Anticoagulation is mandatory in all mechanical valve replacements and, when well-managed, permits both embolism and valve thrombosis to be kept at low levels. Here, we report the case of a male patient who has survived 30 years without anticoagulation following aortic valve replacement with the Björk-Shiley prosthesis.

A 30-year old man underwent a head-on motor accident and was admitted to hospital, where he was successfully treated for several injuries. He was discharged home with slight tricuspid regurgitation, without signs of hemodynamic decompensation, and required no further treatment. Five years later he was admitted to the National Institute of Cardiology due to overt right ventricular failure secondary to tricuspid valve regurgitation. After several unsuccessful attempts at papillary muscle and chordal repair, a 33-mm SJM Biocor valve was implanted. This case history underlines the value of echocardiography, even after blunt chest trauma without cardiac symptoms. A good functional result can be achieved years after onset of traumatic valve insufficiency, but early diagnosis and surgical treatment increases the feasibility of tricuspid valve repair rather than replacement, and also prevents progressive deterioration of right ventricular function.

A 77-year-old man had a large right atrial (RA) thrombus associated with a combined mitral and tricuspid valve disease. Echocardiography showed a large, immobile, non-homogeneous, irregularly surfaced mass in the dilated RA, and prolapse of the anterior mitral leaflet resulting in massive mitral regurgitation. Computed tomography (CT) revealed a laminated structure with calcification and distinct margins, without invasion to the wall of the RA. On the basis of these echocardiographic and CT findings, a diagnosis of combined mitral and tricuspid valvular disease complicated with RA thrombus was made. Removal of the RA thrombus, mitral valve replacement and tricuspid annuloplasty were performed simultaneously, with successful outcome.

Two unusual cases of congenital bicuspid aortic valve associated with aneurysm of the ascending aorta are reported. One patient with a 7-cm ascending aortic dilatation and aortic regurgitation (AR) (II/IV), and another with a 6-cm ascending aorta and AR (III/IV), presented for treatment. Replacement of the ascending aorta and aortic valve repair were performed in both cases. Aortic valve repair included resection of the raphe, leaflet plication and subcommissural annuloplasty. Both patients had satisfactory results in the early postoperative period. Despite the promising outcomes after surgery in these patients, long-term changes in valve function and durability remain unknown. Additional close observation and monitoring are required before the procedure can be recommended as the standard of care.

Background and aim of the study: The predictive value of chronic atrial fibrillation (AF) before percutaneous mitral balloon valvotomy (PMBV) is still under discussion. The effect of the duration of chronic AF on short- or long-term results is not known. Therefore, we analyzed the predictive value of pre-procedural chronic AF and the duration of this rhythm disturbance for short- and long-term outcome after PMBV in patients with mitral valve stenosis. Methods: A total of 140 PMBV procedures was performed in 137 patients with severe mitral stenosis. Sixty-three patients (45%) were in chronic AF; in 40 patients (63%) the AF was of more than one year duration. A successful procedure is defined as PMBV achieved without acute mitral valve replacement, and a mitral valve area after PMBV of > or =1.5 cm2. Results: Patients in chronic AF were significantly older, had a larger left atrial diameter and higher NYHA functional class, compared with patients in sinus rhythm (SR). The success rates of PMBV were 80.5% and 77.6% in patients with SR and AF, respectively (p = NS). Mean follow up was 4.2+/-2.6 years (n = 127). At four years' follow up the event-free survival was 86.5% in patients with SR, and 78.5% in those with chronic AF at baseline (p = 0.031). Multivariate analysis of the entire study population showed the presence of chronic AF to be the only pre-procedural independent predictor for severe mitral regurgitation after PMBV (p = 0.030), as well for an event (p = 0.039) and restenosis (p = 0.034) during follow up. The risk for an event or restenosis during follow up increased seven-fold when chronic AF at baseline was present for more than one year (p = 0.010). Conclusion: Pre-procedural chronic AF is an independent predictor for unfavorable outcome at short- and long-term follow up after PMBV. A longer duration of AF further increases the risk of an event or restenosis.

