- Email: [email protected]
Usefulness of Preoperative Cardiac Dimensions to Predict Success of Reverse Cardiac Remodeling in Patients Undergoing Repair for Mitral Valve Prolapse
Usefulness of Preoperative Cardiac Dimensions to Predict Success of Reverse Cardiac Remodeling in Patients Undergoing Repair for Mitral Valve Prolapse Leonidas V. Athanasopoulos, MDa, Siobhan McGurk, BSa, Zain Khalpey, MDa, James D. Rawn, MDa, Jan D. Schmitto, MDa,b, Laurens W. Wollersheim, MDa, Ann M. Maloney, BAa, and Lawrence H. Cohn, MDa,* Mitral valve repair for mitral regurgitation (MR) is currently recommended based on the degree of MR and left ventricular (LV) function. The present study examines predictors of reverse remodeling after repair for degenerative disease. We retrospectively identified 439 patients who underwent repair for myxomatous mitral valve degeneration and had both pre- and postoperative echocardiographic data available. Patients were categorized based on left atrial (LA) diameter and LV diameter standards of the American Society of Echocardiography. The outcome of interest was the degree of reverse remodeling on all heart dimensions at follow-up. Mean age was 57 – 12 years, and 37% of patients were women. Mean preoperative LV end-diastolic diameter was 5.8 – 0.7 cm, LV end-systolic diameter 3.5 – 0.6 cm, LA 4.7 – 0.7 cm, and median ejection fraction 60%. Median observation time was 81 months, and time to postoperative echocardiography was 38 months. Overall, 95% of patients had normal LV diastolic dimensions postoperatively, 93% normal LV systolic dimensions, and 37% normal LA dimensions. A Cox regression analysis showed that moderate (odds ratio [OR] 2.1, 95% confidence interval [CI] 1.3 to 3.4) or severe preoperative LA dilatation (OR 2.7, 95% CI 1.7 to 4.4), abnormal preoperative LV end-systolic dimensions (OR 1.3, 95% CI 1.1 to 1.5), and age in years (OR 1.02, 95% CI 1.01 to 1.03) were predictive of less reverse remodeling on follow-up. In conclusion, preoperative LV end-systolic dimensions and LA dilatation substantially affect the likelihood of successful LA remodeling and normalization of all heart dimensions after mitral valve repair for MR. These findings support early operation for MR before the increase in heart dimensions is nonreversible. Ó 2014 Elsevier Inc. All rights reserved. (Am J Cardiol 2014;113:1006e1010)
The regurgitant jet of blood from mitral regurgitation (MR) overloads the left atrium (LA) and the left ventricle (LV).1 LV loading is critical in understanding the remodeling process.1 The remodeling due to MR is different from that seen in other left-sided valve diseases such as mitral stenosis, aortic stenosis, or aortic regurgitation in that it manifests with higher LV radius/thickness ratios and lower mass/volume ratios.2 Reverse LA and LV remodeling may happen early after surgical correction of MR,3e6 but longstanding MR left untreated leads to deterioration of LV function and impacts the potential for reverse remodeling. Thus, the degree of disease progression is an important consideration when weighing the value of surgical intervention. The present study evaluates long-term outcomes in a series of patients who underwent mitral valve repair for MR due to degenerative disease (prolapse). We sought to a Division of Cardiac Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts and bDivision of Cardiothoracic, Transplantation and Vascular Surgery, Hannover Medical School, Hannover, Germany. Manuscript received September 17, 2013; revised manuscript received and accepted December 4, 2013. See page 1010 for disclosure information. *Corresponding author: Tel: (617) 732-6569; fax: (617) 264-6369. E-mail address: [email protected] (L.H. Cohn).
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2013.12.009
investigate whether preoperative echocardiographic data or other factors may predict the degree of reverse remodeling seen at long-term follow-up. Methods With approval from our internal review board, we identified all 824 patients who underwent minimally invasive mitral valve repair from August 1996 to November 2010. Our inclusion criteria were patients with MR due to myxomatous mitral valve degeneration, normal coronary arteries, availability of a preoperative echocardiogram within 1 year of the surgery date, and a follow-up echocardiogram at least 6 months after surgery. Subjects with previous mitral valve surgery or mitral stenosis were excluded; 439 patients met our inclusion criteria. Clinical and surgical data were collected at the time of presentation, extracted from hospital electronic records or by chart review, and coded to Society of Thoracic Surgeons Adult Cardiac Database, version 2.52, specifications. Mortality data were collected during routine postoperative follow-up and supplemented by a query of the Social Security Death Index. Our main outcome of interest was reverse remodeling of the heart, defined as each of the following being within www.ajconline.org
Valvular Heart Disease/Preoperative Cardiac Dimensions Predict Reverse Remodeling
