Balancing Stenosis and Regurgitation During Mitral Valve Surgery in Pediatric Patients

Transcription

1 Balancing Stenosis and Regurgitation During Mitral Valve Surgery in Pediatric Patients Abdullah A. Alghamdi, MD, Bobby Yanagawa, MD, PhD, Steve K. Singh, MD, Ari Horton, MD, Osman O. Al-Radi, MD, and Christopher A. Caldarone, MD Department of Surgery, Division of Cardiovascular Surgery, Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada; Department of Cardiac Sciences, King Abdulaziz Cardiac Center, National Guard Health Affairs, Riyadh; and Department of Surgery, King Abdulaziz University, Jeddah, Saudi Arabia Background. The objectives of this study were to evaluate the composite outcome of reoperation or death for mitral valve repair and replacement and to determine the relative importance of regurgitation and stenosis during mitral valve repair to guide intraoperative decision making. Methods. All consecutive patients undergoing firsttime mitral valve surgery from 1990 to 2008 at our institution were reviewed. Those with atrioventricular septal defects were excluded. Intraoperative transesophageal echocardiography was used to assess mitral repairs. Residual mitral valve stenosis (mean gradient) and regurgitation (Grade I less than mild; Grade II mild to less than moderate; and Grade III moderate or greater) were assessed. Results. One hundred-fifty patients were included with median age of 47 months (range, 0.03 to 228 months) and median weight of 14 kg (range, kg). Of these, 18 (12%) patients underwent replacement, and 132 (88%) patients underwent primary repair. In-hospital mortality was 6%. Of the patients reviewed, 21% underwent reoperation (repair or replacement) or died. Multivariable proportional hazard modeling adjusted for weight, valve replacement, postrepair mean gradient and grade of regurgitation showed postrepair mitral valve regurgitation to be the predominant predictor of death or reoperation. Compared with patients with mild to moderate mitral regurgitation, those with stenosis, but Grade I regurgitation, were associated with improved outcomes. Conclusions. Within the limits of our analysis, residual mild mitral valve stenosis was associated with less hazard than mild to moderate regurgitation. Taken together, our data suggest that an intraoperative strategy to minimize mitral valve regurgitation is rational, even at the expense of mild mitral stenosis. (Ann Thorac Surg 2011;92:680 4) 2011 by The Society of Thoracic Surgeons Outcomes after pediatric mitral valve (MV) replacement are inferior in comparison to a perfect mitral repair and superior to a poor repair [1, 2]. Between these extremes is a breakpoint at which mitral repairs have similar outcomes to replacements. Because a surgeon can choose to make repeated attempts to improve a repair, or bail out, with a mitral replacement (in a patient with an appropriately size annulus), knowledge of this relationship could be used to inform intraoperative decision making. Material and Methods Study Population After approval of the research ethics board, all consecutive pediatric cardiac surgery patients (age 19 years) who underwent first-time mitral valve surgery (repair or replacement) at the Hospital for Sick Children between 1990 and 2008 were reviewed. Patients with associated diagnoses of atrioventricular septal defect, functional Accepted for publication March 15, Address correspondence to Dr Alghamdi, The Hospital for Sick Children, 555 University Ave, Ste 1525, Toronto, ON, M5G 1X8, Canada; abdullah.alghamdi@utoronto.ca. single ventricle and transposition of great arteries were excluded. Intraoperative Measurement Intraoperative transesophageal echocardiography was used to assess the integrity of mitral valve repair. Quantitative assessment of mitral valve stenosis was estimated based on the mean gradient across the valve and was further stratified in the following manner: Grade I being mean gradient less than 5 mm Hg; Grade II being mean gradient from 5 mm Hg to less than 10 mm Hg; and Grade III being mean gradient 10 mm Hg or more. The mitral valve regurgitation was qualitatively assessed based on the regurgitant jet and was stratified in the following manner: Grade I being less than mild; Grade II being mild to less than moderate; and Grade III being moderate or greater. Statistical Methods The primary outcome was a composite of mitral valve reoperation or death after primary mitral valve procedure. Continuous data were presented as medians with ranges, and categorical data were