Background and aim of the study: Ostium primum atrial septal defect (ASD I) is very rarely observed in the adults. Although ASD I was first corrected surgically almost 50 years ago, the efficacy of surgical treatment in adults has not been well documented. Thus, the long-term outcome of patients aged over 20 years and having surgical repair of ASD I was examined. Methods: Fifteen patients (10 males, five females; mean age 31.4+/-13.1 years; range: 20 to 56 years) who had surgical repair of ASD I between 1982 and 2000 were followed. All patients were examined physically and underwent chest radiography, ECG and echocardiography (cross-sectional and Doppler) before and after surgery. Autologous pericardium was used to close the defect in 14 patients, and a direct suture was used in one patient. Mitral valvuloplasty (repair of the cleft) was performed in 12 patients, and De Vega annuloplasty in eight. Results: There were no hospital deaths. In one case a pacemaker was implanted five days after surgery because of complete heart block. Preoperatively, nine patients (60%) were in NYHA classes III and IV; at the end of follow up, 14 (93.3%) were in classes I and II. Preoperatively, 13 patients had sinus rhythm, and one each had atrial fibrillation (AF) and rhythm from the atrioventricular sinus. During follow up, three patients developed AF which was treated successfully with electrical cardioversion. The preoperative mean cardiac volume index of 695 +/- 216 ml/m(2) was reduced significantly after repair to 523 +/- 108 ml/m(2) (p < 0.05). Before surgery, mitral regurgitation was observed (severe in five cases, moderate in seven, mild in three). Postoperatively, a residual intracardiac shunt was identified in one case. Postoperative mitral regurgitation was noted in six patients (moderate in two, mild in four). The right ventricular dimension was decreased significantly, from 5.0 +/- 1.5 mm before surgery to 3.2 +/- 0.6 mm after repair (p < 0.001). Conclusion: Patients aged over 20 years with ASD I benefit from surgical repair of the defect.

Background and aim of the study: Residual or recurrent tricuspid regurgitation (TR) has been reported after several types of surgical repair. The development of late TR is an important complication of left heart surgery. The results of De Vega annuloplasty were compared with those obtained after Carpentier-Edwards ring (CE ring) annuloplasty in patients with secondary TR.  Methods: The records of 45 patients who underwent surgery for secondary TR between January 1995 and July 2000 were reviewed retrospectively. Twenty-eight patients underwent De Vega annuloplasty, and 17 had a CE ring annuloplasty. The groups were similar with respect to associated cardiac lesions. No significant preoperative differences were observed in NYHA functional class, TR grade, and pulmonary artery pressure between the two groups. Results: One CE patient died of left ventricular dysfunction after postoperative bleeding. The 28 De Vega patients and remaining 16 CE patients had an uneventful recovery, and were discharged. Tricuspid ring size after repair was similar between groups. Mean (+/- SD) follow up in the entire patient cohort was 39+/-23 months (range: 6 to 75 months). TR recurrence was rated as grade II or III in 13 patients (45%) after De Vega annuloplasty, but was grade II or III in only one patient (6%) patient after CE ring annuloplasty. There was a significant difference in TR recurrence between the groups (p = 0.027), but no significant difference in NYHA class. Conclusion: CE ring annuloplasty significantly decreased the recurrence of TR; thus, CE ring annuloplasty is superior to De Vega annuloplasty in patients with secondary TR.