normal parameters: LV end-diastolic diameter, LV endsystolic diameter, and LA diameter. These were collected by a manual review of the echocardiographic report. All echocardiographies were performed transthoracically. The preoperative echocardiography was the study most proximate to the surgery. If multiple echocardiographic studies existed, the latest normal or the earliest abnormal study was used. Only 1 preoperative and 1 postoperative echocardiogram for each patient was collected in the research data file. Grade of MR was assessed using color-flow Doppler as none or trivial (0), trace (1þ), mild (2þ), moderate (3þ), and severe (4þ). We grouped patients based on the preoperative severity of abnormality of heart dimensions. For LV end-systolic diameter, the grouping was based on a preoperative value of 4 cm (normal) or >4 cm (abnormal). LV end-diastolic diameter and LA groupings were derived from the American Society of Echocardiography’s recommendations for chamber quantification.7 For men, normal LV end-diastolic diameter, 4.2 to 5.9 cm; mildly abnormal LV end-diastolic diameter, 6.0 to 6.3 cm; moderately abnormal LV end-diastolic diameter, 6.4 to 6.8 cm; severely abnormal LV end-diastolic diameter, 6.9 cm; normal LA, 3.0 to 4.0 cm; mildly abnormal LA, 4.1 to 4.6 cm; moderately abnormal LA, 4.7 to 5.2 cm; and severely abnormal LA, 5.3 cm. For women, normal LV end-diastolic diameter, 3.9 to 5.3 cm; mildly abnormal LV end-diastolic diameter, 5.4 to 5.7 cm; moderately abnormal LV end-diastolic diameter, 5.8 to 6.1 cm; severely abnormal LV end-diastolic diameter, 6.2 cm; normal LA, 2.7 to 3.8 cm; mildly abnormal LA, 3.9 to 4.2 cm; moderately abnormal LA, 4.3 to 4.6 cm; and severely abnormal LA, 4.7 cm. Successful postoperative remodeling for each of the heart dimensions was defined as measurements equal or the upper normal value; for men, LV end-diastolic diameter 5.9 cm and LA 4.0 cm, and for women, LV enddiastolic diameter 5.3 cm and LA 3.8 cm. Normal postoperative LV end-systolic diameter for all the samples was defined as 4.0 cm. Our secondary outcome of interest was survival over the study observation period. Because we excluded reoperations and early mortality, the end point for survival was the time in months from operation until (1) the first postoperative echocardiography indicating heart abnormality, (2) death, or (3) alive at the end of the study observation period, with normal postoperative echocardiographic data. The study period ended on June 30, 2011. Categorical data are expressed as numbers and percentages and were compared using Fisher’s exact test. Continuous variables are presented as mean and SD or median and interquartile range for normally or nonnormally distributed continuous data, respectively, and were compared using t test with Levine test for homogeneity of variance or MannWhitney U test, as appropriate. Multivariate analysis of survival with complete reverse heart remodeling was estimated using a forward stepwise Cox regression model. An event was the first of either an echocardiogram showing any dimension abnormal or report of death. Tested variables were derived from available literature, clinical expertise, and variables that differed between groups on exploratory univariate analysis or could reasonably be expected to impact our outcomes. Interaction terms were examined, and colinearity diagnostics were
1007
Table 1 Characteristics and operative and in-hospital outcomes of 439 patients with mitral valve repair Characteristic Preoperative characteristic Age (yrs) Age 65 yrs Women Body surface area New York Heart Association class III or IV Diabetes mellitus Chronic obstructive pulmonary disease Renal failure Atrial fibrillation Echo data 3þ MR LV diameter/diastole LV diameter/systole LA dimension Ejection fraction (%) Operative data No ring used Ring size Bypass time (minutes) Crossclamp time (minutes) Postoperative outcome Transfused with red cells Units per transfused patient Renal failure New-onset renal failure Atrial fibrillation New-onset atrial fibrillation Died during follow-up Follow-up time (mo)
Study Cohort, n ¼ 439 Mean n n Mean n n n n n
57.0 12.1 121 (27.6) 164 (37.4) 1.9 0.2 88 (20.0) 8 (1.8) 26 (5.9) 6 (1.4) 35 (8.0)
n Mean Median Mean Median Mean Median Median
438 (99.8) 5.77 0.7 5.73 (5.4e6.1) 3.52 0.6 3.50 (3.2e3.9) 4.69 0.7 4.60 (4.2e5.0) 60 (55e65)
n Median Median Median
5 34 115 81
(1.1) (26e38) (98e139) (69e98)
n Median n n n n n Median
130 2.00 27 22 137 110 20 81
(29.6) (1e3) (6.2) (5.0) (31.2) (25.1) (4.6) (53e116)
Number of cases presented with the percentage in parentheses. Means are presented with SDs and medians with interquartile ranges, except for ring size, which is a range.
evaluated. Statistical analyses were performed with SPSS, version 13.0 (SPSS Inc., Chicago, Illinois). Many cases had individual measures missing at random. Because deleting patients with missing values introduces biases that may not be identified or adequately controlled and affects generalizability of findings, we decided to impute missing data rather than conduct a complete-case analysis.8,9 Preoperative LV end-diastolic diameter was missing for 89 patients (18%), LV end-systolic diameter missing for 107 (22%), and LA missing in 105 patients (21%). All patients had at least 2 of these data points present. Missing data were estimated using multiply imputed linear regression approach with random-error component; for each imputed measure, a minimum of 5 data sets (maximum of 20) were created until stability in the standard errors was achieved. The regression models included all covariates, including outcome variables.10,11 Compared with known data, the correlation for imputed LV end-systolic diameter was R2 ¼ 0.988, LV enddiastolic diameter was R2 ¼ 0.988, and LA R2 ¼ 0.993. Imputed measures were only used when that variable was considered a covariate for that analysis. Outcome measures
1008
The American Journal of Cardiology (www.ajconline.org) Table 2 Reverse remodeling across heart dimensions at follow-up compared with baseline echocardiography Preoperative Category
Patients (n)
Absolute % Reduction, Mean SD
Patients With Normal Dimensions at Follow-Up, n (%)
LV end-diastolic dimension at follow-up LV end-diastolic enlargement Normal Mild Moderate Severe
Figure 1. Scatter plot of preoperative LV dimensions in diastole and systole, grouped by LA dimensions. The vertical line denotes the boundary for abnormal LV end-systolic diameter. The Horizontal long-dash line is the threshold for abnormal LV end-diastolic diameter for women, the shortdash line is the threshold for men. Abnormal thresholds are the lower boundaries of moderate enlargement.
