Reoperations after primary aortic valve replacement

Third-Time Aortic Valve Replacement: Patient s and Operative Outcome Kasra Shaikhrezai, MD, MRCS, Giordano Tasca, MD, FETCS, Mohamed Amrani, PhD, FETCS, Gilles Dreyfus, MD, FETCS, and George Asimakopoulos, MD, PhD Department of Cardiac Surgery, Harefield Hospital, Royal Brompton and Harefield NHS Trust, London, United Kingdom Background. Reoperative cardiac surgery is being performed with increasing frequency. Third-time aortic valve surgery remains a rare procedure. We retrospectively analyzed the outcome of third-time aortic valve replacement (AVR) at our institution. Methods. Between 1990 and 2005, 49 patients underwent third-time AVR. Data analyzed included preoperative patient characteristics, type of preexisting aortic valve prosthesis, prosthetic valve pathology necessitating third-time AVR, postoperative morbidity and mortality, and echocardiographic data. Results. The mean age was 47.4 17 years. The mean interval between the first and second operation was 8.7 5.7 years, and between the second and third operation it was 10.2 5.6 years. Prosthetic valves at the time of second AVR included 32 homografts (65.4%), 11 mechanical prostheses (22.4%), and 6 xenografts (12.2%). At third-time AVR, 29 patients (59.2%) received a homograft or autograft, 12 (24.5%) received a mechanical valve, and 8 (16.3%) received a xenograft. In-hospital mortality was 4.1%. The mean follow-up was 80 69 months. Freedom from reoperation was 84% 6% at 5 years and 65% 11% at 10 years. Long-term survival was 79% 6% at 5 years and 73% 7% at 10 years. Multivariate analysis showed that age, female sex, and postoperative high left ventricular mass were factors associated with decreased long-term survival. Mean left ventricular mass decreased from 320 133 g to 263 102 g at 1 year postoperatively (p 0.01). Conclusions. Third-time AVR can be performed with low operative mortality, low cumulative operative mortality, and satisfactory long-term survival and freedom from reoperation. The procedure results in significant regression of left ventricular mass. (Ann Thorac Surg 2010;89:479 84) 2010 by The Society of Thoracic Surgeons Reoperations after primary aortic valve replacement (AVR) have been reported with increasing frequency in recent years. Second-time AVR was previously considered as a high-risk procedure, but operative mortality and morbidity is declining and, in some studies, does not exceed those after primary AVR [1 3]. Published guidelines for reporting morbidity and mortality after valve operations divide causes of reoperation into structural valvular deterioration, nonstructural dysfunction, valve thrombosis, and surgical valvular endocarditis [4]. The choice of prosthetic valve type is often controversial, particularly for reoperation. Mechanical prostheses are resistant to structural deterioration and dysfunction but carry considerable long-term morbidity owing to their requirement for anticoagulation therapy [5]. Structural valve deterioration affects primarily bioprostheses and tends to increase 10 to 15 years after implantation [6]. Bioprosthetic valves and homografts, in particular, have been used widely in our institution for their favorable hemodynamic profile and reduced need for anticoagulation therapy. Second-time AVR after primary implantation of a homograft is safe and durable [2]. Third-time valve surgery, Accepted for publication April 14, 2009. Address correspondence to Dr Asimakopoulos, Level 4, Dolphin House, Bristol Royal Infirmary, Bristol, BS2 8HW, United Kingdom; e-mail: george.asimakopoulos@uhbristol.nhs.uk. however, remains a much rarer event. Literature reporting patient outcome after third-time AVR remains very limited. Patient groups are usually small and comprise heterogeneous subpopulations of larger series [1, 7]. The objective of this study was to review our experience with patients undergoing third-time AVR over a 15-year period. Material and Methods This study involved retrospective data analysis of all patients who underwent third-time AVR at the Cardiothoracic Department of Harefield Hospital, UK, between January 1990 and December 2005. All the clinical data were collected retrospectively from patients record and our computerized database that stores information prospectively in line with the appended Minimum Dataset (MDS) defined by the Society of Cardiothoracic Surgeons of Great Britain and Ireland [8]. Institutional Review Board approval was obtained for this study, and the informed consent requirement was waived. During the study period, 49 patients underwent thirdtime AVR at Harefield Hospital. During the same period, 1,155 patients underwent first-time and 162 patients underwent second-time AVR. Data analyzed included preoperative patient characteristics, type of preexisting aortic valve prosthesis, prosthetic valve pathology neces- 2010 by The Society of Thoracic Surgeons 0003-4975/10/$36.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2009.04.044

