Risk Factors Of Renal Injury In Gout Patients And Establishment Of Clinical Prediction Models

Background:Gout complicating renal injury is common,and renal injury can further impair the excretion of blood uric acid,they promote each other.Without interfering in time,it will gradually develop into end-stage renal disease,which can seriously reduce the patients'quality of life.The reasons of renal injury caused by gout include:direct effects of uric acid,combined effects of concurrent diseases,effects of improper living habits,etc.The interactions among various factors accelerate the progression of disease and lead to different complications together.It is necessary to fully understand the risk factors of kidney injury of patients with gout,predict the possibility of its occurrence,and strengthen the prevention and treatment of gout and kidney injury.These can help improve the patients'quality of life,reduce complications and mortality.Objective:To explore the risk factors of renal injury in gout patients and establish clinical prediction modelsMethods:Retrospectively analyze of 781 cases of patients with gout in the First Hospital of Jilin University from January 2013 to December 2016.These cases are divided into renal injury group and normal group according to Glomerular Filtration Rate(GFR),and the general data and laboratory indicators between the two groups are compared.Intergroup analysis are used on factors with statistically significant differences.Statistical analysis uses SPSS 25.0.The two sets of measurement data are firstly tested for the normality and the homogeneity of the variance.The independent sample t-test is applied in measurement data with the homogeneity and the homogeneity of the variance,and the t'-test is applied in measurement data with the normality but not the homogeneity of the variance.These results are expressed as mean±standard deviation(x±s).Mann-Whitney U test is applied in measurement data without conformity and homogeneity of variance,and the results are expressed as median(interquartile range).The comparison of the enumeration data uses thec~2test.Analysis of variance is applied in intergroup analysis.Factors with statistic different in single factor analysis are selected for multiple factor logistic forward stepwise regression analysis.P<0.05indicatesastatisticaldifference.Finally,logisticclinical prediction models are established,and 536 cases of patients with gout in the First Hospital of Jilin University from January 2017 to December2018 are set up as validation data set to test the efficacy of models.The comparison of diagnostic performance uses ROC analysis.Results:1.General information:Among the 781 cases of gout patients,759 are males,accounting for 97.2%.22 women,accounting for 2.8%.Age of onset is 13 to 88 years old,the average is 43.80±13.67 years old.Body mass index is 18.10 to 40.23,the average is 26.32±3.85.Course of gout is1 to 34 years,the average is 8.36±6.30 years.The highest value of blood uric acid is 431 to 1361?mmol/L,the average is 603.69±120.51?mmol/L.Past medical history:125 cases of diabetes,accounting for 16.0%.418cases of hypertension,accounting for 53.5%.266 cases of smoking,accounting for 34.1%.229 cases of alcohol abuse,accounting for 29.3%.Laboratory examination:189 cases of abnormal urine protein,accounting for 24.2%.512 cases of dyslipidemia,accounting for 65.6%.474 cases of renal injury,accounting for 60.7%.There is no significant difference in the validation data set between the above indicators.2.Correlations:(1)Single factor analysis showed that the proportion of female patients with renal injury is higher than that of normal group,and the difference is statistically significant(P<0.01).Weight,body mass index,albumin and hemoglobin levels in the renal injury group are lower than those in the normal group,and the difference is statistically significant(P<0.01).Age of onset,course of gout,highest value of blood uric acid,prevalence of hypertension,course of hypertension,proportion of alcohol abuse,proportion of abnormal urine protein,blood urea nitrogen levels are higher in the renal injury group than that in the normal group,the difference is statistically significant(P<0.01).There are no significant differences in height,prevalence of diabetes,smoking history,course of alcohol abuse,daily alcohol consumption,total cholesterol,triglyceride,high-density lipoprotein,and low-density lipoprotein between the two groups(P>0.05).(2)Intergroup analysis of age of onset,body mass index,hypertension and urine protein shows that the incidence of renal injury increased with the age of onset,and the hypertension groups of grade 1,2,and 3have higher incidence of renal injury than that in the non-hypertensive group.The incidence of renal injury in the urine protein positive groups±,+,++,+++are significantly higher than that in the urine protein negative group.The difference between body mass index groups has no statistical significance.(3)Multiple factor logistic regression analysis shows that age of onset,course of gout,history of blood uric acid,hypertension,abnormal urine protein,blood urea nitrogen levels are related to the occurrence of renal injury in gout patients,and the greater the age of the onset,the longer the course of gout,the higher the highest value of blood uric acid,and the higher the blood urea nitrogen level,the greater the risk of renal injury.There is a high risk of renal injury when patients with hypertension and abnormal urine protein.3.Model condition:The clinical prediction model established by the age of onset,the course of gout,the highest value of blood uric acid,hypertension,urine protein,and blood urea nitrogen has an AUC value of0.891(95%CI:0.869~0.913)in the establishment data set,the optimal threshold for predicting renal injury is 0.621.The AUC value of the model in the validation data set is 0.719,(95%CI:0.676~0.761),and the optimal threshold is 0.819.After adding gender and body mass index to the model,the establishment clinical prediction model has an AUC value of 0.893(95%CI:0.871~0.915)in the established data set,and the optimal threshold for predicting renal injury is 0.538,in the validation data set.The model has an AUC value of 0.720,(95%CI:0.678~0.763)and an optimal threshold of 0.845.Conclusions:1.For gout patients,the age of onset,the course of gout,the highest value of blood uric acid,hypertension,urine protein qualitative and blood urea nitrogen levels are risk factors for renal injury.Gout patients with greater age of onset,longer course of gout,higher the highest value of blood uric acid,higher blood urea nitrogen level,hypertension and abnormal urine protein,have greater risk of renal injury.2.The clinical prediction model based on the age of onset,the course of gout,the highest value of blood uric acid,hypertension,urine protein,blood urea nitrogen,gender,body mass index as predictors is valuable for predicting renal injury in gout patients.