Background and aim of the study: Glutaraldehyde may promote calcification in xenograft tissue by the action of toxic aldehyde group residues involved in the cross-link process. Post-fixation treatment with homocysteic acid (HA) neutralizes this toxicity by bonding aldehyde groups, and enhances biocompatibility on the basis of strongly electronegative sulfonic groups. Previous studies in a rat subcutaneous model showed significant long-term mitigation of mineralization of glutaraldehyde-fixed pericardium treated with HA. This study aimed to assess the anticalcific efficacy of HA in a valvular implant in growing sheep, and establish if the tricuspid position is suitable for testing replacement bioprosthetic valves. Methods: Eleven stented 25 mm Pericarbon bioprostheses (seven HA-treated, four standard) were implanted in the tricuspid position of growing sheep. Infective endocarditis occurred in four prostheses. Among the remaining seven, three (two HA-treated, one standard) were explanted at 91 days (mid-term), and four (two HA-treated, two standard) at 140-141 days (long-term). All explants were studied by gross, X-ray, light, transmission and scanning electron microscopy, as well as by atomic absorption spectroscopy. Results: No histological and ultrastructural difference in tissue preservation were observed between HA-treated and standard Pericarbon bioprostheses, either in the mid or long term. The mean calcium content of mid-term HA-treated explants was 9.55 mg/g compared with 16.26 mg/g in mid-term standard explants. Only one late standard explant failed as a result of severe stenosis caused by massive dystrophic calcification. Among four late explants, two showed significant increase in mineralization (HA-treated, 87.45 mg/g; standard, 181.20 mg/g), while two showed calcium contents similar to those in mid-term explants (HA-treated, 11.96 mg/g; standard, 17.32 mg/g). Conclusion: Post-fixation treatment with HA preserves structural properties after tricuspid implantation in growing sheep. The tricuspid implant in the sheep model failed to reproduce remarkable accelerated progressive calcification in all xenografts so as to demonstrate a significant difference between HA and standard explants. The tricuspid position for testing replacement bioprosthetic valves should be abandoned, and investigations repeated with the prosthesis in the mitral position.

Background and aim of the study: Orthotopic valve replacement in large animals is an important component of the preclinical assessment of bioprosthetic valves. To provide the most useful preclinical information, the development of models that parallel clinical practice patterns is essential. Therefore, we sought to develop a technically feasible and reproducible model for chronic evaluation of stentless bioprosthetic aortic valves implanted as aortic root replacements in juvenile sheep. Methods: Juvenile domestic sheep (mean age 21+/-2.28 weeks; range: 17-26 weeks) underwent aortic root replacement using standard cardiopulmonary bypass (CPB) and surgical techniques. Animals were implanted with 19 mm (n = 21), 21 mm (n = 18) or 23 mm (n = 4) bioprostheses from two different manufacturers, and followed for 150 days. Animals surviving at least 150 days were considered long-term survivors; those which died prior to postoperative day (POD) 31 were considered operative deaths. Results: Forty-three animals underwent aortic root replacement. The mean CPB time was 91+/-20 min (range: 62-149 min); mean cross-clamp time was 63+/-13 min (range: 39-95 min). Thirty-five animals (81%) survived the first 30 days of the study period. Five deaths occurred at POD 0 due to anastomotic complications. One death occurred each on POD 3, 6, and 26 as a result of prosthesis size mismatching, thromboembolic complications, and endocarditis, respectively. There were five late deaths. Twenty animals survived the minimum 150-day study period, and 12 were sacrificed at 183+/-17 days. Six animals remain alive at 151+/-0.98 days, and one animal died each on POD 184 and 190. The remaining 10 animals are not yet 150 days from their operation. Currently, all are well at 102+/-34 days (range: 33-140) days. Conclusion: These data suggest that long-term evaluation of stentless aortic bioprostheses implanted as aortic root replacements can be accomplished using juvenile sheep.

Background and aim of the study: The use of flexible structures within cardiovascular prostheses such as valves, stents and vascular grafts has been proposed as a means of more closely modeling native mechanics, and thereby reducing the biomechanical problems associated with rigid materials. However, the design of such materials has been hampered by the paucity of quantitative information on the in-vivo behavior of such structures. The aim of this study was to explore the use of 3D ultrasound imaging coupled with finite element analysis (FEA) as a tool to estimate deformation and forces imposed in vivo on a novel bioprosthetic valve design. Methods: The method was first tested using in-vitro static loading conditions, where good agreement between displacements seen on video and those obtained from application of the identical force within the finite element program was seen. The method was then tested in a porcine model with valves implanted in the mitral position. Images of the deforming annular ring were obtained over the cardiac cycle using 3D intravascular ultrasound; these images were fed into the FEA program for calculation of reaction forces. Results: Results in vitro showed that a force of 2.7-8.0 Newtons (N) was required to produce a deformation of between 1.0 and 3.0 mm in the radial direction. A time history of deformation and force around the ring of the valve stent could be obtained for the in-vivo conditions. These results revealed a maximum deformation of 0.5-1.7 mm along the short axis (anteroposterior) of the mitral valve. Coupled to this, a peak reaction force of 4.4-13.9 N was found at the points corresponding to maximal deflection. Both deformation and reaction force reached maximum during atrial contraction. Conclusion: This method provides an accurate means of estimating deformation and corresponding forces imposed in vivo on intracardiac prostheses. The results provide information on the dynamic behavior of the mitral valve annulus. Such information should be useful in the design of flexible cardiovascular prostheses.

