cardiac valve prostheses, valve-related deaths and complications have remained a substantial source of morbidity and mortality. This follow-up study was carried out in April, 1986, in a group of 816 patients who received 876 St. Jude Medical prostheses between October, 1977, and October, 1984. One hundred fifty-five patients (37%) of the 422 having aortic valve replacement (AVR) and 88 (29%) of the 300 having mitral valve replacement (MVR) had concomitant coronary artery bypass grafting, and 94 patients (12%) had miscellaneous procedures performed in addition to valve replacement. Sixty-three patients died within the operative period, an operative mortality of 7.7%. Follow-up was complete for 98% of the patients; the mean follow-up was 42 months. At that time, 87 more patients had died, a late mortality of 11.9%. There were 13 valve-related deaths (4 in the AVR group, 7 in the MVR group, and 2 in the miscellaneous group); the causes of death were thromboembolism (7), anticoagulant-related hemorrhage (2), valve thrombosis (2), prosthetic valve endocarditis (1), and perivalvular leak (1). There were 37 thromboembolic episodes (1.78/100 patient-years) and 67 anticoagulant-related hemorrhages (3.22/100 patient-years), 46% of which necessitated transfusion. The incidence of valve thrombosis was 0.6% (4 patients). Perivalvular leak and prosthetic valve endocarditis occurred in 7 patients (1.0%) and 5 patients (0.7%), respectively. Reoperation was carried out in 6 patients (0.8%), but only for perivalvular leak. There were no deaths, no complications, and no clinical evidence of malfunction of the prosthetic valve.(ABSTRACT TRUNCATED AT 250 WORDS)     

Mitral annulus morphologic and functional analysis using real time tridimensional echocardiography in patients submitted to unsupported mitral valve repair

Introduction

Mitral valve repair is the treatment of choice to correct mitral insufficiency, although the literature related to mitral valve annulus behavior after mitral repair without use of prosthetic rings is scarce.

Objective

To analyze mitral annulus morphology and function using real time tridimensional echocardiography in individuals submitted to mitral valve repair with Double Teflon technique.

Methods

Fourteen patients with mitral valve insufficiency secondary to mixomatous degeneration that were submitted to mitral valve repair with the Double Teflon technique were included. Thirteen patients were in FC III/IV. Patients were evaluated in preoperative period, immediate postoperative period, 6 months and 1 year after mitral repair. Statistical analysis was made by repeated measures ANOVA test and was considered statistically significant P<0.05.

Results

There were no deaths, reoperation due to valve dysfunction, thromboembolism or endocarditis during the study. Posterior mitral annulus demonstrated a significant reduction in immediate postoperative period (P<0.001), remaining stable during the study, and presents a mean of reduction of 25.8% comparing with preoperative period. There was a significant reduction in anteroposterior and mediolateral diameters in the immediate postoperative period (P<0.001), although there was a significant increase in mediolateral diameter between immediate postoperative period and 1 year. There was no difference in mitral internal area variation over the cardiac cycle during the study.

Conclusion

Segmentar annuloplasty reduced the posterior component of mitral annulus, which remained stable in a 1-year-period. The variation in mitral annulus area during cardiac cycle remained stable during the study.     

The case for early surgical treatment of left-sided primary infective endocarditis. A collective review.

Sixteen cases of infective endocarditis requiring valve replacement with a prosthesis are described. The indication for surgery was intractable congestive heart failure in all. Fourteen patients underwent aortic valve replacement, with five hospital deaths secondary to cardiac decompensation and one late death related to recurrent periprosthetic insufficiency. Two patients who required mitral valve replacement did well, with no early or late deaths. A collected review of 293 cases of left-sided primary infective endocarditis reported in the English literature confirms the feasibility, advisability, and life-saving potential of urgent surgical intervention in patients with persistent or progressive congestive heart failure, irrespective of the activity of the infective process or the duration of antibiotic therapy. All infected tissue should be resected, and Teflon bolsters should be used to minimize the incidence of periprosthetic leaks. The incidence of prosthetic infection is minimal, contrasted with the fatal potential of procrastination to achieve microbial cure.     

