| Objective: The aim of this study were to analyze the incidence and regression of non-alcoholic fatty liver disease(NAFLD)in the Jinchang cohort,to explore the factors affecting the incidence and regression of NAFLD,and establish the risk prediction model for the incidence and regression of NAFLD to identify high-risk groups for effective monitoring and timely intervention.Methods: Based on baseline and phase I and II follow-up data from the Jinchang cohort,this study used incidence and regression rates to describe NAFLD incidence and regression in different subgroups.The independent factors influencing the incidence and regression of NAFLD were studied by Cox proportional hazard model.The least absolute shrinkage and selection operator(LASSO)was used to screen the variables,and the prediction models for NAFLD incidence and regression risk were constructed using the Cox regression method in the training set,and the predictive value of the models were verified by the validation set.The prediction models were visualized using nomograms,and the area under the ROC curve and calibration curve were used to evaluate the discrimination and calibration of the models.Decision curve analysis(DCA)was used to evaluate the clinical benefit of the model,and a risk stratification system was later established based on the prediction scores of nomogram.Results: 1.A total of 22,702 study participants without NAFLD were included at baseline and followed for 58,477.49 person-years,during which 2,172 new cases of NAFLD occurred,with an incidence rate of 37.14/1000 person-years.The incidence of NAFLD was higher in the male population(43.38/1000 person-years)than in the female population(29.97/1000 person-years)(P<0.05).2.Multi-factor Cox regression analysis of NAFLD incidence showed that the risk of NAFLD incidence was increased by 37% in the age group 45-59 years than the age group 18-44 years(HR=1.30,95%CI: 1.14-1.49).The risk of NAFLD incidence was increased by 28%(HR=1.28,95%CI: 1.17-1.39)in those with monthly household income ≥2000 yuan compared to those with less than 2000 yuan.The risk of developing NAFLD was 2.46 times(HR=2.46,95%CI: 2.23-2.71),and 4.39 times(HR=4.39,95%CI: 3.79-5.08)higher in overweight and obese people than in people with normal BMI,respectively.The risk of NAFLD was increased by 22%(HR=1.22,95%CI: 1.10-1.35),44%(HR=1.44,95%CI: 1.23-1.69),and 21%(HR=1.21,95%CI: 1.08-1.36)in those with hypertension,diabetes,and hyperuricemia,respectively,than in those without.The risk of NAFLD was increased by 28%(HR=1.28,95%CI: 1.15-1.43)and24%(HR=1.24,95%CI: 1.09-1.40)in people with elevated ALT and GGT compared to the normal population.The risk of NAFLD was increased by 86%(HR=1.86,95%CI:1.69-2.05)in people with elevated TG compared to the normal population.3.Nomogram prediction model for the risk of NAFLD were constructed based on BMI,hypertension,diabetes,hyperuricemia,ALT and TG.The area under the ROC curves were 0.724,0.729 and 0.734 for 2-year,3-year and 4-year in the training set,respectively.The area under ROC curves were 0.722,0.743 and 0.763 for 2-year,3-year and 4-year in the validation set,respectively.The calibration curves showed good agreement in both training and validation sets.Decision curve analysis(DCA)showed the clinical utility of the nomogram for predicting the risk of NAFLD incidence.The risk of developing NAFLD was significantly higher in the high-risk group(519/2209,23.5%;score ≥95.57)than in the low-risk group(406/9182,4.4%;score ≤31.33)in different risk strata(P < 0.05).4.A total of 6,227 study participants with NAFLD at baseline were followed up for 15,012.70 person-years,during which a total of 2,076 NAFLD were regressions,with a regression rate of 138.28/1000 person-years.The regression rate of NAFLD was higher in the male population(145.24/1000 person-years)than in the female population(121.71/1000 person-years)(P<0.05).5.Multi-factor Cox regression analysis of NAFLD regression showed that the likelihood of NAFLD regression was 30% higher in those aged 18-44 years than in those aged 60 years and older(HR=1.30,95%CI: 1.14-1.49).The likelihood of NAFLD regression was 65% higher in men than in women(HR=1.65,95%CI: 1.44-1.88).The likelihood of NAFLD regression was 13% lower for those with junior high school or less education than in those with high school or more(HR=0.87,95%CI: 0.79-0.96).Non-tea drinkers were 22% more likely to have NAFLD regression than tea drinkers(HR=1.22,95%CI: 1.11-1.34).The likelihood of NAFLD regression was 96% higher in those with normal BMI than in those with obesity(HR=1.96,95%CI: 1.71-2.25),and21% higher in those with overweight than in those with obesity(HR=1.21,95%CI:1.07-1.37).NAFLD regression was 49% more likely in those with normal uric acid than in those with high uric acid(HR=1.49,95%CI: 1.33-1.66).NAFLD regression was 68%more likely in those with normal ALT than in those with abnormalities(HR=1.68,95%CI: 1.51-1.87).The likelihood of NAFLD regression was increased by 34%(HR=1.34,95%CI: 1.11-1.63),35%(HR=1.35,95%CI: 1.23-1.49)and 26%(HR=1.26,95%CI: 1.13-1.40)in the TC,TG and HDL-C normal population than in the abnormal population,respectively.6.NAFLD regression prediction nomogram were constructed based on gender,BMI,hyperuricemia,ALT and TG.The area under the ROC curve in the training set at2,3 and 4 years was 0.710,0.658 and 0.636,respectively.The area under the ROC curve in the validation set was 0.742,0.680 and 0.680 for the prediction model at 2,3and 4 years,respectively.The calibration curves showed good agreement in both the training and validation sets.The DCA showed the clinical utility of the nomogram in predicting the likelihood of NAFLD regression.The probability of NAFLD regression was significantly higher(P<0.05)in the high-scoring group(304/492,61.79%;score≥264.95)than in the low-scoring group(545/2340,23.29%;score ≤186.75).Conclusions: 1.The incidence of NAFLD in the Jinchang cohort population was moderate,and the incidence was higher in the male population than in the female population.2.Age 45-59 years,monthly household income ≥2000 yuan,overweight and obesity,hypertension,diabetes,hyperuricemia,elevated ALT,GGT and TG were risk factors for the development of NAFLD.Elevated DBIL was a protective factor for the development of NAFLD.3.The regression rate of NAFLD in the Jinchang cohort population was at a high level,and the regression rate of NAFLD in the male population was higher than that in the female population.4.Age 18-44 years,male,no tea consumption,normal BMI,normal uric acid z,normal ALT,TC,TG and HDL-C were favorable factors for NAFLD regression.Educational level of junior high school and below were unfavorable factors for NAFLD regression.5.The prediction models of NAFLD incidence and regression constructed in this study have good predictive ability and clinical utility.The risk stratification system can effectively distinguish between low,medium and high-risk populations,which is important for early identification of high-risk populations and appropriate adjustment of disease prevention and control measures. |
PDF Full Text Request