Mechanism of and Risk Factors for Reoperation After Mitral Valve Repair for Degenerative Mitral Regurgitation

Transcription

1 Circulation Journal Official Journal of the Japanese Circulation Society Advance Publication by-j-stage Mechanism of and Risk Factors for Reoperation After Mitral Valve Repair for Degenerative Mitral Regurgitation Takashi Murashita, MD; Yukikatsu Okada, MD, PhD; Hiroshi Fujiwara, MD, PhD; Hideo Kanemitsu, MD, PhD; Naoto Fukunaga, MD; Yasunobu Konishi, MD; Ken Nakamura, MD; Yoshito Sakon, MD; Tadaaki Koyama, MD, PhD Background: We reviewed our 20-year experience of mitral valve (MV) repair for degenerative mitral regurgitation (MR) and analyzed the mechanisms and risk factors of reoperation. Methods and Results: Six hundred and fifty-four patients who underwent MV repair for degenerative MR between 1991 and 2010 were retrospectively reviewed. The mean follow-up duration was 7.5±4.9 years. Late echocardiography was obtained at a fixed schedule. Standard procedures for MV repair were resection and suture for posterior prolapse, artificial chordal reconstruction for anterior prolapse and a combination of them for prolapse of both. Ring annuloplasty was performed in most cases. We encountered 2 early and 23 late reoperations. Fifteen (60%) were valve-related failure, 9 (36%) were procedure-related failure and 1 was unknown. Valve-related failure was characterized with recurrence of MR due to new prolapse region and progression of mitral stenosis due to leaflet thickening. For them, reoperation was performed at 7.6 years and 14.3 years after the initial operation, respectively. Multivariate analysis identified preoperative left ventricular diastolic diameter and residual regurgitation at discharge as risk factors for reoperation. Conclusions: Valve-related failure occurred late due to slow progression of degenerative disease. Long-term followup after surgery is mandatory. Given that progression and severity of MR were identified as risk factors for reoperation, early surgical intervention is desirable for degenerative MR. Key Words: Mitral valve; Surgery For degenerative mitral regurgitation (MR), mitral valve (MV) repair has been the standard operation and excellent long-term outcomes have been reported. 1 3 Some previous studies reported that MV repair is superior to replacement in terms of long-term survival, cardiac function and incidence of valve-related adverse events. 4 6 Reoperation after MV repair, however, has been reported to occur at a linear rate of up to 3.0% per year. 7,8 Not many reports exist on the detailed mechanisms of reoperation after MV repair David et al reported that anterior leaflet (AL) prolapse was the only independent predictor of reoperation, and anterior and bi-leaflet prolapse, age and ejection fraction <40% were independent predictors of recurrent MR after MV repair. 12 Shimokawa et al reported that AL prolapse, preoperative atrial fibrillation and non-use of annuloplasty ring were independent predictors of recurrent MR. 11 They reported that the main mechanism of recurrent MR was progressive degeneration such as leaflet thickening and prolapse. We also reported long-term outcomes of MV repair for bi-leaflet prolapse. In that study, involvement of AL did not become a risk factor for reoperation or recurrent MR. 13 The aim of this study was to investigate the long-term outcome of MV repair for degenerative MR in a single institution and analyze the mechanisms and risk factors of reoperation after MV repair. Methods The data analysis for this retrospective study was approved by the Institutional Review Board of Kobe City Medical Center General Hospital and the Board waived the need for patient consent. Patients From January 1991 to December 2010, 1,138 patients underwent MV surgery (MV repair, n=949; MV replacement, n=189) Received January 22, 2013; revised manuscript received March 25, 2013; accepted April 18, 2013; released online May 29, 2013 Time for primary review: 13 days Department of Cardiovascular Surgery, Kobe City Medical Center General Hospital, Kobe, Japan Mailing address: Takashi Murashita, MD, Department of Cardiovascular Surgery, Kobe City Medical Center General Hospital, Minatozimaminami, Chuo-ku, Kobe , Japan. tmurashita@kcho.jp ISSN doi: /circj.CJ All rights are reserved to the Japanese Circulation Society. For permissions, please cj@j-circ.or.jp