480 SHAIKHREZAI ET AL Ann Thorac Surg THIRD-TIME AORTIC VALVE REPLACEMENT 2010;89:479 84 sitating third-time AVR, postoperative morbidity and mortality, and echocardiography data. Patients undergoing concomitant further cardiac surgical procedures were included in the study. There was 100% completeness of follow-up. All interventions were performed through resternotomy, with no reentry injury. This was followed by ascending aorta cannulation, right atrial or bicaval cannulation, mild systemic hypothermia, and administration of cold crystalloid or blood cardioplegia. The aortic valve was assessed through aortotomy, and the decision regarding type of aortic prosthesis was made. Intraoperative transesophageal echocardiography was performed in all procedures after 1997. Follow-up was performed in the outpatient department on a 6-monthly or annual basis. Definitions Definition of morbidity, mortality, and consequences of morbid events were reported according to published guidelines [4]. Early death was defined as in-hospital death or death within 30 days from the day of operation. Structural valvular deterioration (SVD) was defined as an intrinsic abnormality of the valve causing stenosis or regurgitation but not associated with infection or thrombosis. Nonstructural dysfunction was any abnormality of the aortic valve that was not intrinsic to the valve itself, such as pannus or paravalvular leak, in the absence of endocarditis or thrombosis. Prosthetic valve endocarditis was any infection involving aortic valve prosthesis. The diagnosis was based on Duke s criteria and confirmed at reoperation [9]. A neurologic event was any new temporary or permanent neurologic deficit diagnosed clinically and assessed by computer tomography. Statistical Analysis Continuous variables were expressed as mean SD values unless stated otherwise. Categorical variables were expressed as percentage. One-way analysis of variance (ANOVA) for repeated measurement was used to evaluate the left ventricular mass (LVM) regression over the study period. If the F value was significant and, since the variance was homogeneous, the Tukey s multiple comparison test was used. Cumulative probability values of survival and operation free survival were estimated by the Kaplan-Meier method, reported as mean SEM, and compared with the log-rank test. The effect of the preoperative, operative, and postoperative variables on survival was assessed with the Cox proportional hazard model. The variables with a p value less than 0.1 were inserted in the final models. The variables tested in the models were as follows: (1) preoperative variables were age, sex, body mass index, New York Heart Association (NYHA) functional class, hypertension, diabetes mellitus, coronary artery disease, history of myocardial infarction, LV fractional shortening, history of heart failure, arteriopathy, chronic renal insufficiency, chronic obstructive pulmonary disease, sinus rhythm, and LVM; (2) perioperative variables were urgent/emergent operation, aortic crossclamp time, etiology of valve disease, type of prosthesis implanted (stentless bioprosthesis, stented bioprosthesis, mechanical prosthesis, or homograft/ross procedure; and (3) the postoperative variable was LVM at 3 months and 1 year after surgery. The data were statistically analyzed with the use of SPSS 13.0 (SPSS, Chicago, IL), and the statistical significance was assumed when a p value was less than 0.05. Results Preoperative and Perioperative s Preoperative and perioperative characteristics are presented in Tables 1 and 2. The mean age at the time of the third-time AVR was 47.4 17 years. The mean interval between the first and second AVR was 8.7 5.7 years, and between the second and third AVR, it was 