Author + information
- Received April 15, 2014
- Revision received May 21, 2014
- Accepted June 12, 2014
- Published online November 1, 2014.
- George Petrov, MD, MSc∗,†∗ (, )
- Elke Dworatzek, PhD†,‡,
- Tabea Marie Schulze, MD (cand)∗,
- Michael Dandel, MD∗,†,
- Georgios Kararigas, PhD†,‡,
- Shokufeh Mahmoodzadeh, PhD†,‡,
- Christoph Knosalla, MD∗,†,
- Roland Hetzer, MD∗,† and
- Vera Regitz-Zagrosek, MD∗,†,‡
- ∗Department of Cardiothoracic and Vascular Surgery, Deutsches Herzzentrum Berlin, Berlin, Germany
- †DZHK (German Centre for Cardiovascular Research), partner site Berlin, Berlin, Germany
- ‡Institute of Gender in Medicine and Center for Cardiovascular Research, Charité Universitaetsmedizin Berlin, Berlin, Germany
- ↵∗Reprint requests and correspondence:
Dr. George Petrov, Department of Cardiothoracic and Vascular Surgery, Deutsches Herzzentrum Berlin, DHZB, Augustenburger Platz 1, 13353 Berlin, Germany.
Objectives The purpose of this study was to test whether adaptive or maladaptive remodeling is associated with survival in women and men after aortic valve replacement (AVR).
Background Women with isolated aortic valve stenosis (AS) develop more concentric left ventricular hypertrophy (LVH) than men in similar disease states. We recently reported less up-regulation of profibrotic genes at AVR and faster LVH regression post-operatively in women than in men, suggesting that there are sex differences in the adaptation to pressure overload and its regression.
Methods The study cohort included 128 patients (age 70.0 ± 9.6 years, 49% women) undergoing AVR for AS. Echocardiography was obtained before and 4.0 ± 1.6 years after surgery. Factor analysis was used to classify LVH as adaptive (combining smaller left ventricular [LV] mass/diameters and greater relative wall thicknesses) or maladaptive. Myocardial tissue samples from the LV septum were obtained during AVR to analyze cardiac fibrosis and associated key molecular regulators.
Results Before AVR, LVH was classified as adaptive in 62% of women and 45% of men (p < 0.050). Four years after AVR, adaptive LVH was observed in 75% of women and 49% of men (p < 0.031). At surgery, more cardiac fibrosis was present in men compared with women (p < 0.05). Higher levels of transforming growth factor beta 1 (p < 0.01), SMAD2 phosphorylation (p < 0.001), and periostin expression (p < 0.05) were found in men than in women. Women with maladaptive LVH had worse survival than women with adaptive LVH (p < 0.050), whereas the pattern of LVH did not affect survival in men (p < 0.307).
Conclusions Women more frequently exhibit adaptive LV remodeling with less fibrosis than men. Maladaptive LVH is associated with worse survival in women. Thus, sex should be considered as a strong modulating factor when management of patients with AS is discussed.
- aortic valve replacement
- aortic valve stenosis
- cardiac fibrosis
- hypertrophy regression
Aortic stenosis (AS) is a common valvular disease that affects both sexes. More than 12,000 patients undergo aortic valve replacement (AVR) in Germany per year; at least 70% of them are age ≥60 years (1). In AS, left ventricular hypertrophy (LVH) develops as a response to pressure overload (PO). It is considered primarily a compensatory response to reduce wall stress and to maintain systolic function of the heart. However, pressure-induced LVH also initiates a series of molecular events that may lead to a vicious circle and finally to myocardial fibrosis, dilation, and heart failure (2). LVH is not uniform in patients with a similar degree of AS, and its regression after surgical correction is variable. Genetic predisposition of the patient probably interferes with hemodynamic load to determine the pattern of LVH, its regression, and clinical course after surgery (3,4).