2 Advance Publication by-j-stage MURASHITA T et al. Table 1. Preoperative Clinical and Echocardiographic Characteristics All (n=654) AL prolapse (n=308) PL prolapse (n=346) P-value Mean age (years) 56.4± ± ±12.0 <0.001 Male 385 (58.9) 190 (61.7) 195 (56.4) BSA (m 2 ) 1.60± ± ± Hypertension 251 (38.8) 81 (26.3) 170 (49.1) <0.001 Hyperlipidemia 39 (6.0) 7 (2.3) 32 (9.2) <0.001 Diabetes mellitus 27 (4.1) 12 (3.9) 15 (4.3) COPD 3 (0.5) 2 (0.6) 1 (0.3) Coronary artery disease 32 (4.9) 11 (3.6) 21 (6.1) Hemodialysis 2 (0.3) 2 (0.6) History of cardiac operation 8 (1.2) 4 (1.3) 4 (1.2) Urgent/emergency operation 5 (0.8) 0 5 (1.4) NYHA functional class I II 483 (73.9) 226 (73.4) 257 (74.3) III IV 171 (26.1) 82 (26.6) 89 (25.7) Prolapse region Anterior 76 (11.6) 76 0 Posterior 346 (52.9) Both 232 (33.5) Persistent atrial fibrillation 155 (23.7) 83 (26.9) 72 (20.8) LVDd (mm) 55.4± ± ± LVDs (mm) 33.9± ± ±6.7 <0.001 LVEF (%) 66.5± ± ±7.6 <0.001 LAD (mm) 46.6± ± ± Systolic PAP (mmhg) 43.6± ± ± Regurgitant volume (ml) 75.0± ± ± Regurgitant fraction (%) 46.9± ± ± Effective regurgitant orifice, cm ± ± ± TR grade Mild or less 479 (73.2) 226 (73.4) 253 (73.1) Moderate-severe 175 (26.8) 82 (26.6) 93 (26.9) Data given as n (%) or mean ± SD. AL, anterior leaflet; BSA, body surface area; COPD, chronic obstructive pulmonary disease; LAD, left atrial diameter; LVDd, left ventricular diastolic diameter; LVDs, left ventricular systolic diameter; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; PAP, pulmonary artery pressure; PL, posterior leaflet; TR, tricuspid regurgitation. Results Patient Characteristics Patient characteristics and preoperative echocardiographic data are listed in Table 1. The region of prolapse was anterior only in 76 patients (11.6%), posterior only in 346 (52.9%) and both leaflets in 232 patients (35.5%). Patients who had anterior prolapse and both leaflet prolapse were combined into the AL prolapse group and compared with the posterior leaflet (PL) prolapse group. Of note, patients in the PL group were older, and had a higher rate of hypertension and hyperlipidemia. In terms of cardiac function, patients in the AL group had larger left ventricular systolic diameter, and lower left ventricular ejection fraction. Patients in the PL group had higher systolic pulmonary artery pressure. The preoperative symptoms, inciat Kobe City Medical Center General Hospital. Among patients who underwent MV repair, 96 patients had Carpentier s type I disease, 690 patients had type II disease and 163 patients had type III disease. Among type II patients, we excluded 36 patients who underwent aortic valve procedures. As a result, 654 patients who underwent MV repair for type II disease were enrolled in this study. Follow-up Examinations and Management We followed up the patients at the outpatient clinic or via telephone survey, and the follow-up was completed in 647 patients (98.9%). The mean follow-up duration was 7.5±4.9 years. Postoperative echocardiographic follow-up was generally performed before discharge and at the outpatient clinic at 1, 5, 10, and 15 years after operation. The mean length of echocardiographic follow-up was 5.5±4.8 years. Follow-up echocardiographic data were obtained in 558 patients at 1 year, in 360 patients at 5 years, in 202 patients at 10 years and in 65 patients at 15 years after operation. Statistical Analysis The continuous data in this study are expressed as mean ± SD and range. Categorical variables were compared with the chi- square or Fisher s exact tests, and continuous variables were compared with unpaired t or Wilcoxon tests. Survival and freedom from events were calculated with the Kaplan-Meier method. Univariate and multivariate Cox hazard regression analysis were used to identify predictors for reoperation. Statistical analysis was performed with StatView (SAS Institute, Cary, NC, USA).

