Early and mid-term outcomes of aortic annular enlargement: a systematic review and meta-analysis

Introduction

Patient-prosthesis mismatch (PPM) is widely recognized as a significant factor impacting clinical outcomes following prosthetic valve implantation. In the context of surgical aortic valve replacement (SAVR), PPM, whether moderate or severe, has been shown to increase both all-cause mortality and cardiac-related mortality (1). In the current era, patients with PPM continue to have reduced long-term survival, as well as an increased risk of rehospitalizations for heart failure (2,3), with some studies also suggesting an increased risk of re-replacement of the aortic valve (3).

To minimize the risk of PPM, the largest possible prosthetic valve should be implanted in each patient. When the native aortic root is small, i.e., at increased risk of PPM, an important approach is to enlarge the aortic annulus before implanting a prosthetic valve. This technique is referred to as aortic annular enlargement (AAE) and includes a variety of techniques, each differing in terms of either the location of the annular incision or the extent of the incision. These techniques include the posterior incisions of Nicks (4,5), Manouguian (5,6), the Nunez modification to the Manouguian (5,7), and the Y-incision described by Yang et al. (8). Additionally, the anterior annular incision with a right ventricular outflow tract (RVOT) incision, the Konno procedure (9), is often reserved for congenital heart disease and adult congenital heart disease applications. Despite the increasing importance of addressing PPM, the most recent valvular heart disease guidelines do not address when or if AAE should be performed (10,11).

There is mounting evidence at experienced centers (8,12,13) that AAE procedures are safe adjuncts to SAVR that do not increase perioperative morbidity and mortality (8,12-15). Despite the increasing experience with AAE at high-volume centers, there is an absence of high-quality evidence related to the long-term results of AAE. There are no comparative studies of AAE versus SAVR without AAE that report mean follow-up periods of 10 years or more. With the literature available, it is unclear how AAE influences the mid- and long-term outcomes of SAVR (14,16).

The most recent meta-analysis examining mid-term survival after AAE was completed by Sá et al. in 2022 (16). Kaplan-Meier curves were required for their quantitative synthesis to generate individual patient data (IPD) using one method of IPD extraction by Liu and colleagues (17). Therefore, their review excluded seven studies due to the absence of Kaplan-Meier curves (16). The other relevant meta-analyses were completed by Yu et al. in 2019 (14) and Sá et al. in 2021 (15). While Yu et al. (14) examined mid-term mortality with five studies published up to 2018, Sá et al. (15) limited their analysis to the perioperative outcomes of AAE. Thus, this systematic review features the most up-to-date and inclusive meta-analysis on the impact of AAE on both the perioperative and mid-term outcomes after SAVR.

Methods

This systematic review is based on a protocol registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD 42023461543). The protocol was developed according to the Cochrane Handbook for Systematic Reviews of Interventions (18), and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Protocols 2015 (PRISMA-P 2015) statement (19), with consultation from a health sciences librarian at the Gerstein Science Information Centre at the University of Toronto.

Literature search strategy

OVID MEDLINE, OVID Embase, and Cochrane Library were searched comprehensively with no limits on the publication time period or language. The search was completed on August 3, 2023. Search terms included “aortic annular enlargement, aortic root enlargement, and aortic valve replacement”, along with relevant synonyms. The reference lists of included studies were reviewed to retrieve additional eligible studies. Grey literature sources were not searched. The search strategy was developed in collaboration with a health sciences librarian at the Gerstein Science Information Center.

Eligibility criteria

Randomized controlled trials (RCTs), controlled (non-randomized) clinical trials, and comparative observational studies were eligible for inclusion. Non-comparative observational studies, case reports, conference proceedings, abstracts, commentaries, letters to the editor, and unpublished work were excluded. The population was limited to adult patients, 18 years or older, who underwent SAVR. Studies that included concurrent procedures were eligible for inclusion, except those that included aortic root replacement with bioprosthetic or mechanical valves, homograft root replacement, the Ozaki procedure, and the Ross procedure. Any study that included patients with a prior aortic root replacement or Ross procedure was also excluded. To be eligible for inclusion, each comparative study needed to have a clearly defined intervention group that underwent SAVR with AAE, and a clearly defined comparator group that underwent SAVR without AAE. Eligible AAE procedures included the following techniques: Nicks (4,5), Manouguian (5,6), Nunez modification to the Manouguian (5,7), Y-incision (8), Konno (9), and any other aortic annular incision that did not require coronary button mobilization and reimplantation. To be eligible for inclusion, each study needed to report on at least one of the outcomes of interest through at least 5 years of follow-up. This was confirmed through a full-text review of the potentially eligible studies by two independent reviewers. The primary outcome of interest was all-cause mortality. Relevant secondary outcomes included cardiac mortality, aortic valve reintervention, structural valve deterioration and non-structural valve dysfunction, valve thrombosis, infective endocarditis, major bleeding, stroke, and rehospitalization for heart failure. While this review intended to examine the long-term results following AAE, due to the absence of studies with mean follow-up lengths beyond 10 years, only mid-term outcomes were assessed.

Data extraction and critical appraisal

Search results were de-duplicated in EndNote (Berkeley, California, USA) and were uploaded to Covidence (Covidence, Melbourne, Australia), an online platform that facilitates de-duplication, record screening, and data extraction for systematic reviews. Title and abstract screening were performed in Covidence by two independent reviewers. Disagreements were resolved by consensus, involving a third reviewer if consensus could not be reached. The records that remained after title and abstract screening underwent a full-text review by two independent reviewers. Data were extracted by two independent reviewers and included study design, patient demographics, surgical techniques, perioperative surgical outcomes, and long-term outcomes of interest. The data extraction form is available on request. Two of the included studies contained Kaplan-Meier curves that required digitization (20,21). This was performed using a web-based Shiny application previously developed by Liu and colleagues to facilitate the digitization and reconstruction of IPD from published Kaplan-Meier curves (17). Risk of bias was assessed in duplicate according to the Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) tool, as all the eligible studies were of non-randomized design (18,22,23). An overall rating of low risk of bias is uncommon within the ROBINS-I methodology as this would mean that the observational study being evaluated would be comparable to a well-designed RCT examining the same question. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework was used to determine the overall quality of evidence (24,25). This was completed by two reviewers based on consensus. Results are reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement (26).