Sex has been described to have a major impact on the development of LVH in pressure overload (5,6). Women with AS develop a more concentric form of LVH with smaller ventricular diameters, greater relative wall thicknesses, and a better systolic function, whereas men with comparable degree of pressure overload are characterized by a more eccentric form of LVH with more ventricular dilation and fibrosis, decreased ejection fraction, and heart failure (7). Our own previous investigations in patients with AS showed at the time of AVR higher messenger ribonucleic acid expression of fibrosis-associated genes, that is, collagen I and III, MMP-2, and MMP-9, in men compared with women (8).
Relief of mechanical obstruction by AVR leads to hemodynamic improvement and reversal of LVH. Sufficient evidence exists that LVH regression not only is a passive process, triggered by reduction in afterload, but also requires an activation of a specific gene expression program (9). Regression of LVH starts early after AVR but may go on for years (10). Even though the time course of hypertrophy regression has been extensively studied, few studies have considered the influence of sex on the post-operative reversibility of LVH, and the overall evidence of systematic sex differences remains limited (8,11). We recently demonstrated that early LVH regression after AVR is more pronounced in women than in men (8).
Survival after AVR depends on a number of clinical variables, including pattern of pre-operative LVH, severity of myocardial dysfunction, and cardiac fibrosis (12). There are only few studies analyzing sex differences in the post-operative survival after AVR (13–15). Linking sex differences in pre-operative LVH to regression and survival after AVR may improve our understanding of their relevance and offer new hints for optimal timing of interventions. We, therefore, continued our previous investigations on sex-specific LVH regression after AVR (8). The present study was undertaken to assess whether the pattern of myocardial adaptation in women and men with AS is associated with long-term survival after AVR.
We included patients who were hospitalized at the German Heart Institute Berlin for AVR due to isolated AS. Exclusion criteria were emergency surgery, second AVR, transcatheter aortic valve implantation, left ventricular (LV) ejection fraction ≤40%, coronary artery narrowing ≥50%, coincident significant congenital or acquired valvular defects other than AS, pre-existing inflammatory heart diseases, cardiomyopathies, uncontrolled hypertension (blood pressure ≥160/≥100 mm Hg) or other extracardiac comorbidities (i.e., human immunodeficiency virus, metabolic disorders, malignancies, and so on), or language problems that might affect the clinical outcome or limit informed consent. The study was started in 2005; patients were followed prospectively and invited between 2011 and 2012 for follow-up echocardiography. The initial response rate was 83%. In follow-up responders, reasons for refusal to participate in echocardiography included residency abroad or in other federal states of Germany (n = 10), physical invalidity (n = 5), death before follow-up was conducted (n = 19), or refusal without indication of reasons (n = 4). Cardiac morphology and function were assessed about a week before AVR and 4.0 ± 1.6 years post-operatively, following the recommendations of the American Society of Echocardiography (16). Follow-up echocardiography was obtained in a representative subgroup of 68 survivors (43% women; p < 0.097 for sex difference; see Online Table S1-A for patients’ characteristics), and the cumulative hazard to have an echocardiography at follow-up did not vary between women and men (p < 0.117). Written informed consent was obtained from all patients, and the study was approved by the ethical committee of Charité Universitaetsmedizin Berlin.
Mean overall follow-up was 3.2 years in women (95% confidence interval [CI]: 2.7 to 3.7 years) and 2.9 years in men (95% CI: 2.5 to 3.4 years, p < 0.362). Primary clinical endpoints were LVH regression and death. Before the study database was closed in December 2012, a query to all local residents’ registration offices (all residents in Germany are registered) was performed to assign unknown deaths.
Myocardial tissue samples were harvested from the LV septum during AVR in a representative subgroup of 43 AS patients (49% women; see Online Table S1-B for patients’ characteristics) and compared with LV septum biopsies from nondiseased donor hearts (n = 23; 52% women, see Online Table S2 for nondiseased subjects’ characteristics) that could not be used for transplantation because of age or logistic problems with the recipients. Fibrosis was quantitated by histology; transforming growth factor beta-1 (TGF-β1), SMAD family member 2 or 3 (SMAD2/3), and periostin protein expression by western blotting (see the Online Appendix for details). Detailed methodology has been described previously (17,18).