10.2 5.6 years. Urgent or emergent operations were performed in 18 cases (36.7%). Twenty-six patients (53.1%) presented at NYHA class III or IV. Endocarditis was the definite diagnosis of aortic valve prosthesis failure in 12 cases (24.5%), whereas SVD was the indication for AVR in 37 (75.5%). There were no aortic valves displaying nonstructural dysfunction. In the SVD patient population, there were 26 patients (70.3%) with aortic regurgitation, 6 (16.2%) with AV stenosis, and 5 (13.5%) with mixed valve disease. There are no complete data available on the initial diagnosis for the first AVR. Postoperative s Postoperative characteristics are presented in Table 2. Twenty-three patients (47%) received a homograft. There were 6 autografts (12.2%), 12 mechanical (24.5%), 6 porcine stentless (12.3%), and 2 porcine stented prostheses (4%). Four patients (8.1%) received concomitant CABG, while 1 patient (2%) had concomitant mitral valve replacement. One patient suffered a transient ischemic attack; there were no cases of mediastinitis or myocardial infarction among the surviving patients. The most common postoperative complication was new atrial fibrillation in 7 patients (14.3%). There were 2 hospital deaths (4.1%; 95% confidence interval: 1.1 to 13.7). A 52-year-old man patient received a 23-mm mechanical prosthesis and single vessel CABG (CABG 1); low cardiac output syndrome developed, and he died of multiorgan failure on the fourth postoperative day. A 66-year-old man underwent emergency implantation of a homograft for prosthetic valve endocarditis on an unstented xenograft; there was an extensive abscess and complete destruction of aortic valve annulus, and he died of low cardiac output syndrome within 24 hours. Further Aortic Valve Surgery and Long-Term Survival Ten patients (20.4%) received a further, fourth, AVR during the follow-up period. Of these patients, 4 went on to receive a fifth aortic valve procedure. Table 3 demonstrates the type of implanted valve in the fourth and the fifth AVR. Freedom from reoperation was 84% 6at5

Ann Thorac Surg SHAIKHREZAI ET AL 2010;89:479 84 THIRD-TIME AORTIC VALVE REPLACEMENT 481 Table 1. Preoperative and Perioperative s Table 2. Perioperative and Postoperative s Total no. of patients 49 Age, years Mean SD 47.4 17 Median (interquartile difference) 48 (33 59) Male 43 (87.7%) Body mass index, kg/m 2 24.1 6.8 Body surface area, m 2 1.86 0.28 Interoperative intervals, years, median (interquartile difference) Between first and second AVR 8 (6 11) Between second and third AVR 10 (6 13) First operation Homograft 15 (31%) Valvotomy 10 (20%) Xenograft 3 (6%) Mechanical 2 (4%) Unspecified 19 (39%) Second operation Homograft 32 (65.4%) Mechanical 11 (22.4%) Xenograft 6 (12.2%) Operative priority Elective 31 (63.3%) Urgent 12 (24.4%) Emergent 6 (12.3%) Atrial fibrillation 11 (22.4%) Creatinine 200 mmol/dl 2 (4%) Cerebrovascular accident 6 (12.2%) COPD/asthma 4 (8.1%) Peripheral vascular disease 1 (2%) Hypertension 10 (20.4%) Diabetes mellitus 1 (2%) NYHA functional class 0 II 23 (46.9%) NYHA functional class III IV 26 (53.1%) Endocarditis 12 (24.5%) Structural valvular deterioration 37 (75.5%) AV stenosis 6 (16.2%) AV regurgitation 26 (70.3%) AV mixed disease 5 (13.5%) Cardiopulmonary bypass time, minutes 169 64 Cross-clamp time, minutes 124 43 Intraoperative intra-aortic balloon pump 4 (8.1%) AV aortic valve; AVR aortic valve replacement; COPD chronic obstructive pulmonary disease; NYHA New York Heart Association. years and 65% 11 at 10 years. Long-term survival was 79% 6% at 5 years and 73% 7% at 10 years. Third operation Homograft/autograft 29 (59.2%) Mechanical 12 (24.5%) Xenograft stentless 6 (12.3%) Xenograft stented 2 (4%) Reexploration for bleeding 1 (2%) Reintubation 1 (2%) Renal dialysis 4 (8.1%) Postoperative intra-aortic 3 (6.1%) balloon pump Septicemia 2 (4%) Cerebrovascular accident New atrial fibrillation 7 (14.3%) Hospital deaths (95% 2 (4.1%; 1.1% to 13.7%) confidence interval) Echocardiography s Table 4 presents echocardiographic findings 3 months and 1 year after the third AVR. Fractional shortening and LV dimensions remained statistically unchanged while LVM regressed significantly from 320 133 g preoperatively to 249 87 g after 1 year (p 0.014). Regression Analysis Univariate and multivariable analysis showed that increasing age, female sex, and postoperative high