Statistical analysis

Analyses were performed using Review Manager (RevMan version 5.4; Cochrane Collaboration, Oxford, UK) and random effects models, which incorporated between-trial heterogeneity and provided wider and more conservative confidence intervals (CIs) when heterogeneity was present (27). We assessed statistical heterogeneity among trials using I², which is defined as the percentage of total variability across studies attributable to heterogeneity rather than chance. Published guidelines categorized I² values as low (25% to 49%), moderate (50% to 74%), and high (≥75%) heterogeneity (28). For peri-operative outcomes, relative risks (RRs) were used to pool binary outcomes, and the mean difference (MD) was employed for continuous outcomes. When required, the method of Wan et al. (29) was used to convert continuous variables reported as medians and interquartile ranges, or ranges to means and standard deviations. For mid-term outcomes with different follow-up periods between groups, we pooled hazard ratios (HRs) or, if not provided, incidence rate ratios (IRRs) as approximations of the HR on the logarithmic scale using the generic inverse variance method in Review Manager. IRRs for each study were calculated either (I) as the ratio of the Kaplan-Meier survival-curve mortality estimates for each group, with standard error estimated using either the log-rank survival curve P value when available, or alternatively using the standard errors of the survival-curve mortality estimates and the ratio of means method (30,31); or otherwise (II) as the ratio of reported events divided by the group-specific patient-years of follow-up when the group-specific mean follow-up durations were provided, with standard error on the logarithmic scale estimated as the square root of the sum of the reciprocals of the event rates (32). Individual trial and pooled summary results were reported with 95% CIs. Separate sub-groups were created for propensity-score matched or risk-adjusted observational data and unmatched/unadjusted observational data. The a priori-determined sensitivity analyses included studies at moderate versus serious and critical risk of bias, studies with both moderate and serious risk of bias versus critical risk of bias, and studies with and without concomitant procedures. An additional sensitivity analysis was performed to assess the impact of the Rao et al. study, as the procedures used in the AAE cohort were markedly heterogeneous (12). Uncertainty for the pooled binary and continuous outcomes is represented by 95% CIs. Differences between subgroups were assessed using Z-tests. P<0.05 was taken as statistically significant.

Results

Literature search

The search strategy retrieved 2,765 records. After de-duplication, 2,210 unique records remained. Title and abstract screening were performed in duplicate, identifying 139 potentially eligible studies that underwent full-text review by two independent reviewers. Overall, 32 potentially eligible studies (12,13,20,21,33-51) were identified (52-60), including 17 studies (13,45-60) that were excluded because they did not include any information on at least one of the mid-term outcomes of interest through 5 years of follow-up. Consequently, 15 unique studies (12,20,21,33-44) remained and were included in data extraction and quantitative synthesis. The screening process is summarized in the PRISMA trial flow diagram (Figure 1).

Figure 1 Preferred reporting items for systematic reviews and meta-analyses flow diagram. AAE, aortic annular enlargement; SAVR, surgical aortic valve replacement.

Quality of evidence

All 15 included studies are observational and non-randomized (12,20,21,33-44). Of the included studies, five compared propensity-matched groups (34,36,38,41,43), two employed case-control designs to define their reference SAVR groups (34,42), and four reported adjusted mid-term outcomes of interest (21,37,43,44). Notably, Tam and colleagues described two distinct cohorts of patients within the same study—patients who underwent isolated SAVR with or without AAE, and patients who underwent SAVR combined with coronary artery bypass grafting (CABG) with or without AAE (43). As a result, these cohorts were extracted independently, and then combined in the pooled analyses. Only three studies were based on multicentre patient data (12,21,43); the rest reported single-center outcomes (20,33-42,44).

Risk of bias was assessed for each outcome of interest within the included studies according to the ROBINS-I framework (Figure 2 and Figure S1, and Appendix 1) (18,22,23). None of the included studies within our systematic review were deemed to have an overall low risk of bias. Only three included studies reported on outcomes at moderate risk of bias (38,41,43). Mid-term mortality was deemed to be at moderate risk of bias in the studies by Shih et al., Tam et al., and Okamoto et al. (38,41,43). Cumulative incidence of aortic valve reintervention was assessed to be at moderate risk of bias in the study reported by Tam and colleagues (43). All five studies that reported on heart failure-related endpoints were at serious or critical risk of bias for that outcome (12,21,37,38,43). The remaining studies and their other reported outcomes of interest were at serious or critical risk of bias (12,20,21,33-37,39,40,42,44).

Figure 2 ROBINS-I assessment for mortality, aortic valve reintervention, and heart failure. *, distinct secondary cohort reported within the same publication; **, long-term reoperation outcome was assumed to be related to aortic valve reintervention. L, low risk of bias; M, moderate risk of bias; S, serious risk of bias; C, critical risk of bias; NI, no information; AoV, aortic valve; CHF, congestive heart failure; NYHA, New York Heart Association; ROBINS-I, Risk Of Bias In Non-randomized Studies of Interventions.

Publication bias was assessed with visual analysis of the funnel plot for the primary outcome, mid-term mortality (Figure 3), with no indication of significant asymmetry.

Figure 3 Funnel plot for mid-term mortality. SE, standard error.