Statistics were performed using SPSS Statistics, version 21 (IBM, Armonk, New York). All of the data in the tables are presented as mean ± SD or percentages unless otherwise indicated. Sex comparisons were performed using unpaired Student t test or Mann-Whitney U test as appropriate. Categorical variables were tested using Pearson chi-square test. Two-way analysis of variance with Bonferroni post hoc testing was used to assess sex differences in cardiac fibrosis and associated key molecular regulators of patients with AS versus samples from nondiseased hearts. Kaplan-Meier and Cox models were used to analyze the survival data. Comparisons of survival curves were on the basis of the Breslow test. A value of p < 0.05 was considered statistically significant.
Factor analysis including LV mass, left ventricular internal diameter at end-diastole (LVID[d]), relative wall thickness (RWT), transvalvular or transprosthetic mean pressure gradient (Pmean) and peak pressure gradient (Ppeak), as well as body size and mass was used to analyze LVH under conditions of chronic PO and its regression after AVR. The factor analysis identified 1 multidimensional index for LVH (LVH index), 1 for PO (PO index), and 1 for body size and mass. The LVH index was correlated with LV mass (factor loading 0.676) and LVID(d) (factor loading 0.937) and inversely correlated with RWT (factor loading −0.861). The PO index was correlated with Pmean (factor loading 0.971) and Ppeak (factor loading 0.970). The body size and mass index was correlated with height (factor loading 0.856) and weight (factor loading 0.839). Because the factor analysis was conducted as a principal component analysis, the LVH, PO, and body size and mass indexes are orthogonal (uncorrelated). Thus, the observed variance in cardiac morphology that is included in the LVH index is statistically independent from differences in afterload, body size, and mass. We used the LVH index to classify LVH into adaptive or maladaptive patterns. Because the LVH index is a z-transformed variable, negative LVH index values reflect adaptive stages of LVH, with smaller LV mass and LVID(d) and greater RWT, whereas positive LVH index values reflect maladaptive stages of LVH with adverse alteration of LV morphology. The null value served as a parameterized reference to discriminate between adaptive and maladaptive LVH patterns. Kaiser-Guttman criterion was applied to ascertain the number of extracted factors (Online Tables S3-A and S3-B).
The study population comprised 128 patients (49% women). Women and men were at comparable age, had equal symptoms and New York Heart Association functional class at referral for AVR, and comparable cardiovascular risk profile as reflected by similar prevalence of diabetes mellitus, hypertension, and hyperlipidemia. Cardiac medications were equally administered in women and men. Coronary artery sclerosis, chronic obstructive pulmonary disease, or cerebrovascular accident were observed in a small portion of the study population, and there were no significant sex differences. Women compared with men had smaller body size and lower renal function. Mechanical and larger prosthetic valves were more frequently implanted in men than in women. Sex differences abolished after the effective orifice area of the implanted prosthetic valves was indexed to body surface area, and there was no evidence of significant sex differences in the occurrence of prosthesis-patient mismatch after AVR (Table 1).
At the time of AVR, myocardial fibrosis, quantified in biopsies from the LV septum, was significantly more pronounced in men with AS compared with nondiseased men and women with AS (Figures 1A and 1B). We found in men, but not in women with AS, a significant increase in TGFβ-1 protein expression (Figure 2A) and phosphorylation of SMAD2 (Figure 2B). Phosphorylation of SMAD3 was not altered in a sex-specific manner (Figure 2C). Periostin was significantly increased in both sexes with AS, but was more highly expressed in men compared with women with AS (Figure 2D).
