Original Investigation Impact of the Triglycerides to High-Density Lipoprotein Cholesterol Ratio on the Incidence and Progression of CKD: A Longitudinal Study in a Large Japanese Population Kazuhiko Tsuruya, MD, PhD,1,2,3 Hisako Yoshida, PhD,1 Masaharu Nagata, MD, PhD,2 Takanari Kitazono, MD, PhD,2 Kunitoshi Iseki, MD, PhD,3 Chiho Iseki, PhD,4 Shouichi Fujimoto, MD, PhD,3 Tsuneo Konta, MD, PhD,3 Toshiki Moriyama, MD, PhD,3 Kunihiro Yamagata, MD, PhD,3 Ichiei Narita, MD, PhD,3 Kenjiro Kimura, MD, PhD,3 Masahide Kondo, MD, MSc, PhD,3 Koichi Asahi, MD, PhD,3 Issei Kurahashi, PhD,5 Yasuo Ohashi, PhD,6 and Tsuyoshi Watanabe, MD, PhD3 Background: The impact of the triglycerides to high-density lipoprotein cholesterol (TG:HDL-C) ratio on chronic kidney disease (CKD) is unclear. Study Design: Longitudinal cohort study. Setting & Participants: 124,700 participants aged 39 to 74 years in the Japanese Specific Health Check and Guidance System, including 50,392 men, 74,308 women, 102,900 without CKD, and 21,800 with CKD. Predictor: Quartiles of TG:HDL-C ratio. Outcomes & Measurements: Changes in estimated glomerular filtration rate (eGFR) and urinary protein excretion during the 2-year study period. Incident CKD in participants without CKD, and progression of CKD in participants with CKD. Results: In the entire study population, higher quartile of TG:HDL-C ratio at baseline was significantly associated with greater decline in eGFR and increase in urinary protein excretion during the 2-year study period, even after adjustment for confounding factors. A higher ratio was associated with higher risk of incident CKD in participants without CKD and higher risk of rapid decline in eGFR and increase in urinary protein excretion in participants with CKD. Higher TG:HDL-C ratio was more strongly associated with decline in eGFR (P for interaction 5 0.002) and with incident CKD (P for interaction 5 0.05) in participants with diabetes than without diabetes. Limitations: Short observation period and single measurement of all variables. Conclusions: A higher TG:HDL-C ratio affects the decline in eGFR and incidence and progression of CKD in the Japanese population. Am J Kidney Dis. -(-):---. ª 2015 by the National Kidney Foundation, Inc. INDEX WORDS: Chronic kidney disease (CKD); estimated glomerular filtration rate (eGFR); urinary protein excretion; proteinuria; kidney disease progression; dyslipidemia; triglycerides to high-density lipoprotein cholesterol ratio; lipid nephrotoxicity; diabetes; Japanese population.

A

bnormalities of lipid metabolism have been identified as a possible cause of progression of chronic kidney disease (CKD). However, the precise mechanism of the effect of such abnormalities remains unknown.1,2 In the current era, aggressive reduction of low-density lipoprotein cholesterol (LDL-C) levels is recommended and can be achieved in most patients with dyslipidemia by treatment with 3-hydroxy-3methylglutaryl-coenzyme A reductase inhibitors, also known as statins. High serum triglyceride (TG) and low

high-density lipoprotein cholesterol (HDL-C) levels have come to be recognized as residual cardiovascular (CV) risk factors. Thus, in various disorders of lipid metabolism, high TG and low HDL-C levels (the dominant type of dyslipidemia in patients with CKD) have recently attracted attention and have been shown to predict the occurrence of ischemic heart disease, myocardial infarction, and CV mortality.3,4 TG:HDL-C ratio, which has been considered to be a useful marker of insulin resistance,5-8 reportedly

From the Departments of 1Integrated Therapy for Chronic Kidney Disease and 2Medicine and Clinical Science, Graduate School of Medical Sciences, Kyushu University, Fukuoka; 3 Steering Committee for “Design of the Comprehensive Health Care System for Chronic Kidney Disease (CKD) Based on the Individual Risk Assessment by Specific Health Checkup,” Fukushima; 4Dialysis Unit, University Hospital of the Ryukyus, Okinawa; 5iAnalysis LLC; and 6Department of Integrated Science and Engineering for Sustainable Society, Chuo University, Tokyo, Japan.

Received December 19, 2014. Accepted in revised form May 5, 2015. Address correspondence to Kazuhiko Tsuruya, MD, PhD, Department of Integrated Therapy for Chronic Kidney Disease, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan. E-mail: [email protected]  2015 by the National Kidney Foundation, Inc. 0272-6386 http://dx.doi.org/10.1053/j.ajkd.2015.05.011

Am J Kidney Dis. 2015;-(-):---

1

Tsuruya et al

correlates with small dense LDL-C level and LDL particle size.9-11 Recent studies have shown that TG:HDL-C ratio may be a better predictor of CV events3,8,12 and CV mortality13,14 than other lipid parameters, including TG level, LDL-C level, and total cholesterol to HDL-C ratio. We recently reported the association between TG:HDL-C ratio and the prevalence of CKD in a large Japanese population.15 However, the crosssectional study design limited the interpretation of causality between TG:HDL-C ratio and prevalence of CKD. In the present study, we hypothesized that an elevated TG:HDL-C ratio is involved in the decline in estimated glomerular filtration rate (eGFR) and the incidence and progression of CKD. We thus examined associations between TG:HDL-C ratio at baseline and changes in eGFR and the incidence and progression of CKD during a 2-year prospective longitudinal study.

initiated in 2008 by the Japanese government, and it promotes early diagnosis and intervention strategies for the prevention of metabolic syndrome. Data were collected from 1,120,101 individuals who participated in the health checkups in 2008 to 2011. Inclusion and exclusion criteria for participants are shown in the flow chart in Fig 1. Of 1,120,101 participants, we selected 261,958 with data for eGFR and urinary protein level from 2 time points at a 2-year interval. We then excluded 136,676 participants without essential data including age; sex; systolic blood pressure; diastolic blood pressure; body mass index (BMI); waist circumference; levels of hemoglobin A1c (HbA1c), LDL-C, HDL-C, and TG; information on smoking, alcohol consumption, and exercise habits; histories of stroke and heart disease; and medications for hypertension, diabetes mellitus, and dyslipidemia. We also excluded participants with a history of kidney disease and those with eGFR , 15 mL/ min/1.73 m2. Finally, data from 124,700 participants (50,392 men and 74,308 women) aged 39 to 74 years were analyzed in the present study. This study was conducted in accordance with the Private Information Protection Law and ethics guidelines for epidemiology research published by the Ministry of Health, Labour and Welfare in 2008.