Baseline demographics

Meta-analyses of baseline characteristics (Table 1 and Table S1) were performed to assess for differences between groups and the effectiveness of matching in the relevant studies (Figures S2-S31). Prior to adjustment or matching, patients who underwent AAE at the time of SAVR were younger (MD, −1.72 year; 95% CI: −2.61 to −0.82), less likely to be male sex (RR, 0.72; 95% CI: 0.63 to 0.81), and had higher body mass index (BMI; MD, 1.80 kg/m²; 95% CI: 0.44 to 3.16), at the same body surface area (BSA; MD, −0.01 m²; 95% CI: −0.03 to 0.01). They were less likely to have chronic renal failure (RR, 0.87; 95% CI: 0.77 to 0.99), coronary artery disease (RR, 0.92; 95% CI: 0.86 to 0.98), and preoperative atrial fibrillation (RR, 0.77; 95% CI: 0.69 to 0.86). They were more likely to have diabetes (RR, 1.13; 95% CI: 1.10 to 1.16), and a history of prior SAVR (RR, 4.54; 95% CI: 2.45 to 8.44). Despite having a slightly higher preoperative left ventricular ejection fraction (LVEF; MD, 0.87%; 95% CI: 0.11% to 1.62%), they tended to have a smaller preoperative aortic valve area (MD, −0.05 cm²; 95% CI: −0.08 to −0.02), including when indexed to BSA [indexed effective orifice area (iEOA); MD, −0.03 cm²/m², 95% CI: −0.05 to −0.01], a smaller aortic annular diameter (MD, −1.36 mm; 95% CI: −2.12 to −0.59), and were more likely to present with predominantly stenotic aortic valve disease (RR, 1.03; 95% CI: 1.01 to 1.05). There were no significant differences regarding BSA, cerebrovascular disease, chronic obstructive pulmonary disease (COPD), smoking, dialysis, hypertension, dyslipidemia, peripheral vascular disease, congestive heart failure/reduced LVEF, New York Heart Association (NYHA) class III–IV, mean NYHA class, elective versus urgent/emergent surgery, Society of Thoracic Surgeons (STS) risk score, prior cardiac surgery, peak aortic gradient, mean aortic gradient, or bicuspid aortic valve. When examining only the studies with matching or adjusted outcomes, almost all significant baseline differences disappeared, with the only exceptions being that patients undergoing AAE had higher preoperative BMI (MD, 1.24 kg/m²; 95% CI: 0.18 to 2.31), with no significant difference in their BSA, and were less likely to have a bicuspid aortic valve (RR, 0.64; 95% CI: 0.43 to 0.95).

Of the 15 included studies, only five described attempting to standardize the size of the native aortic annulus between the SAVR with AAE and SAVR without AAE groups, including two matched/adjusted studies (37,41) and three unmatched and unadjusted studies (33,35,39). Kulik et al. described both groups as having a native annulus that would have necessitated a size 21 prosthesis or smaller (37). Shih et al. incorporated the aortic valve area into their propensity score matching model (41). Beckmann et al. defined both groups as having a projected iEOA ≤0.89 cm²/m² when measured intraoperatively (33). Correia et al. defined both groups as having an implanted prosthesis size of 21 mm or smaller (35). Penaranda et al. defined both groups as having an annulus that would only accept a maximum valve size of 19 mm prior to any annular enlargement being performed (39).

Intraoperative details

AAE was performed through a variety of techniques (Table 2). The most common approaches were the Nicks, and the Manouguian procedures. Only one study (20) described the use of the Nunez technique in combination with the Nicks root enlargement. None of the included studies described the use of the Konno or Y-incision techniques. Importantly, two of the three largest multicentre studies did not capture the AAE technique within their study data (21,43). In both cases, this was due to limitations of the databases used in each of these studies; Mehaffey and colleagues used the STS Adult Cardiac Surgery Database (21), while Tam and colleagues used the CorHealth Ontario Cardiac Registry in combination with the Canadian Institute of Health Information Discharge Abstract Database to collect procedural data for each patient (43). Finally, in the multicentre study reported by Rao and colleagues, there was marked heterogeneity within the proposed aortic root enlargement group (12). Only 27 of the 90 patients in the group underwent a confirmed AAE, with three other patients undergoing an aortic root replacement within the group, and others within the group undergoing either a sinotubular junction (STJ) enlargement or a sinus of Valsalva patch augmentation.

The indication(s) for AAE were infrequently reported within the included studies (Table 3). When indications were reported, they were often listed as possible considerations that could be weighed at the surgeon’s discretion at the time of the operation. Only the study by Sakamoto and colleagues described an objective criterion, aortic annulus smaller than a size 21 valve sizer, without indicating that the decision could also be influenced by surgeon preference (40). Correspondingly, the intraoperative results of the AAE procedures, i.e., the extent of annular enlargement achieved, were also infrequently described. The studies that did report the extent of annular enlargement described an implanted valve, at most, one-to-two valve sizes larger than the initial intraoperative measurement of the aortic root (20,33,35,37,40,42).

Operative details, including valve type, sizing, and rates of concomitant procedures, were pooled (Figures S32-S39). In the matched or adjusted studies, there were notable procedural differences between the AAE and SAVR groups. The patients undergoing AAE were less likely to receive a mechanical valve (RR, 0.80; 95% CI: 0.68 to 0.93), and required both longer cardiopulmonary bypass (MD, 21.33 min; 95% CI: 9.69 to 32.97) and aortic cross-clamp (MD, 19.25 min; 95% CI: 10.17 to 28.33) times. In the unmatched and unadjusted studies, patients receiving AAE were less likely to receive both concomitant mitral valve surgery (RR, 0.55; 95% CI: 0.39 to 0.78) and concomitant tricuspid valve surgery (RR, 0.27; 95% CI: 0.10 to 0.73). Implanted valve size in the AAE group was lower, but only in the matched/adjusted studies (MD, −0.67 mm; 95% CI: −1.09 to −0.25). Only one matched study described concomitant mitral and tricuspid valve surgeries, and these were well-balanced after propensity matching (38). Notably, there was no significant difference in the rate of concomitant CABG observed between groups, in either the matched/adjusted studies or the unmatched/unadjusted studies.

