Comparison of bioprosthetic valves in primary isolated aortic valve replacement: a nationwide study^†

Abstract

OBJECTIVES

This study aimed to comprehensively compare the early and long-term clinical outcomes of bovine pericardial valve (BOV) and porcine valve (POV), specifically during the primary isolated aortic valve replacement (AVR), using data from the Korean National Health Insurance Service (NHIS) database.

METHODS

Using the Korean NHIS claimed data, all adult patients (age ≥19) who underwent primary isolated AVR using bioprosthetic valve from 2003 to 2019 were identified and propensity-score matching (PSM) analysis was performed.

RESULTS

Overall, 5470 patients with BOV (n = 3947, group B) or POV (n = 1,523, group P) were enrolled, of whom 814 pairs were matched in a 1:1 ratio using PSM analysis. Early postoperative mortality and morbidities were comparable between the groups before and after PSM, considering inter-hospital clustering. The cumulative incidence of all-cause mortality was higher in group P than in group B in the total cohort (group B 5.3%/patient-years vs group P 6.4%/patient-years, adjusted hazard ratio: 1.20, P = 0.002), whereas those differences in all-cause mortality disappeared in PSM analysis (P = 0.24). The cumulative incidences of late stroke, reoperation and infective endocarditis were not significantly different between the groups in PSM populations.

CONCLUSIONS

In Korean national database cohort patients with primary isolated AVR, there was no significant difference in the early- and long-term clinical outcomes between BOV and POV.

Open in new tabDownload slide

INTRODUCTION

In surgical aortic valve replacement (AVR), bioprosthetic valves are core implants, especially those used in elderly patients because they do not require lifelong anticoagulation and have sufficient durability [1–3]. Considering the increasing trend of valve-in-valve transcatheter aortic valve implantation, it is expected that the use of bioprosthetic valves during the first AVR will also continue to rise, especially in patients requiring AVR at a younger age than the current guideline’s recommendations [4, 5]. There are 2 types of bioprosthetic valves: bovine pericardial valves (BOV) and porcine valves (POV). Previous studies have demonstrated that the BOV showed better haemodynamics, a lower frequency of prosthesis-patient mismatch, and improved left ventricular mass regression after AVR [6–9]. However, the debate over long-term clinical outcomes continues, with conflicting results on whether the BOV or the POV is superior in the aortic position [10–13]. This study aimed to compare the early and long-term clinical outcomes of BOV and POVs in patients with primary isolated AVR using the Korean National Health Insurance Service-National Health Information Database (NHIS-NHID).

MATERIALS AND METHODS

Ethics statement

The study protocol was approved by the Institutional Review Board of Seoul National University College of Medicine and Seoul National University Hospital as an exempt study on 5 March 2021 (Approval number: E-2103-021-1201).

Patient selection

Data were obtained from the claims of the Korean NHIS, which is a single insurer managed by the Korean government providing healthcare insurance coverage to >97% of residents [14, 15]. This database contains a complete set of medical claims and healthcare information, including patients’ demographic data, administered clinical procedures, diagnosis codes for underlying diseases, and survival information. All diagnoses were coded based on the International Classification of Diseases, 10th Revision [14].

Using the Korean NHIS-NHID data, adult patients (age ≥19) who underwent AVR from 2003 to 2019 were identified using the relevant procedure code ‘Aortic valve replacement’. Patients who underwent AVR using a bioprosthetic valve were selected (Table 1). Sutureless or rapid-deployment valves were excluded. Also, patients who received St Jude Trifecta valves (Abbott, Green Oaks, IL, USA), which were recently withdrawn from the market due to issues with early failure, were also excluded from the analysis. After selection, patients who met the following criteria were excluded from the analysis: (i) any concomitant cardiac procedures, (ii) history of infective endocarditis and (iii) history of open cardiac surgery (Supplementary Material, Table S1). Finally, the study population was classified into 2 groups based on the bioprosthesis used during primary isolated AVR.

Preoperative comorbidities were evaluated based on the diagnostic codes administered within one year before surgery (Supplementary Material, Table S2). Chronic renal failure (CRF) requiring dialysis was evaluated separately using the associated procedure codes. The Charlson comorbidity index was calculated by summing the weights assigned to 17 comorbidities based on diagnosis codes [16] (Supplementary Material, Table S3).