were present for all cases, except in 21 patients (<5%) missing postoperative LA diameter. Missing outcome variables were not imputed. Results The baseline characteristics of our study cohort are presented in Table 1. Mean age at surgery was 57 12.1 years, and 28% were 65 years; 164 of 439 (40%) were women, and 20% were in New York Heart Association class III or IV. At baseline, the median LV end-diastolic diameter was 5.73 cm (5.4 cm to 6.1 cm) and LA was 4.60 cm (4.2 cm to 5.0 cm). Over the study period, there were 20 late deaths. Median survival time was estimated at 14.3 years (95% confidence interval [CI] 13.9 to 14.5). Concomitant procedures included patent foramen ovale closure (24 of 439, 5%), atrial septal defect closure (4 of 439, 1%), and tricuspid valve repair (14 of 439, 3%). Five patients had concomitant Maze. Alt. Cox maze III. procedures. A lower ministernotomy was done for 86% of patients (380 of 439); 49 (11%) had a right parasternal incision, 8 (2%) a right thoracotomy, and 2 (0.5%) an upper ministernotomy. In most cases, a posterior leaflet resection or modified resection and sliding annuloplasty (332 of 439, 76%) was performed and an annuloplasty ring implanted (434 of 439, 99%). Of patients with ring annuloplasty, 417 of 439 (95%) received a Cosgrove-Edwards Ring (Edwards Lifesciences, Irvine, California). Table 1 presents the operative and in-hospital outcomes for our cohort. We assigned our patients into groupings based on their preoperative heart dimensions. For LV end-diastolic measures, our a priori criteria were that anything greater than and including the lower range of moderate enlargement (for men, 6.4 cm and for women, 5.8 cm) was considered abnormal enlargement and dimensions below that were normal. For LV
209 12.5 11.5 203 (97.1) 117 17.4 9.7 111 (94.9) 80 21.1 9.5 73 (91.3) 33 25.2 8.3 28 (84.8) LV end-systolic dimension at follow-up
LV end-systolic enlargement Normal Abnormal
370 69
LA dilation Normal Mild Moderate Severe
57 108 121 132
6.5 20.9 352 (95.1) 19.1 16.7 55 (79.7) LA diameter at follow-up 4.8 6.1 9.3 16.2
19.7 15.3 14.1 13.5
41 63 36 24
(71.9) (58.3) (29.8) (18.2)
end-systolic measures, any value >4 cm was abnormal and 4 normal. For LA, we used the lower range of moderate enlargement (4.7 cm for men and 4.3 cm for women) as the threshold for abnormal, whereas measurements below that were considered normal. At the time of surgery, 74% (326 of 439) of patients met normal LV end-diastolic criteria, 84% (370 of 439) met normal LV end-systolic criteria, and 39% (170 of 439) met normal LA criteria. Figure 1 shows the relation between the preoperative LV end-systolic iameter, LV end-diastolic diameter and LA dimensions. The correlation between LV end-diastolic and end-systolic diameters is very high (R2 ¼ 0.471, p <0.001). However, the correlation between either and LA dimensions is R2 0.08. As can be seen in Figure 1, most patients with abnormal LA dimensions presented with normal or mild enlargement of the LV end-diastolic diameter (68%, 181 of 266) and LV end-systolic diameter (83%, 221 of 266). Overall, 95% of patients had normal LV end-diastolic diameter at follow-up, 93% had normal LV end-systolic diameter, but only 37% exhibited normal LA size, and 35% were normal on all measures. Table 2 presents the degree of reverse remodeling and the percentage of patients with dilatation at follow-up, within each category of preoperative dilatation. The trend noted for each of LV end-diastolic diameter, LV end-systolic diameter, and LA was that for increasing degrees of preoperative dilatation, the degree of reverse remodeling, as indicated by the percent difference of the postoperative measurements from the baseline measurements, was likewise increased. However, the percentage of patients who met normal criteria in each group decreased; patients having severe preoperative LV end-diastolic diameter enlargement were less likely to normalize LV enddiastolic diameter postoperatively than the patients with mild enlargement (odds ratio [OR] 0.34, 95% CI 0.11 to 0.99). Regarding LA measurements, 58% of those with mild
Valvular Heart Disease/Preoperative Cardiac Dimensions Predict Reverse Remodeling
1009
Table 3 Performance of the Cox regression analysis of survival with reverse heart remodeling Significant Predictors
Preoperative LA dilation (vs none) Mild Moderate Severe Age (in yrs, per yr) Preoperative LV end-systolic dimensions (continuous)
Model performance Figure 2. Cox proportional hazards regression analysis of survival with normal heart dimensions at echocardiographic follow-up, grouped by the degree of preoperative LA dilation. Overall model performance: 2 log likelihood ¼ 3,328.945, chi-square ¼ 54.464, df ¼ 5, p 0.001.
enlargement, 30% with moderate (OR 0.59, 95% CI 0.46 to 0.73), and 18% with severe enlargement (OR 0.51, 95% CI 0.40 to 0.65) did not have dilatation at follow-up. We used a Cox proportional hazards analysis to evaluate survival with complete heart reverse remodeling (Figure 2). Tested variables included gender, body surface area, preoperative ejection fraction and heart dimensions, pre- and postoperative atrial fibrillation, New York Heart Association class, chronic lung disease, renal failure, diabetes, etiology, and bypass time (Table 1). Significant predictors included moderate (OR 2.1, 95% CI 1.3 to 3.4) or severe preoperative LA dilatation (OR 2.7, 95% CI 1.7 to 4.4), abnormal preoperative LV end-systolic dimensions (OR 1.3, 95% CI 1.1 to 1.5), and age in years (OR 1.02, 95% CI 1.01 to 1.03) (Table 3). Discussion One of our observations was that the percentage of change in postoperative remodeling correlated to preoperative dimensions, with larger LV dimensions experiencing larger percent changes. Therefore, percent change may not be a fair marker for successful reverse remodeling, as has been done in previous studies.3,12 Instead, we chose to use whether postoperative measurements had returned to normal parameters as the measure of successful outcome. It is striking that although 95% and 93% of the patients had normal LV end-diastolic diameter and LV end-systolic diameter, respectively, at follow-up, only 37% of them had normal LA dimensions. The LA wall is thinner than the LV wall,13 which may make dilation less reversible. These factors explain our Figure 1, which indicates that meaningful LA dilation often occurs while ventricular dimensions remain normal, as well as our observation that 95% of those in the moderate LV end-diastolic enlargement group were normal at follow-up compared with only 30% of those with baseline moderate LA dilatation. On multivariate analysis, degree of LA dilation, LV endsystolic diameter, and age (in years) were predictors of reverse
p
Hazard Ratio
95% CI Lower Upper
0.001 0.532 0.002 0.001 0.002 0.009
1.180 2.115 2.734 1.017 1.264
0.703 1.305 1.698 1.006 1.054
2 Log Chi-Square df Likelihood 2,742.531 52.925 5
Noncontributing factors tested Diabetes mellitus (binary) Body surface area (continuous) Bypass time (continuous) Atrial fibrillation* (categorical) Chronic lung disease (binary) Preoperative renal failure (binary) Gender (binary) New York Heart Association class (categorical) Preoperative ejection fraction (%, continuous)
1.979 3.428 4.402 1.027 1.514 p 0.001 p 0.105 0.238 0.362 0.372 0.496 0.509 0.632 0.724 0.926
An event was either the earliest abnormal echocardiogram recorded or death. Time was calculated in months from the date of operation to the event or study follow-up if event free. Inclusion criterion was p 0.05; exclusion criterion was p >0.10. * Atrial fibrillation was a categorical value with 4 possible values: none, history of atrial fibrillation preoperatively, atrial fibrillation both preoperatively and postoperatively, and new-onset atrial fibrillation postoperatively.