Perioperative outcomes

Perioperative outcomes were also assessed via meta-analyses (Figures S40-S55). In the unmatched and unadjusted studies, AAE patients were less likely to have severe PPM (iEOA ≤0.65 cm²/m²; RR, 0.61; 95% CI: 0.40 to 0.93), moderate or severe PPM (defined as iEOA ≤0.85 cm²/m² in most studies; RR, 0.70; 95% CI: 0.58 to 0.84), and were at increased risk of chest reopening (RR, 1.10; 95% CI: 1.01 to 1.20). Notably, they were also at increased risk of perioperative mortality (RR, 1.34; 95% CI: 1.02 to 1.76), and prolonged mechanical ventilation/other respiratory complications (RR, 1.67; 95% CI: 1.23 to 2.26). However, when only the matched or adjusted studies were considered, the risks of perioperative mortality (RR, 1.06; 95% CI: 0.69 to 1.61), and prolonged ventilation/other respiratory complications (RR, 1.61; 95% CI: 0.75 to 3.47) were not significantly higher in the AAE group. In both the unadjusted/unmatched and the matched/adjusted studies, there were no significant differences identified regarding the risk of perioperative stroke, myocardial infarction, permanent pacemaker implantation, intensive care unit (ICU) length of stay, hospital length of stay, deep sternal wound infection, postoperative iEOA, moderate PPM, peak/mean transprosthetic gradient at discharge or paravalvular leak. The only perioperative complication that was found to be statistically significant in the matched and adjusted studies, was an increased risk of chest reopening in the AAE group (RR, 1.58; 95% CI: 1.13 to 2.21). This was primarily due to the results of Tam et al. (43), which accounted for 89% of the weighting for this matched/adjusted pooled outcome. Without the study from Tam et al. (43), the pooled outcome for the risk of chest reopening in the remaining matched/adjusted studies was no longer statistically significant (RR, 0.97; 95% CI: 0.36 to 2.65).

Assessment of primary and secondary endpoints

The only outcomes of interest with sufficient data to allow for pooled analysis were the mid-term mortality (Figure 4), aortic valve reintervention (Figure 5), and heart failure (Figure 6). The other outcomes of interest were reported by too few studies to provide meaningful pooled estimates of effect (Figures S56-S61).

Figure 4 Meta-analysis for mid-term mortality. Mean duration of follow-up in round brackets for AAE + SAVR vs. SAVR groups; method used to calculate hazard ratio or incident rate ratio in square brackets. The ≤3 and >3 years HRs provided in Mehaffey et al. were replaced with an average HR as the pooled HR is essentially unchanged. AAE, aortic annular enlargement; SAVR, surgical aortic valve replacement; SE, standard error; IV, inverse variance; CI, confidence interval; KM, Kaplan-Meier; HR, hazard ratio; CABG, coronary artery bypass graft; RoM, ratio of means.

Figure 5 Meta-analysis for aortic valve reintervention. Mean duration of follow-up in round brackets for AAE + SAVR vs. SAVR groups; method used to calculate hazard ratio or incident rate ratio in square brackets. Tam et al., Mehaffey et al., and Rao et al. provided hazard ratios, and Shih et al. provided group-specific follow-up. For Sakamoto et al., Prifti et al., and Yousef et al., where no group-specific follow-up or hazard ratio was provided, equal follow-up was assumed to calculate incident rate ratios. AAE, aortic annular enlargement; SAVR, surgical aortic valve replacement; SE, standard error; IV, inverse variance; CI, confidence interval; CABG, coronary artery bypass graft; KM, Kaplan-Meier; HR, hazard ratio.

Figure 6 Meta-analysis for heart failure. Mean duration of follow-up in round brackets for AAE + SAVR vs. SAVR groups; method used to calculate hazard ratio or incident rate ratio in square brackets. AAE, aortic annular enlargement; SAVR, surgical aortic valve replacement; SE, standard error; IV, inverse variance; CI, confidence interval; NYHA, New York Heart Association; CHF, congestive heart failure; HR, hazard ratio; CABG, coronary artery bypass graft; KM, Kaplan-Meier; pts, patients.

Mid-term mortality was reported by nine studies with matched groups or adjusted outcomes (21,34,36-38,41-44) and six studies without matching or adjustment (12,20,33,35,39,40). The unmatched/unadjusted cohorts within six of the studies with matching/adjustment were also available and were included in the synthesis of unmatched/unadjusted studies (21,36,37,42-44). Of note, the study by Tam and colleagues yielded an independent secondary cohort, SAVR with CABG both with and without AAE, that contained both matched/adjusted and unmatched/unadjusted outcome data for mid-term mortality (43). The estimates from the primary and secondary cohorts were combined in the pooled analyses for mid-term mortality. The study by Mehaffey and colleagues, with a median follow-up of 3.3 years, provided two separate HRs for mid-term mortality, up to 3 years of follow-up, and greater than 3 years of follow-up (21). As the primary interest of the review was mid-term mortality, we elected to consolidate the two HRs into an average HR. Importantly, the pooled HR was unchanged when the two HRs were replaced by the average HR. Overall, there was no significant difference in the mid-term mortality observed between groups in either the unmatched/unadjusted (HR, 0.91; 95% CI: 0.80 to 1.03; P=0.12; I²=63%) or matched/adjusted (HR, 1.03; 95% CI: 0.95 to 1.11; P=0.49; I²=20%) studies.