3 Reoperation After MV Repair Advance Publication by-j-stage dence of persistent atrial fibrillation and grade of tricuspid regurgitation were similar between the 2 groups. Surgical Techniques We approached through a median sternotomy in most of the patients. Standard cardiopulmonary bypass techniques were used, including bicaval cannulation. Myocardial protection was achieved with antegrade and retrograde intermittent cold blood cardioplegia. Left atrial incision was applied in most patients. The techniques of MV repair have been described by Carpentier 14 and by David et al. 15 Prolapse of the PL was mostly corrected by resection and suture of mitral leaflets. Prolapse of the AL was mostly corrected by chordal replacement with polytetrafluoroethylene (Gore-Tex; WL Gore and Associates, Flagstaff, AZ, USA) sutures. For mitral annuloplasty, we used a flexible Duran Annuloplasty ring (Medtronic, Minneapolis, MN, USA) in 513 patients (78.4%), a Carpentier-Edwards Classic annuloplasty ring (Edwards Lifesciences, Irvine, CA, USA) in 26 patients (4.0%), a Carpentier-Edwards Physio II Annuloplasty ring (Edwards Lifesciences) in 25 patients (3.8%) and other rings in 14 patients (2.1%). During the operation, we routinely performed transesophageal echocardiography to determine whether a second pump run and intraoperative re-repair were necessary. Concomitant procedures were performed in 243 patients (37.2%). The majority of operations were performed by 1 surgeon (Y.O.) The details of surgical procedures are listed in Table 2. Early Outcomes There were seven 30-day deaths (1.1%). The causes of death were low cardiac output in 3, left ventricular rupture in 2, stroke in 1 and bowel ischemia in 1. We encountered 2 early reoperations during admission after first surgery. The causes of these were ring dehiscence in 1 and severe hemolysis in 1. We performed re-repair successfully in both cases. Excluding operative Table 2. Details of Surgical Procedure Operative procedure for mitral valve Leaflet resection and suture 548 (83.8) Artificial chordal reconstruction 296 (45.3) Folding plasty 59 (9.0) Auto-pericardial patch 30 (4.6) Ring annuloplasty 578 (88.3) Concomitant procedures Tricuspid annuloplasty 162 (24.8) Maze procedure 87 (13.3) Coronary artery bypass grafting 35 (5.4) Closure of ASD, PFO 34 (5.2) Cardiopulmonary bypass time (min) 147±48 Aortic cross-clamp time (min) 104±36 Second-pump run 47 (7.2) Data given as n (%) or mean ± SD. ASD, atrial septal defect; PFO, patent foramen ovale. mortality and early reoperation cases, MR grade at discharge was trivial in 551 (85.4%), mild in 81 (12.6%) and moderate in 13 (2.0%). Reoperation We encountered 25 reoperations in all. Among them, 2 were performed during the first admission. The other 23 were performed during follow-up. Among 25 reoperations, 6 were for AL prolapse, 10 were for PL prolapse and 9 were for prolapse of both leaflets. There was no significant difference in rate of freedom from reoperation in terms of region of prolapse (logrank P=0.297); the 5-year freedom from reoperation rate was 96.0±2.3% for AL, 98.7±0.6% for PL and 98.1±1.0% for both; the 10-year freedom from reoperation rate was 90.6±4.3% for AL, 96.1±1.5% for PL and 96.1±1.6% for both; and the 15- Figure 1. Freedom from reoperation. Anterior, anterior prolapse; both, prolapse of both leaflets; posterior, posterior prolapse.

