[A Cohort Study on the Correlation Between Body Mass Index Trajectories and New-Onset Non-Alcoholic Fatty Liver Disease].

S Wang,J Zhang,H L Zhang,G L Zhu,Y,S L Wu

DOI: https://doi.org/10.3760/cma.j.cn501113-20190629-00230

2020-01-01

Abstract:Objective: To explore the correlation between the body mass index (BMI) trajectories and new-onset non-alcoholic fatty liver disease (NAFLD) so as to provide a scientific basis for the prevention and treatment of NAFLD. Methods: A total of 16388 observation subjects that met the inclusion criteria in the Kailuan study were used to form a cohort study. According to the BMI values of the observed subjects during annual physical examinations from 2006 to 2007, 2008 to 2009 and 2010 to 2011, SAS Proc Traj was used to determine four different BMI trajectories groups, namely, the low-stable medium-stable, medium-high and high-stable group. NAFLD incidence in each group was followed up during annual physical examinations from 2012 to 2013, 2014-2015 and 2016-2017. A total of 14998 observation subjects were finally included in the statistical analysis. The cumulative incidences of NAFLD differences in the four groups were compared. The Cox's proportional hazards regression model was used to analyze the correlation between different BMI trajectories and new-onset NAFLD. One-way analysis of variance was used to compare the intergroup difference of measurement data, and pairwise comparisons were conducted. LSD test was used for the homogeneity of variance. Dunnett's T3 test was used for heterogeneity of variances. χ (2) test was used to compare the count data, and the difference of NAFLD cumulative incidence rate between the different BMI trajectories groups was compared by log-rank test. Results: (1) the cumulative incidence of NAFLD was increased with the increase of BMI trajectories, which were 31%, 47%, 63%, 77%, respectively, and the difference was statistically significant (P < 0.01). (2) after adjusting for multiple confounding factors such as age and gender with the Cox's proportional hazards regression model, the risk of NAFLD in the BMI medium stable, medium-high, and high stable group was still 1.757 times [95% confidence interval (CI): 1.589 ~ 1.942], 2.612 (95%CI: 2.353 ~ 2.900), 3.566 (95%CI: 3.129 ~ 4.064) of the low-stable group (P < 0.01). Conclusion: The risk of NAFLD increases with increase of BMI trajectories, and long-term high levels of BMI are independent risk factors for the onset of NAFLD.