Clinical Evaluation and Laboratory Measurements

METHODS Study Population This longitudinal cohort study was conducted as part of a prospective ongoing project titled “Study on the design of the comprehensive health care system for CKD based on the individual risk assessment by Specific Health Checkups” and was based on data obtained from the Japanese Specific Health Check and Guidance System. This annual health check program was

All participants completed a self-administered questionnaire that documented medical history, current medications, smoking habit (current smoker or not), alcohol consumption (daily drinker or not), and regular exercise habit. Participants’ height and weight were measured, and BMI was calculated (kg/m2). For these measurements, participants wore light clothing without shoes. Blood pressure measurement and blood and urine sampling were performed at each participant’s local medical institute, as stipulated by the health check program.

All participants who took part in the health checkup from 2008 to 2011 (n = 1,120,101)

Participants with data on eGFR & proteinuria in 2008 (n = 552,353)

Participants without data on eGFR or proteinuria in 2008 (n = 567,748) Participants with data on eGFR & proteinuria in 2009 (n = 242,386)

Participants with data on eGFR & proteinuria in 2010 (n = 217,938)

Participants with data on eGFR & proteinuria in 2011 (n = 44,020)

Participants with data on eGFR & proteinuria at 2 points with an interval of 2 years (n = 261,958) Exclusion of participants without essential data (n = 136,676) * Participants with all essential data (n = 125,282) Exclusion of participants with history of kidney diseases or those with eGFR < 15 mL/min/1.73 m2 (n = 542)

Study 1

Linear regression analysis for change in eGFR and logistic regression analysis for increase in urinary protein excretion

Participants for this study (n = 124,700)

Participants without CKD at baseline (n = 102,900)

Study 2

Logistic regression analyses for incident CKD, low eGFR, and proteinuria

Participants with CKD at baseline (n = 21,800)

Study 3

Logistic regression analyses for progression of CKD

Figure 1. Flow chart of study participants. *Essential data: age; sex; systolic blood pressure; diastolic blood pressure; levels of hemoglobin A1c, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and triglycerides; waist circumference; information on smoking, alcohol consumption, and exercise habits; histories of stroke and heart disease; and medication for hypertension, diabetes mellitus, and dyslipidemia. Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate. 2

Am J Kidney Dis. 2015;-(-):---

Impact of TG:HDL-C Ratio on CKD Blood samples were collected after participants had fasted overnight, and blood was analyzed using an automated clinical chemical analyzer within 24 hours of sampling. All blood analyses were conducted at a local, rather than a central, laboratory. Methods used for blood analyses were not calibrated between laboratories; however, the Japan Society of Clinical Chemistry recommended methods for laboratory testing several years ago, and these recommendations have been widely adopted by laboratories across Japan. The enzymatic method was used to measure serum creatinine in fresh blood samples. LDL-C, HDL-C, and TG levels were determined enzymatically. HbA1c level was expressed as a National Glycohemoglobin Standardization Program equivalent value, which was calculated according to the following formula: HbA1c (%) 5 HbA1c (Japan Diabetes Society) (%) 1 0.4%.

Definitions CKD, Low eGFR, Proteinuria, Hypertension, Diabetes, Obesity, and TG:HDL-C Ratio Urinary protein excretion was examined by dipstick testing and categorized into 5 degrees; –, 6, 11, 21, and 31. Proteinuria was defined as urinary protein $ 11. CKD was defined as eGFR , 60 mL/min/1.73 m2 and/or the presence of proteinuria. eGFR was calculated using the following equation: eGFR (mL/ min/1.73 m2) 5 194 3 serum creatinine (mg/dL)-1.094 3 age (years)-0287 3 0.739 (for women).16 Low eGFR was defined as eGFR , 60 mL/min/1.73 m2. Hypertension was defined as systolic blood pressure $ 140 mm Hg and/or diastolic blood pressure $ 90 mm Hg or self-reported use of antihypertensive drugs. Diabetes mellitus was defined in accordance with the guidelines of the American Diabetes Association17: fasting glucose concentration $ 126 mg/dL, HbA1c level $ 6.5%, or selfreported use of antihyperglycemic drugs. Obesity was defined as BMI $ 25 kg/m2. TG:HDL-C ratio was calculated as TG level (mg/dL) divided by HDL-C level (mg/dL). Male and female participants were separately grouped into quartiles based on TG:HDL-C ratios. TG:HDL-C ratios for the quartile groups are shown in Table S1 (provided as online supplementary material).

Decline in eGFR and Increase in Urinary Protein Excretion Decline in eGFR was assessed using the following 3 definitions according to the guidelines provided by the KDIGO (Kidney

Disease: Improving Global Outcomes) group18 and a recent report by the CKD Prognosis Consortium19: (1) decline in GFR category ($90 [G1], 60-89 [G2], 45-59 [G3a], 30-44 [G3b], 15-29 [G4], and ,15 [G5] mL/min/1.73 m2), defined as a certain decrease in GFR category accompanied by a $25% decrease in eGFR from baseline18; (2) rapid decline in eGFR, defined as sustained decline in eGFR . 5 mL/min/1.73 m2 per year18; and (3) .30% decline in eGFR, defined as .30% decrease in eGFR from baseline.19 In addition to the mentioned outcomes, we examined the increase in urinary protein excretion, defined as a certain increase in semiquantitative urinary protein (–, 6, 11, 21, and 31) on dipstick testing in this report.

Study Protocol We performed the following 3 studies in this report. Eligibility of participants, sample size, definition of outcomes, and number of participants who met the outcomes in each study are shown in Table 1.

Study 1 (all 124,700 participants) First, we examined associations between sex-specific quartiles of TG:HDL-C ratio and changes in eGFR or increase in urinary protein excretion during the 2-year study period in all participants.

Study 2 (102,900 participants without CKD) Next, to elucidate the impact of TG:HDL-C ratio on incident CKD, we examined associations between sex-specific quartiles of TG:HDL-C ratio and incident CKD, low eGFR, and proteinuria after 2 years in 102,900 participants without CKD at baseline.

Study 3 (21,800 participants with CKD) Third, to elucidate the impact of TG:HDL-C ratio on progression of CKD, we examined the association between sex-specific quartiles of TG:HDL-C ratio and decline in eGFR or increase in urinary protein excretion after 2 years in 21,800 participants with CKD at baseline.