Background and aim of the study: Photo-oxidation treatment of porcine stentless bioprostheses (Photofix) was compared with glutaraldehyde fixation, with either AOA (Freestyle valve) or Tween-80 (Edwards Prima Plus valve). Methods: Six valves of each type were implanted in juvenile sheep, in the pulmonary position. Valves were explanted after three or six months and examined macroscopically, by X-radiography, and by light and transmission electron microscopy. Calcium content was determined by atomic absorption spectrometry. Results: The cusps of all valves were free of calcification, and had normal histology and function. Calcium contents (median +/- IQR) were 0.63+/-0.45, 0.73+/-1.46 and 0.46+/-1.42 microg/mg for the Photofix, Freestyle and Prima Plus valves, respectively (p = NS). Calcium contents of the aortic wall portions were 0.71+/-1.27 (Photofix), 10.78+/-77.22 (Freestyle) and 28.70+/-66.53 (Prima Plus) (p <0.05 for Photofix versus Freestyle or Prima Plus). Conclusion: Photo-oxidation of a porcine stentless valve prevents calcification not only in the cusps, but also in the aortic wall portion.

Cardiac involvement in childhood brucellosis is rare and, when present, findings are usually seen in acute rheumatic fever with endocarditis. We report a successfully treated case of Brucella endocarditis in which the aortic valve was affected. Medical therapy did not cure this patient who, due to hemodynamic deterioration, required valve replacement. This case report suggests that successful management of Brucella endocarditis requires a combination of medical and early surgical therapy.

Background and aim of the study: Most previous studies on surgery for endocarditis included cases of both prosthetic and native valve endocarditis (NVE), which clearly differ in their course and prognosis.

Methods: Between February 1979 and July 2000, 199 patients underwent surgical treatment for NVE. The native valve was replaced in 185 patients, and repaired in 14. Correlations between previous heart disease and site of NVE, site and etiology, site and embolism, and etiology and extracardiac infection were analyzed. Viability of echocardiography to identify NVE features was compared with surgical inspection. Risk factors for hospital, late mortality and recurrence were assessed by multivariate analysis, including preoperative, operative and postoperative variables. Results: Streptococci were found in 52 cases (26.1%), mostly spreading from splanchnic (n = 10) and cutaneous (n = 9) infections; Staphylococci were found in 43 cases (21.6%), mostly from cutaneous foci (n = 20). Tricuspid valve endocarditis was due to Staphylococcus sp. in all documented cases (11/15; p <0.001), while Streptococcus sp. predominated in the mitral (p = 0.026) position. Echocardiography was reliable in identifying anatomic features of NVE, except for periannular abscess (13 versus 31 found at surgery; p <0.001). Hospital mortality was 6% (n = 12): preoperative NYHA class IV (p = 0.034) and emergency surgery (p = 0.001) were significant predictors of mortality. Average follow up was 57.2+/-4.4 months and was 96.3% complete. There were 27 late deaths (15%). Actuarial survival was 75.7% at 10 years and 60.3% at 20 years. Low social status (p = 0.022), diabetes mellitus (p = 0.003), fever at time of surgery (p = 0.041) and persistent postoperative fever (p = 0.0002) were predictors for late mortality. Endocarditis recurred in 28 patients (15.5%), with 74.1% freedom at 10 years and 68.7% at 20 years. Drug abuse (p = 0.001), mitral endocarditis (p = 0.002), positive blood cultures (p = 0.0003), homograft implantation (p = 0.0012) and persistent postoperative fever (p <0.0001) were significant predictors for recurrence. Conclusion: Preoperative conditions of patients with NVE significantly affect early and long-term outcomes. Prognosis is worsened by delayed diagnosis and operation during the active septic process.

