Association between remnant cholesterol and progression of bioprosthetic valve degeneration

Abstract

Aims

Remnant cholesterol (RC) seems associated with native aortic stenosis. Bioprosthetic valve degeneration may share similar lipid-mediated pathways with aortic stenosis. We aimed to investigate the association of RC with the progression of bioprosthetic aortic valve degeneration and ensuing clinical outcomes.

Methods and results

We enrolled 203 patients with a median of 7.0 years (interquartile range: 5.1–9.2) after surgical aortic valve replacement. RC concentration was dichotomized by the top RC tertile (23.7 mg/dL). At 3-year follow-up, 121 patients underwent follow-up visit for the assessment of annualized change in aortic valve calcium density (AVCd). RC levels showed a curvilinear relationship with an annualized progression rate of AVCd, with increased progression rates when RC >23.7 mg/dL (P = 0.008). There were 99 deaths and 46 aortic valve re-interventions in 133 patients during a median clinical follow-up of 8.8 (8.7–9.6) years. RC >23.7 mg/dL was independently associated with mortality or re-intervention (hazard ratio: 1.98; 95% confidence interval: 1.31–2.99; P = 0.001).

Conclusion

Elevated RC is independently associated with faster progression of bioprosthetic valve degeneration and increased risk of all-cause mortality or aortic valve re-intervention.

Graphical Abstract

Graphical Abstract.

(Upper panel) Diagram of the study. All patients completed 5-year follow-up following the study protocol, and the median follow-up time was 8.80 years [95% confidence interval (CI): 8.67–9.57 years]. (Lower panel) RC was calculated as TC minus LDL-c minus HDL-c. Cholesterol of Lp(a) is included in the LDL-c measurement. SAVR, surgical aortic valve replacement; IQR, interquartile range; MDCT, multi-detector computed tomography; RC, remnant cholesterol; CM, chylomicron; VLDL, very low-density lipoprotein; IDL, intermediate-density lipoprotein; LDL, low-density lipoprotein; Lp(a), lipoprotein(a); HDL, high-density lipoprotein; OR, odds ratio; HR, hazard ratio; HVD, haemodynamic valve deterioration.

Introduction

Bioprosthetic valves (BPVs) are now used in most surgical aortic valve replacement (SAVR) procedures.^1,2 The noticeable shift towards BPVs from mechanical valves could be explained by the aging of the target population, the freedom from lifelong anticoagulation, and the improved haemodynamic performance of BPVs. However, BPVs are prone to irreversible leaflet fibrosis and calcification, leading to BPV degeneration, which is a major clinical concern.³ Underlying mechanisms and risk factors for the process have yet to be fully understood, and there is no treatment to retard this progression nowadays.

Identifying risk factors of BPV degeneration could lead to novel therapeutic approaches. Histological findings had suggested that BPV calcification may share similar pathways with native calcific aortic valve stenosis.^4,5 In addition, observational studies found a potential association between total cholesterol and the risk of BPV calcification.^6–9 Due to the absence of evidence from randomized trials and the inconsistent findings from limited retrospective studies, pharmacological lowering of low-density lipoprotein cholesterol (LDL-c) to prevent BPV calcification remains controversial.^8,10–12 Meanwhile, a growing number of observational and Mendelian studies suggest an additive contribution of remnant cholesterol (RC), the cholesterol content of triglyceride (TG)-rich lipoproteins, to the pathogenetic processes of atherosclerotic cardiovascular disease and native aortic valve stenosis, regardless of LDL-c.^13–18 Therefore, RC might have a role in the pathogenetic mechanisms underlying BPV degeneration. Altogether, we hypothesize that RC is associated with BPV degeneration and could act as a readily available biomarker for routine surveillance and a promising target for pharmacological treatment.

The objective of this study was to evaluate the impact of RC on both the progression of BPV degeneration and clinical outcomes in patients who underwent SAVR.

Methods

Study population

From 2008 to 2010, 204 consecutive patients with isolated BPV SAVR were prospectively recruited at the Quebec Heart & Lung Institute. Two-hundred and three patients received fasting blood sample analyses, multi-detector computed tomography (MDCT), and complete Doppler echocardiography evaluation at the baseline visit, which is a median of 7.0 years [interquartile range (IQR): 5.1–9.2] after SAVR. A follow-up visit with multimodality imaging (MDCT and echocardiography) was completed in 121 eligible patients (60% of the entire cohort) at a median of 3.0 (IQR: 2.8–3.2) years following the baseline visit (Graphical abstract). Reasons for the exclusion and comparison of baseline characteristics between patients with and without follow-up imaging were presented in the Supplementary data online, Supplementary methods and Supplementary data online, Table S1. The study protocol has been described previously.^19,20 The inclusion and exclusion criteria are detailed in Supplementary data online, Supplementary methods. The protocol was approved by the institutional review board, and signed informed consent was obtained for all patients.

Determination of lipid profile

Fasting plasma samples of the 203 patients were collected at the baseline visit. Complete lipid profile including total cholesterol (TC), TG, high-density lipoprotein cholesterol (HDL-c), apolipoprotein A-I (ApoA-I), and apolipoprotein B (ApoB) was measured using automated techniques standardized with the Canadian Reference Laboratory. lipoprotein(a) (Lpa) was measured with chemiluminescent immunoassays. The definition of RC, TC minus LDL-c minus HDL-c, has been frequently used in recent correlative studies.^14,21,22 LDL-c was calculated by the Martin/Hopkins equation when TG ≤400 mg/dL.²³ For the only patient with TG >400 mg/dL, i.e. TG = 590.08 mg/dL, LDL-c was measured directly. The Martin/Hopkins equation provides an adjustable ratio [very low-density lipoprotein (VLDL)/TG] based on each subject’s lipid profile other than a fixed value used in the Friedewald equation (VLDL/TG = 5), as recommended by current American Heart Association/American College of Cardiology guidelines.^23–25

Calcification progression on MDCT

BPV leaflet calcification was quantified by MDCT. Specifically, the aortic valve calcium (AVC) load was estimated offline by the Agatston method on Aquarius iNtuition (TeraRecon, Inc., Foster City, CA) and then indexed to the cross-sectional area of the aortic annulus measured by echocardiography to calculate the AVC density (AVCd).²⁶ Then, the annualized change in AVCd (AU/cm²·year) was used to assess the progression of bioprosthetic calcification. (Aortic annulus kept constant for individuals at both baseline and follow-up.) Baseline and follow-up MDCT acquisition were done by experienced technicians, and all measurements were performed, within the VaRMI core laboratory, by the same experienced cardiologist (B.Z.), masked to the clinical and laboratory data. Particular attention was paid to distinguish artefacts of prosthesis stents, and only the calcification area of aortic valve leaflets was included. Twenty patients were randomly selected to ascertain intra- (intraclass correlation efficient = 0.972, P < 0.001) and inter- (M.-A.C. vs. B.Z.) (intraclass correlation efficient = 0.916, P < 0.001) observer variability.

Haemodynamic evaluation on echocardiography

Mean transprosthetic gradient (MG) was calculated using the modified Bernoulli Equation. The dynamic change in MG (mm Hg), reflecting haemodynamic progression of BPV, was defined as (MG at follow-up−MG at baseline). Prosthetic valve regurgitation severity was classified as mild, moderate, or severe by a multi-parametric approach described by Zoghbi et al.²⁷ BPV effective orifice area was derived using the standard continuity equation. Prosthesis–patient mismatch (PPM) was graded according to body mass index-adjusted cut-points.²⁸ All measurements were performed and reviewed by the same experienced cardiologist (H.M.). The assessor was masked to the clinical and laboratory data.

Clinical endpoint

The primary clinical endpoint was a composite of all-cause mortality or aortic valve re-intervention (redo-SAVR or valve-in-valve implantation). The secondary clinical endpoint was all-cause mortality. Mortality and re-intervention data were obtained from the Quebec Institute of Statistics as previously described.¹⁹ All patients completed 5-year follow-up pre-specified in the research protocol and then a clinical follow-up at an interval of 3 years continued up to death or valvular re-intervention. The reverse Kaplan–Meier median follow-up time was 8.8 years [95% confidence interval (CI): 8.7–9.6 years].

Statistical analysis

Data were expressed as median (IQR), mean (SD), or percentage, as appropriate. Normality was assessed by the Shapiro–Wilk test. Differences between groups were compared using the unpaired Student’s t-test or the Wilcoxon test for continuous variables and the χ² test or the Fisher exact test for categorical variables. Following the categorization of lipid levels in previous studies,^29,30 when analysed as categorical variables, RC was dichotomized according to the upper tertile (23.7 mg/dL; Tertile 3 vs. Tertiles 1 and 2).

Uni- and multivariable linear regressions were used to evaluate the impact of RC on AVCd progression. Clinically relevant variables clarified in previous publications and variables with a P-value < 0.10 in univariable regression were included as covariates, including age, sex, time from AVR to baseline visits, smoking, hypertension, renal insufficiency, LDL-c, lipid-lowering therapy, baseline MG, baseline AVCd, and PPM.³¹ Referring to previous studies,^14,32–34 body mass index (BMI) and diabetes were not considered, as they are known to be causally related to the RC level. Penalized splines were employed to capture the pattern of the association between RC and the annualized progression of AVCd. Piecewise linear models were then used to quantify the associations of RC with AVCd, with a cut-point value of upper tertile in RC.³⁵ Furthermore, traditional lipid biomarkers (LDL-c, TC, ApoB, ApoA-I, and Lpa) were separately added to the AVCd progression model to explore possible interaction with RC.

In survival analyses, the pattern of the association between RC and the clinical endpoint was initially examined with RC modelled as a spline term. Survival curves were built by using the Kaplan–Meier estimator and compared by using the log-rank test. Cox proportional hazards models were performed to estimate the hazard ratios (HRs) and 95% CIs for the association between RC and clinical outcomes, with RC modelled as both continuous and dichotomous variables. Adjustment for ApoB, ApoB/ApoA-I, and LDL-c was performed separately. The proportional hazards assumption was examined using Schoenfeld residuals. As an exploratory analysis, discordance between RC and LDL-c in association with clinical endpoint was also assessed: individuals were categorized into four groups based on the top tertile of RC (23.7 mg/dL) and LDL-c (98.5 mg/dL). Moreover, sensitivity analyses were performed in individuals on lipid-lowering treatment at baseline. Interactions between RC and lipid-related covariates (BMI, diabetes, coronary artery disease, hypertension, lipid-lowering treatment) were also examined.

A P-value < 0.05 was considered statistically significant. All analyses were performed by using R software, version 3.5.2 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Population characteristics

The study population was representative of patients who underwent SAVR in our institution.³⁶ Among the 203 patients, mean age was 67.4 ± 7.9 years, and 70% were men. The prevalence of cardiovascular comorbidities was high: 71% hypertension, 48% coronary heart disease, and 22% diabetes. Most patients (82%) received lipid-lowering medications (77% statins; 4% fibrates; 6% ezetimibe) at the time of the baseline visit. The median RC was 20.6 (17.5–25.3) mg/dL, and the median LDL-c was 82.6 (67.5–103.9) mg/dL. Baseline characteristics of the study population by RC levels are shown in Table 1. Supplementary data online, Table S2 further presents the characteristics of patients with 3-year multimodality imaging follow-up according to the levels of RC. No significant difference in AVCd level was found among various bioprosthesis type or bioprosthesis size (type: P = 0.967; size: P = 0.927).

Table 1.

Baseline characteristics of the study population by RC levels

	RC levels			P-value
	All patients (n = 203)	Tertiles 1 and 2a (n = 136; 67%)	Tertile 3b (n = 67; 33%)
Clinical parameters
Age, mean (SD), years	67.4 (7.9)	68.0 (7.63)	66.1 (8.4)	0.12
Male, n (%)	141 (69.5)	100 (73.5)	41 (61.2)	0.10
Body mass index, mean (SD), kg/m2	28.4 (5.3)	27.7 (5.3)	29.8 (5.1)	0.007
Hypertension, n (%)	144 (70.9)	93 (68.4)	51 (76.1)	0.33
Coronary artery disease, n (%)	97 (47.8)	61 (44.9)	36 (53.7)	0.30
Myocardial infarction, n (%)	25 (12.3)	18 (13.2)	7 (10.4)	0.73
Diabetes, n (%)	44 (21.7)	30 (22.1)	14 (20.9)	0.99
Smoking, n (%)	125 (61.6)	79 (58.1)	46 (68.7)	0.19
COPD, n (%)	23 (11.3)	17 (12.5)	6 (9.0)	0.61
Atrial fibrillation, n (%)	44 (21.7)	31 (22.8)	13 (19.4)	0.71
Previous stroke, n (%)	40 (19.7)	27 (19.9)	13 (19.4)	1.00
Renal insufficiency, n (%)	15 (7.4)	9 (6.6)	6 (9.0)	0.57
Evaluation timing
Time from AVR to baseline visits, median (IQR)	7.0 (5.1–9.2)	7.0 (4.8–9.2)	7.0 (5.7–10.4)	0.37
Time from baseline to follow-up visits, median (IQR)	3.0 (2.9–3.3)	3.1 (2.9–3.3)	3.0 (2.9–3.3)	0.31
Coronary artery bypass graft, n (%)	71 (35)	46 (33.8)	25 (37.3)	0.74
Medication
ACEI, n (%)	63 (31.0)	40 (29.4)	23 (35.3)	0.55
ARB, n (%)	56 (27.6)	36 (26.5)	20 (30.0)	0.76
β-Blockers, n (%)	96 (47.3)	58 (42.6)	38 (56.7)	0.09
Anti-lipid, n (%)	166 (81.8)	114 (83.8)	52 (77.6)	0.38
Statin, n (%)	157 (77.3)	112 (82.4)	45 (67.2)	0.29
Fibrates, n (%)	7 (3.6)	2 (1.5)	5 (7.6)	0.04
Ezetimibe, n (%)	12 (5.9)	3 (2.2)	9 (14.3)	0.003
Laboratory data
Total cholesterol, median (IQR), mg/dL	157.0 (137.2–188.3）	149.2 (131.2–174.2)	175.5 (156.2–202.4)	<0.001
Triglycerides, median (IQR), mg/dL	111.6 (86.8–150.6)	92.6 (78.4–111.9)	172.8 (150.6–202.9)	<0.001
HDL cholesterol, median (IQR), mg/dL	50.6 (43.3–59.0)	53.7 (47.2–62.3)	45.6 (38.1–52.2)	<0.001
LDL cholesterol, median (IQR), mg/dL	82.6 (67.5–103.9)	76 (64.9–92.4)	100.7 (82.5–129.7)	<0.001
Remnant cholesterol, median (IQR), mg/dL	20.6 (17.5–25.3)	18.4 (16–20.6)	28.2 (25.5–32.5)	<0.001
Lipoprotein(a), median (IQR), mg/dL	16.5 (5.1–57)	18.2 (5.5–59.5)	10.1 (3.8–49.9)	0.37
Apolipoprotein A-I, median (IQR), g/L	1.5 (1.4–1.7)	1.5 (1.4–1.7)	1.4 (1.3–1.7)	0.05
Apolipoprotein B, median (IQR), g/L	0.6 (0.5–0.7)	0.6 (0.5–0.7)	0.7 (0.7–0.9)	<0.001
Echocardiography
Peak aortic jet velocity, mean (SD), m/s	2.6 (0.6)	2.6 (0.6)	2.6 (0.6)	0.75
Mean gradient, mean (SD), mm Hg	14.6 (7.5)	14.4 (7.7)	14.9 (7.0)	0.66
Aortic valve area, mean (SD), cm2	1.3 (0.4)	1.3 (0.4)	1.3 (0.4)	0.50
Indexed aortic valve area, mean (SD), cm2/m2	0.7 (0.2)	0.7 (0.2)	0.7 (0.2)	0.37
Aortic regurgitation				0.78
None, n (%)	130 (65.7)	84 (63.6)	46 (69.7)
Mild, n (%)	58 (29.3)	41 (31.1)	17 (25.8)
Moderate, n (%)	10 (5.1)	7 (5.3)	3 (4.5)
Prosthesis–patient mismatch				0.59
None or mild, n (%)	76 (37.4)	48 (35.3)	28 (41.8)
Moderate, n (%)	79 (38.9)	56 (41.2)	23 (34.3)
Severe, n (%)	48 (23.6)	32 (23.5)	16 (23.9)
LVEF, mean (SD), %	63.6 (9.2)	63.5 (9.0)	63.8 (9.7)	0.85
MDCT
AVC, median (IQR), AU	30.3 (0.0–105.4)	32.6 (0.0–95.8)	27.9(5.9–157.5)	0.21
AVCd, median (IQR), AU/cm2	8.8 (0.0–33.5)	8.8 (0–27.3)	7.6 (1.6–55.7)	0.21
Surgical data, n (%)
Bioprosthesis type				0.78
Stentless	57 (28.8)	18 (26.9)	39 (29.8)
Stented porcine	62 (31.3)	20 (29.9)	42 (32.1)
Stented pericardial	79 (39.9)	29 (43.3)	50 (38.2)
Bioprosthesis size, mm				0.58
19	6 (3)	3 (4.5)	3 (2.3)
21	32 (16.2)	14 (20.9)	18 (13.7)
23	61 (30.8)	22 (32.8)	39 (29.8)
25	64 (32.3)	18 (26.9)	46 (35.1)
27	29 (14.6)	8 (11.9)	21 (16)
29	6 (3.0)	2 (3.0)	4 (3.1)

^aRC levels ≤23.7 mg/dL.

^bRC levels >23.7 mg/dL.

COPD, chronic obstructive pulmonary disease; ACEI, angiotensin-converting enzyme inhibitors; ARB, angiotensin receptor blockers; LVEF, left ventricular ejection fraction.

RC levels and progression of AVCd

In patients who undergone progression assessment (n = 121), the patients in the top tertile RC values had similar AVCd to those with lower RC values at baseline [6.4 (0.8–22.2) vs. 6.6 (0.0–21.6) AU/cm², P = 0.601]. During a 3-year follow-up, the median annualized progression rate of AVCd was higher in patients in the top tertile RC values, as compared with those in the lower tertiles [3.49 (1.11–10.04) vs. 1.89 (0.00–5.93) AU/cm²·year; P = 0.026] (Table 2). A curvilinear relationship between continuous RC levels and annualized progression rate of AVCd was identified (Figure 1A): the progression rate of AVCd increased faster as the RC levels were high. After adjusting for traditional lipid biomarkers, RC was still identified as an independent predictor of bioprosthetic calcification progression (see Supplementary data online, Table S3).

Table 2.

Association between RC and calcification progression

	Univariate analysis			Multivariable analysisa
	Est.β ± SE	Std.β	P-value	Est.β ± SE	Std.β	P-value
Progression in AVCd: △AVCd (AU/cm2·yr)
RC top tertileb	8.68 ± 3.66	0.22	0.018	9.01 ± 4.17	0.22	0.033
RC level, mg/dLc	0.82 ± 0.23	0.31	<0.001	0.88 ± 0.25	0.33	<0.001
Piecewise linear model
RC ≤23.7 mg/dL	−0.29 ± 0.51	−0.11	0.575	−0.60 ± 0.60	−0.23	0.320
RC＞23.7 mg/dL	1.84 ± 0.76	0.47	0.017	2.29 ± 0.85	0.58	0.008

^aAdjusted for age, sex, time from AVR to baseline visits, smoking, hypertension, renal insufficiency, low-density lipoprotein-cholesterol, lipid-lowering therapy, baseline mean gradient, baseline aortic valve calcification density, and prosthesis–patient mismatch.

^bRC as a categorical variable.

^cRC as a continuous variable.

Est.β, coefficient estimate; Std.β, standardized coefficient.

Figure 1.

Association of continuous RC with disease progression measured by aortic valve calcification density (A) and mortality/reintervention (B). (A) The progression rate of AVCd was relatively flat until 23.7 mg/dL of RC (dotted lines) and then began to increase rapidly. (B) Restricted cubic splines (with three knots) demonstrated the shape and strength of the association between RC and mortality/re-intervention. The curves are presented with 95% CIs. The shaded area below indicates the density of the population.

RC levels and clinical outcomes

During a median follow-up of 8.8 (8.7–9.6) years, 133 patients experienced the primary clinical outcome (43 cardiovascular deaths, 56 non-cardiovascular deaths, and 46 re-interventions). RC concentration was associated with the risk of clinical events (Figure 1B). The Kaplan–Meier survival curve revealed that patients in the top tertile RC levels had a significantly higher rate of clinical events (Figure 2). In multivariable Cox regression, each 10 mg/dL increase in RC levels was associated with 38% and 36% increase of primary clinical outcome [hazard ratio (HR):1.38, 95% confidence interval (CI): 1.10–1.74, P = 0.005] and secondary clinical outcome (HR:1.36, 95% CI: 1.02–1.81, P = 0.037), respectively (Table 3; see Supplementary data online, Table S4). RC levels were also identified as one of the most contributive factors in the model (see Supplementary data online, Figure S1). The association between RC levels and clinical outcome remained consistent after further adjustment for LDL-c, ApoB, and ApoB/ApoA-I (all P < 0.001) (see Supplementary data online, Table S5). The method used to calculate AVCd did not modify the results (see Supplementary data online, Figure S2).

Figure 2.

Impact of RC on mortality and re-intervention. Adjusted for age, sex, time from AVR to baseline visits, hypertension, smoking history, chronic obstructive pulmonary disease, renal insufficiency, lipid-lowering treatment, baseline mean gradient, baseline aortic valve calcification density, and LDL-c.

Table 3.

Univariate and multivariable analyses of risk factors for mortality or re-intervention

	Univariate		Multivariable
	Crude HR (95% CI)	P-value	Adjusted HR (95% CI)	P-value	Adjusted HR (95% CI)	P-value
Age, years	1.03 (1.00–1.05)	0.030	1.04 (1.00–1.06)	0.014	1.03 (1.00–1.06)	0.017
Sex	0.98 (0.67–1.42)	0.903	1.36 (0.78–1.85)	0.176	1.35 (0.86–2.11)	0.187
Time from AVR to baseline visits, years	1.05 (1.00–1.10)	0.065	1.07 (1.02–1.13)	0.013	1.07 (1.02–1.13)	0.008
Hypertension	0.86 (0.59–1.26)	0.445	0.81 (0.55–1.27)	0.325	0.88 (0.58–1.33)	0.540
Smoking history	1.03 (0.72–1.45)	0.888	0.93 (1.02–2.23)	0.731	0.95 (0.63–1.43)	0.809
COPD	1.97 (1.24–3.15)	0.004	2.12 (1.23–3.35)	0.005	2.02 (1.20–3.39)	0.008
Renal insufficiency	1.82 (1.02–3.24)	0.043	1.58 (0.78–2.82)	0.152	1.53 (0.81–2.88)	0.189
Lipid-lowering treatment	0.87 (0.57–1.34)	0.529	0.81 (0.42–1.28)	0.432	0.83 (0.48–1.43)	0.494
MG, 5 mm Hg increase	1.26 (1.12–1.41)	<0.001	1.22 (1.05–1.37)	0.003	1.20 (1.05–1.37)	0.006
AVCd, 10 AU/cm2 increase	1.06 (1.04–1.08)	<0.001	1.06 (1.03–1.08)	<0.001	1.06 (1.03–1.08)	<0.001
LDL-C, 10 mg/dL increase	1.00 (0.95–1.06)	0.960	0.93 (0.89–1.03)	0.070	0.95 (0.88–1.02)	0.153
RC, top tertilea	1.45 (1.02–2.06)	0.041	1.98 (1.31–2.99)	0.001	—	—
RC, 10 mg/dL increase	1.29 (1.04–1.60)	0.019	—	—	1.38 (1.10–1.74)	0.005

^aRC as a categorical variable.

AVR, aortic valve replacement; COPD, chronic obstructive pulmonary disease; MG, mean gradient.

The exploratory discordance analyses revealed that patients with high RC/low LDL-c (n = 30, 14.78%) presented an increased risk of mortality or valve re-intervention (HR: 1.85; 95% CI: 1.12–3.06; P = 0.017), while those with low RC/high LDL-c (n = 30, 14.78%) did not (HR: 0.79; 95% CI: 0.45–1.37; P = 0.399) (see Supplementary data online, Figure S3). Results were consistent in the assessment of discordance between RC and ApoB or ApoB/ApoA-I with regards to the increased risk of clinical endpoint (see Supplementary data online, Figures S4 and S5). No interactions with lipid-related covariates and lipid-lowering treatment were found with regards to the associations between RC and clinical outcomes (see Supplementary data online, Tables S6–S8 and Figures S6 and S7).

Discussion

The present study expands the knowledge on the risk of elevated RC levels in BPV degeneration. We demonstrate for the first time that increased circulating levels of RC are associated with faster structural valve deterioration and worse prognosis in patients with aortic bioprostheses independent of LDL-c levels and relevant risk factors. Moreover, compared with LDL-c, ApoB, and ApoB/ApoA-I, RC presents a greater contribution to the disease progression and risk of death and re-intervention.

Potential mechanism of RC on degenerative progression

Increasing evidence has shown that RC is highly atherogenic particles.^18,37 RC and LDL-c were associated with a similar risk of cardiovascular disease per unit difference in ApoB among 654 783 participants. Recent Mendelian randomization studies further revealed that lifelong exposure to high TG levels and RC levels increases the risk of native aortic stenosis.^13,38 Similar to native aortic stenosis, a lipid-mediated inflammatory process is known to contribute to BPV degeneration, and ‘valvular-metabolic risk’ is thus established.^4,39,40 In our study, we reported for the first time that RC is associated with faster BPV degeneration, higher mortality, and re-intervention rate. Based on the atherogenic-like nature of early BPV degeneration and the established causal relationship of RC with atherosclerotic cardiovascular diseases, RC could be postulated as one of the main culprits of BPV degeneration and thus have the potential to be an additional biomarker aiding traditional apoB particles.¹⁴

Our previous work demonstrated an association between elevated ApoB, ApoB/ApoA-I ratio, and the risk of BPV degeneration (defined as an increase in MG ≥10 mm Hg and/or worsening of regurgitation ≥1/3 class).²⁰ In the present study, following a more standardized and comprehensive approach evaluating BPV degeneration, we found that RC outperformed ApoB or ApoB/ApoA-I ratio as a predictor of calcification progression and long-term clinical outcomes. Several hypotheses may elucidate the mechanism behind the new finding. Firstly, compared with total ApoB particle concentration, the content of cholesterol across lipoprotein particles may contribute more to the disease progression and long-term prognosis. As RC could contain up to four-fold greater cholesterol content per particle than LDL,¹⁵ the higher cholesterol/ApoB molar ratio might clarify, at least partly, the stronger association of RC with BPV denegation than ApoB and ApoB/ApoA-I. Secondly, RC could stay longer in circulation and have the capacity to penetrate the inflow and outflow surface of valve leaflet.^4,39 Then it could be directly taken by macrophages without any modification,^41,42 resulting in the enhanced inflammatory process, foam cells formation, extracellular matrix disruption, and bioprostheses calcification.^4,39,43 Thirdly, independent of cholesterol content itself, RC may reflect low-grade inflammation and other atherogenic mechanisms relating to increased apolipoprotein C3 (ApoC3) or angiopoietin-like protein 3 (ANGPTL3). Hence, while ApoB/ApoA-I ratio reflects the balance between proatherogenic vs. antiatherogenic lipoprotein particles, RC may better echo other atherogenic and calcification mechanisms behind the counteraction between LDL-c and HDL-c, particularly in patients with normal or low levels of LDL-c.

RC and structural valve deterioration

Of note, in aortic bioprostheses, levels of calcification were substantially lower compared with native aortic valves when reaching the comparable severity of haemodynamic dysfunction. Given that an AVCd severity cut-point of 58 AU/cm² is already associated with catastrophic clinical outcomes,¹⁹ accelerated progression of calcification triggered by high RC levels should cause alarm and attention. In fact, a recent study showed that even a ‘microscopic’ calcification assessed by ¹⁸F-fluoride positron emission tomography–computed tomography imaging is strongly associated with future bioprosthetic dysfunction.⁴⁴ Moreover, the higher risk of death or valve re-intervention in patients with elevated RC levels also lend credence to the clinical significance of a calcification progression owing to RC in the present study. The process of BPV degeneration accelerated when RC levels were higher than 23.7 mg/dL, as evidenced by a calcification progression of 3.5 AU/cm² per year. Interestingly, the cut-off value of RC in our study is in line with previous relevant studies. There are associations of elevated RC levels (i.e. higher than 23.2, 23.4, 23.6, or 25.0 mg/dL) with atheroma progression in patients with coronary artery disease^21,45,46 and the progression of diabetic nephropathy in Type 1 diabetes.⁴⁷ Still, further studies with larger sample sizes are needed to validate the cut-off value of RC in patients with BPVs.

