The Kailuan Study is an ongoing prospective community-based cohort study conducted in Tangshan, China. All participants in the Kailuan Study are employees and retirees of the Kailuan Group. Details of the study design and procedure have been described elsewhere [9]. At baseline, 101,510 participants (81,110 males and 20,400 females; aged 18-98 years) were recruited, underwent clinical and laboratory examinations, and completed a questionnaire interview (June 2006 to October 2007) at 11 hospitals affiliated with the Kailuan Group. Subsequent examinations involving anthropometric, cardiovascular risk factor measures, and self-reported questionnaires (including income, educational level, drinking and so on) occurred approximately biennially.

We excluded participants without baseline information on WHtR or hs-crp (n=4193), and under the following conditions including hs-crp>10 mg/L (n=4020), alcohol consumption more than 70g/week for females or 140g/week for males (n=16660), and missing data on ultrasound examination (n=255). Additionally, participants were excluded if a history of NAFLD or other known history of chronic liver disease such as autoimmune hepatitis or viral hepatitis (HBsAg positive or anti-HCV antibody positive, etc.) or those using hepatotoxic drugs (n=18081), a history of cancer and cardiovascular disease (n=1813) or failed to finish a series of assessments(n=5492). Finally, a total of 50776 individuals were enrolled in the present study (Figure 1).

Fig. 1.

Flow chart of participants

(0.27MB).

The study was performed according to the guidelines of the Helsinki Declaration and was approved by the Ethics Committee of Kailuan General Hospital (approval number: 2006-05). All participants were agreed to take part in the study and provided informed written consent.

During the visits, anthropometric measurements including height, weight, waist circumference (WC), and blood pressure (BP) were recorded according to a standard protocol. Height and WC were averaged to 0.1 cm and weight was averaged to 0.1 kg. BMI is calculated as weight (kg) divided by height squared (m2). WHtR is calculated by dividing each WC by the participant's height. In accordance with World Health Organization classification, abdominal obesity was defined as WHtR ≥0.5 and non-abdominal obesity was defined as WHtR <0.5 [10].

Blood samples from antecubital vein were obtained in the morning after a night of fasting and then transfused into vacutainers containing ethylene diamine tetraacetic acid (EDTA). Samples were centrifuged at 3,000 g for 10 min at room temperature. After separation, the blood samples were immediately frozen at -80 °C for storage for further laboratory examinations. The level of hs-crp was measured using a commercial, high-sensitivity nephelometry assay (Cias Latex CRP-H, Kanto Chemical Co. Inc), with a detection limit of 0.1 mg/L. In-house intra- and inter-assay CVs for hs-crp were 6.53% and 4.78%, respectively. All the blood variables were measured using an autoanalyzer (Hitachi 747; Hitachi) at the central laboratory of the Kailuan hospital (Tangshan, China). According to baseline hs-crp levels and American Heart Association cardiovascular disease risk cutpoints: hs-crp <1 mg/L, low; hs-crp 1–3 mg/L, moderate; and hs-crp ≥3 mg/L, elevated [11]. And hs-crp ≥3 mg/L was considered as chronic inflammation.

Due to the limitation of research conditions at that time, serum insulin level was not directly collected and HOMA-IR could not be evaluated. Insulin sensitivity was evaluated using the triglyceride-glucose index (TyG index). And the TyG index was calculated as ln (fasting triglyceride [mg/dL] × fasting blood glucose [mg/dL]/2).

Follow-up ended at the first record of NAFLD event, all-cause death, at the last date of follow up or at the end of follow-up on 31 December 2013, whichever came first. Diagnosis of NAFLD was based on the presence of at least two of the following three abnormal findings after excluding the individuals with excess alcohol intake(more than 70 g/week for females or 140g/week for males) and autoimmune hepatitis or viral hepatitis according to the National Workshop on Fatty Liver and Alcoholic Liver Disease, Chinese Society of Hepatology: (1) diffusely increased echogenicity of the liver relative to the kidney; (2) ultrasound beam attenuation; (3) poor visualization of intrahepatic structures [12]. Abdominal ultrasonography was performed by using a high-resolution B-mode topographical ultrasound system with a 3.5 MHz probe (PHILIPS HD-15, US).

Participants were classified as “current” or “never” according to whether they smoked at least one cigarette weekly for over one year. In terms of physical activity, participants were divided based on whether they took activity more than three times a week with each time lasting more than 30 min or not. Education level and general information of the participants were collected from questionnaires as baseline in 2006. Diabetes mellitus was defined based on a fasting blood glucose value ≥7.0 mmol/L or the participant had previous medical treatment for diabetes or a history of diabetes. Hypertension was defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg or the participant had previous administration of antihypertensive medications or a history of hypertension.