Background and aim of the study: Porcine bioprosthetic valves have excellent hemodynamics and do not require anticoagulation, but have limited durability. Cusp tearing is a major cause of bioprosthetic valve failure. It has been suggested that the mechanism of bioprosthetic valve failure is stiffening by calcification, which leads to elevated stresses and secondary collagen fiber damage and leaflet tearing. This thesis was tested in explanted porcine bioprostheses. Methods: A total of 60 explanted porcine bioprosthetic valves was tested mechanically, and 15 explanted valves were examined grossly and histologically. Circumferentially and radially oriented samples of cusp tissue were tested uniaxially in a materials testing machine and compared with five controls. Results: Mean (+/-SD) duration of implantation was 10.9+/-5.6 years. Circumferential specimens from explants were less extensible than controls (11.0+/-5.5% versus 24.5+/-2.8% strain, p <0.001), and failed at lower tensions (973+/-733 versus 3075+/-911 N/m, p = 0.001) and at lower strains (21.2+/-8.1% versus 47.3+/-7.1% strain, p <0.001). Radial specimens from explants were less extensible (28.7+/-6.8% versus 39.2+/-5.9% strain, p = 0.002) and failed at lower strains (60.3+/-17.3% versus 112.2+/-24.9% strain, p <0.001) than the controls. The stiffness of the explants was unchanged from controls in both circumferential and radial samples. There were no differences between explants and controls in radial and circumferential stiffness, and in radial failure strength. Calcification was mild and diffuse in most of the tested samples. Tears were found in areas without calcific deposits, along with breaks in collagen fiber bundles. Conclusion: These results do not support the thesis that calcification stiffens glutaraldehyde-fixed porcine bioprostheses, except when the entire cusp is transformed into a solid mass of mineral. Rather, leaflet tears may develop as a result of accumulated mechanical damage that is independent of calcification.

Background and aim of the study: With increasing use of homograft and autograft aortic valves for aortic valve replacement (AVR), more patients will be presenting for aortic valve reoperation due to structural degeneration of the homograft or autograft valve. Management options include homograft re-replacement, which may require extensive surgery, versus AVR with a mechanical valve or a stented xenograft. Here, results are reported in 18 consecutive patients who underwent aortic valve re-replacement (AVreR) after previous homograft or autograft insertion. Methods: Between May 1976 and March 2001, 18 patients underwent AVR after previous homograft (n = 16) or autograft (n = 2) insertion. The homograft or autograft had been implanted as a full root in eight patients (44%), as a mini-root in one (6%), and in the subcoronary position in nine (50%). Indication for the reoperation was structural valve degeneration (n = 14; 72%) in one occasion combined with aneurysm of the homograft, or endocarditis (n = 4; 22%), and seven (39%) presented as a non-elective procedure. The median interval between the two operations was 5.4 years (range: 0.3-10.8 years). Results: Fourteen patients (78%) received either a mechanical valve (n = 12; 67%) or a stented xenograft valve (n = 2; 11%). Four others (22%) required root re-replacement with either another homograft (n = 3) or a mechanical valved conduit (n = 1) for endocarditis (n = 2) or an associated aneurysm (n = 2). Overall hospital mortality was 11% (n = 2) due to stroke (n = 1) or respiratory failure (n = 1). Two patients died 3.1 and 7.0 years after the procedure. Conclusion: Aortic valve reoperation after previous homograft or autograft implantation is a rare operation and presents a high-risk group. A simplified approach was preferred by utilizing mechanical or stented xenograft valves at reoperation, while homograft re-replacement was reserved for endocarditis or an associated aneurysm.

