The implantation of bioprostheses is the preferred. Influence of Prosthesis Patient Mismatch on Diastolic Heart Failure After Aortic Valve Replacement

Influence of Prosthesis Patient Mismatch on Diastolic Heart Failure After Aortic Valve Replacement Shahab Nozohoor, MD, Johan Nilsson, MD, PhD, Carsten Lührs, MD, Anders Roijer, MD, PhD, and Johan Sjögren, MD, PhD Departments of Cardiothoracic Surgery and Cardiology, Lund University Hospital, Lund, Sweden Background. Bioprostheses for supraannular placement have been developed to optimize the hemodynamic performance after aortic valve replacement. To evaluate the potential benefit of this design, we analyzed the influence of prosthesis patient mismatch on diastolic function and left ventricular mass regression and evaluated the clinical performance of the Sorin Soprano and Medtronic Mosaic in the aortic position. Methods. A total of 372 patients underwent aortic valve replacement between July 2004 and February 2007, receiving either a Sorin Soprano (n 235) or a Medtronic Mosaic (n 137) prosthetic valve. Echocardiographic and clinical data were collected prospectively, and follow-up was performed in April 2007. Multivariate analyses were used to identify differences in hemodynamic performance, diastolic function, left ventricular mass regression, and predictors of impaired survival. Kaplan Meier survival curves and log-rank tests were used to compare postoperative outcomes. Results. The 30-day mortality was 1.7% (4 of 235 patients) in the Sorin Soprano group and 2.9% (4 of 137 patients) in the Medtronic Mosaic group (p 0.473). Neither prosthesis patient mismatch nor type of prosthesis was a significant predictor of early or late mortality. Diastolic heart failure was a predictor of poor survival (p 0.004); however, the recovery of diastolic function was not significantly influenced by prosthesis patient mismatch. Neither moderate (indexed effective orifice area < 0.85 cm 2 /m 2 ) nor severe (indexed effective orifice area < 0.65 cm 2 /m 2 ) prosthesis patient mismatch resulted in a significantly impaired left ventricular mass regression. Conclusions. Prosthesis-patient mismatch was not an independent predictor of poor survival, impaired left ventricular mass regression, or recovery of diastolic function. The Sorin Soprano and the Medtronic Mosaic bioprostheses demonstrated comparable hemodynamic performance and excellent clinical outcome without signs of structural valve deterioration during follow-up. (Ann Thorac Surg 2008;85:1310 8) 2008 by The Society of Thoracic Surgeons Accepted for publication Dec 27, 2007. Address correspondence to Dr Sjögren, Department of Cardiothoracic Surgery, Heart and Lung Center, Lund University Hospital, Lund, SE-221 85, Sweden; e-mail: johan.sjogren@med.lu.se. The implantation of bioprostheses is the preferred choice for aortic valve replacement (AVR) in the elderly as a result of the shorter life expectancy and reduced need for anticoagulation therapy. However, despite progress in the design and construction of these valves, the hemodynamic performance is not yet comparable to that of the native aortic valve. Prosthesis patient mismatch (PPM) has been suggested to occur when the effective orifice area (EOA) of the inserted prosthesis is smaller than the EOA of the patient s native valve [1]. Previous studies have shown that short-term and longterm survival may be influenced by PPM [2, 3]. Furthermore, it has been reported that PPM may lead to impaired recovery of left ventricular (LV) systolic function and poor LV mass regression [4]. Impaired LV mass regression may be detrimental inasmuch as persistent LV hypertrophy is one of the known causes of diastolic heart failure (DHF) [5]. Diastolic heart failure has been demonstrated in 50% of patients in the elderly population, thereby causing symptomatic congestive heart failure despite a normal LV ejection fraction [6]. The pathophysiologic condition of heart failure in aortic stenosis derives from increased pressure overload on the LV resulting in diastolic dysfunction, in addition to systolic dysfunction, or even before systolic function deteriorates. Dineen and colleagues [7] demonstrated that the LV ejection fraction was preserved in 60% of patients with aortic stenosis and DHF. Studies on DHF after AVR and the relation between PPM and diastolic dysfunction are scarce [8, 9]. Prosthesis patient mismatch may be avoided by aortic root enlargement or the implantation of stentless prostheses or homografts. However, these procedures increase the surgical complexity [10], and their outcome is dependent on the individual surgeon s experience. Third-generation bioprostheses designed for complete supraannular implantation offer another alternative, as the stent is positioned with potentially less disturbance of aortic blood flow [11]. The Sorin Soprano is a thirdgeneration bovine pericardial bioprosthesis designed for complete supraannular implantation and has been avail- 2008 by The Society of Thoracic Surgeons 0003-4975/08/$34.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2007.12.069

