Abstract: Objective To construct a risk prediction model for gout disease activity in male patients based on tumor markers and renal function indicators. 
Methods Clinical data of 118 male gout patients admitted to Xi′an Fifth Hospital from January 2021 to December 2023 were retrospectively analyzed. They were randomly divided into training set (n=94) and verification set (n=24) according to an 8∶2 ratio, and divided into two groups according to disease activity: acute stage (n=41) and remission stage (n=53). Among them, the training set had acute stage group (n=41) and remission stage group (n=53), and the verification set had acute stage group (n=10) and remission stage group (n=14). Tumor markers ［carbohydrate antigen 125 (CA125), carbohydrate antigen72-4 (CA72-4), prostate specific antigen (PSA), pro-gastrin-releasing peptide (proGRP)］ and renal function indexes［serum creatinine (Scr), cystatin C (Cys-C), blood urea nitrogen (BUN), blood β2 microglobulin (β2-MG), blood uric acid (UA)］ were detected in all patients. The levels of tumor markers, renal function indexes and other clinical data of the two groups were compared, and the independent factors affecting the disease activity of the patients were analyzed. The area under the receiver operating characteristic (ROC)curve (AUC) was used to analyze the predictive efficiency of the prediction model for the disease activity of patients. 
Results 〖JP2〗The levels of CA125 ［(14.25±2.85)kU/L vs. (12.67±2.53)kU/L］〖JP〗, CA72-4［(3.41±0.85)kU/L vs. (2.11±0.42)kU/L］, proGRP［(71.32±17.83)ng/L vs. (42.65±10.66)ng/L］, UA［(566.43±113.28)μmol/L vs. (372.71±74.54)μmol/L］, Cys-C ［(1.21±0.25) mg/L vs. (0.77±0.26）mg/L］, β2-MG［(3.41±1.05)mg/L vs. (1.86±0.62)mg/L］, erythrocyte distribution width (RDW) ［(13.84±1.54)% vs. (12.67±1.41)%］ and platelet-lymphocyte ratio (PLR) ［(165.24±33.05) vs. (148.43±29.69)］ in acute stage were higher than those in remission stage, and the differences were significant (P＜0.05). Multivariate stepwise Logistic regression analysis showed that CA72-4 level (OR=2.989, 95%CI: 1.164-7.673), proGRP level (OR=3.678, 95%CI: 1.571-8.610), Cys-C level (OR=3.162, 95%CI: 1.773-5.637), β2-MG level (OR=5.236, 95%CI: 2.621-10.458) and UA level (OR=4.543, 95%CI: 2.778-7.430) were independent risk factors for disease activity (P＜0.05). The C-index was 0.844 (95%CI: 0.759-0.929) in the nomogram constructed based on the above influencing factors, and the correction curve for predicting disease activity was close to the ideal curve (P＞0.05). ROC of the training set showed that the prediction sensitivity of the model was 87.80%, the specificity was 84.90%, and the AUC was 0.896 (P＜0.05). ROC of the validation set showed that the sensitivity was 85.40%, the specificity was 86.80%, and the AUC was 0.861 (P＜0.05). 
Conclusion The nomogram prediction model based on CA72-4, proGRP, UA, Cys-C and β2-MG levels has a good predictive value for identifying high-risk male gout patients with disease activity. 

Key words: gout, male, disease activity, tumor markers, renal function, risk model