Unmatched/unadjusted aortic valve reintervention was reported by seven studies (12,20,21,40,41,43,44). Two of the seven studies also reported matched or adjusted results (21,43). There was no significant difference in aortic valve reintervention observed between groups in either the unmatched/unadjusted studies (HR, 1.08; 95% CI: 0.85 to 1.39; P=0.53; I²=0%) or the matched/adjusted studies (HR, 0.98; 95% CI: 0.75 to 1.27; P=0.86; I²=0%).

Unadjusted/unmatched congestive heart failure was reported by four studies (12,21,37,43). Three of the four studies also reported matched or adjusted results (21,37,43), along with another study that reported only propensity-matched results (38). There was no significant difference in heart failure observed between groups in either the unmatched/unadjusted studies (HR, 1.10; 0.998 to 1.21; P=0.06; I²=0%) or the matched/adjusted studies (HR, 1.06; 95% CI: 0.86 to 1.30; P=0.58; I²=25%).

The overall quality of evidence for each outcome of interest was assessed using the GRADE methodology and is presented in the summary of findings table (Table 4) (24,25). For both mid-term mortality and aortic valve reintervention, the quality of evidence was low and very low in the matched/adjusted and the unmatched/unadjusted subsets, respectively. For heart failure, the quality of evidence was very low in both the matched/adjusted and the unmatched/unadjusted subsets. In the case of the matched or adjusted subsets, their ratings resulted from the inherent limitations of unblinded and non-randomized study designs. While for the unmatched and unadjusted subsets, the serious and critical risk of bias associated with multiple included studies warranted an additional downgrade to very low-quality evidence. Importantly, the matched/adjusted subset for heart failure was also downgraded to very low quality due to the presence of studies at serious and critical risk of bias (Table 4 and Table S2).

Sensitivity analyses

Sensitivity analyses were performed to assess the impact of the Rao 2023 study (12), the inclusion of concomitant procedures, and the studies at various risk of bias levels (Tables 5,6 and Figures S62,S63). The sensitivity analyses were limited to mid-term mortality and aortic valve reintervention, as there were too few included studies in the heart failure outcome to warrant additional hypothesis testing. The pooled results for both mid-term mortality and aortic valve reintervention did not differ with regards to the presence or absence of the Rao 2023 study (12), concomitant procedures, or the removal of either studies only at critical risk of bias or studies at both serious and critical risk of bias.

Discussion

As is consistent with the current understanding of AAE procedures, the results of this meta-analysis attest to their perioperative safety. The findings of no increased risk of perioperative mortality, myocardial infarction, permanent pacemaker implantation, or stroke when AAE is performed in appropriately matched patients, align with the previous work of Yu et al. (14) and Sá et al. (15). Similarly, this synthesis is aligned with the previous work of Yu et al. (14) and Sá et al. (16) that did not demonstrate a difference in mid-term mortality in appropriately matched patients. However, this review is the first to describe the mid-term risks of aortic valve reintervention and heart failure after AAE. It is also the first synthesis to identify an increased risk of chest reopening after AAE procedures that were present within matched groups. This finding was primarily driven by the increased risk of chest reopening in the secondary cohort of one study, i.e., SAVR with CABG with or without AAE (43). While Tam et al. (43) have theorized that this may have been due to the addition of AAE to a more complex operation, i.e., SAVR with CABG, this finding warrants further exploration, ideally through well-matched comparative studies with detailed descriptions of concomitant procedures.

Despite the increasing use of AAE during SAVR, there remains a paucity of long-term data concerning the impact of AAE on SAVR. For the studies that do have a mid-term follow-up, the reported outcome domains are sparse, with only enough data at this time to derive pooled estimates for all-cause mortality, aortic valve reintervention, and heart failure. A few of the many mid- and long-term outcomes that can factor into the decision to perform an AAE include cardiac mortality, stroke, and structural valve deterioration. Outcomes such as these are not available to patients and their surgeons in the context of AAE. At best, there is indirect evidence of the long-term viability of AAE procedures. When performed in high-volume centers of expertise or examined in syntheses (14,15) with appropriate adjustment to account for meaningful differences in baseline risks between patient populations, there appears to be no added perioperative morbidity or mortality due to AAE (8,12-15). When these procedures are successfully performed, the iEOA is either restored to that of a comparator group with a native annulus that can accommodate the same valve size without requiring augmentation, or the annular enlargement cohort exceeds the iEOA of a comparator group that received a valve that was sized too small relative to their BSA. Given the growing understanding of the risks posed by PPM, i.e., increased risk of mortality (2,3), heart failure rehospitalization (2,3), and aortic valve reintervention (3), a successful AAE cohort would be expected to either reach the equivalent survival to a comparator group with an appropriately sized valve or superior survival versus one with significant PPM.

