| Objective:(1)To investigate and analyze the prevalence trend of hyperuricemia(HUA)in part of the physical examination population in Guilin city from 2011 to 2017.(2)To analyze the risk factors for HUA and establish simple prediction models for HUA by using the nomogram model and CT model.To test the discrimination ability,accuracy and clinical practicability of the two models,so as to provide some help for the prevention and intervention of HUA.Methods :(1)Statistical analysis was performed on 213,562 patients(112,868 males,100,694females)with serum uric acid and other test results from the physical examination information system of the Physical Examination Department of the Affiliated Hospital of Guilin Medical College from January 2011 to December 2017 to analyze the prevalence trend of HUA.(2)From July to October 2017,10,286 cases(5,983 males and 4,303 females)were selected as the subjects for general clinical data collection,physical examination and laboratory examination to analyze the risk factors of HUA.And the above 10,286 samples were divided randomly into a training set(7,169 cases)and a validation set(3,117 cases)at a ratio of7:3.In the training set,we performed univariate and multivariate logistic regression analysis based on patients’ basic information and non-invasive examinations,including age,gender,Body Mass Index(BMI)and prevalence of hypertension,to establish a nomogram model.At the same time,the classification regression tree algorithm of decision tree model was used to construct a classification tree(CT)model.And the two models were verified and evaluated in the validation set.Receiver operating characteristic(ROC)curve,calibration curve and decision curve analysis(DCA)were used to test the discrimination ability,accuracy and clinical applicability of the two models.Results:(1)The total prevalence of HUA was 23.2%.After standardization by age and gender,the prevalence of HUA from 2011 to 2017 was 16.4%,20.5%,23.5%,23.7%,22.1%,25.7% and 25.5%,respectively,and the prevalence of HUA in 2016 and 2017 were significantly higher than that in2011(P<0.05).(2)From 2011 to 2017,the prevalence of HUA in males was significantly higher than that of females(P<0.05).What’s more,the prevalence of HUA in males was the highest in the age group of 30 to 49 years old,while in females increased gradually with age after 30 years.(3)Multiple non-conditional logistic regression analysis showed that gender(male),advanced age,high BMI,hypertension,high serum creatinine(Scr),high blood urea nitrogen(BUN),high triglyceride(TG),high low-density lipoprotein cholesterol(LDLc)were risk factors for HUA(P<0.05),while high high-density lipoprotein cholesterol(HDLc)was the protective factor(P<0.05).(4)Age,gender,BMI and prevalence of hypertension were all related to the occurrence of HUA.In terms of discrimination ability,the area under the ROC curve(AUC)of the nomogram model was 0.745 and 0.720 in the training and validation sets,respectively.The AUC of the CT model was 0.737 and 0.715 in the two sets,respectively,but were not statistically different.The nomogram model had a higher specificity(63.48%)while the CT model had a higher sensitivity(75.82%).The calibration curves and DCAs of the two models performed well on accuracy and clinical practicality,and remained stable on the calibration curves.Conclusions:(1)In recent years,the prevalence of HUA in some physical examination populations in Guilin City showed an overall upward trend.(2)Gender(male),advanced age,high BMI,hypertension,high Scr,high BUN,high TG and high LDLc were the risk factors for HUA,and high HDLc was the protective factor.(3)HUA can be predicted based on patient’s basic information and non-invasive examination,including age,gender,BMI and hypertension.In this study,the relevant parameters of the HUA simple prediction model established by the nomogram model and the CT model proved that these two models have certain practical value. |
PDF Full Text Request