Evaluation of early and follow-up clinical outcomes

Operative mortality was defined as all-cause death within the 30 days after surgery or before discharge during the index admission for AVR. Postoperative stroke was defined as a newly diagnosed stroke after the AVR during the index hospitalization or during the follow-up in the cohort without the previous history of stroke. Postoperative bleeding was defined as patients who were administered the procedure code ‘Bleeding control after cardiac surgery’ during the index hospitalization. Reoperation was defined as the addition of AVR or a redo-AVR procedure code during the follow-up period.

Patients were followed up from the date of the index AVR and censored at the first occurrence of an event of death, or the end of the study period (31 December 2020), whichever occurred first. Because the NHIS is a single compulsory health insurance system for the Korean population, it was anticipated that follow-up would be close to 100%.

Statistical analysis

Statistical analyses were performed using SAS Enterprise Guide 7.15 (SAS Institute, Cary, NC) and R software, version 4.03 (R Foundation for Statistical Computing, Vienna, Austria) with RStudio software version 1.1.463 (RStudio, PBC, Boston, MA). Continuous variables are expressed as the mean (standard deviation) for normally distributed data or the median with interquartile range for data that were not normally distributed. Categorical variables are presented as numbers with percentages of participants. To estimate the propensity score during the propensity-score matching (PSM), we used a logistic regression model that included the following 18 covariates: age, sex, Charlson comorbidity index, year of surgery, type of aortic valve disease, cancer, hypertension, dyslipidaemia, chronic lung disease, coronary artery disease, cerebrovascular disease, atrial fibrillation, peripheral vascular disease, renal disease, congestive heart failure, diabetes mellitus, chronic liver disease and dialysis requirements. To analyse clinical outcomes that could not be differentiated from a previous history or an underlying condition before the index admission, we established separate patient cohorts for stroke and acute kidney injury to minimize the possibility of overestimating the occurrence rate. To minimize the potential bias stemming from institutional variations, we combined PSM with the exact matching of medical institutions. PSM was conducted using a 1:1 nearest-neighbourhood approach (greedy matching) with a caliper width of 0.05. The balance between the 2 groups after PSM was evaluated using standardized mean differences (SMD). An SMD < 0.1 was considered negligible when the groups. The chi-square test was used for categorical variables and the Wilcoxon rank-sum test was used to compare continuous variables between groups before matching. The McNemar test or Bowker’s test and Wilcoxon signed-rank test were used to compare categorical and continuous variables between the matched groups, respectively.

Early clinical outcomes were analysed using multivariate logistic regression analysis and adjusted odds ratios were derived after adjusting for variables that were not balanced between the groups (SMD > 0.1).

For long-term clinical outcomes, Gray’s test was used to evaluate differences in cumulative incidence functions while considering competing risks. The incidence of long-term clinical outcomes per 100 patient-years was calculated using Poisson regression models. To estimate the effect of covariates on the cumulative incidence, Fine-Gray proportional subdistribution hazard models were used to compare the hazards of stroke, late reoperation, infective endocarditis and progression to CRF by considering any death as a competing event [17]. Adjusted hazard ratios (aHR) were derived after adjusting for variables that were not balanced (SMD > 0.1) between the groups in the entire population. We used a sandwich-type robust variance estimator of standard errors to consider the clustering effect at the hospital level when calculating standard errors during the analysis of the entire population. All tests were 2-tailed, and a P-value < 0.05 was considered statistically significant.

RESULTS

Patient selection

According to the exclusion criteria mentioned above, a total of 5470 with primary isolated AVR using bioprosthetic valve were identified and enrolled in the present study (Fig. 1). The study population was classified into BOV (group B, n = 3947) and POV (group P, n = 1523) groups.

The median follow-up duration was 5.9 (standard deviation 4.0) years (interquartile range 2.7–8.6 years).

Bioprosthetic valves used in primary isolated AVR

The use of POV during isolated primary AVR was lower than that of BOV throughout the study period. The proportion of POV implantation gradually increased until 2013; however, it decreased from 2015 (POVs used from 2010 to 2014; mean = 149, from 2015 to 2019; mean = 60) (Supplementary Material, Fig. S1).

The used POV and BOV during the study period are shown in Table 1.

Preoperative characteristics

There were no significant differences in mean age between the groups at the time of surgery. The incidence of aortic stenosis was slightly higher in group B than in group P. The proportion of comorbidities was well balanced between the 2 groups, except for dyslipidaemia (P < 0.001). After PSM, all covariates were well balanced between the 2 groups (SMD < 0.1) (Table 2).

Early clinical outcomes

There was no significant difference in the operative mortality between the groups in both the total cohort and PSM analysis (P = 0.22 and 0.19, respectively). Also, no significant differences were found in the occurrence of bleeding reoperation, stroke (both ischaemic and haemorrhagic), early PPM implantation or acute kidney injury requiring renal replacement therapy between the 2 groups in both total cohort and PSM analysis (Table 3).

Long-term clinical outcomes

All-cause mortality

The cumulative incidence of all-cause mortality was significantly higher in group P than in group B in the overall cohort (aHR: 1.20, P = 0.002) (Table 4, Fig. 2A). However, there was no significant difference in the cumulative incidence of all-cause mortality between the 2 groups in the PSM analysis (P = 0.24) (Table 5, Fig. 2B).

This finding was consistent in the subgroup analysis of both the first half (2003–2009) and second half (2010–2019) periods (Supplementary Material, Table S4).

Stroke, reoperation, infective endocarditis and progression to CRF

There were no significant differences during the follow-up period in the total or in PSM patients in late stroke (both ischaemic and haemorrhagic), infective endocarditis, and progression to CRF (Tables 4 and 5).

The cumulative incidence of reoperation was significantly higher in group P than in group B in the overall cohort (aHR: 1.59, P = 0.003) (Table 4, Fig. 2C). However, there was no significant difference in the cumulative incidence of reoperation between the 2 groups in the PSM analysis (P = 0.29) (Table 5, Fig. 2D).

In the subgroup analysis, during the first half (2003–2009), there was no significant difference in the cumulative incidence of reoperation between the 2 groups in both the overall cohort and PSM analysis. However, in the second half (2010–2019), the cumulative incidence of reoperation was significantly higher in group P compared to group B in the overall cohort, but no significant difference was observed between the 2 groups in the PSM analysis (Supplementary Material, Table S4).

DISCUSSION

The main finding of the present study was that the long-term all-cause mortality and reoperation rate were higher in the POV group than in the BOV group during primary isolated AVR; however, these did not persist after matching. The cumulative incidence of infective endocarditis was comparable in the total cohort and in the PSM patients.

Despite the long history of valve replacement using bioprosthetic valves, controversies persist over whether BOV or POV is better. Various studies have compared long-term survival according to the prosthetic valve leaflet material. A study based on the England-Wales National Registry demonstrated no significant difference in survival outcomes between patients undergoing AVR with BOV or POV, irrespective of concomitant coronary artery bypass grafting [18]. Recently, Kim et al. reported that both cardiovascular and non-cardiovascular mortality rates were not statistically different between the 2 types of valves during AVR in a large NHIS-NHID-based study using data similar to those of the present study [19]. However, that study did not exclude patients who underwent concomitant procedures or those who received St Jude Trifecta valves (Abbott, Green Oaks, IL, USA), which were withdrawn in 2023 due to its higher incidence of early structural valve deterioration (SVD). In contrast, a study based on the Swedish population registry demonstrated better long-term survival with POV than with BOV in patients who underwent AVR with or without coronary artery bypass grafting [11].

Long-term durability remains a significant concern in heart valve surgery involving bioprosthetic valves, particularly in the aortic position. The aortic position typically demonstrates shorter durability compared to the mitral position, primarily due to the high systolic and diastolic pressures, which contribute to SVD and increase the likelihood of reoperation. The SVD of a bioprosthetic valve usually progresses through 2 main mechanisms. In BOV, leaflet thickening and stiffening with calcification are the main causes of SVD, resulting in gradually progressive valve stenosis. Cusp tears or perforations are the main causes of SVD in POV, resulting in acute valve insufficiency [7, 11, 20]. In a previous study conducted at our institution, patients undergoing AVR with POV demonstrated a higher incidence of moderate or SVD and valve haemodynamic deterioration compared to those with BOV. However, no significant difference was observed in the reoperation rates between the 2 groups [6]. Conversely, previous studies based on the Swedish population registry and NHIS-NHID also showed a higher reoperation rate in patients who underwent AVR with POV [11]. However, most previous studies had factors that could affect long-term results, such as the inclusion of patients with concomitant procedures or infective endocarditis.