The overall literature regarding AAE is poorly defined. Most studies do not report preoperative aortic annular dimensions, including the high-powered database studies that are often limited in that they lack the granularity of individual patients’ echocardiographic data. Matching patients in the annular enlargement and comparator groups by their native aortic annular dimensions is also rarely described. As such, it is rarely possible to determine whether the expected outcome is for the annular enlargement cohort to reach equivalence to an appropriately sized comparator or exceed the performance of a group with a significant PPM. The decision of when to perform AAE is similarly unclear. Although the adverse effects of PPM continue to be recognized, most studies either do not list objective decision-making criteria, such as predicted PPM, or when they do, they qualify the criteria with the decision remaining subject to surgeon discretion. When even the best available studies are subjected to this uncertainty, the possibility of unmeasured known and unknown confounders multiplies. The finding that patients undergoing AAE are less likely to receive mechanical valves within the matched and adjusted studies is perhaps a signal that alternate means of avoiding the unfavorable hemodynamics of a mismatched bioprosthesis are being employed in comparator groups, thereby diminishing the potential benefits seen with AAE procedures. Finally, the definition of a successful AAE is equally uncertain. In the rare studies where the annular increase is reported, it is often conservative, with one to two valve sizes at most (20,33,35,37,40,42). With new techniques (8,61) yielding annular enlargement to the extent of three to five valve sizes, one must wonder whether a single valve size increase is enough, and whether the studies that do not report their annular dimensions are achieving any annular increase at all. An illustration of this technical variability can be seen wherein patients undergoing AAE in the matched or adjusted studies were more likely to receive a smaller valve size. Importantly, the same AAE methods were described in both subsets. Despite the numerous techniques described for AAE, their central principle is the alleviation of PPM, and it is this principle that is often unable to be assessed within the existing literature.

There are inherent methodologic limitations within this systematic review. Firstly, all the included studies were non-randomized, leaving a significant possibility of confounding, particularly with regard to the selection of patients undergoing AAE. While some studies reported mid-term secondary outcome data for stroke, structural valve deterioration, non-structural valve dysfunction, infective endocarditis, or major bleeding, they lacked the specificity in terms of the outcome descriptions and the requisite breadth of data across the dataset to be able to enter quantitative syntheses. As the included studies were published from 1997 to 2023, there is additionally an era effect that can be expected in terms of both the evolution of prosthetic aortic valve technologies, as well as the surgical volumes and technical developments with the various AAE techniques at both the center and surgeon levels.

The quality of available observational studies remains poor and randomized trials are unlikely. Collaborative multicentre prospective studies with clear decision-making criteria for AAE and a priori determined benchmarks of technical success, including the number of valve sizes gained, and the expected post-operative transprosthetic gradients, would be able to better assess the impact of AAE procedures on the long-term outcomes of SAVR. It is likely that the exact technique of AAE is less important than the successful upsizing of the prosthetic valve and avoidance of PPM. With regards to propensity matching, selecting comparator patients based on preoperative annular size may yield a much more informative comparison than matching based on the size of the prosthetic valve implanted. Patients matched by implanted valve size would also likely be matched to BSA, and thus would not be expected to have a meaningful difference in PPM, a potential driver of their mid- and long-term outcomes (2,3).

Conclusions

Despite the variability in technical success amongst the studies reviewed and inherent issues with generalizability from single-center, non-randomized, observational studies, particularly those that select patients for AAE without formal criteria, AAE remains an important technique to address the challenge of SAVR in the small aortic root. SAVR with AAE does not appear to be associated with increased perioperative morbidity or mortality. There is no conclusive indication that AAE enhances mid-term survival, freedom from reoperation after SAVR, or freedom from heart failure. When considering mid- to long-term outcomes, it is important to consider what the definition of success would be for AAE procedures. It is critical to be able to understand whether an AAE has succeeded in alleviating PPM, and what the natural history of a particular comparator group is, to contextualize the technical innovations and refinements of AAE to come.

Acknowledgments

We would like to acknowledge Julia Martyniuk at the Gerstein Science Information Centre at the University of Toronto for their invaluable support in developing the search strategy and protocol for this work.

Funding: None.

Footnote

Conflicts of Interest: D.V. is supported by the Canadian Institutes of Health Research (CIHR) Vanier Canada Graduate Scholarship. M.O. is partially supported by the Munk Chair in Advanced Therapeutics and the Antonio & Helga DeGasperis Chair in Clinical Trials and Outcomes Research. The other authors have no conflicts of interest to declare.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Head SJ, Mokhles MM, Osnabrugge RL, et al. The impact of prosthesis-patient mismatch on long-term survival after aortic valve replacement: a systematic review and meta-analysis of 34 observational studies comprising 27 186 patients with 133 141 patient-years. Eur Heart J 2012;33:1518-29. [Crossref] [PubMed]
2. Dismorr M, Glaser N, Franco-Cereceda A, et al. Effect of Prosthesis-Patient Mismatch on Long-Term Clinical Outcomes After Bioprosthetic Aortic Valve Replacement. J Am Coll Cardiol 2023;81:964-75. [Crossref] [PubMed]
3. Fallon JM, DeSimone JP, Brennan JM, et al. The Incidence and Consequence of Prosthesis-Patient Mismatch After Surgical Aortic Valve Replacement. Ann Thorac Surg 2018;106:14-22. [Crossref] [PubMed]
4. Nicks R, Cartmill T, Bernstein L. Hypoplasia of the aortic root. The problem of aortic valve replacement. Thorax 1970;25:339-46. [Crossref] [PubMed]
5. Nezić D, Knezević A, Borović S. Surgical techniques for posterior aortic root enlargement. J Thorac Cardiovasc Surg 2008;135:1401-2; author reply 1402-4. [Crossref] [PubMed]
6. Manouguian S, Seybold-Epting W. Patch enlargement of the aortic valve ring by extending the aortic incision into the anterior mitral leaflet. New operative technique. J Thorac Cardiovasc Surg 1979;78:402-12.
7. Nuñez L, Aguado MG, Pinto AG, et al. Enlargement of the aortic annulus by resecting the commissure between the left and noncoronary cusps. Tex Heart Inst J 1983;10:301-3.
8. Yang B, Ghita C, Makkinejad A, et al. Early outcomes of the Y-incision technique to enlarge the aortic annulus 3 to 4 valve sizes. J Thorac Cardiovasc Surg 2024;167:1196-1205.e2. [Crossref] [PubMed]
9. Konno S, Imai Y, Iida Y, et al. A new method for prosthetic valve replacement in congenital aortic stenosis associated with hypoplasia of the aortic valve ring. J Thorac Cardiovasc Surg 1975;70:909-17.
10. Vahanian A, Beyersdorf F, Praz F, et al. 2021 ESC/EACTS Guidelines for the management of valvular heart disease: Developed by the Task Force for the management of valvular heart disease of the European Society of Cardiology (ESC) and the European Association for Cardio-Thoracic Surgery (EACTS). Rev Esp Cardiol (Engl Ed) 2022;75:524. [Crossref] [PubMed]
11. Otto CM, Nishimura RA, Bonow RO, et al. 2020 ACC/AHA Guideline for the Management of Patients With Valvular Heart Disease: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2021;143:e72-e227. [Crossref] [PubMed]
12. Rao V, Linick JA, Reardon MJ, et al. Early and late effects of aortic root enlargement: Results from the Pericardial Surgical Aortic Valve Replacement Pivotal Trial: A multicenter, prospective clinical trial. JTCVS Open 2023;13:54-74. [Crossref] [PubMed]
13. Rocha RV, Manlhiot C, Feindel CM, et al. Surgical Enlargement of the Aortic Root Does Not Increase the Operative Risk of Aortic Valve Replacement. Circulation 2018;137:1585-94. [Crossref] [PubMed]
14. Yu W, Tam DY, Rocha RV, et al. Aortic Root Enlargement Is Safe and Reduces the Incidence of Patient-Prosthesis Mismatch: A Meta-analysis of Early and Late Outcomes. Can J Cardiol 2019;35:782-90. [Crossref] [PubMed]
15. Sá MPBO, Zhigalov K, Cavalcanti LRP, et al. Impact of Aortic Annulus Enlargement on the Outcomes of Aortic Valve Replacement: A Meta-analysis. Semin Thorac Cardiovasc Surg 2021;33:316-25. [Crossref] [PubMed]
16. Sá MP, Van den Eynde J, Amabile A, et al. Late Outcomes After Aortic Root Enlargement During Aortic Valve Replacement: Meta-Analysis With Reconstructed Time-To-Event Data. J Cardiothorac Vasc Anesth 2022;36:3065-73. [Crossref] [PubMed]
17. Liu N, Zhou Y, Lee JJ. IPDfromKM: reconstruct individual patient data from published Kaplan-Meier survival curves. BMC Med Res Methodol 2021;21:111. [Crossref] [PubMed]
18. Higgins JPT, Thomas J, Chandler J, et al., editors. Cochrane Handbook for Systematic Reviews of Interventions. February 2022. Available online: www.training.cochrane.org/handbook [accessed 15 June 2023].
19. Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015;4:1. [Crossref] [PubMed]
20. Prifti E, Bonacchi M, Baboci A, et al. Hemodynamics of 17-mm vs. 19-mm St. Jude Medical Regent and annulus enlargement. Asian Cardiovasc Thorac Ann 2015;23:670-83. [Crossref] [PubMed]
21. Mehaffey JH, Hawkins RB, Wegermann ZK, et al. Aortic Annular Enlargement in the Elderly: Short and Long-Term Outcomes in the United States. Ann Thorac Surg 2021;112:1160-6. [Crossref] [PubMed]
22. Sterne JA, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ 2016;355:i4919. [Crossref] [PubMed]
23. Sterne JAC, Higgins JPT, Elbers RG, et al. Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I): detailed guidance. 20 October 2016. Available online: http://www.riskofbias.info [accessed 16 January 2024].
24. Schünemann HJ, Brennan S, Akl EA, et al. The development methods of official GRADE articles and requirements for claiming the use of GRADE - A statement by the GRADE guidance group. J Clin Epidemiol 2023;159:79-84. [Crossref] [PubMed]
25. Schünemann H, Brożek J, Guyatt G, et al., editors. Handbook for grading the quality of evidence and the strength of recommendations using the GRADE approach. October 2013. Available online: https://gdt.gradepro.org/app/handbook/handbook.html [accessed 1 February 2024].
26. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372: [Crossref] [PubMed]
27. DerSimonian R, Laird N. Meta-analysis in clinical trials revisited. Contemp Clin Trials 2015;45:139-45. [Crossref] [PubMed]
28. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-60. [Crossref] [PubMed]
29. Wan X, Wang W, Liu J, et al. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 2014;14:135. [Crossref] [PubMed]
