Mitral valve repair is the treatment of choice for most. Hemolysis After Mitral Valve Repair: Mechanisms and Treatment

Hemolysis After Mitral Valve Repair: Mechanisms and Treatment Buu-Khanh Lam, MD, Delos M. Cosgrove III, MD, Sunil K. Bhudia, MD, and A. Marc Gillinov, MD Department of Thoracic and Cardiovascular Surgery, Cleveland Clinic Foundation, Cleveland, Ohio Background. The objectives of this study were to determine the mechanisms of hemolysis after mitral valve repair and to determine outcomes after surgical treatment (mitral replacement or re-repair). Methods. Between 1981 and 2002, 32 patients (mean age, 58 years) presented with hemolytic anemia after mitral valve repair for degenerative, rheumatic, or ischemic mitral regurgitation (MR). Three types of annuloplasty were used at the initial mitral valve repair: Cosgrove-Edwards, Carpentier-Edwards, and bovine pericardial (Perigard). The diagnosis and the mechanisms of hemolysis were investigated with laboratory testing and echocardiography. Results. Median interval from initial mitral valve surgery to diagnosis of hemolysis was 3 months (range, 1 week to 4 years). At presentation, mean hematocrit was 27.5% 4.9% and 22 patients (69%) required transfusion. Echocardiographic findings varied. Twenty-four patients (77%) had grade 3 or 4 MR. Mitral regurgitant jet types included fragmentation (11 patients, 34%), acceleration (10, 31%), slow deceleration (5, 16%), collision (4, 13%), and free jet (2, 6%). Mitral valve replacement was performed in 28 patients, mitral valve re-repair in 3, and 1 patient did not undergo reoperation. At reoperation the mitral valve repair was physically intact in 25 of 31 patients (81%). There were 2 hospital deaths in patients having reoperation (6%). Actuarial survival was 95% at 1 year and 85% at 5 years. In 1 patient recurrent mechanical hemolysis developed caused by a perivalvular leak after mitral valve replacement. Conclusions. Hemolysis is a mode of failure of mitral valve repair. Patients with hemolysis generally present within 3 months of mitral valve repair. Although echocardiographic features varied, most patients had highgrade MR and regurgitant jets that fragmented or accelerated. Mitral valve replacement yields favorable outcomes for patients with hemolysis after mitral valve repair. (Ann Thorac Surg 2004;77:191 5) 2004 by The Society of Thoracic Surgeons Mitral valve repair is the treatment of choice for most patients with mitral regurgitation (MR). Hemolytic anemia is a rare complication after mitral valve repair [1 10]. In patients with previous mitral valve replacement, echocardiographic characterization of mechanisms of hemolysis has helped to establish a clinical causal link [10, 11]. However, in patients having mitral valve repair the mechanisms of hemolysis have not been well characterized. The objectives of this study were to determine mechanisms of hemolysis after mitral valve repair and to determine outcomes after surgical treatment (mitral replacement or re-repair). Material and Methods Between 1981 and 2002, 32 patients presented with hemolytic anemia after mitral valve repair. After other possible causes of hemolysis were ruled out, the persistent anemia in these patients was attributed to traumatic recurrent or residual MR. Accepted for publication July 25, 2003. Address reprint requests to Dr Gillinov, Department of Thoracic and Cardiovascular Surgery, The Cleveland Clinic Foundation, 9500 Euclid Ave, Desk F25, Cleveland, OH 44195; e-mail: gillinom@ccf.org. Hemolytic Anemia Hemolysis was clinically diagnosed by the constellation of persistent severe anemia (hemoglobin 10 g/dl, hematocrit 33%), elevated lactate dehydrogenase ( 440 U/L), reduced serum haptoglobin ( 37 mg/dl), and the presence of schistocytes, fragmented cells, and polychromasia on peripheral blood smear. Etiology of MR and Initial Mitral Valve Repair Standard criteria were used to classify the etiology of mitral valve dysfunction. Mitral regurgitationa was classified as degenerative if intraoperative findings included myxomatous changes of the leaflets, leaflet prolapse/flail, annular dilatation, and/or ruptured chordae. Rheumatic MR was characterized by thickening of the valvular/ subvalvular apparatus, restriction of valve motion, and/or commissural fusion. Ischemic mitral hregurgitation was deemed present in patients with a previous myocardial infarction, normal appearing leaflets and chords, restricted leaflet motion, and annular dilation. At the time of original mitral valve repair 10 patients (31%) had degenerative disease, 9 (28%) had ischemic disease, 8 (25%) had rheumatic disease, 3 (10%) had endocarditis, 1 (3%) had hypertrophic obstructive cardiomyopathy, and 1 (3%) had a congenital malformation. 