A Longitudinal Analysis On The Relationship Between Dynamic Changes Of Serum ALT,AST Levels And Metabolic Syndrome

Objective: At present,the prevalence of metabolic syndrome(Met S)is increasing year by year,which has become a public health problem.With the thorough knowledge of Met S,many studies showed the association between liver enzymes and the incident risk of Met S.Our study aim to analyze the impact of dynamic changes in alanine aminotransferase(ALT)and aspartate aminotransferase(AST)levels on the incident risk of Met S accurately.Methods: The health check-up data during March 2010 to April 2016 was collected from the database of a hospital medical examination center of Dalian and then established a seven-years dynamic physical examination database.The Met S was defined according to the Joint Interim Statement criteria 2009.The participants were divided into two groups,subjects who occurred Met S and without Met S,based on whether they developed Met S during the follow-up period.The basic descriptive analysis were used SPSS17.0.The Joint model and the Cox model were fitted by the R3.2.3 package JM.The optimal models which included all statistical significant independent variables were selected by the minimum akaike information criterion(AIC)value.Results: 1.A total of 4491 subjects(1497 male and 2994 female,median baseline age: 37 years,range: 19 to 80 years)who did not have Met S at their baseline visit were enrolled in this cohort study.In the follow-up periods,833 incident cases were developed Met S,the incidence of Met S was 18.55%(the incidence density was 7.45 per 100 person-years)and was 30.39%(14.32 per 100 person-years)in males and 12.63%(4.73 per 100 person-years)in females.The incidence of Met S in male was higher than that in female(P<0.05).From baseline to follow-up,the average levels and the abnormal rates of ALT and AST in Met S group were higher than non-Met S group(ALT:Z:-21.287,P=0.000;AST:Z:-11.764,P=0.000).The abnormal rates of ALT and AST in Met S group were higher than that in non-Met S group,regardless of their gender.The annual average rates of change in the two liver markers in Met S group were higher than that in non-Met S group(ALT:Z:-7.826,P=0.000;AST:Z:-6.256,P=0.000).2.In the Cox model,a unit increased in LNALT0 represents a 1.751-fold(95% CI: 1.532,2.000)increase in the relative risk of Met S.However,the independent variable LNAST0 did not remained in the final model because its impact on Met S was no statistically significant(P>0.05).3.In the Joint model established by LNALT,the LME sub-model showed that the average level of LNALT in the female who aged 37 years(median years)was 2.775U/L.For male of the same age,the average level of LNALT was 0.403U/L higher than female.The LNALT level for female was increased 0.010U/L for every 1 year added to baseline age,however,the increased rate of LNALT with baseline age in male was lower than that in female.The follow-up time was increased by 1 year,the LNALT level was increased 0.008U/L for 37 years old female,at baseline.The negative interaction between follow-up time and gender indicated that the LNALT level increased with time was faster in female.The Cox sub-model showed that the increased RR for developing Met S was 1.017(95% CI: 1.012,1.023)for every 1 year added to baseline age.The male had a 1.974-fold(95% CI: 1.671,2.332)risk of Met S relative to female.The association coefficient ? was statistically significant(P<0.05),which indicated 1 unit longitudinal increase in LNALT represented as 3.626-fold(95% CI: 2.721,4.831)increase in the RR of Met S.4.In the Joint model constructed by LNAST,the LME sub-model showed that the average level of LNAST in the female who aged 37 years was 2.962U/L.For male of the same age,the average level of LNAST was 0.135U/L higher than female.The LNAST level for female was increased 0.010U/L for every 1 year added to baseline age,however,the increased rate of LNAST with baseline age in male was lower than that in female.The follow-up time was no statistically significant and not remained in the final sub-model.The Cox sub-model showed the effects of baseline age and gender on Met S were similar to the results in the model with LNALT.However,the association coefficient ? was no statistically significant(P>0.05),which represented the effect of dynamic changes in LNAST level on Met S was not find in this model temporarily.Conclusion: In conclusion,the longitudinal increment of individuals' ALT level over time increased the incidence risk of Met S,even the level was in the normal reference range(?40U/L).However,the impact of dynamic changes of individuals' AST level on Met S was not find and it remains to be further studied.