Ann Thorac Surg NOZOHOOR ET AL 2008;85:1310 8 PROSTHESIS PATIENT MISMATCH AND DIASTOLIC FUNCTION 1311 Table 1. Preoperative Characteristics a Characteristic Sorin Soprano (n 235) Medtronic Mosaic (n 137) p Value Sex (female) 107 (46%) 53 (39%) 0.198 NYHA III 103 (44%) 70 (51%) 0.175 NYHA IV 21 (9%) 17 (12%) 0.286 Diabetes mellitus 32 (14%) 23 (17%) 0.406 COPD 31 (13%) 21 (15%) 0.566 PHT 60 mm Hg 16 (7%) 5 (4%) 0.203 PVD 32 (14%) 22 (16%) 0.519 Age at implant (y) 76.4 6.3 76.7 5.8 0.712 BSA (m 2 ) 1.85 0.2 1.86 0.2 0.608 BMI (kg/m 2 ) 26 4 26 4 0.712 EuroSCORE 8.5 3 8.0 2 0.143 Preoperative creatinine ( mol/l) 97 59 95 38 0.796 a Values are number and percentage, or mean standard deviation. BMI body mass index; BSA body surface area; COPD chronic obstructive pulmonary disease; NYHA New York Heart Association functional class; PHT pulmonary hypertension (systolic pressure 60 mm Hg); PVD peripheral vascular disease. able for commercial use since 2003. However, data on clinical outcome and hemodynamic performance are limited [11 13]. The aim of this study was to evaluate the influence of PPM in modern bioprostheses, with respect to the recovery of LV diastolic function and LV mass regression, and to present the mid-term outcome for the Sorin Soprano bioprosthesis after AVR. Patients and Methods Patient Population Between July 2004 and February 2007, 372 patients underwent AVR with the Sorin Soprano bovine pericardial bioprosthesis (Sorin Biomedica Cardio SpA, Saluggia, Italy; n 235) and the Medtronic Mosaic porcine bioprosthesis (Medtronic Inc, MN; n 137). Patients undergoing concomitant coronary artery bypass grafting were included in the study. Exclusion criteria were double valve replacement, aortic dissection, surgery on the aortic arch, and procedures including deep hypothermia with circulatory arrest. Valve disease was mainly classified as degenerative calcification (n 341), active endocarditis (n 8), or repeat AVR (n 5), and in the remaining cases the disease was mixed (n 18). Preoperative, perioperative, and postoperative variables were prospectively collected and entered into the department s computerized cardiac surgical database. In addition, medical records were retrospectively reviewed, or contact was made with the patient or patient s physician when necessary. The preoperative characteristics of the study patients included in the study are summarized in Table 1. The study protocol was approved by the Ethics Committee for Clinical Research at Lund University, Sweden. Surgical Procedure Aortic valve replacement was performed using standard cardiopulmonary bypass with mild hypothermia and intermittent cold blood cardioplegic cardiac arrest. The perioperative data are presented in Table 2. The choice between pericardial or porcine bioprostheses was according to surgeons preference. The International Organization for Standardization (ISO) specification concerning labeling of valve sizes (ISO/CD 5840) recommends that labeled prosthesis size should represent the tissue annulus diameter of the patient into whom the valve is to be implanted. At present, the only manufacturer following this recommendation is Sorin Biomedica Cardio. Because of the differences in manufacturers labeling of the two prostheses studied, a thorough analysis of prosthesis size was conducted by measuring the outer diameter of the prosthetic sizers provided by the manufacturer with a neutral instrument (Digimate digital caliper, model CD-15B; Mitutoyo, Aberden, UK). The average value of three consecutive measurements was used. Echocardiography Measurements were performed with a Philips I33 echocardiograph (Andover, MA) with a 2.5-MHz broadband transducer. The LV mass was calculated using the formula LV mass 1.04 [(LVIDd IVSd LPWDd)3 (LVIDd) 3] 13.6, and was normalized to the body surface area. (LVIDd left ventricular internal diameter in diastole; IVSd intraventricular septal diameter; LPWDd left posterior wall diameter in diastole.) Left Table 2. Perioperative Data a Variable Sorin Soprano Medtronic Mosaic p Value Concomitant CABG 119 (51%) 93 (68%) 0.001 Emergency operation 8 (3.4%) 2 (1.5%) 0.263 Moderate valvular 70 (30%) 39 (29%) 0.787 calcification Severe valvular calcification 125 (53%) 87 (64%) 0.053 Moderate aortic calcification 49 (21%) 33 (24%) 0.468 Severe aortic calcification 16 (7%) 11 (8%) 0.662 Aortic annuloplasty 13 (6%) 6 (4%) 0.626 PPM in vivo (0.85 cm 2 /m 2 ) b 146 (84%) 71 (74%) 0.038 PPM in vivo (0.65 cm 2 /m 2 ) b 63 (36%) 44 (46%) 0.131 Cross-clamp time (min) 89 28 98 30 0.003 CPB (min) 124 38 136 43 0.009 Transvalvular gradient Peak 76 26 80 27 0.143 Mean 45 16 46 19 0.558 AVA (cm 2 ) 0.69 0.22 0.70 0.26 0.813 Annulus diameter 20.6 2.5 22.6 2.0 0.001 a Values are number and percentage, or mean standard deviation. b The analysis is based on patients where EOA in vivo was available. AVA aortic (native) valve area; CABG coronary artery bypass grafting; CPB cardiopulmonary bypass time; EOA effective orifice area; in vivo EOA derived by echocardiography; PPM prosthesis patient mismatch.