presented as proportions. Analysis of variance was used to compare continuous variables, and a chi-square test was used to com by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg ALGHAMDI ET AL 2011;92:680 4 PEDIATRIC MITRAL VALVE SURGERY 681 Table 1. Characteristics of Patients Who Underwent Primary Mitral Valve Procedures Variable Repair (n 132) Replacement (n 18) p Value Age (months) 43 ( ) 116 ( ) 0.15 Males 65 (49%) 9 (50%) 0.9 Weight (Kg) 14 ( ) 24 (2.8-73) 0.47 Infant ( Age <1 Year) 42 (32%) 4 (22%) 0.4 Obstructive lesion a 37 (28%) 5 (28%) 0.9 Need for mitral valve 19 (14.3%) 2 (11%) 0.7 reoperation In-hospital death 5 (4%) 4 (22%) 0.01 Overall death (early & late) 14 (10.6%) 4 (22%) 0.15 a Obstructive lesion is defined as history of previous or additional surgery for left ventricular outflow tract obstructive lesion including: subaortic stenosis, aortic stenosis, supra-aortic stenosis and aortic arch hypoplasia and aortic coarctation. pare categorical variables. Kaplan-Meier statistics were used to compare time to reoperation or death. A log-rank test was used to compare survival curves. An associated p-value of less than 0.05 was considered significant. A multivariable proportional hazard model was used to identify independent predictors of death or reoperation. Variables were kept in the adjusted statistical model if they were statistically significant in univariable analysis or if they were considered clinically important. Relevant variables were tested for collinearity, and a decision was made to keep the more clinically relevant variables (eg, body weight and age at time of surgery). The final proportional hazard model was assessed for linearity assumption and need for data transformation using the method of cumulative sums of Martingale residuals [3]. Interaction terms between variables were tested where clinically appropriate. Where there was a violation of linearity assumption, data transformation with graphic representation was used. Where continuous variables needed dichotomized, the area under receiver operating characteristics (ROC) curve was used to achieve the best cut-off value. The statistical software system SAS, version 9.2(SAS Corp, Cary, NC) was used for all statistical analyses. Results Analysis included 150 patients. The 150 patients underwent 179 mitral valve procedures, of which 150 were as primary procedures and 29 were as repeated procedures. Of the 150 primary mitral valve procedures, 132 (88%) procedures were mitral valve repair, and 18 (12%) were primary mitral valve replacement. At the time of first mitral valve operation, median age was 47 months (range, months), and the median weight was 14 kg (range, kg). Of the 29 repeated procedures, 14 had a single reoperation, 6 had a second reoperation and 1 had a third reoperation. Among the 29 repeated procedures, 7 were mitral valve repairs, and 22 were mitral valve replacements. The etiology of mitral valve disease included: dysplasia of the mitral valve (n 101), myxomatous disease (n 26), endocarditis (n 12) and rheumatic disease (n 11). The predominant valve dysfunction was regurgitation (n 92), stenosis (n 20) and mixed regurgitation and stenosis (n 38). Of the 150 patients who underwent primary mitral valve procedure, the overall in-hospital mortality was 9 (6%). Thirty-two (21%) patients underwent mitral valve reoperation (21) or died (11; Table 1). Among the 132 patients who underwent mitral valve repair, frequency of reoperation or death increased with the increasing grades of postrepair intraoperative regurgitation: 9% in Grade I, 22% in Grade II and 80% in Grade III (Fig 1). Similarly, but less significantly, the frequency of reoperation or death increased with increasing postrepair intraoperative stenosis: 14% in Grade I, 29% in Grade II and 57% in Grade III (Fig 2). The interaction between postrepair regurgitation and stenosis was not statistically significant (ie, increasing grades of stenosis were not observed with less grades of regurgitation). However, when the adjusted hazard was graphically represented, taking into account regurgitation and stenosis, a clinically important observation was noted in those with Grade II regurgitation (ie, mild to less than Fig 1. Kaplan-Meier Survival curves (freedom from reoperation or death) for the different grades of postrepair intraoperative mitral valve regurgitation. Postrepair intraoperative mitral valve regurgitation grades: Grade I less than mild, Grade II mild to less than moderate and Grade III moderate or more.

