Effects of mitral valve disease etiology on the outcomes of mechanical and biological valve replacement: retrospective cohort study

Abstract

Introduction:

The choice of an artificial mitral valve (MV) is a crucial clinical decision that affects the long-term survival and quality of life of patients. However, current guidelines recommend selecting MV based on patient age and life expectancy at the time of mitral valve replacement (MVR), without considering the etiology of MV disease. This study aimed to investigate whether MV disease etiology should be considered when choosing a valve for MVR and to evaluate the impact of MV disease etiology on long-term patient survival.

Methods:

Using data (2002–2018) from Taiwan’s National Health Insurance Research Database, the authors conducted a nationwide retrospective cohort study to compare the biological and mechanical valves in terms of all-cause mortality as the primary outcome. The inverse probability of the treatment weighting method was used to reduce the effects of the confounding factors. The following etiologies were assessed: infective endocarditis, rheumatic heart disease, ischemic mitral regurgitation, and degenerative mitral regurgitation.

Results:

In patients aged below 70 years, it was observed that mechanical valves demonstrated an association with benefits compared to biological valves in the context of survival. In patients with infective endocarditis aged below 72 years, mechanical valves were associated with survival benefits, but not in those with stroke during hospitalization. These valves were also found to be linked with survival advantages for patients with rheumatic heart disease aged below 60 years and for those with degenerative mitral regurgitation aged below 72 years. However, no age-dependent effects of valve type on all-cause mortality were observed in patients with ischemic mitral regurgitation.

Conclusion:

The etiology of MV disease appears to be important in the selection of a suitable MV and determination of a cutoff age for mechanical and biological MVR.

Introduction

Highlights

• We used nationwide data in this population-based retrospective cohort study to clarify whether mitral valve (MV) disease etiology should be considered when selecting valves for mitral valve replacement (MVR) and to evaluate the effects of MV disease etiology on long-term patient survival.

• When the effects of MV disease etiology are not considered, mechanical valves outperformed biological valves in terms of survival in patients aged below 70 years.

• The underlying etiology of MV disease can have a profound impact on long-term survival.

• Patients who underwent MVR due to ischemic mitral regurgitation had a lower probability of long-term survival than patients with infective endocarditis, rheumatic heart disease, or degenerative mitral regurgitation.

• The etiology of MV disease is an important factor in selecting an appropriate valve and determining the cutoff age for mechanical and biological MVRs.

• Survival benefits with mechanical valves held for infective endocarditis patients under 72 years without stroke, rheumatic heart disease patients under 60 years, and degenerative mitral regurgitation patients under 72 years. Mechanical valves offer limited survival benefits and are linked to higher risks of stroke and bleeding for patients with ischemic mitral regurgitation.

The prevalence of mitral valve (MV) disease, a major cardiovascular condition¹, has been increasing worldwide². Structural abnormalities in MVs may be associated with various pathologies such as infective endocarditis (IE), rheumatic heart disease (RHD), age-related MV calcification, degenerative changes, ischemic heart disease (IHD), cardiomyopathy, and congenital malformations. These abnormalities may lead to valvular stenosis, regurgitation, or both³. MV repair and replacement (MVR) are standard therapies for symptomatic patients. A growing number of surgeons are advocating for valve repair as a viable approach to addressing MV diseases^3–5, particularly degenerative mitral regurgitation (DMR)^3,4 and IE-related mitral regurgitation⁶. However, MVR, which prolongs patient survival, remains the treatment of choice if valve repair is not possible⁷.

Studies on the selection of prosthetic valves for MVR (i.e. mechanical vs. biological) have reported inconclusive findings^4,8–10, because each valve type has its own advantages and disadvantages. Mechanical valves are highly durable; however, they necessitate lifelong use of anticoagulants, which are associated with the risk of hemorrhage and thromboembolism^7,11. Although biological valves can overcome this problem, their durability is limited. Patient age and life expectancy at the time of MVR are crucial for valve selection (mechanical vs. biological). Current clinical guidelines recommend mechanical valves for patients aged below 65 years. Although the underlying cause of MV disease may affect long-term clinical outcomes^12,13, the etiology of MV disease is not currently considered when selecting valves or determining a cutoff age for mechanical and biological MVR.