A case of thrombosed St. Jude Medical valve 16 years after initial mitral valve replacement]

We report successful surgery for a thrombosed St. Jude Medical (SJM) valve 16 years after the initial mitral valve replacement even under conditions of satisfactory anticoagulation therapy. A 61-year-old-female had intermittent claudication and was admitted to our hospital for examination. The prosthetic valve sounds were normal to auscultation and the left ankle-pressure index was decreased to 0.6. Transthoracic echocardiography revealed no mitral regurgitation and a mean mitral valve gradient of 6-7 mmHg. Furthermore, transesophageal echocardiography revealed that one of the leaflets of the prosthetic valve was entirely immobilized at the closing position and a mobile soft tissue mass, 5 mm in diameter, was detected at the atrial side of the obstructed leaflet. Although 96,0000 IU of urokinase was administered intravenously for a week, we could not confirm any change in leaflet mobility. At the time of surgery, the posterior leaflet of the SJM valve, which was implanted at an anatomical orientation, was obstructed at the closing position with old and fresh thrombi. We decided upon replacement with a CarboMedics 29 M prosthetic valve. Postoperative medication consisted of warfarin plus low-dose aspirin. Generally, valve thrombosis occurs within 5 years after valve replacement. However, valve thrombosis is possible even in a reliable SJM valve and as long as 16 years after replacement. Therefore, the implantation of an SJM valve at an anti-anatomical orientation might lower the incidence of valve thrombosis in addition to life-long anticoagulation therapy.     

The Clinical Life History of Explanted Prosthetic Heart Valves

This report analyzes 118 prosthetic heart valves obtained from 97 patients at reoperation (96) or at postmortem examination (22). The number obtained from the mitral, aortic, and tricuspid positions were 78, 32, and 8, respectively. Duration of implant ranged from one day to 12.3 years. Valves showing the least long-term wear were the Starr-Edwards metal strut-silicone ball and the Björk-Shiley. Moderate long-term durability was provided by the Beall and Starr-Edwards cloth-covered composite-seat prostheses while short-term durability was given by Hancock and Carpentier valves. Reoperation for valve-related causes was performed for 46 of 47 Beall valves, which demonstrated stenosis, hemolysis, and incompetence from component wear, 6 of 27 Björk-Shiley prostheses for valve thrombosis or thromboembolism or both, and 11 of 17 porcine prostheses because of calcification (4) or cusp perforation or rupture. Analyses of wear and fatigue of mechanical valves demonstrated that use of ultrahard materials (pyrolyte carbon, titanium, stellite 21) provided superior durability in contrast to polymeric solids or fabrics with poor abrasion and impact characteristics. Further, cloth and disc wear were evident as early as 0.5 year after implant and appeared to be complete by 4 years. Completeness of healing after 24 months was not related to the type of fabric material used or its construction.This study suggests that mechanical valves made from hard materials have long durability when properly implanted and require fastidious prophylaxis against infection and thromboembolism. The findings of early cusp perforation or rupture in the aortic position and leaflet calcification, stiffening, or disruption in the mitral position for porcine prostheses suggest that frequent and careful examinations of patients with these prostheses are required to detect early signs of stenosis or incompetence and that early reoperation is required before catastrophic valve failure necessitates emergency prosthetic valve replacement.     

Novel method of valve replacement with interrupted sutures using protractor rings for the partitioning of a prosthetic sewing ring.

Newly devised protractor rings that can equally partition a prosthetic valve sewing ring into 10-20 acrs have been used on 46 patients who underwent 59 aortic, mitral, and/or tricuspid valve replacement. After an appropriate number of interrupted mattress sutures were placed along the host annulus, one of the 11 protractor rings was chosen and used to copy the partition onto the sewing ring with a tissue pen. The interrupted mattress sutures kept on a suture holder were then passed through the prosthetic sewing ring directly from one end to the other. Based on ex vivo experiments in simulating valve replacement, it was evident that time required for placing sutures around the sewing ring was significantly shortened by the use of such protractor rings. The use of these protractor rings may thus be useful in decreasing valve replacement time as well as in potentially simplifying operative procedures.     

Tricuspid valve replacement with the St. Jude Medical valve

A study was conducted on 20 patients who underwent tricuspid valve replacement (TVR) with the St. Jude Medical (SJM) valve. Isolated TVR was performed on 9 patients, and additional mitral, or mitral and aortic valve replacements were performed on 11 patients. Four patients (20%) died in the early postoperative period, but there were no deaths related to the SJM valve in the tricuspid position. The mean follow-up period of the 16 survivors was 74.4 months, and there have been no deaths during the follow-up period. The postoperative actuarial survival rate was 80%, 10 years after surgery. Three patients, representing 0.25%/patient-months, developed valve thrombosis, the valve thrombosis-free rate being 72.8%, 10 years after surgery, while entrapment of a leaflet by endothelial pannus was found in one patient, representing 0.08%/patient-months. Thus, the incidence of all prosthetic valve-related complications was 0.34%/patient-months, and the postoperative complication-free rate was 65.3%, 10 years after surgery. The medium-term follow-up study of TVR with the SJM valve revealed no prosthetic valve-related deaths and a relatively low incidence of prosthetic valve-related complications. However, as with other mechanical valves, valve thrombosis was a major risk posed by the SJM valve in the tricuspid position.     