3 682 ALGHAMDI ET AL Ann Thorac Surg PEDIATRIC MITRAL VALVE SURGERY 2011;92:680 4 Fig 2. Kaplan-Meier Survival curves (freedom from reoperation or death) for the different grades of postrepair intraoperative mitral valve stenosis. Postrepair intraoperative mitral valve stenosis grades: Grade I mean gradient less than 5 mm Hg, Grade II mean gradient 5 to less than 10 mm Hg and Grade III 10 mm Hg or more. moderate). In this group (ie, those with postrepair intraoperative Grade II regurgitation), increasing grade of stenosis was associated with less hazard than that of regurgitation (Fig 3). The clinical relevance of this observation suggests that, when trading between mild stenosis and mild regurgitation, improving mitral regurgitation at the expense of mild mitral stenosis is reasonable, as mild stenosis is associated with less hazard (ie, reoperation or death) than that of increasing regurgitation. For the entire cohort (N 150, primary repair and replacement), the adjusted proportional hazard model included body weight at the time of operation, mitral valve replacement, postrepair stenosis (mean gradient in mm Hg) and grade of mitral valve regurgitation (Grades I, II and III). The year of surgery and overall interaction terms between regurgitation and stenosis were not significant, in turn, they were eliminated from the final model. Body weight was tested for the linearity assumption, and the relationship with Fig 3. Graphic representation of the adjusted interaction between postrepair stenosis and grade of regurgitation. Postrepair regurgitation grades: Grade I less than mild, Grade II mild to less than moderate and Grade III moderate or more. the hazard function was not linear (Fig 4). Therefore, body weight was dichotomized to two values with 13 kg as the cut-off value. This value was chosen based on the statistics of ROC, where 13 kg yielded the maximum area under ROC curve. Postrepair regurgitation was a predominant predictor of reoperation or death. Those with postrepair intraoperative mitral valve regurgitation of Grade III (moderate or more) had a hazard value that was not different from that of valve replacement (Table 2). Comment The optimal outcome for mitral repair is restoration of valvular function in the absence of residual stenosis and regurgitation. The advantages of repair include: allowance for valve growth, avoidance of anticoagulation and prosthesis-related complications, and association with more favorable outcomes replacement [1, 2, 4-14]. A perfect repair of the mitral valve, however, is not always possible, and achievable repairs are associated with a range of quality in terms of residual regurgitation and stenosis. These achievable repairs must be compared with mitral valve replacement in the subset of patients in whom the mitral annulus is of sufficient size to permit insertion of prosthesis. The relationship between the quality of mitral valve repair and late outcomes in the context of comparison with mitral valve replacement has not been extensively studied in the pediatric population. Clarification of this relationship would help inform intraoperative decision making when a surgeon must choose between repeated attempts to improve an imperfect repair versus proceeding with mitral valve replacement. In this study, 150 patients were included, and 18 underwent primary mitral valve replacement. The analysis was adjusted for multiple variables, including body weight and residual valve lesions. Of the most notable findings from this study is, first, that intraoperative postrepair mitral valve regurgitation was a predominant independent predictor of reoperation or death. Residual Grade I regurgitation (ie, less than mild) was a strong

4 Ann Thorac Surg ALGHAMDI ET AL 2011;92:680 4 PEDIATRIC MITRAL VALVE SURGERY 683 Fig 4. Graphic representation of the adjusted relationship between hazard function and variables included in the final model. (A) Mitral valve replacement. (B) Body weight represented as non-linear with restricted cubic spline function. (C) Postrepair intraoperative mitral valve regurgitation grades: Grade I less than mild, Grade II mild to less than moderate and Grade III moderate or more. (D) Postrepair mean gradient. protective factor from reoperation or death, whereas Grade III residual regurgitation was a strong risk factor for reoperation or death. Patients with Grade III regurgitation had outcomes that were not different for the primary valve replacement group. Table 2. Final Adjusted Proportional Hazard Model for the Entire Cohort Variable Hazard Ratio p Value Weight less than 13 kg at the time of surgery Primary mitral valve replacement compared to grade II postrepair regurgitation Postrepair intraoperative mitral valve mean gradient Grade I postrepair regurgitation compared to grade II Grade III postrepair regurgitation compared to grade II Postrepair regurgitation grades: Grade I less than mild, Grade II mild to less than moderate and Grade III moderate or more. Grade II regurgitation is the reference group. Hazard ration is for the composite outcome of reoperation or death. The other notable and clinically important finding from this study is that patients with residual Grade II regurgitation (ie, mild to less than moderate) are a unique subset, in that the interaction between stenosis and regurgitation was significant and the hazard of increasing stenosis was less than that of increasing regurgitation. This could inform the intraoperative decision making, where the surgeon can trade between mild stenosis and regurgitation