Therefore, we conducted a nationwide retrospective cohort study to compare mechanical and biological MVR in terms of long-term survival. We hypothesized that the etiology of MV disease influences the effects of valve type on long-term survival. We further sought to determine a suitable cutoff age for the selection of valves (mechanical or biological) according to the etiology of MV disease.

Methods

Data source

Relevant data were obtained from the Taiwan National Health Insurance Research Database (NHIRD). The Taiwan National Health Insurance (NHI) program, launched in 1995, currently includes data on ∼99.9% of all residents of Taiwan^14,15. The NHIRD contains all registration and claims data, including patients’ personal information, ambulatory and inpatient claims, diagnostic codes [International Classification of Diseases, Ninth Revision or Tenth Revision, Clinical Modification (ICD-9-CM or ICD-10-CM, respectively)], medical facility registry data, prescriptions, and surgical records. All NHIRD data were deidentified and anonymized; therefore, informed consent was not mandatory for using NHIRD data. This study was approved by the Institutional Review Board of our institution.

Study population

We included patients who underwent MVR at medical centers in Taiwan between January 1, 2001 and December 31, 2018. Eligible patients were identified based on procedural codes (ICD-9-CM before 2016 and ICD-10-CM after 2016) and NHI reimbursement and supply codes. We included only patients who underwent surgery at medical centers because multitudinous valve options are available at these centers. Internal validation was performed to confirm the accuracy of the data used for patient inclusion¹². Initially, 10 531 patients who underwent MVR during the study period were identified from the database. The exclusion criteria were as follows: lack of demographic data (age and sex), received both repair and replacement surgeries for MV (e.g. valve replacement after a failed valve repair), history of valve surgery, age below 20 years, MV disease associated with etiologies other than those of interest, and the combined use of biological and mechanical valves for MVR. After the exclusion criteria were applied, 9201 patients remained; of them, 3253 were implanted with biological valves and 5948 were implanted with mechanical valves (Fig. 1A). To ensure the quality of the research, the authors followed the STROCSS guidelines, Supplemental Digital Content 1, http://links.lww.com/JS9/C115 for reporting surgical studies of different designs¹⁶. In addition, distinctions between the present retrospective cohort study and a randomized controlled trial were delineated through a comparative table using the target-trial emulation approach (Supplementary Table 1, Supplemental Digital Content 2, http://links.lww.com/JS9/C116).

Figure 1.

The inclusion and exclusion of the study patients (A), the proportion of bioprosthesis valve among mitral valve replacement surgeries across the study period (B) and the all-cause mortality of patients with different etiologies (C). DMR, degenerative mitral regurgitation; IE, infective endocarditis; RHD, rheumatic heart disease; IMR, ischemic mitral regurgitation. *‘Combined use of biological and mechanical valves’ refers to the use of either mechanical or biological valves initially in mitral valve replacement valve surgery, but the use of another type of valve later due to various clinical considerations.

MV disease etiology

We developed an algorithm to determine the etiology of MV disease in each patient. First, patients who received an inpatient diagnosis of IE, RHD, ischemic mitral regurgitation (IMR), or DMR during index hospitalization were classified accordingly. Patients who received two or more of the aforementioned four diagnoses during the index hospitalization were reclassified as having a single diagnosis, with priority assigned as follows: IE, RHD, IMR, and DMR. The rationale for prioritization was as follows. IE is a condition of acute nature, and it significantly influences the prognosis of valve replacement. In cases where there were two or more concurrent diagnoses, with IE being one of them, the prognostic impact of IE was considered to surpass that of the other identified conditions. Some studies indicated that the diagnostic code for IE is precise. Consequently, we prioritized the identification of IE based on its diagnostic accuracy and consequential prognostic implications for valve replacement. RHD is a prevalent cardiovascular condition globally, with Taiwan having a well-defined diagnostic code specifically for RHD. The diagnosis of ischemic heart disease relied on a patient’s history of myocardial infarction or documented involvement in percutaneous coronary intervention or coronary artery bypass graft surgery. Furthermore, for patients who did not receive any of the four diagnoses during the index hospitalization, we referred to their previous inpatient diagnoses if available. Patients who received none of the four diagnoses during the index hospitalization or previous hospitalization were excluded from the analysis. The electronic medical records of our hospital were used to validate the accuracy of the algorithm. The surgical records and discharge notes of patients who underwent MV surgery at the study centers between 2002 and 2018 were reviewed to determine etiology. We then compared the diagnoses with the results of the algorithm in terms of accuracy in etiology classification as follows: IE (93.2%), RHD (94.9%), IMR (87.5%), and DMR (88.4%) (results not shown).

Covariates

The following baseline data were collected: year of index surgery, demographics and comorbidities of patients, cumulative volume of valve surgery at each hospital, and information on additional concomitant surgeries. The year of index surgery was considered to adjust for the effects of a secular trend. Demographic characteristics included age, sex, urbanization level, and monthly income. Concomitant surgeries included aortic valve surgery, tricuspid valve surgery, coronary artery bypass, the maze procedure, and aortic surgery. Comorbidities included stroke, diabetes mellitus, and 11 other comorbidities. The inclusion of a comorbidity required at least two outpatient diagnoses or one inpatient diagnosis of a comorbidity in the previous year. For patients with IE, we additionally collected data on the status of IE (active or prior) and history of illicit drug use^17,18. For patients with RHD, we additionally collected data on the history of percutaneous transvenous mitral commissurotomy and type of MV disease (mitral regurgitation or stenosis)¹⁹.

Outcome variables

The outcome variables included in-hospital and delayed outcomes. The primary outcome of interest was all-cause mortality during follow-up. Relevant data were obtained from the Taiwan Death Registry (Health and Welfare Data Science Center database), which contains information on the dates, places, and causes of death of individuals. The in-hospital outcomes included stroke, re-operation for bleeding, new-onset dialysis, deep wound infection, and 30-day mortality during the entire course of hospitalization. Delayed outcomes included all-cause mortality, repeat MV surgery, IE, stroke, major bleeding, and composite valve complications. Data on date, place, and cause of death were linked to the Taiwan Death Registry. The aforementioned in-hospital and delayed outcomes were defined according to the definitions in the literature^12,17–19. For the analysis of late outcomes (follow-up outcomes), patients were followed up from the day of discharge until the day of outcome occurrence, day of death, or end of relevant database entry (December 31, 2018), whichever occurred first.

Statistical analysis

To adjust for the effects of potential confounders in this observational study, we used the inverse probability of treatment weighting (IPTW) method, with the average treatment effect estimated based on propensity scores. The IPTW method was used for the entire cohort and for patients with each of the four etiologies. Propensity scores were derived using a multivariate logistic regression model, in which MVR performed with a mechanical or biological valve was the dependent variable. All baseline data were included in the logistic regression model to obtain propensity scores; the follow-up year was replaced with the year of the index surgery. To prevent extreme weights from affecting the results, we truncated the outlier weights to the 99th percentile of the weights. A between-group standardized difference of below 0.2 indicated a nonsubstantial difference.

Using a logistic regression model, we compared in-hospital outcomes between patients who received a mechanical valve and those who received a biological valve. The risk of all-cause mortality was compared between the two prosthesis groups using the Cox proportional-hazard model. The risk of nonfatal delayed outcomes (e.g. redo MV surgery) was compared between the groups using the Fine–Gray subdistribution hazard model, in which all-cause mortality during follow-up was regarded as a competing risk. Notably, the study group inclusion was the only explanatory variable in the regression models. To explore the age-dependent effects of valve type on all-cause mortality, we used a Cox proportional-hazard model including an interaction term for age (modeled as a natural spline) and valve type (biological vs. mechanical) in the matched study population. Continuous age was treated as a flexible restricted cubic spline with four knots at the 5th, 35th, 65th, and 95th percentiles. Subsequently, this model was used to obtain the hazard ratios (HRs; biological valve vs. mechanical valve) and the corresponding 95% CIs for the etiologies.

Statistical significance was defined as a two-sided P value below 0.5 indicated statistical significance. Restricted cubic spline modeling was performed using R (version 4.0.2; R Project for Statistical Computing) and the “rms” package. All other statistical analyses were performed using SAS (version 9.4; SAS Institute Inc., Cary, North Caroline).

Results

Patient characteristics

In total, 10,531 patients underwent MVR during the study period. Of these, 9201 were included in this study: 3253 underwent biological MVR and 5948 underwent mechanical MVR. Table 1 presents the baseline characteristics of the patients before and after IPTW. Patients who received a biological valve (mean age: 65.5±12.1 y) were older than those who received a mechanical valve (mean age: 57.7±13.3 y). Biological valve recipients were more likely to have hypertension and undergo a concomitant maze procedure than mechanical valve recipients (Table 1). After IPTW, no substantial difference was observed between the biological and mechanical valve recipients in terms of covariates (absolute standardized difference <0.2). We further grouped the patients according to disease etiology (IE, RHD, IMR, or DMR). Supplementary Tables 2–5, Supplemental Digital Content 2, http://links.lww.com/JS9/C116 present the baseline characteristics of patients who underwent MVR for IE, RHD, IMR, and DMR, respectively. Notably, for all MV diseases, the proportion of biological valve use increased throughout the study period (Fig. 1B).

Table 1.

Baseline characteristics of patients who received mitral valve replacement with biological versus mechanical prosthesis before and after weighting.

	Before weightinga			After weightingb
Variables	Biological (n=3253)	Mechanical (n=5948)	STD	Biological	Mechanical	STD
Etiology
Infective endocarditis (IE)	546 (16.8)	958 (16.1)	0.02	15.9	16.6	-0.02
Rheumatic heart disease (RHD)	822 (25.3)	2176 (36.6)	-0.25	27.3	31.8	-0.10
Ischemic mitral regurgitation (IMR)	655 (20.1)	941 (15.8)	0.11	18.8	17.1	0.04
Degenerative mitral regurgitation (DMR)	1230 (37.8)	1873 (31.5)	0.13	38.0	34.6	0.07
Year of the index surgery
2001–2003	181 (5.6)	1136 (19.1)	-0.42	12.9	15.0	-0.06
2004–2006	191 (5.9)	1282 (21.6)	-0.47	13.4	16.3	-0.08
2007–2009	399 (12.3)	1315 (22.1)	-0.26	17.5	19.0	-0.04
2010–2012	649 (20.0)	973 (16.4)	0.09	18.6	17.1	0.04
2013–2015	896 (27.5)	766 (12.9)	0.37	19.9	17.5	0.06
2016–2018	937 (28.8)	476 (8.0)	0.56	17.7	15.1	0.07
Age (years)	65.5±12.1	57.7±13.3	0.62	61.4±21.2	60.0±16.1	0.08
Age group (years)
<50	333 (10.2)	1556 (26.2)	-0.42	18.2	21.2	-0.08
50–64	998 (30.7)	2,450 (41.2)	-0.22	35.9	38.0	-0.04
65–79	1594 (49.0)	1701 (28.6)	0.43	39.1	35.2	0.08
≥80	328 (10.1)	241 (4.1)	0.24	6.8	5.6	0.05
Male sex	1612 (49.6)	3142 (52.8)	-0.07	50.7	50.9	0.00
Urbanization level
Low	464 (14.3)	834 (14.0)	0.01	14.1	14.2	0.00
Moderate	829 (25.5)	1583 (26.6)	-0.03	24.3	25.9	-0.04
High	1259 (38.7)	2226 (37.4)	0.03	38.9	37.6	0.03
Very high	701 (21.5)	1305 (21.9)	-0.01	22.7	22.3	0.01
Monthly income, NTD
Tertile 1	970 (29.8)	1943 (32.7)	-0.06	32.3	31.9	0.01
Tertile 2	542 (16.7)	1084 (18.2)	-0.04	17.2	17.4	0.00
Tertile 3	1741 (53.5)	2921 (49.1)	0.09	50.5	50.7	0.00
Comorbid conditions
Old stroke	399 (12.3)	601 (10.1)	0.07	11.6	10.6	0.03
Diabetes mellitus	846 (26.0)	1208 (20.3)	0.14	23.3	22.2	0.03
Hypertension	1821 (56.0)	2683 (45.1)	0.22	50.9	48.3	0.05
Heart failure	1437 (44.2)	2309 (38.8)	0.11	41.8	39.9	0.04
Coronary arterial disease	1305 (40.1)	2325 (39.1)	0.02	40.4	39.4	0.02
Atrial fibrillation	1635 (50.3)	2821 (47.4)	0.06	48.7	47.8	0.02
Chronic kidney disease	794 (24.4)	1012 (17.0)	0.18	21.3	19.3	0.05
Dialysis	180 (5.5)	206 (3.5)	0.10	4.8	4.2	0.03
Chronic obstructive pulmonary disease	435 (13.4)	675 (11.3)	0.06	13.0	11.9	0.03
Liver cirrhosis	88 (2.7)	126 (2.1)	0.04	2.3	2.2	0.01
Major bleeding history	572 (17.6)	830 (14.0)	0.10	15.5	14.4	0.03
Gastrointestinal bleeding history	489 (15.0)	694 (11.7)	0.10	13.0	12.3	0.02
Myocardial infarction history	252 (7.7)	344 (5.8)	0.08	6.7	6.2	0.02
Cumulative hospital volume of valve surgery between 2001 and 2018
First quartile	726 (22.3)	1399 (23.5)	-0.03	22.9	22.8	0.00
Second quartile	982 (30.2)	1225 (20.6)	0.22	25.5	23.8	0.04
Third quartile	735 (22.6)	1277 (21.5)	0.03	23.9	22.4	0.04
Fourth quartile	810 (24.9)	2047 (34.4)	-0.21	27.7	31.0	-0.07
Additional valve surgery
Aortic valve repair or replacement	655 (20.1)	1200 (20.2)	0.00	10.6	16.6	-0.18
Tricuspid valve repair or replacement	707 (21.7)	1030 (17.3)	0.11	20.2	18.4	0.05
Additional surgery
Coronary arterial bypass surgery	563 (17.3)	910 (15.3)	0.05	17.2	16.0	0.03
Maze	882 (27.1)	817 (13.7)	0.34	20.9	18.5	0.06
Aorta surgery	53 (1.6)	85 (1.4)	0.02	1.3	1.4	-0.01
Follow-up year	4.0±3.7	7.2±5.2	-0.72	5.5±7.1	6.2±6.1	-0.11

^aData are presented as frequency (percentage) or mean±SD.

^bData are presented as percentage or mean±SD.

NTD, New Taiwan Dollar; STD, standardized difference.

Primary outcome: all-cause mortality

The mean follow-up duration was 6.1±4.7 years. During this period, the cumulative post-MVR mortality rate was 41.4%. As shown in Figure 1C, the mortality rate was higher in patients with IMR (63.0%) than in those with IE (36.5%), RHD (34.4%), or DMR (39.3%). Biological MVRs were associated with higher mortality than mechanical MVRs as follows: IE (HR: 1.14; 95% CI: 1.01–1.30), RHD (HR: 1.28; 95% CI: 1.16–1.41), and DMR (HR: 1.29; 95% CI: 1.19–1.31). In patients with IMR, valve type had no significant effect on mortality (Table 2).

Table 2.

Late outcomes of patients who received mitral valve replacement with biological versus mechanical prosthesis in the IPTW-adjusted cohort

Outcome/etiology	Total	Biological	Mechanical	HR or SHR (95% CI) of biological	P value
All-cause mortality
IE	36.5	37.4	35.8	1.14 (1.01–1.30)	0.036
RHD	34.4	35.4	33.7	1.28 (1.16–1.41)	<0.001
IMR	63.0	64.6	61.6	1.08 (0.98–1.18)	0.105
DMR	39.3	42.9	35.8	1.29 (1.19–1.40)	<0.001
The whole cohort	41.4	43.8	39.3	1.24 (1.19–1.31)	<0.001
Redo mitral valve surgery
IE	5.6	9.0	2.6	4.05 (2.78–5.88)	<0.001
RHD	5.0	8.9	2.0	5.74 (4.26–7.73)	<0.001
IMR	1.5	2.2	0.9	2.45 (1.30–4.60)	0.006
DMR	3.3	4.8	1.9	2.64 (1.93–3.60)	<0.001
The whole cohort	3.8	5.9	1.9	3.45 (2.89–4.12)	<0.001
Infective endocarditis
IE	4.9	6.0	4.0	1.58 (1.12–2.23)	0.010
RHD	2.2	2.3	2.1	1.22 (0.83–1.78)	0.314
IMR	2.1	2.2	2.0	1.08 (0.65–1.77)	0.772
DMR	2.5	2.6	2.4	1.06 (0.77–1.46)	0.721
The whole cohort	2.7	2.9	2.5	1.20 (1.000–1.45)	0.0495
Stroke
IE	9.5	9.0	10.0	0.95 (0.74–1.21)	0.667
RHD	14.7	13.5	15.6	0.95 (0.82–1.11)	0.519
IMR	12.5	13.7	11.4	1.23 (1.003–1.51)	0.047
DMR	11.8	11.1	12.5	0.88 (0.76–1.03)	0.102
The whole cohort	12.3	11.6	12.9	0.93 (0.86–1.02)	0.123
Major bleeding
IE	8.2	6.2	9.9	0.66 (0.50–0.87)	0.003
RHD	11.8	10.1	13.0	0.85 (0.72–1.01)	0.068
IMR	10.2	9.1	11.2	0.80 (0.63–1.00)	0.051
DMR	9.4	7.5	11.3	0.64 (0.54–0.76)	<0.001
The whole cohort	10.1	8.3	11.6	0.73 (0.66–0.81)	<0.001
Composite valve complicationa
IE	14.4	12.7	15.8	0.83 (0.68–1.01)	0.068
RHD	21.3	19.5	22.6	0.95 (0.84–1.08)	0.444
IMR	18.8	19.4	18.2	1.07 (0.91–1.27)	0.407
DMR	17.6	15.9	19.2	0.80 (0.71–0.91)	<0.001
The whole cohort	18.4	16.9	19.6	0.88 (0.82–0.94)	<0.001

Data not specified are expressed as percentage.

^aAnyone with stroke and major bleeding.

DMR, degenerative mitral regurgitation; HR, hazard ratio; IE, infective endocarditis; IMR, ischemic mitral regurgitation; IPTW, inverse probability of treatment weighting; RHD, rheumatic heart disease; SHR, subdistribution hazard ratio.

Secondary outcomes

The in-hospital outcomes are listed in Supplementary Table 6, Supplemental Digital Content 2, http://links.lww.com/JS9/C116. Regarding delayed outcomes, biological valve recipients were more likely to require MV redo surgery than mechanical valve recipients [subdistribution HR (SHR): 3.45; 95% CI :2.89–4.12]. Among the biological valve recipients, only those with IE had an elevated risk of IE recurrence during the follow-up period (SHR: 1.58; 95% CI: 1.12–2.23). Finally, biological MVR was associated with lower risks of major bleeding (SHR: 0.73; 95% CI: 0.66–0.81) and composite valve complications (SHR: 0.88; 95% CI: 0.82–0.94) than mechanical MVR (Table 2).

Age-dependent HR for all-cause mortality

Among patients aged below 70 years, mechanical valve recipients survived longer than biological valve recipients did (Fig. 2). In IE patients, no age-dependent effects of valve type were observed on all-cause mortality (Fig. 3A). However, IE is associated with an increased risk of preoperative ischemic and hemorrhagic stroke²⁰, which may influence the valve selection. This prompted us to exclude patients who had stroke at the time of admission (9.9% of the cohort). Subsequent analyses revealed that mechanical valves were associated with survival benefits in patients aged below 72 years (Fig. 3B). Mechanical valves were also associated with survival benefits in patients with RHD aged below 60 years (Fig. 4A) and those with DMR aged below 72 years (Fig. 4C). However, for patients with IMR, no age-dependent effect of the valve type was noted on all-cause mortality (Fig. 4B).

Figure 2.

The age-dependent hazard ratio of all-cause mortality for bioprosthetic valves relative to mechanical valves in the inverse probability of treatment weighting-adjusted cohort.

Figure 3.

The age-dependent hazard ratio of all-cause mortality for bioprosthetic valves relative to mechanical valves in patients whose etiology were infective endocarditis (IE) (A) and further excluded the patients with stroke during the index admission (B) in the inverse probability of treatment weighting (IPTW)-adjusted cohort.

Figure 4.

The age-dependent hazard ratio of all-cause mortality for bioprosthetic valves relative to mechanical valves in patients whose etiology were rheumatic heart disease (RHD) (A), ischemic mitral regurgitation (IMR) (B), and degenerative mitral regurgitation (DMR) (C) in the inverse probability of treatment weighting (IPTW)-adjusted cohort.

Discussion

Among younger patients (below 70 y) who underwent MVR during the study period, the use of mechanical valves was associated with a lower risk of mortality during follow-up compared to biological valve recipients. The etiology of MV disease may considerably affect the long-term survival of patients. Patients with IMR were less likely to survive long-term than those with IE, RHD, or DEG. Furthermore, etiology is an important factor in determining the cutoff age for valve selection. In patients with IE, RHD, or DMR, mechanical valves were associated with a higher survival rate than biological valves in younger recipients. The cutoff age below which mechanical valves were deemed advantageous was lower for patients with RHD (∼60 y) than for those with IE or DMR (∼72 y). In patients with IHR, the choice of valve did not significantly affect the long-term survival.

When the effects of MV disease etiology are not considered, a biological valve is a reasonable choice for patients aged above 70 years. This notion was supported by the findings of several high-quality studies^7,21,22. Practical guidelines based on observational studies regard patient age at MVR as the primary factor in valve selection. Current clinical guidelines recommend using mechanical valves for patients aged below 65 years and biological valves for older patients³. However, this selection process largely disregards disease etiology^12,13. This prompted us to conduct the present population-based study on the association between MV disease etiology and age-dependent long-term survival.

IE, an endocardial surface infection, is commonly associated with acute MV failure^23,24. In cases where MVR is indicated for IE, such as prosthetic valve endocarditis and extended infection on more than one leaflet²⁴, clinicians must consider the benefits of mechanical valves over biological valves. This has been a contentious issue for decades. However, some studies have reported no significant difference between the two devices when other risk factors were not considered^25,26, other studies have advocated the use of mechanical valves to reduce the risk of disease recurrence and to improve the likelihood of long-term survival^27,28. In the present study, we observed a difference between mechanical and biological valves in terms of long-term survival (Fig. 3A). Regardless of the patient age, mechanical valves were associated with more advantages than biological valves. We performed further analyses excluding patients who had a stroke during hospitalization, as these patients had an elevated risk of bleeding and coagulation. Our findings suggest that in IE patients younger than ∼72 years of age without concomitant stroke, mechanical valves are linked with greater survival benefits when compared to biological valves (Fig. 3B).

For patients with RHD, mechanical valves are generally recommended over biological devices^29–31; however, the optimal cutoff age for valve selection remains to be established. In the present and previous studies¹², we found that mechanical valves are associated with long-term survival benefits in patients aged below 60 years. The situation is similar in patients with DMR; however, for these patients, the cutoff age for long-term survival benefits of mechanical valves is ∼72 years.

IMR, which is a severe complication of coronary artery disease, results from the chronic pathological remodeling of the left ventricle and is associated with high morbidity and mortality risks³². MV repair and MVR are effective treatments for IMR; however, increasing evidence suggests that the risk of disease recurrence is lower after MVR than after MV repair^33,34. Furthermore, patients undergoing MVR survive for as long as or longer than those undergoing MV repair^33–35. In our patients with IMR, valve selection (mechanical or biological) did not have a significant effect on long-term survival.

Moreover, the expected survival duration was shorter in patients with IMR than in those with other MV diseases (Fig. 1C). Thus, the low postoperative life expectancy of patients with IHD may partially offset the durability of the mechanical valves. In younger patients (∼42–74 y old), mechanical valves are associated with slightly advantage in terms of survival duration; however, this benefit is insufficient to infer that these valves are inherently superior. Overall, mechanical valves did not provide significant survival benefits. Studies have reported that mechanical valves are associated with an elevated risk of stroke and bleeding^7,13. Thus, considering the postoperative quality of life, biological valves appear to be a viable alternative for patients with IHD.

Our findings (Fig. 1B) demonstrate a distinct preference for the utilization of bioprosthetic valves in MVR over the last two decades. Despite our research data and previous study^7,36 suggesting a mortality benefit of mechanical valves in most patients under 70 years of age, to avoid subjecting patients to lifelong anticoagulation, bioprosthetic valves are often considered the preferable option in cases where MV repair is not possible. In addition, anticipated advances in valve-in-valve (ViV) technology may facilitate the popularity of bioprosthetic valves.

