Lipoprotein Particle Profiles Compared With Standard Lipids in the Association With Subclinical Aortic Valve Calcification in Apparently Healthy Japanese Men

Thien Vu; Akira Fujiyoshi; Takashi Hisamatsu; Aya Kadota; Maryam Zaid; Hiroyoshi Segawa; Keiko Kondo; Sayuki Torii; Yoshihisa Nakagawa; Tomoaki Suzuki; Tohru Asai; Katsuyuki Miura; Hirotsugu Ueshima; for the Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA) Research Group

doi:10.1253/circj.CJ-20-1090

Abstract

Background: Risk factors for atherosclerotic disease including dyslipidemia have been shown to be associated with aortic valve calcification (AVC). Nuclear magnetic resonance (NMR)-measured lipoprotein particles, low-density and high-density lipoprotein particles (LDL-p, HDL-p) in particular, have emerged as novel markers of atherosclerotic disease; however, whether NMR-measured particles are associated with AVC remains to be determined. This study aimed to examine the association between NMR-based lipoprotein particle measurements and standard lipids with AVC. The primary variables of interest were LDL-p (nmol/L), HDL-p (μmol/L), LDL-cholesterol, and HDL-cholesterol (both in mg/dL).

Methods and Results: A community-based random sample of Japanese men aged 40–79 years examined in 2006–2008, in Shiga, Japan was studied. Presence of AVC was defined as an Agatston score >0. Lipoprotein particles were measured using NMR spectroscopy. In the main analysis, multivariable-adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) for the prevalence of AVC across the higher quartiles of lipids in reference to the lowest ones were obtained. Of 874 participants analyzed, 153 men had AVC. Multivariable-adjusted ORs of prevalent AVC for the highest vs. the lowest quartile were significantly elevated for LDL-p (OR, 2.20; 95% CI: 1.23–3.93) and LDL-cholesterol (OR, 2.16; 95% CI: 1.23–3.78). In contrast, neither HDL-p nor HDL-cholesterol was associated with AVC.

Conclusions: The association of prevalent AVC with NMR-based LDL-p was comparable to that with LDL-cholesterol.

Degenerative calcific aortic valve disease is one of the most common heart diseases worldwide, and its prevalence has increased in developed countries as populations have aged.¹^–³ If mild valve obstruction exists, the disease seems to progress inevitably to hemodynamically more altered states. Once the disease becomes symptomatic and severe, the prognosis without surgery is dismal.⁴ The subclinical early stage of degenerative calcific aortic valve disease is characterized by aortic valve calcification (AVC). The pathophysiology of AVC involves an active process, similar to that of atherosclerosis, characterized by 3 primary processes: lipid accumulation, inflammation, and calcification.⁵^–⁸

Nuclear magnetic resonance (NMR)-based lipoprotein assessment that simultaneously measures the concentrations and sizes of various lipoprotein particles , presents novel markers of atherosclerosis and/or cardiovascular disease.⁹^–¹² Although some of these measures are correlated with corresponding conventional lipid measures, some studies suggest that NMR-based measures have a role beyond the conventional lipids. For example, an observational study showed that low-density lipoprotein particle (LDL-p)-lowering therapy guided by change in LDL-p was more cost-effective than a low-density lipoprotein cholesterol (LDL-c)-guided approach in preventing cardiovascular disease.¹³ With regard to atherosclerosis, we have previously reported that LDL-p and high-density lipoprotein particle (HDL-p) concentrations were more robust independent markers of subclinical atherosclerosis compared to the corresponding conventional cholesterols in a community-based sample.¹⁰^,¹¹

However, whether NMR-based lipoprotein particles are associated with AVC remains to be determined. In this cross-sectional study, we primarily examined an association of AVC with LDL-c, HDL-c and the corresponding NMR-based lipoproteins; that is, LDL-p and HDL-p. We selected those 4 lipid measures as the primary objectives of our investigation because of their clinical importance and the wealth of literature.

Methods

Study Participants

The Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA) is an ongoing prospective population-based study conducted in Japan. The study design and recruitment details have been reported previously.¹⁴^,¹⁵ In brief, we randomly selected and invited 2,379 Japanese men aged 40–79 years, residents of Kusatsu city, Shiga, Japan, from 2006 to 2008, based on the Basic Residents’ Register of the city. A total of 1,094 men agreed to participate (participation rate, 46%), with their written informed consent obtained. For the present study, we excluded those participants who had any reported history of aortic valve surgery (n=4), or myocardial infarction or stroke (n=69). Because we used Friedewald’s formula for estimating the serum concentration of LDL-c, we also excluded those who had serum triglycerides (TG) concentration ≥400 mg/dL (n=17) as the formula is not applicable for such individuals.¹⁶ We further excluded those with an estimated glomerular filtration rate (eGFR) of <30 mL/min/1.73 m² using serum creatinine by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation modified for the Japanese (n=3),¹⁷^,¹⁸ those who used lipid-lowering medication (n=118), and those with missing pertinent information, which left 874 participants for analyses. The study conforms to the Declaration of Helsinki and was approved by the Institutional Review Board of Shiga University of Medical Science, Otsu, Japan.

Measurements

Data on medical history and lifestyle factors were collected from each participant using a self-administered questionnaire, and trained technicians confirmed the completed questionnaire with participants. Height, weight, blood pressure, and a variety of other measures were collected. Blood pressure was measured by using an automated sphygmomanometer (BP-8800; Omron Colin, Tokyo, Japan).

Blood samples were obtained early in the clinic visit after 12-h fasting.¹⁴^,¹⁵ Serum was separated by centrifugation (3000 revolutions per min, for 15 min) at 4℃ within 90 min. A portion of the samples was sent for routine laboratory tests, including those for standard lipids and glucose. Concentration of blood glucose was determined from sodium fluoride-treated plasma using a hexokinase glucose-6-phosphate-dehydrogenase enzymatic assay. Concentrations of serum total cholesterol (TC) and triglycerides were determined using enzymatic assays, and that of HDL-c was measured using a direct method (Determiner- C-TC, Determiner-C-TGL, Determiner-L HDL-c, respectively; Kyowa Medix, Tokyo, Japan). Measurements were standardized according to guidelines from the Center for Disease Control and Prevention/Cholesterol Reference Method Laboratory Network (CDC/CRMLN). We used Friedewald’s formula to estimate LDL-c levels¹⁶ with TG <400 mg/dL. Non-HDL-c was calculated by subtracting HDL-c from TC.

The remaining serum samples were stored at −80℃. Then, a portion of them was shipped on dry ice to LipoScienceInc. (Raleigh, NC; now LabCorp, Burlington, NC, USA) for NMR-based lipoprotein particles measurement. NMR spectroscopy was performed to quantify the particle concentrations of LDL and HDL.¹⁹^,²⁰ Additionally, particle concentrations were further determined for 3 LDL subclasses (intermediate-density lipoprotein [IDL], 23–27 nm; large, 21.3–23 nm; small, 18.3–21.2 nm), and 3 HDL subclasses (large, 8.8–13 nm; medium, 8.2–8.8 nm; and small, 7.3–8.2 nm). Weighted average particle sizes of LDL and HDL were also calculated.

Glycated hemoglobin (HbA1c) was measured using a latex agglutination assay according to the standardized method of the Japanese Diabetes Society (JDS). We then converted JDS values to those of the National Glycohemoglobin Standardization Program (NGSP) using the following formula recommended by the JDS: HbA1c (NGSP) = 1.02 × HbA1c(JDS) + 0.25 (%).²¹

C-reactive protein (CRP) was measured by nephelometry using a BN II analyzer with an inter-assay coefficient of variation ranging from 4.5% to 4.6%.² eGFR was calculated according to the CKD-EPI equation modified for the Japanese as follows:¹⁷^,¹⁸

eGFR (mL/min/1.73 m²) = 141 × (serum creatinine/0.9)^(−0.411) × 0.993^(Age) × 0.813 for those with a serum creatinine ≤0.9 mg/dL (79.6 µmol/L).

eGFR (mL/min/1.73 m²) = 141 × (serum creatinine/0.9)^(−1.209) × 0.993^(Age) × 0.813 for those with a serum creatinine >0.9 mg/dL (79.6 µmol/L).

Aortic Valve Calcification Assessment

The detailed method for cardiac computed tomography (CT) in SESSA was reported previously.²^,²² In brief, the aortic valve was assessed by either a GE-Imatron C150 Electron Beam Tomography system (EBCT; GE Medical Systems, South San Francisco, CA, USA) for participants examined from May 2006 through to August 2007, or a 16-row multi-detector row CT system (MDCT, Aquilion-16^TM; Toshiba Medical Systems, Tochigi, Japan) for participants examined thereafter.²³ A DICOM workstation and AccuImage software (AccuImage Diagnosis, San Francisco, CA, USA) were used to quantify calcium scores. The calcium score was calculated by multiplying the lesion area by a density factor derived from the maximal Hounsfield unit within this area, as described by Angaston et al.²³ AVC was identified according to the methods from the Multi-Ethnic Study of Atherosclerosis (MESA); this included any calcified lesion residing within the aortic valve leaflets.²⁴^–²⁶ Calcification of the aortic annulus, aortic sinuses, ascending aorta, or coronary arteries was excluded. All CT images were assessed by 1 trained medical technologist masked to the clinical information of the participants.

Statistical Analysis

Participants’ characteristics are shown using means and standard deviations (SDs) for continuous variables with a bell-shaped distribution, medians and interquartile ranges for those with a skewed distribution, and percentages for categorical variables. Difference in characteristics by the presence or absence of AVC was evaluated using the unpaired Student’s t-test, Mann-Whitney U-test, or Chi-squared test, as appropriate. We calculated age-adjusted Spearman’s rank correlation coefficients between the measured lipid levels.

Body mass index (BMI) was calculated as weight (kg) divided by height squared (m²). Hypertension was defined as systolic blood pressure (SBP) ≥140 mmHg, diastolic blood pressure (DBP) ≥90 mmHg, or use of antihypertensive medications. Diabetes mellitus was defined as fasting plasma glucose ≥126 mg/dL, HbA1c (NGSP) ≥6.5%, or the use of diabetic medication. Hypertriglyceridemia was defined as triglycerides ≥150 mg/dL. Prevalent AVC was defined as a calcium score >0.²⁴^,²⁵

In the main analyses, we divided each lipid measure into quartiles, and computed multivariable adjusted odds ratios (ORs) and 95% confidence intervals (CIs) of the prevalent AVC, using logistic regression, for the upper 3 quartiles (Q2–Q4) in reference to the lowest quartile (Q1). Quartiles of each lipid measure were used because there are no established cut-off points for those measures in relation to AVC. In the logistic regression model, we adjusted for the following risk factors: age, smoking (pack-years), drinking status (current, past, never), BMI, HbA1c, SBP, medication status (hypertension and diabetes), in addition to eGFR, CRP, and CT-type (EBCT/16-MDCT). A P value for trend across the quartiles was obtained by inserting a variable that takes median values for each quartile into a model.

In sensitivity analyses, we treated lipid measures as continuous, and calculated adjusted ORs and 95% CIs of prevalent AVC per 1-SD higher or lower lipids.

As post-hoc analyses, we conducted the following additional analyses: (1) we repeated analyses after stratification by presence/absence of metabolic syndrome (MetS), hypertension, diabetes mellitus and hypertriglyceridemia. The rationale for those stratified analyses was to explore potential interaction by those factors on the association between AVC and lipid measures.²⁷ For the definition of MetS, we used both revised National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) criteria²⁸ and the Japanese criteria²⁹ (See the Supplementary File for detailed definitions). In conducting the stratified analyses, we excluded covariates that defined stratifying factor from regression models. Take stratification by MetS for example, we excluded BMI, systolic blood pressure, HbA1c, medications for blood pressure and diabetes from our multivariable adjusted model. The P value for interaction was calculated by inserting the interaction term and based on the Wald Chi-squared test (In quartile model testing interaction, we treated the quartiles as rank); (2) for completeness, we extended the exposure of interest to non-HDL-cholesterol and triglycerides, subclasses of LDL-p (large, small, and IDL), HDL-p (large, medium, and small), and size (nm) of LDL and HDL themselves; (3) we conducted sensitivity analysis repeating the main analysis including those with TG ≥400 mg/dL for the purpose of a comparison of HDL-c and HDL-p. We did not conduct the same sensitivity analysis for the comparison between LDL-c and LDL-p because the validity of Friedewald-based LDL-c concentration is uncertain in those with TG ≥400 mg/dL. Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA). A 2-tailed P value <0.05 was considered significant.

Results

Of the 874 men (aged 40–79 years) we analyzed, the mean (SD) age of our participants was 63.3 (10.0) years. One-hundred and fifty-three (17.5%) men had prevalent AVC (Table 1), and they tended to be older, more likely to have hypertension and diabetes and to take medication(s) for hypertension and/or diabetes. With regard to the standard lipids, those with prevalent AVC tended to have a higher concentration of LDL-c and a lower concentration of triglycerides. In NMR-measured lipid levels, the corresponding concentration of LDL-p tended to be higher and that of HDL-p tended to be lower than those without AVC (Table 1). By using an age-adjusted Spearman’s correlation matrix, it was found that the LDL-p concentration was strongly correlated with LDL-c and non-HDL-c (r=0.81 and 0.82, respectively) (Supplementary Table 1). In a multivariable-adjusted logistic regression model, older age and higher SBP, in addition to greater LDL-c concentration, were independently associated with prevalent AVC (Supplementary Table 2).

Table 1. Characteristics of Study Participants With or Without AVC (Shiga Epidemiological Study of Subclinical Atherosclerosis [SESSA] 2006–2008)

Characteristic	Overall (N=874)	No AVC (n=721)	Prevalent AVC (n=153)	P value
Age, years	63.3 (10.0)	61.8 (10.0)	70.3 (6.5)	<0.001
BMI, kg/m²	23.4 (2.9)	23.4 (2.9)	23.3 (3.1)	0.904
Smoking status, %				0.984
Current	33.3	33.4	32.7
Past	49.9	49.8	50.3
Never	16.8	16.8	17
Smoking,^† pack-years	23.8 (5.0–43.0)	23.1 (5.0–42.0)	27.2 (8.4–47.0)	0.095
Drinking status, %				0.065
Current	77.4	78.6	71.2
Past	5.3	4.6	8.5
Never	17.3	16.8	20.3
Systolic blood pressure, mmHg	135.8 (19.2)	134.3 (18.8)	142.5 (19.7)	<0.001
Diastolic blood pressure, mmHg	79.5 (11.0)	79.6 (11.1)	79.5 (10.5)	0.917
Anti-hypertensive medication, %	25.1	22.5	37.3	<0.001
Hypertension, %	50.9	47.4	67.3	<0.001
HbA1c, %	5.4 (5.2–5.8)	5.4 (5.1–5.7)	5.6 (5.3–6.0)	<0.001
Medication for diabetes, %	7.9	6.5	14.4	0.001
Diabetes, %	12.7	11.1	20.3	<0.001
eGFR, mL/min/1.73 m²	74.9 (10.3)	75.9 (9.9)	70.0 (10.6)	<0.001
CRP, mg/dL	0.4 (0.2–0.9)	0.4 (0.2–0.8)	0.5 (0.3–1.0)	0.023
TC, mg/dL	209.1 (34.0)	208.7 (33.2)	211.0 (37.6)	0.490
LDL-c, mg/dL	126.4 (31.6)	125.3 (31.0)	131.2 (34.2)	0.036
HDL-c, mg/dL	59.3 (17.2)	59.8 (17.4)	57.0 (16.2)	0.065
TG, mg/dL	100.0 (74.0–145.0)	100.0 (73.0–145.0)	98.0 (77.0–143.0)	0.002
Non HDL-c, mg/dL	149.8 (34.9)	148.9 (34.4)	154.0 (37.0)	0.101
LDL-p (nmol/L)	1,292.5 (386.8)	1,280.7 (379.1)	1,348.3 (417.9)	0.049
Large LDL-p (nmol/L)	640.4 (286.7)	626.4 (280.8)	706.39 (305.3)	0.002
Small LDL-p (nmol/L)	535.1 (412.6)	534.8 (411.3)	536.6 (420.1)	0.960
IDL (nmol/L)	93.5 (42.0–165.0)	97.0 (42.0–169.0)	80.0 (40.0–148.0)	0.122
HDL-p (μmol/L)	34.1 (6.6)	34.6 (6.6)	31.6 (6.0)	<0.001
Large HDL-p (μmol/L)	6.6 (4.3–9.7)	6.6 (4.3–9.7)	6.7 (4.2–9.8)	0.861
Medium HDL-p (μmol/L)	7.3 (4.9–10.3)	7.5 (5.1–10.7)	6.4 (4.0–9.1)	<0.001
Small HDL-p (μmol/L)	18.9 (5.2)	19.1 (5.1)	17.7 (5.4)	0.002
LDL-p size (nm)	20.9 (0.6)	20.9 (0.6)	21.0 (0.6)	0.102
HDL-p size (nm)	9.3 (0.5)	9.3 (0.5)	9.4 (0.6)	0.204

The presence of AVC was defined as an Agatston score >0. Values are expressed as mean (standard deviation) for continuous variables with approximately normally distribution or by median (interquartile range) with skewed distribution and % for categorical variables. Differences in characteristics were evaluated by using the unpaired Student’s t-test, Wilcoxon rank sums test, or Chi squared test. Hypertension was defined as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or use of antihypertensive medications. Diabetes was defined as fasting plasma glucose ≥126 mg/dL, HbA1c (National Glycohemoglobin Standardization Program) ≥6.5%, or the use of diabetic medication. ^†The number of participants with and without aortic valve calcification were 149 and 717, respectively, owing to missing information. AVC, aortic valve calcification; BMI, body mass index; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; HbA1c, glycated hemoglobin; HDL-c, high-density lipoprotein cholesterol; HDL-p, high-density lipoprotein particle; IDL, intermediate-density lipoprotein; LDL-c, low-density lipoprotein cholesterol; LDL-p, low-density lipoprotein particle; TC, total cholesterol; TG, triglycerides.

In main analyses, the highest quartiles (Q4) of LDL-c and LDL-p were associated with prevalent AVC. ORs (95% CIs) of prevalent AVC in Q4, in reference to Q1, were 2.16 (1.23–3.78) and 2.20 (1.23–3.93) respectively, after adjusting for age, BMI, smoking, drinking, SBP, HbA1c, medications for blood pressure and diabetes, eGFR, CRP, and CT-type (Table 2). Neither HDL-c nor HDL-p was associated with prevalent AVC.

Table 2. Multivariable-Adjusted OR for Prevalent AVC Across the Quartile of Primary Lipid Indices in Apparently Healthy Japanese Men (Aged 40–79 Years: 2006–2008, SESSA)

	1^st quartile (ref)	2^nd quartile	3^rd quartile	4^th quartile	P value for trend
	Range [Adjusted OR (95% CI)]
LDL-c, mg/dL	34.6–105.6	106.0–123.6	123.8–146.8	147.4–275.8	0.006
LDL-c, mg/dL	Reference	1.11 (0.62–1.95)	1.12 (0.64–1.96)	2.16 (1.23–3.78)	0.006
HDL-c, mg/dL	25.0–46.0	47.0–56.0	57.0–68.0	69.0–150.0	0.323
HDL-c, mg/dL	Reference	0.91 (0.53–1.55)	0.98 (0.56–1.70)	0.73 (0.39–1.33)	0.323
LDL-p, nmol/L	297.0–1,009.0	1,010.0–1,273.0	1,275.0–1,543.0	1,544.0–3,156.0	0.011
LDL-p, nmol/L	Reference	1.44 (0.81–2.56)	1.18 (0.65–2.11)	2.20 (1.23–3.93)	0.011
HDL-p, μmol/L	13.9–29.7	29.8–33.3	33.4–37.5	37.6–68.9	0.297
HDL-p, μmol/L	Reference	0.81 (0.50–1.34)	0.70 (0.40–1.24)	0.79 (0.43–1.43)	0.297

ORs were adjusted for age, BMI, smoking (pack-years), drinking, systolic blood pressure, HbA1c, medications for blood pressure and diabetes, eGFR, CRP, and CT-type. The presence of AVC was defined as an Agatston score >0. P values for trend across the quartiles were calculated by inserting the median value for each quartile. CI, confidence interval; OR, odds ratio. Other abbreviations as in Table 1.

In the sensitivity analyses treating lipid measures continuous, the strength of association with prevalent AVC per 1-SD higher value was the largest for LDL-c (adjusted OR=1.40, 95% CI: 1.15–1.72), followed by LDL-p (1.33, 1.09–1.63) (Table 3). HDL-p, but not HDL-c, was significantly inversely associated with prevalent AVC, such that 1 SD lower HDL-p was associated with 1.28 (1.02–1.61) higher OR.

Table 3. Multivariable-Adjusted OR for the Presence of AVC per 1 Standard Deviation Higher (or *Lower) Value of the Primary Lipid Indices, in Apparently Healthy Japanese Men (Aged 40–79 Years: 2006–2008, SESSA)

	OR	95% CI	P value
LDL-c, mg/dL	1.40	1.15–1.72	0.001
HDL-c*, mg/dL	1.12	0.90–1.40	0.320
LDL-p, nmol/L	1.33	1.09–1.63	0.005
HDL-p*, μmol/L	1.28	1.02–1.61	0.033

The results of post-hoc analyses were as follows (Supplementary Tables 3–10): (1) in stratified analyses, LDL-c tended to be more strongly associated with AVC among participants with MetS than in those without it (Supplementary Tables 3,4). By contrast, in participants with hypertension, the association was lower compared to that with LDL-p (Supplementary Table 5). Similar to participants with MetS, the association of LDL-c tended to be stronger in those with diabetes (Supplementary Table 6), and those with hypertriglyceridemia (Supplementary Table 7) as compared to those without them. However, a statistically significant difference across the strata was observed only in the quartile model of hypertriglyceridemia stratification (Supplementary Table 7). Regarding HDL, the direction of association between HDL-p and AVC tended to be opposite by the presence/absence of MetS in the quartile models with a significant interaction by Japanese MetS criteria (P value for interaction=0.012, Supplementary Table 4). However, no such interaction was observed in a continuous model using either of the MetS definitions, and their directions of association were the same as the main results. Otherwise, there was no clear evidence of interaction by any of the stratifying factors. Likewise, according to continuous models, there was no clear evidence supporting stronger associations of LDL-p or HDL-p compared to those of corresponding conventional lipids in any of the strata (Supplementary Tables 3,5,6); (2) as for extended lipid indices, non-HDL-c showed a significant linear trend across the quartile (P value for trend=0.03) (Supplementary Table 8). In the continuous model, 1 SD higher non-HDL-c was associated with an adjusted OR of 1.39 (1.13–1.71). TG was not associated with prevalent AVC in either of the models (quartile, continuous). None of the NMR-based LDL, or HDL subclasses and their particle sizes were significantly associated with AVC in quartile models. In continuous models, only large LDL-p was significantly positively associated with AVC (OR=1.23, 1.02–1.49) (Supplementary Table 9); (3) including the participants with TG ≥400 mg/dL (n=9) resulted in no significant association of AVC with HDL-c or with HDL-p in either of the models (quartile, continuous), although the point estimates in the continuous model indicated a stronger association with HDL-p compared to HDL-c (Supplementary Table 10), which is similar to the main results.

Discussion

In this population-based, cross-sectional study of apparently healthy Japanese men, we found positive associations of LDL-p and LDL-c with prevalent AVC, independent of conventional cardiovascular disease risk factors. In contrast, an association of prevalent AVC was observed only with HDL-p in a continuous model and not with HDL-c. To the best of our knowledge, this is the first study that directly compared NMR-measured lipoproteins with standard lipids in relation to AVC in a Japanese population.

AVC is not uncommon in a general elderly population, and it may lead to more advanced aortic valvular disease.¹^–³ Identifying risk factors for subclinical AVC is of the utmost importance to establish new therapeutic targets that might allow us to halt or at least slow the progression of disease. The positive association between LDL-c and AVC in our study is consistent with the results of previous studies.²⁶^,³⁰ Our study further suggests that LDL-p in addition to LDL-c may be a possible therapeutic target for preventing aortic valve calcification and consequent functional deterioration.

NMR-based lipoprotein parameters, LDL-p in particular, have been suggested as alternative lipid measures in assessing coronary artery diseases risk.³¹ LDL-p is an alternate LDL measure, and although it is highly correlated with LDL-c, the cholesterol amount carried by each LDL particle varies. Higher levels of circulating LDL-p may increase the chances that these particles may contribute to the oxidative transformation of lipids within the aortic valve, promoting valvular inflammation and remodeling, and thus cause disease progression. A study showed that statin use, guided by changes in LDL-p, was more cost-effective than a LDL-c-guided approach in preventing cardiovascular disease.¹³

Some observational studies suggested that statin therapy may slow the progression of AVC.³²^,³³ A meta-analysis of 4 randomized controlled trials found no differences in the clinically relevant outcomes such as mean pressure gradient, valve area, freedom from valve replacement and death from cardiovascular causes between the statin group and the placebo group,³⁴ but the authors of the meta-analysis stated that “this issue is not over, mainly because the available evidence is based on studies which have limitations such as follow-up, randomization process, sample size and very elderly participants with many comorbidities”. Indeed, the analyzed patients (n=269) in the meta-analysis had a mean age of 58 years, with asymptomatic mild aortic valve stenosis defined by maximum aortic valve velocity of 2.5–4.0 m/s. Agreeing with those authors, we think it is premature to conclude that there is no potential benefit for those in the subclinical stage of AVC to keep their lipids within the optimum range in hopes of preventing/delaying progression.

Comparison of Conventional Lipids and NMR-Based Lipoproteins

In comparing LDL-c and LDL-p in association with prevalent AVC, our results were sensitive to the choice of model. The point estimate (OR) in the 4^th quartile for LDL-c in reference to 1^st quartile was slightly smaller compared to the corresponding OR for LDL-p in the main model, whereas the opposite was the case in the continuous model. The post-hoc stratified analyses showed the association between LDL-c and AVC was stronger in individuals with hypertriglyceridemia in the quartile model, although the interaction by hypertriglyceridemia was not statistically supported in the continuous model (P value for interaction=0.262). Furthermore, no clear evidence supported a stronger association of LDL-p with AVC compared with that of LDL-c in any of the strata we examined. Taken together, LDL-p was similar but not clearly superior to LDL-c in the association with prevalent AVC.

Evidence of the potential role of HDL in AVC is scarce. Although serum HDL-c level has been shown to be inversely associated with hemodynamic progression of calcific aortic valve disease,³⁵ no studies have compared the strength of the association between HDL-c and NMR-based HDL-p in relation to AVC. Overall, the relationship of HDL-c or HDL-p with AVC was weaker compared to those of LDL counterparts in the main analyses. Among them, the association with HDL-p seemed be slightly stronger than HDL-c in the continuous model. Otherwise, post-hoc analyses did not show any different pattern of association from the main results. Some studies suggested that serum HDL-p concentration was more closely related to the performance rate of cholesterol efflux in the reverse cholesterol transport pathway than HDL-c;¹²^,³⁶^,³⁷ however, whether this is translated into HDL-p being a better predictor for AVC remains uncertain, thus, warranting further study.

Other Conventional Lipids and Subclasses of LDL-p, HDL-p and Their Size

In post-hoc analyses, we observed a strong association between non-HDL-c and AVC, but not between TG and AVC. Non-HDL-c was shown to be similar to, or even better than, LDL-c in predicting cardiovascular disease incidence and mortality,³⁸^,³⁹ and our finding is consistent with those reports. In the literature, there is no single subclass of NMR-measured LDL/HDL and their size that is consistently associated with clinical atherosclerotic diseases. In our post-hoc analysis, only large LDL-p was significantly positively associated with AVC. Given that this was one statistically significant finding of all the post-hoc analyses we added, its interpretation should be made carefully. Further study is needed to examine whether our finding will be reproduced in a different setting, and to elucidate the potential role of the subclasses of NMR-based LDL-p, HDL-p and their sizes in the association with AVC.

Limitations and Strengths

Our study has several limitations. First, the study design was cross-sectional, so causality in any relationship could not be established; however, we excluded participants with any history of valvular surgery, myocardial infarction and stroke, and those on lipid medication, to minimize reverse causality. In addition, we believe it is unlikely that our participants changed their behavior as a result of their AVC status because we did not inform them of their status, due to its unclear clinical significance. Also, our results may not be applicable to women or non-Japanese individuals given our studied sample was limited to apparently healthy Japanese men. However, this may also offer an advantage in limiting the confounding factors associated with gender, race and genetic variation, as well as diet and lifestyle. The strengths of the study include community-based random sampling of apparently healthy men enhancing generalizability, and masked assessment of the key component, AVC, from the participants’ characteristics to minimize information bias, as well as the standardized measurements of relevant parameters, including laboratory data, and a moderate sample size.

In conclusion, among LDL and HDL measures, both LDL-cholesterol and LDL-particle levels were positively associated with presence of AVC, and the strengths of association seemed similar in this community-based sample of asymptomatic Japanese men not using lipid-lowering medication.

Author Contributions

T.V., A.F., K.M. designed the study. T.V. performed analysis, interpreted the results, and drafted the manuscript. A.F. revised the manuscript. All authors contributed to the manuscript’s critical review and approval of the final version.

Acknowledgments

The authors would like to thank the investigators, staff, and study participants of Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA) for their outstanding dedication and commitment. The name of the members of the SESSA research group are provided in the Supplementary File.

Sources of Funding

This study has been supported by the Uehara Memorial Foundation; by the Young Investigator Grant of the Shimane University; by Grants-in-aid for Scientific Research (A) 13307016, (A) 17209023, (A) 21249043, (A) 23249036, (A) 25253046, (A) 15H02528, (B) 26293140, and (B) 18H03048 from the Ministry of Education, Culture, Sports, Science, and Technology, Japan; by the National Institute of Health in USA (R01HL068200), and by Glaxo-Smith Kline GB.

The present study was initiated and analyzed by the authors. The funding sources listed above had no role in the study design, collection, analyses, and interpretation of the results.

Disclosures

All authors declare no conflicts of interest.

IRB Information

The study was approved by the Institutional Review Board of Shiga University of Medical Science, Otsu, Japan.

Supplementary Files

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-20-1090

References

1. Hulin A, Hego A, Lancellotti P, Oury C. Advances in pathophysiology of calcific aortic valve disease propose novel molecular therapeutic targets. Front Cardiovasc Med 2018; 5: 21.
2. Hisamatsu T, Miura K, Fujiyoshi A, Kadota A, Miyagawa N, Satoh A, et al. Serum magnesium, phosphorus, and calcium levels and subclinical calcific aortic valve disease: A population-based study. Atherosclerosis 2018; 273: 145–152.
3. Lerman DA, Prasad S, Alotti N. Calcific aortic valve disease: Molecular mechanisms and therapeutic approaches. Eur Cardiol 2015; 10: 108–112.
4. Peeters FECM, Meex SJR, Dweck MR, Aikawa E, Crijns HJGM, Schurgers LJ, et al. Calcific aortic valve stenosis: Hard disease in the heart: A biomolecular approach towards diagnosis and treatment. Eur Heart J 2017; 39: 2618–2624.
5. Koos R, Mahnken AH, Sinha AM, Wildberger JE, Hoffmann R, Kühl HP. Aortic valve calcification as a marker for aortic stenosis severity: Assessment on 16-MDCT. AJR Am J Roentgenol 2004; 183: 1813–1818.
6. Liu F, Coursey CA, Grahame-Clarke C, Sciacca RR, Rozenshtein A, Homma S, et al. Aortic valve calcification as an incidental finding at CT of the elderly: Severity and location as predictors of aortic stenosis. AJR Am J Roentgenol 2006; 186: 342–349.
7. Pohle K, Maffert R, Ropers D, Moshage W, Stilianakis N, Daniel WG, et al. Progression of aortic valve calcification: Association with coronary atherosclerosis and cardiovascular risk factors. Circulation 2001; 104: 1927–1932.
8. Shimizu K, Yamamoto M, Koyama Y, Kodama A, Sato H, Kano S, et al. Usefulness of routine aortic valve calcium score measurement for risk stratification of aortic stenosis and coronary artery disease in patients scheduled cardiac multislice computed tomography. Int J Cardiol Heart Vasc 2015; 9: 95–99.
9. Otvos JD, Mora S, Shalaurova I, Greenland P, Mackey RH, Goff DC Jr. Clinical implications of discordance between low-density lipoprotein cholesterol and particle number. J Clin Lipidol 2011; 5: 105–113.
10. Zaid M, Fujiyoshi A, Miura K, Abbott RD, Okamura T, Takashima N, et al. High-density lipoprotein particle concentration and subclinical atherosclerosis of the carotid arteries in Japanese men. Atherosclerosis 2015; 239: 444–450.
11. Zaid M, Miura K, Fujiyoshi A, Abbott RD, Hisamatsu T, Kadota A, et al. Associations of serum LDL particle concentration with carotid intima-media thickness and coronary artery calcification. J Clin Lipidol 2016; 10: 1195–1202.e1191.
12. Mackey RH, Greenland P, Goff DC Jr, Lloyd-Jones D, Sibley CT, Mora S. High-density lipoprotein cholesterol and particle concentrations, carotid atherosclerosis, and coronary events: MESA (Multi-Ethnic Study of Atherosclerosis). J Am Coll Cardiol 2012; 60: 508–516.
13. Grabner M, Winegar DA, Punekar RS, Quimbo RA, Cziraky MJ, Cromwell WC. Cost effectiveness of achieving targets of low-density lipoprotein particle number versus low-density lipoprotein cholesterol level. Am J Cardiol 2017; 119: 404–409.
14. Kadota A, Miura K, Okamura T, Fujiyoshi A, Ohkubo T, Kadowaki T, et al. Carotid intima-media thickness and plaque in apparently healthy Japanese individuals with an estimated 10-year absolute risk of CAD death according to the Japan Atherosclerosis Society (JAS) guidelines 2012: The Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA). J Atheroscler Thromb 2013; 20: 755–766.
15. Ueshima H, Kadowaki T, Hisamatsu T, Fujiyoshi A, Miura K, Ohkubo T, et al. Lipoprotein-associated phospholipase A2 is related to risk of subclinical atherosclerosis but is not supported by Mendelian randomization analysis in a general Japanese population. Atherosclerosis 2016; 246: 141–147.
16. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499–502.
17. Horio M, Imai E, Yasuda Y, Watanabe T, Matsuo S. Modification of the CKD Epidemiology Collaboration (CKD-EPI) equation for Japanese: Accuracy and use for population estimates. Am J Kidney Dis 2010; 56: 32–38.
18. Imai E, Yasuda Y, Makino H. Japan Association of Chronic Kidney Disease Initiatives (J-CKDI). Japan Med Assoc J 2011; 54: 403–405.
19. Jeyarajah EJ, Cromwell WC, Otvos JD. Lipoprotein particle analysis by nuclear magnetic resonance spectroscopy. Clin Lab Med 2006; 26: 847–870.
20. Otvos JD, Jeyarajah EJ, Cromwell WC. Measurement issues related to lipoprotein heterogeneity. Am J Cardiol 2002; 90: 22i–29i.
21. Kashiwagi A, Kasuga M, Araki E, Oka Y, Hanafusa T, Ito H, et al. International clinical harmonization of glycated hemoglobin in Japan: From Japan Diabetes Society to National Glycohemoglobin Standardization Program values. J Diabetes Investig 2012; 3: 39–40.
22. Hisamatsu T, Fujiyoshi A, Miura K, Ohkubo T, Kadota A, Kadowaki S, et al. Lipoprotein particle profiles compared with standard lipids in association with coronary artery calcification in the general Japanese population. Atherosclerosis 2014; 236: 237–243.
23. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990; 15: 827–832.
24. Katz R, Wong ND, Kronmal R, Takasu J, Shavelle DM, Probstfield JL, et al. Features of the metabolic syndrome and diabetes mellitus as predictors of aortic valve calcification in the Multi-Ethnic Study of Atherosclerosis. Circulation 2006; 113: 2113–2119.
25. Linefsky JP, O’Brien KD, Sachs M, Katz R, Eng J, Michos ED, et al. Serum phosphate is associated with aortic valve calcification in the Multi-Ethnic Study of Atherosclerosis (MESA). Atherosclerosis 2014; 233: 331–337.
26. Owens DS, Katz R, Takasu J, Kronmal R, Budoff MJ, O’Brien KD. Incidence and progression of aortic valve calcium in the Multi-Ethnic Study of Atherosclerosis (MESA). Am J Cardiol 2010; 105: 701–708.
27. Tehrani DM, Zhao Y, Blaha MJ, Mora S, Mackey RH, Michos ED, et al. Discordance of low-density lipoprotein and high-density lipoprotein cholesterol particle versus cholesterol concentration for the prediction of cardiovascular disease in patients with metabolic syndrome and diabetes mellitus (from the Multi-Ethnic Study of Atherosclerosis [MESA]). Am J Cardiol 2016; 117: 1921–1927.
28. Alberti KG, Zimmet P, Shaw J. The metabolic syndrome: A new worldwide definition. Lancet 2005; 366: 1059–1062.
29. Taki K, Nishio K, Hamajima N, Niwa T. Metabolic syndrome defined by new criteria in Japanese is associated with increased liver enzymes and C-reactive protein. Nagoya J Med Sci 2008; 70: 1–9.
30. Owens DS, Katz R, Johnson E, Shavelle DM, Probstfield JL, Takasu J, et al. Interaction of age with lipoproteins as predictors of aortic valve calcification in the Multi-Ethnic Study of Atherosclerosis. Arch Intern Med 2008; 168: 1200–1207.
31. Cromwell WC, Otvos JD, Keyes MJ, Pencina MJ, Sullivan L, Vasan RS, et al. LDL particle number and risk of future cardiovascular disease in the Framingham Offspring Study: Implications for LDL Management. J Clin Lipidol 2007; 1: 583–592.
32. Shavelle DM, Takasu J, Budoff MJ, Mao S, Zhao XQ, O’Brien KD. HMG CoA reductase inhibitor (statin) and aortic valve calcium. Lancet 2002; 359: 1125–1126.
33. Rosenhek R, Rader F, Loho N, Gabriel H, Heger M, Klaar U, et al. Statins but not angiotensin-converting enzyme inhibitors delay progression of aortic stenosis. Circulation 2004; 110: 1291–1295.
34. Thiago L, Tsuji SR, Nyong J, Puga ME, Gois AF, Macedo CR, et al. Statins for aortic valve stenosis. Cochrane Database Syst Rev 2016; 9: Cd009571.
35. Olgun Küçük H, Küçük U, Demirtaş C, Özdemir M. Role of serum high density lipoprotein levels and functions in calcific aortic valve stenosis progression. Int J Clin Exp Med 2015; 8: 22543–22549.
36. Tan HC, Tai ES, Sviridov D, Nestel PJ, Ng C, Chan E, et al. Relationships between cholesterol efflux and high-density lipoprotein particles in patients with type 2 diabetes mellitus. J Clin Lipidol 2011; 5: 467–473.
37. Linsel-Nitschke P, Jansen H, Aherrarhou Z, Belz S, Mayer B, Lieb W, et al. Macrophage cholesterol efflux correlates with lipoprotein subclass distribution and risk of obstructive coronary artery disease in patients undergoing coronary angiography. Lipids Health Dis 2009; 8: 14.
38. Sigdel M, Yadav BK, Gyawali P, Regmi P, Baral S, Regmi SR, et al. Non-high density lipoprotein cholesterol versus low density lipoprotein cholesterol as a discriminating factor for myocardial infarction. BMC Res Notes 2012; 5: 640.
39. Sniderman AD, Williams K, Contois JH, Monroe HM, McQueen MJ, de Graaf J, et al. A meta-analysis of low-density lipoprotein cholesterol, non-high-density lipoprotein cholesterol, and apolipoprotein B as markers of cardiovascular risk. Circ Cardiovasc Qual Outcomes 2011; 4: 337–345.

Corresponding author

Register with J-STAGE for free!