1312 NOZOHOOR ET AL Ann Thorac Surg PROSTHESIS PATIENT MISMATCH AND DIASTOLIC FUNCTION 2008;85:1310 8 ventricular hypertrophy was defined as the LV mass normalized to body surface area of greater than 131 g/m 2 in males and greater than 100 g/m 2 in females. The LV ejection fraction was determined using the area-length method. Left ventricular inflow was measured in pulsedwave Doppler mode at the tip of the mitral valve leaflets, and the following variables were recorded: peak velocity of early (E) and late (A) filling, and deceleration time of the E wave velocity. Diastolic function was graded as normal, impaired relaxation, pseudonormal, or restrictive filling. Normal was defined as E wave velocity greater than A wave and normal-sized left atrium. Impaired relaxation was defined as E wave velocity lower than A wave and deceleration time greater than 250 ms. Pseudonormalization was defined as E wave velocity higher than A wave, enlarged left atrium, and the E/A ratio greater than 1. The E wave was correlated to the e= from the tissue Doppler of the medial atrioventricular plane and the ratio was calculated. In patients with E greater than A, an E/e= ratio greater than 15 was considered pathologic, indicating that the patient had a moderate diastolic dysfunction with a pseudonormal inflow of the LV. Restrictive filling was defined as an E/A ratio greater than 2 and deceleration time less than 150 ms. The average of three consecutive beats was used. Pressure gradients (peak and mean) of the bioprostheses were calculated from continuous-wave Doppler measurements using the modified Bernoulli equation. Left ventricular outflow tract peak and mean velocities were calculated from pulsed-wave Doppler measurements. The time-velocity integral of the transaortic flow and LV outflow tract flow was calculated. The EOA of the prosthesis was estimated by multiplying the time-velocity integral ratio between LV outflow tract and the prosthesis by the area of the LV outflow tract, using the continuity equation. Moderate PPM was defined as indexed EOA less than 0.85 cm 2 /m 2 and severe PPM as indexed EOA less than 0.65 cm 2 /m 2. Follow-Up Early mortality was defined as all-cause mortality within 30 days after the surgery. Valve-related and cardiacrelated death was classified in accordance with the Guidelines for Reporting Morbidity and Mortality after Cardiac Valvular Operations [14]. Postdischarge survival data and cause of death were obtained from the National Board of Health and Welfare (Socialstyrelsen, Sweden) or, if necessary, from patient records. Follow-up was 100% complete and included 441.8 patient-years. The mean follow-up time was 1.19 0.76 years (median, 1.14; range, 0 to 2.7 years). Echocardiographic data were available for 81% of the patients at a mean of 4.5 6.9 months. The number of patients investigated with complete echocardiographic data at 3, 6, and 12 months was 99, 83, and 50, respectively. Table 3. Preoperative and Postoperative Hemodynamic Data a Variable Sorin Soprano Medtronic Mosaic p Value Preoperative variables LVEF 0.30 0.50 71 (30%) 47 (34%) 0.413 LVEF 0.30 22 (9%) 7 (5%) 0.140 Moderate LVH 47/119 (39%) 28/64 (44%) 0.577 Severe LVH 9/119 (8%) 3/64 (5%) 0.454 LVDF Pseudonormal 30/77 (39%) 15/49 (31%) 0.340 Restrictive 5/82 (6%) 2/51 (4%) 0.585 LVMI (g/m 2 ) 141 43 151 43 0.246 Postoperative variables LVEF 0.30 0.50 38/190 (20%) 24/108 (22%) 0.650 LVEF 0.30 12/190 (6%) 2/108 (2%) 0.080 Moderate LVH 64/173 (37%) 34/102 (33%) 0.540 Severe LVH 14/173 (8%) 13/102 (13%) 0.210 LVDF Pseudonormal 56/112 (50%) 29/75 (39%) 0.127 Restrictive 5/112 (4%) 1/75 (1%) 0.404 Transprosthetic gradient Peak 28 9 32 11 0.003 Mean 16 5 17 6 0.036 EOA (cm 2 ) 1.35 0.35 1.36 0.42 0.833 EOAi in vivo 0.73 0.19 0.74 0.24 0.737 EOAi in vitro 0.85 0.11 0.78 0.16 0.001 LVMI (g/m 2 ) 130 42 121 36 0.227 a Values are number and percentage, or mean standard deviation. EOA effective orifice area; EOAi indexed effective orifice area; in vitro EOA according to the manufacturer; in vivo EOA derived by echocardiography; LVDF left ventricular diastolic function; LVEF left ventricular ejection fraction; LVH left ventricular hypertrophy; LVMI left ventricular mass index. Statistical Analysis Categorical variables were expressed as percentages and continuous variables were expressed as mean standard deviation. Student s t test was used to evaluate continuous variables. Categorical data were compared using the 2 test or Fisher s exact test when the expected frequency was less than 5. For categorical variables not following a normal distribution, a nonparametric test (Mann- Whitney U test) was used. The Wilcoxon signed-rank test and the McNemar test were used to assess differences between two related dichotomous variables. Multivariate analysis was performed using stepwise Cox proportional hazard regression analysis to determine independent predictors of long-term survival. The inclusion criterion for each outcome was a probability less than 0.200, and the limit for stepwise backward elimination was a probability less than 0.100. Survival was analyzed in a timerelated manner using the Kaplan Meier method, and differences in survival were analyzed with the log-rank test. Risk-adjusted survival was assessed with Cox proportional hazard regression analysis. Missing values were replaced using the probability imputation technique [15]. Statistical significance was defined as a probability less than 0.05. Statistical analysis was performed with the SPSS statistical software package (SPSS 15.0, Chicago, IL).