Continuous variables were compared using analysis of variance or the Kruskal-Wallis test according to distribution, and categorical variables were compared with the chi-square test. A correlation between WHtR and hs-crp was analyzed by Pearson's chi-squared test.

To investigate the relationship of abdominal obesity and chronic inflammation with NAFLD, the participants were stratified into groups based on WHtR (<0.5 and >0.5) and hs-crp levels (<1 mg/L, 1-3 mg/L, ≥3 mg/L). In the analysis of the combined effect, all subjects were assigned into four groups: non-abdominal obesity and non-chronic inflammation (G1), abdominal obesity and non-chronic inflammation (G2), non-abdominal obesity and chronic inflammation (G3), abdominal obesity and chronic inflammation (G4). G1 was used as a control group. The incidence of NAFLD was calculated. Hazard ratios (HR) and 95% confidence intervals (CI) of WHtR, hs-crp alone, and their combination on NAFLD were analyzed by the Cox proportional hazard model. And the Cox models and found that the proportional hazards assumption was satisfied. Finally, exposure factors were introduced into the multivariate model as an interaction term to test the interaction between abdominal obesity and chronic inflammation [13]. We calculated the relative excess risk due to interaction (RERI), proportion of disease attributable to interaction (AP) and synergy index for interaction (SI) [9,14-15]. Interaction was considered significant when the 95%CI of RERI, AP and SI did not comprise 0.0 and 1, separately. In the Cox and interaction models, the adjusted variables included age, gender, BMI, TyG index, smoke (never/current), physical activity (yes/no), hypertension (yes/no), diabetes mellitus (yes/no), high school or above (yes/no). Subgroup analyses were performed based on age, sex, BMI and diabetes mellitus. All analyses were done with SAS (version 9.4), at a two-tailed alpha level of 0.05.

Among the 50776 subjects included in the statistical analysis, there were 38169 men and 12607 women, with a mean age of 49.40±12.57 years, a mean WHtR of 0.51±0.06, and a median hs-crp of 0.69 (0.25-1.72) mg/L. The baseline characteristics of the participants stratified by different abdominal obesity status and chronic inflammation status were shown in Table 1. The correlation coefficient between WHtR and hs-crp was r=0.175.

The mean follow-up period was 5.59±1.79 years involving 15451 cases (including 11088 men and 4363 women). The overall incidence of NAFLD was 54.46 per 1000 person-years (for male and for female). The G4 group had highest incidence of NAFLD (table 2).

As shown in Table 2, the results of multivariate Cox proportional hazard analysis indicated the risk of NAFLD. Compared with those with hs-crp<1 mg/L group, risk of NAFLD was significantly higher in those with hs-crp 1-3 mg/L group (HR 1.14, 95% CI 1.10-1.18) and hs-crp ≥3 mg/L group (HR 1.26, 95% CI 1.21-1.32). After correction for the same confounding factors, the risk of NAFLD in the WhtR>0.5 group (HR 1.31, 95% CI 1.26-1.37) was higher, compared with those with WhtR≤0.5 group.

The participants were classified into four groups based on the levels of WHtR and hs-crp. The result of the multivariate Cox proportional hazard analysis indicated that after the adjustments of confounding factors, the HR in the G4 group (HR 1.60, 95% CI 1.52-1.70) was increased compared with that in the G1 group. In addition, the HR in the G4 group was higher than that in the G2 (HR 1.21, 95%CI 1.11-1.32) group and the G3 (HR 1.32, 95%CI 1.27-1.38) group (Table 2).

Regarding previous study on factors associated with NAFLD, age, sex, obesity which is usually presented as BMI index and diabetes mellitus were considered as potential factors. We therefore performed four subgroup analyses to observe the effect of factors on development of NAFLD. As shown in Table 3, the result of the multivariate Cox proportional hazard model indicated that the risk in the G4 group in the male and female was increased 1.45 times (95% CI 1.36-1.55) and 1.94 times (95% CI 1.75-2.16) compared with the G1 group, respectively. The results of the remaining three subgroup analyses are similar.

After the adjustment conducted for other factors, the following three measured values of additive interaction were all significant: RERI 0.28, 95% CI 0.19-0.37; AP 0.17, 95% CI 0.12-0.22; SI 1.86, 95% CI 1.44-2.39. In addition, the multiplicative interaction was statistically significant (P <0.05). All indicated a significant synergistic interaction between WHtR and hs-crp on the risk of NAFLD.