Reverse left ventricular remodeling after aortic valve replacement for aortic stenosis: a systematic review and meta-analysis

Reverse left ventricular (LV) remodeling after aortic valve replacement (AVR), in patients with aortic stenosis, is well-documented as an important prognostic factor. With this systematic review and meta-analysis, we aimed to characterize the response of the unloaded LV after AVR. We searched on MEDLINE/PubMed and Web of Science for studies reporting echocardiographic findings before and at least 1 month after AVR for the treatment of aortic stenosis. In total, 1,836 studies were identified and 1,098 were screened for inclusion. The main factors of interest were structural and dynamic measures of the LV and aortic valve. We performed a random-effects meta-analysis to compute standardized mean differences (SMD) between follow-up and baseline values for each outcome. Twenty-seven studies met the eligibility criteria, yielding 11,751 patients. AVR resulted in reduced mean aortic gradient (SMD: −38.23 mmHg, 95% CI: −39.88 to −36.58, I2=92%), LV mass (SMD: −37.24 g, 95% CI: −49.31 to −25.18, I2=96%), end-diastolic LV diameter (SMD: −1.78 mm, 95% CI: −2.80 to −0.76, I2=96%), end-diastolic LV volume (SMD: −1.6 ml, 95% CI: −6.68 to 3.51, I2=91%), increased effective aortic valve area (SMD: 1.10 cm2, 95% CI: 1.01 to 1.20, I2=98%), and LV ejection fraction (SMD: 2.35%, 95% CI: 1.31 to 3.40%, I2=94.1%). Our results characterize the extent to which reverse remodeling is expected to occur after AVR. Notably, in our study, reverse remodeling was documented as soon as 1 month after AVR.


Introduction
Aortic stenosis (AS) is the most common acquired valvopathy in the Western world (1).Its incidence increases with age, and its prevalence is expected to rise in the future (2).
AS is not an isolated valve disease but a more complex and broad pathology involving the myocardium.AS progression is associated with left ventricular (LV) remodeling, which is the myocardial response to increased afterload (2).Initially, LV remodeling is a compensatory response to a persistent obstacle to systolic ejection.The sustained increased pressure and hemodynamic load lead to the classical development of LV hypertrophy.This initial adaptation allows for a reduction in wall stress and maintenance of cardiac output.After this stage, persistent obstruction leads to maladaptive LV remodeling, causing gradual deterioration of diastolic and systolic functions (1).Clinically, this process can translate into various symptoms, including death due to heart failure or arrhythmic events (2).In other words, maladaptive LV response negatively impacts the prognosis of AS patients regarding survival and cardiovascular events (3).
The only effective treatment for severe AS is aortic valve replacement (AVR), which can be performed either surgically (SAVR) or percutaneously via transcatheter AV implantation (TAVI).AVR aims to eliminate the LV obstruction and ultimately revert this inadequate LV response (2).After AVR, the extension of the achieved reverse LV remodeling is a major determinant of symptoms and outcomes (2).Its prognostic importance has been reported in several randomized trials (2,4,5).Transthoracic echocardiography (TTE) is the gold standard method to characterize AS severity, LV remodeling, and LV reverse remodeling after AVR.These LV adaptations comprise several changes in echocardiographic parameters, such as LV mass, cavity dimensions and volumes, wall thicknesses, and left ventricular ejection fraction (LVEF) (1).Unfortunately, data to predict LV response after AVR are lacking.
In this systematic review and meta-analysis, we aim to assess the extent of left ventricular remodeling at pre-determined time points post-procedure in patients with aortic stenosis who underwent AVR.The measured variables of interest included effective aortic valve area (AVA), mean aortic gradient (MAG), left ventricular mass (LVM), LVEF, and end-diastolic left ventricular diameter (EDLVD) and volume (EDLVV).

Eligibility and search strategy
This systematic review and meta-analysis was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement (6).
The literature search was conducted on 15 March 2022 in two electronic databases: MEDLINE (through PubMed) and Web of Science.The search was conducted with no restrictions on language or year of publication.Full details of the search are presented in Table 1.
Studies were included if they reported echocardiographic findings before and at least 1 month after SAVR or TAVI for the treatment of AS.This time interval was chosen to allow acute changes after the procedure to resolve and for reverse remodeling to occur (7).Furthermore, patient evaluation had to be performed at pre-determined time points post-procedure, i.e., at either 1, 3, 6, or 12 months.
Studies also needed to report at least one outcome variable of interest for the measurement of the left ventricle reverse remodeling to be included, namely, left ventricular dimensions or ejection fraction.
We excluded all non-human studies, case-control studies, case reports, and reviews.Studies without a predefined follow-up period and with fewer than 100 patients were also excluded.

Study selection, data collection process, and study outcomes
Two investigators (FSN and CAM) independently reviewed each study by title and abstract and then by full-text reading.

