The Ansung cohort of the KoGES was established to investigate the genetic and environmental etiology of common complex diseases in Koreans. Participants of the Ansung cohort were recruited in 2001 to 2002, and biennial examinations are ongoing, with the latest follow-up performed in 2015 to 2016. The KoGES data are open to the public and the details of the study design and procedures were described previously [20]. Sera collected for oral glucose tolerance tests for participants in 2008 were available; thus, the visit in 2008 was considered as the baseline in this study. Using serum samples collected in 2008, sera of 1,537 participants were used to measure MIS and AhRL levels. We excluded 26 people whose yearly estimated glomerular filtration rate (eGFR) change could not be calculated. Therefore, 1,511 participants were included in this analysis. This study was approved by the ethics committee of the Nowon Eulji Medical Center, Eulji University (IRB Number: 2019-06-014). Written informed consent was obtained from all participants.

Serum AhRL bioactivity was determined as described previously [17,21]. pGL4-DRE-luc (puromycin+)/pRL-mTK double-positive stable cells and heat-inactivated serum were utilized [17]. The AhRL assay was similar to the Chemical Activated Luciferase gene eXpression (CALUX) assay, except that it utilized different recombinant cell lines and an organic solvent extraction-free sample preparation method [22]. All cell-based assays were performed in duplicate on blinded samples. The AhR bioactivity of serum sample-treated cells was presented as fold-induction compared to the AhR bioactivity of the 10% charcoal-stripped sera (CSS)-treated control. AhRL was expressed as 2,3,7,8-tetrachlorodibenzodioxin equivalents (TCDDeq, pM), in which a 0.1-fold induction of AhR bioactivity was equivalent to 0.37 pM TCDDeq [17]. The intra- and inter-assay coefficients of variation for AhRL were less than 5.0% [18].

Levels of MIS in serum samples were evaluated by measuring intracellular ATP content (MIS-ATP) and ROS generation (MIS-ROS), as described previously [17,21]. pRL-mTK-transfected mouse Hepa1c1c7 cells (5×10⁴/well) in a 96-well plate were treated with 10 μL heat-inactivated serum samples for 48 hours. The intracellular ATP content was determined using a CellTiter-Glo luciferase kit (Promega, Madison, WI, USA) with the output normalized to Renilla luciferase activity (% control). The ROS level was determined using 5-(and-6)-chloromethy l–2′, 7′-dichlorodihydrofluorescein diacetate and acetyl ester (CM-H₂DCFDA; Molecular Probes, Eugene, OR, USA). Both intracellular ATP and ROS levels were expressed as percentages of CSS-treated control (% control). The intra- and inter-assay coefficients of variation for these methods were less than 6.0% [18]. MIS-ATP ranged from 50.7% to 166.4% of control, while MIS-ROS ranged from 82.9% to 174.3%. Since intracellular ATP content decreases while ROS level increases in mitochondrial dysfunction, decreasing MIS-ATP and increasing MIS-ROS signified increased MIS exposure [1,6].

The eGFR was calculated using the equation from the Chronic Kidney Disease Epidemiology Collaboration for serum creatinine measurements [23]. At least two eGFR measurements were required to estimate the yearly eGFR change. For each individual, the yearly eGFR change (mL/min/1.73 m²/year) was determined with a regression coefficient using linear regression analysis. The mean±standard deviation follow-up period for eGFR measurements was 6.9±1.9 years. The definition of rapid decline in kidney function was a yearly eGFR change <−4 mL/min/1.73 m²/year [24].

Clinical data were downloaded from the KoGES database with permission. KoGES investigators measured anthropometric parameters and blood pressure (BP) by standard methods. Body mass index (BMI) was calculated as (weight in kg)/(height in m²). Waist circumference (WC, cm) was measured at the narrowest point between the lower rib and the iliac crest (measured to the nearest 0.1 cm), and the average of three repeated measurements was recorded. Blood and urine samples were obtained in the morning after at least 8 hours of overnight fasting. The specimens were sent to a central laboratory (Seoul Clinical Laboratories, Seoul, Korea). The fasting plasma concentrations of glucose, creatinine, and total cholesterol were measured using a Hitachi 747 chemistry analyzer (Hitachi Ltd., Tokyo, Japan) in a central laboratory. The hemoglobin A1c (HbA1c) level was determined using high-performance liquid chromatography (Variant II, BioRad Laboratories, Hercules, CA, USA). High-sensitivity C-reactive protein (hsCRP) level was measured using an immunoradiometric assay (ADVIA 1650, Bayer Diagnostics, Tarrytown, NY, USA), and plasma insulin concentration was measured using a radioimmunoassay (Linco kit, St. Charles, MO, USA) in a central laboratory. The homeostasis model assessment of insulin resistance (HOMA-IR) index was calculated following the original description [25]. Hypertension was defined as a systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg or physician’s diagnosis of hypertension. Diabetes was defined as fasting plasma glucose level ≥126 mg/dL, HbA1c ≥6.5%, or physician’s diagnosis of diabetes. Proteinuria was measured by dipstick urinalysis (URISCAN Pro II, YD Diagnostics Corp., Yongin, Korea) and defined when the result of dipstick urinalysis test was ≥1+. Tobacco smoking and alcohol drinking were defined as current smokers and drinkers, respectively.