Statistical Analyses Independent 2-sample t tests and c2 tests were used to analyze continuous and categorical variables, respectively. We used a linear regression model to compare mean values of possible risk factors among the quartile groups in each sex. The multivariable-adjusted

Table 1. Eligibility of Participants, Sample Size, and Definition of Outcome in Each Study Eligibility of Participants

Sample Size

Definition of Outcome

Study 1

Both CKD and non-CKD

124,700

Study 2

Non-CKD

102,900

Study 3

CKD

2-y change in eGFR Increase in urinary protein excretiona Incident CKDb Incident low eGFRc Incident proteinuriad Decline in GFR categorye Rapid decline in eGFRf .30% decline in eGFRg Increase in urinary protein excretiona

21,800

Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; GFR, glomerular filtration rate. a Increase in urinary protein excretion was defined as certain increase in category of semiquantitative urinary protein (–, 6，11, 21, 31) on dipstick testing. b CKD was defined as eGFR , 60 mL/min/1.73 m2 and/or proteinuria. c Proteinuria was defined as urinary protein $ 11 on dipstick testing. d Low eGFR was defined as eGFR , 60 mL/min/1.73 m2. e Decline in GFR category ($90 [G1], 60-89 [G2], 45-59 [G3a], 30-44 [G3b], 15-29 [G4], ,15 [G5] mL/min/1.73 m2) was defined as certain decrease in GFR category accompanied by $25% decrease in eGFR from baseline. f Rapid decline in eGFR was defined as sustained decline in eGFR . 5 mL/min/1.73 m2 per year. g .30% decline in eGFR was defined as .30% decrease in eGFR from baseline. Am J Kidney Dis. 2015;-(-):---

3

Tsuruya et al least-square mean values of DeGFR were calculated using analysis of covariance. Univariable, age- and sex-related, or multivariableadjusted odds ratios (ORs) and 95% confidence intervals (CIs) for incident CKD, low eGFR, and proteinuria, as well as progression of CKD, were determined using a logistic regression model adjusted for potential confounding covariates. The confounding variables used for adjustment included age, sex, waist circumference, hypertension, obesity, diabetes mellitus, current smoking habit, daily alcohol consumption, regular exercise habit, histories of stroke and heart disease, and medication for dyslipidemia. We tested for heterogeneity between subgroups by adding a multiplicative interaction term to our statistical model. We performed stratified analyses of the association between TG:HDL-C ratio and incidence of CKD by sex and the presence of diabetes mellitus, hypertension, or obesity because these factors are known as risk factors for both dyslipidemia and CKD. They were also examined in our previous study,15 in which significant interactions were observed between TG:HDL-C ratio and CKD due to diabetes mellitus or hypertension, but not due to obesity. All statistical analyses were performed with JMP, version 11.0, software (SAS Institute Inc).

RESULTS Study 1 (all 124,700 participants) Table 2 shows clinical features of the 124,700 participants according to sex-specific quartiles of TG:HDL-C ratio. Age, BMI, waist circumference, blood pressure, and levels of HbA1c, LDL-C, and TG increased with higher TG:HDL-C ratios. The frequency of hypertension, diabetes mellitus, obesity, current smoking habit, histories of stroke and heart disease, and medication for hypertension, diabetes mellitus, and dyslipidemia also increased, whereas HDL-C levels and frequencies of daily alcohol consumption and regular exercise decreased with higher TG:HDL-C ratios. Correlations between these parameters and TG:HDL-C ratio at baseline were similar in both male and female participants with the

Table 2. Clinical Features of All Participants According to Sex-Specific Quartiles of TG:HDL-C Ratio

No. Age, y Male sex BMI, kg/m2 Waist circumference, cm Systolic BP, mm Hg Diastolic BP, mm Hg Fasting blood glucose, mg/dL Hemoglobin A1c, % LDL-C, mg/dL HDL-C, mg/dL TG, mg/dL TG:HDL-C ratio Hypertension Diabetes mellitus Obesity Current smoker Daily drinker Regular exercise History of stroke History of heart disease Medication for hypertension Medication for diabetes mellitus Medication for dyslipidemia eGFR, mL/min/1.73 m2 eGFR , 60 mL/min/1.73 m2 Proteinuria CKD

Q1

Q2

Q3

Q4

M: ,1.26 F: ,0.97

M: 1.26-1.96 F: 0.97-1.45

M: 1.97-3.14 F: 1.46-2.24

M: .3.14 F: .2.24

P for Trend

31,175 62.5 6 8.0 12,598 (40.4) 21.9 6 2.9 79.5 6 8.6 126 6 18 75 6 11 95 6 16 5.6 6 0.5 115 6 27 77 6 15 60 6 15 0.80 6 0.21 11,334 (36.4) 2,155 (6.9) 4,196 (13.5) 3,195 (10.3) 8,464 (27.2) 14,241 (45.7) 915 (2.9) 1,591 (5.1) 6,905 (22.2) 1,192 (3.8) 3,895 (12.5) 77.0 6 15.3 3,347 (10.7) 1,165 (3.7) 4,274 (13.7)

31,168 63.4 6 7.4 12,592 (40.4) 23.0 6 3.1 83.0 6 8.6 128 6 17 76 6 11 96 6 18 5.7 6 0.6 126 6 28 65 6 12 87 6 17 1.36 6 0.25 13,531 (43.4) 2,599 (8.3) 7,175 (23.0) 3,838 (12.3) 6,945 (22.3) 13,833 (44.4) 954 (3.1) 1,725 (5.5) 8,684 (27.9) 1,273 (4.1) 5,007 (16.1) 75.4 6 15.2 4,031 (12.9) 1,344 (4.3) 5,055 (16.2)

31,184 63.7 6 7.2 12,602 (40.4) 23.8 6 3.2 85.3 6 8.6 130 6 17 77 6 11 98 6 19 5.8 6 0.6 132 6 29 57 6 10 117 6 25 2.07 6 0.43 14,987 (48.1) 3,318 (10.6) 9,718 (31.2) 4,371 (14.0) 6,132 (19.7) 13,233 (42.4) 1,050 (3.4) 1,762 (5.7) 9,946 (31.9) 1,556 (5.0) 5,899 (18.9) 74.4 6 15.2 4,583 (14.7) 1,583 (5.1) 5,743 (18.4)

31,173 63.2 6 7.5 12,600 (40.4) 24.6 6 3.3 87.4 6 8.3 132 6 17 78 6 11 101 6 24 5.9 6 0.7 132 6 32 48 6 9 202 6 100 4.38 6 2.75 16,494 (52.9) 4,445 (14.3) 12,710 (40.8) 5,591 (17.9) 5,947 (19.1) 12,585 (40.4) 1,027 (3.3) 1,738 (5.6) 10,829 (34.7) 1,807 (5.8) 6,426 (20.6) 73.9 6 15.7 5,243 (16.8) 2,110 (6.8) 6,728 (21.6)

,0.001 0.9 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 0.002 0.008 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001

Note: Study 1; N 5 124,700. Values for continuous variables given as mean 6 standard deviation; values for categorical variables, as number (percentage). Hypertension was defined as systolic BP $ 140 mm Hg, diastolic BP $ 90 mm Hg, or self-reported use of antihypertensive drugs. Diabetes was defined in accordance with American Diabetes Association guidelines as fasting glucose level $ 126 mg/dL, hemoglobin A1c level $ 6.5%, or self-reported use of antihyperglycemic drugs. Obesity was defined as BMI $ 25 kg/m2. Proteinuria was defined as urinary protein $ 11 on dipstick testing. CKD was defined as eGFR , 60 mL/min/1.73 m2 and/or proteinuria. TG:HDL-C ratio was calculated as TG level (mg/dL) divided by HDL-C level (mg/dL). Conversion factors for units: glucose in mg/dL to mmol/L, 30.05551; LDL-C and HDL-C in mg/dL to mmol/L, 30.02586; TG in mg/dL to mmol/L, 30.01129. Abbreviations: BMI, body mass index; BP, blood pressure; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; Q, quartile; TG, triglycerides. 4