30. Friedrich JO, Adhikari NK, Beyene J. Ratio of means for analyzing continuous outcomes in meta-analysis performed as well as mean difference methods. J Clin Epidemiol 2011;64:556-64. [Crossref] [PubMed]
31. Friedrich JO, Adhikari NK, Beyene J. The ratio of means method as an alternative to mean differences for analyzing continuous outcome variables in meta-analysis: a simulation study. BMC Med Res Methodol 2008;8:32. [Crossref] [PubMed]
32. Higgins JPT, Thomas J, Chandler J, et al., editors. Cochrane Handbook for Systematic Reviews of Interventions version 6.4 (updated August 2023). Section 6.7.1 Effect Measures for Counts and Rates. Available online: www.training.cochrane.org/handbook [accessed 20 February 2024].
33. Beckmann E, Martens A, Alhadi F, et al. Aortic valve replacement with sutureless prosthesis: better than root enlargement to avoid patient-prosthesis mismatch? Interact Cardiovasc Thorac Surg 2016;22:744-9. [Crossref] [PubMed]
34. Chauvette V, Sénéchal M, Barrette V, et al. Annulus root enlargement during redo aortic valve replacement: Perioperative results and hemodynamic impact. J Card Surg 2020;35:2158-64. [Crossref] [PubMed]
35. Correia PM, Coutinho GF, Branco C, et al. Long-term follow-up of patients undergoing aortic root enlargement for insertion of a larger prosthesis. Eur J Cardiothorac Surg 2016;50:82-8. [Crossref] [PubMed]
36. Haunschild J, Scharnowski S, Mende M, et al. Aortic root enlargement to mitigate patient-prosthesis mismatch: do early adverse events justify reluctance? Eur J Cardiothorac Surg 2019; [Crossref]
37. Kulik A, Al-Saigh M, Chan V, et al. Enlargement of the small aortic root during aortic valve replacement: is there a benefit? Ann Thorac Surg 2008;85:94-100. [Crossref] [PubMed]
38. Okamoto Y, Yamamoto K, Sugimoto T, et al. Early and Late Outcomes of Aortic Valve Replacement with Aortic Annular Enlargement: A Propensity Analysis. Thorac Cardiovasc Surg 2016;64:410-7. [Crossref] [PubMed]
39. Penaranda JG, Greason KL, Pislaru SV, et al. Aortic root enlargement in octogenarian patients results in less patient prosthesis mismatch. Ann Thorac Surg 2014;97:1533-8. [Crossref] [PubMed]
40. Sakamoto Y, Hashimoto K, Okuyama H, et al. Prevalence and avoidance of patient-prosthesis mismatch in aortic valve replacement in small adults. Ann Thorac Surg 2006;81:1305-9. [Crossref] [PubMed]
41. Shih E, DiMaio JM, Squiers JJ, et al. Outcomes of aortic root enlargement during isolated aortic valve replacement. J Card Surg 2022;37:2389-94. [Crossref] [PubMed]
42. Sommers KE, David TE. Aortic valve replacement with patch enlargement of the aortic annulus. Ann Thorac Surg 1997;63:1608-12. [Crossref] [PubMed]
43. Tam DY, Dharma C, Rocha RV, et al. Early and late outcomes following aortic root enlargement: A multicenter propensity score-matched cohort analysis. J Thorac Cardiovasc Surg 2020;160:908-919.e15. [Crossref] [PubMed]
44. Yousef S, Brown JA, Serna-Gallegos D, et al. Impact of Aortic Root Enlargement on Patients Undergoing Aortic Valve Replacement. Ann Thorac Surg 2023;115:396-402. [Crossref] [PubMed]
45. Aka SA, Orhan G, Unal S, et al. Functional results in aortic root enlargement. Heart Surg Forum 2004;7:E160-3. [Crossref] [PubMed]
46. Castro LJ, Arcidi JM Jr, Fisher AL, et al. Routine enlargement of the small aortic root: a preventive strategy to minimize mismatch. Ann Thorac Surg 2002;74:31-6; discussion 36. [Crossref] [PubMed]
47. Dhareshwar J, Sundt TM 3rd, Dearani JA, et al. Aortic root enlargement: what are the operative risks? J Thorac Cardiovasc Surg 2007;134:916-24. [Crossref] [PubMed]
48. FouadAHassanMMagedMAortic Valve Replacement in Patients with Small Aortic Annulus: New Pericardial Stented Valves vs Aortic Root Enlargement Procedures.Authorea 2021. doi: .
49. Ghoneim A, Bouhout I, Demers P, et al. Management of small aortic annulus in the era of sutureless valves: A comparative study among different biological options. J Thorac Cardiovasc Surg 2016;152:1019-28. [Crossref] [PubMed]
50. Shibata Y, Abe T, Sekine S, et al. Supraannular aortic root enlargement in patients with small aortic annulus: efficacy and safety. Ann Thorac Cardiovasc Surg 1998;4:262-5.
51. Takakura H, Kurosawa H, Nakano M, et al. Enlarging procedure for small aortic annulus. J Cardiol 1995;26:105-8.
52. Kitamura M, Yamazaki K, Ohta T, et al. Left ventricular function on exercise after surgical treatment of small aortic annuli. Cardiovasc Surg 1995;3:583-5. [Crossref] [PubMed]
53. Mubarak Y, Abdel Jawad AAR. Aortic Root Enlargement in patients with small aortic annulus undergoing double valve replacement. Is it justified? A retrospective Comparative cohort study. Heart Surg Forum 2021; [Crossref]
54. Oshima H, Usui A, Akita T, et al. Aortic valve replacement for the small aortic annulus. Kyobu Geka 2006;59:269-75.
55. Prakash S, Agarwal S, Dutta N, et al. A comparative study of surgical treatment of small aortic root with or without aortic root enlargement using a single prosthesis type. J Cardiovasc Med (Hagerstown) 2010;11:836-42. [Crossref] [PubMed]
56. Sant'anna JR, de Bacco FW, Santánna RT, et al. Aortic valve replacement with anterior and posterior enlargement of small aortic annulus is comparable to surgery with normal annulus. Rev Port Cardiol 2006;25:593-602.
57. Sim EK, Orszulak TA, Schaff HV, et al. Influence of prosthesis size on change in left ventricular mass following aortic valve replacement. Eur J Cardiothorac Surg 1994;8:293-7. [Crossref] [PubMed]
58. Stoger GI, Marenchino RG, Rosi E, et al. Aortic annulus enlargement vs. Conventional surgery in patients with small aortic annulus undergoing aortic valve replacement. Rev Argent Cardiol 2017;85:30-4.
59. Wang R, Chen X, Xu M, et al. Clinical research of aortic valve replacement in small aortic annulus. Zhonghua Wai Ke Za Zhi 2014;52:131-4.
60. David TE, Uden DE. Aortic valve replacement in adult patients with small aortic annuli. Ann Thorac Surg 1983;36:577-83. [Crossref] [PubMed]
61. Yang B, Ghita C, Palmer S. Y-incision Aortic Root Enlargement With Modified Aortotomy Upsizing the Annulus by 5 Valve Sizes. Ann Thorac Surg 2022;114:e479-81. [Crossref] [PubMed]