LVM were factors associated with decreased long-term survival (Table 5). Long-term mortality doubled for every 10-year increase in age. It increased by 70% for every 50 g increase of LVM. Comment This retrospective analysis presents unique data after third-time AVR performed at our institution over a 15-year period. Third-time AVR was performed on a relatively young group of patients (mean age, 47.4 years) who were predominately male. Two thirds of the patients had received a homograft as second-time AVR, and an equal percentage of the procedures were elective. Implantation of homograft aortic valve was justified with either patient preference of not being on a regimen of warfarin or warfarin administration contraindication. Table 3. Type of Implanted Valve in the Fourth and Fifth Operations Fourth operation Homograft/autograft 4 (40%) Mechanical 4 (40%) Xenograft stentless 0 Xenograft stented 2 (20%) Fifth operation Homograft/autograft 1 (25%) Mechanical 3 (75%) Xenograft stentless 0 Xenograft stented 0

482 SHAIKHREZAI ET AL Ann Thorac Surg THIRD-TIME AORTIC VALVE REPLACEMENT 2010;89:479 84 Table 4. Echocardiographic Findings Variables Preoperative Third Month One Year p Fractional shortening, % 33 9.7 29 10 30 10 NS LVEDD, mm 61.3 11 57.2 11 56.2 10 NS LVESD, mm 41.8 11 41.0 12 39.8 11 NS Left ventricular mass, g 320 133 263 102 a 249 87 b 0.01 One-way analysis of variance of the echocardiographic findings. a Versus preoperative p value 0.046; b Versus preoperative p value 0.014. LVEDD left ventricular end-diastolic diameter; LVESD left ventricular end-systolic diameter; NS not significant. Moreover, homograft was the favorable bioprosthetic valve in our institution, according to the previous acceptable results [2, 10]. Structural valvular deterioration presented the most common indication for a third-time AVR, and prosthetic valve endocarditis was present in almost 25% of patients. Structural valvular deterioration was attributable to a malfunctioning regurgitant homograft in the majority of patients. A homograft was used in nearly half of the patients undergoing third-time AVR whereas the remaining patients received a mechanical or a variety of bioprosthetic valves. The procedure was carried out nonelectively in 18 patients (36.7%), and severe dyspnea was present in 26 (53.1%). Early mortality was 4.1%. The main strength of this study is that it represents a unique population of patients. The study suggests that thirdtime AVR can be carried out with relatively low morbidity and mortality. Variables associated with increased long-term mortality were advanced age, female sex, and LV hypertrophy. The early mortality in this study was comparable to that in previously published data on second-time AVR. Figures range from 3% to 14.6% [1 3, 7, 11 19] and are within the confidence intervals for early mortality in our report. The highest mortality (14.6%) was observed among patients treated with first reoperative AVR for prosthetic valve endocarditis [14]. Mortality was 22.6% after emergency replacement of a degenerated aortic valve bioprosthesis and only 1.4% for elective reoperation in one series [18]. Mortality after 144 second-time AVRs with a homograft was 3.4% in our institution [2]. Piehler and coworkers [15] published the largest report to date on reoperative valve surgery. Their three-institution retrospective study included 1,068 patients undergoing AVR, with 9.2% hospital mortality. Early mortality in a group of 162 patients undergoing second-time AVR was 5%, not statistically different from early mortality for primary AVR [16]. Kumar and colleagues [7] published the clinical outcome of 178 patients undergoing secondtime or more AVR. The overall 30-day mortality was 12.3%. In a different report, the use of a mechanical prosthesis to replace first-time AVR with a homograft resulted in 4.3% early mortality [3]. In a recent study, mortality was 3.7% after second AVR and 13% among 23 patients undergoing third-time AVR [1]. We could not identify further published early mortality data on thirdtime AVR. Interestingly, mortality of AVR after previous CABG was 16.6% in a study of 145 patients [20], although, in a different report, CABG at the time of the first AVR did not significantly affect mortality after second-time AVR [15]. In view of only 2 early deaths, we were unable to perform meaningful regression