Serum High-Density Lipoprotein Cholesterol Levels and the Risk of Kidney Function Decline: The Japan Specific Health Checkups (J‑SHC) Study

Abstract

Aims: Both low and high serum levels of high-density lipoprotein cholesterol (HDL-C) were reported to be associated with adverse kidney outcomes. However, this association has not been well investigated in the general Japanese population.

Methods: This nationwide longitudinal study used data from the Japan Specific Health Checkups Study conducted between 2008–2014. The association between serum HDL-C levels and 40% decline in estimated glomerular filtration rate (eGFR) was analyzed using Cox regression analysis. Trajectories of eGFR were compared using mixed-effects model.

Results: Among 768,495 participants, 6,249 developed 40% decline in eGFR during the median follow-up period of 34.6 (interquartile range: 14.8–48.4) months. Using serum HDL-C levels of 40–59 mg/dL as a reference, the adjusted hazard ratios (95% confidence intervals) for the kidney outcome of serum HDL-C levels of ＜40, 60–79 and ≥ 80 mg/dL were 1.26 (1.14–1.39), 0.91 (0.86–0.96), and 0.86 (0.78–0.93), respectively. Restricted cubic spline analysis showed that HDL-C levels of less than approximately 60 mg/dL were associated with an increased risk of kidney outcomes. Subgroup analysis showed that baseline eGFR and proteinuria modified the effects of serum HDL-C levels on kidney outcomes. The mixed-effects model showed that the lower category of HDL-C level was associated with a higher eGFR decline rate (p for interaction ＜0.001).

Conclusions: Low HDL-C levels were associated with kidney function decline; however, high HDL-C levels were not associated with adverse kidney outcomes in the general Japanese population.

See editorial vol. 32: 405-406

Introduction

High-density lipoprotein (HDL) has several protective effects on atherosclerosis, including reverse cholesterol transport (RCT) and anti-inflammatory, antioxidative, and antithrombotic effects¹⁾. Serum HDL-cholesterol (HDL-C) level is a useful predictor for cardiovascular disease (CVD), since it has been reported that low serum HDL-C levels are associated with CVD and mortality^2-5). However, serum HDL-C levels do not always reflect HDL function^{6, 7)}, and several studies have shown that high levels were also associated with adverse outcomes^{4, 8)}.

Chronic kidney disease (CKD) has been widely recognized as a public health concern, with its global prevalence reported to be 9.1%⁹⁾. In individuals with CKD, lipid metabolism is altered by many factors^{10, 11)}. Low serum HDL-C levels are one of the characteristic dyslipidemias in patients with CKD¹²⁾, and CKD is reported to be associated with an increased risk for new onset of low HDL cholesterolemia among the general population¹³⁾. This increased incidence is explained by several mechanisms, such as decreased apolipoprotein (apo) A-I synthesis by the liver and defective lecithin-cholesterol acyltransferase (LCAT) concentration and activity, which are involved in impaired RCT in patients with CKD^{10, 11)}. Moreover, decreased HDL-C levels may also interactively affect the progression of kidney disease, as has been reported by several studies^{14, 15)}. In fact, several studies reported that decreased serum HDL-C levels were associated with kidney disease progression^16-20). Notably, this was inconsistent among papers, but some reports stated that very high levels of serum HDL-C were also associated with increased risk of kidney dysfunction^{17, 18)}. There are several differences in genetic or acquired factors affecting HDL between Japan and Western countries²¹⁾; the association between HDL-C levels and adverse kidney outcomes could be modified by race or severity of CKD stage^{22, 23)}. In a Japanese study, Kawachi et al. reported that a low serum HDL-C level was a significant predictor of CKD progression, especially in females aged ＜70 years; however, their sample size was relatively small²⁴⁾. At present, little evidence is available regarding the relationship between serum HDL-C levels and the progression of kidney dysfunction in the Japanese population.

Aim

We aimed to investigate the association between serum HDL-C levels and adverse kidney outcomes in the general Japanese population.

Materials and Methods

Study Population

Data were collected from the Japan Specific Health Checkups (J-SHC) study. This study is based on the specific nationwide health check-up and guidance system known as Tokutei-Kenshin, which was initiated in April 2008. Study participants were Japanese citizens aged 40–74 years and underwent annual health checkups at pre-assigned clinics or hospitals between 2008–2014. The data were anonymously provided and included in this analysis. This study was conducted in accordance with the Private Information Protection Law and Ethical Guidelines for Epidemiology Research published by the Japanese Ministry of Health, Labour, and Welfare in 2008. The details of this study were described previously²⁵⁾. The number of regions participating in the J-SHC study has increased to include data from 27 prefectures: Hokkaido, Yamagata, Miyagi, Fukushima, Tochigi, Ibaraki, Chiba, Saitama, Tokyo, Kanagawa, Niigata, Ishikawa, Fukui, Nagano, Gifu, Osaka, Hyogo, Okayama, Tokushima, Kochi, Fukuoka, Nagasaki, Saga, Kumamoto, Oita, Miyazaki, and Okinawa. The exclusion criteria were as follows: only one visit during the study period, missing serum creatinine levels and HDL-C levels at baseline, and serum creatinine levels of ＜0.3 mg/dL or ≥ 8 mg/dL.

Measurements

All participants completed a self-administered questionnaire to document their medical history, current medications, smoking habits, and alcohol consumption. Height and weight were measured, and body mass index (BMI) was calculated. Following the specific health checkup program, blood pressure measurements and urine and blood sampling were performed at each participant’s local medical institute after a 10-hour overnight fast. The urine dipstick test for proteinuria included five categories of results: (−), (±), (1+), (2+), and (3+); proteinuria was defined as 1+ or higher. The exposure of interest was serum HDL-C levels, which was categorized into four groups: ＜40 mg/dL, 40–59 mg/dL, 60–79 mg/dL, and ≥ 80 mg/dL. The outcomes were 40% decline in estimated glomerular filtration rate (eGFR) from baseline and trajectory of eGFR over the study period. The eGFR was calculated using the following equation by the Japanese Society of Nephrology²⁶⁾:

eGFR (mL/min/1.73 m²)=194×serum creatinine (mg/dL)^−1.094×age (years)^−0.287 (×0.739 for females).