2004 by The Society of Thoracic Surgeons 0003-4975/04/$30.00 Published by Elsevier Inc doi:10.1016/s0003-4975(03)01455-3

192 LAM ET AL Ann Thorac Surg HEMOLYSIS AFTER MITRAL VALVE REPAIR 2004;77:191 5 Table 1. Repair Techniques No. of 32 Annuloplasty Cosgrove-Edwards 22 69 Bovine pericardial (Periguard) 7 22 Carpentier-Edwards 2 6 Chordal transfer 8 26 Leaflet debridement 7 22 Cleft closure 3 9 Commissural plication 3 9 Chordal shortening 2 6 Placement of Goretex chordae 1 3 Repair techniques are summarized in Table 1. One patient did not receive an annuloplasty ring at the time of initial mitral valve repair. Additional procedures included coronary revascularization in 11 patients (34%), aortic valve replacement in 4 (13%), tricuspid valve repair in 3 (9%), and aortic valve repair in 2 (6%). Echocardiographic Evaluation of Hemolysis and Mitral Valve Function In patients with hemolysis after mitral valve repair, preoperative transthoracic and intraoperative transesophageal echocardiography were performed to determine MR etiology and grade, jet direction, and mechanisms of hemolysis. Garcia and associates [11] previously characterized regurgitant jets causing hemolysis by studying their hydrodynamic properties and this classification system was employed. The following patterns were described: (1) fragmentation: the regurgitant jet is divided by a solid structure such as a suture, ruptured chord, or dehisced annuloplasty ring; (2) collision: there is sudden deceleration of the regurgitant jet due to direct impact on a solid structure such as an annuloplasty ring or pledget, which sharply alters the trajectory of the jet; (3) rapid acceleration: a regurgitant jet originates from a small orifice ( 2 mm in diameter) such as a leaflet perforation or narrow region of para-ring dehiscence with no direct impact on a solid structure; (4) free jet: the regurgitant jet (generally with a central trajectory) originates from a wide orifice ( 2 mm in diameter) and is not constrained by a solid structure until it extends to the dome of the left atrium; and (5) slow deceleration: an eccentric regurgitant jet originates from a large eccentric orifice and adheres to the left atrial wall from its point of origin. As per our institutional protocol, all patients who had a mitral valve repair underwent echocardiographic evaluation of the repair before discharge. In this study, the echocardiographic information was gathered retrospectively from the Department of Cardiology echocardiography database. Follow-Up Patients were followed up systematically at 2-year intervals with a mailed questionnaire, a telephone interview, or examination at the Cleveland Clinic. This information was entered into the Cardiovascular Information Registry at the Cleveland Clinic Foundation; the Institutional Review Board of the Cleveland Clinic Foundation has approved use of this Registry for clinical research. Follow-up for survival was supplemented by use of the Social Security Index search system. For patients discharged from hospital, follow-up was 100% complete; mean follow-up interval was 6 4.3 years (range, 0.4 to 17). Data Analysis Simple descriptive statistics were used to summarize the data. Continuous variables were presented as mean standard deviation. Statistical transformations of continuous variables were performed when necessary to facilitate data analysis. Categorical data were described using frequencies and percentages. Dichotomous and ordinal variables were cross-tabulated and exploratory analysis was performed by 2 methods. The small sample size and paucity of events precluded