analysis for detection of predicting variables. Multivariate predictors of hospital death after second-time AVR, published previously, include advanced age, male sex, renal insufficiency, and nonelective operation [11]. Further identified multivariate predictors include history of chronic obstructive pulmonary disease, LV ejection fraction less than 50% [7], LV ejection fraction less than 25%, concomitant surgical procedures [2], NYHA class, peripheral vascular disease, and endocarditis [1, 13, 15, 16]. Reoperation itself was not significant predictor of hospital mortality [1]. Data from more than 400,000 valve procedures collected in The Society of Thoracic Surgeons database, however, demonstrated that reoperation influences operative mortality with an odds ratio of 1.61 [21]. When reoperation is indicated, early referral is likely to result in low operative mortality by reducing the average NYHA class and the rate of nonelective procedures. Long-term actuarial survival was 73% at 10 years after third-time AVR in our study. This figure was 82% after a second homograft [2]. The 5-year survival after second reoperation AVR for prosthetic valve endocarditis was 59% among a small group of 19 patients [14]. Reoperation-free survival was 84% 6% at 5 years and 65% 11% at 10 years in our study. Freedom from reoperation was 82% at 10 years in the analysis published by Hasnat and coworkers [2]. We identified advanced age, female sex, and LV hypertrophy as independent predictors of late mortality after third-time AVR. Age and sex are well-known risk factors for midterm and long-term survival after AVR [10, 21]. A long-term survival is affected by poor LV function after reoperative homograft [2]. Table 5. Multivariable Cox Proportional Hazard Model Variable p Hazard Ratio 95% CI Age 0.034 2.1 1.1 4.3 Female sex 0.007 38.5 2.7 540.8 Left ventricular mass 0.037 1.7 1.03 2.8 The hazard ratio of the age variable represents the increase in the risk of mortality per 10 years increase in age. For left ventricular mass, the hazard ratio represents the increase of the risk for 50 g increase of left ventricular mass.

Ann Thorac Surg SHAIKHREZAI ET AL 2010;89:479 84 THIRD-TIME AORTIC VALVE REPLACEMENT 483 In this study, third-time AVR resulted in echocardiographic improvement with significant LVM regression as a consequence. Mean LVM regressed by 23% within the first year and by 25% at the last echocardiogram available. We also showed a statistically nonsignificant trend for reduction in the LV end-diastolic diameter. Similarly, LVM index decreased significantly by 16% after secondtime homograft [2]. Left ventricular hypertrophy was a risk factor for long-term mortality in this study. Similarly, LV hypertrophy has previously been shown to be a risk factor after primary AVR. Nearly half of our patients received a homograft as third-time AVR. All of these patients had previously had homografts as the second-time procedure. The mean durability for this type of prosthesis was 10.5 5.6 years and resembles further reports [3]. Homograft has been reported to have optimal hemodynamics with a low risk of prosthetic valve endocarditis or patient-prosthesis mismatch [3, 10, 22]. Structural valvular deterioration was the main indication for reoperation in our population. Homograft SVD usually starts at 8 to 10 years after operation; therefore, we believe that patients were reoperated on without excessive delay. The risk of a reoperation has several implications in the decision regarding which surgical procedure to adopt, reparative or replacement, and what type of prosthesis to choose, biologic or mechanic. There are several recently reported series of AVR reoperations showing low operative mortality with a risk profile that is similar to the one before the primary AVR. Indeed, advanced NYHA class and reduced ejection fraction were predictors for operative mortality in redo patients; thus, it is crucial that patients be referred for surgery as soon as valve dysfunction is detected. In our previous experience, early mortality after first-time AVR amounts to 5%, followed by 0.6% mortality per patient-year, resulting in a total of 17% valve-related mortality over 20 years postoperatively [10]. Certain limitations of this retrospective analysis should be mentioned. The patient population analyzed in this report is relatively small and