The Angell-Shiley porcine xenograft

A 4-year clinical experience with fresh allograft tissue valves prompted a trial of 0.5% buffered glutaraldehyde as a valve fixative and sterilant. Tanned allograft and porcine xenograft valves were inserted into experimental animals, and, beginning in 1970, similar valves were implanted in a series of patients now totaling 312. The clinical results are excellent. The 5-year valve-related mortality is 6% for patients who had mitral valve replacement and 16% for those with aortic valve replacement. To date, the incidence of thromboembolism is 1.3% per patient-year, and valve-related morbidity and mortality for the combined groups is 27.4%. Valve stent design has evolved from symmetrically configured metal to anatomically molded plastic. The maintenance of natural valve configuration has optimized leaflet coaptation and support, decreased tissue stress, and eliminated valve-stent dehiscence and tissue rupture seen in valves deformed to fit symmetrical stents. Stent design, controlled glutaraldehyde solutions, and fixation techniques have improved leaflet flexibility and reduced valve orifice to annulus diameter ratios, thus producing transvalvular gradients comparable to both mechanical and modified orifice tissue valves. To date, tissue failure, observed in only 1.0% (3 of 287) of patients, is the result of calcification (2 patients) and cusp rupture due to incomplete fixation (1 patient).     

Long-term results of valve replacement with the CarboMedics prosthetic heart valve

This study examined the long-term results of the CarboMedics prosthetic heart valve. Between July 1990 and May 2000, 256 CarboMedics valve prostheses (120 single aortic, 62 single mitral, and 34 double aortic/mitral) were implanted into 216 patients (mean age 57.8 years). The mean follow-up was 57.7 +/- 36 months (maximum 133 months) with a total of 1,038.4 patient years. Patient survival, including operative deaths, was 85.1% at 10 years. Linearized rates for various complications were thromboembolism 0.39% per patient year, bleeding events 0.39% per patient year; prosthetic valve endocarditis 0.29% per patient year, nonstructural dysfunction 0.87% per patient year, and reoperation 0.39% per patient year. There were no valve thrombosis or structural valve failures in this series. The probabilities of freedom from thromboembolism and reoperation at 10 years were 96.3% and 93.1%, respectively. In conclusion, the CarboMedics valve can be used satisfactorily with a low incidence of thromboembolism and valve thrombosis.     

Repeat heart valve surgery: risk factors for operative mortality.

BACKGROUND: Patients undergoing repeat heart valve operations are a diverse population. We assessed risk factors for operative mortality in patients undergoing a first heart valve reoperation. METHODS: A retrospective review of hospital records was performed for 671 patients who underwent first repeat heart valve operations between 1969 and 1998. Univariable and multivariable analyses were performed. RESULTS: Operative mortality was 8.6%. Mortality fell each decade to 4.8% in the most recent period (adjusted chi(2) for linear trend P <.0005). Mortality increased from 3.0% for reoperation for a failed repair or reoperation at a new valve site to 10.6% for prosthetic valve dysfunction or periprosthetic leak and to 29.4% for endocarditis or valve thrombosis. Concomitant coronary artery bypass grafting was associated with a mortality of 15.4% compared with 8.2% when it was not required. Mortality for aortic valve replacement was 6.4%, mitral valve replacement 7.4%, aortic and mitral valve replacement 11.5%, tricuspid valve replacement 25.6%, periprosthetic leak repair 9.1%, and isolated valve repair 2.2%. Among 336 patients requiring replacement of prosthetic valves, mortality was 26.1% for replacement of a mechanical valve compared with 8.6% for replacement of a tissue valve (P <.0005). Multivariable analyses identified year of reoperation, age, coronary artery bypass grafting, indication, and replacement of a mechanical valve rather than a tissue valve as significant explanatory variables for operative mortality. CONCLUSIONS: Heart valve reoperations can be performed with an acceptable operative mortality. However, we have identified several categories of patients in whom reoperation carries an increased risk.     

Prosthetic valve thrombosis: twenty-year experience at the Montreal Heart Institute