identification of early or late predictors of mortality by multivariable regression analyses. Nonparametric survival analysis was performed using the product-limit method of Kaplan and Meier. All analyses were performed using the SAS statistical software (SAS version 8.2; SAS, Cary, NC). Results The median interval from mitral valve repair to presentation with hemolytic anemia was 3 months (range, 0.17 to 54) whereas that from diagnosis of hemolysis to reoperation was 2.5 months (range, 0.07 to 20). At presentation with hemolytic anemia, the patients mean age was 59 15 years. Comorbidities are listed in Table 2. For the 31 patients undergoing reoperation, symptoms included fatigue in 30 patients (97%), dyspnea (30 patients, 97%), shortness of breath on exertion (31, 100%), syncope (2, 7%), and palpitations (12, 39%). Hemolytic Anemia The mean hemoglobin and hematocrit at time of discharge were 9.9 1.4 g/dl and 29.7% 3.7% respectively. On readmission, the mean hemoglobin and hematocrit were 8.9 1.3 g/dl and 27.5% 4.9%. Mean lactate dehydrogenase (LDH) was 1,834 1,505 U/L, median haptoglobin was 6 mg/dl, and mean reticulocyte percentage was 4.3% 3.9%. Peripheral blood smear and urine analysis showed changes consistent with hemolysis (Table 3). Twenty-two patients (69%) were transfused an average of 5.4 3.3 U of blood as outpatients; in addition, 20 (63%) were transfused an average of 2.6 1.2 U of blood in the hospital before their reoperation. Mechanisms of Hemolysis Although echocardiographic findings varied, all patients had recurrent or residual MR; regurgitation was 3 or 4 in 77% but only 1 or 2 in 23%. Mitral regurgitant jets were examined with echocardiography and classified with regard to hydrodynamic properties and direction.

Ann Thorac Surg LAM ET AL 2004;77:191 5 HEMOLYSIS AFTER MITRAL VALVE REPAIR Table 2. Patient Comorbidities No. of 32 Cardiac NYHA I 3 10 II 17 53 III 10 31 IV 2 6 MR severity 2 7 23 3 6 19 4 18 58 Previous MI 13 41 Hypertension 22 69 Atrial fibrillation 6 19 Noncardiac Smoker 19 59 Peripheral vascular disease 9 28 Diabetes 3 9 COPD 6 19 Renal disease 2 6 COPD chronic obstructive pulmonary disease; MI myocardial infarction; MR mitral regurgitation; NYHA New York Heart Association. The mechanisms responsible for hemolysis were fragmentation of the mitral regurgitant jet in 11 patients (34%), rapid acceleration in 10 (31%), slow deceleration in 5 (16%), collision in 4 (13%), and free jet in 2 (6%). The jet was posterior in 17 patients (55%), central in 10 (33%), anterior in 2 (6%), and complex in 2 (6%). Mitral regurgitation after mitral valve repair were attributed to progression of rheumatic heart disease in 8 patients (25%), recurrent ischemic MR in 6 (19%), mechanical trauma (ring dehiscence, leaflet perforation, ruptured neochordae) in 6 (19%), incomplete repair of degenerative disease in 5 (16%), de novo or recurrent bacterial endocarditis in 5 (16%), and indeterminate in 2 (6%). At reoperation the mitral valve repair was physically intact in 25 of 31 patients (81%). Intactness was defined as the absence of ring dehiscence (partial or complete), suture-related tear of valve tissue, and suture or chord break. Findings in patients with a disrupted repair included new mitral valve leaflet perforations and partial Table 3. Laboratory Testing for Hemolysis Peripheral blood smear Schistocytes 85 Fragmented cells 94 Polychromasia 100 Urine analysis Hemosiduria 100 Hemoglobinuria 91 Urobilinogen 33 dehiscence of the mitral ring in 2, isolated annuloplasty dehiscence in 1, rupture of a mitral edge-to-edge repair in 1, ruptured Goretex neochordae in 1, and new posterior leaflet perforation in 1. Treatment of Hemolysis Thirty-one of 32 patients had mitral reoperation. In 28 patients (90%) the hemolysis was treated by mitral valve replacement. Twelve received a bioprosthesis and 16 received a mechanical valve. Three (9%) underwent rerepair of the mitral valve. The first re-repair patient retained his initial annuloplasty (Cosgrove band) but underwent closure of a leaflet perforation, commissure closure, and papillary muscle plication The second patient s initial annuloplasty (Cosgrove band) had partially dehisced and was replaced with a one size smaller Cosgrove band and a leaflet perforation was closed. The third patient had a bovine pericardial (Perigard) annuloplasty band that had partially dehisced; it was replaced with a new Perigard band. The single patient who did not undergo reoperation had a low level of hemolysis, which was successfully managed medically. Postoperative complications included cerebrovascular accident in 2 patients (6%), respiratory failure (2 patients, 6%), and bleeding (1 patient, 3%). Postoperative transfusions were necessary in 19 patients (61%) with a mean of 3.5 2.3 U of blood. The mean hemoglobin and hematocrit on discharge were 9.6 0.9 g/dl and 29.6% 2.6% respectively. Mean length of stay from surgery to discharge was 8.4 4.5 days. Late Outcomes There were 2 (6%) hospital deaths, both related to a perioperative stroke. There were 6 late deaths (19%); 3 patients had myocardial infarctions, 2 had noncardiac causes of death, and 1 died of malignant tachyarrhythmia. Two patients had recurrent hemolysis. One patient had a paravalvular leak from a mechanical valve requiring a third operation to repair the leak. The second patient s hemolysis was caused by an autoimmune reaction to procainamide, and that resolved with medical treatment. The single patient who did not receive reoperation has continued hemolysis. For patients still alive at a mean follow-up of 6 4.3 years, 15 patients (63%) were in New York Heart Association class I, 7 (29%) were in class II, and 2 (8%) were in class III. This intermediate length of follow-up is driven by the fact that 19 of 32 patients (59%) were operated on after 1995. The actuarial survival was 95% at 1 year and 85% at 5 years. Follow-up transthoracic echocardiograms showed that 16 patients (67%) had no MR, 7 (29%) had 1 MR, and 1 (4%) had 2 MR. Comment 193 Hemolysis is a mode of early repair failure in patients having mitral valvuloplasty. We anticipate the incidence to be low but do not have sufficient data to quantify it at this time. The main emphasis of our review was to look at

194 LAM ET AL Ann Thorac Surg HEMOLYSIS AFTER MITRAL VALVE REPAIR 2004;77:191 5 mechanisms of hemolysis and its treatment. Patients with hemolysis generally presented within the first 6 months of surgery. The observations in this study demonstrate that hemolysis can be associated with a variety of different echocardiographic findings after mitral valve repair. Seventy-seven percent of patients had 3 or 4 MR and the regurgitant jets usually fragmented or accelerated. However, 23% of patients had only 1 or 2 MR. Therefore, in a patient with a persistently low or falling hematocrit and recurrent MR, the clinician should have a high index of suspicion for hemolysis, particularly if regurgitation is severe and the jet fragments or accelerates. In 35% of patients hemolysis was caused by incomplete initial repair or technical error resulting in disruption of the repair; such failures are preventable. Treatment by mitral valve replacement was effective at eliminating hemolysis. The association between hemolysis and heart valve disease (native or prosthetic) has been known for more than 40 years. In 1961 Sayed and associates [12] reported a case of hemolysis in a patient with an ostium primum atrial septal defect and unrepaired mitral cleft. Subsequently, others observed that turbulent flow associated with aortic and mitral prosthetic paravalvular leaks created shearing stresses that were damaging to red blood cells [13, 14]. It is likely that improvement and more routine utilization of perioperative echocardiography at our institution over the last decade has led to increased diagnosis of postoperative hemolysis. In recent reports of hemolysis after mitral valve repair, the focus has been on exploring possible mechanisms of hemolysis. These latter have included protruding paravalvular suture material that provided an impact site for red blood cells [4, 8], whiplash motion of residual free-floating chordae tendinae [5], partial annuloplasty dehiscence [6], regurgitant jet colliding with an undehisced rigid annuloplasty ring [9], or a pledget [2] and central jets colliding with the atrial wall [7]. Hemolysis can occur in a variety of settings after mitral valve repair. In all cases there is recurrent MR, although the regurgitation need not be severe to cause clinically important hemolysis. In our review, hemolysis was not associated with any preoperative or postoperative echocardiographic variables, including mild degrees of MR. A controlled study would be required to determine predictors of hemolysis. Using transesophageal echocardiography and fluid hydrodynamic simulation, Garcia and associates [11] defined several hydrodynamic patterns associated with mitral valve hemolysis. The majority of patients with hemolysis in their study had a mitral prosthesis and only 3 patients had mitral valve repair. More recently, Yeo and associates [10] reported the hydrodynamic patterns of 13 patients with hemolysis after mitral valve repair; they observed that most patients with hemolysis had hydrodynamic patterns associated with high shear stresses (fragmentation, collision, and rapid acceleration). Comparison of patients with hemolysis and controls demonstrated that high shear hydrodynamic patterns (fragmentation, collision, and rapid acceleration) were more likely to cause hemolysis than low shear hydrodynamic patterns (slow acceleration and free jet). Similarly in the current study, high shear hydrodynamic patterns (fragmentation, collision, and rapid acceleration) caused 88% of our patients hemolysis whereas 12% were attributed to low shear hydrodynamic patterns (free jet and slow acceleration). Although all patients had MR, the mitral valve repair was physically intact in 81% of patients. Similarly, Yeo and associates [10] reported that 67% of their repairs were intact. However, the majority of our patients (77%) had 3 and 4 MR. Recurrent MR was caused by a variety of factors. In 25% of patients, progression of rheumatic heart disease led to acceleration regurgitant jets and hemolysis. In 19%, a failed repair led to traumatic hemolysis and in 16% an incomplete repair in degenerative disease also led to hemolysis; these latter two categories represent preventable causes of recurrent MR and hemolysis. In another 16%, bacterial endocarditis led to new structural damage of the mitral valve and hemolysis. These various etiologies of recurrent MR created turbulent flow and regurgitant jets that usually fragmented or accelerated. Others have observed that patients with failed repairs (dehisced rings, tears, flail segments) or incomplete repairs (redundant chordae, redundant sutures, clefts) have a greater propensity for hemolysis due to residual turbulent flow [9]. Progression of rheumatic heart disease has caused hemolysis in some patients; the mechanism is thought to be acceleration of regurgitant jets through an increasingly stenosed mitral valve. Mitral valve hemolysis has been observed with both the flexible band and the rigid ring; in this review we did not find that ring type predicted hemolysis. More data would be required to perform a just comparison. Irrespective of hydrodynamic mechanism or etiology, 88% of patients were treated by mitral valve replacement. Other series have described a similar treatment strategy [1, 10]. The hemolysis resolved in all patients treated by mitral valve replacement except for 1 patient who developed a paravalvular leak. Only 3 patients were rerepaired. At 6 years follow-up, 92% of surviving patients were in NYHA class I or II and 96% had either no MR or 1 MR. Actuarial survival was 95% at 1 year and 85% at 5 years. Conclusions Although echocardiographic findings varied in patients with hemolysis after mitral valve repair, most patients had high grade MR and jets that fragmented or accelerated. Such echocardiographic observations should suggest the possibility of hemolysis in a patient with persistently low or falling hematocrit after mitral valve repair. Regurgitant jets were caused by recurrent or residual traumatic MR, which was associated with incomplete initial repair or technical error in 35% of patients. Most patients required blood transfusions and presented early with symptoms. Mitral valve replacement was the predominant treatment and yielded favorable outcomes.