Background and aim of the study: Few long-term data are available on the Carpentier-Edwards Standard bioprosthesis in the mitral position. As for other bioprostheses, patient age at the time of implant is the main risk factor for structural deterioration, but no published report has analyzed the life-span of these bioprostheses with respect to this parameter. Methods: A series of 139 patients who underwent mitral valve replacement with the Carpentier-Edwards Standard bioprosthesis between 1978 and 1987 was reviewed. Mean age at implant was 59.6+/-14.7 years (range: 17-79 years). Follow up was 98.4% complete; total follow up was 1,078.7 patient-years (pt-yr) (mean 8.4+/-4.1 years). Mean follow up in the subgroup of patients alive at the time of the survey was 10.4+/-3.4 years. Results: Structural valve deterioration (SVD) occurred in 30 patients, with mean time to onset of deterioration 9.0+/-2.7 years (median 8.7 years). This time was independent of age at the time of implantation. Analysis by age group (< or =35, 36-50, 51-60, 61-65, 66-70, >70 years) showed deterioration to be more frequent in younger subjects (linear rates 7.9, 6.0, 3.3, 2.4, 0.6 and 0.4% pt-yr, respectively). Over the age of 65 years, the risk of SVD no longer varied with age, and was a rare complication. Conclusion: The mean time to onset of SVD was independent of patient age at the time of implant. After 65 years, the risk of SVD was low, without any significant variation. The Carpentier-Edwards Standard bioprosthesis may be used in the mitral position in subjects aged over 65 years, and with a low risk of deterioration.

Background and aim of the study: The St. Jude Medical (SJM) Regent heart valve is a new bileaflet prosthetic valve modified from the currently marketed SJM mechanical valve, with a modified external profile that results in a larger geometric orifice area without changing the existing design of the pivot mechanism or blood contact surface areas. The aim of the present study was to report the early hemodynamic and clinical results of an on-going multicenter trial investigating the clinical performance of the Regent mechanical aortic valve prosthesis. Methods: The early results from 204 patients at 11 centers in North America and Europe who underwent implantation of a Regent mechanical aortic valve prosthesis are described. Clinical status was prospectively recorded, and echocardiography with Doppler performed at hospital discharge, and at two and six months and one year postoperatively. Results: Follow up to date is 109.3 patient-years (average follow up 0.5+/-0.4 years per patient; range: 0 to 1.7 years). NYHA class improved for the group, and there were low rates of clinical adverse effects. Echocardiographic mean pressure gradient at six months was 13.8+/-10.3, 7.4+/-4.1, 5.4+/-3.2, 5.2+/-2.8, 3.4+/-2.3 and 3.6 mmHg, respectively, for 19, 21, 23, 25, 27 and 29 mm valves; effective orifice area was 1.5+/-0.6, 2.0+/-0.7, 2.4+/-1.0, 2.5+/-0.7, 3.6+/-1.4 and 4.8 cm2, respectively. There was a statistically significant decrease in left ventricular mass index between early postoperative (169.1+/-57.7 g/m2) and six months follow up (137.2+/-42.7 g/m2, delta = -30.1+/-42.5 g/m2, p <0.0001). Conclusion: The SJM Regent aortic valve has excellent associated hemodynamics with rapid and significant left ventricular mass regression. In all cases of adverse events, rates fell within Objective Performance Criteria guidelines. Long-term clinical assessment is on-going.

Background and aim of the study: Brisk hemolysis due to perivalvular leak is usually an indication for valve re-replacement. Repeated surgery after multiple previous valve operations is associated with high mortality, morbidity and failure rates. The present study evaluated the role of erythropoietin (EPO) administration in deferring or obviating the need for repeated surgery. Methods: Three patients (two men, one woman; age range 62-76 years) with two mechanical valves each and two to four previous heart valve operations, who suffered from severe mechanical hemolytic anemia, were given subcutaneous EPO for 15-17 months. Results: A marked reduction in red blood cell consumption was achieved with a weekly EPO dose of 18,000 U in two patients, both of whom also had mild or moderate kidney malfunction. A third patient with normal renal function and extreme hemolysis showed a transient, partial response to 30,000 U of EPO per week, and eventually needed a fifth operation. Conclusion: EPO may defer or even obviate the need for repeated valve surgery in patients with severe hemolysis due to perivalvular leak, especially those with inadequate EPO response, such as those with renal malfunction.