Table 4. Preoperative and Postoperative Doppler Echocardiographic Variables a Size 18 (18.00 mm) 20 (19.96 mm) 22 (21.89 mm) 24 (23.91 mm) (n 63) (n 92) (n 62) (n 18) Sorin Soprano Preop Postop Preop Postop Preop Postop Preop Postop Mean gradient 50 17 19 6(n 48) b 44 16 15 5(n 71) b,c 43 15 14 5(n 49) b,c 34 13 12 4(n 14) b Peak gradient 82 28 33 10 (n 48) b 76 25 27 8(n 70) b,c 73 23 25 8(n 49) b 58 17 23 7(n 14) b EOA (cm 2 ) 0.65 0.2 1.19 0.3 (n 45) b 0.7 0.2 1.3 0.2 (n 68) b,c 0.7 0.2 1.4 0.4 (n 47) b,c 0.7 0.2 1.7 0.6 (n 13) b LVMI (g/m 2 ) 134 33 123 44 (n 14) 135 44 127 40 (n 33) 145 39 133 38 (n 23) 239 6 144 64 (n 7) c EOAi (cm 2 /m 2 )... 0.70 0.2 (n 45)... 0.71 0.1 (n 68)... 0.73 0.2 (n 47)... 0.89 0.3 (n 13) PPM (%) EOAi... 89... 88 d... 81 d... 62 d ( 0.85 cm 2 /m 2 ) EOAi ( 0.65 cm 2 /m 2 )... 40... 38 d... 36 d... 15 d Size 19 (16.58 mm) 21 (18.51 mm) 23 (20.56 mm) 25 and 27 (22.55 and 24.06 mm) (n 8) (n 47) (n 52) (n 28, 2) Medtronic Mosaic Preop Postop Preop Postop Preop Postop Preop Postop Mean gradient 66 5 21 8(n 5) b 54 19 18 7(n 34) b,c 41 20 16 6(n 42) b,c 39 11 16 5(n 21) b Peak gradient 84 33 39 12 (n 5) b 90 28 33 13 (n 34) b,c 75 26 30 10 (n 42) b 70 17 30 9(n 21) b EOA (cm 2 ) 0.5 0.1 1.3 0.4 (n 5) b 0.6 0.2 1.3 0.5 (n 31) b 0.7 0.2 1.3 0.3 (n 39) b 0.9 0.3 1.6 0.4 (n 21) b,c LVMI (g/m 2 ) NA 146 44 (n 2) 150 47 112 29 (n 17) b 150 40 112 29 (n 20) b 156 47 147 43 (n 11) c EOAi (cm 2 /m 2 )... 0.8 0.2 (n 5)... 0.8 0.3 (n 31)... 0.7 0.2 (n 39)... 0.8 0.2 (n 21) PPM (%) EOAi... 60... 68 d... 82 d... 71 d ( 0.85 cm 2 /m 2 ) EOAi ( 0.65 cm 2 /m 2 )... 40... 42 d... 59 d... 29(p 0.03) a Values are mean standard deviation. b p 0.05 preoperative versus postoperative. c p 0.05 between valve sizes postoperatively. d Not significant between valve sizes. EOA effective orifice area; EOAi effective orifice area index; LVMI left ventricular mass index; NA data not available; Postop postoperative; Preop preoperative; PPM prosthesis patient mismatch. Ann Thorac Surg NOZOHOOR ET AL 2008;85:1310 8 PROSTHESIS PATIENT MISMATCH AND DIASTOLIC FUNCTION 1313

1314 NOZOHOOR ET AL Ann Thorac Surg PROSTHESIS PATIENT MISMATCH AND DIASTOLIC FUNCTION 2008;85:1310 8 in LV mass regression between patients with moderate PPM (p 0.535) or severe PPM (p 0.653) and patients without PPM. Neither moderate (p 0.726) nor severe PPM (p 0.353) was a predictor of impaired diastolic or systolic LV function (p 0.519 and p 0.083, respectively) postoperatively. Early Mortality and Morbidity The overall 30-day mortality was 1.7% (4 of 235 patients) in the Sorin Soprano group and 2.9% (4 of 137 patients) in the Medtronic Mosaic group (p 0.473). None of the early deaths occurred during the operation. Causes of 30-day mortality were cardiac-related in 3 of 8 patients and noncardiac in 5 of 8 of the patients (gastrointestinal hemorrhage in 1 patient, cerebrovascular insult in 2 patients, ischemic bowel disease in 1 patient, renal failure in 1 patient). The Medtronic Mosaic prosthesis was more frequently implanted in patients undergoing concomitant coronary artery bypass grafting (p 0.001), resulting Fig 1. Preoperative and postoperative pattern of left ventricular diastolic function in patients undergoing aortic valve replacement. Results Hemodynamic Evaluation The hemodynamic performance of the prostheses after surgery is summarized in Tables 3 and 4. The reduction of both peak and mean transvalvular gradients after AVR resulted in a significant reduction of LV mass index, 144 43 g/m 2 (preoperatively) versus 126 40 g/m 2 (postoperatively; p 0.001; n 127). Patients receiving the Sorin Soprano valve exhibited an absolute mean reduction of LV mass index of 11 g/m 2 (95% confidence interval, 0.11 to 21.7; p 0.048) and the corresponding value for patients receiving the Medtronic Mosaic valve was 30 g/m 2 (95% confidence interval, 15.5 to 45.2; p 0.001). The overall pattern of improvement in LV diastolic function after AVR is presented in Figure 1. The characteristics of patients with preoperative diastolic dysfunction are presented in Table 5. The implantation of a Sorin Soprano or Medtronic Mosaic prosthesis was not a predictor of postoperative improvement of diastolic (p 0.714) or systolic LV function (p 0.276). Prosthesis Patient Mismatch The incidence of PPM with the two prostheses is presented in Tables 2 and 4. Patients with moderate PPM had significantly higher mean (16.5 5.5 mm Hg versus 14.0 5.8 mm Hg; p 0.004) and peak (29.9 9.6 mm Hg versus 25.7 9.5 mm Hg; p 0.005) transprosthetic gradients than patients without PPM. Similarly, patients with severe PPM had significantly higher mean (17.7 5.3 mm Hg versus 14.9 5.6 mm Hg; p 0.001) and peak (32.8 9.3 mm Hg versus 26.6 