Study On The Associations Between BMI And Insulin Resistance In Non-diabetic Population

Aims:Insulin resistance(IR)is defined as a disease,which cells could not respond to insulin normally,resulting in hyperinsulinemia.IR becomes a serious public health problem,which is a major risk factor for many diseases associated with obesity.Most studies focused on the association between IR and obesity based on the general population and diabetic population,whereas few studies based on non-diabetic population.The study focused on the association between Body Mass Index(BMI)and IR of the population in non-diabetes to provide theoretical justification for the prevention of the IR in the target population.Methods:This study is divided into two parts,the first part of the subjects comes from National Health and Nutrition Examination Survey(NHANES)cycles in the years2007～2016.Of these,this study investigated the association between BMI and IR in subjects aged≥20 years old and BMI change and IR in subjects aged≥50 years old.BMI changes are divided into three time periods,including the BMI in 25 years old compared to current,10 years ago and 1 years ago,baseline is BMI at age 25 years old;age is divided into three categories including 50～59 years,60～69 years,70～80 years;the four categories of BMI changes including normal or overweight to obesity,overweight or obesity throughout,normal to overweight and normal throughout.IR is evaluated by the homeostasis model assessment for IR(HOMA-IR).Under complex weighting conditions,continuous variables of non-normal distribution are expressed by median and inter-quartile range and are compared using Wilcoxon rank sum test.The categorical variables are expressed by frequency and composition ratio,and are compared usingχ~2 test.Statistical analyses were performed by R 3.6.1 and SPSS 24.0.The second part of the single nucleotide polymorphisms(SNP)associated with BMI are published in 2015 of journal of Nature including 322,154 participants from European and the SNPs associated with HOMA-IR are published in 2011 of journal of Nature Genetics including 46,186 participants from European are used for Two-sample Mendelian randomization(TSMR)analysis.The methods including inverse-variance weighted(IVW),Mendelian Randomization-Egger(MR-Egger)regression and Weighted Median Estimator(WME)and are implemented by the MR-base platform.Results:1.In population≥20 years old,compared normal BMI,the odds ratio(OR)value and CI in obesity and overweight population are 12.624(9.678,16.467)and 3.183(2.364,4.286),respectively.2.68 SNPs associated with BMI and HOMA-IR are found by TSMR.IVW method shows the evidence of causal association of BMI and HOMA-IR(β=0.184,standard error=0.022,P<0.001).MR-Egger regression and WME also provide evidence for causal association of BMI and HOMA-IR.3.The effect of BMI changes on IR at 25 years old and current,for the 50～59years group,normal or overweight to obesity,overweight or obesity throughout and normal to overweight are 8.894 times,3.742 times and 2.674 times compared with normal throughout;for the 60～69 years group,the above three ways of BMI change are 26.189 times,9.632 times and 8.886 times,respectively;for the 70～80 years group,the above three ways of BMI change are 6.765 times,4.545 times and 3.387times,respectively.The BMI in 25 years old compared with 1 years ago and 10 years ago have similar outcome.Conclusions:1.The risk of IR in obesity population is higher than that of the overweight population compared with population with normal BMI.The TSMR supports the causal association between BMI and HOMA-IR;2.The BMI change of 25 years old and current has the highest risk of IR;3.The 60～69 years group has the highest risk of IR;4.BMI change from normal or overweight to obesity has the highest risk of IR.