This population-based, cross-sectional study was conducted in Community Health Centers of Songjiang District, Shanghai, China during June 2016 and December 2017. Details of sample methods in this study have been described elsewhere . Briefly, we used a multistage, stratified, clustered sampling method to collect health related data from 31 neighborhood committees and 16 administrative villages in 4 study community sites, including Zhongshan, Xinqiao, Sheshan, and Maogang. Exclusion criteria were as follows: unable or unwilling to provide a written informed consent form; pregnancy; previously diagnosed critical illness, including cancer, stroke, coronary artery diseases, cirrhosis, chronic hepatitis, cardiorespiratory failure, and hyper-or hypothyroidism; or have got organ transplantation or on dialysis therapy. For the present analysis, a total of 37,670 adults aged 20 to 74 years who were natives of Shanghai municipality or those have lived in Shanghai for at least 5 years were enrolled in the present study. Among these, we excluded participants who violated the inclusion criteria (n=4,271), and who had no serum creatinine (Scr) measurement (n=264) or miss data for a physical examination or a questionnaire survey, or laboratory measurements (n=1,839). The final analysis included 31,296 participants (Fig. 1). The study protocol was approved by the Ethics Committee of the Fudan University, School of Public Health (IRB#2016-04-0586) and complied with the principles of the Declaration of Helsinki. Informed written consents were obtained from all participants before data collection.
Trained interviewers used a structured questionnaire through face-to-face interviews or extracting from electronic medical records to obtain data on all study participants. Information about sociodemographic characteristics (age, sex, marital status, educational level, and working status), medical histories (such as self-reported history of T2DM, hypertension, and cancer), and lifestyle factors (smoking status, alcohol consumption, and physical activity) was collected. We used the International Physical Activity Questionnaire to assess the physical activity. Smoking status was defined as >1 cigarette per day and lasting >6 months, and alcohol consumption was defined as alcohol intake at least 3 times per week and lasting for at least half a year. Smoking status and alcohol consumption have been classified as never, former, or current.
Anthropometrical data were obtained from all participants, including height, weight, and waist circumference (WC), and were measured in duplicate when participants were wearing light clothing without shoes. The mean values of these measurements were calculated for further analysis. Body mass index (BMI) was defined as body weight in kilograms divided by height in meters squared (kg/m2). Blood pressure (BP) was consecutively measured three times using an electronic sphygmomanometer, and the mean values were used for analysis.
Prior to the investigation, all participants were asked to fasting overnight for at least 8 hours, and fasting venous blood specimens were collected to perform laboratory measurements in DiAn medical laboratory center. Serum total cholesterol, TG, high-density lipoprotein (HDL) cholesterol, and low-density lipoprotein (LDL) cholesterol levels were measured using an automatic biochemical analyzer (Roche Cobas C501). Fasting plasma glucose (FPG) level was measured by glycokinase method using Roche P800 biochemical analyzer. Scr level was measured using enzymatic methods by Roche C702 automatic biochemical analyzer. HbA1c level was determined using high pressure liquid chromatography (TOSOH G8 automatic biochemical analyzer).
Kidney function assessment
The estimated glomerular fltration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation for Chinese population :
where Scr is the serum creatinine (mg/dl), k is 0.7 for females and 0.9 for males, α is -0.329 for females and -0.411 for males, min indicates the minimum of Scr/k or 1, and max indicates the maximum of Scr/k or 1.
Decreased eGFR was defined as an eGFR value below 60 mL/min/1.73 m2. The CKD classification was in accordance with the National Kidney Foundation , and we classified GFR stages into 4 categories as follows: normal eGFR, ≥90mL/min/1.73 m2; mildly decreased eGFR, 60-89 mL/min/1.73 m2; moderately decreased eGFR, 30-59 mL/min/1.73 m2; and severely decreased eGFR, 15-29 mL/min/1.73 m2. In addition, eGFR was evaluated using the Modification of Diet in Renal Disease (MDRD) Study Equation for Chinese population  in a sensitivity analysis:
Definitions of triglyceride waist phenotype, T2DM, and hypertension
Participants were classified into four groups according to the following cut-off points : (1) NTNW, normal serum TG level (<1.7 mmol/L) and normal WC (<90 cm for men and <80 cm for women); (2) NTGW, normal serum TG level and enlarged WC (≥90 cm for men and ≥80 cm for women); (3) HTNW, elevated serum TG level (≥1.7 mmol/L) and normal WC; (4) HTGW, elevated serum TG level and enlarged WC. In the sensitivity analyses, HTGW phenotype was also defined as elevated serum TG level (≥2.0 mmol/L) along with enlarged WC (≥90 cm for men and ≥85 cm for women) , or a TG level ≥2.0 mmol/L and WC ≥90 cm for men or a TG level ≥1.5 mmol/L and WC ≥85 cm for women .
The definition of T2DM was in accordance with the American Diabetes Association criteria : FPG level ≥7.0 mmol/L, HbA1c concentration ≥6.5%, or previously diagnosed type 2 diabetes mellitus (T2DM). The diagnosis of hypertension was according to the Seventh Joint National Committee Report on Detection, Evaluation, and Treatment of High Blood Pressure guidelines (JNC-7) : systolic BP ≥140 mmHg, diastolic BP ≥90mmHg, or previously diagnosed hypertension.
We accounted for a complex sample survey design, and the results were weighted in the present study. Continuous variables were presented as means ± standard deviation (SD) or median with interquartile ranges. Categorical variables were expressed as number and percentage. We compared the differences between decreased eGFR and non-decreased eGFR using student’s t test or Mann-Whitney U test for comparisons of continuous data and c2 test for categorical data. We used weighted logistic regression models to determine the association of triglyceride waist phenotypes and decreased eGFR, and odd ratios (ORs) and 95% confidence intervals (CIs) were calculated, with NTNW as the reference group. Multiple models were adjusted for age, sex (men vs. women), marital status (married vs. unmarried/divorced/widowed), educational level (0-6, 7-12, and >12 years), working status (retired vs. not retired), smoking status (never, former, and current), alcohol consumption (never, former, and current), physical activity, BMI, systolic BP, HDL cholesterol, and FPG.
In the sensitivity analyses, we repeated the analyses and estimated GFR by the use of the MDRD equation or redefined the triglyceride waist phenotypes on the basis of the previous recommended criteria. We performed stratified analyses and potential effect modifications by sex, age (<60 years, ≥60 years), BMI (<24 kg/m2, ≥24 kg/m2), and presence or absence of T2DM or hypertension. In addition, we investigated the associations of four phenotype groups with the severity of decreased eGFR, including mildly, moderately, and severely decreased eGFR with using weighted multinomial logistic regression models, respectively by treating the normal eGFR group as the control group, and the same confounding factors as above were adjusted for the analyses.
All analyses were performed using SAS 9.4 version (Institute Inc., Cary, NC, USA). P values of less than 0.05 (two-sided) were considered to be statistically significance.