1. Introduction
Chronic kidney disease (CKD) is a worldwide health problem that leads to a higher risk of dialysis, hospitalization, cardiovascular morbidity, and mortality [
1,
2,
3]. The prevalence of CKD has been reported to be approximately 13% in the adult population [
4,
5]. There has been a recent dramatic increase in the prevalence of end-stage renal disease (ERSD), which is the final stage of CKD, because of the improvement in dialysis techniques and quality of medical care [
6]. The number of chronic dialysis patients has been increasing over the last three decades, and the number of new dialysis patients has continuously increased in Japan [
5]. Therefore, the latent CKD population appears to be enormous in Japan. Preventive plans mounted against CKDs will be essential to reduce the increase of patients with ESRD and consequent economic burdens. Recently, CKD has been considered as a risk factor for not only ERSD but also cardiovascular disease, even in the early stages of renal function [
7]. Hence, it is very important to identify the risk factors of CKD to promote its prevention and early detection. In addition to CKD, non-alcoholic fatty liver disease (NAFLD) is one of the most common chronic liver diseases, and it represents a significant health burden worldwide. The prevalence of NAFLD has been estimated to be between 20% and 30% in the general population [
8]. NAFLD includes a broad range of conditions, such as simple steatosis and non-alcoholic steatohepatitis, which could progress to cirrhosis and hepatocellular carcinoma. In addition, NAFLD is considered to be an independent risk factor for cardiovascular disease [
9,
10,
11]. NAFLD has been strongly associated with obesity and has been considered a phenotype of metabolic syndrome (Mets) in the liver [
12]. The components of Mets including obesity, diabetes, hypertension, and dyslipidemia are also important risk factors for the development and progression of CKD [
1,
13]. The possibility that CKD and NAFLD are linked has attracted attention in research interest since NAFLD is frequently accompanied by risk factors for CKD, such as insulin resistance, visceral obesity, type 2 diabetes, and Mets. Several studies have been reported that NAFLD increases the incidence of CKD [
14,
15,
16,
17,
18]. In contrast, a recent cohort study found no relationship between NAFLD and CKD [
19]. It is unclear whether the link between NAFLD and CKD is mediated by common risk factors, or whether NAFLD independently contributes to increasing the risk for CKD. Therefore, this study aimed to determine whether NAFLD itself is independently associated with CKD in a general population.
2. Materials and Methods
2.1. Study Design and Population
Our cross-sectional study used data from 4637 Japanese individuals (2461 men and 2176 women), who had consecutively undergone comprehensive annual examinations including ultrasonography for employee or resident health check-ups at Nara Health Promotion Center during 2012. Of these, 912 were excluded from the study because they were hepatitis B surface antigen- or anti-hepatitis C virus antibody-positive or consumed alcohol (men and women: >30 and >20 g/day of alcohol, respectively). The remaining 3725 subjects (1751 men, 1974 women; mean age: 53.4 ± 9.1 years, age range: 17–89 years) were included in this study.
This study was approved by the Ethics Committee of Nara Medical University (approval no. 1814-2). Informed consent was obtained in the form of opt-out from all participants prior to study.
2.2. Clinical and Laboratory Assessments
We used a standardized questionnaire to collect data including alcohol consumption, smoking history and status, and medical history. Smoking history and status were defined by using the Brinkman Index and answering the question whether you are a past or current smoker. The Brinkman Index measures the number of cigarettes smoked per day times smoking years. All subjects were examined following a 12-hour overnight fast. The body mass index (BMI) was calculated as weight (kg)/squared height (m2). Obesity was defined as a BMI ≥25 kg/m2 according to the criteria established by the Japan Society of Obesity. Resting blood pressure was measured using an automatic sphygmomanometer. Hypertension was defined as a systolic blood pressure (SBP) ≥140 mmHg, diastolic blood pressure (DBP) ≥90 mmHg, or current use of antihypertensive medications.
A high triglyceride level was defined as ≥150 mg/dL, and a high low-density lipoprotein (LDL) cholesterol level was defined as ≥140 mg/dL. Diabetes mellitus was defined as being present in patients on medication for diabetes mellitus and/or in those with a fasting blood glucose level ≥126 mg/dL. The glomerular filtration rate (GFR) was estimated from the Japanese Society of Nephrology CKD Practice Guide: estimated glomerular filtration rate (eGFR) (mL/min/1.73 m2) = 194 × (serum creatinine level (mg/dL)) − 1.094 × (age (y)) − 0.287. The product of this equation was multiplied by a correction factor of 0.739 for women. CKD was defined as an estimated GFR <60 mL/min/1.73 m2. Hyperuricemia was defined as ≥7 mg/dL.