remodeling at follow-up (Table 3 and Figure 2). This is consistent with the finding by Cho et al3 that age, sinus rhythm, predominant MR, and preoperative LA volume predicted early LA remodeling. Antonini-Canterin et al12 also identified preoperative LA volume along with blood pressure, postoperative mean mitral valve pressure gradient, and severity of residual MR as predictors of 15% postoperative reduction in LA dimension within 6 months of surgery. In our study, ejection fraction did not predict LA dimension at follow-up. In contrast, preoperative LV end-systolic diameter was predictive of both LV and LA reverse remodeling. This suggests that LV end-systolic diameter may be the better marker to decide whether to intervene with mitral valve repair. Our findings indicate that when MR remains untreated surgically until LV diameters are moderate to severely abnormal or LA moderately or severely dilated, successful reverse remodeling becomes less likely. It is thus important that all dimensions be taken into account along with the MR grade in the decision to refer for surgery, not simply the MR grade alone or with ejection fraction, as is a common practice. The American College of Cardiology/American Heart Association guidelines for referring a patient for mitral valve surgery call for chronic severe MR and mild-to-moderate LV dysfunction (4 cm, class I indication).14 Yet, our findings suggest that clinically important LA dilation may
1010
The American Journal of Cardiology (www.ajconline.org)
precede abnormal LV dimensions; thus, intervening before marked LA abnormality may help improve overall reverse remodeling postoperatively. Our data echo the findings of Shafii et al,15 who concluded “The best response occurs when surgery is performed before left heart dilation, LV hypertrophy, or LV dysfunction develop.” Such an approach may yield benefits in the immediate postoperative period by minimizing the risk of atrial fibrillation and strokes; elevated risks of each are associated with a dilated LA.16 The present study is subject to the limitations inherent in a single-center retrospective study design. Because our study excluded patients with reoperations or early mortality, it may not be generalizable to patients at high risk for each. Although our method retained as many eligible patients as possible, the differences inherent between our study cohort and patients who did not meet study criteria also affect interpretation of our results; we may overestimate the incidence of successful reverse remodeling in the larger population undergoing mitral valve repair. Follow-up echocardiography is not routine patient care; thus, patients with low clinical suspicion of disease or without symptomatology may be underrepresented, creating a bias overestimating abnormality. Serial follow-up echocardiographic data were not included in the study, thus constraining our ability to analyze remodeling as a timedependent factor. This potentially overestimates reverse remodeling at each time point, producing a right shift in the curves on Figure 2. Further study, including clinical studies with prospective designs, will be needed to confirm our findings. Future confirmation studies should record serial echocardiographic data to yield more precise estimates. Disclosures The authors have no conflicts of interest to disclose. 1. Carabello BA. The current therapy for mitral regurgitation. J Am Coll Cardiol 2008;52:319e326. 2. Carabello BA. The relationship of left ventricular geometry and hypertrophy to left ventricular function in valvular heart disease. J Heart Valve Dis 1995;4(Suppl 2):S132eS138. 3. Cho DK, Ha JW, Chang BC, Lee SH, Yoon SJ, Shim CY, Cho JR, Kim JS, Choi EY, Rim SJ, Chung N. Factors determining early left atrial reverse remodeling after mitral valve surgery. Am J Cardiol 2008;101:374e377. 4. Suri RM, Schaff HV, Dearani JA, Sundt TM 3rd, Daly RC, Mullany CJ, Sarano ME, Orszulak TA. Determinants of early decline in ejection fraction after surgical correction of mitral regurgitation. J Thorac Cardiovasc Surg 2008;136:442e447.
5. Stevens LM, Basmadjian AJ, Bouchard D, El-Hamamsy I, Demers P, Carrier M, Perrault LP, Cartier R, Pellerin M. Late echocardiographic and clinical outcomes after mitral valve repair for degenerative disease. J Card Surg 2010;25:9e15. 6. Ashikhmina EA, Schaff HV, Suri RM, Enriquez-Sarano M, Abel MD. Left ventricular remodeling early after correction of mitral regurgitation: maintenance of stroke volume with decreased systolic indexes. J Thorac Cardiovasc Surg 2010;140:1300e1305. 7. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ, Chamber Quantification Writing Group, American Society of Echocardiography’s Guidelines and Standards Committee, European Association of Echocardiography. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiog 2005;18:1440e1463. 8. Rubin DB. Multiple imputation after 18þ years. J Am Stat Assoc 1996;91:473e489. 9. Little RJA. Regression with missing Xs—a review. J Am Stat Assoc 1992;87:1227e1237. 10. Moons KGM, Donders RART, Stijnen T, Harrell JFE. Using the outcome for imputation of missing predictor values was preferred. J Clin Epidemiol 2006;59:1092e1101. 11. Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, Wood AM, Carpenter JR. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 2009;338:b2393. 12. Antonini-Canterin F, Beladan CC, Popescu BA, Ginghina C, Popescu AC, Piazza R, Leiballi E, Zingone B, Nicolosi GL. Left atrial remodelling early after mitral valve repair for degenerative mitral regurgitation. Heart 2008;94:759e764. 13. Braunwald E. Mitral regurgitation: physiologic, clinical and surgical considerations. N Engl J Med 1969;281:425e433. 14. Bonow RO, Carabello BA, Chatterjee K, de Leon AC Jr, Faxon DP, Freed MD, Gaasch WH, Lytle BW, Nishimura RA, O’Gara PT, O’Rourke RA, Otto CM, Shah PM, Shanewise JS, Smith SC Jr, Jacobs AK, Adams CD, Anderson JL, Antman EM, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Lytle BW, Nishimura R, Page RL, Riegel B. ACC/AHA 2006 guidelines for the management of patients with valvular heart disease a report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines (writing committee to revise the 1998 guidelines for the management of patients with valvular heart disease) developed in collaboration with the Society of Cardiovascular Anesthesiologists endorsed by the Society for Cardiovascular Angiography and Interventions and the Society of Thoracic Surgeons. J Am Coll Cardiol 2006;48:1e148. 15. Shafii AE, Gillinov MA, Mihaljevic T, Stewart W, Batizy LH, Blackstone EH. Changes in left ventricular morphology and function after mitral valve surgery. Am J Cardiol 2012;110:403e408. 16. Abhayaratna WP, Seward JB, Appleton CP, Douglas PS, Oh JK, Tajik AJ, Tsang TSM. Left atrial size. Physiologic determinants and clinical applications. J Am Coll Cardiol 2006;47:2357e2363.