BACKGROUND: Prosthetic valve thrombosis is a life-threatening complication. We reviewed the incidence, risk factors, and treatment strategies of this rare complication. METHODS: From February 1981 through January 2001, 5430 valve operations were performed in 4924 patients at the Montreal Heart Institute. Of this cohort, 39 patients presented with prosthetic valve thrombosis and had complete follow-up data obtained from our prospective valve clinic database. RESULTS: In this series 82% of patients were women, and the mean age was 58 +/- 11 years. The underlying pathology involved the mitral valve in 75% of cases. Most prosthetic valve thromboses occurred with mechanical prostheses (95%). The time interval from first valve replacement to prosthetic valve thrombosis was 39 +/- 42 months. The most frequent clinical presentation was severe congestive heart failure (44%). On prosthetic valve thrombosis presentation, the international normalized ratio was less than 2.5 in 54%, with inadequate anticoagulation management in 26% and poor compliance in 26%. Eighty-two percent of patients underwent a surgical procedure, consisting of thrombectomy in 47%, mitral valve replacement in 47%, and aortic valve replacement in 6% of patients. The 30-day operative mortality and total in-hospital mortality after prosthetic valve thrombosis were 25% and 41%, respectively. The 10-year actuarial survival after prosthetic valve thrombosis was 46% +/- 10%. CONCLUSION: Inadequate level of anticoagulation is the most important factor involved in the pathogenesis of prosthetic valve thrombosis. The overall mortality rate despite surgical treatment remains high. This study underscores the importance of meticulous surveillance of anticoagulation therapy in patients with prosthetic valves.     

Prosthetic valve thrombosis—should thrombolysis be considered as first line of management?

Background

Traditionally prosthetic valve thrombosis has been managed with emergency surgery. However, there is growing evidence that thrombolysis is a suitable alternative. In this study, we aim to examine the safety and efficacy of thrombolysis in the management of prosthetic valve thrombosis (PVT).

Material and methods

This retrospective study was carried out at a single center between June 2010 and June 2014. All patients presenting with PVT were included in the study. All patients were treated with thrombolysis using our institutional protocol.

Results

Nineteen patients presented with 24 episodes of PVT. Overall, 73.6 % (14/19) of the patients were successfully treated with thrombolysis leading to clinical as well as haemodynamic improvement. Four (21 %) patients were taken directly for an operative intervention. The remaining 15 patients had a trial of thrombolysis which was successful in all cases. However, four of these patients had a recurrence within 2 weeks and were offered a second trial of thrombolysis. Three of these patients had a satisfactory outcome, and one patient was taken for surgery. Overall, 73.6 % (14/19) of the patients were successfully treated with thrombolysis. There was one incidence of intracerebral bleed leading to death. Mitral valve was the commonest valve involved in our study.

Conclusions

Thrombolysis is safe and may be considered as primary treatment strategy for first episode of prosthetic valve thrombosis. In recurrent thrombosis as well, a trial of thrombolysis can be given.

     

The CarboMedics Prosthetic Heart Valve in the Mitral Position: Results of the Multicenter International Trial

A bstract Isolated mitral valve replacement using the CarboMedics prosthetic mitral valve (CarboMedics, Inc., Austin, TX) was studied in 13 centers in the United States, Canada, and Scandinavia between 1987 and 1993 in 428 patients with a mean age of 57 ± 14 years. Actuarial survivals at 1, 2, and 5 years were 88.2% and 75.8%. Freedom from events at 5 years were 94.8% for major thromboembolism, 96.6% for thrombosis, and 96.7% for endocarditis. Linearized morbidity rates (events per 100 patient years) for events at > 30 days postoperatively were .51 thrombosis, 1.1 major thromboembolism, and .37 endocarditis. These results at early to mid-term compare favorably with the first generation bileaflet valve.     

Mechanical valve replacement in children and teenagers

A previous study from this unit showed that only 19% of children with mitral bioprostheses were free from complications after 7 years and prompted us to review the performance of new-generation mechanical prostheses implanted in the same population group. In a 5-year period (1980-1985), 352 patients 20 years old and younger (mean age 15.3 +/- 4 years) with rheumatic valvular disease had 177 mitral, 62 aortic and 113 double (mitral + aortic) valve replacements with Medtronic-Hall or St. Jude prostheses. The overall early mortality was 6.3%. All survivors, followed up for a total of 1171 patient-years, received oral anticoagulation. The late mortality for mitral, aortic and double valve replacement was 4.1% per patient-year, 4.3% per patient-year and 8.0% per patient-year, respectively (P less than 0.05), and was valve-related in 46% of the cases. Twenty-nine patients, all but 2 in the mitral and double valve replacement groups, were reoperated upon (2.5% per patient-year), mainly for infective endocarditis (34%), for prosthetic thrombosis (33%) and for bland periprosthetic leak (31%). The incidence of thrombotic obstruction was 1.1% per patient-year: mitral valve replacement, 1.0% per patient-year; aortic valve replacement 0.5% per patient-year; and double valve replacement, 1.7% per patient-year; P less than 0.05) and was fatal in 33% of the cases. Major systemic thromboembolism occurred at the rate of 1.4% per patient-year, similar in the three groups. The incidence of prosthetic endocarditis was 0.9% per patient-year.(ABSTRACT TRUNCATED AT 250 WORDS)