Risk of bias assessment
We used the Study Quality Assessment Tool for Observational Cohort Studies from the National Institutes of Health to categorize several domains for all the eligible studies.The overall risk of bias was independently assigned to each study by two investigators (FSN, CAM) and classified into "good," "fair," and "poor", as detailed in Table 3.

Statistical analysis
We performed a random-effects meta-analysis using the restricted maximum likelihood approach to compute pooled mean differences (MD) or standardized mean differences (SMD) between post-followup and baseline values for each outcome.Heterogeneity was assessed by the Cochran Q statistic p-value and the I 2 statistic: a p-value <0.10 and an I 2 >50% were considered to represent substantial heterogeneity.Sources of heterogeneity were explored using univariable meta-regression models, with tested covariates including the publication year, mean age of the participants, percentage of females, average BSA, percentage of patients in NYHA classes III/IV, and percentage of patients with other comorbidities such as hypertension, diabetes, and coronary heart disease.In addition, we performed subgroup analyses for the follow-up period and the initial LVEF (classes were categorized into two groups: lower than 50% and higher than 50%).All statistical analyses were performed using the meta package of R software (35,36).

Results
In total, 1,836 publications were identified through our search of MEDLINE/PubMed (944 records) and Web of Science (892 records) databases.After removing the duplicates, 1,098 records remained.Following the title and abstract screening, we selected 67 articles for full-text review.After excluding articles that did not meet the inclusion criteria, we ended up with 27 primary studies (see Figure 1 for the PRISMA 2020 flow diagram and Table 4 for a summary table of the included studies) .
Since some studies contained more than one distinct population, the search yielded 39 independent patient cohorts.The studies were published between 1998 and 2020, assessing 11,751 patients who completed echocardiographic assessment before and at least 1 month post-AVR.

Effective aortic valve area and mean aortic gradient
While this work is related to left ventricular remodeling after AVR, we chose to start by reporting measures related to AVR, such as aortic valve area and gradient.This ensures that the studies assessed comparable conditions and demonstrated similar improvements after valve obstruction is resolved.By doing so, we aimed to establish a consistent baseline for analyzing left ventricular remodeling parameters.
Our meta-analytical results indicate that, after AVR, there was an increase in the effective aortic valve area and a decrease in the mean aortic gradient.Based on 26 cohorts (n ¼ 6,726 at baseline, Figure 2), the pooled SMD for effective aortic valve area was 1.10 cm 2 (95% CI: 1.01-1.20,p , 0:0001, I 2 ¼ 98%, Cochran's Q p-value , 0:0001), corresponding to a significant increase after AVR, albeit with substantial heterogeneity.Univariate metaregression identified publication year, age, hypertension, NYHA class III or IV, DM, type of AVR, and EF >50% as potential moderators of heterogeneity (see Supplementary Table S1 for subgroup and heterogeneity analysis and Supplementary Table S2 for meta-regression).
The mean aortic gradient was assessed in 33 cohorts (n ¼ 10,480 patients at baseline, Figure 3).The pooled SMD for mean aortic gradient was À38:23 mmHg (95% CI: À39:88 to À36:58 mmHg, p , 0:0001, I 2 ¼ 92%, Cochran's Q p-value , 0:0001), indicating a significant decrease after AVR, but with substantial heterogeneity.Univariate meta-regression identified publication year and coronary artery disease as potential moderators of heterogeneity (see Supplementary Table S1 for subgroup and heterogeneity analysis and Supplementary Table S2 for meta-regression).Subgroup analyses showed a trend for differences according to follow-up periods (p ¼ 0:06; Figure 3) but not according to the type of AVR (p ¼ 0:16).

Parameters on left ventricular reverse remodeling 3.2.1 Left ventricular mass
LVM change after AVR was analyzed in 14 cohorts (Figure 4).The pooled SMD for LVM was À37:24 g (95% CI: À49:31 to À25:18, p , 0:0001; I 2 ¼ 96%, Cochran's Q p-value , 0:0001), indicating a significant decrease after AVR, albeit with substantial heterogeneity.Performing subgroup analysis according to follow-up periods, significant differences were observed (p ¼ 0:007).However, the values involved were relatively small (and may represent different samples evaluated at various time points and not a cohort evaluated prospectively through time): LVM reduction of 27 g at 1 month, 16 g at 3 months, 70 g at 6 months, and 34 g at 12 months.Performing subgroup analysis according to the type of AVR, no significant differences were observed (p ¼ 0:49).
Univariate meta-regression identified publication year and DM as potential moderators of heterogeneity (see Supplementary Table S1 for subgroup and heterogeneity analysis and Supplementary Table S2 for meta-regression).
Univariate meta-regression identified publication year and NYHA classification III or IV as potential moderators of heterogeneity (Supplementary Table S1 for subgroup and heterogeneity analysis and Supplementary Table S2 for meta-regression).