4 Advance Publication by-j-stage MURASHITA T et al. Figure 2. Freedom from recurrence of severe mitral regurgitation. Anterior, anterior prolapse; both, prolapse of both leaflets; posterior, posterior prolapse. Table 3. Details of Reoperation n Duration between first operation and second operation (median, range) Procedures for reoperation Valve related 15 New prolapse years (6.0 years 16.1 years) ReMVP 5 Leaflet thickening years (5.6 years 17.8 years) ReMVP 1; MVR 3 Endocarditis years (2 months 11.3 years) MVR 4 New onset of AF years (6.5 years 16.2 years) ReMVP 2 Procedure related 9 Hemolysis years (9 days 7.8 years) ReMVP 3; MVR 3 Ring dehiscence 2 2 weeks, 3 months ReMVP 2 Tear of suture site 1 1 month ReMVP Unknown years MVR AF, atrial fibrillation; MVR, mitral valve replacement; ReMVP, re-mitral valve repair. year freedom from reoperation rate was 85.3±6.6% for AL, 96.1±1.5% for PL and 92.9±3.5% for both (Figure 1). There was no significant difference in rate of freedom from recurrence of severe MR in terms of the region of prolapse (logrank P=0.143; Figure 2). Details of Reoperation We divided the causes of reoperation into valve related and procedure related (Table 3). The leading cause of reoperation was valve related, which was found in 15 cases. The detailed causes were as follows: recurrent severe MR due to new prolapse region in 5, progression of mitral stenosis in 4, endocarditis in 4 and new onset of atrial fibrillation with moderate MR in 2. The second cause of reoperation was procedure related, which was found in 9 cases. The detailed causes were as follows: severe hemolysis in 6, dehiscence of a ring in 2 and tear at the suture site in 1. In 1 case, redo MV replacement was performed at another hospital for unknown cause. In terms of the procedure for reoperation, 14 re-mv repairs and 11 MV replacements were performed. There were 2 inhospital deaths at the time of redo surgery (8.0%). The cause of in-hospital death was low output syndrome in 1 and mediastinitis in 1. Risk Factors for Reoperation Univariate analysis was performed to determine the risk factors for reoperation (Table 4). Moderate MR at discharge, preoperative left ventricular diastolic diameter, regurgitant volume and use of prosthetic ring were identified as risk factors for reoperation. On multivariate analysis, moderate MR at discharge (hazard ratio [HR], 3.246; 95% confidence interval [CI]: , P=0.012) and left ventricular diastolic diameter (HR,

5 Reoperation After MV Repair Advance Publication by-j-stage Table 4. Risk Factors for Reoperation HR 95% CI P-value Univariate Age Hypertension Posterior prolapse Preoperative AF Postoperative AF Moderate MR at discharge LVDd LVDs LVEF LAD Systolic PAP Regurgitant volume TR grade Use of ring Second pump-run Multivariate Moderate MR at discharge LVDd Regurgitant volume Use of ring CI, confidence interval; HR, hazard ratio; MR, mitral regurgitation; TR, tricuspid regurgitation. Other abbreviations as in Tables 1, ; 95% CI: , P=0.003) were identified as risk factors for reoperation. Discussion Many reports have concluded that MV repair has a favorable outcome in degenerative MR and that repair is superior to replacement in terms of long-term survival, cardiac function and incidence of valve-related adverse events. 4 6,16 MV repair, however, is sometimes associated with failure of repair, which leads to MV replacement at the time of initial operation or redo operation for MV at a later date. Because redo operation is accompanied by higher operative risk than the initial operation, the superiority of MV repair should be shown by a lower or equal failure rate than replacement with prosthetic valve. To assess the durability of valve repair, late echocardiography is mandatory. We basically followed all the patients at the outpatient clinic and routinely obtained late echocardiography at a fixed schedule. That led to a relatively high follow-up rate and large number of late echocardiographic data. At Kobe City Medical Center General Hospital we aggressively attempted MV repair for degenerative MR, and the repair rate exceeds 99%. In this study, we reviewed our more than 20 years experience of MV repair, which was performed mostly by 1 operator. The standard procedure for degenerative MR is resection and suture for PL, artificial chordal replacement for AL and annuloplasty rings in most cases. The present study showed that the rate of freedom from reoperation of the overall cohort was 98.2±0.6% at 5 years and 95.5±1.1% at 10 years after the operation, which was better than or equivalent to previous reports. The rate of freedom from recurrent severe MR of the overall cohort was 98.3±0.5% at 5 years and 95.2±1.1% at 10 years after the