Association Of Empirical Dietary Index For Insulin Resistance And Odds Of Obesity And Weight Change Among Adults: A Crosssectional Study

ObjectiveWe developed and evaluated empirical dietary index for insulin resistance（EDIR）to verify the hypothesis of disease mechanism,and assess the insulin resistance potential of usual diets and lifestyles,using reduced rank regression.We investigated the association between adherence to EDIR and risk of obesity and weight change.We examined whether diet could influence the risk of obesity and weight change through modulates insulin resistance at the population level.MethodsDietary data are obtained in Lu’an City,Anhui Province cross-sectional study with a validated semi-quantitative FFQ that assessed diet intake during the previous 1 year.We entered forty FFQ-derived food groups in reduced rank regression,and defined most predictive of the ratio of triglyceride and high-density lipoprotein cholesterol（TG:HDLC）as insulin resistance indices.EDIR sore derived by stepwise linear regression models in the training set of 2137 subjects age≥18 years.This score was further validated in a validation set of 300 participants.We evaluated the validity of insulin resistance indices using multivariable-adjusted linear regression analyses to calculate relative concentrations of biomarkers.We used the non-conditional logistic regression model to estimate the Odds ratios（OR）and 95% CI of obesity and weight change associated with EDIR score.The nonlinear relationship of EDIR score with risk of obesity and weight change was assessed using restricted cubic spline regression.ResultsA total of nine food intake components were included in EDIR,including four proinsulin resistance foods（i.e.red meat,refined grains,other aquatic products and other vegetables）,which were positively associated with insulin resistance levels.Five antiinsulin resistance foods（i.e.pickles,rice wine,dark yellow vegetables,green vegetables and condiments）,which were inversely associated with insulin resistance levels.In multivariate-adjusted models in validation set,the EDIR were significantly associated with TG:HDL-C.The TG:HDL-C concentration in the highest quintile of the EDIR was0.42（95%CI: 0.12,0.95,Ptrend<0.001）.After adjusting for potential confounders,the multivariate-adjusted ORs（95% CIs）for participants in the highest quintile of EDIR,as compared with those in the lowest quintile,were 1.31（95%CI 1.00-1.71,Ptrend=0.018）for center obesity defined by WHR.We did not find association between adherence to EDIR and general obesity（OR=0.78,95%CI 0.55-1.09,Ptrend=0.100）and weight change（OR=0.95,95%CI 0.76-1.19,Ptrend=0.733）.The association between EDIR score and risk of center obesity appeared stronger in male（Per 1-SD）,compared with female（male and female,Pinteraction <0.002）.Restricted multivariate cubic spline analyses showed non-significantly linear or nonlinear associations between EDIR scores and risk of obesity and weight change.ConclusionEDIR was novel hypothesis-driven empirically dietary pattern,derived dietary indices assess dietary quality on the basis of insulin resistance potential.EDIR robust associations with the insulin resistance response biomarkers,which suggest usefulness in assessing the ability of whole diets and sustain insulin secretion.EDIR plays an important role in the validation of etiological hypotheses in nutritional epidemiology and the dietary prevention and treatment of diseases.Diets with a higher insulin resistance potential were associated with elevated risk of center obesity.These findings provide strong evidence that insulin resistance may partially mediate the association between diet and center obesity risk.Guidelines and interventions highlighting the importance of reducing or avoiding insulin resistance diet may therefore promote health and longevity.