Our study had some limitations. First, because this was a retrospective study conducted using claims data, the probability of selection bias and effects of unmeasured confounders cannot be ruled out. The NHIRD does not include information on lifestyle, which may influence the prognosis of cardiovascular disease. However, because randomized controlled trials conducted by balancing mechanical and biological valve distributions cannot be considered an acceptable approach for comparing long-term outcomes for this topic, we think that using real-world clinical data represents a viable and valuable research methodology³⁷. Second, the diagnostic codes used to identify preexisting comorbidities did not provide information on disease severity or duration; moreover, missing and inaccurate data were likely¹⁵. Nevertheless, we validated the diagnostic codes by analyzing relevant electronic medical records¹². Third, in the nonrandomized controlled trial study, we acknowledge that, despite employing careful statistical methods, we can only mitigate the impact of confounding variables to a certain extent and recognize the potential for residual confounding. Therefore, we characterized the correlation between valve type and mortality as associative, avoiding implications of a causal relationship. Finally, this was a population-based study conducted in Taiwan. The generalizability of our findings is limited. These findings might have been influenced by the incidence of RHD, which differs between the developed and developing countries. In Taiwan, the incidence of illicit drug use-related IE is low and relatively few patients with IMR undergo heart transplantation. The prevailing conditions in this context may exhibit variances when compared to other countries. Consequently, the outcomes of this study may not be universally representative.

In conclusion, when considering MV disease, in the case of a combination of four etiologies, the choice of a mechanical valve over a bioprosthetic valve is associated with a long-term survival advantage for patients under 70 years of age. The etiology of MV disease is an important contributor to overall outcomes, and thus should be considered in the selection of the valve type and the determination of the cutoff age for valve selection. Among patients with IE without concomitant stroke, RHD, or DMR, mechanical valve recipients survived longer than biological valve recipients did. Specifically, the cutoff age for patients with RHD was lower than that for patients with IE or DMR. The long-term survival was shorter in patients with IMR than in those with other etiologies. Valve type exerted a limited or no effect on survival duration. Notably, structural degeneration of biological valves can be remedied by performing a transcatheter ViV procedure, a technique expected to have significant progress and development. Taken together, the results revealed that, in addition to age, MV disease etiology and ViV procedure should be considered when comparing mechanical and biological valves in terms of survival benefits.

Ethical approval

This study was approved by the Institutional Review Board of our institution (IRB No. 202201859B0).

Consent

All data were obtained from the Taiwan National Health Insurance Research Database (NHIRD). All NHIRD data were deidentified and anonymized; therefore, informed consent was not mandatory or using NHIRD data. This study was approved by the Institutional Review Board of the Chang Gung Memorial Hospital, Taiwan.

Sources of funding

This work was supported by a grant from Chang Gung Memorial Hospital, Taiwan CFRPG3M0041(CYC) and CORPG3M0371, CMRPG3L0101, CMRPG3L0102, CMRPG3L0103, CFRPG3M0011, BMRPD95(SWC). This work was also supported by the Ministry of Science and Technology grant MOST-110-2314-B-182A-114 (SWC). The authors have reported that they have no relationships relevant to the contents of this paper to disclose. This study was based on data from the NHIRD provided by the NHI administration, Ministry of Health and Welfare of Taiwan, and managed by the National Health Research Institutes of Taiwan. However, the interpretation and conclusions contained in this paper only represent the authors.

Author contribution

S.W.C. was responsible for conceptualization, data curation, funding acquisition, investigation, and project administration. The authors acknowledge their roles in the final manuscript as follows: C.Y.C., F.C.C., C.P.L., and C.H.Y.: investigation, methodology, and validation. Y.H.C., V.C.C.W., and Y.T.C.: data curation. P.H.C. and A.H.C.: resources and supervision. C.Y.C. and S.W.C.: manuscript writing. All authors critically reviewed and approved the final version of the author disclosure form.

Conflicts of interest disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the International Journal of Surgery.

Research registration unique identifying number (UIN)

Registry used: Research Registry Research Registry unique Identifying Number: researchregistry

Guarantor

Shao-Wei Chen.

Data availability statement

The data that support the findings of this study are obtained from the Taiwan National Health Insurance Research Database (NHIRD). Restrictions apply to the availability of these data, which were used under license for this study.

Provenance and peer review

Not commissioned, externally peer-reviewed.

Supplementary Material

js9-110-3495-s001.docx ^{(33.2KB, docx)}

js9-110-3495-s002.docx ^{(70.4KB, docx)}