operation, which also was satisfactory. We suspect that optimal techniques for individual cases and use of intraoperative transesophageal echocardiography reduced the incidence of incomplete repair. During our initial stage of learning MV repair before 1991, we had a couple of experiences of mild residual regurgitation increasing and requiring reoperation within the first postoperative year. Thus we applied a strict cutoff for residual regurgitant signal of <2 cm 2. Moreover, to avoid intravascular hemolysis after repair, regurgitant jet signals impinging on the ring are not considered acceptable. A small percentage of patients (2.0%), however, had residual moderate MR at discharge, which led to a risk factor for late reoperation. Jamieson et al reported freedom from structural valve deterioration after MV replacement with bioprosthesis at 56.0±4.1% at 18 years for the year age group. 17 Aoyagi et al reported long-term durability of mechanical prosthetic valve, although it requires lifelong anticoagulation. 18 The present outcomes showed that MV repair was superior to replacement in terms of durability and avoidance of anticoagulation. Not many reports have described detailed mechanisms of failure after MV repair. Dumont et al reported that valve-related failure occurred at a median of 5.4 years after operation and was caused by progressive disease in >90% and endocarditis in 10%. Procedure-related failure occurred at a median of 19 days after operation and was caused by suture dehiscence in >40%, rupture of shortened chordae in approximately 20%, systolic anterior motion in approximately 20% and hemolysis in approximately 20%. 9 Shimokawa et al reported that valve-related failure occurred at a median of 5.7 years after operation and that the main mechanisms of recurrent regurgitation were leaflet thickening in approximately 50%, leaflet prolapse in 30% and dehiscence in approximately 15%. 11 The present outcomes were similar to these reports. Valve-related failure comprised 60% of reoperations and the causes included new prolapse, leaflet thickening leading to mitral stenosis and endocarditis. Procedure-related failure comprised 36% of reoperations, and the causes included hemolysis, ring dehiscence and tear of suture site. Relatively early surgical intervention was required

6 Advance Publication by-j-stage MURASHITA T et al. for procedure-related failure. In contrast, redo surgery for valve-related failure was carried out at a later date. Reoperations for recurrent severe regurgitation due to new prolapse region and leaflet thickening were performed at a median of 7.6 years and 14.3 years after the initial operation, respectively. Shimokawa et al reported that leaflet thickening was especially noted in patients with AL prolapse, 11 but in the present study, leaflet thickening was seen in both ALs and PLs. In 4 out of 5 cases, a new prolapse region emerged on the same side of the leaflet that was previously repaired. As previous studies implied, the present data also show that degenerative disease might slowly progress even after successful MV repair. We believe very long-term follow-up (>10 years) is mandatory after mitral surgery. Regarding the procedure for reoperation, Dumont et al reported that MV replacement was performed in 64% and rerepair was performed in 36%. 9 Zegdi et al reported the outcomes of redo operation for failure of MV repair, and noted that there were no perioperative deaths after re-repair, whereas perioperative death occurred in approximately 10% of patients after redo replacement. 10 In the present study, we could re-repair for new prolapse region, but it was necessary to replace the MV in most cases of leaflet thickening and in all cases of endocarditis. The in-hospital mortality of redo operation was 8.0%, which was noted only in cases of MV replacement. As the previous study indicated, we believe that we should attempt re-repair at the time of reoperation rather than replacement. David et al reported that AL prolapse was the only independent predictor of reoperation, and anterior and bi-leaflet prolapse, age and ejection fraction <40% were independent predictors of recurrent MR after MV repair. 