Abstract: Objective To analyze the correlation between longitudinal dynamic changes of uric acid and gout in male and female populations so as to provide theoretical basis for the early prevention and control of gout. Methods The physical examination data of healthy people from a hospital in Xian, Shannxi Province during 2007 and 2014 were collected. The joint model including linear mixed effects model and Cox proportional risk model was adopted to explore the relationship between uric acid level and gout based on the correlation parameter α, which was run by R 3.6.2 software. Then, based on the normal range of uric acid, grades were classified according to the number of observations. Cox proportional risk model was used to analyze the effects of dynamic changes of uric acid on the incidence of gout. Results A total of 1,144 subjects were enrolled in our study. During the follow-up, 102 individuals developed gout, with 76 males(34.21/10,000/year), and 26 females(16.15/10,000/year). The joint model showed that longitudinal changes in uric acid level were associated with gout incidence in male population(HR=1.014 9, 95%CI: 1.010 6-1.019 2, P<0.001), but there was no correlation between longitudinal changes in uric acid and gout in female population(P>0.05). The Cox model showed that compared with males with normal uric acid level, those with uric acid level returned to normal had the risk of gout increased by 4.458 0-fold(HR=4.458 0, 95%CI: 2.528 0-7.862 0), and by 4.679 0-fold(HR=4.679 0, 95%CI:2.716 6-8.058 0)after uric acid fluctuations. Conclusion The longitudinal fluctuations of uric acid level in healthy male adults could increase the risk of gout. Due attention should be paid to the dynamic changes of uric acid level to reduce the risk of gout.

Key words: Joint model, Uric acid, Gout, Longitudinal data, Survival analysis