The regurgitant jet of blood from mitral regurgitation (MR) overloads the left atrium (LA) and the left ventricle (LV).1 LV loading is critical in understanding the remodeling process.1 The remodeling due to MR is different from that seen in other left-sided valve diseases such as mitral stenosis, aortic stenosis, or aortic regurgitation in that it manifests with higher LV radius/thickness ratios and lower mass/volume ratios.2 Reverse LA and LV remodeling may happen early after surgical correction of MR,3e6 but longstanding MR left untreated leads to deterioration of LV function and impacts the potential for reverse remodeling. Thus, the degree of disease progression is an important consideration when weighing the value of surgical intervention. The present study evaluates long-term outcomes in a series of patients who underwent mitral valve repair for MR due to degenerative disease (prolapse). We sought to a Division of Cardiac Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts and bDivision of Cardiothoracic, Transplantation and Vascular Surgery, Hannover Medical School, Hannover, Germany. Manuscript received September 17, 2013; revised manuscript received and accepted December 4, 2013. See page 1010 for disclosure information. *Corresponding author: Tel: (617) 732-6569; fax: (617) 264-6369. E-mail address: [email protected] (L.H. Cohn).
0002-9149/14/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2013.12.009
investigate whether preoperative echocardiographic data or other factors may predict the degree of reverse remodeling seen at long-term follow-up. Methods With approval from our internal review board, we identified all 824 patients who underwent minimally invasive mitral valve repair from August 1996 to November 2010. Our inclusion criteria were patients with MR due to myxomatous mitral valve degeneration, normal coronary arteries, availability of a preoperative echocardiogram within 1 year of the surgery date, and a follow-up echocardiogram at least 6 months after surgery. Subjects with previous mitral valve surgery or mitral stenosis were excluded; 439 patients met our inclusion criteria. Clinical and surgical data were collected at the time of presentation, extracted from hospital electronic records or by chart review, and coded to Society of Thoracic Surgeons Adult Cardiac Database, version 2.52, specifications. Mortality data were collected during routine postoperative follow-up and supplemented by a query of the Social Security Death Index. Our main outcome of interest was reverse remodeling of the heart, defined as each of the following being within www.ajconline.org
Valvular Heart Disease/Preoperative Cardiac Dimensions Predict Reverse Remodeling
normal parameters: LV end-diastolic diameter, LV endsystolic diameter, and LA diameter. These were collected by a manual review of the echocardiographic report. All echocardiographies were performed transthoracically. The preoperative echocardiography was the study most proximate to the surgery. If multiple echocardiographic studies existed, the latest normal or the earliest abnormal study was used. Only 1 preoperative and 1 postoperative echocardiogram for each patient was collected in the research data file. Grade of MR was assessed using color-flow Doppler as none or trivial (0), trace (1þ), mild (2þ), moderate (3þ), and severe (4þ). We grouped patients based on the preoperative severity of abnormality of heart dimensions. For LV end-systolic diameter, the grouping was based on a preoperative value of 4 cm (normal) or >4 cm (abnormal). LV end-diastolic diameter and LA groupings were derived from the American Society of Echocardiography’s recommendations for chamber quantification.7 For men, normal LV end-diastolic diameter, 4.2 to 5.9 cm; mildly abnormal LV end-diastolic diameter, 6.0 to 6.3 cm; moderately abnormal LV end-diastolic diameter, 6.4 to 6.8 cm; severely abnormal LV end-diastolic diameter, 6.9 cm; normal LA, 3.0 to 4.0 cm; mildly abnormal LA, 4.1 to 4.6 cm; moderately abnormal LA, 4.7 to 5.2 cm; and severely abnormal LA, 5.3 cm. For women, normal LV end-diastolic diameter, 3.9 to 5.3 cm; mildly abnormal LV end-diastolic diameter, 5.4 to 5.7 cm; moderately abnormal LV end-diastolic diameter, 5.8 to 6.1 cm; severely abnormal LV end-diastolic diameter, 6.2 cm; normal LA, 2.7 to 3.8 cm; mildly abnormal LA, 3.9 to 4.2 cm; moderately abnormal LA, 4.3 to 4.6 cm; and severely abnormal LA, 4.7 cm. Successful postoperative remodeling for each of the heart dimensions was defined as measurements equal or the upper normal value; for men, LV end-diastolic diameter 5.9 cm and LA 4.0 cm, and for women, LV enddiastolic diameter 5.3 cm and LA 3.8 cm. Normal postoperative LV end-systolic diameter for all the samples was defined as 4.0 cm. Our secondary outcome of interest was survival over the study observation period. Because we excluded reoperations and early mortality, the end point for survival was the time in months from operation until (1) the first postoperative echocardiography indicating heart abnormality, (2) death, or (3) alive at the end of the study observation period, with normal postoperative echocardiographic data. The study period ended on June 30, 2011. Categorical data are expressed as numbers and percentages and were compared using Fisher’s exact test. Continuous variables are presented as mean and SD or median and interquartile range for normally or nonnormally distributed continuous data, respectively, and were compared using t test with Levine test for homogeneity of variance or MannWhitney U test, as appropriate. Multivariate analysis of survival with complete reverse heart remodeling was estimated using a forward stepwise Cox regression model. An event was the first of either an echocardiogram showing any dimension abnormal or report of death. Tested variables were derived from available literature, clinical expertise, and variables that differed between groups on exploratory univariate analysis or could reasonably be expected to impact our outcomes. Interaction terms were examined, and colinearity diagnostics were
1007
Table 1 Characteristics and operative and in-hospital outcomes of 439 patients with mitral valve repair Characteristic Preoperative characteristic Age (yrs) Age 65 yrs Women Body surface area New York Heart Association class III or IV Diabetes mellitus Chronic obstructive pulmonary disease Renal failure Atrial fibrillation Echo data 3þ MR LV diameter/diastole LV diameter/systole LA dimension Ejection fraction (%) Operative data No ring used Ring size Bypass time (minutes) Crossclamp time (minutes) Postoperative outcome Transfused with red cells Units per transfused patient Renal failure New-onset renal failure Atrial fibrillation New-onset atrial fibrillation Died during follow-up Follow-up time (mo)
Study Cohort, n ¼ 439 Mean n n Mean n n n n n
57.0 12.1 121 (27.6) 164 (37.4) 1.9 0.2 88 (20.0) 8 (1.8) 26 (5.9) 6 (1.4) 35 (8.0)
n Mean Median Mean Median Mean Median Median
438 (99.8) 5.77 0.7 5.73 (5.4e6.1) 3.52 0.6 3.50 (3.2e3.9) 4.69 0.7 4.60 (4.2e5.0) 60 (55e65)
n Median Median Median
5 34 115 81
(1.1) (26e38) (98e139) (69e98)
n Median n n n n n Median
130 2.00 27 22 137 110 20 81
(29.6) (1e3) (6.2) (5.0) (31.2) (25.1) (4.6) (53e116)
Number of cases presented with the percentage in parentheses. Means are presented with SDs and medians with interquartile ranges, except for ring size, which is a range.
evaluated. Statistical analyses were performed with SPSS, version 13.0 (SPSS Inc., Chicago, Illinois). Many cases had individual measures missing at random. Because deleting patients with missing values introduces biases that may not be identified or adequately controlled and affects generalizability of findings, we decided to impute missing data rather than conduct a complete-case analysis.8,9 Preoperative LV end-diastolic diameter was missing for 89 patients (18%), LV end-systolic diameter missing for 107 (22%), and LA missing in 105 patients (21%). All patients had at least 2 of these data points present. Missing data were estimated using multiply imputed linear regression approach with random-error component; for each imputed measure, a minimum of 5 data sets (maximum of 20) were created until stability in the standard errors was achieved. The regression models included all covariates, including outcome variables.10,11 Compared with known data, the correlation for imputed LV end-systolic diameter was R2 ¼ 0.988, LV enddiastolic diameter was R2 ¼ 0.988, and LA R2 ¼ 0.993. Imputed measures were only used when that variable was considered a covariate for that analysis. Outcome measures
1008
The American Journal of Cardiology (www.ajconline.org) Table 2 Reverse remodeling across heart dimensions at follow-up compared with baseline echocardiography Preoperative Category
Patients (n)
Absolute % Reduction, Mean SD
Patients With Normal Dimensions at Follow-Up, n (%)
LV end-diastolic dimension at follow-up LV end-diastolic enlargement Normal Mild Moderate Severe
Figure 1. Scatter plot of preoperative LV dimensions in diastole and systole, grouped by LA dimensions. The vertical line denotes the boundary for abnormal LV end-systolic diameter. The Horizontal long-dash line is the threshold for abnormal LV end-diastolic diameter for women, the shortdash line is the threshold for men. Abnormal thresholds are the lower boundaries of moderate enlargement.
were present for all cases, except in 21 patients (<5%) missing postoperative LA diameter. Missing outcome variables were not imputed. Results The baseline characteristics of our study cohort are presented in Table 1. Mean age at surgery was 57 12.1 years, and 28% were 65 years; 164 of 439 (40%) were women, and 20% were in New York Heart Association class III or IV. At baseline, the median LV end-diastolic diameter was 5.73 cm (5.4 cm to 6.1 cm) and LA was 4.60 cm (4.2 cm to 5.0 cm). Over the study period, there were 20 late deaths. Median survival time was estimated at 14.3 years (95% confidence interval [CI] 13.9 to 14.5). Concomitant procedures included patent foramen ovale closure (24 of 439, 5%), atrial septal defect closure (4 of 439, 1%), and tricuspid valve repair (14 of 439, 3%). Five patients had concomitant Maze. Alt. Cox maze III. procedures. A lower ministernotomy was done for 86% of patients (380 of 439); 49 (11%) had a right parasternal incision, 8 (2%) a right thoracotomy, and 2 (0.5%) an upper ministernotomy. In most cases, a posterior leaflet resection or modified resection and sliding annuloplasty (332 of 439, 76%) was performed and an annuloplasty ring implanted (434 of 439, 99%). Of patients with ring annuloplasty, 417 of 439 (95%) received a Cosgrove-Edwards Ring (Edwards Lifesciences, Irvine, California). Table 1 presents the operative and in-hospital outcomes for our cohort. We assigned our patients into groupings based on their preoperative heart dimensions. For LV end-diastolic measures, our a priori criteria were that anything greater than and including the lower range of moderate enlargement (for men, 6.4 cm and for women, 5.8 cm) was considered abnormal enlargement and dimensions below that were normal. For LV
209 12.5 11.5 203 (97.1) 117 17.4 9.7 111 (94.9) 80 21.1 9.5 73 (91.3) 33 25.2 8.3 28 (84.8) LV end-systolic dimension at follow-up
LV end-systolic enlargement Normal Abnormal
370 69
LA dilation Normal Mild Moderate Severe
57 108 121 132
6.5 20.9 352 (95.1) 19.1 16.7 55 (79.7) LA diameter at follow-up 4.8 6.1 9.3 16.2