in favor of improving the regurgitation even at the expense of mild stenosis. In conclusion, our data suggest that intraoperative postrepair residual mitral valve regurgitation is a dominant predictor of reoperation or death. Our data also suggest that, when residual postrepair regurgitation is mild to less than moderate, it is reasonable to minimize the grade of regurgitation at the expense of mild stenosis. References 1. Caldarone CA, Raghuveer G, Hills CB, et al. Long-term survival after mitral valve replacement in children aged 5 years: a multi-institutional study. Circulation 2001;104:I Oppido G, Davies B, McMullan DM, et al. Surgical treatment of congenital mitral valve disease: midterm results of a repair-oriented policy. J Thorac Cardiovasc Surg2008;135: Lin DY, Wei LJ, Ying Z. Model-checking techniques based on cumulative residuals. Biometrics 2002;58: Almeida RS, Elliott MJ, Robinson PJ, et al. Surgery for congenital abnormalities of the mitral valve at the Hospital

5 684 ALGHAMDI ET AL Ann Thorac Surg PEDIATRIC MITRAL VALVE SURGERY 2011;92:680 4 for Sick Children, London from J Cardiovasc Surg 1988;29: Beierlein W, Becker V, Yates R, et al. Long-term follow-up after mitral valve replacement in childhood: poor event-free survival in the young child. Eur J Cardiothorac Surg 2007;31: Eble BK, Fiser WP, Simpson P, Dugan J, Drummond-Webb JJ, Yetman AT. Mitral valve replacement in children: predictors of long-term outcome. Ann Thorac Surg 859; 76: Galioto FM, Jr, Midgley FM, Shapiro SR, Perry LW, Ciaravella JM, Jr, Scott LP, III. Mitral valve replacement in infants and children. Pediatrics 1981;67: Gunther T, Mazzitelli D, Schreiber C, et al. Mitral-valve replacement in children under 6 years of age. Eur J Cardiothorac Surg 2000;17: Hetzer R, Delmo Walter EBM, Hubler M, et al. Modified surgical techniques and long-term outcome of mitral valve INVITED COMMENTARY Alghamdi and associates [1] have highlighted the delicate balance that can be achieved between mitral stenosis and mitral regurgitation when repairing the mitral valve in children in order to achieve a more durable outcome and avoid mitral valve replacement. They show that greater than mild to moderate mitral regurgitation is poorly tolerated and an independent predictor of reoperation or death. The reduction of mitral regurgitation to less than moderate to mild by the creation of mild mitral stenosis makes for a better long-term repair. This approach of avoiding mitral valve replacement at the expense of a less-than-perfect repair, but one that is well tolerated with balanced regurgitation and stenosis, does allow for growth of the patient and removes the risks of anticoagulation with mitral valve replacement. Deferring mitral valve replacement to the older and larger patient allows a bigger valve to be placed. This article deals with patients of a median age of 4 years, with dysplastic valve morphology rather than the complex morphology seen in infancy, such as parachute mitral valve and where multiple obstructive lesions are present on the left side of the heart (Shone syndrome). Patients in the first few months and year of life are almost certainly more challenging than the cohort in article by Alghamdi and colleagues. However, the approach of achieving balanced mitral stenosis and mitral regurgitation, and a repaired valve, is certainly applicable to this younger, more difficult group. Long-term evaluation outcome for these patients should include the ability of the patient to thrive and develop and the lack of development of pulmonary reconstruction in 111 children. Ann Thorac Surg 2008; 86: Human DG, Joffe HS, Fraser CB, Barnard CN. Mitral valve replacement in children. J Thorac Cardiovasc Surg 1982;83: Kadoba K, Jonas RA, Mayer JE, Castaneda AR. Mitral valve replacement in the first year of life. J Thorac Cardiovasc Surg 1990;100: Kajihara N, Imoto Y, Kan O, et al. Clinical results of commissure plication annuloplasty for mitral regurgitation in children. Surg Today 2009;39: Katogi T, Aeba R, Cho Y, et al. Mitral valve replacement in patients younger than 6 years of age. Jpn J Thorac Cardiovasc Surg 1999;47: Zweng TN, Bluett MK, Mosca R, Callow LB, Bove EL. Mitral valve replacement in the first 5 years of life. Ann Thorac Surg 1989;47: hypertension. The mitral valve hemodynamics were assessed on the operating table under anaesthetic and probably underestimates the degree and severity of mitral stenosis and regurgitation. However, because this assessment mode was uniformly applied to all patients, it would seem to be acceptable. Mitral valve repair in young patients, particularly in the first few months of life, is difficult and challenging. The underlying morphology is complex, and individual variation in morphology from patient to patient makes each case a separate challenge. Although this report is in older children, the concept that a balanced repair that leaves mild to moderate valve regurgitation together with mild mitral stenosis to compensate for what otherwise would be severe mitral regurgitation, with a very bad outcome, seems eminently sensible. William Brawn, FRCS, FRACS Department of Cardiac Surgery Diana, Princess of Wales Children s Hospital Steelhouse Lane Birmingham, West Midlands, B4 6NH United Kingdom sec.brawn@bch.nhs.uk Reference 1. Alghamdi AA, Yanagawa B, Singh SK, Horton A, Al-Radi OO, Caldarone CA. Balancing stenosis and regurgitation during mitral valve surgery in pediatric patients. Ann Thorac Surg 2011;92: by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur