Serum Values of Cholesterol Absorption and Synthesis Biomarkers in Japanese Healthy Subjects: The CACHE Study HEALTHY Analysis

Abstract

Aim: Blood cholesterol absorption and synthesis biomarkers predict cardiovascular risk. This study aimed to determine the values of serum non-cholesterol sterol markers [lathosterol (Latho), campesterol (Campe), and sitosterol (Sito)] in healthy individuals and factors affecting these markers.

Methods: The CACHE Consortium compiled clinical data, including serum Latho (cholesterol synthesis marker), and Campe and Sito (cholesterol absorption markers), by a gas chromatography method in 2944 individuals. Healthy subjects were selected by excluding those with prior cardiovascular disease, diabetes mellitus, hypertension, chronic kidney disease, familial hypercholesterolemia, sitosterolemia, current smokers, those with low (＜17 kg/m²) or high (≥ 30 kg/m²) body mass index (BMI), and those with treatment for dyslipidemia or hyperuricemia. Nonlinear regression stratified by sex was used to examine the associations of cholesterol metabolism markers with age, BMI, and serum lipid levels.

Results: Of 479 individuals selected, 59.4% were female; the median age was 48 years in females and 50 years in males. The three markers showed positively skewed distributions, and sex differences were present. Age was associated positively with Latho, inversely with Campe, but not significantly with Sito. BMI was associated positively with Latho, but not significantly with Campe or Sito. High-density lipoprotein cholesterol (HDL-C) was positively associated with Campe and Sito, but not significantly with Latho. Non-HDL-C was positively associated with the three markers.

Conclusion: Our study results in the healthy subjects help to interpret the non-cholesterol sterol markers for cardiovascular risk assessment in patients with cardiovascular risk factors.

See editorial vol. 30: 1307-1308

Introduction

Blood concentrations of non-cholesterol sterols have been measured as biomarkers of cholesterol absorption and cholesterol synthesis ¹⁾ . Serum concentrations of lathosterol (Latho) and desmosterol have been used for the assessment of hepatic cholesterol synthesis, whereas serum concentrations of campesterol and sitosterol have been utilized for the assessment of intestinal absorption of dietary cholesterol ¹⁾ . These biomarkers have been reported to be closely related to cardiovascular disease ^{2 - 5)} . Subjects with cardiovascular disease revealed higher levels of absorption biomarkers and lower levels of synthesis biomarkers than controls ^{4 , 5)} . As reviewed previously ⁶⁾ , many studies, but not all ^{7 - 11)} , showed that increased levels of cholesterol absorption biomarkers were associated with increased risk of cardiovascular disease. Therefore, knowledge of an individual’s cholesterol absorption and synthesis status could help in the stratification of cardiovascular risk.

Serum or plasma levels of these biomarkers in “healthy” populations have been reported and are summarized in Supplemental Table 1 ^{5 , 9 , 12 - 14)} . However, criteria for the selection of subjects were different among studies, leading to different study populations and different results. The blood concentrations of non-cholesterol sterols are known to be altered in some diseases, such as diabetes mellitus ^{15 , 16)} and kidney diseases ^{17 , 18)} . Furthermore, although the values of biomarkers may be affected by some factors, including age, sex, body mass index (BMI), and serum lipid levels, even in those without apparent diseases, previous studies did not necessarily consider these factors. Thus, we need such information to improve the interpretation of these biomarkers for cholesterol metabolism in healthy individuals.

Supplemental Table 1. Selected prior studies reporting non-cholesterol sterol levels.

	Matthan et al 2013	Dayspring et al 16) 2015	Ishibashi et al 9) 2018	Nunes et al 17) 2020	Yoshida et al 18) 2020
Exclusion criteria	CVD, medication for DL	No information was available regarding FH, medical history, or current medication	CVD, medication for HT, DM or DL	CVD, DM, medication for DL or chronic disease (except HT), BMI ≤ 18.5 or ≥ 35 kg/m2	CVD, SM, CKD, CLD, drinker, taking SUPP., medications for HT, DL or DM, BMI ＜ 17 or ≥ 30 kg/m2
Number of subjects	2616	667,718	256	344	260
Male (%)	56	47	63	100	42
Age (years)		56±15	55±10	56±4	36±11 (M) 36±11 (F)
BMI (kg/m2)			22.8±3.2	25.0±3.0	22.1±3.3 (M) 20.8±2.1 (F)
Blood specimen	Plasma	Plasma	Serum	Plasma	Serum
Measuring method	GC	LC	GC	GC	GC
Synthesis biomarkers
Lathosterol	114.3±1.5 (M)＊ 110.3±1.2 (F)＊		3.05±1.52 μg/ml	1.12±0.59 μg/ml	1.84±0.84 μg/ml (M) 1.44±0.54 μg/ml (F)
Desmosterol	59.4±0.8 (M)＊ 53.8±0.6 (F)＊	0.99±1.24 μg/ml
Absorption biomarkers
Campesterol	225.0±2.9 (M)＊ 214.8±2.5 (F)＊	3.33±1.83 μg/ml	6.24±2.80 μg/ml	1.36±0.96 μg/ml	4.26± 1.57μg/ml (M) 4.79±2.04 μg/ml (F)
Sitosterol	167.7±2.2 (M)＊ 160.2±21.8 (F)＊	2.45±1.39 μg/ml		2.25±1.38 μg/ml	2.12±0.83 μg/ml (M) 2.41±1.09 μg/ml (F)
Associations in analysis	age, BMI, sex, SM, lipids	age, sex, lipids	lipids, vascular function	lipids	sex

Values are mean±standard deviation (other than Matthan et al ) or mean±standard error (Matthan et al ).

^＊Values in the study by Matthan et al are the ratio of non-cholesterol sterol concentration to total cholesterol in μmol/mmol of cholesterol.

Abbreviations: BMI, body mass index; CKD, chronic kidney disease; CLD, chronic liver disease; DL, dyslipidemia; DM, diabetes mellitus; FH,

family history; GC, gas chromatography; HT, hypertension; LC, liquid chromatography; SM, smoking; SUPP, supplement; F, female; M, male.

Aim

In this study, named “the CACHE study HEALTHY analysis”, we examined the serum concentration of Latho as a biomarker for cholesterol synthesis, and campesterol (Campe) and sitosterol (Sito) as biomarkers for cholesterol absorption, to obtain their reference intervals. Moreover, we analyzed the associations of these biomarkers with age, sex, BMI, and serum levels of lipids and lipoprotein cholesterol in healthy subjects.

Methods

Ethical Consideration

The study protocol was reviewed and approved by the Ethics Committee, Osaka City University Graduate School of Medicine, Osaka, Japan (Approval No. 3871), and was registered at UMIN-CTR (UMIN000030635). Further, the protocol of this study was approved by the review board of each participating institution prior to the study.

Clinical Data Collection

The methods of the CACHE study have been described in detail by Shoji et al. ¹⁹⁾ . Briefly, clinical data, including three biomarkers, were collected from 13 research groups in Japan and compiled using the web-based system called Research Electronic Data Capture (REDCap) ^{20 , 21)} . Then, the CACHE population (n=2944) was defined.

Selection of Healthy Subjects

To define healthy subjects in the total CACHE population, we excluded those with pre-existing cardiovascular disease, diabetes mellitus, hypertension, chronic kidney disease (CKD), familial hypercholesterolemia, sitosterolemia, smokers, those with BMI ＜17 or ≥ 30 kg/m², those using medications for dyslipidemia or hyperuricemia, and those treated with low-density lipoprotein (LDL) apheresis.

Diabetes mellitus was defined by a previous diagnosis of diabetes mellitus, use of any antidiabetic medication, fasting plasma glucose (FPG) of 126 mg/dL or higher, or a hemoglobin A1c (HbA1c) value of 6.5% or higher by the National Glycohemoglobin Standardization Program (NGSP), according to the diagnostic criteria by the American Diabetes Association and the Japan Diabetes Society (JDS) ^{22 , 23)} . In the case of JDS HbA1c values in medical records, these values were converted to NGSP values using a conversion formula provided by the JDS ²⁴⁾ .

Hypertension was defined by the use of any antihypertensive medication, systolic blood pressure of 140 mmHg or higher, or diastolic blood pressure of 90 mmHg or higher, according to the criteria by the Japanese Society of Hypertension ²⁵⁾ .

CKD was defined in this study by an estimated glomerular filtration rate (eGFR) lower than 60 mL/min/1.73 m² using the equation for the Japanese ²⁶⁾ . Information on proteinuria was not available in the CACHE dataset.

Familial hypercholesterolemia was diagnosed by the criteria from the Japan Atherosclerosis Society ²⁷⁾ . We excluded subjects with low-density lipoprotein cholesterol (LDL-C) ≥ 180 mg/dL because of suspected familial hypercholesterolemia.

Assay for Non-Cholesterol Sterols

Serum concentrations of Latho, Campe, and Sito were measured by using a gas chromatography method described elsewhere in detail ¹⁴⁾ . Furthermore, the ratio of Campe to Latho (Campe/Latho ratio) was calculated to assess the relative status of cholesterol absorption to that of cholesterol synthesis.

Other Variables

Clinical data were collected from medical records or data sets for research purposes regarding the following items: (1) clinical background, including age, sex, smoking status, high-risk conditions [prior coronary artery disease, prior stroke, prior peripheral artery disease, diabetes mellitus, CKD including dialysis, and familial hypercholesterolemia], and comorbidity, such as hypertension and hyperuricemia; (2) blood tests, including total cholesterol, triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), LDL-C, FPG, HbA1c, serum creatinine, eGFR, uric acid, serum albumin, aspartate transaminase, alanine transaminase, C-reactive protein, and red blood cells, hemoglobin, mean corpuscular volume, white blood cells, and platelet counts; (3) physical examination and vital signs, including height, body weight, BMI, systolic blood pressure, diastolic blood pressure, and pulse rate; (4) medication use, including drugs for dyslipidemia [statin, fibrate, ezetimibe, resin, probucol, omega-3 polyunsaturated fatty acid, nicotinic acid, proprotein convertase subtilisin/kexin type 9 inhibitor, and microsomal triglyceride transfer protein inhibitor], hypertension, diabetes mellitus, and hyperuricemia; and (5) specific treatments including hemodialysis and LDL apheresis.

Regarding lipid parameters, we used the following rules: (1) TG and HDL-C values were used as entered; (2) using total cholesterol (TC), TG, and HDL-C, non-HDL cholesterol (non-HDL-C) was calculated by subtracting HDL-C from TC, and LDL-C was calculated using the Friedewald formula ²⁸⁾ ; (3) if LDL-C measured by a homogeneous assay was entered but TC was not available, the LDL-C by a homogeneous assay was used for analyses, and non-HDL-C was calculated as LDL-C plus TG/5; (4) if LDL-C or non-HDL-C could not be calculated because of missing TG or HDL-C values, it was handled as missing. In the CACHE study HEALTHY analysis, we presented data of TG, HDL-C, non-HDL-C, and LDL-C thus determined.

As we excluded patients with CKD treated with hemodialysis for whom casual blood sampling was done, the laboratory values in this CACHE study HEALTHY analysis were collected in the fasting state in the morning.

Statistical Analyses

To present clinical characteristics of the subjects of this analysis, continuous and categorical variables were summarized by medians (interquartile ranges, IQRs) and numbers (percentages), respectively, and compared by using the Kruskal–Wallis test or Fisher’s exact test, respectively. The associations of serum markers for cholesterol metabolism with age, BMI, HDL-C, non-HDL-C, and TG were analyzed by using nonlinear regression models that considered the restricted cubic spline term for the exposure variable with three knots (10th, 50th, and 90th percentile levels). Adjustments were made for age, BMI, systolic blood pressure, diastolic blood pressure, FPG, HbA1c, and creatinine in each female and male separately. To meet the normal assumption of the regression model, the objective variables were logarithmically transformed and then used in the regression models. In the above regression models, all the missing values were complemented through the multiple imputation methods on the basis of the predictive mean matching approach. All statistical inferences were conducted with a two-sided 5% significance level using R software version 4.0.3.

Results

Selection of Healthy Individuals

Fig.1 shows the selection of participants for the CACHE study HEALTHY analysis. In total, 479 subjects were selected. Table 1 shows the clinical characteristics of the selected study participants by sex. Age was similar in females and males. BMI, systolic blood pressure, and diastolic blood pressure were significantly lower in females than in males. FPG was also significantly lower in females. TC, LDL-C, and non-HDL-C were similar in females and males, but HDL-C was significantly higher in females than in males. TG was significantly higher in males.

Fig.1. Selection of participants for the CACHE study HEALTHY analysis.

The participants of the CACHE study HEALTHY analysis were selected from the CACHE population by excluding the following individuals: those with prior cardiovascular disease, diabetes mellitus, hypertension, chronic kidney disease, familial hypercholesterolemia, sitosterolemia, current smoking, high (≥ 30.0 kg/m²) or low BMI (＜17.0 kg/m²), those treated for dyslipidemia, and those with medication for hyperuricemia.

Abbreviations: CVD, cardiovascular disease; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; LDL-C, low-density lipoprotein cholesterol; BMI, body mass index.

Table 1. Characteristics of participants of this analysis.

Variables	Unit	Total subjects	Missing	Female	Male	p value
Subjects	number	479		285	194
Age	years	49 (36, 59)	0 (0.0)	48 (35, 59)	50 (37, 59)	0.425
BMI	kg/m2	21.8 (20.2, 23.4)	0 (0.0)	21.3 (20.0, 23.1)	22.3 (20.8, 24.3)	＜0.001
SBP	mmHg	118 (110, 127)	0 (0.0)	116 (107, 124)	122 (115, 130)	＜0.001
DBP	mmHg	71 (66, 77)	0 (0.0)	69 (64, 76)	74 (69, 80)	＜0.001
WBC	/μL	5300 (4357, 6400)	37 (7.7)	5400 (4405, 6600)	5200 (4300, 6202)	0.187
RBC	x104/μL	457 (432, 483)	37 (7.7)	440 (421, 459)	485 (467, 514)	＜0.001
FPG	mg/dL	96.0 (90.0, 101.5)	0 (0.0)	94.0 (89.0, 99.0)	98.0 (92.0, 103.0)	＜0.001
HbA1c	%	5.50 (5.20, 5.70)	0 (0.0)	5.50 (5.20, 5.70)	5.40 (5.20, 5.70)	0.591
TC	mg/dL	208.0 (184.0, 229.5)	44 (9.2)	209.5 (184.0, 234.0)	206.0 (184.0, 224.0)	0.101
TG	mg/dL	79.0 (61.0, 108.0)	0 (0.0)	73.0 (58.0, 96.0)	94.5 (70.0, 120.0)	＜0.001
HDL-C	mg/dL	65.1 (55.9, 77.8)	0 (0.0)	70.0 (60.0, 80.8)	59.0 (51.1, 66.9)	＜0.001
LDL-C	mg/dL	124.4 (101.4, 145.8)	0 (0.0)	123.2 (100.0, 46.8)	126.9 (103.6, 144.5)	0.528
non-HDL-C	mg/dL	140.8 (117.7, 165.8)	1 (0.2)	137.8 (114.4, 166.0)	146.4 (124.1, 165.0)	0.086
Albumin	g/dL	4.40 (4.20, 4.60)	237 (49.5)	4.40 (4.20, 4.57)	4.40 (4.20, 4.60)	0.444
AST	U/L	20.0 (18.0, 24.0)	0 (0.0)	20.0 (17.0, 23.0)	22.0 (18.0, 26.0)	＜0.001
ALT	U/L	17.0 (13.0, 22.0)	0 (0.0)	15.0 (12.0, 19.0)	20.0 (15.0, 27.8)	＜0.001
Creatinine	mg/dL	0.68 (0.60, 0.80)	0 (0.0)	0.61 (0.56, 0.67)	0.83 (0.76, 0.91)	＜0.001
eGFR	mL/min/1.73m2	79.6 (71.7, 92.7)	0 (0.0)	81.6 (72.3, 96.0)	77.1 (70.9, 89.3)	0.005
UA	mg/dL	4.80 (3.90, 5.70)	36 (7.52)	4.10 (3.42, 4.77)	5.80 (5.10, 6.60)	＜0.001
CRP	mg/dL	0.03 (0.02, 0.06)	237 (49.5)	0.03 (0.02, 0.06)	0.03 (0.02, 0.06)	0.706

The table gives medians (interquartile ranges) for continuous variables and numbers (percentages) for categorical variables.

Abbreviations: BMI; body mass index, SBP; systolic blood pressure, DBP; diastolic blood pressure, WBC; white blood cell, RBC; red blood cell, FPG; fasting plasma glucose, TC; total cholesterol, TG; triglyceride, LDL-C; low-density lipoprotein cholesterol, HDL-C; high-density lipoprotein cholesterol, AST; aspartate aminotransferase, ALT; alanine aminotransferase, eGFR; estimated glomerular filtration rate, UA; uric acid, CRP:C- reactive protein.

Serum Concentration and Distribution of Sterol Biomarkers

Fig.2 shows the histograms of serum concentrations of Latho, Campe, and Sito, and the Campe/Latho ratios in females and males. All these variables showed positively skewed distributions in both females and males, that is, the modes of the distribution were shifted to the left, with their tails longer on the right side. The reference intervals (2.5th to 97.5th percentile ranges) of these markers are indicated in Fig.2 .

Fig.2. Histograms of serum biomarkers for cholesterol metabolism.

The distribution is shown by density. The 2.5th, 25th, 50th, and 97.5th percentile levels are also indicated.

Abbreviations: Campe, campesterol; Latho, lathosterol.

Associations of Age with Sterol Biomarkers

Fig.3 shows multivariable-adjusted associations of age with Latho, Campe, and Sito stratified by sex. Age was associated positively with Latho, inversely with Campe, and not significantly with Sito. These associations were not linear particularly in males; the association between age and Latho was not apparent in males aged 50 years and older, whereas the association between age and Campe was not apparent in males aged 50 years and younger. The association between age and Sito was an inverted U-shape in males. Campe was lower in females than in males. No significant interaction by sex was found in the association of age with Latho, Campe, or Sito.

Fig.3. Associations of age with biomarkers for cholesterol metabolism by sex.

The associations of age with the biomarkers for cholesterol metabolism are shown in multivariable-adjusted nonlinear regression stratified by sex. The curves and shaded areas indicate estimated means and 95% confidence intervals.

Associations of BMI with Sterol Biomarkers

Fig.4 shows multivariable-adjusted associations of BMI with the three biomarkers stratified by sex. BMI was associated positively with Latho, but not significantly with Campe or Sito. Campe was lower in females than in males. No significant interaction by sex was found in the association of BMI with Latho, Campe, or Sito.

Fig.4. Associations of BMI with biomarkers for cholesterol metabolism by sex.

The associations of BMI with the biomarkers for cholesterol metabolism are shown in multivariable-adjusted nonlinear regression stratified by sex. The curves and shaded areas indicate estimated means and 95% confidence intervals.

Abbreviation: BMI, body mass index.

Associations of Serum Lipids with Sterol Biomarkers

Fig.5 shows the associations of HDL-C with the three biomarkers. HDL-C was positively associated with Campe and Sito, but not significantly with Latho. Campe was lower in females than in males. No significant interaction by sex was found in the association of HDL-C with Latho, Campe, or Sito.

Fig.5. Associations of HDL-C with biomarkers for cholesterol metabolism by sex.

The associations of HDL-C with the biomarkers for cholesterol metabolism are shown in multivariable-adjusted nonlinear regression stratified by sex. The curves and shaded areas indicate estimated means and 95% confidence intervals.

Abbreviation: HDL-C, high-density lipoprotein cholesterol.

Fig.6 shows the associations of non-HDL-C with the three biomarkers. Non-HDL-C was positively associated with the three biomarkers. Campe and Sito levels were significantly lower in females than in males. No significant interaction by sex was found in the association of non-HDL-C with Latho, Campe, or Sito.

Fig.6. Associations of non-HDL-C with biomarkers for cholesterol metabolism by sex.

The associations of non-HDL-C with the biomarkers for cholesterol metabolism are shown in multivariable-adjusted nonlinear regression stratified by sex. The curves and shaded areas indicate estimated means and 95% confidence intervals.

Abbreviation: non-HDL-C, non-high-density lipoprotein cholesterol.

Fig.7 shows the association of TG with the three biomarkers. TG was not significantly associated with any of these markers. Campe was lower in females than in males. No significant interaction by sex was found in the association of TG with Latho, Campe, or Sito.

Fig.7. Associations of TG with biomarkers for cholesterol metabolism by sex.

The associations of TG with the biomarkers for cholesterol metabolism are shown in multivariable-adjusted nonlinear regression stratified by sex. The curves and shaded areas indicate estimated means and 95% confidence intervals.

Abbreviation: TG, triglyceride.

Associations of LDL-C to HDL-C Ratio with Sterol Biomarkers

Fig.8 shows the associations of the LDL-C/HDL-C ratio with the three biomarkers. The LDL-C/HDL-C ratio showed inverse associations with concentrations of Campe and Sito, but not significantly with Latho. Campe was lower in females than in males. No significant interaction by sex was found in the association of the LDL-C/HDL-C ratio with Latho, Campe, or Sito.

Fig.8. Association of LDL-C/HDL-C with biomarkers for cholesterol metabolism by sex.

The associations of the LDL-C/HDL-C ratio with the biomarkers for cholesterol metabolism are shown in multivariable-adjusted nonlinear regression stratified by sex. The curves and shaded areas indicate estimated means and 95% confidence intervals.

Abbreviation: LDL-C/HDL-C, the ratio of low-density lipoprotein cholesterol to high-density lipoprotein cholesterol.

Associations of Age, BMI, HDL-C, Non-HDL-C, and TG with the Campe/Latho Ratio

Fig.9 shows the associations of the Campe/Latho ratio with age, BMI, HDL-C, non-HDL-C, and the LDL-C/HDL-C ratio. The Campe/Latho ratio was associated inversely with age, BMI, and the LDL-C/HDL-C ratio; positively with HDL-C; but did not significantly with non-HDL-C or TG. The Campe/Latho ratio was not different between sexes. No significant interaction by sex was found in these associations.

Fig.9. Associations of Campe/Latho ratio with age, BMI, and serum lipids by sex.

The associations of the Campe/Latho ratio with age, BMI, and serum lipids are shown in multivariable-adjusted nonlinear regression stratified by sex. The curves and shaded areas indicate estimated means and 95% confidence intervals.

Abbreviations: Campe, campesterol; Latho, lathosterol; BMI, body mass index; HDL-C, high-density lipoprotein cholesterol; non-HDL-C, non-high-density lipoprotein cholesterol; LDL-C/HDL-C ratio, the ratio of low-density lipoprotein cholesterol to high-density lipoprotein cholesterol.

Discussion

Using the relatively large sample of the CACHE study, we showed values and distributions of serum Latho, Campe, and Sito concentrations in well-characterized healthy subjects to provide their reference intervals, and examined the associations of these biomarkers with age, BMI, and serum lipids, stratified by sex. These three markers showed positively skewed distributions, and sex differences were present. Age was associated positively with Latho and inversely with Campe and the Campe/Latho ratio, although the associations were nonlinear. BMI was associated positively with Latho and inversely with the Campe/Latho ratio. HDL-C was associated positively with Campe, Sito, and the Campe/Latho ratio, and non-HDL-C was associated positively with Latho, Campe, and Sito. The LDL-C/HDL-C ratio was associated inversely with Campe, Sito, and the Campe/Latho ratio, but not significantly with Latho.

Although there are several previous studies regarding cholesterol absorption and synthesis biomarkers in healthy subjects, the criteria for “healthy” conditions were different. Furthermore, these previous studies did not necessarily consider the possible influence of age, sex, BMI, and serum lipids on the biomarkers of cholesterol metabolism, as shown in Supplemental Table 1 . Our results were similar to some previous reports, but some differences should be discussed.

Distribution of Serum Concentrations of Non-Cholesterol Sterols

The positively skewed distributions of Latho, Campe, and Sito concentrations of this study from Japan agreed with the results of Dayspring et al. ¹²⁾ from the US. Ishibashi et al. ⁹⁾ and Yoshida et al. ¹⁴⁾ also reported serum levels of Latho, Campe, and Sito in healthy subjects in Japan, although the distribution was not clearly presented. The concentrations of these markers are almost comparable between these studies. In contrast, a report from Brazil ¹³⁾ reported a much lower Campe concentration in males (1.368±0.960 µg/mL) than that in our results (median value 4.30 µg/mL). As BMI is inversely associated with Campe, as shown in this study, the difference in Campe levels may be explained by the difference in BMI between studies. Otherwise, the intake of vegetables and fruits may be a factor affecting serum Campe levels, as Campe is a phytosterol that is rich in these foods.

Associations of Age with Non-Cholesterol Sterol Levels

In this study, age was associated positively with Latho and inversely with Campe in both males and females, but not significantly with Sito. These associations of age with these non-cholesterol sterols were not linear, particularly in males, although the interaction by sex was not statistically significant. The observed age-related changes in non-cholesterol sterol levels are in agreement with those of a previous report by Matthan et al. ⁵⁾ but different from those of Dayspring et al.’s report ¹²⁾ which stated that Campe levels in males were gradually lower in higher age groups than those in the 30–50-year-old group; this pattern was similar to our results. However, they reported that the serum concentration of desmosterol, a marker of cholesterol synthesis, was the highest in males aged 30–40 years, and it was lower in higher age groups. This pattern is in sharp contrast with our results and those of Matthan et al. ⁵⁾ .

The reason for this discrepancy in age-related changes in cholesterol synthesis biomarkers may be the differences in populations studied. In the CACHE study HEALTHY analysis, we employed stringent eligibility criteria, excluding those with prior cardiovascular disease, those receiving treatment for dyslipidemia, and those with medical conditions, such as diabetes mellitus, hypertension, and CKD. Matthan et al. ⁵⁾ analyzed data from 1785 participants of the Framingham Offspring Study who had no history of cardiovascular disease and did not take lipid-lowering medication. In contrast, Dayspring et al. ¹²⁾ analyzed data from 667,718 patient blood samples submitted for testing. Only information on age and sex was available; no information on medical history or current medications was available. Therefore, we speculate that the inverse association between age and desmosterol observed by Dayspring et al. ¹²⁾ can be explained by the more common use of statins among those aged more than 30 years.

Our results and those of Matthan et al. ⁵⁾ suggest higher hepatic synthesis and lower intestinal absorption of cholesterol at older age. Age-related hypercholesterolemia is well known in humans ²⁹⁾ and rodents ³⁰⁾ . Sterol-responsive element-binding proteins (SREBPs) regulate metabolic homeostasis including cholesterol metabolism ³¹⁾ . The activation of SREBP2 upregulates the expression of genes responsible for cholesterol synthesis and uptake, such as 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMG-CoAR) and Nieman-Pick C1-like 1 (NPC1L1). In ad libitum fed aged rats, hepatic HMG-CoAR was fully activated, followed by hypercholesterolemia ³²⁾ . Thus, the age-associated higher levels of Latho seem to reflect the age-related activation of hepatic HMG-CoAR in humans also, although the precise molecular mechanisms remain to be elucidated.

The inverse association of age with Campe is a common finding among studies: our study, the Framingham Offspring Study ⁵⁾ , and the study by Dayspring et al. ¹²⁾ . As mentioned above, age-related overexpression of hepatic HMG-CoAR and age-related hypercholesterolemia are well described. Thus, a “compensatory” decrease in intestinal cholesterol absorption may occur. In response to the overexpression of hepatic HMG-CoAR and hypercholesterolemia, cells in peripheral tissues would uptake more cholesterol from the circulating LDL, and cholesterol demand would be decreased in peripheral cells including intestinal epithelial cells. We speculate that the reciprocal associations of age with Latho and Campe reflect the abovementioned age-related increase in cholesterol biosynthesis and compensatory decrease in intestinal cholesterol absorption.

Associations of BMI with Non-Cholesterol Sterol Levels

BMI was positively associated with Latho in both sexes in this study. This was consistent with previous studies ^{33 - 36)} . As caloric restriction decreased hepatic HMG-CoAR and restored hypercholesterolemia via modulation of SREBPs in aged rats ³⁷⁾ , the observed positive association between BMI and Latho may be explained by calorie intake. In contrast, our study did not show a significant association of BMI with Campe or Sito, although previous studies showed significant inverse associations between cholesterol absorption markers and BMI in obesity ^{33 - 36)} , diabetes ^{38 , 39)} or metabolic syndrome ^{40 , 41)} . The apparent discrepancy between studies may be explained by the differences in study populations. Our study excluded those with obesity (BMI ≥ 30 kg/m²) as well as lean subjects (BMI ＜17 kg/m²) and those with diabetes mellitus.

Association of Serum Lipid Parameters with Non-Cholesterol Sterol Levels

Campe and Sito were positively associated with both HDL-C and non-HDL-C levels, whereas Latho was positively associated with non-HDL-C but not significantly with HDL-C. These findings suggest the unequal distribution of Latho between HDL and non-HDL lipoproteins. It is presumable that newly synthesized cholesterol in the liver and a part of its precursor Latho are carried in very low-density lipoprotein (VLDL); they remain in LDL during the metabolism by lipoprotein lipase, and esterified forms of cholesterol and Latho in VLDL–LDL lipoproteins are transported to HDL particles via the action of the cholesteryl-ester transfer protein. These lipoprotein metabolism and lipid exchange processes may explain the observed closer association of Latho with non-HDL-C than with HDL-C.

Our results showing positive associations of Latho and Campe with non-HDL-C were quite different from the results from the Framingham Offspring Study by Matthan et al. ⁵⁾ . They reported inverse associations of non-HDL-C with both Latho and Campe. The discrepancy between the two studies may be explained by the different ways of expressing non-cholesterol sterol levels, namely, we reported the absolute concentration of non-cholesterol sterols, whereas Matthan et al. reported the ratio of non-cholesterol sterol concentration to TC concentration. Both ways of expression have been used to describe the level of non-cholesterol sterols ⁴²⁾ , and it is not established which method is more appropriate. It is true that enzyme activity can be estimated by the ratio of product concentration to the precursor concentration. However, both Latho and cholesterol are downstream of the rate-limiting enzyme HMG-CoAR. Therefore, the ratio of Latho to TC does not estimate the enzyme activity. In addition, Campe is not metabolized into cholesterol. Clearly, further studies are needed to determine which is the more appropriate method to describe the levels of cholesterol metabolism markers, the concentration or the ratio to TC.

Associations of LDL-C/HDL-C Ratio with Non-Cholesterol Sterols

The LDL-C/HDL-C ratio is known to be a better indicator of cardiovascular risk than individual parameters such as LDL-C or HDL-C ^{43 - 47)} . In this study, the LDL-C/HDL-C ratio was associated inversely with Campe, Sito, and the Campe/Latho ratio, but not significantly with Latho. Based on these associations, one may speculate that a higher cholesterol absorption predicts a lower cardiovascular risk. However, the risk of a cardiovascular event does not depend merely on dyslipidemia but also on other factors. More importantly, we did not assess the risk of future CVD in this cross-sectional analysis.

Effects of sex on associations of age with non-cholesterol sterols

As shown in Fig.3 , Campe levels were lower in females than in males, which is consistent with the Framingham Offspring Study results ⁵⁾ , but were different from the study results reported by Dayspring et al. ¹²⁾ , which showed that the Campe concentration was higher in males aged between 30 and 50 years, but lower in males aged 50 years and more. The discrepancy between the two studies may be explained by the important difference in the study populations as mentioned above and the difference in statistical analyses, with and without adjustment for potential confounders. Alternatively, the lower Campe concentration in females than in males observed in the studies by us and Matthan et al. may be due to lower dietary and cholesterol intake in females ⁴⁸⁾ and the age-associated decrease in the positive effect of estrogen on the upregulation of intestinal Niemann-Pick C1-like protein 1 ⁴⁹⁾ .

Study Limitations

There are some limitations to this study. First, the sample size was smaller than those in studies by Matthan et al. ⁵⁾ and Dayspring et al. ¹²⁾ . However, the sample size of 479 in this analysis was sufficient to address the research questions of this study using multivariable-adjusted analysis. Second, the methods of biomarker measurements were different from other studies ( Supplemental Table 1 ) . However, the assay method used in this study was reliable, as shown before ¹⁴⁾ . Third, the results may be confounded by some unmeasured variables, such as menopausal status, gastrointestinal function, and dietary habit, including age-specific dietary patterns ⁵⁰⁾ . Fourth, this study did not intend to examine the predictive value of non-cholesterol sterols for cardiovascular outcomes. Fifth, because of this study’s cross-sectional design, the associations do not necessarily indicate causality.

Conclusions

This study showed reference intervals and distributions of serum Latho, Campe, and Sito, and the Campe/Latho ratio in healthy Japanese individuals. These results of the healthy subjects would help to interpret these non-cholesterol sterol markers for cardiovascular risk assessment in patients with cardiovascular risk factors and for understanding the pathophysiology of any disease by comparative studies with healthy individuals. Further studies are thus needed to utilize these biomarkers in the future work.

Acknowledgements

Part of this study was presented at the 53rd Annual Meeting of the Japan Atherosclerosis Society (October 22-23, 2021, Kyoto, Hybrid version) and at the 18th International Symposium of Atherosclerosis (October 24-27, 2021, Kyoto, Hybrid version).

Funding

This study was supported by a grant to TS from Bayer Yakuhin Ltd. The funder played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Conflict of Interest

Hiroshi Yoshida reported personal fee from Denka Company Ltd and Kowa Company Ltd. Tetsuya Matoba reported personal fee from Bayer Yakuhin Ltd and MSD; and research grant from Amgen and Kowa. Tatsuro Ishida reported personal fee from Bayer Yakuhin Ltd and Kowa Inc. Yasushi Ishigaki reported personal fee from Bayer Yakuhin, Kowa Pharmaceutical Company, MSD, Novartis, Novo Nordisk, Ono Pharmaceutical, Sanofi K.K., and Takeda Pharmaceutica; research grant from Daiichi Sankyo, and Takeda Science Foundation; and Scholarship grant from MSD and Ono Pharmaceutical. Shizuya Yamashita reported personal fee from Kowa. Tomoko Nakagami reported personal fee from Sanwa Kagaku Kenkyusho Co Ltd, Novo Nordisk Pharma Ltd Japan, Eli Lily Japan KK, Sanofi K.K., Sumitomo Pharma, and Boehringer Ingelheim Japan Inc. Tetsuo Shoji reported personal fee and research grant from Bayer Yakuhin Ltd. Other authors reported no financial conflict of interest relevant to this study.