9.3 mm Hg; p 0.001) transprosthetic gradients than patients with moderate PPM or without PPM. There was no significant difference Table 5. Characteristics of Patients With Normal Diastolic Dysfunction or Impaired Relaxation Compared With Patients With Pseudonormalization or Restrictive Filling a Variable Normal or Impaired Relaxation (n 81) Pseudonormal or Restrictive Filling (n 52) p Value Sex (female) 33 (41%) 24 (46%) 0.538 Mortality (30-day) 2 (2.5%) 2 (3.8%) 0.644 Diabetes mellitus 9 (11%) 11 (21%) 0.114 AMI 14 (17%) 12 (23%) 0.411 LVEF 0.50 52 (64%) 24 (46%) 0.040 NYHA IV 5 (6%) 7 (13%) 0.215 CABG 53 (65%) 26 (50%) 0.077 LCOS 7 (9%) 7 (13%) 0.398 Age at implant (y) 76.8 6.0 76.6 6.1 0.898 Postoperative gradient Peak 28 10.2 30 9.7 0.356 Mean 16 6.5 16 5.5 0.943 LVMI (g/m 2 ) preoperative 131 41 162 44 0.008 LVMI (g/m 2 ) postoperative 125 38 130 47 0.641 LVEDD preoperative 49.6 6.6 53.4 8.9 0.055 LVEDD postoperative 50.5 7.5 48.1 5.8 0.154 LA preoperative 44.4 5.9 44.3 7.0 0.927 LA postoperative 46.5 7.4 45.4 8.5 0.571 LOS ICU (h) 41 37 81 187 0.060 Hours on ventilator 14 20 32 82 0.055 Cross-clamp time (min) 95 27 91 37 0.435 CPB (min) 131 35 127 51 0.590 a Values are number and percentage, or mean standard deviation. AMI acute myocardial infarction within 3 months preoperatively; CABG coronary artery bypass grafting; CPB cardiopulmonary bypass time; LA left atrial diameter as measured in end-systole in parasternal long-axis view; LCOS low cardiac output syndrome in intensive care unit; LOS ICU length of stay in thoracic intensive care unit; LVEDD left ventricular end-diastolic diameter; LVEF left ventricular ejection fraction; LVMI left ventricular mass index; NYHA New York Heart Association functional class.

Ann Thorac Surg NOZOHOOR ET AL 2008;85:1310 8 PROSTHESIS PATIENT MISMATCH AND DIASTOLIC FUNCTION 1315 Fig 2. Overall survival after aortic valve replacement with implantation of the Medtronic Mosaic and Sorin Soprano prostheses (p 0.476). in longer aortic cross-clamping time (p 0.003) and cardiopulmonary bypass time (p 0.009; Table 2). Multivariate analysis identified advanced age (p 0.030), preoperative myocardial infarction (p 0.049), diabetes mellitus (p 0.026), female sex (p 0.025), moderate aortic annular calcification (p 0.042), perioperative myocardial infarction (p 0.036), and postoperative cerebrovascular insult (p 0.031) as independent risk factors for 30-day mortality. Early mortality was not affected by moderate (p 1.000) or severe PPM (p 0.565). No significant differences were found when comparing the two valve groups regarding postoperative complications (reoperation for bleeding, p 0.662; postoperative atrial fibrillation, p 0.465; low cardiac output syndrome, p 0.239; renal failure, p 0.121; and length of stay in the intensive care unit, p 0.270). Late Mortality and Outcome No patient required repeated surgery during follow-up owing to prosthetic structural valve deterioration. One patient with a Medtronic Mosaic bioprosthesis underwent prosthetic explant during the follow-up period because of prosthetic valve endocarditis with a subvalvular abscess, but no sign of structural valve deterioration was seen perioperatively. Independent risk factors for late mortality after AVR were preoperative diastolic dysfunction (pseudonormalization; p 0.004), severe postoperative LV hypertrophy (LVH; p 0.001), preoperative cerebrovascular insult (p 0.028), preoperative myocardial infarction (p 0.011), preoperative LV ejection fraction 0.30 to 0.50 (p 0.001), preoperative New York Heart Association functional class IV (p 0.011), postoperative dialysis (p 0.001), and reoperation for bleeding (p 0.006). The actuarial survival at 2 years was 91.4% 2.7% and 90.5% 2.9% for the Sorin Soprano and the Medtronic Mosaic valves, respectively (p 0.476; Fig 2). Multivariable proportional hazard regression analysis for risk factors affecting overall survival after AVR demonstrated no significant difference in survival between the patients receiving the two different prostheses (p 0.480). Comment Despite progress in the construction and design of bioprostheses, the hemodynamic performance still does not match the native aortic valve because parts of the sewing ring and stent construction are positioned within the aortic outflow tract, causing a degree of blood flow obstruction. The resulting functional prosthetic orifice area may lead to a residual transvalvular gradient and an increased incidence of PPM. Garcia and associates [16] demonstrated that PPM after AVR is a predictor of late cardiac death and poor early survival for patients with increased LV mass index, but concluded that the incidence of PPM was reduced with the use of the latest generation of supraannular prostheses. In this study, we assessed a third-generation pericardial bioprosthesis, the Sorin Soprano, available for commercial use in Europe since 2003. This bioprosthesis is implanted in the supraannular position and may therefore be of advantage in patients with small aortic annuli [11, 12]. However, the influence of supraannular prostheses on LV mass regression and incidence of PPM are still under debate. Furthermore, patients with aortic stenosis are known to have LVH and consequently diastolic dysfunction [17]. Very few studies have been performed to investigate the impact of PPM on DHF, and this is the largest series of patients undergoing AVR with the Sorin Soprano at a single center. The present population demonstrated a high incidence of severe PPM (39.8%) and no similar incidence figures have been reported previously. On the contrary, in a previous study by Pavoni and coworkers [11], a low incidence of PPM and lower transvalvular gradients were reported when using the Sorin Soprano valve. However, despite the high incidence of PPM in the present study, postoperative transprosthetic gradients, EOAs, and degree of LV mass regression (Table 4) were similar to those in previous reports [13, 18], supporting the findings of the present study. Furthermore, despite a high incidence, PPM was not found to be a significant predictor of impaired LV mass regression, regardless of severity. Overall, our study demonstrated favorable LV mass regression and excellent clinical outcome, which brings into question the importance of PPM. In contrast to the previous findings of Garcia and colleagues [16], the latest generation of supraannular prostheses seem to be associated with a high incidence of PPM in this study. Improvements in hemodynamic outcome reached statistical significance only in the larger prosthetic sizes, and this finding is in accord with a previous study by Botzenhardt and associates [13]. Diastolic heart failure (defined as pseudonormalization or restrictive filling), in the absence of impaired LV ejection fraction, was present in nearly half of the patients (Table 5). Preoperative LV mass index was significantly higher in this group than in the patients with normal diastolic dysfunction and those with impaired relaxation (p 0.008). Persistent LVH is one of the known causes of DHF [5]. Ina previous study, Bové and colleagues [19] found the preexistence of advanced LVH to be a major obstacle for LV mass regression in spite of an otherwise successful AVR. Therefore, the presence of PPM would intuitively have a negative impact on the recovery of diastolic function. However, no

1316 NOZOHOOR ET AL Ann Thorac Surg PROSTHESIS PATIENT MISMATCH AND DIASTOLIC FUNCTION 2008;85:1310 8 significant relation could be demonstrated between impaired recovery of DHF and PPM. The limited follow-up period and relatively high incidence of advanced LVH in this elderly population may explain the lack of significant improvement in diastolic function despite postoperative LV mass regression. Several risk factors for early and late mortality after AVR with bioprostheses were identified in the present study, confirming the findings of previous studies [20, 21]. However, neither moderate nor severe PPM was found to be an independent risk factor for increased 30-day or late mortality. Furthermore, no significant differences in clinical outcome could be related to prosthetic type. Lund and coworkers [9] have previously reported impaired systolic and diastolic LV function to be independent preoperative predictors of early as well as late mortality after AVR, and suggested the underlying mechanism to be mainly concentric LVH. In contrast, Nakagawa and colleagues [22] found that the occurrence of DHF did not affect postoperative early mortality. In our study, diastolic dysfunction (pseudonormalization) and preoperative advanced LVH were predictors of late mortality. The divergence in the results of previous studies may reflect differences in the duration of follow-up and thus the dynamic remodeling process, which is initiated by the reduction of LV afterload resulting from AVR [19, 23]. Previous studies have demonstrated that the regression of LV mass after AVR peaks after 1 year [9], and that the degree of DHF may deteriorate up to 10 years postoperatively, despite significant LV mass regression [8]. The present study has obvious limitations because of nonrandomized design, although patient variables were prospectively collected in the database. Furthermore, not all the patients underwent postoperative echocardiography, and prosthetic performance was not assessed with echocardiographic exercise testing. The study was limited by its exclusive use of echocardiographic techniques for the diagnosis of diastolic abnormalities. The follow-up time was relatively short, and we cannot rule out that a longer follow-up period might add further information regarding diastolic remodeling. In conclusion, we have demonstrated that PPM was not a predictor of poor survival and did not impair LV mass regression or prosthetic hemodynamics. Both the Sorin Soprano and the Medtronic Mosaic prostheses demonstrated excellent early and late survival, and no sign of structural valve deterioration was identified during follow-up. In our opinion, the complete supraannular design does not necessarily lead to a greater EOA or reduced incidence of PPM. In this context, and in a setting with elderly and physically less active patients, PPM may be of subordinate clinical importance, whereas other covariables may be of greater importance for clinical outcome. The development of preoperative DHF was identified as an independent predictor of poor outcome after AVR. Our findings highlight the importance of appropriate timing of surgical treatment of aortic stenosis as late referral may be associated with irreversible LVH and the development of DHF. The authors would like to thank Håkan Lövkvist, RSKC, Lund University Hospital Statistical Skills Center, for his expert contribution to the statistical evaluation. The echocardiographic examinations of the prostheses were financially supported by Sorin Biomedica Cardio SpA. The authors had full control of the study design, analysis of data, and production of the written report. References 1. Rahimtoola SH. The problem of valve prosthesis-patient mismatch. Circulation 1978;58:20 4. 2. Blais C, Dumesnil JG, Baillot R, Simard S, Doyle D, Pibarot P. Impact of valve prosthesis-patient mismatch on short-term mortality after aortic valve replacement. Circulation 2003; 108:983 8. 3. Rao V, Jamieson WR, Ivanov J, Armstrong S, David TE. Prosthesis-patient mismatch affects survival after aortic valve replacement. Circulation 2000;102(19 Suppl 3):III-5 9. 4. Ruel M, Al-Faleh H, Kulik A, Chan KL, Mesana TG, Burwash IG. Prosthesis-patient mismatch after aortic valve replacement predominantly affects patients with preexisting left ventricular dysfunction: effect on survival, freedom from heart failure, and left ventricular mass regression. J Thorac Cardiovasc Surg 2006;131:1036 44. 5. Fischer M, Baessler A, Hense HW, et al. Prevalence of left ventricular diastolic dysfunction in the community: results from a Doppler echocardiographic-based survey of a population sample. Eur Heart J 2003;24:320 8. 6. Paulus WJ, Tschope C, Sanderson JE, et al. How to diagnose diastolic heart failure: a consensus statement on the diagnosis of heart failure with normal left ventricular ejection fraction by the Heart Failure and Echocardiography Associations of the European Society of Cardiology. Eur Heart J 2007;28:2539 50. 7. Dineen E, Brent BN. Aortic valve stenosis: comparison of patients with to those without chronic congestive heart failure. Am J Cardiol 1986;57:419 22. 8. Gjertsson P, Caidahl K, Farasati M, Oden A, Bech-Hanssen O. Preoperative moderate to severe diastolic dysfunction: a novel Doppler echocardiographic long-term prognostic factor in patients with severe aortic stenosis. J Thorac Cardiovasc Surg 2005;129:890 6. 9. Lund O, Flo C, Jensen FT, et al. Left ventricular systolic and diastolic function in aortic stenosis: prognostic value after valve replacement and underlying mechanisms. Eur Heart J 1997;18:1977 87. 10. Dhareshwar J, Sundt TM III, Dearani JA, Schaff HV, Cook DJ, Orszulak TA. Aortic root enlargement: what are the operative risks? J Thorac Cardiovasc Surg 2007;134:916 24. 11. Pavoni D, Badano LP, Musumeci SF, et al. Results of aortic valve replacement with a new supra-annular pericardial stented bioprosthesis. Ann Thorac Surg 2006;82:2133 8. 12. Gerosa G, Tarzia V, Rizzoli G, Bottio T. Small aortic annulus: the hydrodynamic performances of 5 commercially available tissue valves. J Thorac Cardiovasc Surg 2006;131:1058 64. 13. Botzenhardt F, Eichinger WB, Bleiziffer S, et al. Hemodynamic comparison of bioprostheses for complete supraannular position in patients with small aortic annulus. J Am Coll Cardiol 2005;45:2054 60. 14. Edmunds J, Clark RE, Cohn LH, Grunkemeier GL, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. J Thorac Cardiovasc Surg 1996;112:708 11. 15. Schemper M, Smith TL. Efficient evaluation of treatment effects in the presence of missing covariate values. Stat Med 1990;9:777 84. 16. Garcia Fuster R, Estevez V, Rodriguez I, et al. Prosthesis patient mismatch with latest generation supra-annular prostheses. The beginning of the end? Interact Cardiovasc Thorac Surg 2007;6:462 9.