Ultrasonography was performed with LOGIQ 7 using a 4 MHz convex array transducer (GE Healthcare, Waukesha, WI, USA). Fatty liver was defined by ultrasonography when there was hepatic parenchymal brightness and liver-to-kidney contrast. Since subjects were excluded from the study if they were hepatitis B surface antigen- or anti-hepatitis C virus antibody-positive, or consumed certain amounts of alcohol per day, those diagnosed as having fatty liver were considered to have NAFLD.
2.3. Statistical Analysis
We divided the subjects into two groups: subjects with NAFLD and those without NAFLD. There were potential confounding biases between the groups. In order to balance the bias, we calculated propensity scores (PSs) using the logistic regression model for the following covariates: sex; age; use of antihypertensive medications, antidiabetic medications, and lipid-lowering medications; history of cerebrovascular disease and heart disease; and smoking status (Brinkman Index). Subjects in both groups with similar PSs were matched in a 1:1 ratio using a caliper of 0.25 times the standard deviation of the PS. A standardized difference <0.2 was considered to indicate a negligible difference in the mean or prevalence of covariates between the groups.
Continuous variables were compared between the groups using the Student t-test, and categorical variables were compared using the chi-square test. Logistic regression analysis was used to assess the independent risk factors of CKD. The results are expressed as odds ratios with 95% confidence intervals. A p-value <0.05 was considered to be statistically significant. All calculations were performed using SPSS software (version 25, IBM Corp., Armonk, NY, USA).
4. Discussion
This study showed that NAFLD itself is not an independent risk factor for CKD. The comorbidities of NAFLD such as obesity, hypertension, and hyperuricemia are independently associated with CKD. Logistic regression analysis adjusted for age, sex, and the Brinkman Index demonstrated that obesity, hypertension, and hyperuricemia were independent risk factors for CKD, but NAFLD was not independently associated with CKD. In subjects with NAFLD, obesity, hypertension, and hyperuricemia were independent risk factors for CKD. In addition, the prevalence of CKD was not significantly different between subjects with and without NAFLD in the subgroup analyses stratified by age, sex, and subgroups of subjects with and without the presence of comorbidities including obesity, hypertension, high triglyceride level, high LDL cholesterol level, diabetes, and hyperuricemia.
In this study, obesity and hypertension were independent risk factors for CKD in all subjects, and in subjects with NAFLD. In subjects with NAFLD, the prevalence of CKD was significantly higher in obese subjects than in non-obese subjects, and in subjects with hypertension than in those without hypertension. Obesity is the most common risk factor for NAFLD. Features of Mets are not only highly prevalent in patients with NAFLD, but components of Mets including obesity and hypertension also increase the risk of developing NAFLD [
20]. Obesity is an independent risk factor for CKD, and it is associated with the development of proteinuria and pathologic findings of podocyte hypertrophy and focal segmental glomerular sclerosis even in the absence of hypertension. Hypertension has been defined as a risk factor for CKD. Systemic hypertension is transmitted to intraglomerular capillary pressure, which leads to glomerulosclerosis and loss of kidney function. Herein, hyperuricemia was also an independent risk factor for CKD in all subjects, and in subjects with NAFLD. Hyperuricemia leads to tubular injury, endothelial dysfunction, oxidative stress, and intrarenal inflammation. Additionally, when uric acid is present in the cytoplasm of cells or in the acidic/hydrophobic environment of atherosclerotic plaques, it is converted to a pro-oxidant, and it promotes oxidative stress, including cardiovascular disease through this mechanism, participating in the pathophysiology of human disease [
21]. Recent studies reported that hyperuricemia is an independent predictor of CKD [
22,
23]. Uric acid concentration is an independent risk factor for kidney failure in earlier stages of CKD and has a J-shaped relationship with all-cause mortality in CKD [
24]. A cross-sectional study reported that a higher serum uric acid level was associated with increased CKD prevalence in patients with hypertension [
25]. Furthermore, in a longitudinal study it was demonstrated that a high serum uric acid level increases the risk of obesity. Hyperuricemia is also significantly associated with and independently predicts the risk for NAFLD. The serum uric acid level was significantly associated with the degree of steatosis and severity of lobular inflammation in NAFLD [
26]. Hyperuricemia not only induces hepatic inflammation, but may also cause CKD in NAFLD. One of the most important causes of death in patients with NAFLD is the cardiovascular complications [
27]. As mentioned, CKD is also considered an important risk factor for cardiovascular disease even in the early stages of renal function [
7]. Therefore, NAFLD patients with hyperuricemia do not only have a high risk for CKD but also cardiovascular disease.