Ann Thorac Surg LAM ET AL 2004;77:191 5 HEMOLYSIS AFTER MITRAL VALVE REPAIR References 1. Cerfolio RJ, Orszulak TA, Daly RC, Schaff HV. Reoperation for hemolytic, anaemia complicating mitral valve repair. Eur J Cardiothorac Surg 1997;11:479 84. 2. Dilip KA, Vachaspathy P, Clarke B, et al. Haemolysis following mitral valve repair. J Cardiovasc Surg (Torino) 1992;33: 568 9. 3. Dol YL, Matsumura Y, Yabe T. Haemolytic anaemia after mitral valve repair. Lancet 1996;347:1330 1. 4. Goldberger AL, Orth R, Moores WY. Severe hemolytic anemia after attempted repair of paraprosthetic mitral regurgitation. Am Heart J 1982;104:1381 2. 5. Gupta SC, Suryaprasad AG. Mechanical hemolytic anemia after repair of ruptured chordae tendineae of mitral valve apparatus. Angiology 1979;30:776 9. 6. Mok P, Lieberman EH, Lilly LS, et al. Severe hemolytic anemia following mitral valve repair. Am Heart J 1989;117: 1171 3. 7. Stoschitzky K, Starz I, Anelli-Monti M, et al. Transfusionrequiring haemolytic anaemia after mitral-valve repair. Lancet 1996;347:765. 195 8. Warnes C, Honey M, Brooks N, et al. Mechanical haemolytic anaemia after valve repair operations for non- rheumatic mitral regurgitation. Br Heart J 1980;44:381 5. 9. Wilson JH, Rath R, Glaser R, Panke T. Severe hemolysis after incomplete mitral valve repair [see comments]. Ann Thorac Surg 1990;50:136 7. 10. Yeo TC, Freeman WK, Schaff HV, Orszulak TA. Mechanisms of hemolysis after mitral valve repair: assessment by serial echocardiography. J Am Coll Cardiol 1998;32:717 23. 11. Garcia MJ, Vandervoort P, Stewart WJ, et al. Mechanisms of hemolysis with mitral prosthetic regurgitation. Study using transesophageal echocardiography and fluid dynamic simulation. J Am Coll Cardiol 1996;27:399 406. 12. Sayed HM, Dacie JV, Handley DA, Lewis SM, Cleland WP. Hemolytic anemia of mechanical origin after open heart surgery. Thorax 1961;16:356 60. 13. Nevaril C, Lynch E, Alfrey CJ, Hellums J. Erythrocyte damage and destruction induced by shearing stress. J Lab Clin Med 1968;71:784 90. 14. Rodgers BM, Sabiston DJ. Hemolytic anemia following prosthetic valve replacement. Circulation 1969;39:155 61. Southern Thoracic Surgical Association: Fifty-First Annual Meeting The Fifty-First Annual Meeting of the Southern Thoracic Surgical Association (STSA) will be held November 4 6, 2004, in Cancun, Mexico. Members wishing to participate in the Scientific Program should submit an abstract by April 9, 2004, 5:00 pm Central Daylight Time. Abstracts must be submitted electronically. Instructions for the abstract submission process can be found on the STSA Web site at www. stsa.org; on the CTSNet Web site at www.ctsnet.org; or in the back of the issue of The Annals of Thoracic Surgery. Manuscripts accepted for the Resident Competition must be submitted to the STSA headquarters office no later than September 17, 2004. The Resident Award will be based on abstract, presentation, and manuscript. 2004 by The Society of Thoracic Surgeons Ann Thorac Surg 2004;77:195 0003-4975/04/$30.00 Published by Elsevier Inc