12 Shimokawa et al reported that AL prolapse, preoperative atrial fibrillation and non-use of annuloplasty ring were independent predictors of recurrent MR. 11 The present findings, however, differ from the previous studies. Univariate analysis did not identify age, AL prolapse, left ventricular ejection fraction, preoperative atrial fibrillation and non-use of annuloplasty ring as risk factors for reoperation. In contrast, preoperative left ventricular diastolic diameter and regurgitant volume were the identified risk factors. The reason why left ventricular dilatation and severity of regurgitation influenced the incidence of reoperation is unclear. We suspect the following mechanism. Chronic regurgitant jet produces adverse stimulation on MV leaflets. As MR becomes more severe and mitral leaflets undergo longer exposure to regurgitant jet, the mitral leaflets degenerate further. These adverse changes in mitral leaflets may influence the progression of degenerative disease in the late period, leading to high reoperation rates. Some limitations exist in this study. First, this was a singleinstitution, retrospective study. Second, the patient number was low. Third, clinical and echocardiographic follow-up were not obtained in all patients and the follow-up period was not sufficiently long. Moreover, echocardiographic follow-up was performed on a fixed schedule. The real durability of the procedure remains unknown. Conclusion The main mechanism of reoperation after MV repair was valve-related failure, which was characterized by new prolapse region and leaflet thickening. Because valve-related failure occurred late, due to the slow progression of degenerative disease, long-term follow-up after surgery is mandatory. The risk factors for reoperation were identified as progression and severity of MR, indicating that early surgical intervention is desirable for this type of patient. References 1. Mohty D, Orszulak TA, Schaff HV, Avierinos JF, Tajik JA, Enriquez- Sarano M. Very long-term survival and durability of mitral valve repair for mitral valve prolapse. Circulation 2001; 104(12 Suppl 1): I1 I7. 2. Braunberger E, Deloche A, Berrebi A, Abdallah F, Celestin JA, Carpentier A, et al. Very long-term results (more than 20 years) of valve repair with Carpentier s techniques in nonrheumatic mitral valve insufficiency. Circulation 2001; 104(12 Suppl 1): I8 I Daneshmand MA, Milano CA, Rankin JS, Honeycutt EF, Swaminathan M, Glower DD, et al. Mitral valve repair for degenerative disease: A 20-year experience. Ann Thorac Surg 2009; 88: De Bonis M, Lorusso R, Lapenna E, Kassem S, De Cicco G, Alfieri O, et al. Similar long-term results of mitral valve repair for anterior compared with posterior leaflet prolapse. J Thorac Cardiovasc Surg 2006; 131: Gillinov AM, Blackstone EH, Nowicki ER, Griffin B, Sabik JF 3rd, Svensson LG, et al. Valve repair versus valve replacement for degenerative mitral valve disease. J Thorac Cardiovasc Surg 2008; 135: Jokinen JJ, Hippeläinen MJ, Pitkänen OA, Hartikainen JE. Mitral valve replacement versus repair: Propensity-adjusted survival and qualityof-life analysis. Ann Thorac Surg 2007; 84: Suri RM, Schaff HV, Dearani JA, Sundt TM 3rd, Enriquez-Sarano M, Orszulak TA, et al. Survival advantage and improved durability of mitral repair for leaflet prolapse subsets in the current era. Ann Thorac Surg 2006; 82: Flameng W, Meuris B, Herijgers P, Herregods MC. Durability of mitral valve repair in Barlow disease versus fibroelastic deficiency. J Thorac Cardiovasc Surg 2008; 135: Dumont E, Gillinov AM, Blackstone EH, Sabik JF 3rd, Svensson LG, Mihaljevic T, et al. Reoperation after mitral valve repair for degenerative disease. Ann Thorac Surg 2007; 84: Zegdi R, Sleilaty G, Latrémouille C, Berrebi A, Carpentier A, Deloche A, et al. Reoperation for failure of mitral valve repair in degenerative disease: A single-center experience. Ann Thorac Surg 2008; 86: Shimokawa T, Kasegawa H, Katayama Y, Matsuyama S, Manabe S, Takanashi S, et al. Mechanisms of recurrent regurgitation after valve repair for prolapsed mitral valve disease. Ann Thorac Surg 2011; 91: David TE, Ivanov J, Armstrong S, Christie D, Rakowski H. A comparison of outcomes of mitral valve repair for degenerative disease with posterior, anterior, and bileaflet prolapse. J Thorac Cardiovasc Surg 2005; 130: Okada Y, Nasu M, Koyama T, Shomura Y, Yuzaki M, Murashita T, et al. Outcomes of mitral valve repair for bileaflet prolapse. J Thorac Cardiovasc Surg 2012; 143(4 Suppl): S21 S Carpentier A. Cardiac valve surgery: The French correction. J Thorac Cardiovasc Surg 1983; 86: David TE, Bos J, Rakowski H. Mitral valve repair by replacement of chordae tendineae with polytetrafluoroethylene sutures. J Thorac Cardiovasc Surg 1991; 101: Itagaki S, Adams DH, Anyanwu AC. Triggers for surgical referral in degenerative mitral valve regurgitation. Circ J 2012; 77: Jamieson WR, Gudas VM, Burr LH, Janusz MT, Fradet GJ, Ling H, et al. Mitral valve disease: If the mitral valve is not reparable/failed repair, is bioprosthesis suitable for replacement? Eur J Cardiothorac Surg 2009; 35: Aoyagi S, Tayama K, Okazaki T, Shintani Y, Kono M, Tanaka H, et al. Structural valve deterioration in a starr-edwards mitral caged-disk valve prosthesis. Circ J 2012; 77: