Ann Thorac Surg 2008;85:2026-2029. doi:10.1016/j.athoracsur.2008.02.023
© 2008 The Society of Thoracic Surgeons
Original Articles: Adult Cardiac
Longitudinal Study of the Profile and Predictors of Left Ventricular Mass Regression After Stentless Aortic Valve Replacement
Eric Lim, FRCS (C-Th)a,
Ayyaz Ali, MRCSa,
Panagiotis Theodorou, MBBSa,
Ines Sousa, PhDb,
Hutan Ashrafian, MRCSa,
Themis Chamageorgakis, FRCSa,
Alison Duncan, PhDa,
Michael Henein, FRCPa,
Peter Diggle, PhDb,
John Pepper, FRCSa,*
a Departments of Cardiac Surgery and Echocardiography, Royal Brompton Hospital, London
b Department of Mathematics and Statistics, University of Lancaster, Lancaster, United Kingdom
Accepted for publication February 6, 2008.
* Address correspondence to Prof Pepper, Department of Cardiac Surgery, Royal Brompton Hospital, Sydney St, London SW3 6NP, United Kingdom (Email: m.shah{at}rbht.nhs.uk).
 |
Abstract
|
|---|
Background: The aim of this study was to evaluate the long-term profile and determine the factors that would influence the effect and rate of ventricular mass regression with time after aortic valve replacement with a stentless or a homograft valve.
Methods: We studied 300 patients during a 10-year period with at least a year of follow-up with a total of 1,273 serial echocardiographic measurements. Left ventricular mass was calculated from M-mode recordings and indexed to body surface area. Longitudinal data analysis was performed using a linear mixed effects model.
Results: The mean age (± standard deviation) was 65 (±14) years, consisting of 216 (72%) males. A stentless valve was implanted in 156 (52%), and a homograft in 144 (48%). The median time (interquartile range) to follow-up was 4.7 (2.8 to 6.6) years. The greatest rate of left ventricular mass regression occurred in the first year after surgery. On multivariable modeling, independent predictors of left ventricular mass were valve size (p = 0.011), left ventricular function (moderate impairment, p = 0.418; severe impairment, p = 0.011), and baseline left ventricular mass (middle tercile, p < 0.001; highest tercile, p < 0.001). Only baseline ventricular mass influenced the rate of subsequent left ventricular mass regression; the greatest rate of regression occurred in patients with the highest baseline values of ventricular mass (p < 0.001).
Conclusions: The greatest rate of left ventricular mass regression occurs in the first year with baseline left ventricular mass as the strongest predictor and the only identified variable that influenced the rate of left ventricular mass regression.
|
Introduction
|
|---|
Resolution of left ventricular hypertrophy is an important outcome of aortic valve surgery. Left ventricular mass regression has been associated with improved survival in patients with systemic hypertension. The assumption that left ventricular mass regression will also translate to clinical benefits after aortic valve replacement has generated interest in the use of stentless valves to facilitate this process [1], although this remains to be evaluated in a randomized setting.
To study the characteristics of left ventricular mass regression we reviewed a population of patients who underwent aortic valve replacement with either a porcine stentless valve or homograft (a human stentless valve). Our aims were to appraise the natural profile of left ventricular mass regression during the long term with serial echocardiographic measurements and to determine the factors that would influence the effect and rate of ventricular mass regression with time.
|
Material and Methods
|
|---|
The chairman of the ethics committee approved this study and waived the need for patient consent. This is a single surgeon's (J.P.) consecutive series of patients who underwent aortic valve replacement from 1991 to 2001. The Toronto Stentless Porcine Valve (St. Jude Medical, St. Paul, MN) was the predominant stentless valve in this series, followed by the Freestyle aortic bioprosthesis (Medtronics, Minneapolis, MN). Details of homograft cryopreservation the surgical technique for stentless valve and homograft implantation have been previously published [2].
Data Acquisition
We included all patients during a 10-year period with at least a year of follow-up with serial echocardiographic measurements. Patients with two or more procedures were censored from the time point of the second procedure to ensure that they were analyzed only once. Demographic, operative, and mortality data were obtained from individual hospital notes, death certificates, and autopsy reports.
Echocardiography
Echocardiography was performed routinely for patients on an annual basis. Serial echocardiographic reports were individually retrieved and reanalyzed for this study by one investigator (A.D). Transthoracic echocardiography was performed using the Hewlett Packard Sonos 5500, with a multifrequency transducer and harmonic imaging as appropriate. Two-dimensional echocardiography was performed from the parasternal long-axis and short-axis views and apical four-chamber and two-chamber views. Cross-sectional two-dimensionally guided M-mode recordings were performed using the left parasternal long-axis view with the cursor at the tips of the mitral valve leaflets. Left ventricular minor axis dimensions were measured at end diastole (the onset of the QRS complex) and at end systole (the first high-frequency vibration of the aortic component of the second heart sound on the phonocardiogram, A2, confirmed as synchronous with the onset of the closure artifact on the aortic Doppler record). The M-mode measurements included end-diastolic left ventricular internal diameter, interventricular septal thickness, posterior wall thickness, and end-systolic internal diameter. All measurements were performed using the American Society of Echocardiography (ASE) leading-edge convention [3], and left ventricular mass was derived using the formula described by Devereux and colleagues [4] as follows:
|
where LV mass is left ventricular mass in grams, IVSd is end-diastolic interventricular septum, LVIDd is end-diastolic left ventricular internal diameter, PWTd is end-diastolic posterior wall thickness; left ventricular mass was indexed to body surface area.
Statistical Methods
Baseline characteristics were presented as mean with standard deviation or median with interquartile range for normally and nonnormally distributed measures, respectively. Univariable and multivariable longitudinal data analysis was performed using a linear mixed effects model with random coefficients and an exponential correlation structure [5, 6]. Statistical analyses were performed on S Plus version 6.0 (Insightful, Seattle, WA).
|
Results
|
|---|
We identified 289 patients from January 1, 1991, to January 1, 2001, who underwent aortic valve replacement using either a stentless valve or a homograft with serial echocardiographic data from more than a year of follow-up.
The mean age of our study population was 66 (13) years, consisting of 205 (71%) males. A stentless valve was implanted in 148 (51%), and a homograft was implanted in 141 (49%) patients. In total, 177 (62%) patients were operated on for aortic stenosis, 56 (19%) for aortic regurgitation, and 56 (19%) for mixed aortic valve disease. A summary of patient demographics and operative procedures is presented in Table 1.
The median time to follow-up was 4.0 (1.8 to 6.2) years. The greatest rate of left ventricular mass regression occurred within the first year after surgery; the mean profile is illustrated in Figure 1, where the dots represent yearly averages of data.
View larger version (12K):
[in this window]
[in a new window]
|
Fig 1. Mean left ventricular mass index (LVMI; dashed lines are 95% confidence intervals) as a function of time. The regression model is 4.30 – 0.17 x t for times t up to 1 year, and 4.12 + 0.008 x t for times t after 1 year from surgery, both with plus covariate effects, –0.13 x I(female) + 0.03 x valve size + 0.04 x I(regurgitation) + 0.06 x I(mixed aortic), where I(n) is an indicator function. The fitted model is for a man with a valve size 24.6 mm and preoperative aortic stenosis.
|
|
On multivariable modeling (Table 2), the independent predictors of left ventricular mass were valve size (p = 0.011), left ventricular function (moderate impairment, p = 0.418; severe impairment, p = 0.011; Fig 2), and baseline left ventricular mass (middle tercile, p < 0.001; highest tercile, p < 0.001; Fig 3). The initial highly significant probability value for female sex on univariable analysis was reduced after adjustment for baseline ventricular mass (p = 0.073; Fig 4).
View larger version (5K):
[in this window]
[in a new window]
|
Fig 2. Left ventricular mass index (LVMI) as a function of time by ventricular function. The baseline variables imputed to produce this plot were male sex, a valve size of 23 mm, and a middle tercile for baseline left ventricular mass index. Normal ventricular function is shown by the solid line, moderately impaired function is shown by the dotted line, and poor function is shown by the dashed line.
|
|
View larger version (5K):
[in this window]
[in a new window]
|
Fig 3. Left ventricular mass index (LVMI) as a function of time by baseline ventricular mass. The baseline variables imputed to produce this plot were male sex, a valve size of 23 mm, and good ventricular function. The lowest tercile for baseline ventricular mass is shown by the solid line, the middle tercile for baseline ventricular mass is shown by the dotted line, and the highest tercile for baseline ventricular mass is shown by the dashed line.
|
|
View larger version (4K):
[in this window]
[in a new window]
|
Fig 4. Left ventricular mass index (LVMI) as a function of time by sex. The baseline variables imputed to produce this plot were a valve size of 23 mm, good ventricular function, and a middle tercile for baseline left ventricular mass index. Women are shown by the solid line, and men are shown by the dotted line.
|
|
Only baseline ventricular mass influenced the rate of subsequent left ventricular mass regression. The greatest rate of regression occurred in patients with the highest baseline values of ventricular mass; the probability value associated with the interaction term with time was p = 0.046 for the middle tercile and p < 0.001 for the highest tercile (compared with the lowest tercile) of baseline left ventricular mass index.
|
Comment
|
|---|
Although many studies have been performed to identify the predictors of left ventricular mass regression after aortic valve surgery, few have applied suitable statistical techniques. An important difference in the interpretation of multiple measurements with time (compared with conventional differences at two time points) is that significant predictors of ventricular mass are misinterpreted as predictors that influence the rate of ventricular mass regression.
When multiple measurements are performed with time, differences in the predictors of ventricular mass simply reflect baseline differences, resulting in parallel curves of the measured variable (eg, Fig 2). Factors that truly influence the rate of ventricular mass regression are those that have statistical evidence of an interaction with time, resulting in nonparallel curves (eg, Fig 3).
Predictors of Left Ventricular Mass
The impact of sex has been identified to influence the left ventricular mass regression in many studies [7, 8] , and our results reveal that this is mainly related to lower baseline left ventricular mass index in women, an effect that diminished after adjustment for baseline values of left ventricular mass.
In contradiction to previous reports [7–9] , we identified valve size as an important determinant of left ventricular mass, but increasing valve size was not associated with lower left ventricular mass index [10]. In fact, the opposite was true; a larger valve size was associated with greater left ventricular mass index, possibly because valve size is acting as a surrogate for overall heart size.
A consistent finding was the independent association between ventricular impairment and left ventricular mass. Severe impairment (but not moderate impairment) was associated with increasing left ventricular mass compared with patients with normal left ventricular function. There was no evidence to suggest that the severity of left ventricular impairment influenced the rate of ventricular mass regression, as it was not appreciably different in patients with differing degrees of ventricular impairment (Fig 2).
Predictors of the Rate of Left Ventricular Mass Regression
Baseline left ventricular mass index was the strongest determinant of postoperative left ventricular mass. Although this is consistent with the reports in the literature, our results add that baseline ventricular mass was the only identified variable that influenced the rate of left ventricular mass regression. The greatest rate of regression was observed in patients with the highest baseline left ventricular mass index (Fig 3).
Study Implications and the Profile of Ventricular Mass Regression
The results of our study may be useful to investigators who are planning clinical trials to evaluate left ventricular mass regression after stentless valve surgery, to ensure balance in the variables that affect left ventricular mass. Our results suggest that baseline left ventricular mass, valve size, female sex, and left ventricular function have the potential to influence comparisons among intervention groups. There was no evidence to suggest that the underlying cause for valve surgery (stenosis or regurgitation) affected baseline or follow-up left ventricular mass.