Before AVR, LVH morphology differed between women and men with AS. Absolute LV mass and LVID(d) were significantly larger in men, but sex differences were reduced after adjustment for body surface area. RWT had the tendency to be higher in women, but the sex difference was of borderline statistical significance (Table 2). We, therefore, used factor analysis to combine single normalized parameters of LVH into a composed LVH index. On the basis of the LVH index, LVH was classified as adaptive or maladaptive (see Methods section for more details). Pre-operatively, adaptive LVH prevailed in women (women: 62% vs. men: 45%) and maladaptive LVH in men (women: 38% vs. men: 55%, p < 0.050 for sex difference). This was reflected by slightly greater pre-operative mean LVH indexes in men than in women (women: −0.08, 95% CI: −0.35 to 0.20 vs. men: 0.09, 95% CI: −0.15 to 0.34; p < 0.360) (Figure 3).
After AVR, LVH regression occurred in both sexes, but its pattern was different in women and men. Adaptive LVH was found after AVR more often in women (women: 75% vs. men: 49%), whereas maladaptive LVH characterized predominantly men (women: 25% vs. men: 51%, p < 0.031 for sex difference). The post-operative mean LVH index remained greater in men than in women (women: −0.44, 95% CI: −0.78 to −0.09 vs. men: 0.22, 95% CI: −0.12 to 0.57; p < 0.012) (Figure 3).
Post-AVR survival was similar in both sexes (overall mortality, women: 17% vs. men: 18%, p < 0.883). However, there was an interaction among sex, pattern of LVH, and survival (hazard ratio: 0.12, 95% CI: 0.02 to 0.89, p < 0.038). Women with maladaptive LVH at surgery had significantly worse survival than women with adaptive LVH (p < 0.050) (Figure 4B). In contrast, the pattern of LVH was not associated with survival in men (p < 0.307) (Figure 4A).
In aortic stenosis, maladaptive LV remodeling with eccentric LVH, more profibrotic gene expression, and fibrosis occurs with greater likelihood in men than in women with isolated AS (5–8,19). Our present study is well in agreement with these findings; however, it is the first study to link sex-specific patterns of PO-induced LVH to survival after AVR. An impaired survival was found in women with maladaptive LVH compared with women with adaptive LVH. In contrast, the pattern of LVH did not affect survival in men. This suggests that maladaptive LV remodeling has a stronger negative impact in women than in men.
Sex differences in the pre-operative LV remodeling have implications in the regression of LVH after AVR. We showed in a former phase of the current study that development of more favorable LVH in women was associated with faster regression in the early follow-up after AVR (8). We now go beyond this finding and confirm a more pronounced LVH regression in the long-term follow-up. Furthermore, our study provides evidence that sex differences occurring early after AVR are sustained in the late post-operative course, and thus it documents systematic sex differences in the time course of LVH regression after AVR.
We also show that higher cardiac fibrosis in men with AS compared with women is associated with increased TGF-β1 protein expression and phosphorylation of SMAD2. Periostin, a down-stream target of TGF-β1 signaling and a key regulator of cardiac fibrosis (20), was more highly expressed in men as well. In human failing hearts, up-regulated periostin expression is associated with cardiac fibrosis (21). Higher periostin levels were found in blood from men compared with women with heart failure (22). Furthermore, periostin has already been described as a modulator of cardiac remodeling under PO in mice (23). We now demonstrated sex differences in PO-induced up-regulation of periostin and speculate that lower expression of periostin in the female heart is associated with less cardiac fibrosis, tentatively linked to an adaptive LVH.
Analysis of sex differences in the development and regression of LVH requires the estimation of LV mass, ventricular diameters, wall thicknesses, and their combined effect. We applied in the current study a factor analysis to operationalize LVH as a combined endpoint of LV mass, LVID(d), and RWT changes. Although the factor analysis is commonly used in social sciences, it has only recently been used in biomedical research (24). Factor analysis has the advantages of combining multiple variables into a single index at the patient level, facilitating the use of this index in more complex statistical analyses, and addressing the behavior of complex biological processes better than the use of single variables. With regard to statistical performance, factor analysis of echocardiography data was as efficient as a conventional approach, which we applied earlier, to compare patients’ LV mass and RWT with prognostically-validated reference values from the general population (8). Notably, similar numbers of patients were needed to establish significant sex differences in the development and regression of LVH. Adequate model power beyond controlled study design and meticulous selection is an important prerequisite for the transferability of results.