Background and aim of the study: Prosthetic mitral valve replacement (MVR) is associated with paravalvular leak in up to 12.5% of patients. The influence of the morphology and location of paravalvular leaks on clinical symptoms and degree of leak-related hemolysis is unknown. Methods: Morphology, size, location and number of paravalvular leaks were analyzed in 96 consecutive patients with primary mitral paravalvular leaks. Results: Mitral leak was diagnosed a median of 119 days after primary MVR. A small (1-2 mm) paravalvular leak was found in 41 patients (43%), an intermediate leak (3-5 mm) in 26 (27%), and a large leak (6-15 mm) in 29 (30%). Single leaks were observed in 70 patients (73%), whilst 26 (27%) had multiple leaks. Paravalvular leaks occurred around the entire prosthetic circumference, but were seen predominantly around the mitral commissural areas (76%). The larger the size of the leak, the more symptomatic the patient (p = 0.006); 80% of patients with small leaks were in NYHA classes I and II, whilst 62% with intermediate/large leaks were in NYHA classes III and IV. The number of leaks was not correlated with severity of clinical symptoms. Multiple leaks were more likely to cause significant hemolysis. Patients with preoperative chronic renal insufficiency, postoperative infection or large (>5 cm) left atria were more likely to develop multiple leaks. The size and location of the leaks was surgeon-dependent. Conclusion: Intraoperative transesophageal echocardiography is mandatory to detect possible small leaks and technical errors. Strict monitoring of all MVR patients is necessary for prolonged periods, as the appearance of paravalvular leaks is not necessarily correlated with clinical symptoms. Small paravalvular leaks, in particular, may go unnoticed. As the location and size of the leaks were significantly surgeon-dependent, self-monitoring should be mandatory for all surgeons.

Background and aim of the study: Prostheses used to treat heart valve disease improve patient survival, but have certain disadvantages. Paravalvular leakage (PVL) is a rare complication after mitral valve replacement (MVR), and can impair cardiac function and reduce the patient's functional capacity, depending on the degree of periprosthetic regurgitation.

Methods: Between 1985 and July 1999, 2,502 patients underwent MVR with or without concomitant cardiac procedures. Of these patients, 33 (18 males, 15 females; mean age 39.8+/-15.3 years; range: 12-62 years) had PVL of differing degree. The interval between MVR and observation of PVL was 30.5+/-31.5 months (range: 1-126 months), and the period after diagnosis was 22.6+/-31.5 months (range: 2-114 months). Fourteen patients (42.4%) underwent reoperation (RO group), and 19 (57.6%) were followed medically (ME group). Indications for reoperation were reduction of functional capacity, echocardiographically proven serious mitral regurgitation, and hemolysis. Results: Reoperative mortality was 3.0% (1/33), and late mortality 3.1% (1/32) for all patients. Cumulative survival after PVL was 90.2+/-6.7% at both five and ten years. Annular calcification (33.0%) and infective endocarditis (18.2%) were important predictive factors for development of PVL. Only one patient required second re-do surgery. Univariate and forward stepwise logistic regression analyses showed that there was no predictor for the development of severe PVL requiring a second reoperation. No difference was observed between left ventricular dimensions before and after periprosthetic regurgitation. The only significant finding between groups was an increase in left atrial diameter in RO patients after the development of PVL (p <0.05). Conclusion: Among patients undergoing MVR there are no clinical features to distinguish who will develop severe PVL during follow up. If PVL reduces the patient's functional capacity or causes serious hemolysis, or if severe PVL is evaluated echocardiographically, then reoperation must be performed. Mild or moderate mitral regurgitation without impairment of functional capacity may be followed medically. In asymptomatic patients, enlargement (>5%) of the left atrial diameter following development of moderate PVL may be a valuable criterion for deciding when to reoperate.