Ann Thorac Surg NOZOHOOR ET AL 2008;85:1310 8 PROSTHESIS PATIENT MISMATCH AND DIASTOLIC FUNCTION 1317 17. Aronow WS. Valvular aortic stenosis in the elderly. Cardiol Rev 2007;15:217 25. 18. Dalmau MJ, Maria Gonzalez-Santos J, Lopez-Rodriguez J, Bueno M, Arribas A, Nieto F. One year hemodynamic performance of the Perimount Magna pericardial xenograft and the Medtronic Mosaic bioprosthesis in the aortic position: a prospective randomized study. Interact Cardiovasc Thorac Surg 2007;6:345 9. 19. Bove T, Van Belleghem Y, Francois K, Caes F, Van Overbeke H, Van Nooten G. Stentless and stented aortic valve replacement in elderly patients: factors affecting midterm clinical and hemodynamical outcome. Eur J Cardiothorac Surg 2006; 30:706 13. 20. Edwards FH, Peterson ED, Coombs LP, et al. Prediction of operative mortality after valve replacement surgery. J Am Coll Cardiol 2001;37:885 92. 21. Nozohoor S, Nilsson J, Luhrs C, Roijer A, Sjogren J. The influence of patient-prosthesis mismatch on in-hospital complications and early mortality after aortic valve replacement. J Heart Valve Dis 2007;16:475 82. 22. Nakagawa D, Suwa M, Ito T, Kono T, Kitaura Y. Postoperative outcome in aortic stenosis with diastolic heart failure compared to one with depressed systolic function. Int Heart J 2007;48:79 86. 23. Gjertsson P, Caidahl K, Bech-Hanssen O. Left ventricular diastolic dysfunction late after aortic valve replacement in patients with aortic stenosis. Am J Cardiol 2005;96: 722 7. INVITED COMMENTARY The concept of prosthesis-patient mismatch (PPM) after aortic valve replacement (AVR) remains controversial today as long-term follow-up studies with adequate power to detect survival differences continue to provide discordant results [1 5]. Our previous study in 2000 demonstrated impaired survival after 7 years in patients who had PPM defined as an indexed effective orifice area (EOA) of 0.85 cm 2 /m 2 [1]. Due to the fact that we used the manufacturer s reported in-vitro EOA, we liberalized our definition of PPM compared with the 0.75 cm 2 /m 2 defined by Rahimtoola s [4] original description. Similarly, Pibarot and colleagues [3] reported adverse hemodynamic effects of PPM in 392 patients undergoing stented AVR. Pibarot and colleagues [3] study revealed no survival differences, but important adverse effects of PPM on postoperative hemodynamics and symptom status. Subsequent studies by Pibarot and colleagues examined patients subjected to exercise echocardiography and documented severe hemodynamic compromise in patients with even moderate PPM. The study by Medalion and colleagues [5] reviewed 892 patients who received tissue and mechanical prostheses and used the calculated geometric orifice area to determine if PPM conferred a survival hazard. These authors found that overall mortality was not different at a mean of 5 years in those patients with PPM versus those who received an appropriately sized prosthesis. In this issue of The Annals of Thoracic Surgery, Nozohoor and colleagues [2] reviewed their experience with 372 patients and examined a different, surrogate endpoint, namely, diastolic function. Theoretically, any prosthesis that results in a residual obstruction to forward flow would be expected to create an outflow tract gradient and hence limit regression of left ventricular mass leading to persistent diastolic dysfunction. Indeed, the authors [2] found that PPM resulted in higher postoperative peak and mean transvalvular gradients. However, these differences did not translate into significant changes in left ventricular mass regression or echocardiographic assessments of diastolic function. Intuitively, surgeons always attempt to insert the largest possible prosthesis in a given aortic root, so why is there a controversy? Those who believe that PPM has adverse hemodynamic effects argue for a change in operative strategy: a different prosthesis with more favorable hemodynamics (ie, stentless tissue or mechanical) or an annular enlargement procedure. In contrast, some surgeons argue that the increased operative risk attributable to a more complex operation (whether a stentless valve insertion or an annular enlargement) outweighs any potential risk of PPM. These surgeons point to the lack of convincing data implicating PPM as an independent risk factor for either short-term or long-term survival. Why are these studies negative? The most common shortcoming is inadequate power followed by an inadequate length or quality of follow-up. Indeed, in our previous study of PPM, survival curves did not diverge until 7 years of follow-up and required a sample size of 2,981 patients. Another important limitation is the choice of endpoint. The study of Medalion and colleagues [5] reported all-cause mortality in a population of patients that are at risk for noncardiac related death. Furthermore, few studies report quality-of-life data. Recalling the earlier important study by Pellikka and colleagues [6], even hemodynamically significant aortic stenosis is well tolerated for up to 5 years and only becomes dangerous when patients develop symptoms. Mortality of asymptomatic patients was only 4%. Surgeons would never make a decision regarding surgical intervention without knowledge of a patient s symptomatic status, so why are we so willing to dismiss the concept of PPM without a consideration of a patient s postoperative functional class? The study by Nozohoor and colleagues [2] will no doubt add to the controversy surrounding this important subject; however, based on the previous discussion, the limitations of this study must be placed into context. Clinical follow-up is limited to an average of 1 year and echocardiographic data (determining LV mass and diastolic function), was only performed in 81% of patients at a mean postoperative follow-up of only 5 months. Importantly, echocardiograms were done in the resting state and no attempt was made to examine changes with exercise. In the end, I firmly believe that surgeons act in their patient s best interest, and when a more complex operation is contemplated in a high-risk, elderly patient, the decision to leave a degree of PPM may be justified. However, a similar decision-making process must be extended to the young, active, and large patient who will undoubtedly benefit from a more aggressive surgical approach to their aortic valve disease. 2008 by The Society of Thoracic Surgeons 0003-4975/08/$34.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2008.02.002