Statistical Analysis

Continuous and categorical variables were presented as median and interquartile range (IQR) and as total number and percentage, respectively. The Jonckheere–Terpstra test was used to determine trends across the groups for continuous variables, while the Cochran–Armitage test was used for categorical variables. Associations between the category of HDL-C levels and the incidence of 40% decline in eGFR were analyzed using Kaplan–Meier (log-rank test) and multivariable Cox regression analyses. Furthermore, model results were evaluated using three sets of potential confounders at baseline: Model 1, unadjusted model; Model 2, model adjusted for age and sex; and Model 3, model adjusted for age, sex, BMI, systolic blood pressure, current smoking habits, current drinking status, haemoglobin A1c (HbA1c), triglycerides (TG), low-density lipoprotein (LDL)-C, uric acid, eGFR, proteinuria, medication for hypertension, diabetes and dyslipidemia, and history of heart disease and stroke. Under the assumption of missing at random, missing covariates were imputed using multiple imputation based on predictive mean matching, from which 20 imputed datasets were created. The results from these datasets were combined by Rubin’s rules²⁷⁾. Additionally, regarding serum HDL-C level as a continuous variable, the fully adjusted Cox regression analysis using restricted cubic spline with four knots located at quantiles of 0.05, 0.35, 0.65, and 0.95 was also performed. We performed subgroup analyses to examine effect modifications on the association between serum HDL-C levels and the incidence of 40% decline in eGFR according to age, sex, baseline eGFR, proteinuria, diabetes, alcohol drinking status, and medication for dyslipidemia. To further explore the sex difference in serum HDL-C levels and their effect on the kidney outcome according to sex, we also performed restricted cubic spline anaysis by sex and added the fully adjusted Cox regression analysis with HDL-C categories of ＜30 mg/dL, 30–39 mg/dL, 40–49 mg/dL, 50–59 mg/dL, 60–69 mg/dL and ≥ 70 mg/dL as an exploratory analysis. The trajectories of eGFR (mL/min/1.73 m²/year) over the study period among the categories of HDL-C levels were compared using a linear mixed-effects model with a random intercept and slope. Furthermore, the same variables used in the Cox analysis (Model 3) were selected as time-varying covariates for the mixed-effects model. To investigate the association between the changes in variables (Δ), we calculated the annual Δ for each variable using the following equation: (value at the final visit – value at the first visit) / follow-up period (year). We then performed regression analyses using ΔHDL-C as an explanatory variable and ΔeGFR as a response variable. To confirm the robustness of the main results, we conducted several sensitivity analyses and a complete-case analysis in the fully adjusted Cox regression model, using serum HDL-C level as a categorical variable. To evaluate a milder kidney dysfunction than with 40% decline in eGFR, we also conducted Cox regression analyses with the outcome set to 30% decline in eGFR.

Statistical significance was defined as a p-value ＜0.05. All statistical analyses were performed using R software version 4.1.2 (R Foundation, Vienna, Austria). Multiple imputation was conducted using the “Hmisc” package.

Ethics Approval

All procedures involving human participants were in accordance with the ethical standards of the institutional and/or national research committee at which the studies were conducted (Fukushima Medical University; IRB Approval Number #1485, #2771) and with the 1964 Helsinki Declaration and its later amendments. This study was conducted in accordance with the Ethical Guidelines for Medical and Health Research Involving Human Subjects enacted by the Ministry of Health, Labor, and Welfare of Japan.

Results

Baseline Characteristics of Study Participants

Of 933,488 participants, 768,495 were included in the analysis (Fig.1); the proportion of male participants in this study was 41.8%. At baseline, the median (IQR) of the main demographic and laboratory parameters was as follows: age, 65 (IQR: 59–69) years; BMI, 22.9 (IQR: 21.0–25.1) kg/m²; TG, 101 (IQR: 73–145) mg/dL; LDL-C, 124 (IQR: 105–145) mg/dL; HDL-C, 60 (IQR: 50–72) mg/dL; and eGFR, 74.7 (IQR: 64.5–85.7) mL/min/1.73 m². The dipstick proteinuria was positive in 5% of participants, and the proportion of participants taking medication for dyslipidemia was 15.6%. Baseline characteristics of participants according to the HDL-C category are shown in Table 1. Lower HDL-C levels were associated with older age, male sex, higher BMI, blood pressure, plasma glucose, HbA1c, TG levels, uric acid levels, creatinine levels, positive dipstick proteinuria, current smoking, non-drinker, use of medications for hypertension and diabetes, and a history of stroke and heart disease.

Fig.1.

Flowchart diagram showing study outline

Table 1.Baseline characteristics according to the quantile of high-density lipoprotein (HDL)-cholesterol levels

	Baseline serum HDL-C levels, md/dL				p for trend	Missing (%)
	＜40	40-59	60-79	≥ 80
N	40,446	329,916	293,261	104,872
Age, years	65 [59, 69]	65 [60, 69]	64 [59, 69]	64 [58, 68]	＜0.001	0
Male, n (%)	29,573 (73.1)	169,517 (51.4)	96,240 (32.8)	25,621 (24.4)	＜0.001	0
Body mass index, kg/m2	24.6 [22.8, 26.7]	23.8 [21.9, 25.9]	22.4 [20.6, 24.5]	21.1 [19.4, 23.1]	＜0.001	0.7
Systolic blood pressure, mmHg	130 [120, 140]	130 [119, 140]	128 [116, 140]	126 [114, 138]	＜0.001	3.9
Diastolic blood pressure, mmHg	78 [70, 84]	78 [70, 84]	76 [70, 82]	75 [68, 82]	＜0.001	3.9
Fasting plasma glucose, mg/dL	97 [90, 108]	95 [88, 104]	93 [87, 100]	92 [86, 99]	＜0.001	29.2
Haemoglobin A1c, %	5.3 [5.1, 5.7]	5.3 [5.0, 5.6]	5.2 [5.0, 5.5]	5.1 [4.9, 5.4]	＜0.001	2
Triglycerides, mg/dL	181 [129, 261]	123 [90, 170]	89 [67, 119]	71 [55, 93]	＜0.001	0
LDL-C, mg/dL	117 [97, 137]	128 [108, 148]	124 [105, 145]	117 [98, 137]	＜0.001	0
Uric acid, mg/dL	5.9 [4.9, 6.8]	5.3 [4.5, 6.3]	4.8 [4.1, 5.8]	4.6 [3.9, 5.5]	＜0.001	15.5
Creatinine, mg/dL	0.80 [0.70, 0.90]	0.70 [0.60, 0.82]	0.70 [0.60, 0.80]	0.61 [0.60, 0.70]	＜0.001	0
eGFR, mL/min/1.73 m2	73.2 [63.1, 83.7]	74.1 [64.2, 85.0]	75.0 [64.9, 86.3]	75.7 [65.4, 87.3]	＜0.001	0
Proteinuria, n (%)	3,535 (8.8)	19,247 (5.8)	11,850 (4.0)	3,920 (3.7)	＜0.001	0.2
Current smoker, n (%)	11,638 (29.4)	57,815 (17.9)	33,984 (11.8)	10,422 (10.1)	＜0.001	1.8
Current drinker, n (%)	15,184 (42.2)	133,048 (45.2)	122,359 (46.7)	49,107 (52.2)	＜0.001	10.7
Medication for hypertension, n (%)	14,752 (37.1)	106,402 (32.8)	77,267 (26.8)	22,272 (21.5)	＜0.001	1.6
Medication for diabetes, n (%)	3,582 (9.0)	20,523 (6.3)	11,766 (4.1)	3,241 (3.1)	＜0.001	1.7
Medication for dyslipidemia, n (%)	5,341 (13.5)	51,085 (15.7)	46,160 (16.0)	15,541 (15.0)	0.0013	1.6
History of stroke, n (%)	2,002 (5.5)	12,011 (4.1)	8,057 (3.1)	2,291 (2.4)	＜0.001	10
History of heart disease, n (%)	3,181 (8.8)	19,997 (6.7)	13,952 (5.3)	4,226 (4.5)	＜0.001	10

Abbreviations: eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol.

Association of Serum HDL-C Levels with 40% Decline in eGFR

During a median follow-up period of 34.6 (IQR: 14.8–48.4) months, 6,249 participants experienced 40% decline in eGFR. There was a significant difference in the cumulative incidence between the groups (log-rank p＜0.001; Fig.2). In the unadjusted Cox regression analysis, a lower category with HDL-C levels of ＜40 mg/dL was associated with a higher incidence of 40% decline in eGFR, as compared to the reference category with HDL-C levels of 40–59 mg/dL (hazard ratio [HR], 1.56; 95% confidence interval [CI], 1.42–1.71) (Table 2). Conversely, higher categories with HDL-C levels of 60–79 mg/dL (HR, 0.80; 95% CI, 0.76–0.85) and ≥ 80 mg/dL (HR, 0.72; 95% CI, 0.66–0.78) were associated with decreased risk of 40% decline in eGFR (Table 2). The fully adjusted Cox regression analysis showed a consistent result. In the fully adjusted Cox regression model with restricted cubic spline, HDL-C levels of less than approximately 60 mg/dL were associated with an increased risk of kidney outcomes, but high HDL-C levels were not associated with kidney outcomes (Fig.3). In the subgroup analyses, baseline eGFR and proteinuria modified the association of serum HDL-C levels with kidney outcomes (Table 3). The participants with proteinuria and/or decreased baseline eGFR (participants with CKD) had a stronger impact of decreased HDL-C on the new onset of decline in eGFR than those without CKD.

Fig.2. Kaplan–Meier curves for 40% decline in eGFR across the HDL-C category

Abbreviations: eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol.

Table 2.Hazard ratios (HRs) and 95% confidence intervals (CIs) for 40% decline in estimated glomerular filtration rate (eGFR) and serum HDL-C levels

Model	Hazard ratio (95% confidence interval)
	HDL-C, mg/dL ＜40	40–59	60–79	≥ 80
1 a	1.56 (1.42–1.71)	1.00 (Ref)	0.80 (0.76–0.85)	0.72 (0.66–0.78)
2 b	1.52 (1.38–1.67)	1.00 (Ref)	0.82 (0.78–0.87)	0.75 (0.69–0.82)
3 c	1.26 (1.14–1.39)	1.00 (Ref)	0.91 (0.86–0.96)	0.86 (0.78–0.93)

^a Model 1, crude model.

^b Model 2, adjusted for age and sex.

^c Model 3, Model 1 + adjusted for BMI, systolic blood pressure, current smoking status, current drinking status, HbA1c, TG, LDL-C, uric acid, eGFR, proteinuria, medication for hypertension, diabetes and dyslipidemia, and history of heart disease and stroke.

Abbreviations: eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; BMI, body mass index; HbA1c, haemoglobin A1c; TG, triglycerides; LDL-C, low-density lipoprotein cholesterol.

Fig.3. HRs and 95% CIs for the incidence of 40% decline in eGFR with reference to median HDL-C levels of 60 mg/dL with histogram

Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol.

Table 3.Subgroup analysis of the association between HDL-C category and 40% decline in eGFR

	Hazard ratio (95% confidence interval)				p for interaction
	HDL-C, mg/dL ＜40	40–59	60–79	≥ 80
Age					0.11
≥ 65 yr	1.23 (1.03–1.46)	1.00 (Ref)	0.89 (0.81–0.99)	0.83 (0.70–0.98)
＜65 yr	1.32 (1.12–1.56)	1.00 (Ref)	0.91 (0.83–1.00)	0.81 (0.70–0.94)
Sex					0.11
Male	1.32 (1.16–1.51)	1.00 (Ref)	0.95 (0.86–1.05)	0.93 (0.79–1.10)
Female	1.26 (1.04–1.54)	1.00 (Ref)	0.86 (0.79–0.94)	0.80 (0.70–0.90)
eGFR, mL/min/1.73 m2					0.039
＜60	1.23 (1.01–1.49)	1.00 (Ref)	0.97 (0.84–1.13)	0.90 (0.71–1.15)
≥ 60	1.12 (0.98–1.29)	1.00 (Ref)	0.88 (0.82–0.95)	0.79 (0.71–0.89)
Proteinuria					0.002
Present	1.45 (1.18–1.79)	1.00 (Ref)	1.03 (0.86–1.22)	1.21 (0.93–1.58)
Absent	1.16 (1.02–1.32)	1.00 (Ref)	0.86 (0.80–0.92)	0.75 (0.67–0.84)
Diabetes					0.72
Present	1.30 (1.06–1.60)	1.00 (Ref)	0.88 (0.75–1.03)	1.00 (0.77–1.30)
Absent	1.27 (1.11–1.45)	1.00 (Ref)	0.89 (0.83–0.96)	0.77 (0.69–0.86)
Drinking					0.40
Present	1.16 (0.97–1.38)	1.00 (Ref)	0.88 (0.80–0.98)	0.85 (0.73–0.98)
Absent	1.30 (1.12–1.50)	1.00 (Ref)	0.90 (0.82–0.98)	0.81 (0.70–0.93)
Medication for dyslipidemia					0.24
Present	1.21 (0.92–1.58)	1.00 (Ref)	0.86 (0.73–1.01)	0.80 (0.62–1.03)
Absent	1.27 (1.13–1.44)	1.00 (Ref)	0.88 (0.82–0.95)	0.79 (0.71–0.88)

Abbreviations: eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein.

Sex Difference in Serum HDL-C Levels

The distributions of serum HDL-C levels differed dependeing on sex (Supplementary Fig.1). The median (IQR) HDL-C levels were 55 (46–66) mg/dL for men and 64 (54–75) mg/dL for women, respectively. Restricted cubic spline curves depicting the nonlinear association between continuous HDL-C level and the kidney outcome by sex were shown in Fig.4. There was a significant effect modification of sex (p for interaction =0.031), and the threshold of low serum HDL-C level at which the risk increases was lower for men than for women. This difference in the threshold of low serum HDL-C level was consistent with the result of Cox regression analysis with HDL-C categories of ＜30 mg/dL, 30–39 mg/dL, 40–49 mg/dL, 50–59 mg/dL, 60–69 mg/dL and ≥ 70 mg/dL (Supplementary Table 1). Based on the median values for each sex, the reference HDL-C category was set to 50–59 mg/dL for men and 60–69 mg/dL for women. Compared with the reference HDL-C category of 50–59 mg/dL, the lower HDL-C categories of ＜30 mg/dL and 30–39 mg/dL were associated with an increased risk of the kidney outcome in men. In women, the lower categories of ＜30 mg/dL, 30–39 mg/dL and 40–49 mg/dL were associated with an increased risk of the kidney outcome compared with the reference HDL-C category of 60–69 mg/dL.

Supplementary Fig.1. Histograms showing the distributions of serum HDL-C levels by sex

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

Fig.4. Log hazards and 95% CIs for the incidence of 40% decline in eGFR by sex

Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol.

Supplementary Table 1.Hazard ratios and 95% confidence intervals for 40% decline in eGFR and serum HDL-C levels by sex

Sex	Hazard ratio (95% confidence interval)
	HDL-C, mg/dL ＜30	30–39	40–49	50–59	60–69	≥ 70
Male	3.04 (2.16–4.27)	1.30 (1.11–1.51)	1.09 (0.98–1.23)	1.00 (Ref)	0.96 (0.84–1.09)	1.01 (0.89–1.15)
Female	3.18 (2.25–4.50)	1.36 (1.15–1.61)	1.15 (1.00–1.31)	1.05 (0.92–1.22)	1.00 (Ref)	1.06 (0.92–1.22)

Abbreviations: eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein.

Trajectory of eGFR among the HDL-C Categories

After fully adjusting for clinically relevant factors, the decline rates of eGFR in each category are presented in Table 4. The p-value for the interaction term between the time and HDL-C category was ＜0.001, and eGFR decline rates incrementally increased as the categories of HDL-C decreased (p for trend: ＜0.001).

Table 4. The eGFR decline rate (mL/min/1.73 m²/year) among the categories of HDL-C levels estimated using a linear mixed- effects model with a random intercept and a random slope

The model was adjusted for age, sex, BMI, systolic blood pressure, current smoking status, current drinking status, haemoglobin A1c, TG, LDL-C, uric acid, eGFR, proteinuria, medication for hypertension, diabetes and dyslipidemia, and history of heart disease and stroke.
eGFR decline rate (mL/min/1.73 m2/year) (95% CI)				p for trend
HDL-C, mg/dL ＜40	40–59	60–79	≥ 80
−0.624	−0.598	−0.584	−0.558	＜0.001
(−0.660 to −0.588)	(−0.609 to −0.587)	(−0.613 to −0.556)	(−0.582 to −0.534)

Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; BMI, body mass index; HbA1c, haemoglobin A1c; TG, triglycerides; LDL-C, low-density lipoprotein cholesterol.

Association between ΔHDL-C and ΔeGFR

The medians (IQRs) of ΔHDL-C and ΔeGFR were 0.00 (−2.07 to 2.09) mg/dL/year and −0.37 (−3.00 to 0.76) mL/min/1.73m²/year, respectively. In the crude regression model, ΔHDL-C was not associated with ΔeGFR (β=0.0018, p=0.22). When adjusting for the baseline covariates used in the main analysis, ΔHDL-C was positively associated with ΔeGFR, but the effect was marginal (β=0.012, p＜0.001). When further adjusting for the Δ values of covariates, we obtained a similar result (β=0.0080, p＜0.001).

Sensitivity Analyses

The complete-case analysis in the fully adjusted Cox regression model showed consistent results with the main analysis (Supplementary Table 2). The Cox regression analyses with the outcome set to 30% decline in eGFR also showed similar results (Supplementary Table 3).

Supplementary Table 2.Hazard ratios and 95% confidence intervals for 40% decline in eGFR and serum HDL-C levels calculated by a complete-case analysis

Hazard ratio (95% confidence interval)
HDL-C, mg/dL ＜30	40–59	60–79	≥ 80
1.24 (1.11–1.39)	1.00 (Ref)	0.89 (0.83–0.95)	0.82 (0.74–0.91)

Adjusted for sex, age, BMI, systolic blood pressure, current smoking status, current drinking status, HbA1c, TG, LDL-C, uric acid, eGFR, proteinuria, medication for hypertension, diabetes and dyslipidemia, and history of heart disease and stroke.

Abbreviations: eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; BMI, body mass index; HbA1c, hemoglobin A1c; TG, triglyceride; LDL-C, low-density lipoprotein cholesterol.

Supplementary Table 3.Hazard ratios and 95% confidence intervals for 30% decline in eGFR and serum HDL-C levels

Model	Hazard ratio (95% confidence interval)
	HDL-C, mg/dL ＜40	40–59	60–79	≥ 80
1 a	1.32 (1.24–1.39)	1.00 (Ref)	0.91 (0.89–0.94)	0.89 (0.86–0.93)
2 b	1.39 (1.31–1.47)	1.00 (Ref)	0.88 (0.86–0.91)	0.86 (0.82–0.90)
3 c	1.19 (1.13–1.27)	1.00 (Ref)	0.94 (0.91–0.97)	0.93 (0.89–0.98)

^a Model 1, crude model.

^b Model 2, adjusted for age and sex.

^c Model 3, Model 1 + adjusted for BMI, systolic blood pressure, current smoking status, current drinking status, HbA1c, TG, LDL-C, uric acid, eGFR, proteinuria, medication for hypertension, diabetes and dyslipidemia, and history of heart disease and stroke.

Abbreviations: eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; BMI, body mass index; HbA1c, hemoglobin A1c; TG, triglyceride; LDL-C, low-density lipoprotein cholesterol.

Discussion

In this longitudinal nationwide cohort study, we showed a significant association between serum HDL-C levels and kidney function decline in the Japanese population. As hard endpoints—such as initiation of dialysis—are rarely experienced in this study population with low prevalence of advanced CKD, we used 40% decline in eGFR and slope of eGFR as kidney outcomes²⁸⁾. In the Cox regression model, low HDL-C levels were associated with 40% decline in eGFR, while high HDL-C levels were not associated with the increased risk of the outcome. Additionally, the lower category of HDL-C level was associated with a faster eGFR decline rate in the mixed-effects model, although the difference of eGFR decline rate among the HDL-C categories was relatively small. Regarding the association between ΔHDL-C and ΔeGFR, the multivariable regression analysis yielded the significant but marginal association.

Previous studies demonstrated that low serum HDL-C levels were associated with adverse kidney outcomes, which is consistent with our results^{16-20, 24)}. However, there was no association between serum HDL-C levels and adverse kidney outcomes in some studies^{29, 30)}. This inconsistency may be partially due to a variety of baseline kidney function parameters among the studies; the composition and function of HDL particles, as well as HDL-C concentration, may be different between various stages of CKD²³⁾. In our study, median eGFR at baseline was 74.7 (IQR: 64.5–85.7) mL/min/1.73 m², similar to that reported by Bowe et al.¹⁷⁾. Although we showed that participants with eGFR ＜60 mL/min/1.73 m² had the stronger association of low serum HDL-C levels with decline in eGFR compared to those without in a subgroup analysis, our study included few participants with severe kidney dysfunction. Therefore, it remains unknown whether our results can be applied to those with advanced CKD.

It has been controversial whether low HDL-C concentration is causally related to poor kidney outcomes. As an approach to this issue, it may be useful to meta-analyze the data from trials that targeted HDL-C levels—including niacin and cholesteryl ester transfer protein (CETP) inhibitors—and to investigate whether the intervention had an influence on kidney function³¹⁾. Although the availability of this information is limited, the Randomized Evaluation of the Effects of Anacetrapib through Lipid Modification (REVEAL) trial examined the effect of CETP inhibitor anacetrapib on coronary heart disease. In this trial, anacetrapib increased the incidence of kidney dysfunction (eGFR ＜60 mL/min/1.73 m²) as compared to placebo (11.5% vs. 10.6%, p=0.04)³²⁾. This result may support the argument that there is no causal relationship between lower HDL-C levels and faster kidney function decline. Recently, Mendelian randomization studies are conducted as another approach to investigate the causal association of HDL-C levels with kidney function decline^{33, 34)}. Lanktree et al. reported that a 17-mg/dL genetically higher HDL-C level was associated with an 0.8% higher eGFR with statistical significance, although the impact of HDL-C levels on kidney function was small³⁴⁾. In our study as well, the association between ΔHDL-C and ΔeGFR was marginal. Our results cannot conclude the causal association, but we clearly showed the association between low serum HDL-C levels at baseline and subsequent decline in eGFR. We believe that low serum HDL-C levels can be used as a marker of kidney function decline in patients with normal or mild reduced kidney function.

In this study, the higher category of HDL-C level was associated with a lower risk of 40% eGFR decline or slower eGFR decline rate. In the restricted cubic spline analysis, although the protective effect of high serum HDL-C levels on kidney function decline disappeared, a high serum HDL-C level was not associated with increased risk of kidney function decline. Conversely, some recent studies reported that the association between serum HDL-C levels and kidney outcomes were U-shaped^{17, 18)}. Recently, the importance of HDL function has been highlighted as a predictor of atherosclerotic disease⁶⁾. The major role of HDL is RCT, in which excess cellular cholesterol is transported from peripheral tissues to small HDL particles, esterified by LCAT to generate larger HDL particles, and transported to the liver via apoB-containing lipoprotein by cholesteryl ester transfer protein (CETP)⁶⁾. In cellular cholesterol efflux in RCT, membrane transporters—such as members of the ABC transporter family, ABCG1 and ABCA1, and scavenger receptor BI (SR-BI)—are involved⁶⁾. Although HDL has various protective effects against atherosclerosis, high serum HDL-C levels are not always protective⁶⁾. For example, patients with a loss-of-function mutation in SR-BI have a significant increase in serum HDL-C levels but increased cardiovascular risk³⁵⁾. Furthermore, in the clinical trials, agents that elevate serum HDL-C levels—such as niacin and torcetrapib, a CETP inhibitor—failed to demonstrate their protective effects against cardiovascular outcomes^{36, 37)}. Genetic or acquired alterations in the components involving RCT can affect both HDL function and HDL-C levels. In Japan, some types of dysfunctional mutations of CETP are more frequently found in individuals with high HDL-C levels, as compared to Western countries²¹⁾. Additionally, a recent Danish study reported that genetic deficiency of CETP was associated with a lower absolute risk of CVD and mortality, although this study did not evaluate kidney outcomes³⁸⁾. These facts possibly contributed to the inconsistent results between our study and the others.

Subgroup analyses showed that not only baseline eGFR, but also proteinuria, modified the effect of serum HDL-C levels on kidney outcomes (p for interaction: 0.002). The participants with proteinuria and/or decreased baseline eGFR (participants with CKD) had a stronger impact of decreased HDL-C on the new onset of decreased eGFR than those without CKD. However, the mechanism by which CKD affected the association of HDL-C on kidney outcomes remains unclear in this study. The alterations in HDL composition, such as the enrichment in serum amyloid A and apoC-III and depletion in apoA-I and apoA-II, have been reported in patients with CKD¹¹⁾. These changes can impair HDL function, including cholesterol efflux from macrophages and anti-inflammatory and anti-oxidative abilities¹¹⁾. Regarding the effect of proteinuria on HDL, it was reported to increase oxidized lipid concentrations in HDL in a rat model³⁹⁾. Furthermore, accumulation of oxidized phospholipids in HDL particles can impair the anti-inflammatory and anti-oxidative effects of HDL⁴⁰⁾. These facts may explain the interaction between proteinuria and the association of HDL-C levels with kidney outcomes. Although our results may not be applied to patients with advanced CKD as described above, low HDL-C level is supposed to have a stronger influence on kidney function decline in those with mild-to-moderate GFR decline and/or proteinuria compared with those without. Therefore, serum HDL-C level should be carefully monitored especially in patients with mild-to-moderate CKD.

To further explore sex difference in the effect of serum HDL-C levels on kidney outcome, we also examine sex difference in serum HDL-C level and performed additional analyses by sex. As reported previously⁴¹⁾, serum HDL-C level was higher in women than in men on average. In accordance with this difference in serum HDL-C level, these analyses indicated that the threshold of low serum HDL-C level at which the risk of kidney outcome increases was different depending on sex. This finding is similar to the result from the Japan Lipid Intervention Trial, which indicated that the risk of coronary heart disease significantly increased at HDL-C levels below 45 mg/dL in men and below 60 mg/dL in women⁴²⁾. Our finding is novel in that the lower limit of serum HDL-C levels may be different depending on sex even in the association between HDL-C levels and kidney outcome. Although guidelines for prevention of atherosclerotic cardiovascular diseases published by the Japan Atherosclerosis Society recommend the target serum HDL-C values of 40 mg/dL or higher regardless of sex at present⁴³⁾, it might be advisable to set the target value according to sex.

This study has some limitations. First, we cannot identify a causal relationship between serum HDL-C levels and adverse kidney outcomes in this observational study. Furthermore, it did not address the question whether intervention to normalize HDL-C levels had protective effects on kidney function decline. Second, there may be unmeasured confounding factors, such as inflammation or metabolic markers. Third, this study mainly included participants with relatively normal kidney function; therefore, our results may not be applied to those with moderate to severe CKD. Fourth, information on underlying kidney disease were not available in this study. This study can potentially include patients with primary glomerular disease such as glomerulonephritis. Kidney outcomes can be influenced by the underlying diseases. Finally, the enrollment and follow-up in this study were conducted before sodium–glucose cotransporter 2 (SGLT2) inhibitors were launched in Japan. SGLT2 inhibitors have not only renoprotective effects but also multifaceted effects on lipid metabolism. It was reported that empagliflozin significantly improved HDL-C levels^{44, 45)}. Thus, the impact of HDL-C on kidney outcomes may differ depending on the use of SGLT2 inhibitors. However, regardless of these limitations, we firmly believe that this study provides crucial insights into the relationship between serum HDL-C level and kidney function decline.

Conclusion

Low serum HDL-C levels were associated with kidney function decline, while high serum HDL-C levels were not associated with increased risk among the general Japanese population. Although it remains unclear whether normalizing serum HDL-C levels are protective against kidney disease progression, those who have low HDL-C levels should be managed carefully.

Acknowledgements

The authors acknowledge the contributions of the staff members who collected the data and instructed participants with metabolic syndrome at screening centers in the following regions: Hokkaido, Yamagata, Miyagi, Fukushima, Tochigi, Ibaraki, Chiba, Saitama, Tokyo, Kanagawa, Niigata, Ishikawa, Fukui, Nagano, Gifu, Osaka, Hyogo, Okayama, Tokushima, Kochi, Fukuoka, Nagasaki, Saga, Kumamoto, Oita, Miyazaki, and Okinawa.

We would like to thank Editage (www.editage.com) for English language editing.

Notice of Grant Support

This study was supported by the Health and Labor Sciences Research Grants for Research on Design of the Comprehensive Health Care System for Chronic Kidney Disease (CKD) Based on the Individual Risk Assessment by Specific Health Checkup from the Ministry of Health, Labour, and Welfare of Japan and a Grant-in-Aid for Research on Advanced Chronic Kidney Disease (REACH-J), Practical Research Project for Renal Disease from the Japan Agency for Medical Research and Development (AMED), and JSPS KAKENHI Grant Number JP18K11131.

Conflict of Interest

The authors have no conflicts of interest to declare.

Author Contributions

Research idea and study design: TaK, ME, TU, HT, RF, FF, MN, MM, KS, and KT; data acquisition: HY, KI, SF, TsK, TM, KY, IN, MKa, YS, MKo, KA, TW, and KT; data analysis/interpretation: TaK, ME, and KT; statistical analysis: TaK; supervision or mentorship: ME and KT. Each author contributed important intellectual content during manuscript drafting or revision and accepted accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work were appropriately investigated and resolved.

References

• 1)  Rader DJ: Molecular regulation of HDL metabolism and function: implications for novel therapies. J Clin Invest, 2006; 116: 3090-3100
• 2)  Okamura T, Hayakawa T, Kadowaki T, Kita Y, Okayama A and Ueshima H: The inverse relationship between serum high-density lipoprotein cholesterol level and all-cause mortality in a 9.6-year follow-up study in the Japanese general population. Atherosclerosis, 2006; 184: 143-150
• 3)  Ko DT, Alter DA, Guo H, Koh M, Lau G, Austin PC, Booth GL, Hogg W, Jackevicius CA, Lee DS, Wijeysundera HC, Wilkins JT and Tu JV: High-Density Lipoprotein Cholesterol and Cause-Specific Mortality in Individuals Without Previous Cardiovascular Conditions: The CANHEART Study. J Am Coll Cardiol, 2016; 68: 2073-2083
• 4)  Madsen CM, Varbo A and Nordestgaard BG: Extreme high high-density lipoprotein cholesterol is paradoxically associated with high mortality in men and women: two prospective cohort studies. Eur Heart J, 2017; 38: 2478-2486
• 5)  Emerging Risk Factors C, Di Angelantonio E, Sarwar N, Perry P, Kaptoge S, Ray KK, Thompson A, Wood AM, Lewington S, Sattar N, Packard CJ, Collins R, Thompson SG and Danesh J: Major lipids, apolipoproteins, and risk of vascular disease. JAMA, 2009; 302: 1993-2000
• 6)  Rhainds D and Tardif JC: From HDL-cholesterol to HDL-function: cholesterol efflux capacity determinants. Curr Opin Lipidol, 2019; 30: 101-107
• 7)  Ouimet M, Barrett TJ and Fisher EA: HDL and Reverse Cholesterol Transport. Circ Res, 2019; 124: 1505-1518
• 8)  Hirata A, Sugiyama D, Watanabe M, Tamakoshi A, Iso H, Kotani K, Kiyama M, Yamada M, Ishikawa S, Murakami Y, Miura K, Ueshima H and Okamura T: Association of extremely high levels of high-density lipoprotein cholesterol with cardiovascular mortality in a pooled analysis of 9 cohort studies including 43,407 individuals: The EPOCH-JAPAN study. J Clin Lipidol, 2018; 12: 674-684.e675
• 9)  Global, regional, and national burden of chronic kidney disease, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet, 2020; 395: 709-733
• 10)  Shoji T, Abe T, Matsuo H, Egusa G, Yamasaki Y, Kashihara N, Shirai K, Kashiwagi A, Committee of R and Peripheral Arteries JAS: Chronic kidney disease, dyslipidemia, and atherosclerosis. J Atheroscler Thromb, 2012; 19: 299-315
• 11)  Strazzella A, Ossoli A and Calabresi L: High-Density Lipoproteins and the Kidney. Cells, 2021; 10:
• 12)  Hager MR, Narla AD and Tannock LR: Dyslipidemia in patients with chronic kidney disease. Rev Endocr Metab Disord, 2017; 18: 29-40
• 13)  Kosugi T, Eriguchi M, Yoshida H, Tasaki H, Fukata F, Nishimoto M, Matsui M, Samejima KI, Iseki K, Fujimoto S, Konta T, Moriyama T, Yamagata K, Narita I, Kasahara M, Shibagaki Y, Kondo M, Asahi K, Watanabe T and Tsuruya K: Association between chronic kidney disease and new-onset dyslipidemia: The Japan Specific Health Checkups (J-SHC) study. Atherosclerosis, 2021; 332: 24-32
• 14)  Baragetti A, Ossoli A, Strazzella A, Simonelli S, Baragetti I, Grigore L, Pellegatta F, Catapano AL, Norata GD and Calabresi L: Low Plasma Lecithin: Cholesterol Acyltransferase (LCAT) Concentration Predicts Chronic Kidney Disease. J Clin Med, 2020; 9:
• 15)  Vaziri ND, Navab M and Fogelman AM: HDL metabolism and activity in chronic kidney disease. Nat Rev Nephrol, 2010; 6: 287-296
• 16)  Baragetti A, Norata GD, Sarcina C, Rastelli F, Grigore L, Garlaschelli K, Uboldi P, Baragetti I, Pozzi C and Catapano AL: High density lipoprotein cholesterol levels are an independent predictor of the progression of chronic kidney disease. J Intern Med, 2013; 274: 252-262
• 17)  Bowe B, Xie Y, Xian H, Balasubramanian S and Al-Aly Z: Low levels of high-density lipoprotein cholesterol increase the risk of incident kidney disease and its progression. Kidney Int, 2016; 89: 886-896
• 18)  Nam KH, Chang TI, Joo YS, Kim J, Lee S, Lee C, Yun HR, Park JT, Yoo TH, Sung SA, Lee KB, Oh KH, Kim SW, Lee J, Kang SW, Choi KH, Ahn C and Han SH: Association Between Serum High-Density Lipoprotein Cholesterol Levels and Progression of Chronic Kidney Disease: Results From the KNOW-CKD. J Am Heart Assoc, 2019; 8: e011162
• 19)  Muntner P, Coresh J, Smith JC, Eckfeldt J and Klag MJ: Plasma lipids and risk of developing renal dysfunction: the atherosclerosis risk in communities study. Kidney Int, 2000; 58: 293-301
• 20)  Schaeffner ES, Kurth T, Curhan GC, Glynn RJ, Rexrode KM, Baigent C, Buring JE and Gaziano JM: Cholesterol and the risk of renal dysfunction in apparently healthy men. J Am Soc Nephrol, 2003; 14: 2084-2091
• 21)  Yokoyama S: Unique features of high-density lipoproteins in the Japanese: in population and in genetic factors. Nutrients, 2015; 7: 2359-2381
• 22)  Zakai NA, Minnier J, Safford MM, Koh I, Irvin MR, Fazio S, Cushman M, Howard VJ and Pamir N: Race-Dependent Association of High-Density Lipoprotein Cholesterol Levels With Incident Coronary Artery Disease. J Am Coll Cardiol, 2022; 80: 2104-2115
• 23)  Kronenberg F: HDL in CKD-The Devil Is in the Detail. J Am Soc Nephrol, 2018; 29: 1356-1371
• 24)  Kawachi K, Kataoka H, Manabe S, Mochizuki T and Nitta K: Low HDL cholesterol as a predictor of chronic kidney disease progression: a cross-classification approach and matched cohort analysis. Heart Vessels, 2019; 34: 1440-1455
• 25)  Iseki K, Asahi K, Yamagata K, Fujimoto S, Tsuruya K, Narita I, Konta T, Kasahara M, Shibagaki Y, Yoshida H, Moriyama T, Kondo M, Iseki C and Watanabe T: Mortality risk among screened subjects of the specific health check and guidance program in Japan 2008-2012. Clin Exp Nephrol, 2017; 21: 978-985