Additionally, our findings may also explain the inconsistent results in the retrospective studies concerning the effectiveness of LDL-c-lowering therapy in BPV degeneration: RC might be overlooked. In the present study, approximately 80% of patients received statin therapy, and 71% of patients had LDL-c levels <100 mg/dL. Interaction analysis between RC levels and statin treatment and subgroup analysis with patients receiving statin treatment did not change the overall association, indicating that RC could be a major cholesterol fraction contributor to the development of BPV degeneration, regardless of LDL-c level. More importantly, associations of elevated RC with BPV degeneration could potentially suggest new indications for RC-lowering therapy. It is of clinical importance to further explore whether RC-lowering therapies could retard the progression of BPV degeneration in patients with atherosclerotic disease. In fact, post hoc analysis from the Treating to New Targets trial supported a dose-response relationship between intensive statin therapy and the reduction of RC levels and cardiovascular risks, while the effect of the approach on RC levels is modest.⁴⁸ Recent studies demonstrated that there is discordance in the response of LDL-c and RC to statin treatment, as RC is lowered to a less extent than LDL-c (80% relative to LDL-c).⁴⁹ When statin doses are optimized or intolerance to statins exists, ezetimibe and proprotein convertase subtilisin/kexin type 9 inhibitors can be used, but their effects on RC levels are also modest. In contrast, fibrates,⁵⁰ high-dose n-3 fatty acids (icosapent ethyl),⁵¹ and new generation RC lowering therapies (ApoC3⁵² and ANGPTL3⁵³ inhibitors) have a more profound RC-lowering effect.

Clinical implication

For patients with native aortic stenosis, several secondary analyses of randomized controlled trials highlight that the target population for lipid-lowering therapy might be patients with mild aortic stenosis (an initiation phase mediated by valvular lipid deposition).^54,55 However, identifying such patients without any related symptoms turns out to be laborious and difficult in clinical practice. In the context of aortic bioprostheses, owing to the complicated cardiovascular comorbidities and high risk of the target population,¹ stricter and more organized follow-up is usually scheduled, leading to better routine surveillance of the degeneration process. Of vital importance, as opposed to native aortic stenosis, one of the inherent advantages in this particular context is that the aggressive change in lifestyle and prompt pharmacological treatment could be instituted at the time of AVR before the initiation of the lipid-mediated inflammatory process leading to BPV degeneration. With increasing attention towards BPV degeneration, our findings suggest that RC could act as a practical biomarker for clinical surveillance and an appealing treatment target that may delay the progression of BPV degeneration.

Study limitations

Firstly, the relatively low number of patients is one of the main limits of the current study, even though a minimal sample size to ensure the robustness of the model was achieved. Secondly, we used baseline RC levels and could not assess how temporal changes in this biomarker may affect the association with mortality or valve re-intervention. RC was calculated through the standard lipid profile method rather than direct measurement, and head-to-head comparison was not performed. However, indirect measurement of RC is more practical, affordable, and widely accepted in routine practice. Thirdly, blood samples were taken when the patients were in a fasting state while current guidelines endorse non-fasting lipid profiles for routine tests, since it may better reflect atherogenic lipoprotein levels. However, the two approaches are rather complementary than contradictory.⁵⁶ Moreover, the cut-off of RC (23.7 mg/dL) needs to be interpreted with caution as it is mainly based on the distribution of RC in our cohort. As a post hoc analysis from a prospective cohort, we could not rule out the possibility of unidentified confounders; thus, the finding is hypothesis-generating, calling for future verification from large well-designed studies.

Conclusions

In patients with aortic bioprostheses, elevated circulating levels of RC were associated with faster structural valve deterioration and increased risk of all-cause mortality and valve re-intervention, independent of LDL-c and other relevant risk factors. The underlying causal relationship and mechanisms need to be elucidated in further research. Additional studies should elucidate the time-dependent manner of the effect of RC on bioprosthesis degeneration and compare outcomes of therapy guided by RC vs. conventional indicators.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Funding

This study was funded by a research grant (MOP #86666) from the Canadian Institutes of Health Research (CIHR), Ottawa, Ontario, Canada. B.Z. holds a State Scholarship Fund from China Scholarship Council (No.201806210439) and a fund from Capital’s Funds for Health Improvement and Research (CFH 2022-4-4037). P.M. holds a Fonds de Recherche du Québec-Santé (FRQS). P.P. holds the Canada Research Chair in Valvular Heart Diseases from the Canadian Institutes of Health Research. M.-A.C. holds an Early Career Investigator Award in Circulatory and Respiratory Health and the Canada Research Chair on Women’s Cardiac Valvular Health from the Canadian Institutes of Health Research.

Data availability

The data that support the findings of this study are available from the corresponding author (M.-A.C.), upon reasonable request.