Participants
We initially included 310,577 Japanese men and women aged 40-64 years whose health had been periodically assessed by the All Japan Labor Welfare Foundation (Tokyo), between April 2013 and March 2014. Among 310,498 individuals who consented to participate in this study, we excluded 205,730 with missing data and 7,905 who were underweight (BMI <18.5 kg/m2). Thus, we analyzed data from 96,863 participants (69,241 men and 27,622 women).
Written informed consent was obtained from the included individuals to participate in the study and to publish their innominate data. The Medical Ethics Committee at Showa University School of Medicine (Approval No. 2132) and the Ethics Committee at the All Japan Labor Welfare Foundation (Approval No. 3-1-0004) approved the study protocol.
Variables and measurements
We collected the following information from each participant using a self-administered questionnaire recommended by the Japanese Ministry of Health, Labour and Welfare for specific health examinations [7]: age, sex, alcohol consumption (daily, occasionally, none), smoking status (current, previously smoked, never-smoked), and physical activity equivalent to walking at least 60 min per day (yes/no). Alcohol intake was classified as yes (daily and occasional consumption) and no (no alcohol consumption) [8].
Trained staff measured the height and weight of the participants in increments of 0.1 cm and 0.1 kg, respectively, and BMI was calculated as body weight (kg) divided by height squared (m2). Waist circumference (WC) was measured to the nearest 0.1 cm at the level of the umbilicus while standing upright [6]. The WHtR was calculated as WC divided by height. Blood pressure while seated was measured using an HEM-907 automated device (Omron Corporation, Kyoto, Japan).
In accordance with a previous study [9], we defined OB and NW defined as BMI ≥25 and 18.5-24.9 kg/m2, respectively. In addition, the presence and absence of CO were determined as WHtR ≥0.5 and <0.5, respectively [10]. Based on OB and CO status, participants in this study were classified as being of normal weight with (NWCO) and without (NW) CO, and as being obese with (OBCO) and without (OB) central obesity [11].
Venous blood samples with drawn from participants to determine serum values of uric acid, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), triglyceride, blood glucose, and hemoglobin A1c (HbA1c) were stored at 4 ℃, transported to and analyzed at a clinical testing laboratory (SRL Inc., Tokyo, Japan) within 24 hours. Serum uric acid was measured using an enzymatic method (AU5400; Beckman Coulter, Brea, CA, USA). Both HDL-C and LDL-C were determined using a direct method and triglycerides were measured using an enzymatic method (AU5400; Beckman Coulter). Blood glucose values were determined using the hexokinase method (AU5400; Beckman Coulter), and HbA1c was measured using a latex agglutination method (JCA-BM9130; JEOL, Tokyo, Japan).
Hyperuricemia was defined as serum uric acid >7.0 mg/dL in men or ≥6.0 mg/dL in women, or being under medical treatment for hyperuricemia [12-15]. The definition in men was based on Japanese guidelines for the management of hyperuricemia [12]. Serum uric acid levels are lower in women than in men because female hormones decrease them [16, 17]. Thus, we defined hyperuricemia in women as ≥6.0 mg/dL [13-15]. Hypertension was defined as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or under medication for hypertension [18]. Dyslipidemia was defined as LDL-C ≥140 mg/dL, HDL-C <40 mg/dL, triglyceride ≥150 mg/dL, or under medication for dyslipidemia [19]. Diabetes was defined as fasting plasma glucose (≥8 hours after the last caloric intake) ≥126 mg/dL, random plasma glucose ≥200 mg/dL, HbA1c (National Glycohemoglobin Standardization Program) ≥6.5%, or under medication for diabetes mellitus [20].
Statistical analysis
Data for males and females were separately analyzed, because the serum uric acid distribution differed between them. Characteristics were compared between participants with and without hyperuricemia using unpaired t-tests or chi-squared tests. Odds ratios (OR) and 95% confidence intervals (CI) for hyperuricemia were calculated using a logistic regression model that included age, smoking status, physical activity, hypertension, dyslipidemia, and diabetes to control for potential confounding factors [11, 15]. Values with P <0.05 were considered statistically significant. All data were analyzed using JMP version 13.0 (SAS Institute Japan Co., Ltd., Tokyo, Japan).