heterogeneous. Small patient numbers and low complication rates precluded meaningful statistical analysis of various clinical factors. In summary, this report presents outcome after thirdtime AVR over a 15-year period. Results demonstrated low rates of early mortality and complications in this unique patient group. Furthermore, the procedure was followed by significant reduction in LV hypertrophy. We conclude that third-time AVR can be performed safely in selected patients. References 1. Davierwala PM, Borger MA, David TE, Rao V, Maganti M, Yau TM. Reoperation is not an independent predictor of mortality during aortic valve surgery. J Thorac Cardiovasc Surg 2006;131:329 35. 2. Hasnat K, Birks EJ, Liddicoat J, et al. Patient outcome and valve performance following a second aortic valve homograft replacement. Circulation 1999;100(Suppl 2):42 7. 3. Sadowski J, Kapelak B, Bartus K, et al. Reoperation after fresh homograft replacement: 23 years experience with 655 patients. Eur J Cardiothorac Surg 2003;23:996 1001. 4. Edmunds LH, Clark RE, Cohn LH, et al. Guidelines for reporting morbidity and mortality after cardiac valve operations. J Thorac Cardiovasc Surg 1996;112:708 11. 5. Butchart EG, Payne N, Li H, Buchan K, Mandana K, Grunkemeier GL. Better anticoagulation control improves survival after valve replacement. J Thorac Cardiovasc Surg 2002;123: 715 23. 6. Hammermeister K, Sethi GK, Henderson WG, Grover FL, Oprian C, Rahimtoola SH. Outcomes 15 years after valve replacement with a mechanical versus a bioprosthetic valve: final report of the Veterans Affairs randomized trial. J Am Coll Cardiol 2000;36:1152 8. 7. Kumar P, Athanasiou T, Ali A, et al. Re-do aortic valve replacement: does a previous homograft influence the operative outcome? J Heart Valve Dis 2004;13:904 13. 8. Keogh B, Kinsman R. National adult cardiac database report 1990-2000. The Society of Cardiothoracic Surgeons of Great Britain and Ireland, 2000. Available at: www.scts.org. 9. Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings. Duke Endocarditis Service. Am J Med 1994;96:200 9. 10. Lund O, Chandrasekaran V, Grocott-Mason R, et al. Primary aortic valve replacement with allografts over twenty-five years: valve-related and procedure-related determinants of outcome. J Thorac Cardiovasc Surg 1999;117:77 91. 11. Akins CW, Buckley MJ, Daggett WM, et al. Risk of reoperative valve replacement for failed mitral and aortic bioprostheses. Ann Thorac Surg 1998;65:1545 51. 12. Brandao CMA, Pomerantzeff PMA, Souza LR, et al. Multivariate analysis of risk factors for hospital mortality in valvular reoperations for prosthetic valve dysfunction. Eur J Cardiothorac Surg 2002;22:922 6. 13. Jamieson WRE, Burr LH, Miyagishima RT, et al. Reoperation for bioprosthetic aortic structural failure risk assessment. Eur J Cardiothorac Surg 2003;24:873 8. 14. Lytle BW, Priest BP, Taylor PC, et al. Surgical treatment of prosthetic valve endocarditis. J Thorac Cardiovasc Surg 1996;111:198 210. 15. Piehler JM, Blackstone EH, Bailey KR, et al. Reoperation on prosthetic heart valves. Patient-specific estimates of inhospital events. J Thorac Cardiovasc Surg 1995;109:30 48. 16. Potter DD, Sundt TM, Zehr KJ, et al. Operative risk of reoperative aortic valve replacement. J Thorac Cardiovasc Surg 2005;129:94 103. 17. Raanani E, David TE, Dellgren G, Armstrong S, Ivanov J, Feindel CM. Redo aortic root replacement: experience with 31 patients. Ann Thorac Surg 2001;71:1460 3. 18. Vogt PR, Brunner-LaRocca HP, Sidler P, et al. Reoperative surgery for degenerated aortic bioprostheses: predictors for emergency surgery and reoperative mortality. Eur J Cardiothorac Surg 2000;17:134 9. 19. McGiffin DC, Galbraith AJ, O Brien MF, et al. An analysis of valve re-replacement after aortic valve replacement with biologic devices. J Thorac Cardiovasc Surg 1997;113:311 8. 20. Odell JA, Mullany CJ, Schaff HV, Orszulak TA, Daly RC, Morris JJ. Aortic valve replacement after previous coronary artery bypass grafting. Ann Thorac Surg 1996;62:1424 30. 21. Rankin JS, Hammill BG, Ferguson TB, et al. Determinants of operative mortality in valvular heart surgery. J Thorac Cardiovasc Surg 2006;131:547 57. 22. Tasca G, Mhagna Z, Perotti S, et al. Impact of prosthesispatient mismatch on cardiac events and midterm mortality after aortic valve replacement in patients with pure aortic stenosis. Circulation 2006;113:570 6.