When planning the length of follow-up, the results of our analysis suggest that the greatest amount of mass regression occurs within the first year, but ventricular mass regression may continue (at a reduced rate) for up to 4 years. Left ventricular mass regression does not continue indefinitely; after 4 years there was evidence to suggest a slow gradual increase in left ventricular mass.
Potential Limitations
In our study, hypertension alone was not an important predictor of ventricular mass, but the results of the analysis with respect to hypertension were difficult to interpret because of the influence of time, treatment, and the severity of hypertension.
Although others have identified hypertension and valve failure as potential reasons for subsequent increases in left ventricular mass [11], our study was unable to discern whether this was the natural profile of left ventricular mass with time, or developed as a result of late complications. The latter seems unlikely as documented complications in this study were very low.
Conclusions
After aortic valve replacement with homografts and stentless valves, the greatest rate of left ventricular mass regression occurs within the first year. Thereafter the rate of regression reduces for up to 4 years followed by a gradual increase in left ventricular mass. Baseline left ventricular mass was the strongest predictor of subsequent left ventricular mass index and the only identified variable that influenced the rate of left ventricular mass regression.
|
Acknowledgments
|
|---|
The authors gratefully acknowledge the assistance of Alison Duncan and Carin Porter for echocardiographic analyses. This work has been supported by a grant from the Royal Brompton and Harefield NHS Trust Clinical Research Committee.
|
References
|
|---|
- Del Rizzo DF, Abdoh A, Cartier P, Doty D, Westaby S. The effect of prosthetic valve type on survival after aortic valve surgery Semin Thorac Cardiovasc Surg 1999;11:1-8.[Medline]
- Ali A, Lim E, Halstead J, et al. Porcine or human stentless valves for aortic valve replacement?. Results of a 10-year comparative study. J Heart Valve Dis 2003;12:430-435.[Medline]
- Sahn DJ, DeMaria A, Kisslo J, Weyman A. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements Circulation 1978;58:1072-1083.[Abstract/Free Full Text]
- Devereux R, Reichek N. Echocardiographic determination of left ventricular mass in man. Anatomic validation of the method. Circulation 1977;55:613-618.[Abstract/Free Full Text]
- Diggle PJ, Heagerty P, Liang KY, Zeger SL. Analysis of longitudinal data2nd ed.. Oxford: Oxford University Press; 2002.
- Blackstone E. Breaking down barriers: helpful breakthrough statistical methods you need to understand better J Thorac Cardiovasc Surg 2001;122:430-439.[Free Full Text]
- Del Rizzo DF, Abdoh A, Cartier P, Doty D, Westaby S. Factors affecting left ventricular mass regression after aortic valve replacement with stentless valves Semin Thorac Cardiovasc Surg 1999;11:114-120.[Medline]
- Gelsomino S, Frassani R, Morocutti G, et al. Time course of left ventricular remodeling after stentless aortic valve replacement Am Heart J 2001;142:556-562.[Medline]
- Christakis GT, Joyner CD, Morgan CD, et al. Left ventricular mass regression early after aortic valve replacement Ann Thorac Surg 1996;62:1084-1089.[Abstract/Free Full Text]
- Maselli D, Pizio R, Bruno LP, Di Bella I, De Gasperis C. Left ventricular mass reduction after aortic valve replacement: homografts, stentless and stented valves Ann Thorac Surg 1999;67:966-971.[Abstract/Free Full Text]
- Lund O, Emmertsen K, Dorup I, Jensen FT, Flo C. Regression of left ventricular hypertrophy during 10 years after valve replacement for aortic stenosis is related to the preoperative risk profile Eur Heart J 2003;24:1437-1446.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
R. M. Suri, K. J. Zehr, T. M. Sundt III, J. A. Dearani, R. C. Daly, J. K. Oh, and H. V. Schaff
Left ventricular mass regression after porcine versus bovine aortic valve replacement: a randomized comparison.
Ann. Thorac. Surg.,
October 1, 2009;
88(4):
1232 - 1237.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