Continuous variables are expressed as mean±standard deviation and categorical variables are expressed as percentages. Differences were analyzed using one-way analysis of variance (ANOVA) for normally distributed continuous variables and chi-square test for categorical variables. P trends were analyzed by Wilcoxon rank-sum test. Logistic regression analyses were performed to calculate odds ratios (ORs) and 95% confidence intervals for rapid decline in kidney. Multivariate analyses were performed with adjustment for age, sex, behavioral statuses (smoking, drinking, and exercise), metabolic parameters (diabetes, hypertension, BMI, total cholesterol, and hsCRP) and renal parameters (estimated glomerular filtration and proteinuria) based on clinical and statistical relevance. Regarding to exercise, three groups were defined as the number of episodes of exercise per week (none, 1–3 times, 4–7 times a week). P values <0.05 were considered statistically significant. All analyses were performed using Stata Statistical Software Release 16 (StataCorp., College Station, TX, USA).

Of 1,511 participants, the mean age was 60.5 years and 44.1% were men. The prevalence of hypertension and diabetes were 21.3% and 16.7%, respectively. The mean baseline eGFR was 80.3 mL/min/1.73 m², and 5.1% of participants had eGFR <60 mL/min/1.73 m². The prevalence of proteinuria was 1.3%. The median AhRL, MIS-ATP, and MIS-ROS were 2.3 pM TCDDeq, 89.0% of control, and 114.0% of control, respectively. During the mean 6.9-year follow-up, the mean yearly eGFR change was −1.6 mL/min/1.73 m²/year, and 84 participants (5.6%) experienced a rapid decline in kidney function. The baseline characteristics of the study population according to the MIS-ATP quartile are depicted in Table 1. In the lowest quartile group of MIS-ATP, patients were older and had more deleterious metabolic parameters, such as high BP, high level of fasting glucose/HbA1c and fasting insulin, high BMI and large WC, and high level of HOMA-IR and hsCRP (Table 1). In contrast, eGFR was higher in those with higher MIS-ATP levels. Persons with lower MIS-ATP levels showed higher levels of AhRL and MIS-ROS.

Yearly changes of eGFR showed positive correlation with MIS-ATP level, while AhRL and MIS-ROS levels were inversely correlated with yearly eGFR changes (Table 2). Multivariate analyses with adjustment for age, sex, and metabolic parameters, also showed significant correlation between MIS-ATP and yearly change of eGFR (Model 2). However, multivariate analyses with additional adjustment for baseline eGFR and the presence of proteinuria did not show significant correlations between the yearly change of eGFR and MIS-ATP (Model 3). MIS-ROS showed inversed correlation with MIS-ATP, in model 1 (univariate) and model 2 (adjustment for age, sex, and metabolic parameters). The correlation between AhRL level and the yearly change of eGFR was significant only in a univariate analysis, whereas it was not significant in multivariate analyses.

MIS-ATP was associated with decreasing risk of rapidly declining eGFR, defined of decreasing of eGFR more than 4 mL/min/1.73 m²/year (Table 3). Multivariate analyses with adjustments for anthropometric or metabolic parameters also showed significant beneficial effect for risk of rapid declining eGFR. In contrast, MIS-ROS showed significant risk effect, and multivariate analyses also showed significant results. AhRL showed an increasing risk effect for rapidly decreasing eGFR in univariate analysis and multivariate analysis with age and sex as covariates, but there were no significant effect in multivariate analyses with covariates, including metabolic or renal parameters.

Fig. 1 shows risk of MIS-ATP, AhRL and MIS-ROS in each subgroup divided by age, sex, the presence of hypertension and diabetes, and baseline eGFR. Higher MIS-ATP showed significant beneficial effect for risk of rapidly declining eGFR in subgroups of older age, male sex, the absence of hypertension and diabetes, and lower baseline eGFR <80.3 mL/min/1.73 m². Among the factors, lower eGFR had significant interaction term with MIS-ATP for rapid eGFR decline (P for interaction = 0.032). The risk effect of MIS-ROS for rapid eGFR decline was significant in the subgroups of old age, male sex, the presence of hypertension and the absence of diabetes, and lower baseline eGFR. Among them, hypertension showed significant interaction term with MIS-ROS (P for interaction =0.021). The risk of AhRL on rapid eGFR decline were not different between subgroups.