Although many studies have compared BOV and POV, mixed and inconclusive results persist due to the diverse factors influencing long-term outcomes of valve replacement. In this nationwide data-based study, efforts were made to limit variables. First, we conducted a comparative analysis in patients undergoing primary isolated AVR with bioprosthetic valves, excluding those with concomitant heart procedures. Second, we accounted for institutional variations in surgical outcomes by clustering data by institution and matching patients within the same institution during PSM analysis. Third, we excluded patients with valves that were no longer in use including St Jude Trifecta valves (Abbott, Green Oaks, IL, USA), as those were removed from the reimbursement list between 2003 and 2019 to focus on more recent valve products. Finally, with these considerations, the present Korean NHIS-NHID-based study demonstrated no statistically significant difference in long-term clinical outcomes between BOV and POV during primary isolated AVR.

The long-term durability of bioprosthetic valves is determined by many factors, including not only whether the material used for the valve is bovine or porcine but also the way the valve is produced, the surgical skills and the way the patients are managed during follow-up. Ultimately, a randomized prospective study that controls as many factors as possible, including the valve manufacturing process, is needed to determine which of the materials, bovine or porcine, is superior in the aortic valve position.

Study limitations

This study had some limitations. First, this was a retrospective observational study, and the data were extracted from the Korean NHIS-NHID; therefore, some variables that could affect clinical outcomes, such as individual patient characteristics, echocardiographic follow-up data, or risk predicting scores, such as the EuroSCORE II, Society of Thoracic Surgery score or NYHA functional class could not be collected. Although a large study cohort was constructed with strict variable control and PSM analyses were performed to overcome the retrospective nature of this study, some unrevealed bias may remain due to unmeasured confounders affecting the results. Second, the different mechanisms of SVD between the 2 types of bioprostheses and their clinical manifestations could influence the patient’s symptoms and timing of reoperation. As shown in our previous study, the more frequent SVD and valve haemodynamic deterioration did not correlate with more reoperations. Third, the proportion of patients receiving POV varied across different years, which could introduce potential calendar bias. To mitigate this, we divided the surgical time period into 3 intervals and included this as a matching variable, ensuring that patients who underwent surgery within a similar time period were matched to minimize time bias. Also, we also performed a subgroup analysis of first-half and second-half period patients, respectively (Supplementary Material, Table S4). Additionally, sensitivity analyses were performed focusing on patients who underwent surgery in the most recent 5 years for overall mortality, late reoperation and late infective endocarditis (Supplementary Material, Table S5).

CONCLUSIONS

In patients with primary isolated AVR according to the Korean NHIS database, there was no significant difference in early and long-term clinical outcomes, including overall survival and reoperation, between BOV and POV.

Footnotes

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

This study was supported by Seoul National University Hospital (Project number: 0420212110).

Conflict of interest: none declared.

DATA AVAILABILITY

Data will be shared on reasonable request to the corresponding author, with the permission of the National Health Insurance Service of Republic of Korea.

Author contributions

Ji Seong Kim: Conceptualization; Data curation; Funding acquisition; Investigation; Methodology; Project administration; Resources; Validation; Visualization; Writing—original draft; Writing—review & editing. Jinhee Kim: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Software; Supervision; Validation. Yoonjin Kang: Conceptualization; Funding acquisition; Project administration; Resources; Supervision; Validation. Suk Ho Sohn: Conceptualization; Funding acquisition; Methodology; Project administration; Resources; Supervision; Validation. Ho Young Hwang: Conceptualization; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Validation. Kyung Hwan Kim: Conceptualization; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Validation. Mi-Sook Kim: Data curation; Formal analysis; Investigation; Methodology; Software; Supervision; Validation; Visualization; Writing—review & editing. Jae Woong Choi: Conceptualization; Funding acquisition; Investigation; Resources; Supervision; Validation; Writing—original draft; Writing—review & editing

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Marjan Jahangiri, Alfred Kocher and the other anonymous reviewers for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• aHR

  Adjusted hazard ratio

 
• AVR

  Aortic valve replacement

 
• BOV

  Bovine pericardial valves

 
• CRF

  Chronic renal failure

 
• NHIS-NHID

  Korean National Health Insurance Service-National Health Information Database

 
• POV

  Porcine valves

 
• PPM

  Permanent pacemaker

 
• SMD

  Standardized mean differences

 
• SVD

  Structural valve deterioration

Author notes