• 26)  Matsuo S, Imai E, Horio M, Yasuda Y, Tomita K, Nitta K, Yamagata K, Tomino Y, Yokoyama H and Hishida A: Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis, 2009; 53: 982-992
• 27)  White IR, Royston P and Wood AM: Multiple imputation using chained equations: Issues and guidance for practice. Stat Med, 2011; 30: 377-399
• 28)  Baigent C, Herrington WG, Coresh J, Landray MJ, Levin A, Perkovic V, Pfeffer MA, Rossing P, Walsh M, Wanner C, Wheeler DC, Winkelmayer WC and McMurray JJV: Challenges in conducting clinical trials in nephrology: conclusions from a Kidney Disease-Improving Global Outcomes (KDIGO) Controversies Conference. Kidney Int, 2017; 92: 297-305
• 29)  Chen SC, Hung CC, Kuo MC, Lee JJ, Chiu YW, Chang JM, Hwang SJ and Chen HC: Association of dyslipidemia with renal outcomes in chronic kidney disease. PLoS One, 2013; 8: e55643
• 30)  Rahman M, Yang W, Akkina S, Alper A, Anderson AH, Appel LJ, He J, Raj DS, Schelling J, Strauss L, Teal V and Rader DJ: Relation of serum lipids and lipoproteins with progression of CKD: The CRIC study. Clin J Am Soc Nephrol, 2014; 9: 1190-1198
• 31)  Kronenberg F: High-density lipoprotein cholesterol on a roller coaster: where will the ride end? Kidney Int, 2016; 89: 747-749
• 32)  Bowman L, Hopewell JC, Chen F, Wallendszus K, Stevens W, Collins R, Wiviott SD, Cannon CP, Braunwald E, Sammons E and Landray MJ: Effects of Anacetrapib in Patients with Atherosclerotic Vascular Disease. N Engl J Med, 2017; 377: 1217-1227
• 33)  Coassin S, Friedel S, Köttgen A, Lamina C and Kronenberg F: Is High-Density Lipoprotein Cholesterol Causally Related to Kidney Function? Evidence From Genetic Epidemiological Studies. Arterioscler Thromb Vasc Biol, 2016; 36: 2252-2258
• 34)  Lanktree MB, Thériault S, Walsh M and Paré G: HDL Cholesterol, LDL Cholesterol, and Triglycerides as Risk Factors for CKD: A Mendelian Randomization Study. Am J Kidney Dis, 2018; 71: 166-172
• 35)  Zanoni P, Khetarpal SA, Larach DB, Hancock-Cerutti WF, Millar JS, Cuchel M, DerOhannessian S, Kontush A, Surendran P, Saleheen D, Trompet S, Jukema JW, De Craen A, Deloukas P, Sattar N, Ford I, Packard C, Majumder A, Alam DS, Di Angelantonio E, Abecasis G, Chowdhury R, Erdmann J, Nordestgaard BG, Nielsen SF, Tybjærg-Hansen A, Schmidt RF, Kuulasmaa K, Liu DJ, Perola M, Blankenberg S, Salomaa V, Männistö S, Amouyel P, Arveiler D, Ferrieres J, Müller-Nurasyid M, Ferrario M, Kee F, Willer CJ, Samani N, Schunkert H, Butterworth AS, Howson JM, Peloso GM, Stitziel NO, Danesh J, Kathiresan S and Rader DJ: Rare variant in scavenger receptor BI raises HDL cholesterol and increases risk of coronary heart disease. Science, 2016; 351: 1166-1171
• 36)  Boden WE, Probstfield JL, Anderson T, Chaitman BR, Desvignes-Nickens P, Koprowicz K, McBride R, Teo K and Weintraub W: Niacin in patients with low HDL cholesterol levels receiving intensive statin therapy. N Engl J Med, 2011; 365: 2255-2267
• 37)  Barter PJ, Caulfield M, Eriksson M, Grundy SM, Kastelein JJ, Komajda M, Lopez-Sendon J, Mosca L, Tardif JC, Waters DD, Shear CL, Revkin JH, Buhr KA, Fisher MR, Tall AR and Brewer B: Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med, 2007; 357: 2109-2122
• 38)  Nordestgaard LT, Christoffersen M, Lauridsen BK, Afzal S, Nordestgaard BG, Frikke-Schmidt R and Tybjærg-Hansen A: Long-term Benefits and Harms Associated With Genetic Cholesteryl Ester Transfer Protein Deficiency in the General Population. JAMA Cardiol, 2022; 7: 55-64
• 39)  Newman JW, Kaysen GA, Hammock BD and Shearer GC: Proteinuria increases oxylipid concentrations in VLDL and HDL but not LDL particles in the rat. J Lipid Res, 2007; 48: 1792-1800
• 40)  Rosenson RS, Brewer HB, Jr., Ansell BJ, Barter P, Chapman MJ, Heinecke JW, Kontush A, Tall AR and Webb NR: Dysfunctional HDL and atherosclerotic cardiovascular disease. Nat Rev Cardiol, 2016; 13: 48-60
• 41)  Holven KB and Roeters van Lennep J: Sex differences in lipids: A life course approach. Atherosclerosis, 2023; 384: 117270
• 42)  Sasaki J, Kita T, Mabuchi H, Matsuzaki M, Matsuzawa Y, Nakaya N, Oikawa S, Saito Y, Shimamoto K, Kono S and Itakura H: Gender difference in coronary events in relation to risk factors in Japanese hypercholesterolemic patients treated with low-dose simvastatin. Circ J, 2006; 70: 810-814
• 43)  Okamura T, Tsukamoto K, Arai H, Fujioka Y, Ishigaki Y, Koba S, Ohmura H, Shoji T, Yokote K, Yoshida H, Yoshida M, Deguchi J, Dobashi K, Fujiyoshi A, Hamaguchi H, Hara M, Harada-Shiba M, Hirata T, Iida M, Ikeda Y, Ishibashi S, Kanda H, Kihara S, Kitagawa K, Kodama S, Koseki M, Maezawa Y, Masuda D, Miida T, Miyamoto Y, Nishimura R, Node K, Noguchi M, Ohishi M, Saito I, Sawada S, Sone H, Takemoto M, Wakatsuki A and Yanai H: Japan Atherosclerosis Society (JAS) Guidelines for Prevention of Atherosclerotic Cardiovascular Diseases 2022. J Atheroscler Thromb, 2024; 31: 641-853
• 44)  Kovacs CS, Seshiah V, Swallow R, Jones R, Rattunde H, Woerle HJ and Broedl UC: Empagliflozin improves glycaemic and weight control as add-on therapy to pioglitazone or pioglitazone plus metformin in patients with type 2 diabetes: a 24-week, randomized, placebo-controlled trial. Diabetes Obes Metab, 2014; 16: 147-158
• 45)  Taheri H, Chiti H, Reshadmanesh T, Gohari S, Jalilvand A, Arsang-Jang S, Ismail-Beigi F, Ghanbari S, Dadashi M, Asgari A, Mahjani M, Karbalaee-Hasani A and Ahangar H: Empagliflozin improves high-sensitive cardiac troponin-I and high-density lipoprotein cholesterol in patients with type 2 diabetes mellitus and coronary artery disease: a post-hoc analysis of EMPA-CARD Trial. J Diabetes Metab Disord, 2023; 22: 1723-1730