Clinicians are well aware of strong interindividual differences in the amount and pattern of LVH. We now demonstrate that consideration of sex partially reduces this variability. Adaptive LVH is more common in women’s hearts, whereas maladaptive LVH and increased fibrosis are more characteristic of men’s hearts. This is well in agreement with some previous observations from us and others (7,8). We link higher TGF-β1/SMAD2 signaling, periostin expression, and higher cardiac fibrosis level in men tentatively with maladaptive remodeling. A better understanding of the regulatory mechanisms underlying this sex difference offers the potential to identify a protective pathway in the response of the heart to PO. This is of particular interest, because maladaptive LVH in women is associated with worse survival. Our data suggest that determining whether earlier intervention in women—before maladaptive remodeling occurs—is beneficial may be worth evaluating. Prospective and retrospective analysis of large databases may be first steps. The same hypotheses may be tested in men, but may require higher numbers for testing.
Alternatively, sex-specific evidence should be sought to determine whether women and/or men with maladaptive remodeling benefit post-operatively from substances that counteract profibrotic mechanisms, such as angiotensin-converting enzyme inhibitors or aldosterone antagonists.
This study has limited patient numbers and events, but the statistical impact on the power of the analysis is largely compensated by the duration of follow-up and meticulous patient characterization. A further limitation is the small number of patients at follow-up.
We show for the first time that pre-operative LVH is associated with survival in women. We also identified molecular mechanisms underlying sex differences in cardiac fibrosis that may contribute to sex-specific LV remodeling. The link between pattern of remodeling and survival raises the question of optimizing the management of AS in a sex-specific manner, and further evidence, obtained by analysis of suitable databases and well-designed clinical trials, is needed.
The authors thank Jenny Thomas and Karolin Duft for excellent technical help with protein expression analysis, all of the clinical collaborators in the German Heart Institute Berlin, and the patients in the study.
This work was supported by the German Research Foundation (DFG Re662/9-1, FG1054, and DW70/1-1) and DZHK (German Center for Cardiovascular Research). All authors have reported that they have no relationships relevant to the contents of this paper to disclose. Dr. Petrov, Dr. Dworatzek, and Ms. Schulze contributed equally to this work.
- Abbreviations and Acronyms
- aortic stenosis
- aortic valve replacement
- left ventricle
- left ventricular hypertrophy
- left ventricular internal diameter at end-diastole
- pressure overload
- relative wall thickness
- SMAD family member 2 or 3
- transforming growth factor beta 1
- Received April 15, 2014.
- Revision received May 21, 2014.
- Accepted June 12, 2014.
- American College of Cardiology Foundation
- Bruckenberger E.
- Kupari M.,
- Turto H.,
- Lommi J.
- Carroll J.D.,
- Carroll E.P.,
- Feldman T.,
- et al.
- Douglas P.S.,
- Otto C.M.,
- Mickel M.C.,
- Labovitz A.,
- Reid C.L.,
- Davis K.B.
- Villari B.,
- Campbell S.E.,
- Schneider J.,
- Vassalli G.,
- Chiariello M.,
- Hess O.M.
- Petrov G.,
- Regitz-Zagrosek V.,
- Lehmkuhl E.,
- et al.
- Villar A.V.,
- Merino D.,
- Wenner M.,
- et al.
- Hayashida K.,
- Morice M.C.,
- Chevalier B.,
- et al.
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.
- Fliegner D.,
- Schubert C.,
- Penkalla A.,
- et al.
- Kararigas G.,
- Bito V.,
- Tinel H.,
- et al.
- Snider P.,
- Standley K.N.,
- Wang J.,
- Azhar M.,
- Doetschman T.,
- Conway S.J.