Sutureless Biological Aortic Valve Replacement (Su-AVR) in Redo operations: a retrospective real-world experience report of clinical and echocardiographic outcomes

Perioperative data was retrospectively analysed from a prospectively collated database at a single cardiothoracic institution between 2014 and 2021. Due to the retrospective nature of the study, the requirement for ethical approval was waivered by the Research Ethics Office at the Royal Brompton and Harefield Foundation Trust.

Patients were included if they fulfilled two important criteria: i) they had previously undergone aortic valve surgery (non-repair) in adulthood more than 6 months prior to their inclusion in the study; and ii) they were undergoing re-sternotomy and reimplantation of a new aortic valve prosthesis.

Patients who had received both mechanical and biological valves in the prior index procedure were included. Prior aortic root replacements were also included, including homograft procedures.

All indications for redo aortic valve intervention were considered, including structural valve degeneration and endocarditis. Patients were divided into two groups according to the type of AVR prosthesis they were receiving: either i) a sutureless (SU-AVR) Perceval valve; or ii) a Perimount Magnaease prosthesis as a conventional sutured valve (cAVR).

Patients receiving other types of rapid deployment valves were excluded. To ensure a consistent comparison between the groups, other marketed conventional sutured valves, although used in our institution, were excluded from the study.

In the strategic planning of reoperations within our hospital, several crucial factors were meticulously considered to tailor the approach for each patient. These factors encompassed the status of the index valve, including calcification and infection status, as well as the potential risk of coronary sinus sequestration—a complication associated with valve-in-valve procedures. Additionally, the patients' life expectancy was a pivotal determinant in the decision-making process.

In alignment with these parameters, patients were judiciously directed towards either valve-in-valve transcatheter aortic valve replacement (TAVR), mechanical prosthesis, or bioprosthesis interventions. Notably, a noteworthy proportion of individuals within this series were deemed unsuitable for valve-in-valve procedures due to specific characteristics of their index valves. Particularly, those with biological index valves underwent surgical aortic valve replacement for diverse reasons, reflecting the nuanced and individualized nature of the decision-making process in this cohort.

All reoperations were performed via median sternotomy, with the use of an oscillating saw. The strategy of cannulation for cardiopulmonary bypass was patient-specific and subject to the surgeon’s discretion. In general, if vital mediastinal structures were not at risk from sternal re-entry (as assessed from pre-operative cross-sectional imaging, sternal re-entry followed by standard central cannulation was the first-line desirable method. However, femoral cannulation was performed prior to sternotomy i) injury to vital structures was of concern when planning sternotomy on pre-operative cross-sectional imaging; and ii) femoral vessels were of suitable calibre and amenable to cannulation. Once CPB was initiated, systemic normothermia or mild hypothermia (32 °C) if a patent mammary artery bypass was present. Once the aorta was cross clamped, and the heart successfully arrested, myocardial protection was achieved with intermittent antegrade blood cardioplegia (including direct ostial infusion), with or without retrograde cardioplegia, if the coronary sinus was accessible during initial myocardial dissection.

The decision regarding prosthesis or root replacement was made at the surgeon's discretion. In cases where a Perceval valve was chosen, aortotomy was typically performed through a high transverse incision. For patients determined for Perceval implantation, meticulous and precise decalcification of the aortic root was undertaken. The valve was positioned optimally using three guide sutures and then implanted. Post-implantation, no balloon expanding system was utilized; only the valve's functionality was assessed. In instances of neo-sinus/fistulating root endocarditis, patch repair of the aortic root was implemented. Concurrent procedures followed standard techniques.

Echocardiographic assessments were conducted by seasoned non-invasive cardiologists within the echocardiographic department of our institution. In tandem with a comprehensive preoperative evaluation encompassing the left ventricle and aortic valve, subsequent postoperative examinations included valve imaging and assessment of left ventricular function. These evaluations were performed within 30 days post-surgery and subsequently repeated at a median interval of 24 months during the post-discharge period. Importantly, continuity and standardization were ensured by entrusting these evaluations consistently to the same proficient team within our institutional framework.

The cardiac surgical database is locally managed and centrally overseen at a national level, following national guidelines for minimal perioperative data collection, including pre-operative co-variates, detailed operative characteristics and post-operative care, including the record of short-term complications. Early and mid-term outcomes were assessed based on echocardiographic findings and grading of aortic regurgitation.