Discussion
In this study, we assessed the echocardiographic parameters of the unloaded LV after AVR.Notably, LV reverse remodeling was evident at the earliest time point evaluated (1 month after AVR).Several of the evaluated parameters were consistent with reverse remodeling, namely, the significant reduction observed in LVM and EDLVD, and LVEF improvement.A trend for EDLVV reduction was also observed.Our results are consistent with those from Mehdipoor et al. (38), who reported indexed LVM reduction and increased LVEF within 6-15 months after TAVI on 10 primary studies involving 305 patients.
Patient follow-up after AVR typically focusses on monitoring valve hemodynamics over time, specifically the evolution of the effective aortic valve area, gradient, and left ventricular function.Reverse left ventricular remodeling is not commonly assessed in routine clinical practice post-AVR.This is partly due to the lack of established norms for what constitutes "normal" left ventricular remodeling after AVR.This study aimed to establish Finally, it is important to note that, despite its infrequent use, the extent of left ventricular remodeling has significant prognostic implications post-AVR.Patients who do not exhibit improvements in LVEF and reductions in left ventricular mass and dimensions after AVR are at a higher risk for increased cardiovascular events (14,39).In our opinion, further attention should be paid to the predictors of inadequate left ventricular remodeling after AVR, as this may aid in defining other criteria for AVR other than the severity of obstruction and left ventricular function.

Strengths and limitations
To our knowledge, this is the most extensive systematic review and meta-analysis conducted to assess the reverse LV remodeling profile in patients who underwent AVR.We excluded studies without a predefined follow-up period to Frontiers in Cardiovascular Medicine 15 frontiersin.orgobtain the most robust results possible.We performed metaregression and subgroup analyses to explore sources of heterogeneity, identifying several variables in this context.To minimize publication and information bias, we searched different electronic bibliographic databases without applying exclusion criteria based on the date or language of publication and contacted authors whenever relevant information was missing.
Limitations of this meta-analysis are related to three main factors: the inherent source of variability regarding to measurements performed by echocardiography, the incomplete characterization of patients in some of the included studies, and the significant heterogeneity observed in our results.
First, a significant source of variability may be related to the fact that primary studies used TTE as the imaging LV assessment method, which is affected by inter-observer and intra-observer variability that can be a source of heterogeneity.For example, the non-significant reduction in LV volume compared to a significant reduction in LV diameter likely reflects the higher variability in echocardiographic measurements of threedimensional parameters like LV volume, which tend to have a higher standard deviation compared to two-dimensional measurements like LV diameter.This variability could obscure significant findings.An analysis based on studies using CMR to evaluate LV could possibly reduce the heterogeneity across studies.However, it would be an undoubtedly less clinically useful analysis (40)(41)(42).Finally, another possible source of heterogeneity is the presence of prosthesis-patient mismatch (PPM), which could influence the results by leading to worse hemodynamic function and LV reverse remodeling.Our study did not analyze PPM because it was not reported in most studies.
Second, other non-evaluated factors may influence the extent of left ventricular remodeling after AVR.In this work, we showed that LV reverse remodeling may differ according to several patient characteristics, namely, age, hypertension, diabetes, coronary heart disease, and NYHA classification.However, the data available for analysis were sparse on information regarding the severity and duration of aortic stenosis, pre-existing LV remodeling, the presence of atrial fibrillation, associated valvular heart diseases, diastolic function, and patient-prosthesis mismatch that may also contribute to the extent of reverse remodeling.Furthermore, by using a summary or aggregate data from study publications, our meta-analysis may fail to identify patient characteristics that might be significant predictors of adequate LV remodeling.For example, previous works have shown that women have a more favorable LV remodeling after AVR than men (43).However, the available aggregate data were insufficient to characterize the impact of gender on LV reverse remodeling after AVR.Finally, significant heterogeneity among studies was observed.Even though meta-regression and subgroup analysis were performed to identify possible variables that differed between studies and could explain the differences between primary studies, it must be noted that the included studies were mainly observational studies and included patients based on convenient criteria (i.e., patients who underwent AVR at a given institution), which added significant heterogeneity that cannot be controlled using regression techniques.This is the most extensive systematic review and meta-analysis assessing reverse LV remodeling after AVR.Echocardiography demonstrates reverse LV remodeling as soon as 1 month after AVR, with reductions in MAG, LVM, and EDLVD, and improvement in AVA and LVEF.

1 . 2 . 3 . 4 . 5 .
Was the research question or objective in this paper clearly stated?Was the study population clearly specified and defined?Was the participation rate of eligible persons at least 50%?Were all the subjects selected or recruited from the same or similar populations (including the same time period)?Were inclusion and exclusion criteria for being in the study prespecified and applied uniformly to all participants?Was a sample size justification, power description, or variance and effect estimates provided?

FIGURE 2 SMD
FIGURE 2 SMD post-AVR vs. pre-AVR for the aortic valve area.

TABLE 1
Keywords used to perform the query in the two databases used in this study (date of search: 15 March 2022).

TABLE 3
Risk of bias assessment of the included studies.

TABLE 3 Continued
. For the analyses in this paper, were the exposure(s) of interest measured prior to the outcome(s) being measured?

TABLE 4
General characteristics of the included studies.