Am J Kidney Dis. 2015;-(-):---

Impact of TG:HDL-C Ratio on CKD

Sex-specific quartile of TG:HDL-C ratio Q3 Q1 Q2 Q4 0.0

-0.5

-1.0

-1.5

B

Least square mean

Least square mean

A

P for trend < 0.001

Sex-specific quartile of TG:HDL-C ratio Q1 Q2 Q3 Q4 0.0

-0.5

-1.0

P for trend < 0.001

-1.5

Figure 2. Change in estimated glomerular filtration rate (eGFR) according to quartile (Q) of triglycerides to high-density lipoprotein cholesterol (TG:HDL-C) ratio (results of study 1). The 2-year changes in eGFR according to quartiles of TG:HDL-C ratio based on (A) baseline eGFR-adjusted and (B) multivariable-adjusted regression models are shown. Each bar represents the least square mean of the change in eGFR during the 2-year study period (eGFR at 2 years minus eGFR at baseline) and each vertical line represents the standard error of the mean.

exception of age and histories of stroke and heart disease, as shown in our previous study.15 The decline in eGFR during the 2-year study period increased with higher TG:HDL-C ratios. This trend remained statistically significant even after adjustment for various confounding factors by analysis of covariance (Fig 2; Table 3). Stratified analyses according to sex and the presence of hypertension or obesity showed no interaction,

suggesting that the decline in eGFR during the 2 years increased linearly with higher TG:HDL-C ratios regardless of these factors (Fig 3A, C, and D). Significant interactions were observed between TG:HDL-C ratio and diabetes mellitus (P for interaction 5 0.002), suggesting that higher TG:HDL-C ratio was a more relevant factor for the decline in eGFR in participants with than without diabetes mellitus (Fig 3B).

Table 3. Least Square Mean Change in eGFR, and ORs for Increase in Urinary Protein Excretion, by Sex-Specific TG:HDL-C Ratio Quartile, in All Participants

No. of participants Change in eGFR Unadjusted LSM (SE) eGFR-adjusted LSM (SE) Multivariable-adjusted LSM (SE)a Increase in urinary protein excretionb No. of cases (%) Unadjusted OR (95% CI) Age- and sex-adjusted OR (95% CI) Multivariable-adjusted OR (95% CI)a

Q1

Q2

Q3

Q4

M: ,1.26 F: ,0.97

M: 1.26-1.96 F: 0.97-1.45

M: 1.97-3.14 F: 1.46-2.24

M: .3.14 F: .2.24

31,175

31,168

31,184

31,173

20.69 (0.06) 20.28 (0.06) 20.52 (0.12)

20.69 (0.06) 20.64 (0.06) 20.72 (0.12)

20.82 (0.06) 21.00 (0.06) 21.02 (0.12)

20.94 (0.06) 21.22 (0.06) 21.25 (0.12)

2,627 (8.4) 1.00 (reference) 1.00 (reference) 1.00 (reference)

2,848 (9.1) 1.09 (1.03-1.16) 1.09 (1.03-1.15) 1.02 (0.97-1.08)

3,179 (10.2) 1.23 (1.17-1.30) 1.22 (1.16-1.30) 1.10 (1.04-1.16)

3,559 (11.4) 1.40 (1.33-1.48) 1.40 (1.33-1.48) 1.18 (1.12-1.25)

Note: Study 1; N 5 124,700. P for trend for all is P , 0.001. Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; LSM, least-square mean; OR, odds ratio; Q, quartile; SE, standard error; TG, triglycerides. a Multivariable analyses were adjusted for age, sex, waist circumstance, hypertension, obesity, diabetes mellitus, current smoking, daily alcohol consumption, regular exercise habit, history of stroke and heart disease, medication for dyslipidemia, urinary protein excretion, and eGFR. b Increase in urinary protein excretion was defined as certain increase in semiquantitative urinary protein (–, 6，１1, 21, 31) on dipstick testing. Am J Kidney Dis. 2015;-(-):---

5

Tsuruya et al

Female

Male

Sex-specific quartile of TG:HDL-C ratio Q1 Q2 Q3 Q4 0.0

Sex-specific quartile of TG:HDL-C ratio Q1 Q2 Q3 Q4 0.0 Least square mean

Least square mean

A

-0.5 -1.0 -1.5 -2.0

-0.5 -1.0 -1.5 -2.0

P for trend < 0.001

P for trend < 0.001

P for interaction = 0.08

Diabetes (-)

Diabetes (+)

Sex-specific quartile of TG:HDL-C ratio Q3 Q1 Q2 Q4 0.5

Sex-specific quartile of TG:HDL-C ratio Q1 Q2 Q3 Q4 0.5 Least square mean

Least square mean

B

0.0 -0.5 -1.0 -1.5 -2.0 -2.5

0.0 -0.5 -1.0 -1.5 -2.0 -2.5

P for trend < 0.001

P for trend < 0.001

P for interaction = 0.002

Hypertension (-)

Hypertension (+)

Sex-specific quartile of TG:HDL-C ratio Q1 Q2 Q3 Q4 0.5

Sex-specific quartile of TG:HDL-C ratio

0.0 -0.5 -1.0 -1.5 -2.0

0.5 Least square mean

Least square mean

C

Q2

Q3

Q4

0.0 -0.5 -1.0 -1.5 -2.0

P for trend < 0.001

Q1

P for trend < 0.001

P for interaction = 0.1

Obesity (-)

Obesity (+)

Sex-specific quartile of TG:HDL-C ratio Q1 Q2 Q3 Q4 0.0

Sex-specific quartile of TG:HDL-C ratio Q1 Q2 Q3 Q4 0.0 Least square mean

Least square mean

D

-0.5 -1.0 -1.5 -2.0

P for trend < 0.001

-0.5 -1.0 -1.5 -2.0

P for trend < 0.001

P for interaction = 0.6

Logistic regression analysis for increase in urinary protein excretion showed that univariable and ageand sex-adjusted ORs for the outcome in quartiles 2, 3, and 4 were significantly higher than ORs in 6

Figure 3. Change in estimated glomerular filtration rate (eGFR) according to quartile (Q) of triglycerides to highdensity lipoprotein cholesterol (TG:HDLC) ratio stratified by various variables (results of study 1). The 2-year changes in the eGFR based on a multivariableadjusted regression model stratified by (A) sex, (B) diabetes mellitus, (C) hypertension, and (D) obesity are shown. Each bar represents the least square mean of the change in eGFR during the 2-year study period (eGFR at 2 years minus eGFR at baseline) and each vertical line represents the standard error of the mean.

quartile 1. We also calculated ORs for increase in urinary protein excretion after adjustment for age, sex, waist circumference, hypertension, obesity, diabetes mellitus, current smoking habit, daily alcohol Am J Kidney Dis. 2015;-(-):---

Impact of TG:HDL-C Ratio on CKD

consumption, regular exercise habit, histories of stroke and heart disease, medication for dyslipidemia, urinary protein excretion, and eGFR at baseline. The OR for increase in urinary protein excretion progressively increased with higher TG:HDL-C ratios (Table 3). Study 2 (102,900 participants without CKD) We examined the associations of TG:HDL-C ratio with incident CKD, low eGFR, and proteinuria in participants without CKD. Clinical features of the 102,900 participants without CKD according to sexspecific quartiles of TG:HDL-C ratio were similar to 124,700 participants in study 1 except for the histories of stroke and heart disease, which showed no significant trends (Table S2). ORs for these 3 outcomes progressively increased with higher TG:HDL-C ratios (Table 4). Stratified analyses of association between TG:HDL-C ratio and incident CKD according to sex and the presence of diabetes mellitus, hypertension, or obesity revealed that the risk for CKD increased linearly with greater TG:HDL-C ratios regardless of these factors. Moreover, higher TG:HDL-C ratio was a more relevant factor for incident CKD, especially in participants with diabetes mellitus (P for interaction 5 0.05; Fig 4).

Study 3 (21,800 participants with CKD) We then examined associations of TG:HDL-C ratio with decline in eGFR and increase in urinary protein excretion in participants with CKD. Clinical features of the 21,800 participants with CKD according to sexspecific quartiles of TG:HDL-C ratio were similar to the 124,700 participants in study 1 except for histories of stroke and heart disease and frequency of low eGFR, which showed no significant trends (Table S3). Logistic regression analyses showed that univariable and age- and sex-adjusted ORs for these outcomes in quartiles 3 and 4 were significantly higher than ORs in quartile 1. Multivariable-adjusted ORs in quartile 4 were significantly higher than ORs in quartile 1. Significant associations were found between sexspecific quartiles of TG:HDL-C ratio and rapid decline in eGFR or increase in urinary protein excretion, even after adjustment for multivariable relevant factors (Table 5).

DISCUSSION In the present study, we examined the involvement of TG:HDL-C ratio at baseline with 2-year changes in eGFR and increase in urinary protein excretion in the entire population of participants (study 1) and with new-onset CKD, low eGFR, and proteinuria among

Table 4. ORs for Incident CKD, Low eGFR, and Proteinuria, by Sex-Specific TG:HDL-C Ratio Quartile, in Participants Without CKD Q1

Q2

Q3

Q4

M: ,1.22 F: ,0.96

M: 1.22-1.89 F: 0.96-1.43

M: 1.89-3.02 F: 1.43-2.20

M: .3.02 F: .2.20

No. of participants CKD No. of cases (%) Unadjusted OR (95% CI) Age- and sex-adjusted OR (95% CI) Multivariable-adjusted OR (95% CI)a

25,733

25,721

25,730

25,716

1,953 (7.6) 1.00 (reference) 1.00 (reference) 1.00 (reference)

2,322 (9.0) 1.21 (1.13-1.29) 1.18 (1.11-1.26) 1.07 (1.00-1.14)

2,533 (9.8) 1.33 (1.25-1.41) 1.30 (1.22-1.38) 1.10 (1.03-1.17)

2,937 (11.4) 1.57 (1.48-1.67) 1.56 (1.47-1.66) 1.25 (1.18-1.34)

Low eGFR No. of cases (%) Unadjusted OR (95% CI) Age- and sex-adjusted OR (95% CI) Multivariable-adjusted OR (95% CI)a

1,342 (5.2) 1.00 (reference) 1.00 (reference) 1.00 (reference)

1,637 (6.4) 1.24 (1.15-1.33) 1.20 (1.11-1.29) 1.07 (0.99-1.16)

1,739 (6.8) 1.32 (1.22-1.42) 1.27 (1.18-1.37) 1.07 (0.99-1.16)

1,948 (7.6) 1.49 (1.39-1.60) 1.48 (1.38-1.59) 1.20 (1.11-1.30)

Proteinuria No. of cases (%) Unadjusted OR (95% CI) Age- and sex-adjusted OR (95% CI) Multivariable-adjusted OR (95% CI)a

677 (2.6) 1.00 (reference) 1.00 (reference) 1.00 (reference)

774 (3.0) 1.15 (1.03-1.28) 1.14 (1.03-1.27) 1.05 (0.95-1.17)

871 (3.4) 1.30 (1.17-1.44) 1.29 (1.17-1.43) 1.10 (0.99-1.23)

1,095 (4.3) 1.65 (1.49-1.82) 1.65 (1.49-1.82) 1.27 (1.15-1.42)

Note: Study 2; N 5 102,900. P for trend for all is P , 0.001. CKD was defined as eGFR , 60 mL/min/1.73 m2 and/or proteinuria. Low eGFR was defined as eGFR , 60 mL/min/1.73 m2. Proteinuria was defined as urinary protein $ 11 on dipstick testing. Abbreviations: CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; OR, odds ratio; HDL-C, high-density lipoprotein cholesterol; Q, quartile; TG, triglycerides. a Multivariable analyses were adjusted for age, sex, waist circumstance, hypertension, obesity, diabetes mellitus, current smoking, daily alcohol consumption, regular exercise habit, history of stroke and heart disease, medication for dyslipidemia, urinary protein excretion, and eGFR.

Am J Kidney Dis. 2015;-(-):---

7

Tsuruya et al

Figure 4. Odds ratio for incident chronic kidney disease (CKD) according to quartile (Q) of triglyceride to high-density lipoprotein cholesterol (TG:HDL-C) ratio stratified by various variables (results of study 2). Odds ratios for incident CKD according to the quartile of TG:HDL-C ratio stratified by sex, diabetes mellitus, hypertension, and obesity are shown.

those without CKD (study 2). We also examined the involvement of TG:HDL-C ratio at baseline with decline in eGFR and increase in urinary protein excretion in participants with CKD (study 3). All examinations revealed significant involvement of TG:HDL-C ratio in all outcomes even after adjustment for relevant potential confounding factors, indicating that higher TG:HDL-C ratio is an independent risk factor for the incidence and progression of CKD. This finding suggests that elevated levels of small dense LDL-C might induce and aggravate CKD. 8

Previous prospective cohort studies in the general population have shown relationships between various types of dyslipidemia and loss of kidney function. Specifically, high TG, total cholesterol, and non–HDL-C levels; low HDL-C level; and high total cholesterol to HDL-C ratio were significantly associated with increased risk of developing decreased kidney function in the Atherosclerosis Risk in Communities (ARIC) Study,20 Physicians Health Study,21 and Framingham Offspring Study.22 With respect to the relationship between TG:HDL-C ratio and loss of Am J Kidney Dis. 2015;-(-):---

Impact of TG:HDL-C Ratio on CKD Table 5. ORs for Decline in eGFR and Increase in Urinary Protein Excretion, by Sex-Specific TG:HDL-C Ratio Quartile, in Participants With CKD Q1

Q2

Q3

Q4

M: ,1.42 F: ,1.08

M: 1.43-2.21 F: 1.08-1.62

M: 2.21-3.50 F: 1.62-2.54

M: .3.50 F: .2.54

P for Trend

No. of participants Decline in GFR category No. of cases (%) Unadjusted OR (95% CI) Age- and sex-adjusted OR (95% CI) Multivariable-adjusted OR (95% CI)a

5,445

5,455

5,446

5,454

79 (1.5) 1.00 (reference) 1.00 (reference) 1.00 (reference)

105 (1.9) 1.33 (0.99-1.79) 1.33 (0.99-1.79) 1.31 (0.98-1.79)

112 (2.1) 1.43 (1.07-1.91) 1.42 (1.06-1.90) 1.31 (0.97-1.78)

141 (2.6) 1.80 (1.37-2.39) 1.81 (1.37-2.40) 1.51 (1.12-2.04)

,0.001 ,0.001 0.03

Rapid decline in eGFR No. of cases (%) Unadjusted OR (95% CI) Age- and sex-adjusted OR (95% CI) Multivariable-adjusted OR (95% CI)a

314 (5.8) 1.00 (reference) 1.00 (reference) 1.00 (reference)

345 (6.3) 1.10 (0.94-1.29) 1.12 (0.96-1.32) 1.18 (0.99-1.40)

400 (7.3) 1.30 (1.11-1.51) 1.32 (1.14-1.54) 1.39 (1.17-1.65)

405 (7.4) 1.31 (1.13-1.53) 1.31 (1.12-1.52) 1.23 (1.03-1.46)

,0.001 ,0.001 0.01

.30% decline in eGFR No. of cases (%) Unadjusted OR (95% CI) Age- and sex-adjusted OR (95% CI) Multivariable-adjusted OR (95% CI)a

46 (0.8) 1.00 (reference) 1.00 (reference) 1.00 (reference)

60 (1.1) 1.31 (0.89-1.93) 1.30 (0.89-1.92) 1.35 (0.91-2.01)

75 (1.4) 1.64 (1.14-2.39) 1.63 (1.13-2.38) 1.53 (1.04-2.26)

100 (1.8) 2.19 (1.55-3.14) 2.20 (1.56-3.15) 1.73 (1.19-2.54)

,0.001 ,0.001 0.004

Increase in urinary protein excretion No. of cases (%) Unadjusted OR (95% CI) Age- and sex-adjusted OR (95% CI) Multivariable-adjusted OR (95% CI)a

463 (8.5) 1.00 (reference) 1.00 (reference) 1.00 (reference)

533 (9.8) 1.17 (1.02-1.33) 1.16 (1.01-1.32) 1.05 (0.92-1.21)

624 (11.5) 1.39 (1.23-1.58) 1.38 (1.22-1.57) 1.19 (1.04-1.35)

695 (12.7) 1.57 (1.39-1.78) 1.58 (1.40-1.79) 1.28 (1.12-1.46)

,0.001 ,0.001 ,0.001

Note: Study 3; N 5 21,800. Decline in GFR category was defined as certain decrease in eGFR category accompanied by $25% decrease in eGFR from baseline. Rapid decline in eGFR was defined as sustained decline in eGFR $ 5 mL/min/1.73 m2 per year. .30% decline in eGFR was defined as .30% decrease in eGFR from baseline. Increase in urinary protein excretion was defined as certain increase in semiquantitative urinary protein (–, 6，１1, 21, 31) on dipstick testing. Abbreviations: CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; OR, odds ratio; Q, quartile; TG, triglycerides. a Multivariable analyses were adjusted for age, sex, waist circumstance, hypertension, obesity, diabetes mellitus, current smoking, daily alcohol consumption, regular exercise habit, history of stroke and heart disease, medication for dyslipidemia, urinary protein excretion, and eGFR.

kidney function, some cross-sectional studies of the general population, including ours, have recently reported that TG:HDL-C ratio is associated with the prevalence of CKD and proteinuria.15,23-25 However, no prospective community-based cohort studies have been reported until now; only one prospective study of patients with type 2 diabetes mellitus reported an association between TG:HDL-C ratio and increased incidence of both CKD and retinopathy.26 Therefore, to our knowledge, the present study is the first to show the deleterious effects of elevated TG:HDL-C ratio on maintenance of kidney function in the general population. The mechanism of the impact of TG:HDL-C ratio on decline in eGFR and incidence of CKD is considered to involve TG:HDL-C ratio as a marker of LDL particle size. Previous investigations have reported a close relationship between high TG:HDL-C ratio and elevated level of small dense LDL-C,9-11 which are thought to be very atherogenic.27 The Am J Kidney Dis. 2015;-(-):---

effect of dyslipidemia on decreased kidney function was advocated as a “lipid nephrotoxicity” hypothesis by Moorhead et al28 in 1982. This hypothesis was recently updated to include the modification of lipid homeostasis and tissue lipid accumulation by inflammation-induced stress.29 CKD leads to the generation of small dense LDL particles, as well as elevation of plasma levels of intermediate-density lipoproteins and chylomicron remnants.30 These lipoproteins are highly prone to oxidation to lipid peroxides and other secondary oxidation products. Accumulation of oxidized LDL, intermediate-density lipoproteins, and chylomicron remnants stimulates monocytes and macrophages to release proinflammatory cytokines and chemokines and accelerates inflammation.31 Accordingly, systemic inflammation and oxidative stress promote decline in eGFR and the incidence and progression of CKD. Alternatively, data for lipoprotein subclasses in type 1 diabetes and nephropathy are ambiguous with 9