Diabetes is the leading cause of CKD. Among patients with type 2 diabetes mellitus who are initially free of proteinuria, the 20-year risk of diabetic nephropathy is 41% [
28]. In the present study, there was no association between diabetes and CKD in all subjects or subjects with NAFLD. The numbers of subjects with diabetes were only 172 (7.8%) among all subjects and 113 (10.3%) among subjects with NAFLD. Therefore, the sample size was insufficient to elucidate the association between CKD and diabetes in all subjects and subjects with NAFLD by the methods used herein.
Since CKD and NAFLD have similar risk factors, it is necessary to demonstrate the independent association between NAFLD and CKD. A cross-sectional study in the United States showed a significant, positive association between NAFLD and prevalent CKD in univariate analysis. However, this association was attenuated after adjustment for features of Mets [
29]. In another cross-sectional study in Taiwan, mild NAFLD was not significantly associated with CKD, but moderate to severe NAFLD was significantly associated with CKD after adjustment for age, sex, smoking, components of Mets, and the ALT level [
30]. Since there are numerous risk factors for developing CKD including genetics, race, sex, age, smoking, obesity, history of cardiovascular disease, hyperlipidemia, and Mets, it is difficult to adjust for all the factors that affect the prevalence of CKD. In our study, before PS matching, the eGFR was significantly lower, and the serum creatinine level and the prevalence of CKD were significantly higher in subjects with NAFLD than in those without NAFLD. However, those values were not significantly different between subjects with and without NAFLD after PS matching.
Our study has several limitations. First, this study had a cross-sectional design that did not identify whether NAFLD itself develops and progresses to CKD. Second, we used only the eGFR for identification of CKD. However, in clinical practice and epidemiological studies, the eGFR is the most common variable evaluated for CKD, and the eGFR is used for the classification of CKD. Third, we did not assess the relationship between the severity of NAFLD and CKD. Liver biopsy is the gold standard for the assessment of hepatic fibrosis and inflammation severity. However, it has several limitations including sampling variability, invasiveness, complications and cost. Thus, it is recommended for patients who would benefit the most from diagnosis, therapeutic guidance, and prognostic information [
20]. We could not perform liver biopsy because our sample size was large and the majority of the subjects were not eligible for liver biopsy. Qin et al. reported that liver stiffness assessed by transient elastography (TE) is a potential indicator of CKD in patients with NAFLD; such patients with a higher value of liver stiffness have an increased risk of CKD [
31]. Therefore, the risk of CKD may increase with progression in the stage of liver fibrosis. TE is a clinically useful tool for identifying advanced fibrosis in patients with NAFLD. However, it cannot effectively diagnose bridging fibrosis or cirrhosis; therefore, to guide clinical decision-making, liver biopsy must be performed [
32]. Recent studies reported that non-invasive fibrosis markers, the NAFLD fibrosis score (NFS) and fibrosis-4 (FIB-4) index are associated with the presence of CKD [
33,
34,
35]. NFS is based on six variables including age, BMI, hyperglycemia, platelet count, albumin and AST/ALT ratio. FIB-4 index is based on age, platelet count, AST, and ALT. Because age is an independent risk factor for CKD, it is a cofounding factor in the analysis using the NFS and the FIB-4 index. These observations suggest that it may be difficult to accurately examine the relationship between the fibrosis stage of NAFLD and CKD in large studies. Fourth, NAFLD was diagnosed by ultrasonography. Although ultrasonography is a common method for diagnosing hepatic steatosis in clinical practice, it has limited accuracy for the detection of mild steatosis. We did not assess the relationship between the grade of steatosis and CKD. The controlled attenuation parameter (CAP) using TE provides a standardized non-invasive measure of hepatic steatosis [
36] and is efficient to detect even low-grade steatosis [
37]. It may be a useful tool to assess the relationship between the severity of steatosis and CKD. Fifth, we did not exclude autoimmune liver diseases such as autoimmune hepatitis and primary biliary cholangitis. However, due to their low prevalence, this may not significantly affect the results of this study. Lastly, most subjects were asymptomatic and able to engage in their work. Therefore, this study’s results may not be applicable to individuals who have liver cirrhosis or ERSD.