19.7 15.3 14.1 13.5
41 63 36 24
(71.9) (58.3) (29.8) (18.2)
end-systolic measures, any value >4 cm was abnormal and 4 normal. For LA, we used the lower range of moderate enlargement (4.7 cm for men and 4.3 cm for women) as the threshold for abnormal, whereas measurements below that were considered normal. At the time of surgery, 74% (326 of 439) of patients met normal LV end-diastolic criteria, 84% (370 of 439) met normal LV end-systolic criteria, and 39% (170 of 439) met normal LA criteria. Figure 1 shows the relation between the preoperative LV end-systolic iameter, LV end-diastolic diameter and LA dimensions. The correlation between LV end-diastolic and end-systolic diameters is very high (R2 ¼ 0.471, p <0.001). However, the correlation between either and LA dimensions is R2 0.08. As can be seen in Figure 1, most patients with abnormal LA dimensions presented with normal or mild enlargement of the LV end-diastolic diameter (68%, 181 of 266) and LV end-systolic diameter (83%, 221 of 266). Overall, 95% of patients had normal LV end-diastolic diameter at follow-up, 93% had normal LV end-systolic diameter, but only 37% exhibited normal LA size, and 35% were normal on all measures. Table 2 presents the degree of reverse remodeling and the percentage of patients with dilatation at follow-up, within each category of preoperative dilatation. The trend noted for each of LV end-diastolic diameter, LV end-systolic diameter, and LA was that for increasing degrees of preoperative dilatation, the degree of reverse remodeling, as indicated by the percent difference of the postoperative measurements from the baseline measurements, was likewise increased. However, the percentage of patients who met normal criteria in each group decreased; patients having severe preoperative LV end-diastolic diameter enlargement were less likely to normalize LV enddiastolic diameter postoperatively than the patients with mild enlargement (odds ratio [OR] 0.34, 95% CI 0.11 to 0.99). Regarding LA measurements, 58% of those with mild
Valvular Heart Disease/Preoperative Cardiac Dimensions Predict Reverse Remodeling
1009
Table 3 Performance of the Cox regression analysis of survival with reverse heart remodeling Significant Predictors
Preoperative LA dilation (vs none) Mild Moderate Severe Age (in yrs, per yr) Preoperative LV end-systolic dimensions (continuous)
Model performance Figure 2. Cox proportional hazards regression analysis of survival with normal heart dimensions at echocardiographic follow-up, grouped by the degree of preoperative LA dilation. Overall model performance: 2 log likelihood ¼ 3,328.945, chi-square ¼ 54.464, df ¼ 5, p 0.001.
enlargement, 30% with moderate (OR 0.59, 95% CI 0.46 to 0.73), and 18% with severe enlargement (OR 0.51, 95% CI 0.40 to 0.65) did not have dilatation at follow-up. We used a Cox proportional hazards analysis to evaluate survival with complete heart reverse remodeling (Figure 2). Tested variables included gender, body surface area, preoperative ejection fraction and heart dimensions, pre- and postoperative atrial fibrillation, New York Heart Association class, chronic lung disease, renal failure, diabetes, etiology, and bypass time (Table 1). Significant predictors included moderate (OR 2.1, 95% CI 1.3 to 3.4) or severe preoperative LA dilatation (OR 2.7, 95% CI 1.7 to 4.4), abnormal preoperative LV end-systolic dimensions (OR 1.3, 95% CI 1.1 to 1.5), and age in years (OR 1.02, 95% CI 1.01 to 1.03) (Table 3). Discussion One of our observations was that the percentage of change in postoperative remodeling correlated to preoperative dimensions, with larger LV dimensions experiencing larger percent changes. Therefore, percent change may not be a fair marker for successful reverse remodeling, as has been done in previous studies.3,12 Instead, we chose to use whether postoperative measurements had returned to normal parameters as the measure of successful outcome. It is striking that although 95% and 93% of the patients had normal LV end-diastolic diameter and LV end-systolic diameter, respectively, at follow-up, only 37% of them had normal LA dimensions. The LA wall is thinner than the LV wall,13 which may make dilation less reversible. These factors explain our Figure 1, which indicates that meaningful LA dilation often occurs while ventricular dimensions remain normal, as well as our observation that 95% of those in the moderate LV end-diastolic enlargement group were normal at follow-up compared with only 30% of those with baseline moderate LA dilatation. On multivariate analysis, degree of LA dilation, LV endsystolic diameter, and age (in years) were predictors of reverse
p
Hazard Ratio
95% CI Lower Upper
0.001 0.532 0.002 0.001 0.002 0.009
1.180 2.115 2.734 1.017 1.264
0.703 1.305 1.698 1.006 1.054
2 Log Chi-Square df Likelihood 2,742.531 52.925 5
Noncontributing factors tested Diabetes mellitus (binary) Body surface area (continuous) Bypass time (continuous) Atrial fibrillation* (categorical) Chronic lung disease (binary) Preoperative renal failure (binary) Gender (binary) New York Heart Association class (categorical) Preoperative ejection fraction (%, continuous)
1.979 3.428 4.402 1.027 1.514 p 0.001 p 0.105 0.238 0.362 0.372 0.496 0.509 0.632 0.724 0.926
An event was either the earliest abnormal echocardiogram recorded or death. Time was calculated in months from the date of operation to the event or study follow-up if event free. Inclusion criterion was p 0.05; exclusion criterion was p >0.10. * Atrial fibrillation was a categorical value with 4 possible values: none, history of atrial fibrillation preoperatively, atrial fibrillation both preoperatively and postoperatively, and new-onset atrial fibrillation postoperatively.