Tsuruya et al

respect to LDL particles as the main driver.32,33 It is considered that impaired HDL maturation, reverse cholesterol transport, and antioxidant and antiinflammatory actions of HDL might also play important roles in the incidence and progression of CKD in participants with high TG:HDL-C ratios.34 Furthermore, TG:HDL-C ratio was correlated with HDL-C level, which itself is an independent risk factor for the decline in eGFR in all participants and also the incidence and progression of CKD, although its impact was slightly lower than TG:HDL-C ratio in the present study (Tables S4 and S5). In our stratified analyses, high TG:HDL-C ratio was significantly associated with greater 2-year decline in eGFR and incidence of CKD, independent of sex and the comorbid conditions of hypertension, diabetes mellitus, and obesity. However, this association was stronger in participants with diabetes mellitus than in participants without the mentioned comorbid conditions. That is, higher TG:HDL-C ratio was a risk factor for loss of eGFR and incident CKD, especially in participants with diabetes mellitus, implying that a vicious circle exists among atherogenic dyslipidemia, diabetes mellitus, and CKD. The proposed mechanism is as follows. Hyperglycemia increases oxidative stress by mitochondrial superoxide production and forms advanced oxidation protein products and advanced glycation end products, which results in cytokine release. This encourages changes in lipoprotein metabolism, composition, and function, with these alterations then leading to an extremely atherogenic environment, thus perpetuating the vicious circle of accelerated atherogenesis.35 Our study has some limitations. First, the 2-year follow-up was too short to evaluate kidney disease outcomes. However, this limitation seemed to be overcome by the large sample size because similar results were found irrespective of any definition of rapid decline in eGFR in study 3. Second, single measurements of TG, HDL-C, serum creatinine, and urinary protein at baseline and serum creatinine and urinary protein after 2 years could have resulted in the misclassification of exposure, confounders, and outcomes. However, the association observed in the current study could not be attributed to this source of variability because random misclassification of this type would be likely to cause underestimation of study findings and bias results toward the null hypothesis. Third, GFR was not directly measured using the gold-standard method of inulin clearance, but was instead estimated with a serum creatinine–based equation. This could have over- or underestimated the actual GFR in the Japanese general population. Fourth, we did not directly show data regarding small dense LDL-C levels, although we discussed the possible pathogenetic role of small dense LDL. Fifth, 10

we could not evaluate CKD exactly because we had no data for CKD except for eGFR and proteinuria in this study. Sixth, there is a possibility that the people who participated in the health checkups do not capture the characteristics of the community because the response rate for this Specific Health Check and Guidance program was not high. This may have led to selection bias, with this population comprising relatively healthy local inhabitants. The prevalence of diabetes mellitus and hypertension in participants is lower in this study than those in the Japanese general population (10.0% and 45.2% vs 15.6% and 49.2%, respectively).36 However, this bias would underestimate the risk of higher TG:HDL-C ratio for the incidence and progression of CKD because higher TG:HDL-C ratio was a more relevant factor for the decline in eGFR in participants with than without diabetes mellitus, as shown in Fig 3. Thus, we believe that our conclusions are unchanged in regard to the general population. Finally, many individuals were excluded because of missing essential data. The eligible participants in this study are only 10% of the original pool of 1,120,101 individuals who participated in the health checkups from 2008 to 2011 (Fig 1). Therefore, we checked whether there is selection bias by comparison of baseline characteristics between the total 1,120,101 and selected 124,700 participants (Table S6). Notably, mean or median values are almost the same between them, although there are significant differences in most variables due to a property of big data. Thus, we do not think there is selection bias in the present study. There are also several strengths of the present study. First, to our knowledge, this is the first detailed longitudinal study of the impact of TG:HDL-C ratio on decline in eGFR, incidence of CKD, and increase in urinary protein excretion. Second, we evaluated a large general population with sufficient available data, including various definitions of CKD progression. Third, we were able to show the development of lipid nephrotoxicity to some extent because this study was longitudinal in nature. In conclusion, the present study demonstrated that TG:HDL-C ratio affects the incidence and progression of CKD. Further long-term studies with hard kidney disease outcomes such as the development of end-stage renal disease are required to clarify the causative relationship between serum TG:HDL-C ratio and CKD.

ACKNOWLEDGEMENTS We acknowledge the contributions of staff members who collected data and instructed participants with metabolic syndrome at screening centers in the regions of Hokkaido, Yamagata, Miyagi, Fukushima, Tochigi, Ibaraki, Chiba, Saitama, Tokyo, Kanagawa, Niigata, Ishikawa, Fukui, Nagano, Gifu, Osaka, Hyogo, Okayama, Tokushima, Kochi, Fukuoka, Saga, Kumamoto, Oita, Miyazaki, and Okinawa. Follow-up screenings are ongoing. Am J Kidney Dis. 2015;-(-):---

Impact of TG:HDL-C Ratio on CKD Support: This study was supported by Health and Labour Sciences Research Grants for “Research on the positioning of CKD in the Specific Health Check and Guidance System of Japan”, “Study on the appropriate states of Specific Health Checkups and Specific Health Guidance for prevention of CKD progression”, and the “Study on the design of the comprehensive health care system for CKD based on the individual risk assessment by Specific Health Checkups” from the Ministry of Health, Labour and Welfare of Japan. Financial Disclosure: The authors declare that they have no other relevant financial interests. Contributions: Study design: KT, HY, MN; data acquisition: KT, HY, CI, KI, SF, TKonta, TM, KY, IN, KK, MK, KA, IK, YO, TW; data cleaning: CI; data interpretation: KT, HY, MN, IK, YO; statistical analysis: KT, HY, MN, IK, YO; supervision: TKitazono, TW; funding: TW. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved. KT takes responsibility that this study has been reported honestly, accurately, and transparently; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.

SUPPLEMENTARY MATERIAL Table S1: Sex-specific TG:HDL-C ratio quartiles in all participants, without CKD, and with CKD. Table S2: Clinical features of participants without CKD, by sexspecific TG:HDL-C ratio quartile. Table S3: Clinical features of participants with CKD, by sexspecific TG:HDL-C ratio quartile. Table S4: ORs for incident CKD, by sex-specific HDL-C quartile. Table S5: ORs for decline in eGFR and increase in urinary protein excretion, by sex-specific HDL-C quartile, in participants with CKD. Table S6: Clinical features of all 1,120,101 individuals and eligible study participants. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2015.05.011) is available at www.ajkd.org

REFERENCES 1. Ritz E, Wanner C. Lipid changes and statins in chronic renal insufficiency. J Am Soc Nephrol. 2006;17(12)(suppl 3):S226S230. 2. Vaziri ND. Dyslipidemia of chronic renal failure: the nature, mechanisms, and potential consequences. Am J Physiol Renal Physiol. 2006;290(2):F262-F272. 3. Jeppesen J, Hein HO, Suadicani P, Gyntelberg F. Relation of high TG-low HDL cholesterol and LDL cholesterol to the incidence of ischemic heart disease. An 8-year follow-up in the Copenhagen Male Study. Arterioscler Thromb Vasc Biol. 1997;17(6):1114-1120. 4. Robins SJ, Lyass A, Zachariah JP, Massaro JM, Vasan RS. Insulin resistance and the relationship of a dyslipidemia to coronary heart disease: the Framingham Heart Study. Arterioscler Thromb Vasc Biol. 2011;31(5):1208-1214. 5. Giannini C, Santoro N, Caprio S, et al. The triglyceride-toHDL cholesterol ratio: association with insulin resistance in obese youths of different ethnic backgrounds. Diabetes Care. 2011;34(8):1869-1874. 6. Gasevic D, Frohlich J, Mancini GB, Lear SA. The association between triglyceride to high-density-lipoprotein cholesterol Am J Kidney Dis. 2015;-(-):---

ratio and insulin resistance in a multiethnic primary prevention cohort. Metabolism. 2012;61(4):583-589. 7. McLaughlin T, Abbasi F, Cheal K, Chu J, Lamendola C, Reaven G. Use of metabolic markers to identify overweight individuals who are insulin resistant. Ann Intern Med. 2003;139(10): 802-809. 8. Salazar MR, Carbajal HA, Espeche WG, et al. Relation among the plasma triglyceride/high-density lipoprotein cholesterol concentration ratio, insulin resistance, and associated cardiometabolic risk factors in men and women. Am J Cardiol. 2012;109(12):1749-1753. 9. Maruyama C, Imamura K, Teramoto T. Assessment of LDL particle size by triglyceride/HDL-cholesterol ratio in non-diabetic, healthy subjects without prominent hyperlipidemia. J Atheroscler Thromb. 2003;10(3):186-191. 10. Bhalodkar NC, Blum S, Enas EA. Accuracy of the ratio of triglycerides to high-density lipoprotein cholesterol for predicting low-density lipoprotein cholesterol particle sizes, phenotype B, and particle concentrations among Asian Indians. Am J Cardiol. 2006;97(7):1007-1009. 11. Boizel R, Benhamou PY, Lardy B, Laporte F, Foulon T, Halimi S. Ratio of triglycerides to HDL cholesterol is an indicator of LDL particle size in patients with type 2 diabetes and normal HDL cholesterol levels. Diabetes Care. 2000;23(11):1679-1685. 12. Gaziano JM, Hennekens CH, O’Donnell CJ, Breslow JL, Buring JE. Fasting triglycerides, high-density lipoprotein, and risk of myocardial infarction. Circulation. 1997;96(8):2520-2525. 13. Barzi F, Patel A, Woodward M, et al. A comparison of lipid variables as predictors of cardiovascular disease in the Asia Pacific region. Ann Epidemiol. 2005;15(5):405-413. 14. Bittner V, Johnson BD, Zineh I, et al. The triglyceride/ high-density lipoprotein cholesterol ratio predicts all-cause mortality in women with suspected myocardial ischemia: a report from the Women’s Ischemia Syndrome Evaluation (WISE). Am Heart J. 2009;157(3):548-555. 15. Tsuruya K, Yoshida H, Nagata M, et al. Association of the triglycerides to high-density lipoprotein cholesterol ratio with the risk of chronic kidney disease: analysis in a large Japanese population. Atherosclerosis. 2014;233(1):260-267. 16. Matsuo S, Imai E, Horio M, et al. Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis. 2009;53(6):982-992. 17. American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33(suppl 1):S62-S69. 18. Levin A, Stevens PE. Summary of KDIGO 2012 CKD guideline: behind the scenes, need for guidance, and a framework for moving forward. Kidney Int. 2014;85(1):49-61. 19. Coresh J, Turin TC, Matsushita K, et al. Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality. JAMA. 2014;311(24):2518-2531. 20. Muntner P, Coresh J, Smith JC, Eckfeldt J, Klag MJ. Plasma lipids and risk of developing renal dysfunction: the Atherosclerosis Risk in Communities Study. Kidney Int. 2000;58(1):293-301. 21. Schaeffner ES, Kurth T, Curhan GC, et al. Cholesterol and the risk of renal dysfunction in apparently healthy men. J Am Soc Nephrol. 2003;14(8):2084-2091. 22. Fox CS, Larson MG, Leip EP, Culleton B, Wilson PW, Levy D. Predictors of new-onset kidney disease in a communitybased population. JAMA. 2004;291(7):844-850. 23. Kang HT, Kim JK, Kim JY, Linton JA, Yoon JH, Koh SB. Independent association of TG/HDL-C with urinary albumin excretion in normotensive subjects in a rural Korean population. Clin Chim Acta. 2012;413(1-2):319-324. 11

Tsuruya et al 24. Kang HT, Shim JY, Lee YJ, et al. Association between the ratio of triglycerides to high-density lipoprotein cholesterol and chronic kidney disease in Korean adults: the 2005 Korean National Health and Nutrition Examination Survey. Kidney Blood Press Res. 2011;34(3):173-179. 25. Kim JY, Kang HT, Lee HR, Lee YJ, Shim JY. Comparison of lipid-related ratios for prediction of chronic kidney disease stage 3 or more in Korean adults. J Korean Med Sci. 2012;27(12): 1524-1529. 26. Zoppini G, Negri C, Stoico V, Casati S, Pichiri I, Bonora E. Triglyceride-high-density lipoprotein cholesterol is associated with microvascular complications in type 2 diabetes mellitus. Metabolism. 2012;61(1):22-29. 27. Mikhailidis DP, Elisaf M, Rizzo M, et al. “European panel on low density lipoprotein (LDL) subclasses”: a statement on the pathophysiology, atherogenicity and clinical significance of LDL subclasses. Curr Vasc Pharmacol. 2011;9(5):533-571. 28. Moorhead JF, Chan MK, El-Nahas M, Varghese Z. Lipid nephrotoxicity in chronic progressive glomerular and tubulointerstitial disease. Lancet. 1982;2(8311):1309-1311. 29. Ruan XZ, Varghese Z, Moorhead JF. An update on the lipid nephrotoxicity hypothesis. Nat Rev Nephrol. 2009;5(12):713-721.

12

30. Vaziri ND. Role of dyslipidemia in impairment of energy metabolism, oxidative stress, inflammation and cardiovascular disease in chronic kidney disease. Clin Exp Nephrol. 2014;18(2): 265-268. 31. Glass CK, Witztum JL. Atherosclerosis. the road ahead. Cell. 2001;104(4):503-516. 32. Tolonen N, Forsblom C, Thorn L, et al. Lipid abnormalities predict progression of renal disease in patients with type 1 diabetes. Diabetologia. 2009;52(12):2522-2530. 33. Thomas MC, Rosengard-Barlund M, Mills V, et al. Serum lipids and the progression of nephropathy in type 1 diabetes. Diabetes Care. 2006;29(2):317-322. 34. Streja E, Kovesdy CP, Streja DA, Moradi H, KalantarZadeh K, Kashyap ML. Niacin and progression of CKD. Am J Kidney Dis. 2015;65(5):785-798. 35. Krane V, Wanner C. The metabolic burden of diabetes and dyslipidaemia in chronic kidney disease. Nephrol Dial Transplant. 2002;17(suppl 11):23-27. 36. Usui T, Ninomiya T, Nagata M, et al. Angiopoietin-like protein 2 is associated with chronic kidney disease in a general Japanese population: the Hisayama Study. Circ J. 2013;77(9): 2311-2317.

Am J Kidney Dis. 2015;-(-):---