The results underwent analysis and were presented in terms of means and standard deviations. Normal distribution of pre-operative covariates was assessed using the Shapiro–Wilk test. Between-group characteristics were evaluated for statistical differences using either the Student T test or the Wilcoxon Rank Test for non-parametric variables. Multivariate regression models were developed to examine the influence of various covariates on both short- and long-term outcomes. The selection of multivariate regression test parameters was selected by the clinical expertise of the academic surgical team. Additionally, Kaplan–Meier survival analyses with Log-rank tests were employed to determine overall survival, and adjusted odds ratios with 95% confidence intervals (CI) for binary outcomes were calculated. The statistical analyses were conducted using Jamovi (Version 2.4) [Computer Software].

Table 1 presents a comprehensive comparison between the su-AVR group and the cAVR group. In terms of demographics, there was no significant difference observed in the mean age between the cAVR (57.6 ± 2.5 years) and su-AVR (59.6 ± 2.4 years) groups (P = 0.607). Similarly, gender distribution displayed no significant difference, with 36 males in the cAVR group and 16 males in the su-AVR group (P = 0.379). The Body Mass Index (BMI) for patients in the cAVR group was 26.6 ± 0.7, while the su-AVR group had a slightly higher BMI of 28.7 ± 1.0 (P = 0.081).

Regarding cardiac comorbidities, the presence of hypertension was observed in 29 patients from the cAVR group and 18 patients from the su-AVR group, with no significant difference in distribution (P = 0.352). Peripheral Arterial Disease (PAD) was reported in 2 patients in both the cAVR and su-AVR groups, with no significant disparity (P = 0.496). A comparable number of cases of Infective Endocarditis were seen in the cAVR (9 patients) and su-AVR (8 patients) groups (P = 0.252). The New York Heart Association (NYHA) classification demonstrated a median of 3 (range 2–3) in the cAVR group and a median of 2 (range 1–3) in the su-AVR group, showing a statistically significant difference (P = 0.036). The occurrence of more than one previous sternotomy was noted in 4 patients in the cAVR group and 6 patients in the su-AVR group, with no significant distinction (P = 0.250).

In terms of the other comorbidities, the prevalence of Oral Diabetes was recorded in 3 patients from the cAVR group and 4 patients from the su-AVR group (P = 0.542). No cases of Insulin-dependent Diabetes were observed in the su-AVR group, while 2 cases were reported in the cAVR group (P = 0.542). Smoking was found in 22 patients from the cAVR group and 13 patients from the su-AVR group, with no significant discrepancy (P = 0.555). Chronic Obstructive Pulmonary Disease (COPD) was documented in 5 patients in the cAVR group and 6 patients in the su-AVR group, showing no significant difference (P = 0.127). A history of stroke was reported in 3 patients from the cAVR group and 2 patients from the su-AVR group, with no significant variance (P = 0.816). Chronic Kidney Disease (CKD) was observed in 2 patients from both the cAVR and su-AVR groups, displaying no significant difference (P = 0.514).

The Table 2 showcases the varying types of explants within each group. In the cAVR group, there were 6 mechanical valve explants, 22 biological valve explants, 19 native valve explants, 2 autograft explants, 1 homograft explant, and 2 freestyle explants. The su-AVR group displayed 5 mechanical valve explants, 17 biological valve explants, 1 native valve explant, 4 homograft explants, and no cases of autograft or freestyle explants.

The Table 3 demonstrates intraoperative outcomes. The cAVR group exhibited a CPB time of 168.4 ± 11.1 min, while the su-AVR group had a CPB time of 165.9 ± 19.6 min (P = 0.452). Regarding XC time, the cAVR group had a mean time of 108.8 ± 6.3 min, while the su-AVR group demonstrated a mean time of 92.0 ± 8.6 min (P = 0.124). In the cAVR group, 36 cases were elective, 15 were urgent, and 1 was emergency, while in the su-AVR group, 17 cases were elective, 10 were urgent, and none were emergency, displaying no significant difference (P = 0.622). The cAVR group included 12 cases with a valve size of 21, 21 cases with size 23, 14 cases with size 25, and 5 cases with size 27. In contrast, the su-AVR group comprised 10 cases with size 21, 13 cases with size 23, 3 cases with size 25, and 1 case with size 27. A statistically significant difference was noted in valve size distribution between the groups (P = 0.048).

In-hospital mortality was found to be similar between the two groups. Specifically, there were 2 deaths out of 52 patients (3.8%) in the cAVR group and 1 out of 27 patients (3.7%) in the su-AVR group. This difference was not statistically significant (P > 0.05). The incidence of stroke was lower in the cAVR group, but this difference was not statistically significant (1 out of 52 patients (1.9%) in the cAVR group vs. 3 out of 27 patients (11.1%) in the su-AVR group, P > 0.05). No cases of gastrointestinal bleeding were observed in either group, which was attributed to the use of biologic prostheses and the absence of anti-coagulation. However, the su-AVR group had lower rates of respiratory complications (7.4% vs. 19.2%, P > 0.05), the need for dialysis (0% vs. 7.7%, P > 0.05), and pericardial/pleural effusions (3.7% vs. 15.3%, P > 0.05) compared to the cAVR group. In addition, none of the patients in the su-AVR group required postoperative pacemaker implantation, while this was necessary in 4 out of 52 patients (7.7%) in the cAVR group. However, this difference was not statistically significant (P > 0.05) (Table 3).

Table 5 provides a comprehensive analysis of postoperative echocardiographic outcomes between the Conventional Aortic Valve Replacement (cAVR) and Sutureless Aortic Valve Replacement (su-AVR) groups (Table 5, Figs. 1, 2).

Left Ventricular Functions displayed no significant differences between groups. In the cAVR group, 31 patients demonstrated good ejection fraction (EF), 9 patients had moderate EF, and 1 patient had poor EF. Similarly, in the su-AVR group, 19 patients exhibited good EF, 4 had moderate EF, and 1 had poor EF. The mean EF percentages were 56.8 ± 10.0 in the cAVR group and 56.1 ± 11.4 in the su-AVR group (P = 0.537). Prosthesis Functions indicated comparable results. The cAVR group had 34 normally functioning prostheses, 6 with mild aortic regurgitation (AR), 1 with moderate AR, and none with severe AR. Likewise, the su-AVR group had 21 normally functioning prostheses, 2 with mild AR, 1 with moderate AR, and none with severe AR (P = 0.714). Aortic Valve Gradient showed variations. Peak gradients were 23.2 ± 10.6 mm Hg in the cAVR group and 30.5 ± 14.8 mm Hg in the su-AVR group (P = 0.026). Mean gradients were 12.9 ± 6.4 mm Hg in the cAVR group and 16.7 ± 8.4 mm Hg in the su-AVR group (P = 0.045). The distribution of gradients under 20 mm Hg, 20–39 mm Hg, and 40 + mm Hg varied between the groups.

Late Term Follow-up revealed comparable findings. Left Ventricular Functions showed no significant differences, with 31 patients with good EF in the cAVR group and 18 in the su-AVR group. EF percentages were 60.4 ± 6.8 in the cAVR group and 58.5 ± 9.2 in the su-AVR group (P = 0.391). The cAVR group had 29 normally functioning prostheses, 4 with mild AR, none with moderate AR, and none with severe AR. The su-AVR group had 14 normally functioning prostheses, 4 with mild AR, 2 with moderate AR, and none with severe AR (P = 0.116). Aortic Valve Gradient values were similar. Peak gradients were 12.9 ± 7.1 mm Hg in the cAVR group and 15.6 ± 13.6 mm Hg in the su-AVR group (P = 0.360). Mean gradients were 22.8 ± 12.3 mm Hg in the cAVR group and 22.8 ± 13.6 mm Hg in the su-AVR group (P = 0.997). The distribution of gradients within specified ranges also showed similarities between the groups.

Kaplan Meier analysis revealed similar survival rates between the su-AVR and cAVR groups, (crude log rank test p = 0.315) (Fig. 3). A multivariate regression analysis was conducted to assess for predictors of early-term survival, including the choice of valve, age, gender, and pre-operative LV function. The results identified no significant predictors of long-term survival (p > 0.05). Specifically, the choice of valve did not influence survival (hazard ratio 1.22, 95% CI 0.14 – 10.25, p = 0.855). Furthermore, multivariate regression test was employed to determine, predictors of late-term survival. Age, NYHA, Perceval, smoking status, COPD, gender, and HT did not show statistical significance for late-term survival. However, patients with Type 2 Diabetes Mellitus (T2DM) demonstrated a significantly higher hazard ratio of 15.1 (95% CI: 1.593–142.424, p = 0.018), highlighting it as a substantial risk factor for reduced survival.These findings are summarized in Tables 4, 5, 6 and 7.

Our study highlights a crucial finding regarding the lower rate of pacemaker implantation in the su-AVR group. Prior reports have indicated that su-AVR, particularly with the Perceval valve, is associated with a higher incidence of postoperative pacemaker implantation in first-time AVR surgery [8, 9]. Although this has been attributed to the initial part of the learning curve, where over-sizing may be more common, this complication is significant and cannot be overlooked [8, 9]. However, our research suggests that precise suture placement during cAVR is more challenging in cases where the annulus is already distressed and fibrosed, and the use of su-AVR may help circumvent this issue.

Interestingly, our study found a higher proportion of small valves were used in the su-AVR group compared to cAVR; this may explain the disparity in PPM results [10]. It's important to note that in our practice, we do not utilize balloon expansion after sutureless valve implantation, a factor we believe might contribute to our success in mitigating PPM incidence. Overall, our findings suggest that su-AVR may provide benefits in reducing the incidence of pacemaker implantation, especially in patients with a fibrosed annulus.

Redo AVR has traditionally been associated with a higher rate of complications compared to first-time SAVR in the literature, with reported operative mortality rates ranging between 4 and 9% [1] [11]. The formation of adhesions and loss of tissue planes between the heart, mediastinal structures, and sternum after the index SAVR increases the risk of complications during sternal re-entry and dissection, particularly in patients who are older or have comorbidities [2]. These risks are further compounded by the proximity to critical thoracic structures. The sequelae of redo surgical AVR may include stroke, myocardial infarction, new atrial fibrillation, and permanent pacemaker implantation.

In a meta-analysis by Formica et al., valve-in-valve transcatheter replacement was found to be advantageous in terms of complications compared to redo aortic valve surgery but redo aortic valve replacement surgery had advantages in medium and long-term survival [12]. In our study, there was no significant difference in perioperative complications between the su-AVR and cAVR groups. However, there was a reduced frequency of effusion and in-hospital mortality in the su-AVR group. Furthermore, none of the patients in the su-AVR group required dialysis, whereas dialysis was needed in four patients in the cAVR group. This suggests that the reduced operative time afforded by su-AVR may have benefits for reducing post-operative coagulopathy, metabolic disturbance, and ICU complications in redo AVR procedures. Much of this may be attributed to the shorter cross-clamp and CPB times.

In the absence of long-term data on valve durability, the use of su-AVR in patients undergoing redo cardiac surgery is not typically indicated, due to the complexity of the procedure and the associated risk of perioperative complications and mortality.

The choice of bioprosthetic AVR variant has not been found to offer a significant survival advantage, but su-AVR has demonstrated some advantages in high-risk patient groups. Hanedan et al. found that su-AVR was three times more advantageous than cAVR in terms of early-term survival in high-risk patients, although the difference was statistically insignificant (94.7% vs 84.6%) [13]. Similarly, Salmasi et al.'s meta-analysis on su-AVR vs cAVR did not find a significant difference in bleeding and early-term survival, but su-AVR's shorter operative time was advantageous with a shorter ICU stay [14]. Coti et al.'s su-AVR study of 700 cases reported a three-year and five-year survival rate of 91% and 76%, respectively, demonstrating that su-AVR was feasible for mid-term survival, even in redo cardiac surgery patients [15]. Our study supports these findings, showing that su-AVR is comparable to cAVR in terms of short/mid/long-term survival in redo surgeries. While the choice of bioprosthetic AVR variant may not significantly affect survival rates, su-AVR's shorter operative time may lead to fewer complications and improved outcomes.

Prosthetic valve endocarditis is a highly morbid condition and compounds the risk of redo surgery according to most risk-scoring systems. The incidence can be as high as 16% in patients with prior AVR [16] and patients who require redo-surgery have exceptionally high morbidity and mortality rates [17] [18]. Complex procedures, such as prolonged cardiopulmonary bypass time and cross-clamp time, are often required, and these are associated with increased mortality and severe perioperative complications [19].

In a recent study, 29.6% of the su-AVR group and 17.3% of the cAVR group had infective endocarditis. Despite this, in-hospital mortality was observed in only one out of the eight patients with prosthetic valve endocarditis in the su-AVR group, while in-hospital mortality was not observed in any of the nine endocarditis patients in the cAVR group. These findings suggest that the use of su-AVR for redo cases with prosthesis infective endocarditis is effective in achieving acceptable short- and medium-term survival, as reported in recent literature.

Evidence on the status of late hemodynamic in sutureless aortic valve replacement is still needed in the literature[20, 21]. Recent studies have suggested no significant hemodynamic differences early period after surgery between sutureless valves and conventional bioprostheses[20, 21]. This observation extends, although with limited evidence, to redo aortic valve surgeries[22].

Our study aimed to address this gap through a meticulous echocardiographic assessment, evaluating sutureless aortic valve prostheses in redo cases. With a median 24-month follow-up, our findings indicate no notable differences between conventional AVR (cAVR) and sutureless AVR (su-AVR) regarding aortic valve median gradient and peak gradient. These measurements remain well within acceptable clinical ranges for both groups.