remodeling at follow-up (Table 3 and Figure 2). This is consistent with the finding by Cho et al3 that age, sinus rhythm, predominant MR, and preoperative LA volume predicted early LA remodeling. Antonini-Canterin et al12 also identified preoperative LA volume along with blood pressure, postoperative mean mitral valve pressure gradient, and severity of residual MR as predictors of 15% postoperative reduction in LA dimension within 6 months of surgery. In our study, ejection fraction did not predict LA dimension at follow-up. In contrast, preoperative LV end-systolic diameter was predictive of both LV and LA reverse remodeling. This suggests that LV end-systolic diameter may be the better marker to decide whether to intervene with mitral valve repair. Our findings indicate that when MR remains untreated surgically until LV diameters are moderate to severely abnormal or LA moderately or severely dilated, successful reverse remodeling becomes less likely. It is thus important that all dimensions be taken into account along with the MR grade in the decision to refer for surgery, not simply the MR grade alone or with ejection fraction, as is a common practice. The American College of Cardiology/American Heart Association guidelines for referring a patient for mitral valve surgery call for chronic severe MR and mild-to-moderate LV dysfunction (4 cm, class I indication).14 Yet, our findings suggest that clinically important LA dilation may
1010
The American Journal of Cardiology (www.ajconline.org)
precede abnormal LV dimensions; thus, intervening before marked LA abnormality may help improve overall reverse remodeling postoperatively. Our data echo the findings of Shafii et al,15 who concluded “The best response occurs when surgery is performed before left heart dilation, LV hypertrophy, or LV dysfunction develop.” Such an approach may yield benefits in the immediate postoperative period by minimizing the risk of atrial fibrillation and strokes; elevated risks of each are associated with a dilated LA.16 The present study is subject to the limitations inherent in a single-center retrospective study design. Because our study excluded patients with reoperations or early mortality, it may not be generalizable to patients at high risk for each. Although our method retained as many eligible patients as possible, the differences inherent between our study cohort and patients who did not meet study criteria also affect interpretation of our results; we may overestimate the incidence of successful reverse remodeling in the larger population undergoing mitral valve repair. Follow-up echocardiography is not routine patient care; thus, patients with low clinical suspicion of disease or without symptomatology may be underrepresented, creating a bias overestimating abnormality. Serial follow-up echocardiographic data were not included in the study, thus constraining our ability to analyze remodeling as a timedependent factor. This potentially overestimates reverse remodeling at each time point, producing a right shift in the curves on Figure 2. Further study, including clinical studies with prospective designs, will be needed to confirm our findings. Future confirmation studies should record serial echocardiographic data to yield more precise estimates. Disclosures The authors have no conflicts of interest to disclose. 1. Carabello BA. The current therapy for mitral regurgitation. J Am Coll Cardiol 2008;52:319e326. 2. Carabello BA. The relationship of left ventricular geometry and hypertrophy to left ventricular function in valvular heart disease. J Heart Valve Dis 1995;4(Suppl 2):S132eS138. 3. Cho DK, Ha JW, Chang BC, Lee SH, Yoon SJ, Shim CY, Cho JR, Kim JS, Choi EY, Rim SJ, Chung N. Factors determining early left atrial reverse remodeling after mitral valve surgery. Am J Cardiol 2008;101:374e377. 4. Suri RM, Schaff HV, Dearani JA, Sundt TM 3rd, Daly RC, Mullany CJ, Sarano ME, Orszulak TA. Determinants of early decline in ejection fraction after surgical correction of mitral regurgitation. J Thorac Cardiovasc Surg 2008;136:442e447.
5. Stevens LM, Basmadjian AJ, Bouchard D, El-Hamamsy I, Demers P, Carrier M, Perrault LP, Cartier R, Pellerin M. Late echocardiographic and clinical outcomes after mitral valve repair for degenerative disease. J Card Surg 2010;25:9e15. 6. Ashikhmina EA, Schaff HV, Suri RM, Enriquez-Sarano M, Abel MD. Left ventricular remodeling early after correction of mitral regurgitation: maintenance of stroke volume with decreased systolic indexes. J Thorac Cardiovasc Surg 2010;140:1300e1305. 7. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ, Chamber Quantification Writing Group, American Society of Echocardiography’s Guidelines and Standards Committee, European Association of Echocardiography. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiog 2005;18:1440e1463. 8. Rubin DB. Multiple imputation after 18þ years. J Am Stat Assoc 1996;91:473e489. 9. Little RJA. Regression with missing Xs—a review. J Am Stat Assoc 1992;87:1227e1237. 10. Moons KGM, Donders RART, Stijnen T, Harrell JFE. Using the outcome for imputation of missing predictor values was preferred. J Clin Epidemiol 2006;59:1092e1101. 11. Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, Wood AM, Carpenter JR. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 2009;338:b2393. 12. Antonini-Canterin F, Beladan CC, Popescu BA, Ginghina C, Popescu AC, Piazza R, Leiballi E, Zingone B, Nicolosi GL. Left atrial remodelling early after mitral valve repair for degenerative mitral regurgitation. Heart 2008;94:759e764. 13. Braunwald E. Mitral regurgitation: physiologic, clinical and surgical considerations. N Engl J Med 1969;281:425e433. 14. Bonow RO, Carabello BA, Chatterjee K, de Leon AC Jr, Faxon DP, Freed MD, Gaasch WH, Lytle BW, Nishimura RA, O’Gara PT, O’Rourke RA, Otto CM, Shah PM, Shanewise JS, Smith SC Jr, Jacobs AK, Adams CD, Anderson JL, Antman EM, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Lytle BW, Nishimura R, Page RL, Riegel B. ACC/AHA 2006 guidelines for the management of patients with valvular heart disease a report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines (writing committee to revise the 1998 guidelines for the management of patients with valvular heart disease) developed in collaboration with the Society of Cardiovascular Anesthesiologists endorsed by the Society for Cardiovascular Angiography and Interventions and the Society of Thoracic Surgeons. J Am Coll Cardiol 2006;48:1e148. 15. Shafii AE, Gillinov MA, Mihaljevic T, Stewart W, Batizy LH, Blackstone EH. Changes in left ventricular morphology and function after mitral valve surgery. Am J Cardiol 2012;110:403e408. 16. Abhayaratna WP, Seward JB, Appleton CP, Douglas PS, Oh JK, Tajik AJ, Tsang TSM. Left atrial size. Physiologic determinants and clinical applications. J Am Coll Cardiol 2006;47:2357e2363.